playlist stringclasses 160
values | file_name stringlengths 9 102 | content stringlengths 29 329k |
|---|---|---|
MIT_6034_Artificial_Intelligence_Fall_2010 | MegaR6_Boosting.txt | The following content is provided under a Creative Commons license. Your support will help MIT OpenCourseWare continue to offer high quality educational resources for free. To make a donation or view additional materials from hundreds of MIT courses, visit MIT OpenCourseWare at ocw.mit.edu. PROFESSOR: Good morning, everyone. Today, we're going to talk about boosting. Boosting is pretty awesome, and it's not as hard as it might seem. It's actually pretty easy, as long as you do it right. So let's take a look at this boosting problem. Just like with ID trees, you may have noticed there's the ID tree problem where there's a graph, and all of the tests are x and y-axis tests on that graph. And then there's the ID tree problem where there are a lot of crazy different classifiers about characteristics that are discrete. And then all of the sort of ID tree stumps, if you will, are built out of those discrete qualities. This is an example of a boosting problem of the second type with a bunch of discrete qualities, like evil, emo, transforms, sparkly. And it has an numerical quality number of romantic interests. So it's one of basically the two kinds of boosting problems that you might see. And now, there is the two-axis Cartesian problem. A good example of that is the hours of sleep versus coffee problem, which I, for one, am planning on doing in tutorial on Monday so that you guys sort of get a sense of both types of problems. This one looks big. I think it looks-- I mean, it has ten different vampires or non-vampires to classify. It has a whole bunch of possible classifiers. But if you do it right, you can do it fast. So here's the prompt for the problem. Let's see. After graduating MIT, you get a job working for Van Helsing and Sommers, a famous vampire-hunting consulting agency. Gabriel Van Helsing, one of the two founders, once attended several 6.034 lectures as a guest, and he remembers Professor Winston's vampire identification tree lecture. He assigns you the task of creating a superior classifier for vampires by using boosting on the following data. So we've got the ID number, which we can use to just write things out in shorthand. We've got the name of several vampires and non-vampires. Then you see whether they're a vampire or not. That's whether their classification is a plus or minus for the quality of vampire. After that, there's a bunch of possible ways to classify whether they're a vampire or not. There's whether or not they're evil, whether or not they're emo, whether or not they transform, and whether or not they're sparkly, as well as the number of romantic interests that they have. So for instance, on the one hand, you have Dracula, who's evil, but he's not emo. He can transform into a bat or a cloud of mist. He does not sparkle, and he has five romantic interests-- those three vampire chicks at the beginning, Wilhelmina Murray, and Lucy Westenra. So on the other hand, you have Squall Leonhart, who's the protagonist of Final Fantasy VII, is extremely emo and doesn't have any of the other characteristics. And he's not a vampire. However, he's a nice counterexample for a possible rule that all emo people are vampires because he's very, very emo, and he's not a vampire. So how will we go about tackling this problem with boosting? Well, there's a whole bunch of different classifiers. And if you think this is all of them-- like evil, emo, transforms, sparkly. romantics interests, and true-- it's actually only half of them. The other half are the opposite versions, but we'll ignore them for now. So if you look at these, you can probably figure out what they mean-- evil equals yes means vampire is what we're saying here-- except maybe true. You might be saying, why is there one that just says true? The one that just says true says that everybody is a vampire. You might think, oh, that sucks. But it's not that bad since seven of the 10 samples are vampires. The key, crucial thing about boosting is that for any possible classifier, like classifying on the evil dimension-- which actually sounds like some kind of weird place that you'd go in a comic book. But classifying on the emo dimension or whatever, as long as it's not a 50-50 split of the data, you're guaranteed to be able to use it in some way for boosting. If there is a 50-50 split, it's like flipping a coin, so it's useless. Because if you had some other thing, like gender equals male or female-- and let's say that was 50-50. It's not. But let's say it was 50-50 between vampire and non-vampire-- it's a useless classifier because it would be just the same as flipping a coin. You'd get no information. Now, you might say, wait a minute. What about classifiers that get worse than 50-50? What about them? Might not they be even worse than a 50-50 classifier? I claim a classifier that gets less than 50-50 is still better than a classifier that gets exactly a 50-50 split. Is there a question? AUDIENCE: Yeah. In the ID tree example, somebody said you used 50-50 classifiers and played around. And after you already produced the elements per set, then you only used 50-50 classifiers [INAUDIBLE] per set. PROFESSOR: So the question is, in the ID tree example, you might use 50-50 classifiers in later rounds if, for instance, there's a 50-50 classifier except for that most of the things off of that side have been already removed. Let's say there's 20 data points, and there's a classifier that splits it 10 and 10. And it gets half plus half minus on both sides. But all the pluses from the right side have been removed by some other classifier. You might use it. That's true. But in boosting, you will never use something that's a 50-50 classifier. You never use something that has exactly a 50-50 chance of being correct. Because if it has a 50-50 chance of being correct, it's useless. And if it has a 50-50 chance of-- no, sorry. Let me specify again. You'll never use something that has a 50-50 chance of giving you the right answer given the weights. That's very, very important. And that may be what your question was getting at. As I'm about to show you and as Patrick told you in the lecture, in later rounds of boosting, you change the weights of each of the 10 data points. At first, you start with all weights being 1/10. The weights have add up to 1. In this case, you never, ever choose a classifier that gets five of them right and five of them wrong. In the later rounds, you'll never, ever choose a classifier that gets half of the weight wrong-- exactly half of the weight wrong. But half of the weight may not be half of the data points. So it's possible to choose a classifier that gets half of the data points wrong if it doesn't get half of the weight wrong. And that's similar to an ID tree when you've already gotten things right before. Because you'll see that the weight is going to go to the ones you got wrong. So I'm not saying that you should throw out right away anything that gets five of the points wrong. Hell, you shouldn't even throw out right away something that gets seven of the points wrong. It's possible-- possible-- that you can get seven of the points wrong, or getting less than half of the weight wrong if those other three points are really, really annoying to get right. And we'll see that later on. But for insight, at every step along the way, we're willing to choose any classifier that doesn't get 50-50. However, we want to choose the classifier that gets the most of the weight right. By most of the weight, at first, we mean most of the points right. Later, we will mean exactly what I said-- most of the weight. And if you don't understand that, it's sometimes hard to get it right away when Patrick just lectures through it, introduces a new concept. If you don't understand that, you'll see what I mean when we go through, all right? So my point I was making before is, what about things that get less than half of the weight right? Well, those are always OK because you can just flip them around, use their inverse, and that gets more than half of the weight right. It's sort of like-- yeah, it's sort of like my girlfriend always tells me that she is more than 50% likely to choose the wrong direction when you're trying to go between two places, which I'm kind of skeptical of. But I said, if that's really true, then we can just go wherever you didn't say to go, and we'll be more likely to go the right way. So you're actually really good at finding the place that we want to go. And then she's like, no, that won't work. Because then I'll know that you're going to do that, and I'll double say the wrong way. And then you'll go the wrong way again. But that not withstanding, you can see you can apply the same concept to boosting. And that's why underneath of this, I have all the opposite versions of all these tests. So what should we be doing to solve this problem more quickly? First, let's figure out which data points are misclassified by each of these classifiers. In other words, if we say all the evil things are vampires and all the non-evil things are not vampires, what do we get wrong? And if we do that for every classifier, then that'll make it faster later on. Because later on, we're going to go through classifiers, and we're going to have to add up the ones they got wrong. So this chart over here is going to be extremely useful. And on the test that this appeared in, they even made you fill it in to help yourself out. So let's see. If we said that all the evil equals yes are vampires and all the evil equals no are not vampires, then-- I'll do the first one for you. So we get all of the non-vampires correct because they are all evil equals no. But unfortunately, we get Angel, Edward Cullen, Saya Otonashi, and Lestat de Lioncourt wrong because they are vampires, and they're evil equals no. Apparently, Lestat is iffy. But I never read those books, and the Wikipedia article said that in the end, he wasn't that evil. So there we go. Evil equals yes misclassifies 2, 3, 4, and 5. All right, so let's try emo equals yes. I'll have someone else do it. So let's see if you guys got it. So if we say that all the emo people are vampires and all the non-emo people are not vampires, what do we get wrong? AUDIENCE: 1, 6, 7, 9. PROFESSOR: 1, 6, 7, 9-- that's exactly right, and fast. Good. We get 1, 6, 7, and 9 wrong. 1, 6, and 7 are wrong because they are not emo, but they're vampires. 9 is wrong because Squall is emo, and he is not a vampire. Good. OK, what if we said that exactly the transforming characters are vampires and the ones that do not transform are not vampires? Which ones will we get wrong? AUDIENCE: [INAUDIBLE]. PROFESSOR: Transforms is the next one over. AUDIENCE: All of the ones [INAUDIBLE]. PROFESSOR: So which ones would we get wrong if we said that said transforms yes were vampires and transforms no were not vampires? AUDIENCE: We'd get 8 wrong. PROFESSOR: We'd definitely 8 wrong because she's not a vampire. AUDIENCE: [INAUDIBLE]. PROFESSOR: Well, no. It's actually on there. AUDIENCE: And 3 and 4 we'd also get wrong. PROFESSOR: Yep. AUDIENCE: As well as 5. PROFESSOR: Yes, exactly. OK. Oh, man. You didn't see the chart. You were just like, hmmm. You saw the left. You just said, hm, which one of these are the transforming characters? OK, that's pretty hardcore. AUDIENCE: [LAUGHTER]. PROFESSOR: But yeah, 3, 4, 5, and 8. No, no, it's definitely given to you. That would be like the worst test ever for international students. Ah, if you don't know these ten characters as vampires, you lose. All right, so what about that sparkly equals yes is a vampire, and if it's not sparkly, it's not a vampire? This is guaranteed not to go well. What do you think it's going to get wrong? AUDIENCE: For sparkly? PROFESSOR: Yeah, sparkly equals yes are the only vampires. AUDIENCE: [INAUDIBLE] wrong. Angel's going to be wrong. Saya-- so 1, 2, 4, 5, 6, 7, and 8. PROFESSOR: And 8-- yes, that's right. It gets 1, 2, 4, 5, 6, 7, and 8 wrong. That's pretty awful. But dammit, it gets Edward Cullen right. And he's hard to get correct due to the fact that he's not very much like a vampire. He's more of a superhero who says he's a vampire. OK, so next, number of romantic interest greater than two. So if they have more than two romantic interest, they're a vampire. And otherwise, they're not a vampire. So which ones would that get wrong? Hm? AUDIENCE: 3 and 10. PROFESSOR: Just 3 and 10, that's right. Because Circe had Odysseus. She had Telemachus. Actually, she had that guy she turned into a woodpecker. She had that other guy who was a sea god who caused her to turn Scylla into the nine-headed thing, and probably at least one other person. So Circe it gets wrong. And it also gets Edward Cullen wrong because he only has one. So 3 and 10. You can tell I thought about this problem when I was writing it up. I wrote this one. All right, number of romantic interest greater than four. So it's a little bit different this time. Now you have to have at least four romantic interests-- or actually, greater than four, but there are none that are exactly four-- to be classified as a vampire. Which ones do you think it's going to get wrong? AUDIENCE: 3, 4, 10. PROFESSOR: Yup, it is going to get 3, 4, and 10 wrong. Because now, you run into the fact that Saya has that blond [INAUDIBLE] guy, Haji, and Kai. So the last of the positive ones-- because I claim I can derive all the negative ones if you guys give me the positive ones. The last of the positive ones is everybody's a vampire. Who does that get wrong? AUDIENCE: 8, 9, 10. PROFESSOR: Yes, OK. Now, I can derive all the negative ones from this without a sweat. Evil equals no-- well, it's 1, 6, 7, 8, 9, 10 without looking at the chart. Raise your hand if you see why it's 1, 6, 7, 8, 9, 10 without looking at the chart. Raise your hand if you don't. Nobody, OK-- wait. One hand. OK, I saw another one back there too later. They were just more tentative. OK, it's the complement of A because A is evil equals yes is a vampire. It gets 2, 3, 4, and 5 wrong. So therefore, evil equals no is a vampire is guaranteed to get all the opposite ones. AUDIENCE: Oh, we could have looked at that too? PROFESSOR: Yeah, we're looking here, but we're not looking at the big chart there. AUDIENCE: You can look at any? PROFESSOR: Oh, yeah. If you can't look at anything, then you're screwed-- unless not only are you as hardcore as this guy, but you've also memorized the numbers. All right, so emo equals no is going to be 2, 3, 4, 5, 8, 10. Transforms equals no is 1, 2, 6, 7, 9, 10. Sparkle equals no is 3, 9, 10. Romantic interest less than two is everything except 3 and 10-- 1, 2, 4, 5, 6, 7, 8, 9. 1, 2, 5, 6, 7, 8, 9. And then finally, everything but 8, 9, and 10, so 1, 2, 3, 4, 5, 6, 7. All right, so when we started off, we know what everything gets wrong. I then make a bold claim, because there are n of these, which is 14. I make the claim that there are only six that, in your wildest dreams, you would ever possibly even consider using ever. And the rest, you would never, ever use. Question? AUDIENCE: Yeah, I just have a question about the number of romantic interests. PROFESSOR: Yes. AUDIENCE: You negated it without an equals on either side. PROFESSOR: That's true. That works only because there are [INAUDIBLE] 2 or 4. But it should have been negated with a less than or equal. I'm copying off of the quiz. But yes. I noticed that this morning when I was putting myself through my pace. I'm like, there's not a less than or equal to. Wait a minute. And then, oh, wait. It doesn't have any 2's or 4's. Actually, I don't remember writing them all as 5's and 3's. It's possible that somebody else in the post-editing process changed them all to be about the same number, and then changed the less than or equal tos to be less confusing. It's possible I had Circe at 4, and there was an equal to somewhere. And they were like, forget it. Because I can't think of the fifth romantic interest for her. So yes, normally, you would have to negate it with an equal to sign, but there happened to not be any things that are equal to 4 or 2 here. So they get away with it this time. But it's good practice. So I'm claiming that in our wildest dreams, we'd only ever want to use six of these ever. And the other eight, forget it. So let's see. I will call on people at random to-- the first people obviously are getting it really easy-- to tell me which of these you think that you might ever want to use. Give me one you might ever want to use. AUDIENCE: E. PROFESSOR: Why, E? Of course. That's the best one. Yes, that's one that you might ever want to use. I'll circle the ones that you might ever want to use. E-- it only gets 3 and 10 wrong. That's amazing. It's like the best class of [INAUDIBLE]. OK, so give me another one that you might ever want to us. AUDIENCE: F. PROFESSOR: F. Let's see. F-- F is great. It only gets three wrong. Do people agree that you would ever want to use F? AUDIENCE: No. PROFESSOR: Everyone's saying no. Why not? AUDIENCE: It's like E, except worse. PROFESSOR: It's like E, except worse. It's guaranteed at every step, no matter what the weights are, to have a worse accuracy than E. It is definitely good. If E wasn't around, it would be one of our best classifiers of all. But actually, F is not one of the six. This is why I had them write on the test that there were six, because people might not have found all six. Because people who did figure out not to include F might not have figured out to include some of the ones you want to include. So-- AUDIENCE: I don't understand why you can't use F. PROFESSOR: Why you can't use F, OK. So we start off with 1/10 weight for all our data points. But let's say during our time of boosting that all ten of the data points have now different weights. So we'll call whatever the weight of 3 is, which you're going to get wrong-- you want to minimize the error, right? So that weight of 3, which goes into the error of the E, is x. The weight of 10 can be y. So if you're thinking of choosing 3, you know you're-- oh, sorry. If you're thinking of choosing E, your error is x plus y. AUDIENCE: Sure. PROFESSOR: If you're thinking of choosing F, your error is x plus y plus z, where z is the error of 4. And since you're never going to have a negative weight, x plus y plus z is always greater than x plus y. AUDIENCE: You can't choose something without the 3 and the 10 anymore because you already chose E. PROFESSOR: That's-- yes, you would probably choose something that didn't get the three and the ten wrong. But you would certainly never choose F ever because it's always worse than E. In fact, this is exactly the process that will allow you to find the correct six. And by "will," I mean "can." And by "can," I mean, let's see if you guys get it. Give me another one of the six that you might keep. AUDIENCE: K. PROFESSOR: K is the claim-- sparkly. K, I'm going to say will lose for the same reason as F. It's 3, 9, and 10. It's essentially similar to 3, 4, and 10. So-- oh, by the way, we should not be only going for the ones with the fewest incorrect. You need to be going for ones that do not have something that is strictly better. In this case, 3 and 10 wrong is strictly better than getting 3, 9, and 10 wrong. Question? AUDIENCE: Oh, I was going to say transform. PROFESSOR: You were going to say transform. You are going to be correct. Transforms is one of the ones we need, C. 3, 4, 5 and 8-- there's nothing down here that gets fewer than those wrong. There's nothing that gets us 3, 4, 5 wrong, for instance. Yeah, there's no way to get 3, 4 wrong without getting either 10 wrong, or 5 and 8 wrong. AUDIENCE: So why is the 8 like that? PROFESSOR: What? AUDIENCE: Why not G? PROFESSOR: Why not G? Why not? Why not G? Let's include G, too. AUDIENCE: Oh. We don't have to do it all? PROFESSOR: We need six. No, I just said give me any, and someone gave me the easiest one, E. Question? AUDIENCE: B PROFESSOR: Why not B? B looks great. I love B. Let's include B. Does someone else want to give another one that they want to include? AUDIENCE: A. PROFESSOR: A. Why not A? Sure. I mean, it's hard to see down here because there might be something better on the bottom. But yeah, there's not. So let's include A. Why not A? I love A. A's great. OK. So that is now five. There's one more that we need. It's by far the hardest one to find. Find me one more that there's nothing better than it. There's nothing that has a strict subset of the same ones wrong. AUDIENCE: Question. PROFESSOR: What? AUDIENCE: Sorry, can we quickly-- why would you choose A before we chose C? PROFESSOR: OK, why would you choose A before you've chosen C? Let's say 8 was a real problem for you. And you were just getting-- let's say that 3, 4 and 5, they weren't that bad. They weren't that bad. They weren't that bad. OK, you got them wrong here with transforms. You chose C. But sometime later, 8 was just by far your issue. All right? 3, 4, and 5, and 8-- 3, 4 and 5 are much smaller weights than 8. And then after you've got 3, 4, 5, and 8 wrong, 3, 4, and 5 were still not that bad. And 8 still was a high number. And then sometime later down the line, you're looking at things. And you're saying, you know what? I really don't want to get 8 wrong again, but I don't mind if I get 3, 4 and 5 wrong. Maybe I'll get it wrong with 2, which I've never gone wrong yet. Actually, none of the ones we've circled here get 2 wrong, so it's probably not that bad to get 2 wrong. So that's why-- because it doesn't get 8 wrong. If A was 2, 3, 4, 5, 8, you'd never take it. Do you see what I mean? Oh, did someone raise their hand? Did someone find it? AUDIENCE: I just have a question. You can use the same reasoning for choosing K, right? Because after E, we could have chosen A and said that 9 only a little different-- PROFESSOR: But it's strictly worse. AUDIENCE: Sorry, sorry, I meant [INAUDIBLE] E to 9 and 10, and then we could have chosen 3, 9, and 10, right? Because-- PROFESSOR: But we chose to use E. Because getting only 3, 10 wrong is better than getting 3, 9, and 10 wrong in any universe. You pick. You see what I mean? It might not be that much worse. It might be only a little bit worse to choose K, but it's always worse. So it-- question? AUDIENCE: [INAUDIBLE] the right one by reasoning [INAUDIBLE] we need one that doesn't have 3 in it. Because right now, [INAUDIBLE] get 3 wrong. PROFESSOR: OK, that's a pretty good insight. What are you thinking about? AUDIENCE: Well, I'm trying to justify D. PROFESSOR: You're trying to justify D. AUDIENCE: [INAUDIBLE]. PROFESSOR: D is huge. It gets more than half of them wrong. But you know what? It gets 3 right. You know what? It gets 10 right. And unlike our things that get 3 and 10 right, which is B, it also gets 9 right. D is the last classifier. You got it. It's hard to choose one that has seven of them wrong, but D is the last one you might pick. It turns out there's nothing better than this for classifying correctly those annoying data points of Edward Cullen and Squall, and also Circe, who's not that annoying, but she tends to be a problem when romance is concerned. So these are the six that we might use. We can now ignore the rest of them forever-- or at least until someone reuses this problem or something like that. But we can ignore everything except A, B, C, D, E, G. In fact, why did I even bring that up? All the ones we want are on the front. I'll bring it back down. Then I'll cross this off with reckless abandon. That even broke off a piece of my chalk. Now, these are the ones that we're actually thinking about using. There is a chart over here already prepared to do some boosting with these six classifiers, all right? So let's give it a try. Now, remember, in boosting, we always try to choose whichever classifier has the least errors. Is there a question? AUDIENCE: Sorry, yeah. Before we move on, can you say again a little more slowly what exactly we were looking for when we were choosing our classifiers, like something about the subset? PROFESSOR: You generally want to take classifiers. So I'll tell you what let's you cross off a classifier. That may be a better way to do it. You can cross off a classifier as useless if-- and by the way, this is only useful if you can do it faster than just wasting your time looking at all of them. Because if you can't cross off some of them as useless-- usually on the test, they won't make you. You can just waste your time and have 14 instead of six possibilities every step of the boosting. But take a look at this-- 1, 2, 3, 4, 5, 6, 7. Then see. Do you have anything here that has a strict subset of these wrong? Oh, look. 2, 3, 4, 5 is a strict subset. This can be crossed off. 1, 2, 5, 6, 7, 8, 9-- anything that's a strict subset? Yes, 1, 6, 7, 9. So it can be crossed off. 1, 2, 4, 5, 6, 7, 8, 9. Let's see. 1, 6, 7, 9 is a strict subset. 3, 9, 10. 3, 10 is a strict subset. 1, 6, 7, 9-- a strict subset. 3, 4, 5, 8 is a strict subset, as is 2, 3, 4, 5. 1, 6, 7, 9 is a strict subset. And up here, 3, 4, 10-- 3, 10 is a strict subset. But the others don't have one, even 1, 2, 4, 5, 6, 7, 8. In general, you want to keep them. You want to keep every classifier you might use. The only ones you'll never use are ones that there's something else that's just better always by having a strict subset of them wrong. Hopefully, that was more clear. It's tricky. Very few people realize-- we're brave enough to take sparkly even when it got seven things wrong. So let's start out some boosting. This wasn't boosting. This was setting yourself up. But it was setting yourself up with the knowledge of how boosting works. Less knowledge, less search. Now we only have to search six things. Ah, I mean more knowledge, less search, not less knowledge, less search. So we start off with all weights being equal. And since there's ten data points, all ten of the data points are weighted 1/10. OK, we're now weighting all of them equally. Since we're weighting all of them equally, when we want to find the classifier that gets the least error, we just want to find the one that gets the fewest points wrong. Which one is that? That's our friend E, the first one that people realized was a good one. So we're going to choose classifier E. What's our error? It's just the sum of the ones we get wrong. So what's our error this time? It's 1/5. We got points 3 and 10 wrong. They both have a weight of 1/10. 1/10 plus 1/10 is 1/5. So I'll put 1/5, and alpha. Alpha is sort of a vote that will be used at the very end to aggregate our classifier. Alpha is 1/2 natural log of 1 minus the error over the error. However, I have a little trick for you. It's not that impressive of a trick, but it's a little fun. So since error is 1/2-- sorry, not error-- alpha. Alpha is 1/2 natural log of 1 minus error over error. If error is 1/x, then alpha is 1/2 natural log of x minus 1. That just follows from the math. It's a little shortcut. If error is in the form of 1/x, then it's just 1/2 natural log of x minus 1, which means since this is in the form of 1/5, everyone, alpha is? 1/2 ln 4. OK, 1/2 ln 4. So now we come to the part in boosting that many people consider the hardest part, and I'm going to show you how to do it more easily. This is that part where we try to make all the ones we got right, we change their weights to be 1/2. And all of the ones we got wrong, we change their weights to be 1/2. Here is my automated process. It's called the numerator stays the same method. Here's how it works. Here's our ten data points. Their current weight is 1/10, all of them. We're about to re-weight them for the next step. So you agree they're all 1/10? They're equal, to start off. So step one-- erase the denominators. Screw fractions. I don't like them. There's division, multiplication. It's a pain. I just want to add whole numbers. That's what we're going to do. So which ones do we get wrong? 3 and 10. Circle those. All right. Add the numbers in the circles and multiply by 2. What does that give you? 4. That's the new denominator. AUDIENCE: Do you always multiply by 2, or just-- PROFESSOR: You always multiply by 2. Add the numbers not in the circles. Multiply by 2. What does that give you? AUDIENCE: 16. PROFESSOR: 16. That's the new denominator. The final, crucial step so that we can do this again next round is by far the most mathematically complicated thing here because we have to actually do something with fractions, but it's not too bad-- is we then change everything to be with the same denominator. So the 1/4's become? AUDIENCE: 4/16. PROFESSOR: 4/16. All right. I can also uncircle these for next-- ah. I hit the that button. All right. New weights-- 1/16, 1/16, 4/16, 1/16, 1/16, 1/16, 1/16, 1/16, 4/16. Note, the weights add up to 1. The ones you got wrong add up to 1/2. The ones you got right add up to 1/2. You're happy. So now that you get 4/16 of an error for getting 3 wrong, 4/16 of an error for getting 10 wrong, take a look at these six. I'm not going to call on someone, just whoever's good at math and can add these up more quickly-- just 3 and 10 count as 4. All the others count as 1. Add them up. Tell me which one's the lightest. What did you say? AUDIENCE: I'd go with B. PROFESSOR: You'd go with B. It doesn't get 3 wrong. That sounds pretty good to me. Does everyone else like B as well? I like it. I mean, all of our ones that don't get 3 wrong or 10 wrong, we're only looking at B and D. And D has seven. B has four. So B is the best. B gets 4/16 wrong. Does everyone see that? Because even getting one of 3 or 10 wrong is as bad as all the ones that B gets wrong because of the new weights. So cool. Let's choose B. That's right. And I sort of gave it away. What's the error that B has? It has four wrong, each of which are worth 1/16. The error is? What? 4/16, or 1/4, whichever is your favorite, which means that the alpha is? AUDIENCE: 1/2 ln 3. PROFESSOR: 1/2 ln 3. Bingo. Final round. OK, what did we get wrong? We got 1, 6, 7, and 9 wrong. Oh yeah, we can erase the denominators. All right. What are the numbers in the circles, summed up, multiplied by 2? AUDIENCE: 8. PROFESSOR: That's 8-- 1/8. And what about the numbers not in the circle, summed up, multiplied by 2? AUDIENCE: 24. PROFESSOR: That's right-- 24, which means I'm going to have to change all the numbers in the circle to 3/24, except I guess I don't because this is the last round. But if I was going to do another round-- let's prepare in case we where, change all of these to 3/24. Besides, it makes it easier to calculate which one is the best possible classifier because you can just use the numerators and sort of add them up. So while I'm writing that up, you guys figure out which one you like for classifier and call it out to me when I'm done. 3/24. 1/24. 4/24. 1/24. 1/24. 3/24. 3/24. 1/24. 3/24-- wait. I'm off by one here. 3, 1, 4-- AUDIENCE: It's because w1 is not assigned to anything. So w2 is really w1. PROFESSOR: Aha. You're right. w1 is not assigned to anything, so w2 is really w1 Yeah? AUDIENCE: There's an extra 1/16 between w1, w2. There's an extra 1/16. PROFESSOR: Yes, that's true. OK, well-- AUDIENCE: We get it. PROFESSOR: You get it. H, so what is the best H? You get it because it's right here. See? The process is so foolproof, even a fool like me can get it right while they have the chart wrong. All right, so what's the best classifier? AUDIENCE: C. PROFESSOR: You say C. I say that seems pretty reasonable. It only gets 3, 4, 5, and 8 wrong. Does anyone else get a different answer? AUDIENCE: A. PROFESSOR: Someone else gets A. I like A. Who said A? A lot of people said A. Well, let's figure it out. So A gets 1, 5, 6, 7. C gets 4, 5, 6, 7. They're, in fact, equal. Tie-break goes to the lower letter because that's what we said. In fact, I didn't tell you, but that's what we said. Question? AUDIENCE: So when we were deciding which classifier to use, can we only look at the weights, or do we also have to look at the [INAUDIBLE]-- PROFESSOR: Ignore all previous rounds. The question is, do you only look at the current weights when determining a classifier? Or do you look at the previous rounds as well? You have to ignore the previous rounds. Trust me. They will be used later in the vote. But it's sort of like tainting the jury a little bit to use the previous rounds when you're doing the current round. Because you want to start fresh with these new weights, get a new classifier. And then later, everyone will get to make their vote. So you only do it based on the current weights. AUDIENCE: We don't take any consideration if the last round's 6 was wrong or anything. PROFESSOR: Nope, although the weights take into consideration is when it's wrong, it's going to increase. AUDIENCE: OK. PROFESSOR: Question? AUDIENCE: Could you theoretically reuse a classifier? PROFESSOR: The question is, could you theoretically reuse a classifier? Answer-- you absolutely can. When that happens, it essentially gets extra weight because you used it again. But you can never, ever use it twice in a row. Here's why. Let's say that we want to use-- which was the one we used last over there? B? Let's say we wanted to use B again. What does it give us? 50-50. If we wanted to use B and then B-- 3, 6, 9, 12 wrong. Always guaranteed to give you 50-50, which is the only way that you can be sure you'll never use it. In fact, that's by design. You could reuse it, but not twice in a row. It could be used later on down the stream. And it will be used. Because if you do seven rounds, one of them has to be reused. It just gives more weight to whichever one is reused. But yes, A wins against C. C was a perfectly good answer as well. Question? AUDIENCE: Wait, can you reuse [INAUDIBLE]? PROFESSOR: What? AUDIENCE: Instead of A or C. PROFESSOR: OK. If you could reuse, why doesn't he pick E? E gets eight out of 24 wrong. It's one worse than A and C. That's the only reason. Next step will probably use A-- or sorry. Well, next step, we'll probably use E, frankly-- although maybe not, because we got 3 wrong on A. But pretty soon, we would use E again because E's pretty awesome. But anyway, here we used A. And we said we got 7/24 wrong. Oh, man, we can't use my little shortcut. So the answer, it has to be 17/7-- or 1/2 natural log of 17/7. So there we go. Now, we have to ask, what is the final classifier that we created from all these things? All we do is we sum up all the classifiers we chose. And we multiply them times their weight, alpha. So 1/2 ln 4 times E, whether or not E returns true, plus 1/2 ln 3 times B plus 1/2 ln 17/7 times A, is our final classifier, where E returns a plus 1 if E thinks it's a vampire, and a minus 1 if E think it's not. Same for B and A, all right? And then we take the sign of this. And I don't mean sine and cosine. I mean just, just is it positive or negative? OK? So the question now on the exam is, how many of the ten data points do you get right if we used this? Let's give it a look see. E is-- so we have romantic interest greater than 2. We have emo yes. And we have evil yes. So oh my gosh, logarithms, they're sometimes annoying. Do we have to actually add them up? I claim we don't. Here's a nice special case of having three logarithms on the board. One of two things is true. Either one of those three logarithms is so large that it's bigger than the other two combined, in which case, if that one returns a positive or a negative, it's just positive or negative because that one's big. Or one is not that large, and in which case, any two can dominate the other one, and so is just equivalent to a majority vote. So I claim we never have to add them when there's only three. You guys see what I mean? Like, let's say one of them was 1/2 log of a billion, and the others were 1/2 log of 3 and 1/2 log of 4. Obviously, whatever the 1/2 log of a billion says, which is multiplied by 1/2 log of a billion, is it's just going to be that, and the others will be ignored. However, if it's not the case that one of them is larger than the other two combined, then it's a simple vote between the three, because any two can out-vote the other one if they work together. And in this case, let's see, 17/7 is not quite 3. However, log of 4 is certainly not better than log of 3 plus log of 17/7. It's not even-- log of 4 is equal to log of 2 plus log of 2. And these are both bigger than log of 2. That's rules of logs-- log of 4 equals log of 2 squared, and you can take the 2 out. So these are not big enough that one of them is bigger than the other two combined. So it's just going to be a simple vote. So let's go through. Dracula-- OK, he's got tons of his little vampyrettes. He's not emo, so E gets it right. He's not emo. So that gets it wrong. But he is evil. That gets it right. Two out of three vote that he's a vampire-- correct. Next. Angel-- OK, well, he was in a long running series. He's got plenty of romantic interests. So that gets it right. He's certainly emo. That gets it right. And even though he's not evil, two out of three says he's a vampire, so correct. Next, Edward Cullen-- well, Twilight, here we come. Let's see. He only has one romantic interest, so that gets it wrong. OK. He's emo, so that gets it right. But he's not evil, so two wrong. So Edward's not a vampire according to our final classifier. But he is. So we got one of the data points wrong. You guys see that? Because two out of three of our classifiers here said that he was not a vampire. All right, let's see. Saya-- well, she has more than two romantic interests. And she's emo. So even though she's not evil, we get it right. OK? Let's see. Lestat-- he also has may love interests, is emo, and is not evil. So you get it right. OK, Bianca is evil with many love interests. Even though she's not emo, two out of three, you get it right. All right, Carmilla-- I'm going to call her Karnstein-- is basically exactly the same as Bianca with the number of romantic interests fixed the way it is. So she will always do the same thing that Bianca does. It's why 6 and 7 always travel together. So we get it right. Sailor Moon is supposed to be not a vampire. So her number of love interests say that she's not a vampire because she only has one. The fact that she's not evil and not emo says that actually, she's perfectly not a vampire. They all agree. And that's correct. Squall has only one love interest, Rinoa. And he's not evil, both of which of say he's not a vampire. But he is emo. But two out of three says he's not a vampire. We get it correct. And Circe, despite her many romantic interests which says she might be a vampire, is neither evil nor emo, and is not a vampire. So we got everything right except Edward Cullen, which perhaps says more about Stephenie Meyers writing than about our boosting. All right, final question-- Wesley Wyndham, a fellow consultant, has noticed a few correlations between some of the classifiers you used. He suggests using a new set of weak classifiers consisting of a pair of your classifiers that are logically anded and ored together. For instance, two of the new classifiers would be emo equals yes or evil equals yes, or sparkly equals no and transforms equals yes. So that would cut out Sailor Moon from the transforms cloud. He believes that you'll be able to classify large vampire data sets-- larger than this one, anyway-- more quickly with fewer rounds of boosting using his system. Do you agree or disagree with Wesley? Explain your arguments. So this was, you know, the tough concept question. Does anyone have just an instinctual thing other than like, oh, man, it's Wesley. He must be wrong. AUDIENCE: You'll probably use fewer rounds of boosting because you have more classifiers. But you'll have to search through more classifiers. PROFESSOR: Aha, that is the rare full-point answer. Very few people realize that Wesley was partially right. They either said something about him being completely wrong, which was wrong, or said that he was completely right. Yes, it will use fewer rounds of boosting because of the fact that you can-- essentially, one of these boosting already does is sort of gets things to vote together in an [INAUDIBLE] fashion. So it will use approximately half the number of rounds of boosting by being able to combine into two. But there's a lot of ands and ors. There's in fact n choose 2, where n is the number of vampires. And since using half the number of rounds but taking n choose 2 time for each round is not less time. So that's exactly correct. Not that many people got full credit on that one because sometimes, they were seduced by Wesley's idea. Or they were just like, it's Wesley. He's wrong-- or just some other funny answer. Any questions about boosting? Question? AUDIENCE: How do you know how many rounds of boosting to take? PROFESSOR: The question is, how do you know how many rounds of boosting to do? The answer is-- so on the quiz, it tells you that you have three. In real life, you might want to just kind of keep it running until it converges. That's one possibility-- keep it running until it converges to an answer, and it doesn't do anything anymore. Patrick has a little widget on the 6.034 website, I think, that lets you plunk down some data points and run boosting on them. And you can see that eventually, it converges. The boosting converges to an answer, and it doesn't change. AUDIENCE: What converges? PROFESSOR: Basically, the-- not the classifiers you picked, of course, or the weights. But what converges is which ones of your data set you get correct. Because he does his in two-dimensional space rather than like this. And he shows you the lines that boosting is drawing between classification, and colors things in green and red or something like that. And eventually, it converges where the lines are and which ones it's getting right. It generally converges to getting everything correct. And when that happens, then you can stop. But that's a good question. And it's not always that easy in the real world. You have to sometimes just say, this is enough for me. I've given it n number of rounds. And that's much more the number of classifiers, so maybe it won't get anything better. |
MIT_6034_Artificial_Intelligence_Fall_2010 | MegaR5_Support_Vector_Machines.txt | The following content is provided under a Creative Commons license. Your support will help MIT OpenCourseWare continue to offer high-quality educational resources for free. To make a donation or view additional materials from hundreds of MIT courses, visit MIT OpenCourseWare at ocw.mit.edu. PROFESSOR: Hope everyone had a great Veteran's Day break yesterday, spending it, of course, as I check my test audience, spending it mostly doing your [INAUDIBLE] research or PSETs. But at least one person got to watch TV. So at least one person got to have a real break, and that's something truly amazing and special. So now we're going to talk about SVMs. They're pretty much the hardest thing in 6.034. However, in recent years a few shortcuts have popped up that will sometimes allow you to solve the question, depending on what they're asking for, without solving some vast ugly set of equations with a vast ugly number of unknowns. So I'm going to show that to you guys. I'm also going to try to explain all the alphabet soup that's in [INAUDIBLE] and what all the letters stand for because it took me a few times going through SVMs. It took me a few times going through SVMs to actually find out for sure what all those letters stood for. And if you guys figure it out first try, that's going to be great, and you guys will be just fine. So let's take a look at the problem that's perhaps most optimized for using some of the shortcuts to solving it and not putting up all the equations. Then I will-- not because I'm sadistic, but because I'm being nice, I will force you with me to solve some of the things they didn't ask for us to solve so that you can see that we can't get away with everything without doing some of the harder stuff. And of course, definitely ask questions as always, but this time even more so. You guys, well, if you were looking around, you saw nobody in this entire lecture hall raised their hand that they are already set and ready and know SVMs. So if you have a question, maybe everybody else does. So let's go. We'll start right here. As always, pretend that I can draw, and that therefore all the pluses and minuses are only on integer coordinates. So we are asked in this problem to circle the support vectors. Draw the edges of the street and then the dotted line in the middle that separates them, the separator, as a dashed line. And then to give w and b. So what are w and b? Well, there's a few important equations in SVMs that we really hope-- and I'm going to tell you we're lucky in this because we don't have to-- but we really hope that we don't have to use because they provide a huge number of variables. So one of those crucial equations is that for a plus support vector, w vector dot x plus, the plus support vector, plus b equals 1. w dot x minus plus b equals minus 1. And w dot that dotted line-- I don't know, we'll call it dot dot dot-- plus b equals 0. So what does this mean? There are a lot of vectors. Well, I mean, we're usually in two-dimensional space, so we can basically just say that there's two components of this w vector, w1 and w2. And they're just two coefficients in a linear equation. So for instance, what we're interested in finding, this dot dot dot line, we'll just call that x, so with nothing on it. Actually, maybe that'll be easier. This is equivalent to saying w1x1 plus w2x2 plus b equals 0, where x1 is this, and x2 is this. We would possibly call them x and y. So one way to think about it is w1, we'll call it a, ax plus-- call w2 b-- by. Oh, don't call it b. Well, ax plus cy plus b equals-- I'll put this all in parentheses-- this is basically an equation like this. Or y equals negative a over c x minus b over c. It's basically y equals mx plus b. Does everyone see that? This thing that we're looking for, this w dot x plus b equals 0, is the equation of a line in Cartesian coordinates. It just looks uglier. Normally, when we're doing all this solving for w and b, we would have to put in tons of equations, plug in all of the support vectors in there. And we'd have to use these little devils called alphas. Alphas essentially-- if it wasn't clear in the lecture, which it usually isn't completely clear to everyone, wasn't clear to me completely-- alphas, the way I like to think about them, the alphas in this problem is they are the weight of how significant any particular point on the graph is towards creating the boundary. The higher the alpha is, the more that that point narrows in the boundary. The lower the alpha is, the less that that point narrows in the boundary, the wider the road can be. And if that point doesn't do anything, if that point is irrelevant and could be removed and it wouldn't affect the boundary, the alpha is? Everyone? AUDIENCE: Zero. PROFESSOR: Zero. Well, that was one person, but you can suffice for everyone. The alpha is 0. And that means if it's not a support vector, if it's not one of the vectors on the boundary lines, it will always have an alpha of 0 because it doesn't affect. So keeping that in mind, there's a few fun and important equations about alphas that we'll need if we're solving many equations for many unknowns, which hopefully we won't have to do. The sum over the positive alphas equals the sum over the alphas-- the negative points. And this is true over all the points. But since all of the alphas are 0, except for the support vectors, it also means the alphas of the positive support vectors are equal to the alphas of the negative support vectors. Additionally, our old buddy, the w vector, is equal to the sum over all i that are plus vectors of wi alpha i minus m over j minus vectors of wj alpha j. Now, all of these equations can be used in a bloody mess to figure out the answer to what we're trying to find, which is circles-- well, actually, they can't be used as circle support vectors and draw the dotted line. But once we do that, all these equations can be used in a bloody mess to give us the next thing that we want, which is w and b. So fortunately, there's another way to get w and b. If you guys really want, at the end of the hour we can also try to derive w and b using the many equations in many unknowns, but it's a bit painful. We'll try to do it the cool way. So let's start off. This is the one we're looking at. We need to find where the support vectors are. So the first thing we need to do is simply eye it. Fortunately, on the test, there will always be ones that you can eye if you're supposed to circle the support vectors. There's obviously some number of pluses and some number of minuses. I say obviously, but maybe not. But hopefully obviously, and we'll find out because I'm going to call on random people. So give me a positive support vector. AUDIENCE: Um, going to the one that looked like [INAUDIBLE]? PROFESSOR: Which plus sign, [INAUDIBLE]? AUDIENCE: One on the right. PROFESSOR: The one all the way on the right. Yeah, that plus sign is a positive support vector. That's good. All right? Excellent. Now, give me a negative support vector. That one? No? AUDIENCE: Yeah, sorry. PROFESSOR: Ah, no problem. Give me a negative support vector. AUDIENCE: I should definitely ask you, what's a support vector? [LAUGHTER] PROFESSOR: That is a good question. The question is, what is a support vector? How many other people will admit to having this question? See? You're not alone. OK. Before I go on, I'm going to assume-- you guys make sure I'm correct-- Monday was, just being sure so I can tailor based on this. Monday was the support vector machine lecture. But it was also very difficult to follow. That's what I usually expect. So what is a support vector? Well, all these pluses and minuses, if we were me, and if, I guess-- yeah, if we were me and if I was describing this problem, the one that we work out in class, I would call them points because they're on the graph. They're points. They're data points. But however, in more difficult versions of this problem that have n dimensions, where n is some ridiculous number of dimensions that you're never going to graph. Like say some of the research I'm doing now, I could use support vector machines on some of these articles that I'm reading about cyber events to try to figure out if there's a real event or if it's just someone complaining about how we're really vulnerable or something like that and no event actually happened. So the reason why they call these guys vectors is when you're not able to graph them on a Cartesian plane, there's still this long vector of many different dimensions. Right now, though, these points represent the vectors. This is very simple. It's easier to view them this way. But for instance, that plus at negative 1, 2 represents the fact that there is a vector going in the direction of negative 1, 2 with a magnitude such that it reaches negative 1, 2. So all these points are just a point representation of a vector. You probably, in any class that worked with vectors, saw this, saw vectors being represented as points. Question? AUDIENCE: Always from the respect to the origin? PROFESSOR: Yes. The question is always with respect to the origin. The answer is canonically, when vectors are represented as points, yes, it's always with respect to the origin. So that's the basic idea is that all these points are vectors. So what are support vectors? Well, you could call them support points for this case. But the reason we call them support vectors is again, in the generalized case that you might be doing in the real world with real AI, you're going to have a giant vector. And it's not just going to be points on a graph. Well, usually. So the support vectors, the support points, we found one of them correctly. It's this guy. They're going to be the ones that again, they don't have an alpha of zero. They're the ones that bind in the, as Petra calls it, the road, the boundary lines. They're going to be on the edge of plus. Whichever direction we draw it, this plus is the edge of the plus region. If we made this the edge of the plus region and everything on this side is plus and everything on this side is minus, we'd be screwed because there's two pluses on the other side of that. Generally, when trying to find a support vector, you do something a little bit similar to my crazy method of doing nearest neighbors, and try to find a plus-minus pair that's close to each other. Sometimes though, it's not just two points because sometimes if you try to draw the simple-minded thing, which is the perpendicular bisector of the two points, you get screwed because there's another point in your way. So now that I've given away a clue, let's go-- and hopefully that made sense to you guys. The support vectors are the ones on the edges that are just barely a plus for sure, or just barely a minus for sure. Let's go back. Can you give me a negative support vector? AUDIENCE: The top one? PROFESSOR: Hmm? AUDIENCE: The top negative point? PROFESSOR: The one on the top left? Yes. And does anyone think that there's a third support vector? Well, let's simple-mindedly try the thing that-- remember, support vectors always attempt to have the widest possible space between the pluses and minuses that they can. So let's simple-mindedly try to do the perpendicular bisector and see if screws us over. So when we simple-mindedly do the perpendicular bisector, it goes through here like this. And it's just fine. So these are the only two support vectors. And there's our divider line. So we're on the home stretch. But we have to find w and b. In olden days, we would find w and b by plugging in w dot the plus support vector, plus b equals 1. Oh, that's very crucial. These w dot x plus x minus are only true equaling 1 or negative 1? Or only true for support vectors? It's always true that w dot any positive point plus b will be some positive number. But it won't always be 1. In fact, it will always be greater than 1 up over here. It will always be less than -1 down over there. In olden days, we would plug in -1, 2 into this equation. We would plug in 3, -2 into this equation. We'd plug in alpha plus equals alpha minus in its sums. And since there's only 1 plus 1 minus, we'd know they were equal. And then we'd fidget around with this w equation. However, there is a better way to do it. And so let's use this cheap strategy to solve this version of the SVM. Here's how. First, and I know I didn't draw these completely straight. Sorry. But can anyone, by looking at-- this is three, -2. 2 And this is -1, 2. Can anyone tell me what the equation-- you can do y equals mx plus b. Can anyone tell me what the equation of the dotted line is supposed to be if I was good at drawing? AUDIENCE: [SEVERAL ANSWERS] PROFESSOR: People say y equals x minus 1. And I say yes, y equals x minus 1. So therefore, the pluses would be y is greater than or equal to x minus 1 indeed. So we've already seen that w dot x plus b somehow can be converted into this form. Right? So therefore, if we have y equals x minus 1, then we know that we have we have w dot x plus b equals 0. Let's do that here. So we know that w1 x1. We can even call it x and y. I think it'll be fine. No one will come after us. w2 y plus b equals 0. But we also know that y equals x minus 1, which means that if y equals x minus 1, then according to this thing we have over here, then negative w1 over w2 equals-- So we know that negative w1 over w2-- and we have -b over w2. So y equals x minus 1. And if we solve this equation to make it look like this, we would have y equals negative w1 over w2 Minus b over w2. So we know that in some way, shape, or form-- we know that then therefore, w1 over w2 is some scalar multiple of minus 1. And we know that b over w2 is, in fact, some scalar multiple of positive 1. Scalar multiple, what's a scalar multiple? Well, why is it a scalar multiple? Why isn't it just going to be negative 1 or positive 1? Just because in this equation, we can multiply the entire equation by any number and it will still have the same boundary line. You guys see that? Oh, there's an x here. If we multiplied everything, since it's all divided by w2. If we double w2, but also doubled b and w1, it would be the exact same equation. Do you guys agree? So there's, in fact, infinitely many possible equations. You say, well, great, Mark. You've figured out what form it is. So you figured out that w1 over w2 equals some scalar multiple of negative 1. So it's negative 1 times-- what's everyone's favorite letter? AUDIENCE: k. PROFESSOR: k. Negative 1 times k. And we figured out that b over w2 is-- I guess we can just do negative -- is positive k. But what's k? How are we going to figure it out? Well, it's a good question. And I will tell you how. I will assert the following fact as true without proof. Then I will not prove it. 1 over the magnitude of w, which is this vector here with w1 and w2, equals this where this is that line that I just drew, the line from here to this point. 1 over the magnitude of w equals this. Therefore, since 1 over the magnitude of w equals this, and this equals, I believe, 2 root 2, because we're going over to, down to, Pythagorean Theorem, 2, root 2. So therefore, flip everything over. Magnitude of w equals 1 over 2 root 2. So therefore, magnitude of w equals root 2 over 4. But why are we OK? Well, how do we calculate the magnitude of w? Do people know, in general, magnitudes of vectors? Generally, for these vectors, we do it by the square root of the sum of the components squared. So the square root of w1 squared plus w2 squared equals root 2 over 4. But that's not all. That's not all, we say, because we know from this over here that the ratio of w1 and w2 is-- AUDIENCE: [SEVERAL ANSWERS] PROFESSOR: Yeah, the ratio of w1 and w2 is going to be-- actually, sorry. I shouldn't put a k here. I realize I probably have been confusing you guys a lot. w1 over w2 is just -1. B Over w2 is just 1. That's just a fact. There's no k. The k is to determine what w1 and w2 are. So w1 equals -k. And w2 equals positive k. And b equals also positive k. By the way, here's a question for you. Could I have put the negative sign on w2 and b instead of on w1? So many people said yes. That's a very smart answer. Actually, no, because of the fact that the pluses are on the negative x-axis. It's just a little trick I picked up. When one of them is negative and the other one isn't, follow the pluses. So we know that w1 is -k, w2 is positive k, and b Is positive k. w1 over w2 is -1. b over w2 is positive 1. So what do we know about the ratio of w1 and w2? It's equal to -1. And that means that when we square it, w1 squared equals w2 squared. So therefore, this is the square root of 2 w1 squared, which equals root 2 w1. Well, actually no, it doesn't equal root 2 w1 because w1 is actually negative. So it's negative root 2 w1. It doesn't matter. The point is that if that equals root 2 over 4, then w1 is-- everyone? AUDIENCE: Negative 1/4. PROFESSOR: Negative 1/4. Bingo. And if w1 is -1/4, 1 everything else falls into place. What are w2 and b? Everyone? AUDIENCE: Positive 1/4. PROFESSOR: Positive 1/4. We got it. We're done with this part of problem. However, bonus. Let's come to the alphas, which they didn't ask you to calculate. Actually, you know what? We'll do the alphas if we have enough time, since they didn't actually ask you to calculate them. However, my recommendation is since there's only one alpha-plus and one alpha-minus, they must be equal from this equation, since the sum of the alpha-plus equals the sum of alpha-minus. And so therefore, w equals the sum of w-- sorry, this should be an x. Of course, there's not a million w's in this equation. The sum of the positive data points times their alpha is minus the negative data points times their alphas. So we're looking at here -1/4, 1/4 equals-- what do we got here? Positive point negative 1, 2? So we've got alpha, alpha of that point negative 1, 2 minus alpha of that minus point. And what is that? It's 3, -2. 3, -2. So if both of the alphas which are equal were 1, we'd have -4, 4. But we want -1/4, 1/4. So actually both of the alphas are 1/16. And that's the answer. We'll do that more in depth if we have time. But we won't. So let's do number two. So let's go into faster mode. Number two, very similar to number one in many ways. But as you can see, one of the main things that they added an extra minus sign at 2, -1. So I think we can all agree that this will still be our plus-- Actually, they added another plus sign there, too. So maybe this plus sign is a support vector. But it's not. This plus sign is a support vector. What do you guys think about the new negative sign? Will it become a support vector since it is strictly closer to the pluses? Yep, you're right. OK, so this is a very beautiful division because if I do this correctly, which I didn't, but if we pretend that I did. Then the dotted line is-- AUDIENCE: [INAUDIBLE]. PROFESSOR: y equals x. OK, so with the dotted line at y equals x, then just like we did up here, we know that if y equals x plus 0, we know that first of all, b equals 0. Second of all, we know that if y equals x, then we know that -w1 over w2 equals 1. The pluses are still on the left and up, so we know that w1 is some negative number, -k, and w2 is some positive number k. Great. How are we going to figure it out? Well, let's call this d for distance, or whatever you want to call it. So 1 over w equals d. d In this case is not 2 over 2. Can everyone tell what d is here? AUDIENCE: [INAUDIBLE] PROFESSOR: It's actually-- so it goes over 2 and 1. So it should be 1 1/2 root 2 since this width distance, which is twice as much, goes over 3 and 3, which is 3 root 2. So it's 1 1/2 root 2. I don't like putting in decimals and stuff there. So we'll say that 2 over magnitude of w equals 2d equals 3 root 2. So therefore, right, Pythagorean, one, two, three, one, two, three, 3 root 2. So therefore, magnitude of w equals-- let's see. Switch them over. We should get root 2 over 3. And if magnitude of w is root 2 over 3, we can do our same trick from before, square root of 2 w2 squared equals root 2 over 3. And this is just root 2 times w2 equals root 2 over 3. So therefore, w2 is? 1/3. And w1 is? -1/3. Bingo. We've got w1. We've got w2. We know that b was zero because obviously it's 0. It's y equals x. And we're done. That was fast. The alphas might taken longer. Actually, the alphas on this one are more of a pain in the ass than anywhere else because let's take a look at this one if you can see it. We've added in yet some new points. We've got this point up here and this point down there. So I think pretty clearly this plus and minus are the closest to each other. But what happens if we take the perpendicular bisector between these two and do like this? This plus is in the middle. So therefore, this plus is going to have to also be a support vector. So we can't just draw this line. We have to include this. What's our best division? Vertical lines, that's right. Vertical lines just so. And that means that the equation of our boundary, the dotted line here, y-axis. So the equation of the boundary with the y-axis, then b equals 0. And hell, w2 equals 0. So the only thing that is not-- AUDIENCE: Wait, w1 [INAUDIBLE]. PROFESSOR: w2 equals 0. So w2 equals 0. b equals 0. But w1 is not equal to zero because it's just the equation of the y-axis. So we therefore know that the equation is just w1 times x equals 0. So it's w1 times x equals 0. And we know that that just means essentially that x is going to be some k. It's also going to be negative because of the fact that the pluses are still on the left. Then we're going to have to figure out what that k is. We'll use our old trick-- by this point old, hopefully. One over magnitude of w equals d. This time d is just 1. So therefore, magnitude of w equals 1. There's only one component in w. So therefore w1 is AUDIENCE: -1. PROFESSOR: -1 because the plus is on the left. Do people see that? Not too bad. This one's easy to calculate the w. But it's not as easy to get all the alphas. But let's move on to a new and even more fun-- maybe not-- question, which is this guy. As you can see-- well maybe not. This is a one dimensional vector. These vectors only have a single dimension. So it just looks like a number line here. That dimension varies from it looks like -9 to positive 9. It just has one component. You don't have to worry about any of these crazy magnitudes with two components, just everything as a single component. However, it's obvious that a linear basis line is going to completely screw us up here, since lines at this point are just like, grunk, all these are pluses. All these are minuses. Well, great, that doesn't get them all. So how are we going to do it? Well, we're going to use what is usually perhaps the hardest thing in SVMs, but in this case is not going to be too bad for us. We're going to use a kernel. Now, based on how little I understood kernels the first time I took this class, I'm guessing that you guys would like to have some explanation on these kernels. You probably saw them. You remember the kernels from Patrick's lecture vaguely? There's this phi. And then there's this k. And then they get really complicated. OK, so here's how the kernel works. The basic idea is this. And I'll write it over here. Oh, wow, there's more stuff. OK. I'll write it right here. The basic idea is this. We're taking the normal space, which is this number line or it could be any kind of normal space, and we're going to take a vector, we're going to put into it a function called phi. And phi of vector x brings x into some new dimension. Phi, or "phee" if you like it better, is usually a nasty piece of work and something you never, ever want to look at. Sometimes it's not too bad. Phi is the function that brings it into the new dimension. OK? And when you brought the data into a new dimension, sometimes you can just cut a straight line in that dimension and you'll just be happy. However, something that was noted by the very, very smart inventor of support vector machines is that you don't actually need to work with function phi, even if phi is an absolutely horrible monstrosity, because of the fact that you never need to know what all these vectors x actually are in the new space, at least not directly. In none of these equations up here do we ever use x by itself. However, we do use x being dot product with something else. So he figured out a very sneaky and excellent shortcut. OK, so-- oh, I shouldn't use x1 and x2. I'll use x and z. So if you have two vectors, x and z, which are in a regular space, you put them into this function called the kernel. Then it will tell you phi x dotted with phi z. And if you have that, you don't need phi. Does everyone see that? Does everyone see why we don't need phi? Look at all these equations up here. We have never looked at x by itself in these vector equations at least. Now, calculating alphas, yeah, that gets a little bit fuzzy. Also, you may ask, why would you do this? You can't calculate the alphas. It turns out that actually, other than for these very simple linear problems, human minds cannot calculate the alphas. In fact, you run a very complicated quadratic optimization. In fact, finding out the best alphas is the thing that you hill climb on when you're doing SVMs in the real world. You say, all right, I'll run my algorithm when I know there's only one peak, which is very, very good because it's quadratic optimization. Let me figure out the alphas. So in fact, it doesn't matter that you can't use these alpha equations to figure out the alphas if you only know the kernel function and not the phi function because normally, the computer figures out the alphas for you with quadratic optimization. Just in these simple problems, we know you can calculate the alphas. So we have the kernel, which basically gives us the dot product of the things in the new space. So being that as it may, I'll give you the kernel here. I'd like you give me phi. Someone got an idea, whose name was Susan Q. Random Student, apparently. She got an idea that if we had a kernel for x and z-- actually, they're not vectors, I guess. There just single components. And the kernel equals cosine Pi over 4x times cosine Pi over 4 z plus sine Pi over 4x plus sine Pi over 4z. So that is the new dot product. Oh, wait, sorry. I put one of the z's not inside the parentheses. That was silly of me. So cosine of the quantity pi Over 4x times cosine of the quantity Pi over 4z plus sine of quantity Pi over 4x plus sine of quantity Pi over 4z is the new product. So that begs the question. This is an easy one so we can calculate the phi. What is phi of x? We're actually taking it from one dimension and we may be playing around with it a lot to get this. And this thing has become a new dot product. It replaces dot product. And remember, the dot product for scalars would have just been multiplying two numbers together. So it actually makes it a little bit more complicated. Does anyone think they know the phi? Oh, we got one. What do you think? AUDIENCE: [INAUDIBLE] PROFESSOR: You mean two common vectors, two dimensional? AUDIENCE: The two points. PROFESSOR: Absolutely. That's exactly correct. How would you have solved this on the actual quiz if you're not our brave volunteer? Well, that k, if you squint at it-- not very much actually-- is pretty much a dot product between cosine of Pi over 4 and sine Pi over 4. I mean, look at it. Remember, if the dot product of x and z vectors is x1, z1 plus x2, z2-- so that basically is x1, z1 plus x2 z2. Oh, this should have been a times. Yeah, this should have been a times. Sorry. There's a plus there. Anyone who missed it because of that, my bad. That's should have been a times. That should have been a times up there. It's cosine Pi over 4x cosine Pi over 4z plus sine Pi over 4x times sine Pi over 4z. So yeah, it's basically the dot product between cosine Pi over 4x and sine Pi over 4x. Bingo. All right, last thing. Well, we're not done yet because we're going to maybe ask some questions. And then we're going to see if we can calculate those alphas. But last thing, let's graph in this new dimension all the points. So obviously, cosines and sines, so we're going to get results between 1 and -1. Let's see. Maybe I can graph it-- did I write on all these? Wait, maybe this one. No, people drew weird stick figures there. OK. Oh, yeah, this one's kind of messy. But we'll do it on this. OK, so this is 1, -1, 1/2, -1/2, 1, -1, -1/2, 1/2. OK? So given that, let's try to graph all these points on this number line into this brave new dimension by using their cosine times Pi over 4. So all right, great. So let's do the pluses first. The plus at 0 is cosine 0, sine 0. So what is that? AUDIENCE: 1, 0. PROFESSOR: That's 1, 0. That's right. In fact, the 8 and the -8 are also that times 3. The 8 and the -8 are also that because then it's just 2 Pi minus 2 Pi, which both cosine and sine are periodic. OK, great. What about the 1? Well, that's cosine Pi over 4, sine Pi over 4. And what's that? AUDIENCE: [SEVERAL ANSWERS] PROFESSOR: Yeah, it's root 2 over 2, root 2 over 2. So that's something like here, we'll say. And in fact, and the 9 and the -7 are also that. So there's three of these two. What about the -1? That's cosine negative pi over 4, sine Pi over 4. AUDIENCE: [SEVERAL ANSWERS] PROFESSOR: That's right. The x value is positive root 2 over 2. And the y value is negative. And again, there's three of them. All right, great. Now let's do the minuses. There's the minus at 3, which is also the same as the minus at -7. The minus at 3 is cosine 3 Pi over 4, sine 3 Pi over 4. Which one is that? AUDIENCE: [INAUDIBLE]. PROFESSOR: Yeah, that's going to be in the second quadrant. The cosine is going to be negative. But the sine is going to be positive. And so we get 3 points here. And as you may have predicted, the other one, the 5 Pi over 4, is in the third quadrant. We get 3 points here, Where are the support vectors? AUDIENCE: Question. PROFESSOR: Question? AUDIENCE: I understand where you're getting at the three quantities of pluses in the first of the quadrants. But according to the [INAUDIBLE] line there, you want that the total of four-- maybe the values. PROFESSOR: Oh, you're right. There's only two. Good call. There's two negatives here. And there's two negatives here. Good call. It doesn't change the problem. In fact, if we just graph more points, there might have been more. But that's a very subtle and important distinction. There are two negatives. But otherwise, yeah, these are graphed correctly. Does anyone see where the support vectors are? AUDIENCE: The first two. AUDIENCE: Maybe the top two. PROFESSOR:So the top two, the minus and plus. We'll try to do the perpendicular bisector. Let's see it. That works. But guess what? These guys are on the same line. So we'd better circle them. So actually, the question is what isn't a support vector? Only this. Only those three. Question? AUDIENCE: Couldn't you have just done this in one dimension? I mean, you just showed that those ran on the same lines. So you really didn't need the cosine term and a sine term. You could have proved all this with just the cosine. PROFESSOR: All right, the question is couldn't we have done this in one dimension. All we do is only the cosine. So if we did only the cosine, then they would've still been easily divisible. The answer is, absolutely, we could have. However, the question did not because Susan Q. Random Student decided to do cosine and sine. But yes, if we had said you, [INAUDIBLE] student, find a phi that will work for this, you could have found a phi that was just cosine. That would have been easier. However, it's important to be a little work with what somebody else gives you. In this case, they gave you that transformation, which yeah, was wasteful with an extra dimension. You didn't need the sine because you didn't need the y-axis really here. You just needed the x. Does everyone see this, how this works? You can maybe transform dimensions? The main hardest part is they'll usually give you a k and ask for a phi or give you a phi and ask for a k. But it's not too bad. Just remember, if they give you a phi, do a dot product with it. And if they give you a phi that's just one component, dot product of one component, just multiply them together. Easy enough. If they give you a k, treat it as a dot product and try to reverse engineer. It's usually something like this that's easy to reverse engineer. I really haven't seen it where it's not. So it often looks like the scariest problem. But it's usually not too bad to go between phis and k's. Does anyone have any questions on anything that we did on support vector machines? Question. AUDIENCE: So what's the intuition behind the w? We solved it and figured out numbers and integrations with it. But what is it in relation to-- PROFESSOR: Question is, what is intuition? What is w? W Is the dividing line. It is the drop dead dividing line. When I say the drop dead dividing line, you like those big, bold solid lines over there. Those are your pretty certain lines. Everything past that was a minus. In your training, it's that everything past the big bold line there was a plus in your training stuff. But the dotted line is the one you're really going to use in the test data. In the test data, when push comes to shove, you might get something if it's inside of that gutter. And if it's on the, say, of that one up there, if it's on the upper left side of that dotted line, you're going to call it a plus. So that dotted line is your decision boundary. And that is basically the idea. And in fact, the way that the algorithm would do it on the computer is it would quadratically optimize the alphas, which messes around with the dotted line. And by quadratically maximizing the alphas-- you see how the alphas add up to a w. It just checks it around. And eventually, it finds, oh, making the alpha of this one 0 makes it a better optimization. You're trying to get the widest possible road. It would eventually come out to this. This is trivial for a human to eyeball. But some real problems with 200 data points that have to get one or two of them wrong, classified, and you may be using a quadratic kernel or something that, you can't do that. You just can't. Well, maybe can, in which case you should be getting a MacArthur Fellowship or something like that. But the computer can. And the basic idea is when it comes down to it, it figures out the alphas, that the best w for the widest road. And the w s your decision boundary. Good question. Any other questions about our old friend, SVM? I have a question for you. After seeing this-- and let's pretend that they only asked to solve for w's, b's, these kind of kernels and phi, which are the typical things-- how many people now think that you can go through and work an SVM problem? All right, we've got a few. We've got a happy few. Band of brothers. Maybe eight people raised their hand there. That's good. How many people know what a support vector is now? That's really good. Because if that's all you learned from today's recitation, it's still good. It really is. I'm telling you. I had to take two classes on this and then TA it before I really, really understood it. So you guys are ahead of me. All right, take care. Have a great weekend. And we'll see you for boosting and vampires next week. |
MIT_6034_Artificial_Intelligence_Fall_2010 | MegaR3_Games_Minimax_AlphaBeta.txt | The following content is provided under a Creative Commons license. Your support will help MIT OpenCourseWare continue to offer high quality educational resources for free. To make a donation or view additional materials from hundreds of MIT courses, visit MIT OpenCourseWare at ocw.mit.edu. PROFESSOR: Today we're going to be talking about games. And I know you guys as well as I hope I do. The main thing that you guys want to talk about with games is how to do that alpha-beta thing. Because it's pretty confusing. And it's easy to get lost in a corner or something. Whereas doing the regular minimax, in my experience, most 6034 students can do that. And they do it right pretty much all the time. However, we're going to focus on all the different components of games. And I put up two provocative silver star ideas up on the board, which will come into play here. The Snow White principle is a new name. And it has never been revealed until today. Because I made up name recently. So you will be the first people to hear it and decide if it works better than the term "grandfather clause" for the thing that I'm trying to describe. Because most grandfathers don't eat their children. So here we've got a beautiful game tree. It has nodes from A through R. This is our standard game tree from 6034. We've got a maximizer up at the top who's trying to get the highest score possible. The minimizer is her opponent. And the minimizer is trying to get to the lowest score possible. And it's really unclear who wins or loses at each point. They're just trying to get it to the highest or the lowest score. All right, so let's do a refresher. Hopefully the quiz didn't put people into such panic modes that they forgot Monday's lecture. So let's make sure that we can do regular minimax algorithm on this tree and figure out the minimax value at A. So let's see how that works. All right, as you guys remember, the game search when using regular minimax is essentially a depth first search. And at each level, it chooses between all of the children whichever value that the parent wants. So here after it would choose the maximum of K and L, for instance. But that's getting ahead of ourselves. Because it's a depth first search. So we best start at the top. I'll help you guys up for a while. So we're doing A. We need the maximum of B, C, D, depth first search. We go to B. We're looking for the minimum of E and F. So having looked at E, our current minimum of E and F is just 2 for the moment. So this is going to be less than or equal to 2. All right, then we go down to F, which is a maximizer. And its children are K and L. So now I'm going to start making you guys do stuff. So what do you think? What is going to be the minimax value at F? The minimax value at F, what will that be? AUDIENCE: [INAUDIBLE] PROFESSOR: So that level is a maximizer-- max, min, max. F is a maximizer. K and L themselves are minimizers. But they're pretty impotent minimizers. Because they don't get to choose. They just have to do K or L. So the minimax value is three. And yeah, the path it would like to go is K. So we'll say that the minimax value here is 3. It's in fact exactly equal to 3. So if this is 3 and this is 2, then everyone, we know that the value of B is? AUDIENCE: [INAUDIBLE] PROFESSOR: I hear 3 and 2. Which one is it? AUDIENCE: 2. PROFESSOR: 2, that's right. So the value here is 2. Great, let's go down into this branch. So C is going to be the minimum of G and 6. But we don't see that yet. Because we're doing a depth first search. It's going to be the minimum of G. Now we need the maximum of M and N. We're going to need the minimum. M is the minimum of Q and R. So let's switch sides. The minimum of Q and R is? AUDIENCE: 1. PROFESSOR: Let's see. That's right, it's 1. So M has a value of 1. But I'm going to stay over here. Because M has a value of 1. Knowing that, then we know that G has a value of? AUDIENCE: 7 PROFESSOR: That's right. 7 is higher than 1. And since G is a 7, we now know going up to C that C has a value of? AUDIENCE: 6. PROFESSOR: Yes, C has a value of 6. That's the minimum 6 and 7. So now I'm going to go back down. Because we've done one of the other sub-trees. This is a 6. All right, great. Now we're going to go down to D. Hopefully it won't be too bad. These things usually aren't terrible. Because they're made to be pruned a lot in alpha-beta. So let's see, in D, we go down to I. And that's just a 1. We go down to J. And let's see, what's the minimax value of J? AUDIENCE: It's 20. PROFESSOR: That's right. 20 is the maximum of 20 and 2. Great, so what's the minimax value of D? Everyone said it-- 1. All right, so what's the minimax value at A? AUDIENCE: 6. PROFESSOR: 6 is right. 6 is higher than 2. It's higher than 1. Our value is 6. And our path is-- everyone-- A, C, H. That's it. Great, is everyone good with minimax? I know that usually a lot of people are. There's usually a few people who aren't. So if you're one of the people who would like some clarifications on minimax, raise your hand. There's probably a few other people who would like some too. OK. AUDIENCE: When you're doing this minimax, whatever values are not showing, you keep going down the tree and then just look at whether you're trying to find the minimax. And just whatever values you get you go back up one? PROFESSOR: Yes. The question was, when you go to do the minimax-- and let's say you got E was 2, and you know that B is going to be less than or equal to 2, but you don't know F yet. The question is, do you go down the tree, find the value at F, and then go back up? The answer is yes. By default, we use a depth first search. However, in non alpha-beta version, just regular minimax, it turns out it probably doesn't matter what you do. I suggested doing a depth first search to get yourself in the mindset of alpha-beta. Because order is very, very important in alpha-beta. But here, I don't know, you could do some weird bottom up search. Whatever you want, it's going to give you the right answer unless it asks what order they evaluated. But here's a hint. The order they're evaluated in is depth first search order. So without even doing anything, E, K, L, Q, R, N, H, I, O, P are the order of starting evaluation in this tree. AUDIENCE: [INAUDIBLE] PROFESSOR: So the question is, nodes like M and G we don't have to put values next to. Technically, if we were doing this very formally, and we couldn't remember, and I wasn't up there among the people, we would put 1 there. So at M, we would put a 1. But people remembered that. So we didn't do it. But then at G, we would put a 7. So if we were writing it out very formally, we would have a 1 and a 7. And at this D, we would have a 1. And then at the A, we would put a 6. And then that's the answer. Also, we've even put things like less than or greater than part way along the way. However, I believe that our alpha-beta search is going to definitely fulfill everyone's quota of pedantically putting lots of numbers next to nodes on a game tree. And so once you've done alpha-beta, if you can do it correctly, you'll think, oh, minimax, oh, those were the days. It's going to be easy. Because alpha-beta is a little bit more complicated. There's a lot of things that trip people up here. For alpha-beta, however, I will erase some of these numbers for the moment. They're still right. But we do it a little differently. So what do alpha-beta and beta add to this formula? Well, this is all sort of a winning formula, except for it's not. Because it takes too long. But it's a very nice formula. U is the maximizer, say. I would try to think, if I do this, what's he going to do? And then if he does that, what am I going to do? And then what is he going to do if I do that, et cetera, et cetera, all the way to the bottom. With alpha and beta, we add in what I like to call nuclear options. I'd like in this game of maximizer minimizer-- you can think of it as like the Cold War or the Peloponnesian War, except the Peloponnesian War didn't have nukes, so probably the Cold War. And in the Cold War, or any situation where you're up against an adversary who-- actually, this doesn't really work as well for the Cold War. But in any situation where you're up against an adversary whose only goal in life is to destroy you, you always want to find out what the best thing you can possibly do is if they hit that button and send nukes in from Cuba, or if they send fighter pilots, or whatever is going on. So the idea of alpha and beta is that they are numbers that represent the fail-safe, the worst case. Because obviously in the Cold War, sending nukes was not a good plan. But presumably, us sending nukes would be better than just being attacked and killed. So the alpha and beta represent the worst possible outcome you'd be willing to accept for your side. Because right now, you know you're guaranteed to be able to force the conflict to that point or better. So the alpha is the nuclear option, the fail-safe, of the maximizer. Nuclear options-- alpha is maximizer's nuclear option. And beta is the minimizer's nuclear option. So we ask ourselves-- and people who were paying attention at lecture or wrote stuff down know the answer already-- what could we possibly set to start off? Before we explore the tree and find anything, what will we set as our nuclear option, as our sort of fail-safe? We could always fall back on this number. So you could set 0. You could try to set some low number for the maximizer. Because if you set a high number for the maximizer as its fail-safe, it's going to be really snooty and just say, oh, I won't take this path. I already have a fail-safe that's better than all these paths. If you set, like, 100, you have no tree. Our default usually, in 6034, is to set negative infinity for alpha or negative some very large number if you're doing it in your lab. So if we set negative infinity as a default for alpha, that negative infinity is basically maximizer loses. So the maximizer goes in thinking, oh my god, if I don't look at this game tree, I automatically lose. He's willing to take the first path possibly presented. And that's why that negative infinity is a good default for alpha. Anyone have a good idea what a good default for beta is, or just remember? Positive infinity, that's right. Because the minimizer comes in, and she's like, oh crap, the maximizer automatically wins if I don't look at this node here. That makes sure the maximizer and the minimizer both are willing to look at the first path they see every time. Because look on this tree. If 10 was alpha, the maximizer would just reject out of hand everything except for P. And then we wouldn't have a tree. The maximizer would lose, because he would be like, hmm, this test game is very interesting. However, I have another option-- pft, and then you throw over the table. That's 10 for me, because you have to pick up the pieces. I don't own this set. I don't know. So that is why we set negative infinity and positive infinity as the default for alpha and beta. So how do alpha and beta propagate? And what do they do? The main purpose of alpha and beta is that, as we said, alpha-- let's say we have some chart of values. Alpha, which starts at negative infinity, is the worst that the maximizer is willing to accept. Because they know they can get that much or better. It starts out, that's the worst thing you can have. So it's not a problem. Infinity is the highest that the minimizer is willing to accept. That's beta. As you go along, though, the minimizer sees, oh, look at that. I can guarantee that at best the maximizer gets 100. Haha, beta is now 100. The maximizer says, oh yeah? Well I can guarantee that the lowest you can get me to go to is 0. So it's going to be 0. And this keeps going on until maybe at 6-- note, not drawn to scale. Maybe at 6, the maximizer said, haha, you can't make me go lower than 6. And the minimizer says, aha, you can't make me go higher than 6. And then 6 is the answer. If you ever get to a point where beta gets lower than alpha, or alpha gets lower than beta, then you just say, screw this. I'm not even going to look at the remaining stuff. I'm going to just prune now and go somewhere else that's less pointless than this. Because if the alpha gets higher than the beta, what that's saying is the maximizer says, oh man, look at this, minimizer. The lowest you can make me go is, say, 50. And the minimizer says, that's strange. Because the highest that you can make me go is 40. So something's generally amiss there. It usually means that one of the two of them doesn't even want to be exploring that branch at all. So you prune at that point. All right, so given that that's what we're looking for, how do we move the alphas and betas throughout the tree? There's a few different ways to draw them. And some of them I consider to be very busy. Probably in recitation and tutorial you will see a way that's busier and has more numbers. Technically, every node has both an alpha and a beta. However, the one that that node is paying attention to is the alpha, if it's a maximizer, and the beta if it's a minimizer. So I generally, for my purposes, only draw the alpha out for the maximizer and only draw the beta out for the minimizer. Very rarely, but it happens, they'll sometimes ask you, well, what's the beta of this node, which is a maximizer node? So it's good to know how it's derived. But I think that it wastes your time to write it out. That's my opinion. We'll see how it goes. So the way that it works, the way that alpha and beta works, is the Snow White principal. So does everyone know the story of Snow White? So there's a beautiful princess. There's an evil queen stepmother. Mirror mirror on the wall, who's the fairest of them all, finds out that it's the stepdaughter. So much like in the real world, in Snow White, the stepdaughter, Snow White, had the beauty of her parents. She inherited those. However, much like in the real world, maybe or perhaps not, the stepmother had an even better plan. She hired a hunter to sort of hunt Snow White, pull out Snow White's heart, and feed it to her so that she could gain Snow White's beauty for herself. How many people knew that version of the story? A few people. That's the original version of the story. Disney didn't put that in. The hunter then brought the heart of a deer, which I think in Disney the hunter did kill a deer arbitrarily, but it was not explained that that's why he was doing it. So in alpha-beta, it's just like that. By which I mean you start by inheriting the alpha and beta of your parents. But if you see something that you like amongst your children, you take it for yourself-- the Snow White principle. So let's see how that goes. Well, I told you guys that the default alpha was-- AUDIENCE: Negative infinity. PROFESSOR: Negative infinity. So here alpha is negative infinity. And I told you that the default beta was positive infinity. We're doing a depth first search here. All right, beta is infinity. All right, so we come here to E. Now, we could put an alpha. But I never put an alpha or a beta for one of the terminal nodes. Because it can't really do anything. It's just 2. So as we go down, we take the alpha and beta from our parents. But as we go up to a parent, if the parent likes what it sees in the child, it takes it instead. So I ask you all the question, would the minimizer prefer this 2 that it sees from its child or its own infinity for a beta? AUDIENCE: 2 PROFESSOR: It likes the 2. That's absolutely right. So 2. All right, great, so now we go down to F. What is F's alpha? Who says negative infinity? Who says 2? No one-- oh, you guys are good. It's negative infinity. Technically, it also will have a beta of 2. But we're ignoring the beta. And the alphas that have been progressing downward from the parents-- negative infinity. That's why I called it the grandfather clause before. Because you would often look up to your grandparent to see what your default number is. So we get an alpha of negative infinity. We then go down to the K. It's a static evaluation. And now I'm going to start calling on people individually. So hopefully people paid attention to the mob, who were always correct. All right, so we go down to K. And we see a 3. F is a maximizer node. So what does F do now? AUDIENCE: Switches its alpha to 3. PROFESSOR: Yes, switches its alpha to 3, great. All right, so that's already quite good. It switches alpha to 3. It's very happy. It's got a 3 here. That's a nice value. So what does it do at L, the next node? It's gone to K, went back up to F. Depth first search, the next one would be L, right? AUDIENCE: [INAUDIBLE] PROFESSOR: Well, technically F could take L's value of 0 if it liked it better than 3. But it's a maximizer. So does it want to take that? AUDIENCE: [INAUDIBLE] PROFESSOR: OK, that technically would be correct. But I'm sorry. I burdened you with a trick question. In fact, we don't look at L at all. Does everyone see that? I'll explain. The alpha at F has reached 3. But the beta at B is 2. So B looks down and says, wait a minute. If I go down to F, my enemy's nuclear option, my enemy is the worst it can be for-- the best it can be for me is 3. F is trumpeting it around. I was thinking of eating his heart, or whatever, but I didn't want to. But it's going to be 3. It's going to be 3 or higher down there at F. There's no way I want that. I already have my own default escape plan. And that's 2. That's going to be better than whatever comes out of that horrible F. So screw it. And we never look at L. Does everyone get that? That is the main principle of alpha-beta pruning. If you see an alpha that's higher than the beta above it-- as I said, if alpha goes up above the beta-- or if you see a beta, like if there's a beta down here, and it's lower than the alpha above it, prune it. Stop doing that. And the question is, who prunes? Who decides that you don't look at L? The person who is thinking not to look at L is always up higher by at least two levels. So up here, B is saying, hmm, I don't want to look at L. Because F is already so terrible for me that it's just beyond belief. If this is 100, it might be 100. Even if it's lower, I'm still going to get a three. There's a sanity check that I've written that I sort of came up with just in case you're not sure that you can skip it. Because on a lot of these tests, we ask you, which one's do you evaluate, which ones do you skip, right? Or we just say, which ones do you evaluate, and you don't write the ones that you skip. Here's my sanity test to see if you can skip it. Ask yourself, if that node that I'm about to skip contained a negative infinity or some arbitrarily small number, negative infinity being the minimizer wins, would it change anything? Now that I've answered that, if it contained a positive infinity, would it change anything? If the answer is no both times, then you're definitely correct in pruning it. So look at that 0. If it was a negative infinity, minimizer wins, what would happen? The maximizer would say, I'm not touching that was a 10 foot pole, choosing 3. The minimizer would say, screw that, I'll take E. Let's say it was a positive infinity. The maximizer would say, eureka, holy grain, I win. The minimizer would say, yeah, if I'm a moron, and go down to F, and then would go to E and take 2. So no matter what was there, the minimizer would go to E. And you could say, well, what if it was exactly 2? But still the maximizer would choose K. The minimizer would go to E. So there's no reason to go down there. We can just prune it off right now. Does everyone agree, everyone see what I'm talking about here? Great, so we're now done with this branch. Because beta is 2. So now we're up at old grandpappy A. And he has an alpha of negative infinity. Everyone, what will he do? He'll take the 2. It's better than negative infinity for him. It's not wonderful. But certainly anything is better than an automatic loss. All right, now our highest node is a 2. So let's keep that in mind for our alpha. OK, so let's go over here. Let's see, so what will be the value at C? What will be the beta value? AUDIENCE: [INAUDIBLE] PROFESSOR: You go back to which one? To G. I'm not at G yet. I'm actually just starting the middle branch. So I'm going to C. And what's going to be its starting beta before I go down? AUDIENCE: Infinity. PROFESSOR: Infinity, that's right-- default value. It's easier than it seemed. All right, so yes, beta is equal to infinity. This should be better erased. I think it's confusing people. Great, OK, so beta is equal to infinity at C. Now we go down depth first search to G. What's going to be our alpha at G? AUDIENCE: Minus infinity. PROFESSOR: Ahh, it would seem so. However, take a look up at the great-grandpappy A. It seems to have changed to 2. So this time it's 2. Why is it 2 instead of negative infinity? Why can we let A be so noxious and not start with saying, oh, I automatically lose? Well, A knows that no matter how awful things get in that middle branch, he can just say, screw the whole middle branch. I'm going to B. That's something that the minimizer can't do. And we have to start at infinity for the minimizer. But the maximizer can. Because he has the choice at the top. Does everyone see that? He can just say, oh, I'm not even going to C. Yeah, shows you. I'm going to A and taking the 2. So therefore alpha is actually 2 at G. All right, great, so we've got an alpha that's 2 at G. We're going to go down to M. It's a minimizer. All right, what's going to be our beta value at M? AUDIENCE: [INAUDIBLE] PROFESSOR: Or which is the beta default, minus or positive infinity? What would be the minimizer? AUDIENCE: Positive. PROFESSOR: Positive infinity, that's right. M is going to be a positive infinity for beta. Again, it picks it up from C. Great, now we get to some actual values. So we're at some actual values. We are at Q. So what's going to happen at M when M sees that Q is 1? AUDIENCE: [INAUDIBLE] PROFESSOR: What is beta? It says infinity. I'm sorry, it's hard to read. Beta is infinity at M. AUDIENCE: OK, so it's going to minimize, right? So it's going to be like, OK, [INAUDIBLE]. PROFESSOR: That's right. So they're going to put beta to 1. Because it sees Q. Great, so my next question is, what's going to happen at R? AUDIENCE: [INAUDIBLE] PROFESSOR: Very smart. You've detected my trap. The question is, does it look at R? The answer is, no. It doesn't look at R. Why doesn't it look at R? Does everyone see? Yeah, alpha is now greater than the beta below it. Beta has gotten lower than alpha. This is the same thing I was talking about before, when we figured out that the alpha here is 2. The maximizer says, wait a minute. The maximizer G says, if I go to M, the best I'm getting out of this is 1. Because if this is negative infinity, the minimizer will choose it. If this is positive infinity, he'll choose 1. The best I'm going to get out of here is 1. If that's the case, I might as well have just gone to B and not even gone to C. So I'm not going to go to M. I'll go to N, maybe. Maybe N is better. Does everyone see that? Great, so let's say that the maximizer does go to N. So what's going to happen with this alpha? AUDIENCE: [INAUDIBLE] PROFESSOR: That's right, it's going to be 7. 7 is better than 2. And the maximizer has control to get to that seven, at least if it gets to G. All right, now the minimizer at C-- we'll do everyone this time. The minimizer at C, seeing that 7, what does the minimizer do? Anyone? So it sees the 7. What does it do to its beta? It takes the 7-- better than infinity, anyway. And yeah, then it checks H. And everybody, again, what happens at H? It takes the 6. It's lower than 7. All right, now we'll go back to having people do it on their own. Well, all the way back to the top, what does A do when it sees the 6 coming out of C? AUDIENCE: Changes to 6. PROFESSOR: Changes to 6, that's right. Alpha equals 6. Great-- homestretch, people, homestretch. So the minimizer, everyone, has a beta of infinity. And if I wasn't a static node, it would have an alpha of 6. But it is a static node. So it just has a value of 1. So since it has a value of 1, everyone, the beta becomes 1. And what next, everyone? AUDIENCE: Prune. PROFESSOR: Prune, that's right. Why prune? Well, this time it's A himself who can prune. A says, well darn, if I go to D, I'm going to get 1 or something even worse than 1. I might as well take my 6 while I have it, prune all the rest all the way down. Everyone see that? Everyone cool with that? It's not too bad if you take it one step at a time. We did it. Our question is, which nodes are evaluated in order? Our answer is, everyone-- E, K, Q, N, H, I. OK, not so obvious, I guess. A few people followed me. But it is E, K, Q, N, H, I. It's just depth first order. And we pruned some of them away. Great, so that is alpha-beta. Any questions about that before I give some questions about progressive deepening? All right, we've got a bunch. So first question. AUDIENCE: [INAUDIBLE] nodes like F, B, C, and D? PROFESSOR: The question is, when asked for the order of evaluation, are we excluding F, B, C, and D? The answer is we're talking about here static evaluation. The static evaluator is a very important and interesting function. And I'll get back to something a few students have asked me about the static evaluator later and try to explain what it is. It's basically the thing that pops out those numbers at the bottom of the leaves. So when we ask, what is the order of nodes that were statically evaluated, we mean leaves only. That's a good question. Any other questions? Let's see, there was one up here before. But it's gone. It might have been the same one. Question? AUDIENCE: So a similar question. When you say, static nodes, that just means the leaf nodes? PROFESSOR: Means the leaf nodes, that's right. The question is, does static nodes mean the leaf nodes. The answer is yes. AUDIENCE: And so static evaluation is when you compare the value of a static node to something? PROFESSOR: Static evaluation is when you get that number, the static node. Let me explain. Unless someone else has another question about alpha-beta, let me explain static values. Because I was about to do that. There is a question about alpha-beta. I'll come back to both of yours after I answer this. AUDIENCE: You were mentioning [INAUDIBLE]. And I'm a little bit confused. If you're looking at one node, and you're seeing either grab the value from the grandparent or grab it from the-- PROFESSOR: So it always starts-- the question is, what is the Snow White principle? How does it work? Every node always starts off with taking the value of the same type, alpha or beta, from its grandparent. It always starts that way. Now, you say, why the grandparent? Wouldn't it take it from the parent? It actually does. But I'm not drawing out the alphas at all the minimizer levels. Because they don't do anything. They're only even there to pass them down. So all of the values pass down, down, down, down, down to begin. Every node, in fact, starts off with its grandparents with its parents' values, OK? But then when the node sees a child, it's completely done evaluating. It's finished. It can't be in the process. Let's say C. When C sees that G is completely done with all of its sub-branches and is ready to return a value, or if it's just a static evaluation, then it's automatically completely done. Because it has no children. A static value like K of 3 is automatically completely done. It's got a 3. Similarly, when we came back to G after going to N, and we knew that the value was 7, that was completely done. The value was definitely 7. There was no other possibilities. AUDIENCE: That's after looking at the children, right? PROFESSOR: Yes. So once you're done with all the children of G, then G comes up and says, guess what? Guess what, guys? So technically before that, you would have said that G's alpha is greater than or equal to 1 when we looked at Q. And then we looked at M. We'd say, it's equal exactly to 7. We're done here. And then at that point, when it's fresh and ripe and has all of its highest value or its best value, that's when the parent can eat its heart and gain that value itself. So that's when C says, for instance, oh man, I have an infinity. I really like that 7 better. And it takes the 7. But then it saw H. And it said, oh man, that's a 6. That's even better than 7. So it took the 6. AUDIENCE: So shouldn't the alpha take 7 then? PROFESSOR: So alpha takes 6. Because C is a minimizer. C took the 7 from G, but then right after that C saw H and took the 6. Because 6 is even lower than 7. And then alpha took the 6. Because 6 was higher than 2. AUDIENCE: So it's not going to look below the branch? PROFESSOR: Yeah, the problem is that the maximizer doesn't have control there. The minimizer has got control at C. And the minimizer is going to make sure it's as low as possible. The maximizer at A, his only control, or her only control, is the ability to send either way to B or C or D. And then at that point, at C, the minimizer gets to choose if we go to G or H. And it's never going to choose G. Because G is higher than H. All right, awesome, was there another question? All right, let's go back to static evaluations. When I first took this class, I had some weird thoughts about static evolutions. I heard some students ask me this. I almost got a question about it onto one of the tests, but it was edited to some other weird question that was m to the b to the d minus 1 or something like that at the last minute. So I'm going to pose you guys the actual question that would have been on one of the older test, which is the following. I had a student who came to me and said, you know, [INAUDIBLE], when we do this alpha-beta pruning, and all this other stuff, we're trying to assume that we're really saving that much time by getting rid of a few static evaluations. In fact, when we do progressive deepening, we're always just counting, how many static evaluations do we have to do? And he said, I look at these static evaluations. And there's just a 3 there. It takes no time to do the static evaluation. It's on the board. It takes much longer to do the alpha-beta. It's faster by far to not do alpha-beta. So I then tried to explain to that student. I said, OK, we need to be clear about what static evaluations are. You guys get it easy. We put these numbers on the board. A static evaluation-- let's say you're playing a game like chess. Static evaluation takes a long time. When I was in 6170, Java [INAUDIBLE], the class that used to exist, we had a program called Anti-Chess where I used my 6034 skills to write the AI. And the static evaluator took a long time. And we were timed. So getting the static evaluator faster, that was the most important thing. Why does it take a long time? Well, the static evaluator is an evaluation of the board position, the state of the game, at a snapshot of time. And that's not as easy as just saying, oh, here's the answer. Because in chess, first of all, not only did I have to look at how many pieces I had, what areas that I controlled. Also-- well, it was anti-chess. But that's not withstanding. Let's pretend it's regular chess. I also had to look, if it was in regular chess-- and I still had to do this in anti-chess-- if my king was in check. And what that meant is I had to look at all of my opponent's moves, possible moves, to see if anyone of them could take my king. Because in regular chess, it's illegal to put your king into check. So you better not even allow that move. And regardless, getting into checkmate is negative infinity for you. So it takes a really long time to do static evaluations, at least good ones, usually. You want to avoid them. Because they're not just some number on the page. They are some function you wrote that does a very careful analysis of the state of the game and says, I'm good to heuristically guess that my value is pi, or some other number, and then rates that compared to other states. Does that make sense to everyone? So the answer to the hypothetical question that might have been on the old test, when the person said, I've got this great idea where you do tons of static evaluation, and you don't have to do this long alpha-beta, is, don't do that. The static evaluations actually take a long time. Does that clear it up for people who asked me before about what is a static evaluation, why are the leaf nodes called static? And you might ask, why are some of these static just arbitrarily? The answer is, when you're running out of time to expand deeper, and you just need to stop that stage of the game-- maybe it's just getting too hairy, maybe it's spreading out too much, you have some heuristic that says, this is where I stop for now-- it's a heuristic guess of the value. It's kind of like those heuristic values in the search tree. It's a guess of how much work you have left to get to the goal. Here, you say, well, I wish I could go deeper. But I just don't have the time. So here's how I think I'm doing at this level. It's not always right. And that's going to lead us into the answer to one of the questions about progressive deepening. So I'll put up the progressive deepening question really quickly. So the question is this. Let me see, this is a maximizer-- yes. Suppose that we do progressive deepening on the tree that is only two levels deep. What is progressive deepening in a nutshell if you don't remember from the lecture? The idea is this. In this tree, it doesn't work. But in trees that actually branch like 2 to the n, it doesn't take that much time to do some of the top levels first and then move on to the bottom levels. Just do them one at a time. So let's say we only did it up through J. We only did the top two levels of the tree. We'd like to reorder the tree so that alpha-beta can prune as much as it possibly can, at least we hope. So let's pretend that we had a psychic awesome genius friend who told us that the static values when we went up to two levels-- remember, when we go to two levels, F, G, and J have to get a static value, right? Because we're not going down. We do a static evaluation. They get the exact correct numbers-- 3, 7, and 20. Genius, brilliant. All right, so if that happens, what is the best way that we could reorder that tree? Oh yeah, so it's A, B, C, D with values of 2, 3, 7, 6, 1, 20. I'll draw that. This is the non-reordered tree. Let's see, so it's 2, 3, 7, 6, 1, 20. So what's the best way to reorder? Well, first of all, does anyone remember what Patrick said when he talked about progressive deepening? Usually no one does, so don't worry about it. Because at that time you guys didn't think, oh, I have to do this for the quiz. You were just thinking, oh man, we've already heard alpha-beta and all this other stuff. And this is just a small fact. But it's a very important fact. And now you know you have to do it for the quiz. So you're probably going to remember it. The way you do it is you try to guess, and you say, which one of these is going to be a winner? Whichever one I think is going to be a winner at that level, I put first. Why is that the case? Well, something interesting you may have noticed here-- whenever you have a winner, like the middle node, or whenever you have whatever is the current best for your alpha, you sort of have to explore out a lot of that area. Like for instance, the left node was our current best at 2. The middle branch was our current best, at that time was 6. It was the total best. We had to explore a good number of nodes. But on the right, we just saw, oh, there's 1. We're done. We cut everything off. In other words, the branch that turns out to be the one that you take, you have to do a pretty good amount of exploration to prove that it's the right one. Whereas if it's the wrong one, you can sometimes with just one node say, this is wrong, done. So therefore, if the one that turns out to be the eventual winner is first of all, then it's really easy to reject all the other branches. Do people see that sort of conceptually a little bit, that if you get the best node right away, you can just reject all the wrong ones pretty quickly? That's our goal. So how can we, quote, "get the right one," the best one right away? Well, here's how we do it. Let's say we're at B. Which one is the minimizer likely to pick assuming that our heuristic is good and that these guesses are pretty much close to the truth? It turns out they're perfect, so this is going to work. So which one will the minimizer pick if it has to choose between E and F, do we think? AUDIENCE: E. PROFESSOR: E, perfect. Which one will it pick between G and H? AUDIENCE: H. PROFESSOR: H. Which one will it pick between I and J? AUDIENCE: I. PROFESSOR: OK, so what we're saying is we think it's going to pick E. We think it's going to pick H. We think it's going to pick I. So first of all, we should put E before F, H before G, and I before J. Because we think it's going to pick those first. Those are probably our best ones to invalidate a poor branch. So now between 2, 6, and 1, which is what we think we're going to get, which one do we think the maximizer is going to take? AUDIENCE: 6. PROFESSOR: 6. Then if it couldn't take 6, what would be its next best choice? 2, then 1. That's just our order-- simple as that. It couldn't be anything easier that evolves really complex trees, a huge number of numbers, and reordering those trees. So C-- you guys told me C, B, D. You told me C, B, D, I think? Yeah, those are the ones the maximizer likes. And then the ones the minimizer likes you told me was H, and before G. Because H is smaller than G. You guys told me E before F. And you guys told me I before J. And you guys would be correct in all regards. We have 6, 7, 2, 3, 1, 20. All the minimizers choose from smallest to highest. The maximizer chooses from highest to lowest of the ones that the minimizers will take. And if we did that, you can see we would probably save some time. Let's see how much time. Let's say we looked at H first. Well, if we looked at H first, we would still have actually had to look at Q and N. However, we would not have had to look at K. Do people see why? If we already knew this branch was 6, as soon as we saw 2 for the beta here-- 2 is less than 6-- we could have pruned. We still would have had to look at I over here. Because you have to look at at least one thing in the new sub-branch. And it actually only would have saved us one node-- oops. So it winds up that in total, how many nodes would we have evaluated if we did that little scheme of reordering? Well, we normally had to do six-- E, K, Q, N, H, I. How many do we evaluate if we do this progressive deepening scheme? How many times do we run the static evaluator, which of course you know the static evaluator takes a long time? Anyone have a guess? I told you the only one we don't evaluate is K. Raise your hand. I won't make anyone give this one. So I said the only one we save on is K. So we still do E, Q, N, H, and I over here. There's two possible answers that I will accept. So you have a higher chance of guessing it. Anyway? Does everyone agree that we did six before? If we didn't do any progressive deepening, we just did E, K, Q, N, H, I. And now we're not doing K. OK, people are saying five. All right, good. That's not the right answer. But it at least shows that you can do taking away the one. We did at least five over here. There's two possible answers, though. Because look over there. In order to do the progressive deepening, we had to do those static evaluations, right? So we either did all those static evaluations and these five-- E, K, Q, N, H, I-- static evaluations. Because we didn't do the K. Or we might have saved ourselves. Because maybe we were smart and decided to cache the static values when we were going down the tree. It's an implementation detail that on this test when we asked that question we didn't say. What I mean by cache is when we did it here and saw that E was a 2, and then here-- oh, we have to do the static value at E. If we were smart, we might have made a little hash table or something and put down 2 so we didn't have to do a static evaluation at E. And if that happened, well, we save E, H, and I, and we do three fewer. Does everyone see that? However, that's still more than six. So it didn't save us time. So you might say, oh, progressive deepening is a waste of time. But it's not. Because this is a very, very small, not very branchy tree that was made so that you guys could easily do alpha-beta and take the quiz, and it wouldn't be bad. If this was actually branching even double at each level, it would have, what, 16 nodes down here at the bottom. Then you would want to be doing that progressive deepening. So now I ask you a conceptual riddle question. It's not really that much of a riddle. But we'll see if anyone wants to answer. Again, I won't call on you for this. According to this test, a student named Steve says, OK, I know I have to pay to do the progressive deepening here. But let's ignore that. Because it's small in a large tree, right? It's not going to take that much. Let's ignore the costs of the progressive deepening and only look at how much we do here. He says, when it comes to performing the alpha-beta on the final level, I'm guaranteed to always prune at least as well or better if I rearrange the nodes based on the best result from progressive deepening. Do you agree? AUDIENCE: [INAUDIBLE] PROFESSOR: Can I repeat it? OK, the question is, ignoring the cost that we pay progressively deepening here-- just forget about it-- at the final step, at the final iteration, the question is, am I guaranteed to do at least as well or better in my alpha-beta pruning when I reorder based on the best order for progressive deepening? Here certainly we did. But the question is, is Steve guaranteed? Answer? AUDIENCE: [INAUDIBLE] PROFESSOR: What did you say? AUDIENCE: [INAUDIBLE] PROFESSOR: That's the answer and the why, which we asked to explain. The answer we got is, doesn't that depend on the heuristic? Perfectly correct. The answer is, no, we're not guaranteed, and it depends on the heuristic. So if we were guaranteed, that would be our heuristic was godlike, like this heuristic. If your heuristic already tells you the correct answer no matter what, don't do game search. Just go to the empty chess board, put all the pieces in the front rows, and run static evaluator on that. And it'll say, oh, it looks like with this game not started that white is stupid, so black will win in 15 turns. And then you're done. And you don't do a search. We know that our heuristic is flawed in some way. It could be very flawed. If it's flawed so badly that it tells us a very bad result of what's actually going to happen, even though we think the minimizer is going to go to H, maybe it's wrong by a lot and it goes to G. It could take us up an even worse path and make us take longer. Question? AUDIENCE: If it's the heuristic, how could you cache the values so you didn't have to recalculate them later? PROFESSOR: The question is, how can you cache the values if it's a heuristic so you don't have to recalculate them later? The answer is, it wouldn't help if there weren't these weird multi-level things where we stop at E for some reason, even though it goes down to five levels. The way you could cache it is it is a heuristic. But it's consistent. And I don't mean consistent from search. I mean it's a consistent heuristic in-- the game state E is, let's say that's the state where I moved out my knight as the maximizer, and the minimizer said, you're doing the knight opening, really, and then did a counterattack. No matter how we get to E, or where we go to get to E, that's always going to be state E. It's always going to have the same heuristic value. It's not like some guy who goes around and just randomly pulls a number out of a hat. We're going to have some value that gives us points based on state E. And it's going to be the same any time we go to state E. Does that make sense? It is a heuristic. But it's always going to give the same value at E no matter how you got to E. But it could be really bad. In fact, you might consider a heuristic that's the opposite of correct and always tells us the worst move and claims it's the best. That's the heuristic that the minimizer program did to our computer, perhaps. In that case, when we do progressive deepening and we reorder, we'll probably get the worst pruning possible. We might not. But we may. So in that case, you're not guaranteed. I hope that's given a few clues. In tutorial, you guys are going to see some more interesting problems that go into a few other details. I at least plan on doing [INAUDIBLE] interesting game problem from last year, which asked a bunch of varied things that are a little bit different from these. So it should be a lot of fun, hopefully, or at least useful, to do the next quiz. So have a great weekend. Don't stress out too much about the quiz. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 12a_Neural_Nets.txt | The following content is provided under a Creative Commons license. Your support will help MIT OpenCourseWare continue to offer high-quality educational resources for free. To make a donation or to view additional materials from hundreds of MIT courses, visit MIT OpenCourseWare at fsae@mit.edu. PATRICK WINSTON: It was in 2010, yes, that's right. It was in 2010. We were having our annual discussion about what we would dump fro 6034 in order to make room for some other stuff. And we almost killed off neural nets. That might seem strange because our heads are stuffed with neurons. If you open up your skull and pluck them all out, you don't think anymore. So it would seem that neural nets would be a fundamental and unassailable topic. But many of us felt that the neural models of the day weren't much in the way of faithful models of what actually goes on inside our heads. And besides that, nobody had ever made a neural net that was worth a darn for doing anything. So we almost killed it off. But then we said, well, everybody would feel cheated if they take a course in artificial intelligence, don't learn anything about neural nets, and then they'll go off and invent them themselves. And they'll waste all sorts of time. So we kept the subject in. Then two years later, Jeff Hinton from the University of Toronto stunned the world with some neural network he had done on recognizing and classifying pictures. And he published a paper from which I am now going to show you a couple of examples. Jeff's neural net, by the way, had 60 million parameters in it. And its purpose was to determine which of 1,000 categories best characterized a picture. So there it is. There's a sample of things that the Toronto neural net was able to recognize or make mistakes on. I'm going to blow that up a little bit. I think I'm going to look particularly at the example labeled container ship. So what you see here is that the program returned its best estimate of what it was ranked, first five, according to the likelihood, probability, or the certainty that it felt that a particular class was characteristic of the picture. And so you can see this one is extremely confident that it's a container ship. It also was fairly moved by the idea that it might be a lifeboat. Now, I'm not sure about you, but I don't think this looks much like a lifeboat. But it does look like a container ship. So if I look at only the best choice, it looks pretty good. Here are the other things they did pretty well, got the right answer is the first choice-- is this first choice. So over on the left, you see that it's decided that the picture is a picture of a mite. The mite is not anywhere near the center of the picture, but somehow it managed to find it-- the container ship again. There is a motor scooter, a couple of people sitting on it. But it correctly characterized the picture as a motor scooter. And then on the right, a Leopard. And everything else is a cat of some sort. So it seems to be doing pretty well. In fact, it does do pretty well. But anyone who does this kind of work has an obligation to show you some of the stuff that doesn't work so well on or doesn't get quite right. And so these pictures also occurred in Hinton's paper. So the first one is characterized as a grill. But the right answer was supposed to be convertible. Oh, no, yes, yeah, right answer was convertible. In the second case, the characterization is of a mushroom. And the alleged right answer is agaric. Is that pronounced right? It turns out that's a kind of mushroom-- so no problem there. In the next case, it said it was a cherry. But it was supposed to be a dalmatian. Now, I think a dalmatian is a perfectly legitimate answer for that particular picture-- so hard to fault it for that. And the last case, the correct answer was not in any of the top five. I'm not sure if you've ever seen a Madagascar cap. But that's a picture of one. And it's interesting to compare that with the first choice of the program, the squirrel monkey. This is the two side by side. So in a way, it's not surprising that it thought that the Madagascar cat was a picture of a squirrel monkey-- so pretty impressive. It blew away the competition. It did so much better the second place wasn't even close. And for the first time, it demonstrated that a neural net could actually do something. And since that time, in the three years since that time, there's been an enormous amount of effort put into neural net technology, which some say is the answer. So what we're going to do today and tomorrow is have a look at this stuff and ask ourselves why it works, when it might not work, what needs to be done, what has been done, and all those kinds of questions will emerge. So I guess the first thing to do is think about what it is that we are being inspired by. We're being inspired by those things that are inside our head-- all 10 to the 11th of them. And so if we take one of those 10 to the 11th and look at it, you know from 700 something or other approximately what a neuron looks like. And by the way, I'm going to teach you in this lecture how to answer questions about neurobiology with an 80% probability that you will give the same answer as a neurobiologist. So let's go. So here's a neuron. It's got a cell body. And there is a nucleus. And then out here is a long thingamajigger which divides maybe a little bit, but not much. And we call that the axon. So then over here, we've got this much more branching type of structure that looks maybe a little bit like so. Maybe like that-- and this stuff branches a whole lot. And that part is called the dendritic tree. Now, there are a couple of things we can note about this is that these guys are connected axon to dendrite. So over here, they'll be a so-called pre-synaptic thickening. And over here will be some other neuron's dendrite. And likewise, over here some other neuron's axon is coming in here and hitting the dendrite of our the one that occupies most of our picture. So if there is enough stimulation from this side in the axonal tree, or the dendritic tree, then a spike will go down that axon. It acts like a transmission line. And then after that happens, the neuron will go quiet for a while as it's recovering its strength. That's called the refractory period. Now, if we look at that connection in a little more detail, this little piece right here sort of looks like this. Here's the axon coming in. It's got a whole bunch of little vesicles in it. And then there's a dendrite over here. And when the axon is stimulated, it dumps all these vesicles into this inner synaptic space. For a long time, it wasn't known whether those things were actually separated. I think it was Raamon and Cahal who demonstrated that one neuron is actually not part of the next one. They're actually separated by these synaptic gaps. So there it is. How can we model, that sort of thing? Well, here's what's usually done. Here's what is done in the neural net literature. First of all, we've got some kind of binary input, because these things either fire or they don't fire. So it's an all-or-none kind of situation. So over here, we have some kind of input value. We'll call it x1. And is either a 0 or 1. So it comes in here. And then it gets multiplied times some kind of weight. We'll call it w1. So this part here is modeling this synaptic connection. It may be more or less strong. And if it's more strong, this weight goes up. And if it's less strong, this weight goes down. So that reflects the influence of the synapse on whether or not the whole axon decides it's stimulated. Then we got other inputs down here-- x sub n, also 0 or 1. It's also multiplied by a weight. We'll call that w sub n. And now, we have to somehow represent the way in which these inputs are collected together-- how they have collective force. And we're going to model that very, very simply just by saying, OK, we'll run it through a summer like so. But then we have to decide if the collective influence of all those inputs is sufficient to make the neuron fire. So we're going to do that by running this guy through a threshold box like so. Here is what the box looks like in terms of the relationship between input and the output. And what you can see here is that nothing happens until the input exceeds some threshold t. If that happens, then the output z is a 1. Otherwise, it's a 0. So binary, binary out-- we model the synaptic weights by these multipliers. We model the cumulative effect of all that input to the neuron by a summer. We decide if it's going to be an all-or-none 1 by running it through this threshold box and seeing if the sum of the products add up to more than the threshold. If so, we get a 1. So what, in the end, are we in fact modeling? Well, with this model, we have number 1, all or none-- number 2, cumulative influence-- number 3, oh, I, suppose synaptic weight. But that's not all that there might be to model in a real neuron. We might want to deal with the refractory period. In these biological models that we build neural nets out of, we might want to model axonal bifurcation. We do get some division in the axon of the neuron. And it turns out that that pulse will either go down one branch or the other. And which branch it goes down depends on electrical activity in the vicinity of the division. So these things might actually be a fantastic coincidence detectors. But we're not modeling that. We don't know how it works. So axonal bifurcation might be modeled. We might also have a look at time patterns. See, what we don't know is we don't know if the timing of the arrival of these pulses in the dendritic tree has anything to do with what that neuron is going to recognize-- so a lot of unknowns here. And now, I'm going to show you how to answer a question about neurobiology with 80% probability you'll get it right. Just say, we don't know. And that will be with 80% probability what the neurobiologist would say. So this is a model inspired by what goes on in our heads. But it's far from clear if what we're modeling is the essence of why those guys make possible what we can do. Nevertheless, that's where we're going to start. That's where we're going to go. So we've got this model of what a neuron does. So what about what does a collection of these neurons do? Well, we can think of your skull as a big box full of neurons. Maybe a better way to think of this is that your head is full of neurons. And they in turn are full of weights and thresholds like so. So into this box come a variety of inputs x1 through xm. And these find their way to the inside of this gaggle of neurons. And out here come a bunch of outputs c1 through zn. And there a whole bunch of these maybe like so. And there are a lot of inputs like so. And somehow these inputs through the influence of the weights of the thresholds come out as a set of outputs. So we can write that down a little fancier by just saying that z is a vector, which is a function of, certainly the input vector, but also the weight vector and the threshold vector. So that's all a neural net is. And when we train a neural net, all we're going to be able to do is adjust those weights and thresholds so that what we get out is what we want. So a neural net is a function approximator. It's good to think about that. It's a function approximator. So maybe we've got some sample data that gives us an output vector that's desired as another function of the input, forgetting about what the weights and the thresholds are. That's what we want to get out. And so how well we're doing can be figured out by comparing the desired value with the actual value. So we might think then that we can get a handle on how well we're doing by constructing some performance function, which is determined by the desired vector and the input vector-- sorry, the desired vector and the actual output vector for some particular input or for some set of inputs. And the question is what should that function be? How should we measure performance given that we have what we want out here and what we actually got out here? Well, one simple thing to do is just to measure the magnitude of the difference. That makes sense. But of course, that would give us a performance function that is a function of the distance between those vectors would look like this. But this turns out to be mathematically inconvenient in the end. So how do you think we're going to turn it up a little bit? AUDIENCE: Normalize it? PATRICK WINSTON: What's that? AUDIENCE: Normalize it? PATRICK WINSTON: Well, I don't know. How about just we square it? And that way we're going to go from this little sharp point down there to something that looks more like that. So it's best when the difference is 0, of course. And it gets worse as you move away from 0. But what we're trying to do here is we're trying to get to a minimum value. And I hope you'll forgive me. I just don't like the direction we're going here, because I like to think in terms of improvement as going uphill instead of down hill. So I'm going to dress this up one more step-- put a minus sign out there. And then our performance function looks like this. It's always negative. And the best value it can possibly be is zero. So that's what we're going to use just because I am who I am. And it doesn't matter, right? Still, you're trying to either minimize or maximize some performance function. OK, so what do we got to do? I guess what we could do is we could treat this thing-- well, we already know what to do. I'm not even sure why we're devoting our lecture to this, because it's clear that what we're trying to do is we're trying to take our weights and our thresholds and adjust them so as to maximize performance. So we can make a little contour map here with a simple neural net with just two weights in it. And maybe it looks like this-- contour map. And at any given time we've got a particular w1 and particular w2. And we're trying to find a better w1 and w2. So here we are right now. And there's the contour map. And it's a 6034. So what do we do? AUDIENCE: Climb. PATRICK WINSTON: Simple matter of hill climbing, right? So we'll take a step in every direction. If we take a step in that direction, not so hot. That actually goes pretty bad. These two are really ugly. Ah, but that one-- that one takes us up the hill a little bit. So we're done, except that I just mentioned that Hinton's neural net had 60 million parameters in it. So we're not going to hill climb with 60 million parameters because it explodes exponentially in the number of weights you've got to deal with-- the number of steps you can take. So this approach is computationally intractable. Fortunately, you've all taken 1801 or the equivalent thereof. So you have a better idea. Instead of just taking a step in every direction, what we're going to do is we're going to take some partial derivatives. And we're going to see what they suggest to us in terms of how we're going to get around in space. So we might have a partial of that performance function up there with respect to w1. And we might also take a partial derivative of that guy with respect to w2. And these will tell us how much improvement we're getting by making a little movement in those directions, right? How much a change is given that we're just going right along the axis. So maybe what we ought to do is if this guy is much bigger than this guy, it would suggest we mostly want to move in this direction, or to put it in 1801 terms, what we're going to do is we're going to follow the gradient. And so the change in the w vector is going to equal to this partial derivative times i plus this partial derivative times j. So what we're going to end up doing in this particular case by following that formula is moving off in that direction right up to the steepest part of the hill. And how much we move is a question. So let's just have a rate constant R that decides how big our step is going to be. And now you think we were done. Well, too bad for our side. We're not done. There's a reason why we can't use-- create ascent, or in the case that I've drawn our gradient, descent if we take the performance function the other way. Why can't we use it? AUDIENCE: Local maxima. PATRICK WINSTON: The remark is local maxima. And that is certainly true. But it's not our first obstacle. Why doesn't gradient ascent work? AUDIENCE: So you're using a step function. PATRICK WINSTON: Ah, there's something wrong with our function. That's right. It's non-linear, but rather, it's discontinuous. So gradient ascent requires a continuous space, continuous surface. So too bad our side. It isn't. So what to do? Well, nobody knew what to do for 25 years. People were screwing around with training neural nets for 25 years before Paul Werbos sadly at Harvard in 1974 gave us the answer. And now I want to tell you what the answer is. The first part of the answer is those thresholds are annoying. They're just extra baggage to deal with. What we really like instead of c being a function of xw and t was we'd like c prime to be a function f prime of x and the weights. But we've got to account for the threshold somehow. So here's how you do that. What you do is you say let us add another input to this neuron. And it's going to have a weight w0. And it's going to be connected to an input that's always minus 1. You with me so far? Now what we're going to do is we're going to say, let w0 equal t. What does that do to the movement of the threshold? What it does is it takes that threshold and moves it back to 0. So this little trick here takes this pink threshold and redoes it so that the new threshold box looks like this. Think about it. If this is t, and this is minus 1, then this is minus t. And so this thing ought to fire if everything's over-- if the sum is over 0. So it makes sense. And it gets rid of the threshold thing for us. So now we can just think about weights. But still, we've got that step function there. And that's not good. So what we're going to do is we're going to smooth that guy out. So this is trick number two. Instead of a step function, we're going to have this thing we lovingly call a sigmoid function, because it's kind of from an s-type shape. And the function we're going to use is this one-- one, well, better make it a little bit different-- 1 over 1 plus e to the minus whatever the input is. Let's call the input alpha. Does that makes sense? Is alpha is 0, then it's 1 over 1 plus 1 plus one half. If alpha is extremely big, then even the minus alpha is extremely small. And it becomes one. It goes up to an asymptotic value of one here. On the other hand, if alpha is extremely negative, than the minus alpha is extremely positive. And it goes to 0 asymptotically. So we got the right look to that function. It's a very convenient function. Did God say that neurons ought to be-- that threshold ought to work like that? No, God didn't say so. Who said so? The math says so. It has the right shape and look and the math. And it turns out to have the right math, as you'll see in a moment. So let's see. Where are we? We decided that what we'd like to do is take these partial derivatives. We know that it was awkward to have those thresholds. So we got rid of them. And we noted that it was impossible to have the step function. So we got rid of it. Now, we're a situation where we can actually take those partial derivatives, and see if it gives us a way of training the neural net so as to bring the actual output into alignment with what we desire. So to deal with that, we're going to have to work with the world's simplest neural net. Now, if we've got one neuron, it's not a net. But if we've got two-word neurons, we've got a net. And it turns out that's the world's simplest neuron. So we're going to look at it-- not 60 million parameters, but just a few, actually, just two parameters. So let's draw it out. We've got input x. That goes into a multiplier. And it gets multiplied times w1. And that goes into a sigmoid box like so. We'll call this p1, by the way, product number one. Out here comes y. Y gets multiplied times another weight. We'll call that w2. The neck produces another product which we'll call p2. And that goes into a sigmoid box. And then that comes out as z. And z is the number that we use to determine how well we're doing. And our performance function p is going to be one half minus one half, because I like things are going in a direction, times the difference between the desired output and the actual output squared. So now let's decide what those partial derivatives are going to be. Let me do it over here. So what are we trying to compute? Partial of the performance function p with respect to w2. OK. Well, let's see. We're trying to figure out how much this wiggles when we wiggle that. But you know it goes through this variable p2. And so maybe what we could do is figure out how much this wiggles-- how much z wiggles when we wiggle p2 and then how much p2 wiggles when we wiggle w2. I just multiplied those together. I forget. What's that called? N180-- something or other. AUDIENCE: The chain rule PATRICK WINSTON: The chain rule. So what we're going to do is we're going to rewrite that partial derivative using chain rule. And all it's doing is saying that there's an intermediate variable. And we can compute how much that end wiggles with respect how much that end wiggles by multiplying how much the other guys wiggle. Let me write it down. It makes more sense in mathematics. So that's going to be able to the partial of p with respect to z times the partial of z with respect to p2. Keep me on track here. Partial of z with respect to w2. Now, I'm going to do something for which I will hate myself. I'm going to erase something on the board. I don't like to do that. But you know what I'm going to do, don't you? I'm going to say this is true by the chain rule. But look, I can take this guy here and screw around with it with the chain rule too. And in fact, what I'm going to do is I'm going to replace that with partial of z with respect to p2 and partial of p2 with respect to w2. So I didn't erase it after all. But you can see what I'm going to do next. Now, I'm going to do same thing with the other partial derivative. But this time, instead of writing down and writing over, I'm just going to expand it all out in one go, I think. So partial of p with respect to w1 is equal to the partial of p with respect to z, the partial of z with respect to p2, the partial of p2 with respect to what? Y? Partial of y with respect to p1-- partial of p1 with respect to w1. So that's going like a zipper down that string of variables expanding each by using the chain rule until we got to the end. So there are some expressions that provide those partial derivatives. But now, if you'll forgive me, it was convenient to write them out that way. That matched the intuition in my head. But I'm just going to turn them around. It's just a product. I'm just going to turn them around. So partial p2, partial w2, times partial of z, partial p2, times the partial of p with respect to z-- same thing. And now, this one. Keep me on track, because if there's a mutation here, it will be fatal. Partial of p1-- partial of w1, partial of y, partial p1, partial of p2, partial of y, partial of z. There's a partial of p2, partial of a performance function with respect to z. Now, all we have to do is figure out what those partials are. And we have solved this simple neural net. So it's going to be easy. Where is my board space? Let's see, partial of p2 with respect to-- what? That's the product. The partial of z-- the performance function with respect to z. Oh, now I can see why I wrote it down this way. Let's see. It's going to be d minus e. We can do that one in our head. What about the partial of p2 with respect to w2. Well, p2 is equal to y times w2, so that's easy. That's just y. Now, all we have to do is figure out the partial of z with respect to p2. Oh, crap, it's going through this threshold box. So I don't know exactly what that partial derivative is. So we'll have to figure that out, right? Because the function relating them is this guy here. And so we have to figure out the partial of that with respect to alpha. All right, so we got to do it. There's no way around it. So we have to destroy something. OK, we're going to destroy our neuron. So the function we're dealing with is, we'll call it beta, equal to 1 over 1 plus e to the minus alpha. And what we want is the derivative with respect to alpha of beta. And that's equal to d by d alpha of-- you know, I can never remember those quotient formulas. So I am going to rewrite it a little different way. I am going to write it as 1 minus e to the minus alpha to the minus 1, because I can't remember the formula for differentiating a quotient. OK, so let's differentiate it. So that's equal to 1 minus e to the minus alpha to the minus 2. And we got that minus comes out of that part of it. Then we got to differentiate the inside of that expression. And when we differentiate the inside of that expression, we get e to the minus alpha. AUDIENCE: Dr. Winston-- PATRICK WINSTON: Yeah? AUDIENCE: That should be 1 plus. PATRICK WINSTON: Oh, sorry, thank you. That was one of those fatal mistakes you just prevented. So that's 1 plus. That's 1 plus here too. OK, so we've differentiated that. We've turned that into a minus 2. We brought the minus sign outside. Then we're differentiating the inside. The derivative and the exponential is an exponential. Then we got to differentiate that guy. And that just helps us get rid of the minus sign we introduced. So that's the derivative. I'm not sure how much that helps except that I'm going to perform a parlor trick here and rewrite that expression thusly. We want to say that's going to be e to the minus alpha over 1 plus e to the minus alpha times 1 over 1 plus e to the minus alpha. That OK? I've got a lot of nodding heads here. So I think I'm on safe ground. But now, I'm going to perform another parlor trick. I am going to add 1, which means I also have to subtract 1. All right? That's legitimate isn't it? So now, I can rewrite this as 1 plus e to the minus alpha over 1 plus e to the minus alpha minus 1 over 1 plus e to the minus alpha times 1 over 1 plus e to the minus alpha. Any high school kid could do that. I think I'm on safe ground. Oh, wait, this is beta. This is beta. AUDIENCE: That's the wrong side. PATRICK WINSTON: Oh, sorry, wrong side. Better make this beta and this 1. Any high school kid could do it. OK, so what we've got then is that this is equal to 1 minus beta times beta. That's the derivative. And that's weird because the derivative of the output with respect to the input is given exclusively in terms of the output. It's strange. It doesn't really matter. But it's a curiosity. And what we get out of this is that partial derivative there-- that's equal to well, the output is p2. No, the output is z. So it's z time 1 minus e. So whenever we see the derivative of one of these sigmoids with respect to its input, we can just write the output times one minus alpha, and we've got it. So that's why it's mathematically convenient. It's mathematically convenient because when we do this differentiation, we get a very simple expression in terms of the output. We get a very simple expression. That's all we really need. So would you like to see a demonstration? It's a demonstration of the world's smallest neural net in action. Where is neural nets? Here we go. So there's our neural net. And what we're going to do is we're going to train it to do absolutely nothing. What we're going to do is train it to make the output the same as the input. Not what I'd call a fantastic leap of intelligence. But let's see what happens. Wow! Nothing's happening. Well, it finally got to the point where the maximum error, not the performance, but the maximum error went below a threshold that I had previously determined. So if you look at the input here and compare that with the desired output on the far right, you see it produces an output, which compared with the desired output, is pretty close. So we can test the other way like so. And we can see that the desired output is pretty close to the actual output in that case too. And it took 694 iterations to get that done. Let's try it again. To 823-- of course, this is all a consequence of just starting off with random weights. By the way, if you started with all the weights being the same, what would happen? Nothing because it would always stay the same. So you've got to put some randomization in in the beginning. So it took a long time. Maybe the problem is our rate constant is too small. So let's crank up the rate counts a little bit and see what happens. That was pretty fast. Let's see if it was a consequence of random chance. Run. No, it's pretty fast there-- 57 iterations-- third try-- 67. So it looks like at my initial rate constant was too small. So if 0.5 was not as good as 5.0, why don't we crank it up to 50 and see what happens. Oh, in this case, 124-- let's try it again. Ah, in this case 117-- so it's actually gotten worse. And not only has it gotten worse. You'll see there's a little a bit of instability showing up as it courses along its way toward a solution. So what it looks like is that if you've got a rate constant that's too small, it takes forever. If you've get a rate constant that's too big, it can of jump too far, as in my diagram which is somewhere underneath the board, you can go all the way across the hill and get to the other side. So you have to be careful about the rate constant. So what you really want to do is you want your rate constant to vary with what is happening as you progress toward an optimal performance. So if your performance is going down when you make the jump, you know you've got a rate constant that's too big. If your performance is going up when you make a jump, maybe you want to increase-- bump it up a little bit until it doesn't look so good. So is that all there is to it? Well, not quite, because this is the world's simplest neural net. And maybe we ought to look at the world's second simplest neural net. Now, let's call this-- well, let's call this x. What we're going to do is we're going to have a second input. And I don't know. Maybe this is screwy. I'm just going to use color coding here to differentiate between the two inputs and the stuff they go through. Maybe I'll call this z2 and this z1 and this x1 and x2. Now, if I do that-- if I've got two inputs and two outputs, then my performance function is going to have two numbers in it-- the two desired values and the two actual values. And I'm going to have two inputs. But it's the same stuff. I just repeat what I did in white, only I make it orange. Oh, but what happens if-- what happens if I do this? Say put little cross connections in there. So these two streams are going to interact. And then there might be some-- this y can go into another multiplier here and go into a summer here. And likewise, this y can go up here and into a multiplier like so. And there are weights all over the place like so. This guy goes up in here. And now what happens? Now, we've got a disaster on our hands, because there are all kinds of paths through this network. And you can imagine that if this was not just two neurons deep, but three neurons deep, what I would find is expressions that look like that. But you could go this way, and then down through, and out here. Or you could go this way and then back up through here. So it looks like there is an exponentially growing number of paths through that network. And so we're back to an exponential blowup. And it won't work. Yeah, it won't work except that we need to let the math sing to us a little bit. And we need to look at the picture. And the reason I turned this guy around was actually because from a point of view of letting the math sing to us, this piece here is the same as this piece here. So part of what we needed to do to calculate the partial derivative with respect to w1 has already been done when we calculated the partial derivative with respect to w2. And not only that, if we calculated the partial wit respect to these green w's at both levels, what we would discover is that sort of repetition occurs over and over again. And now, I'm going to try to give you an intuitive idea of what's going on here rather than just write down the math and salute it. And here's a way to think about it from an intuitive point of view. Whatever happens to this performance function that's back of these p's here, the stuff over there can influence p only by going through, and influence performance only going through this column of p's. And there's a fixed number of those. So it depends on the width, not the depth of the network. So the influence of that stuff back there on p is going to end up going through these guys. And it's going to end up being so that we're going to discover that a lot of what we need to compute in one column has already been computed in the column on the right. So it isn't going to explode exponentially, because the influence-- let me say it one more time. The influences of changes of changes in p on the performance is all we care about when we come back to this part of the network, because this stuff cannot influence the performance except by going through this column of p's. So it's not going to blow up exponentially. We're going to be able to reuse a lot of the computation. So it's the reuse principle. Have we ever seen the reuse principle at work before. Not exactly. But you remember that little business about the extended list? We know that we've seen-- we know we've seen something before. So we can stop computing. It's like that. We're going to be able to reuse the computation. We've already done it to prevent an exponential blowup. By the way, for those of you who know about fast Fourier transform-- same kind of idea-- reuse of partial results. So in the end, what can we say about this stuff? In the end, what we can say is that it's linear in depth. That is to say if we increase the number of layers to so-called depth, then we're going to increase the amount of computation necessary in a linear way, because the computation we need in any column is going to be fixed. What about how it goes with respect to the width? Well, with respect to the width, any neuron here can be connected to any neuron in the next row. So the amount of work we're going to have to do will be proportional to the number of connections. So with respect to width, it's going to be w-squared. But the fact is that in the end, this stuff is readily computed. And this, phenomenally enough, was overlooked for 25 years. So what is it in the end? In the end, it's an extremely simple idea. All great ideas are simple. How come there aren't more of them? Well, because frequently, that simplicity involves finding a couple of tricks and making a couple of observations. So usually, we humans are hardly ever go beyond one trick or one observation. But if you cascade a few together, sometimes something miraculous falls out that looks in retrospect extremely simple. So that's why we got the reuse principle at work-- and our reuse computation. In this case, the miracle was a consequence of two tricks plus an observation. And the overall idea is all great ideas are simple and easy to overlook for a quarter century. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 10_Introduction_to_Learning_Nearest_Neighbors.txt | PROF. PATRICK WINSTON: Well that's the Kodo Drummers. They're a group of about 30 or 40 Japanese people who live in a village on some island off the coast of Japan, and preserve traditional Japanese music. It's an unusual semi communal group. They generally run about 10 kilometers before breakfast, which is served at 5:00 AM. Strange group. Wouldn't miss a concert for the world, although they, alas, don't seem to be coming down to the Boston area very soon. If you go to a concert from the Kodo Drummers--and you should-- and if you're no longer young, you'll want to bring earplugs. Because, as we humans get older the dynamic range control in our inner ear tends to be less effective. So that's why a person of my age might find some piece of music excruciatingly loud, whereas you'll think it's just fine. Because you have better automatic gain control. Just like in any kind of communication device there's a control on how intense the sound gets. Ah, but I go off on a sidebar. Many of you have looked at me in astonishment as I drink my coffee. And you have undoubtedly have been saying to yourself, you know, Winston doesn't look like a professional athlete, but he seemed to have no trouble drinking his coffee. So today's material is going to be pretty easy. So I want to give you the side problem of thinking about how it's possible for somebody to do that. How is it possible? How would you make a computer program that could reach out and drink a cup of coffee, if it wanted a cup of coffee? So that's one puzzle I'd like you to work on. There's another puzzle, too. And that puzzle concerns diet drinks. This is a so-called Diet Coke. Yeah, it's ripe. If you take a Diet Coke and ask yourself, what would a dog think a Diet Coke is for? That's another puzzle that you can work on while we go through the material of the day. So this is our first lecture on learning, and I want to spend a minute or two in the beginning talking about the lay of the land. And then we'll race through some material on nearest neighbor learning. And then we'll finish up with the advertised discussion of sleep. Because I know many of you think that because your MIT students you're pretty tough, and you don't need to sleep and stuff. And we need to address that question before it's too late in the semester to get back on track. All right. So here's the story. Now the way we're going to look at learning is there are two kinds. There's this kind, and there's that kind. And we're going to talk a little bit about both kinds. The kind of the right is learning based on observations of regularity. And computers are particularly good at this stuff. And amongst the things that we'll talk about in connection with regularity based learning are today's topic, which is nearest neighbors. Then a little bit downstream we'll talk about neural nets. And then somewhere near the end of the segment, we'll talk about boosting. And these ideas come from all over the place. In particular, the stuff we're talking about today, nearest neighbors, is the stuff of which the field of pattern recognition-- it's the stuff of which pattern recognition journals are filled. This stuff has been around a long time. Does that mean it's not good? I hope not, because that would mean that everything you learned in 1801 is not good, because the same course was taught 1910. So it has been around a while, but it's extremely useful. And it's the first thing to try when you have a learning problem, because it's the simplest thing. And you always want to try the simplest thing before you try something more complex that you will be less likely to understand. So that's nearest neighbors and pattern recognitions. And the custodians of knowledge about neural nets, well this is sort of an attempt to mimic biology. And I'll cast a lot of calumny on that when we get down there to talk about it. And finally, this is the gift of the theoreticians. So we in AI have invented some stuff, we've borrowed some stuff, we've stolen some stuff, we've championed some stuff, and we've improved some stuff. That's why our discussion of learning will reach around all of these topics. So that's regularity based learning. And you can think of this as the branch of bulldozer computing. Because, when doing these kinds of things, a computer's processing information like a bulldozer processes gravel. Now that's not necessarily a good model for all the kinds of learning that humans do. And after all, learning is one of the things that we think characterizes human intelligence. So if we were to build models of it and understand that we have to go down this other branch, too. And down this other branch we find learning ideas that are based on constraint. And let's call this the human-like side of the picture. And we'll talk about ideas that enable, for example, one-shot learning, where you learn something definite from each experience. And we'll talk about explanation based learning. By the way, do you learn by self explanation? I think so. I had an advisee once, who got nothing but A's and F's. And I said, what are the subjects that you get A's in? And why don't you get A's in all of your subjects? And he said, oh, I get A's in the subjects when I convince myself the material is true. So the learning was a byproduct of self explanation, an important kind of learning. But alas, that's downstream. And what we're going to talk about today is this path through the tree, nearest neighbor learning. And here's how it works, in general. Here's just a general picture of what we're talking about. When you think of pattern recognition, or nearest neighbor based learning, you've got some sort of mechanism that generates a vector of features. So we'll call this the feature detector. And out comes a vector of values. And that vector of values goes into a comparator of some sort. And that comparator compares the feature vector with feature vectors coming from a library of possibilities. And by finding the closest match the comparator determines what some object is. It does recognition. So let me demonstrate that with these electrical covers. Suppose they arrived on an assembly line and some robot wants to sort them. How would it go about doing that? Well it could easily use the nearest neighbor sorting mechanism. So how would that work? Well here's how if would work. You would make some measurements. And it we'll just make some measurements in two dimensions. And one of those measurements might be the total area, including the area of the holes of these electrical covers. Just so you can follow what I'm doing without craning your neck, let me see if I can find the electrical covers. Yes, there they are. So we've got one big blank one, and several others. So we might also measure the hole area. And this one here, this guy here, this big white one has no hole area, and its got the maximum amount of total area. So it will find itself at that point in this space of features. Then we've got the guy here, with room for four sockets in it. That's got the maximum amount of hole area, as well as the maximum amount of area. So it will be right straight up, maybe up here. Then we have, in addition to those two, a blank cover, like this, that's got about 1/2 the total area that any cover can have, so we'll put it right here. And finally, we've got one more of these guys. Oh yes, this one. 1/2 the hole area, and 1/2 the total area. So I don't know, let's see. Where will that go? Maybe about right here. So now our robot is looking on the assembly line and it sees something coming along, and it measures the area. And of course, there's noise. There's manufacturing variability. So it won't be precisely on top of anything. But suppose it's right there. Well it doesn't take any genius human, human or computer, to figure out that this must be one of those guys with maximum area and maximum hole area. But now let's ask some other questions. Where would [TAPPING ON CHALK BOARD], what would that be? Or what would this be? [TAPPING ON CHALK BOARD], and so on. Well we have to figure out what those newly viewed objects are closest to in order to do an identification. But that's easy. We just calculate the distance to all of those standard, platonic, ideal descriptions of things, and we find out which is nearest. But in general, it's a little easier to think about producing some boundaries between these various idealize places, so that we can just say, well which area is the object in? And then we'll know instantaneously to what category it belongs. So if we only had two, like the purple one and the yellow one, it would be easy. Because, we would just construct a line between the two, with a line between the purple and yellow as a perpendicular bisector. And so drawing it out instead of talking about it, if there were only two, that would be the boundary line. Anything south of the dotted line would be purple, and anything north would be yellow. And now we can do this with all the points, right? So we can figure out-- oh could you, Pierre, could you just close the lap top please? So if we want to do this with all these guys it would go something like this-- I better get rid of these dotted x's before they confuse me. Let's see, if these were the only two points, then we would want to construct a perpendicular bisector between the line joining them. And if these two were the only points, I would want to construct this perpendicular bisector. And if these two were the only points, I would want to construct a perpendicular bisector. And if these two points were the only ones involved I'd want to construct-- oh, you see what I'm doing? I'm constructing perpendicular bisectors, and those are exactly the lines that I need in order to divide up this space. And it's going to divide up like this. And I won't say we'll give you a problem like this on an examination, but we have every year in the past ten. To divide up a space and produce-- something we would like to give a name. You know, Rumpelstiltskin effect, when you have a name you get power over it. So we're going to call these decision boundaries. OK so those are the simple decision boundaries, produced in a sample space, by a simple idea. But there is a little bit more to say about this. Because, I've talked about this as if we're trying to identify something. There's another way of thinking about it that's extremely important. And that is this. Suppose I come in with a brand new cover, never before seen. And I only measure, well let's say I only measure the hole area. And the hole area has that value. What is the most likely total area? Well I don't know. But there's a kind of weak principle of, if something is similar in some respects, it's likely to be similar in other respects. So I'm going to guess, if you hold a knife to my throat and back me into a corner, that it's total area is going to be something like that orange cover whole, total area. So this is a contrived example, and I don't make too much of it. But I do want to make a lot of that first principal, over there. And that is the idea that, if something is similar in some respects, it's likely to be similar in other respects. Because that's what most of education is about. Fairy tales, legal cases, medical cases, business cases-- if you can see that there are similar in some respects to a situation you've got now, then it's likely that they're going to be similar in other respects, as well. So when we're learning, we're not just learning to recognize a category, we're learning because we're attempting to apply some kind of precedent. That's the story on that. Well that's a simple idea but does it have any application? The answer is sure. Here's an example. My second example, the example of cell identification. Suppose you have some white blood cells, what might you do? You might measure the total area of the cell. And not the hole area, but maybe the nucleus area. And maybe you might measure four or five other things, and put this thing in a high dimensional space. You can still measure the nearness in a high dimensional space. So you can use the idea to do that. It works pretty well. A friend of mine once started a company based on this idea. He got wiped out, of course, but it wasn't his fault. What happened is that somebody invented a better stain and it became much easier to just do the recognition by brute force. So let's see, that's two examples. the introductory example of the holes of the electrical covers, and the example of cells. And what I want to do now is show you how the idea can reappear in disguised forms in areas where you might not expect to see it. So consider the following problem. You have a collection of articles from magazines. And you're interested in learning something about how to address a particular question. How do you go about finding the articles that are relevant to your question? So this is a puzzle that has been studied for decades by people interested in information retrieval. And here's the simple way to do it. I'm going to illustrate, once again, in just two dimensions. But it has to be applied in many, many dimensions. The idea is you count up the words in the articles in your library, and you compare the word counts to the word counts in your probing question. So you might be interested in 100 words. I'm only going to write two on the board for illustration. So we're going to think about articles from two magazines. Well first of all, what words are we going to use? One word is going to be hack, and that will include all derivatives of hack-- hacker, hacking, and so on. And the other word is going to be computer. And so it would not be surprising for you to see that articles from Wired Magazine might appear in places like this. They would involve lots of uses of the word computer, and lots of uses of the word hack. And now for the sake of illustration, the second magazine from which we are going to draw articles is Town and Country. It's a very tony magazine, and the people who read out Town and Country tend to be social parasites. And they still use the word hack. Because you can talk about hacking, there's some sort of specialize term of art in dealing with horses. So all the Town and Country articles would be likely to be down here somewhere. And maybe they would be one like that when they talk about hiring some computer expert to keep track of the results so the weekly hunt, or something. And now, in you come with your probe. And of course your probe question is going to be relatively small. It's not going to have a lot of words in it. So here's your here's your probe question. Here's your unknown. Which article's going to be closest? Which articles are going to be closest? Well, alas, all those Town and Country articles are closest. So you can't use the nearest neighbor idea, it would seem. Anybody got a suggestion for how we might get out of this dilemma? Yes, Christopher. CHRISTOPHER: If you're looking for word counts and you want to include some terms of computer, then wouldn't you want to use that as a threshold, rather than the nearest neighbor? PROF. PATRICK WINSTON: I don't know, it's a good idea. It might work, who knows. Doug? DOUG: Instead of using decision boundaries that are perpendicular bisectors, if you treated Wired and Town and Country as sort of this like, [INAUDIBLE] targets. And they would look like some [? great radial ?], here. I guess, some radius around curves. If it's within a certain radius then-- PROF. PATRICK WINSTON: Yes? [? SPEAKER 1: Are we, ?] necessarily, have it done with some sort of a [? politidy distance ?] metric? PROF. PATRICK WINSTON: Oh, here we go. We're not going to use any [? politidy distance ?] metric. We're going to use some other metric. SPEAKER 1: Like alogrithmic, or whatnot? PROF. PATRICK WINSTON: Well, algorithmic, gees, I don't know. [LAUGHTER] PROF. PATRICK WINSTON: Let me give you a hint. Let me give you a hint. There are all those articles up there, out there, and out there, just for example. And here are the Town and Country articles. They're out there, and out there, for example. And now our unknown is out there. Anybody got an idea now? Hey Brett, what do you think? BRETT: So you sort of want the ratio. Or in this case, you can take the angle-- PROF. PATRICK WINSTON: Let's be-- ah, there we go, we're getting a little more sophisticated. The angle between what? BRETT: The angle between the vectors. PROF. PATRICK WINSTON: The vectors. Good. So we're going to use a different metric. What we're going to do is, we're going to forget including a distance, and we're going to measure the angle between the vectors. So the angle between the vectors, well let's actually measure the cosine of the angle between the vectors. Let's see how we can calculate that. So we'll take the cosine of the angle between the vectors, we'll call it theta. That's going to be equal to the sum of the unknown values times the article values. Those are just the values in various dimensions. And then we'll divide that by the magnitude of the other vectors. So we'll divide by the magnitude of u, and we'll divide by the magnitude of the art vector to the article. So that's just the dot product right? That's a very fast computation. So with a very fast computation you can see if these things are going to be in the same direction. By the way, if this vector here is actually identical to one of those articles, what will the value be? Well then a cosine will be 0 and we'll get the maximum die of the cosine, which is 1. Yeah, that will do it. So if we use any of the articles to probe the article space, they'll find themselves, which is a good thing to have a mechanism do. OK. So that's just the dot product of those two vectors. And it works like a charm. It's not the most sophisticated way of doing these things. There are hairy ways. You can get a Ph.D. by doing this sort of stuff in some new and sophisticated way. But this is a simple way. It works pretty well. And you don't have to strain yourself, much, to implement it. So that's cool. That's an example where we have a very non-standard metric. Now let's see, what else can we do? How about a robotic arm control? Here we go. We're going to just have a simple arm. And what we want to do is, we want to get this arm to move that ball along some trajectory at a speed, velocity, and acceleration that we have determined. So we've got two problems here. Well let's see, we've got two problems because, first of all, we've got angles, theta 1 and theta 2. It's a 2 degree of 3 of arm, so there are only two angles. So the first problem we have is the kinematic problem of translating the (x,y)-cordinates of the ball, the desired ones, into the theta 1, theta 2 space. That's simple kinematic problem. No f equals ma there. It Doesn't involve forces, or time, or acceleration, anything. Pretty simple. But then we've got the problem of getting it to go along that trajectory with positions, speeds, and accelerations that we desire. And now you say to me, well I've got 801, I can do that. And that's true, you can. Because, it's Newtonian mechanics. All you have to do is solve the equations. There are the equations. Good luck. Why are they so complicated? Well because of the complicated geometry. You notice we've got some products of theta 1 and theta 2 in there, somewhere, I think? You've got theta 2's. I see an acceleration squared. And yeah, there's a theta 1 dot times a theta 2 dot. A velocity times a velocity. Where the hell did that come from? I mean it's supposed to be f equals ma, right? Those are Coriolis forces, because of the complicated geometry. OK. So you hire Berthold Horn, or somebody, to work these equations out for you. And he comes up with something like this. And you try it out and it doesn't work. Why doesn't it work? It's Newtonian mechanics, I said. It doesn't work because we forgot to tell Berthold that there's friction in all the joints. And we forgot to tell him that they've worn a little bit since yesterday. And we forgot that the measurements we make on the lab table are not quite precise. So people try to do this. It just doesn't work. As soon as you get a ball of a different weight you have to start over. It's gross. So I don't know. I can do this sort of thing effortlessly, and I couldn't begin to solve those equations. So let's see. What we're going to do is we're going to forget about the problem for a minute. And we're going to talk about building ourselves a gigantic table. And here's what's going to be on the table. Theta 1, theta 2, theta 3, oops, there are only two. So that's theta 1 again, but it's the velocity, angular velocity. And then we have the accelerations. So we're going to have a big table of these things. And what we're going to do, is we're going to give this arm a childhood. And we're going to write down all the combinations we ever see, every 100 milliseconds, or something. And the arm is just going to wave around like a kid does in the cradle. And then, we're not quite done. Because there are two other things we're going to record. Can you guess what they are? There are going to be the torque on the first motor, and the torque on the second motor. And so now, we've got a whole bunch of those records. The question is, what do we got to do with it? Well here's what we're going to do it. We're going to divide this trajectory that we're hoping to achieve, up into little pieces. And there's a little piece. And in that little piece nothing is going to change much. There's going to be an acceleration, velocity, position. And so we can look those up in the table that we made in the childhood. And we'll look around and find the closest match, and this will be the set of values for the positions, velocities, and accelerations that are associated with that particular movement. And guess what we can do now? We can say, in the past, the torques associated with that particular little piece of movement lie right there. So we can just look it up. Now this method was thought up and rejected, because computers weren't powerful enough. And then, this is the age of recycling, right? So the idea got recycled when computers got strong enough. And it works pretty well, for things like this. But you might say to me, well can it do the stuff that we humans can do, like this? And the answer is, let's look. So this is a training phase, it's going through its childhood. You see what's happening is this. The initial table won't be very good. But that's OK. Because there are only a small number of things that it's important for you to be able to do. So when you try those things it's still writing into the table. So the next time you try that particular motion, it's going to be better at it, because its got better stuff to interpolate [? amongst ?] in that table. So that's why this thing is getting better and better as it goes on. That's as good as I was doing. Pretty good, don't you think? There's just one thing I want to show at the end of this clip just for fun. Maybe you've seen some old Zorro movies? So here's a little set up where this thing has learned to use a lash. So here's the lash, and there's a candle down there. So watch this. Pretty good, don't you think? So how fast does the learning take place? Let me go back to that other slides and show you. So here's some graphs to show you how fast goes, boom. That gives you the curves of how well the robot arm can go along a straight line, after no practice with just some stuff recorded in the memory. And then with a couple of practice runs do give it better values amongst which to interpolate. So I think that's pretty cool. So simple, but yet so effective. But you still might say, well, I don't know, it might be something that can be done in special cases. I wonder if old Winston uses something like that when he drinks his coffee? Well we' ought to do the numbers and see if it's possible. But I don't want to use coffee, it's the baseball season. We're approaching the World Series. We might as well talk about professional athletes. So let's suppose that this is a baseball pitcher. And I want to know how much memory I'll need to record a whole lot of pitches. Is there a good pitcher these days? The Red Socks suck so I don't do Red Socks. Clay Buchholz, I guess. I don't know, some pitcher. And what we're going to do, is we're going to say for each of these little segments were going to record 100 bytes per joint. And we've got joints all over the place. I don't know how many are involved in doing a baseball pitch, but let's just say we have had 100 joints. And then we have to divide the pitch up into a bunch of segments. So let's just say for sake of argument that there are 100 segments. And how many pitches does a pitcher throw in a day? What? SPEAKER 2: In a day? PROF. PATRICK WINSTON: In a day, yeah. This, we all know, is about 100. Everybody knows that they take them out after about 100 pitches. So what I want to know is how much memory we need to record all the pitches a pitcher pitches in his career. So we still have to work on this little bit more. How many days a year does a pitcher pitch? Well, they've got winter ball, and that sort of thing, so let's just approximate it as 100. I don't know, some of these may be a little high, some of the others may be a low. And of course, the career-- just to make things easy-- is 100 years. So that's one, two, three, four, five, six. So we have 10 to the 12th bytes. Is that the hopelessly big to store in here? CHRISTOPHER: 10 to 100 [INAUDIBLE] or just 100 times throwing? PROF. PATRICK WINSTON: 100 pitches in a day-- Christopher's asking some detail-- and what we're gong to do is we're going to record everything there is to know about one pitch, and then we're going to see how many pitches, he pitches in his lifetime. And we're going to record all that. Trust me. Trust me. OK. so we want to know if this is actually a practical scale. And this, by the way, is cocktail conversation, who knows, right? But it's useful to work out these numbers, and know some of these numbers. So the question we have to ask is, how much computation is in there? And the first question relevant to that is, how many neurons do we have in our brain? Volunteer? Neuroscience? No one to volunteer? All right. Well this is a number you should know, because this is what you've got in there. There are 10 to the 10th neurons in the brain, of which 10 to the 11th are in the cerebellum, alone. What the devil do I mean by that? I mean that your cerebellum is so full of neurons that it dwarfs the rest of the brain. So if you exclude the cerebellum, you've got about 10 to 10th neurons. And there about 10 to the 11th neurons in the cerebellum, alone. What's the cerebellum for? Motor control. Interesting. So we're a little short. Oh, but we forget, that's just the number of neurons. We have to count up the number of synapses. Because conceivably, we might be able to adjust those synapses, right? So how many synapses does a neuron have? The answer is, it depends. But the ones in the cerebellum-- I should be pointing back there, I guess-- 10 to the 5th. So if we add all that up we have 10 to the 16th. No problem. It's just that existence proves that you don't have to worry too much about having storage. So maybe our cerebellum functions, in some way, as a gigantic table. And that's maybe how we learn motor skills, by filling up that table as we run around emerging from the cradle, learning how to manipulate ourselves as we go on. So that's the story on arm control. Now all this is pretty straightforward, easy to understand. And of course, there are some problems. Problem number one, what if the space of samples looks like this? [TAPPING ON CHALK BOARD] What's going to happen in that case? Well what's going to happen in that case is that the-- let's see, which values are going to be more important? The x values, right? The y values are spread out all over the place. So you'd like the spread of the data to sort of be the same in all the dimensions. So is there anything we can do to arrange for that to be true? Sure, we can just normalize the data. So we can borrow from our statistics course and say, well, let's see, we're interested in x. And we know that the variance of x is equal to 1 over n times the sum of the values, minus the mean value squared. That's a measure of how much the data spreads out. So now, instead of using x, we can use x prime, which is equal to x over sigma. What's the variance of that going to be? x over sigma sub x. Anybody see, instantaneously, what the variance of that's going be? Or do we have to work it out? It's going to be 1, Work out the algebra for me. It's obvious, it's simple. Just substitute x prime into this formula for variance, and do the algebraic high school manipulation. And you'll see that the variance turns out not to be of this new variable, this transformed variable you want. So that problem, the non uniformity problem, the spread problem, is easy to handle. What about that other problem? No cake without flour? What if it turns out that the data-- you have two dimensions and the answer, actually, doesn't depend on y at all. What will happen? Then you're often going to get screwy results, because it'll be measuring a distance that is merely confusing the answer. So problem number two is the what matters problem. Write it down, what matters. Problem number three is, what if the answer doesn't depend on the data at all? Then you've got the trying to build a cake without flour. Once somebody asked me-- a classmate of mine, who went on to become an important executive in an important credit card company-- asked me if we could use artificial intelligence to determine when somebody was going to go bankrupt? And the answer was, no. Because the data available was data that was independent of that question. So he was trying to make a cake without flour, and you can't do that. So that concludes what I want to say about nearest neighbors. No I want to talk a little bit about sleep. Over there on that left-side branch, now disappeared, we talked about the human side of learning. And I said something about one-shot, an escalation based learning. And what that means is, you don't learn without problem solving. And the question is, how is problem solving related to how much sleep you get? And to answer questions like that, of course, you want to go to the people who are the custodians of the kind of knowledge you are interested in. And so you would say, who are the custodians of knowledge about how much sleep you need? And what happens if you don't get it? And the answer is the United States Army. Because they're extremely interested in what happens when you cross 10 or 12 times zones, and have no sleep, and have to perform. So they're very interested in that question. And they got even more interested after the first Gulf War, which was the most studied war in history, up to that time. Because, there were after action reports they were full of examples like this. The US Forces, in a certain part of the battlefield, and drawn up for the night. And those are Bradley fighting vehicles, there, and back here Abrams tanks. And they're all just kind of settling down for good night's sleep. They've been up for about 36 hours straight, by the way. When, much to their amazement, across their field-of-view came a column of Iraqi vehicles. And both sides were enormously surprised. A firefight broke out. The lead vehicle, over here, on the Iraqi side caught on fire. So these guys, in the Bradley fighting vehicles, went around to investigate, whereupon, these guys started blasting away, in acts of fratricidal fire. And the interesting thing is that all these folks here swore in the after action reports that they were firing straight ahead. And what happened was their ability to put ordnance on target was not impaired at all. But their idea of where the target was, what the target was, whether it was a target, was all screwed up. So this led to a lot of experiments in which people were sleep deprived. And by the way, you think you're a tough MIT student, right? These are Army Rangers. It doesn't get any tougher than this, believe me. So here's one of the experiments that was performed. In those days they had what they called fire control teams. And their job is to take information from an observer, over here, about a target, over here. And tell the artillery, over here, where to fire. So they kept some of these folks up for 36 hours straight. And after 36 hours they all said, we're doing great. And at that time they were bringing fire down on hospitals, mosques, churches, schools, and themselves. Because, they couldn't do the calculations anymore, after 36 hours without sleep. And now you say to me, well I'm a MIT student, I want to see the data. So let's have a look at the data. OK. So there it goes. That's what happens to you after 72 hours without sleep. These are simple things to do. Very simple calculations you have to do in your head, like adding numbers, spelling words, and things like that. So after 72 hours without sleep, your performance relative to what you were at the beginning is about 30%. So loss of sleep destroys ability. [BELL RINGING] Sleep loss accumulates. So you say, well I need eight hours of sleep-- and what you need, by the way, varies-- but I'm going to get by was seven hours of sleep. So after 20 days of one hour's worth of sleep deprivation, you're down about 25%. If you say, well I need eight hours of sleep, but I'm going to have to get by with just six, after 20 days of that, you're down to about 25% of your original capability. So you might say, well does caffeine help? Or naps, naps in this case. And the answer is, yes, a little bit. Some people argue that you get the more affect out of the sleep that you do get if you divide it into two. Winston Churchill always took a three hour nap in the afternoon. He said that way he got a day and a half's worth of work out of every day. He got the full amount of sleep. But he divided it into two pieces. Here's the caffeine one. So caffeine does help. And now you say, well, shoot, I think I'm going to take it kind of easy this semester. And I'll just work hard during the week before finals. Maybe I won't even bother sleeping for the 24 hours before the 6034 final. That's OK. Well let's see what will happen. So let's work the numbers. Here is 24 hours. And that's where your effectiveness is after 24 hours. Now let's go over to the same amount of effectiveness on the blood alcohol curve. And it's about the level at which you would be legally drunk. So I guess what we ought to do is to check everybody as they come in for the 6034 final, and arrest you if you've been 24 hours without sleep. And not let you take any finals again, for a year. So if you do all that, you might as well get drunk. And now we have one thing left to do today. And that is address the original question of, why it is that the dogs and cats in the world think that the diet drink makes people fat? What's the answer? It's because only fat guys like me drink this crap. So since the dogs and cats don't have the ability to tell themselves stories, don't have that capacity to string together events into narratives, they don't have any way of saying, well this is a consequence of desiring not to be fat. Not a consequence of being fat. They don't have that story. And so what they're doing is something you have to be very careful about. And that thing you have to be very careful about is the confusion of correlation with cause. They see the correlation, but they don't understand the cause, so that's why they make a mistake. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 14_Learning_Sparse_Spaces_Phonology.txt | PATRICK WINSTON: So today we're gonna talk about a few miracles of learning in the context of the theme that we're developing here in the class. We started off with a discussion of some basic methods. We talked about nearest neighbors. And we talked about identification trees. And those are kind of basic things that have been around for a long time. Still useful. Still the right things to do when you're faced with a learning problem and you're not sure what method to try. Then we went on to talk about some naive biological mimicry. We talked about neural nets. And we talked about genetic algorithms. And you look at those things and you think and reflect back on what we talked about. And you have to say to yourself, are these nugatory ideas? Perhaps pistareens? Or are they supererogatory ideas that deserve to be center stage? Does anybody know what those words mean? A pistareen? Well, a pistareen is a Spanish coin. It was so small. It was of little worth. These ideas like neural nets, genetic algorithms, I classify them as pistareens because getting them to do something is rather like getting a dog to walk on its hind legs. You can make it happen, but they never do it very well. And you have to think it took a lot of trickery and training to make it happen. So not too personally high on those ideas. But we teach them to you anyway because, of course, we only editorialize part of time and part of time we like to cover what's in the field. Today we're starting a couple of discussions of mechanisms or ideas or things to know about that are quite different. Because now we're going to focus on the problem rather than on the mechanism. And then a later on we're going to talk about deep theory, FIOS, for its own sake. But this week I want to talk about mechanisms that were devised. I want to talk about research that was done. Let me not say mechanisms. Let me say research that was done to attempt an account of some of the things that we humans do well. Sometimes without even knowing that we do it. Now Krishna here tells me his first language was Telugu. Telugu. I once had another student whose first language was Telugu. I said to him, that must be one of those obscure Indian languages. And he said, yes. It's spoken by 56 million people. French is spoken by 52. [LAUGHTER] PATRICK WINSTON: He's going to be our experimental subject. Krishna, if I pluralize words-- you know what it means to pluralize a word. So if I say for example, horse, then if I ask you for the plural you'll say horses. So if I say dog, what's the plural? STUDENT: Then dogs. Or in my language? PATRICK WINSTON: No, no, no. In English. STUDENT: Oh, dogs. PATRICK WINSTON: Well, what about cat? STUDENT: Cats. PATRICK WINSTON: And he got it right. Isn't that a miracle? When did you start speaking English? STUDENT: Second grade. PATRICK WINSTON: Second grade. But he still got it right. But he never learned that he's actually pluralizing those words differently. But he is. So when you pluralize dog, what's the sound that comes after? It's a z sound. Zzzzzz. Dogzzz. If you stick your fingers up here you can probably feel your vocal cords vibrating. If you stick a piece of paper in front of your mouth you'll see it vibrate. But when you say cats, the pluralizing sound is sss, like that. No vocalizing. No vibration of the vocal cords. And old Krishna here learned that rule, as did all of you other non-native speakers of English, effortlessly and without noticing it. You learned it. Buy you always get it right. How can that possibly be? Well, be the end of hour you'll know how that might be. And you'll experience a case study in how questions of that sort can be approached with a sort of engineering point of view. You can say, what if God were an engineer? Or alternatively, what if I were God and I am an engineer? Think about how it might happen that way. So we want to understand how it might be that the machine could learn rules like that. Phonological rules. Not just that one, but all the phonological rules you'd acquire in a course on phonology. That part of speaking that deals with those syllabic and sub-syllabic sounds. The phones of the language. So when Yip and Sussman undertook to solve this engineering problem, both being dedicated engineers, the first thing they did was learn the science. So they went to sit at the foot of Morris Halle, who would develop-- was largely responsible for the development theories of so-called distinctive features. And here's how all that works. You start off with a person who wants to say something. And out that person's mouth comes some sort of acoustic pressure wave. And if I say, hello, George. And you say hello, George. Everybody will understand that we said the same thing. But that acoustic waveform won't look anything alike. It'll be very different for all of us. So it's a miracle that words can be understood. In any case, it goes into an ear. And it's processed. And out comes a sequence of distinctive feature. Vectors. A distinctive feature is a binary variable like is the phone voices or not. That is to say, are your vocal cords vibrating when you say it? If so, then that's plus voiced. If not, it's minus voiced. So according to the original distinctive feature theory and consistent with most of the theories that have been derived since the original one, there are on the order of 14 of these distinctive features that determine which phone you're saying. So if you say ah, that's one combination of these binary features. If you say tuh, that's another combination of these binary features. 14 of them. So how many sounds does that mean, in principle, there could be in a language? SEBASTIAN: 2 to the 14th. PATRICK WINSTON: And what's 2 the 14th, Sebastian? Well, it ought to be about 16,000, don't you think? 2 to the 10th is 1,000. 2 the fourth is 16. So there are about 16,000 possible combination. But no language on Earth has more than 100 phones. That's strange, isn't it? Because some of those choices are probably excluded on physical ground. But most of them are not. So we could have a lot more phones in our language than we actually do. English is about 40. So the sequence of distinctive features could be viewed as then producing meaning after, perhaps, a long series of operations. But in the end, those operations feedback in here because many of the distinctive features are actually hallucinated. We think we heard them, but they're not there. Or they're not even in the acoustic waveform. They're there for the convenience of the phonologist who make rules out of them. It's remarkable how much of this feedback there is, and even injection from other modalities. Many of you may have heard about the McGurk Effect. Here's who the McGurk Effect works. Look at me while I say ga, ga, ga, ga, ga, ga. OK. I said, g-a. Now how about ba, ba, ba, ba. OK. I said ba like a sheep. But if I take the sound I make when I say ba and play it while you're taking video of me saying ga, what do you think you hear? You don't hear ba. Some people report that they hear a d-a sound like da. When I look at it, I can't make any sense out of it. It looks like there's a disconnection between the speech and the video. But it does not sound like ba. But if I shut my eyes and say ba, ba, it's absolutely clear that it's b-a. So what you see has a large influence on what you hear. It's also interesting-- although a side issue-- it's also interesting to note that it's very difficult pronounced things correctly if you don't see the speaker. So many people wonder when they learn foreign languages why they can't speak like a native. And the answer is, they're not watching the mouth of the speaker. I was talking to a German friend once and said, you know, I just can't say the damned umlaut right. And he said, oh, the trouble with you Americans is you don't realize that American cows say moo but German cows say muu. [LAUGHTER] PATRICK WINSTON: And, of course, I got instantly because I could see that the umlaut sounds are produced with protruding lips, which we don't have any sounds an English that require that. Ah, but back to what we know from the phonologists about all this stuff. If you talk to Morris Halle, he will tell you that over here-- I like to think of it as a marionette. There are five pieces of meat down here. And the combination of distinctive features that you're trying to utter are like the control of a marionette on those five pieces of meat. So if you want to say an a sound, the marionette control goes into a position that produces that combination. So let's see. What does that distinctive feature sequence look like for typical word? Well, here's a word. A-e-p-l. Apples. And we can talk about what distinctive features are arrayed in that particular combination of phones. So one of the features that they like to talk about is syllabic. Syllabic. That roughly means, can that sound form the sort of core of a syllable? And the answer is a can, buy these can't. So it's plus, minus, minus, minus. Down here a little ways you'll run into the voiced feature. And for the voiced feature, well, we can do the experiment ourselves. Ahh. Sounds like it's voices to me. Pa. No. That's not voiced. Oo. Yep. Zzz. We already said that was voiced. So that's the combination you see when you utter apples for the voiced feature. Then another one is the continuent one. That roughly says is your vocal apparatus open? Is there no obstruction? And so ahh plus pa is constricted. Oo, open. Zzz, open. So that one happens to run right along with voiced in that particular word. Oh, and there are 14 altogether. But let me just write down one more. The strident one. That says, do you use your tongue to form a little jet of air? So you don't on aa, pa, oo. Buy you do on z. So that gets a plus. So that's a glimpse through a soda straw of what it would like to represent the word apples as a set of distinctive features all arranged in a sequence. So it's a matrix of features. Going down in the columns we have our distinctive features. And going across we have time. So as the first thing Sussman and Yip did in their effort to understand how phonological rules could be learned is to design a machine that would interpret words and sounds and things that you see so as to produce the sounds of the language. So they imagined the following kind of machine. The machine has some kind of mystery apparatus over here that looks out into the world and sees what's there. So I'm looking out in the world and I see two apples. So what this machine might do then is, at some point, decide that there are two apples out there. Then, thinking in terms of these guys as computer engineers, they think in terms of a set of registers that hold values for concepts like noun and verb and plural. And we've not done anything with the machine yet. We've provided no input. So those registers are all empty. Then, up in here, we have a set of words. And they're all kinds of words. Apple is one of them. And those words up there know about how the concept is rendered as a sequence of a phones, that is to say a sequence of distinct features. Then, over here, most importantly, they have a set of constraints. So we'll talk about a particular constrain, the plural constraint. Plural constraint number one. And it's going to reach around and connect itself to some other parts of the machine. Finally, there's a buffer of phones to be uttered. And they're going to flow out this way to the speaker's mouth and get translated into a acoustic wave form. So those are the elements of the machine. Now how are the elements connected together? Well, the words are connected, of course, into the buffer that is used to generate the sound over here on the far left. The plural register is connected to what you see in the world. What you see in the world is connected not only to plural register, but to all of the objects in the word repertoire. This plural constraint here deserves extra attention because it's going to be desirous of actuating itself in the event but the thing observed in the world is plural. There are lots of them. So it's going to be connected then to the plural port. There's going to be a z sound port down here connecting to that file element in the buffer. And finally, over here is going to be a plussed voiced port, which is going to be connected to the second phoneme in the sequence. That's how the machine is going to be arranged. An of course, this is just one of many constraints. But it's a constraint that has a very peculiar property. Information can flow through it in multiple ways. So we think of most programs as having an input and an output. But I try to be careful to draw circles here instead of arrows. Because these are ports and information can flow in any direction along them. What I want to do now is to show you how this machine would react if I suddenly present it with a pair of apples like so. So the assumption is that the vision apparatus comes in and produces the notion, the concept, of two apples. So once that has happened-- that's operation number one-- then information flows from that meaning register up here to the apple word. So that's part of stage number two. Another part of stage number two is information flows along this wire and marks that as plus plural. So operation number one is the activity of the vision system. Activity number two is the flow of information from that vision system into the word lexicon and into this plural register. So far so good. Here's activity number three. This word is also connected to the registers. And information flows along those wires so as to indicate that it's a noun but not a verb. That's part of part number three. At the same time, part number three, information flows down this wire and writes a-p-l into those are elements of the buffer. Now this constraint up here, this box, says, well, I can now see some stuff in that buffer that wasn't there before. So it says, do I see enough stuff on my ports to get excited about expressing values on other ports? Well, let's see. What has it got? It's got the elements in this buffer. Also up here in step three flow the plural thing. So it know that the word is plural. So it says, is this voiced? P is pa. That's not voiced. Is this a z sound. No, that's not as z sound. So it sees what it likes on only one of its three ports. So it says, I'm not going to do anything. I'm [INAUDIBLE]. I'm not in this particular combat. So far so good. What happens next? What happens next is that some time passes. And the elements of the buffer flow to the left toward the speaker's mouth. So we get an a, p, l. Same as we had before, but shifted over. Now what happens? Now what happens is that the l is now in the penultimate position. So information flows up here. Item number four-- oh, I guess that's item number five. Item number four is the leftward flow of the word. So in phase number five, the p is witnessed by this constraint. p is-- sorry, l is witnessed by this constraint. We moved it over one. L is lll. L is voiced. So we have some flow up here like that. That's number five. Now we have voiced and we have plural. And we have nothing here. So there's a great desire of this buffer to have something written into it. So now there's a flow down in there, of z, as item number six. So that's how the machine would work in expressing the idea that there are apples in the field of view. Mmm. Real apples. Not plastic imitations. So that's how the machine works. But all those connections are reversible. So if I hear apples then I get the machine running backwards and my visual apparatus can imagine that there are apples out there. That's how it works. That's just by way of background the machine that they could see it for using the phonological rules once they're learned. All the phonological rules are expressed in these constraints. But since these constraints are such that information can flow in any direction, they deserve to be called propagators. And in the good old days when everyone took 6.001, they learned about propagators as a kind of architecture for building complex systems. But in any event, there's the Sussman-Yip machine. And now comes the big question. How do you learn rule rules like that? Well, what we need is we need some positive examples and some negative examples. And for the simple classroom example I've chosen the same challenge that I presented to Krishna. We're gonna have cats and dogs. So we're gonna look at the distinctive features that are associated with those words. Syllabic. Voiced. Continuent. And strident. Just four of the 14 features that are associated with each of the sounds on those words. Could you close the laptop, please? Just for the distinctive features that are arrayed in those words by way of illustration. So here we have k-a-t-z. Phonetically spelled. And if we work that out, let's see. What is syllabic? That's not. That is. That is. That's not. Voiced? Ka. Nope. Ah. Yep. T. Nope. Z. Yes. That can't be right. Cats. I misspelled it. Because cats. Sss. His a hissing sound but there's no voicing. So that's not as z sound. That's an s sound. So that's not plus voiced. It's minused voiced. Continuent. Let's see. Is my mouth open when I say k? No. Ah? Yes. T? No. S? Yes. And strident. Minus, minus, minus, plus. It's only with the s sound that I have that kind of jet forming with my tongue. Now we can look at dogs. And now we have the z sound as the pluralization. We know that because when we say it, dogzz. Yep. There it comes out as a-- we're only gonna look at the last two columns because they're the only ones that are going to matter to us. So that's plus. And that's minus. Gu, gu, gu, gu. That's plussed. And that's plussed. They're both voiced. Is that right? Dogu? Gu. Gu. Is g sound voiced? Yeah, I didn't think so. G sound is voiced? Look-- oh. Oh, it is voiced buy it's not a continuent. Just like that. Yeah. Cat, dogu zz. Yeah. It is voiced. And it has to be for my example to work out. And that's minus, minus, minus, plus. So what we're interested in is, how come one word gets an s sound and how come the other words gets a z sound? Well, it's a pretty sparse space out there. We've already decided that there are 14,000 possible phonemes and there are only 40 in the language. So that's one thing we can consider. The other thing that we can think is that, well, maybe this is a logical problem. Like the kind of problem you'd face if you were designing a computer. And so Sussman and Yip got stuck for three months thinking about the problem that way. Couldn't make any progress whatsoever. And that happens a lot when you're doing a search. You think you've got a way of approaching it. Try to make it work. You stay up all night. Stay up all night again. Still can't make it work. Eventually, you abandon ship and try something else. So then they began to say, well, let's see. All we care about is the stuff before the two ending sounds. We care about that part of the matrix. And we care about that part of the matrix. And we can ask, in what ways are those things different? And they're different all over the place. That's why they're different words. We can ask the question a little bit differently. And we can say, what can we not care about? And still retain enough of an understanding of how the words are different so as to put the proper plural ending on them. And they worried about that for a long time. Couldn't find a solution. The search space was too big. And then they said, maybe what we ought to do is we ought to think about generalizing this guy here so that we don't care about it. So now we don't care about that guy. And then he went down through here saying, well, let's see when we have to stop generalizing. Because we've screwed everything up and we can no longer keep the z sound words separated from the s sound words. So that eventually distilled itself down to the following algorithm. First thing they did was to collect positive and negative examples. And there's a positive example and a negative example. That's not enough to do it right. But that's enough to illustrate the idea. So the next thing they did was something that's extremely common in learning anything. And that is to pick a positive example to start from. It's actually not a bad idea in learning anything to start with a positive example. So they picked a positive example and they called that a seed. So in our particular case, cats is going to be our seed. And the question we're going to ask is, what are the words that get pluralized like cat? So we've got a positive and negative example. We've picked a seed. And now, the next step is to generalize. And what I mean by generalize is you pick some places in the phoneme matrix that you just don't care about. So you may pick a positive example. And you don't care about it. So you change it to an asterisk or, as demonstrated in the program I'm about show you, a ball. Or you pick one that's negative and you turn it to a ball. Bo. So cats, this seed, becomes a pattern. And in order to pluralize the word this way, you have to match all the stuff in here. But now what we're going to do is we're going to gradually turn some of those elements into don't care symbols until we get to a point where we've not cared about so much stuff that we think that we pluralize that one with an s sound too. So we keep generalizing until we cover, that is to say we admit or match, a negative example. So that's how it works. So we generalize like crazy. And as soon as we cover a negative example, we quit. Otherwise, we just go back up here and generalize some more. And now we've got to pick a search technique to decide which of these guys to actually generalize when. We could pick one at random. And they tried that. It didn't work. So what they decided is that the thing that influences the pluralization most is the adjacent phoneme. And if that isn't the thing that solves the problem, it'll be the one next to that. So in other words, the closer you are, the more likely you are to determine the outcome. So these guys over here are least likely to matter. And those are the ones that are generalized first. So if we do that, what happens? Looks like we're going to come in here and see that there's a big difference between the non-voiced t and the voiced g. But that's only a guess because I've only shown you a fraction of the 14 distinctive features that are involved. So I suppose you like to see a demonstration. Yeah. So there's our 14 features. And that's our seed there, sitting prominently in the display with pluses and minuses indicating the values of the distinctive features for all three of the phones involved. That funny left bracket isn't a mistake. That's just one convention for rendering the ah sound in cat. So it's pretty hard to tell from just that matrix what's going to be the determining feature that separates the positive examples from the negative examples. You notice that there are actually two examples down here. There's cat and duck. Is ducks got an s sound? Ducks? Yep. So dogs and ducks. They both get pluralized with an s sound. And then we have beach doesn't. That's beaches. Dog. We know that's a z. Gun. Gunz. So that's not in the group. So we can run this experiment. Now here we go. We're generalizing like crazy. Generalizing, generalizing, generalizing from left to right. So nothing in the first two columns matters. Now we get to the t. Wow. There it is. So it looks like you pluralize with a s sound. The sss. If, and only if, you're not voiced and you're not strident in the second to the last-- in the last phone of the word that you're trying to pluralize. So that's one phonological rule that the system has learned. And guess what? It's the same rule that's found in phonological textbooks. So now we can try another experiment. So this time we're trying to deal with dog and gun. And our negatives are what was previously positive plus beach, which is still in there as a negative example. So let's see how that one works. Nothing matters except for the last column, the last phone. And now we find out that if the last sound is voiced, then the pluralization gets the z sound, a voiced determinator. And finally, just to deal with beaches. That's beach in it's funny phonetic spelling. So now, if the final sound in the word is strident, if its got this jetty sound-- beach. Beach. Then it gets the ea sound. So let's go back to experiment number one. Because I want to point out one small thing about the way this works. You'll notice that it talks about coverage and excluded down here in the lower left-hand corner. Excluded, well, there are three negative examples, so they better all be excluded. You don't want to cover any of the negatives. But it says coverage, two and two. That's because it actually is doing-- and now we have the vocabulary to say it quickly-- it's doing a beam search through this space. So it's not just doing a depth first search. It's doing a beam search so as to reduce the possibility of overlooking a solution. So it says, oh, the coverage. Both of the beam search elements cover both of the positive examples. And they, in fact, have converged to the same solution. So that's how the Sussman and Yip thing worked. And then the next question to ask is, of course, why did it work? And so the answer, as articulated by Sussman and Yip-- or rather more by Sussman. Or rather more by Yip and a little bit less by Sussman. Yip thinks that it worked because it's a sparse space. And when you have a high dimensional sparse space, it's easy to put a hyperplane into the space to separate one set of examples for another set of examples. So let's consider the following situation. Suppose we have a one-dimensional situation. And we have two white examples and we have two purple examples. Well, too bad for us you can't separate them. Now suppose that this is actually the projection of a two-dimensional space that looks like this. Here are the white examples down here. And here are the purple examples up here. Now it's easy to see that you can separate them with just a line that goes across like that. Now let's take this one more step and suppose that this is actually a projection of a three-dimensional space. It looks like this. This will be dimension one. This'll be two going back there. And this will be three up here. And suppose that the positive examples are right here on this line. Let's say this is-- well, we're gonna draw a little old cube like so. Those are purple examples that are up there. How many ways are there of partitioning the space along those axes? Well, now they're not even just two. They're three. So one way to separate the purple from the white is to draw a hyperplane-- or in this case it's a three dimension, so a plane-- through here on the number three axis. You could also put a plane in on that axis. Or you could do both. So in one case your dividing line would be-- let's see. On the first axis that would be 1/2. And then the don't care. Don't care. Another solution that would be don't care. And then we divide on the number 2 axis with a plane at 1/2 and don't care. Or we could do it with 1/2, 1/2, and don't care. So the higher the dimension of the space, the easier it is sometimes to put in a plane that separates the data. That's why Sussman and Yip think that we use so little of possible phoneme space. Because it makes the thing learnable. That's one possibility. So one explanation for sparse space is learnability. There's another interesting possibility, and that is that if you have a sparse space, high dimensional space with 14 dimensions, and if the 40 points of your language are spread evenly throughout that space-- now let me say it the other way. If they are placed at random in that space, then according to the central limit theorem, then they'll be about equally distant from each other. So it ensures that the phonemes are easily separated when you speak. But if you go to ask a linguist if that's true, they don't know. Because they're not looking at it from a computational point of view. Well, we can look at it from a computational point of view. So I did that. After Sussman and Yip published their paper. And here's the result. This is a diagram that shows all of the phonemes that are separated by exactly one distinctive feature. So if you look over in this corner here, you'll see that the constants-- w and x-- are separated by exactly one distinctive feature. So they're not exactly distant from each other in the space. On the other hand, they are pretty easy to separate relative to the vowels. Which are here in this part of the diagram. Which are all tangled up and the vowels are all close to each other. So guess what? Vowels are much harder to separate than constants. Not surprisingly, because there are many pairs of them that are different. And only one distinctive feature. All right. So now you back up and you say, well, gosh. That's all been sort of interesting. But what does it teach us about how to do science and stuff? And what it teaches us is-- this is an example. Ow. This is an example which we can use to illuminate some of thoughts of David Marr, who I spoke of in a previous lecture, connection with vision. But here's Marr's catechism. I can't spell very well so I won't try to respell it. But this is Marr's catechism. So what Marr said is, when you're dealing with an AI problem, first thing to do is to specify the problem. Gee, that sounds awfully normal. The next thing is to devise a representation suited to the problem. The third thing to do, vocabulary varies, but it's something like determine an approach. Sometimes thought of as a method. And then four, pick a mechanism or devise an algorithm. And, finally, five, experiment. And of course, it never goes linearly like that. You start with the problem and then you go through a lot of loops up here. Sometimes even changing the problem. But that's just the scientific method, right? You start with the problem and you end up with the experiment. But that's not what people in AI, over the bulk of its existence, have tended to do. What they tended to do is to fall in love with particular mechanisms. And then they attempt to apply those mechanisms to every problem. So you might say, well, gee, neural nets are so cool. I think all of human intelligence can be explained with a suitable neural net. That's not the right way to do it. Because that's mechanism envy. You fall in love with mechanism. You try to apply it where it isn't the right thing. This is example starting with the problem and bringing to the problem the right representations, gosh, distinctive features. Once we've got the right representation, then the constraints emerge, which enable us to devise an approach, write an algorithm, and do an experiment. As they did. So this Sussman-Yip thing is an example of doing AI stuff in a way that's congruent with the Marr's catechism. Which I highly recommend. They could have come in here and said, well, we're devotees of the idea of neural nets. Let's see if we can make a machine that will properly pluralize words using a neural net. That's a loser. Because it doesn't match the problem to the mechanism. It tries to force fit the mechanism into some Procrustean bed where it doesn't actually work very well. So what this leaves open, of course, is the question of, well, what is a good representation? And here's the other half Marr's catechism. Characteristic number one is that it makes the right things explicit. So in this particular case, it makes distinctive features explicit. Another thing that Marr was noted for was stereo vision. So in that particular world, discontinuities in the image, when you go across an edge with the things that were made explicit. Once you've got to a representation that makes the right things explicit, you can say, does it also expose constraint? And if you have a representation that exposes constraint, then you're off and running. Because it's constraint that you need in order to do the processing that leads to a solution. So don't have the right representation. If it doesn't expose constraints, you're not going to be able to make a very good model out of it. And finally, there's a kind of localness criteria. If you have a representation in which you can see the right answer by looking at descriptions through soda straw, that's probably a better representation than one that's all spread out. It's true with programs, right? If you can see how they work by looking through a soda straw, you're in much better situation to understand something if you have to look here and there and on the next page and in the next file. So all this is basically common sense. But this is kind of common sense that makes you smarter as an engineer and scientist. Especially as a scientist because if you go into a problem with mechanism envy, you're apt to study mechanisms in a naive way and never reach a solution that will be satisfactory. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 17_Learning_Boosting.txt | PATRICK WINSTON: We've now almost completed our journey. This will be it for talking about several kinds of learning-- the venerable kind, that's the nearest neighbors and identification tree types of learning. Still useful, still the right thing to do if there's no reason not to do the simple thing. Then we have the biologically-inspired approaches. Neural nets. All kinds of problems with local maxima and overfitting and oscillation, if you get the rate constant too big. Genetic algorithms. Like neural nets, both are very naive in their attempt to mimic nature. So maybe they work on a class of problems. They surely do each have a class of problems for which they're good. But as a general purpose first resort, I don't recommend it. But now the theorists have come out and done some things are very remarkable. And in the end, you have to say, wow, these are such powerful ideas. I wonder if nature has discovered them, too? Is there good engineering in the brain, based on good science? Or given the nature of evolution, is it just random junk that is the best ways for doing anything? Who knows? But today, we're going to talk about an idea that I'll bet is in there somewhere, because it's easy to implement, and it's extremely powerful in what it does, and it's the essential item in anybody's repertoire of learning mechanisms. It's also a mechanism which, if you understand only by formula, you will never be able to work the problems on the quiz, that's for sure. Because on the surface, it looks like it'd be very complicated to simulate this approach. But once you understand how it works and look at a little bit of the math and let it sing songs to you, it turns out to be extremely easy. So it's about letting multiple methods work in your behalf. So far, we've been talking about using just one method to do something. And what we're going to do now is we're looking to see if a crowd can be smarter than the individuals in the crowd. But before we get too far down that abstract path, let me just say that the whole works has to do with classification, and binary classification. Am I holding a piece of chalk in my hand, or a hand grenade? Is that a cup of coffee or tea? Those are binary classification problems. And so we're going to be talking today strictly about binary classification. We're not going to be talking about finding the right letter in the alphabet that's written on the page. That's a 26-way choice. We're talking about binary choices. So we assume that there's a set of classifiers that we can draw on. Here's one-- h. And it produces either a minus 1 or a plus 1. So that's how the classification is done. If it's coffee, plus 1. If it's tea, minus 1. Is this chalk, plus one. If it's a hand grenade, minus 1. So that's how the classification works. Now, too bad for us, normally the world doesn't give us very good classifiers. So if we look at the error rate of this classifier or any other classifier, that error rate will range from 0 to 1 in terms of the fraction of the cases got wrong on a sample set. So you'd like your error rate to be way down here. You're dead if it's over there. But what about in the middle? What if it's, say, right there. Just a little bit better than flipping a coin. If it's just a little bit better than flipping a coin, that's a weak classifier. And the question is, can you make a classifier that's way over here, like there, a strong classifier, by combining several of these weak classifiers, and letting them vote? So how would you do that? You might say, well, let us make a big classifier capital H, that works on some sample x, and has its output produces something that depends on the sum of the outputs of the individual classifiers. So we have H1 working on x. We have H2 working on x. And we have H3 also working on x. Let's say three of them, just to start us off. And now let's add those guys up, and take the sign of the output. So if two out of the three of those guys agree, then we'll get an either plus 1 or minus 1. If all three agree, we'll get plus 1 or minus 1. Because we're just taking the sign. We're just taking the sign of the sum of these guys. So this means that one guy can be wrong, as long as the other two guys are right. But I think it's easier to see how this all works if you think of some space of samples, you say, well, let's let that area here be where H1 is wrong, and this area over here is where H2 is wrong. And then this area over here is where H3 is wrong. So if the situation is like that, then this formula always gives you the right answers on the samples. I'm going to stop saying that right now, because I want to be kind of a background thing on the samples set. We're talking about wrapping this stuff over the sample set. Later on, we'll ask, OK, given that you trained this thing on a sample set, how well does it do on some new examples? Because we want to ask ourselves about overfitting questions. But for now, we just want to look and see if we believe that this arrangement, where each of these H's is producing plus 1 or minus 1, we're adding them up and taking the sign, is that going to give us a better result than the tests individually? And if they look like this when draped over a sample set, then it's clear that we're going to get the right answer every time, because there's no area here where any two of those tests are giving us the wrong answer. So the two that are getting the right answer, in this little circle here for H1, these other two are getting the right answer. So they'll outvote it, and you'll get the right answer every time. But it doesn't have to be that simple. It could look like this. There could be a situation where this is H1, wrong answer. This is H2, wrong answer. And this is H3, wrong answer. And now the situation gets a little bit more murky, because we have to ask ourselves whether that area where three out of the three get it wrong is sufficiently big so as to be worse than 1 of the individual tests. So if you look at that Venn diagram, and stare at it long enough, and try some things, you can say, well, there is no case where this will give a worse answer. Or, you might end up with the conclusion that there are cases where we can arrange those circles such that the voting scheme will give an answer that's worst than an individual test, but I'm not going to tell you the answer, because I think we'll make that a quiz question. Good idea? OK. So we'll make that a quiz question. So that looks like a good idea. And we can construct a little algorithm that will help us pick the particular weak classifiers to plug in here. We've got a whole bag of classifiers. We've got H1, we've got H2, we've got H55. We've got a lot of them we can choose from. So what we're going to do is we're going to use the data, undisturbed, to produce H1. We're just going to try all the tests on the data and see which one gives us the smallest error rate. And that's the good guy, so we're going to use that. Then we're going to use the data with an exaggeration of H1 errors. In other words-- this is a critical idea. What we're going to do is we're going to run this algorithm again, but instead of just looking at the number of samples that are got wrong, what we're going to do is we're going to look at a distorted set of samples, where the ones we're not doing well on has exaggerated effect on the result. So we're going to weight them or multiply them, or do something so that we're going to pay more attention to the samples on which H1 produces an error, and that's going to give us H2. And then we're going to do it one more time, because we've got three things to go with here in this particular little exploratory scheme. And this time, we're going to have an exaggeration of those samples-- which samples are we going to exaggerate now? We might as well look for the ones where H1 gives us a different answer from H2, because we want to be on the good guy's side. So we can say we're going to exaggerate those samples four which H1 gives us a different result from H2. And that's going to give us H3. All right. So we can think of this whole works here as part one of a multi-part idea. So let's see. I don't know, what might be step two? Well, this is a good idea. Then what we've got that we can easily derive from that is a little tree looked like this. And we can say that H of x depends on H1, H2, and H3. But now, if that that's a good idea, and that gives a better answer than any of the individual tests, maybe we can make this idea a little bit recursive, and say, well, maybe H1 is actually not an atomic test. But maybe it's the vote of three other tests. So you can make a tree structure that looks like this. So this is H11, H12, H13, and then 3 here. And then this will be H31, H32, H33. And so that's a sort of get out the vote idea. We're trying to get a whole bunch of individual tests into the act. So I guess the reason this wasn't discovered until about '10 years ago was because you've got to get so many of these desks all lined up before the idea gets through that long filter of ideas. So that's the only idea number two of quite a few. Well, next thing we might think is, well, we keep talking about these classifiers. What kind of classifiers are we talking about? I've got-- oh, shoot, I've spent my last nickel. I don't have a coin to flip. But that's one classifier, right? The trouble with that classifier is it's a weak classifier, because it gives me a 50/50 chance of being right. I guess there are conditions in which a coin flip is better than a-- it is a weak classifier. If the two outcomes are not equally probable, than a coin flip is a perfectly good weak classifier. But what we're going to do is we're going to think in terms of a different set of classifiers. And we're going to call them decision tree. Now, you remember decision trees, right? But we're not going to build decision trees. We're going to use decision tree stumps. So if we have a two-dimensional space that looks like this, then a decision tree stump is a single test. It's not a complete tree that will divide up the samples into homogeneous groups. It's just what you can do with one test. So each possible test is a classifier. How many tests do we get out of that? 12, right? Yeah. It doesn't look like 12 to me, either. But here's how you get to 12. One decision tree test you can stick in there would be that test right there. And that would be a complete decision tree stump. But, of course, you can also put in this one. That would be another decision tree stump. Now, for this one on the right, I could say, everything on the right is a minus. Or, I could say, everything on the right is a plus. It would happen to be wrong, but it's a valid test with a valid outcome. So that's how we double the number of test that we have lines for. And you know what? can even have a kind of test out here that says everything is plus, or everything is wrong. So for each dimension, the number of decision tree stumps is the number of lines I can put in times 2. And then I've got two dimensions here, that's how I got to twelve. So there are three lines. I can have the pluses on either the left or the right side. So that's six. And then I've got two dimensions, so that gives me 12. So that's the decision tree stump idea. And here are the other decision tree boundaries, obviously just like that. So that's one way can generate a batch of tests to try out with this idea of using a lot of tests to help you get the job done. STUDENT: Couldn't you also have a decision tree on the right side? PATRICK WINSTON: The question is, can you also have a test on the right side? See, this is just a stand-in for saying, everything's plus or everything's minus. So it doesn't matter where you put the line. It can be on the right side, or the left side, or the bottom, or the top. Or you don't have to put the line anywhere. It's just an extra test, an additional to the ones you put between the samples. So this whole idea of boosting, the main idea of the day. Does it depend on using decision tree stumps? The answer is no. Do not be confused. You can use boosting with any kind of classifier. so why do I use decision tree stumps today? Because it makes my life easy. We can look at it, we can see what it's doing. But we could put bunch of neural nets in there. We could put a bunch of real decision trees in there. We could put a bunch of nearest neighbor things in there. The boosting idea doesn't care. I just used these decision tree stumps because I and everybody else use them for illustration. All right. We're making progress. Now, what's the error rate for any these tests and lines we drew? Well, I guess it'll be the error rate is equal to the sum of 1 over n-- That's the total number of points, the number of samples-- summed over the cases where we are wrong. So gee, we're going to work on combining some of these ideas. And we've got this notion of exaggeration. At some stage in what we're doing here, we're going to want to be able to exaggerate the effect of some errors relative to other errors. So one thing we can do is we can assume, or we can stipulate, or we can assert that each of these samples has a weight associated with it. That's W1, this is W2, and that's W3. And in the beginning, there's no reason to suppose that any one of these is more or less important than any of the other. So in the beginning, W sub i at time [? stub ?] one is equal to 1 over n. So the error is just adding up the number of samples that were got wrong. And that'll be the fraction of samples to that you didn't get right. And that will be the error rate. So what we want to do is we want to say, instead of using this as the error rate for all time, what we want to do is we want to move that over, and say that the error rate is equal to the sum over the things you got wrong in the current step, times the weights of those that were got wrong. So in step one, everything's got the same weight, it doesn't matter. But if we find a way to change their weights going downstream-- so as to, for example, highly exaggerate that third sample, then W3 will go up relative to W1 and W2. The one thing we want to be sure of is there is no matter how we adjust the weights, that the sum of the weights over the whole space is equal to 1. So in other words, we want to choose the weights so that they emphasize some of the samples, but we also want to put a constraint on the weights such that all of them added together is summing to one. And we'll say that that enforces a distribution. A distribution is a set of weights that sum to one. Well, that's just a nice idea. So we're make a little progress. We've got this idea that we can add some plus/minus 1 classifiers together, you get a better classifier. We got some idea about how to do that. It occurs to us that maybe we want to get a lot of classifiers into the act somehow or another. And maybe we want to think about using decision tree stumps so as to ground out thinking about all this stuff. So the next step is to say, well, how actually should we combine this stuff? And you will find, in the literature libraries, full of papers that do stuff like that. And that was state of the art for quite a few years. But then people began to say, well, maybe we can build up this classifier, H of x, in multiple steps and get a lot of classifiers into the act. So maybe we can say that the classifier is the sign of H-- that's the one we picked first. That's the classifier we picked first. That's looking at samples. And then we've got H2. And then we've got H3. And then we've got how many other classifiers we might want, or how many classifiers we might need in order to correctly classify everything in our sample set. So people began to think about whether there might be an algorithm that would develop a classifier that way, one step at a time. That's why I put that step number in the exponent, because we're picking this one at first, then we're expanding it to have two, and then we're expanding it to have three, and so on. And each of those individual classifiers are separately looking at the sample. But of course, it would be natural to suppose that just adding things up wouldn't be enough. And it's not. So it isn't too hard to invent the next idea, which is to modify this thing just a little bit by doing what? It looks almost like a scoring polynomial, doesn't it? So what would we do to tart this up a little bit? STUDENT: [INAUDIBLE]. PATRICK WINSTON: Come again? Do what? STUDENT: [INAUDIBLE]. PATRICK WINSTON: Somewhere out there someone's murmuring. STUDENT: Add-- PATRICK WINSTON: Add weights! STUDENT: --weights. Yeah. PATRICK WINSTON: Excellent. Good idea. So what we're going to do is we're going to have alphas associated with each of these classifiers, and we're going to determine if somebody can build that kind formula to do the job. So maybe I ought to modify this gold star idea before I get too far downstream. And we're not going to treat everybody in a crowd equally. We're going to wait some of the opinions more than others. And by the way, they're all going to make errors in different parts of the space. So maybe it's not the wisdom of even a weighted crowd, but a crowd of experts. Each of which is good at different parts of the space. So anyhow, we've got this formula, and there are a few things that one can say turn out. But first, let's write down the an algorithm for what this ought to look like. Before I run out of space, I think I'll exploit the right hand board here, and put the overall algorithm right here. So we're going to start out by letting of all the weights at time 1 be equal to 1 over n. That's just saying that they're all equal in the beginning, and they're equal to 1 over n. And n is the number of samples. And then, when I've got that, I want to compute alpha, somehow. Let's see. No, I don't want to do that. I want to I want to pick a classifier the minimizes the error rate. And then m, i, zes, error at time t. And that's going to be at time t. And we're going to come back in here. That's why we put a step index in there. So once we've picked a classifier that produces an error rate, then we can use the error rate to determine the alpha. So I want the alpha over here. That'll be sort of a byproduct of picking that test. And with all that stuff in hand, maybe that will be enough to calculate Wt plus 1. So we're going to use that classifier that we just picked to get some revised weights, and then we're going to go around that loop until this classifier produces a perfect set of conclusions on all the sample data. So that's going to be our overall strategy. Maybe we've got, if we're going to number these things, that's the fourth big idea. And this arrangement here is the fifth big idea. Then we've got the sixth big idea. And the sixth big idea says this. Suppose that the weight on it ith sample at time t plus 1 is equal to the weight at time t on that same sample, divided by some normalizing factor, times e to the minus alpha at time t, times h at time t, times some function y which is a function of x, But not a function of time. Now you say, where did this come from? And the answer is, it did not spring from the heart of mathematician in the first 10 minutes that he looked at this problem. In fact, when I asked [INAUDIBLE] how this worked, he said, well, he was thinking about this on the couch every Saturday for about a year, and his wife was getting pretty sore, but he finally found it and saved their marriage. So where does stuff like this come from? Really, it comes from knowing a lot of mathematics, and seeing a lot of situations, and knowing that something like this might be mathematically convenient. Something like this might be mathematically convenient. But we've got to back up a little and let it sing to us. What's y? We saw y last time. The support vector machines. That's just a function. That's plus 1 or minus 1, depending on whether the output ought to be plus 1 or minus 1. So if this guy is giving the correct answer, and the correct answer is plus, and then this guy will be plus 1 too, because it always gives you the correct answer. So in that case, where this guy is giving the right answer, these will have the same sign, so that will be a plus 1 combination. On the other hand, if that guy's giving the wrong answer, you're going to get a minus 1 out of that combination. So it's true even if the right answer should be minus, right? So if the right answer should be minus, and this is plus, then this will be minus 1, and the whole combination well give you minus 1 again. In other words, the y just flips the sign if you've got the wrong answer, no matter whether the wrong answer is plus 1 or minus 1. These alphas-- shoot, those are the same alphas that are in this formula up here, somehow. And then that z, what's that for? Well, if you just look at the previous weights, and its exponential function to produce these W's for the next generation, that's not going to be a distribution, because they won't sum up to 1. So what this thing here, this z is, that's a sort of normalizer. And that makes that whole combination of new weights add up to 1. So it's whatever you got by adding up all those guys, and then dividing by that number. Well, phew. I don't know. Now there's some it-turns-out-thats. We're going to imagine that somebody's done the same sort of thing we did to the support vector machines. We're going to find a way to minimize the error. And the error we're going to minimize is the error produced by that whole thing up there in 4. We're going to minimize the error of that entire expression as we go along. And what we discover when we do the appropriate differentiations and stuff-- you know, that's what we do in calculus-- what we discover is that you get minimum error for the whole thing if alpha is equal to 1 minus the error rate at time t, divided by the error rate at time t. Now let's take the logarithm of that, and multiply it by half. And that's what [INAUDIBLE] was struggling to find. But we haven't quite got it right. And so let me add this in separate chunks, so we don't get confused about this. It's a bound on that expression up there. It's a bound on the error rate produced by that expression. So interestingly enough, this means that the error rate can actually go up as you add terms to this formula. all you know is that the error rate is going to be bounded by an exponentially decaying function. So it's eventually guaranteed to converge on zero. So it's a minimal error bound. It turns out to be exponential. Well, there it is. We're done. Would you like to see a demonstration? Yeah, OK. Because you look at that, and you say, well, how could anything like that possibly work? And the answer is, surprisingly enough, here's what happens. There's a simple little example. So that's the first test chosen. the greens are pluses and the reds are minuses, so it's still got an error. Still got an error-- boom. There, in two steps. It now has-- we can look in the upper right hand corner-- we see its used three classifiers, and we see that one of those classifiers says that everybody belongs to a particular class, three different weights. And the error rate has converged to 0. So let's look at a couple of other ones. Here is the one I use for debugging this thing. We'll let that run. See how fast it is? Boom. It converges to getting all the samples right very fast. Here's another one. This is one we gave on an exam a few years back. First test. Oh, I let it run, so it got everything instantaneously right. Let's take that through step at a time. There's the first one, second one. Still got a lot of errors. Ah, the error rate's dropping. And then flattened, flattened, and it goes to 0. Cool, don't you think? But you say to me, bah, who cares about that stuff? Let's try something more interesting. There's one. That was pretty fast, too. Well, there's not too many samples here. So we can try this. So there's an array of pluses and minuses. Boom. You can see how that error rate is bounded by an exponential? So in a bottom graph, you've got the number of classifiers involved, and that goes up to a total, eventually, of 10. You can see how positive or negative each of the classifiers that's added is by looking at this particular tab. And this just shows how they evolve over time. But the progress thing here is the most interesting. And now you say to me, well, how did the machine do that? And it's all right here. We use an alpha that looks like this. And that allows us to compute the new weights. It says we've got a preliminary calculation. We've got to find a z that does the normalization. And we sure better bring our calculator, because we've got, first of all, to calculate the error rate. Then we've got to take its logarithm, divide by 2, plug it into that formula, take the exponent, and that gives us the new weight. And that's how the program works. And if you try that, I guarantee you will flunk the exam. Now, I don't care about my computer. I really don't. It's a slave, and it can calculate these logarithm and exponentials till it turns blue, and I don't care. Because I've got four cores or something, and who cares. Might as well do this, than sit around just burning up heat. But you don't want to do that. So what you want to do is you want to know how to do this sort of thing more expeditiously. So we're going to have to let them the math sing to us a little bit, with a view towards finding better ways of doing this sort of thing. So let's do that. And we're going to run out of space here before long, so let me reclaim as much of this board as I can. So what I'm going to do is I'm going to say, well, now that we've got this formula for alpha that relates alpha t to the error, then I can plug that into this formula up here, number 6. And what I'll get is that the weight of t plus 1 is equal to the weight at t divided by that normalizing factor, multiplied times something that depends on whether it's categorized correctly or not. That's what that y's in their for, right? So we've got a logarithm here, and we got a sign flipper up there in terms of that H of x and y combination. So if the sign of that whole thing at minus alpha and that y H combination turns out to be negative, then we're going to have to flip the numerator and denominator here in this logarithm, right? And oh, by the way, since we've got a half out here, that turns out to be the square root of that term inside the logarithm. So when we carefully do that, what we discover is that it depends on whether it's the right thing or not. But what it turns out to be is something like a multiplier of the square root. Better be careful, here. The square root of what? STUDENT: [INAUDIBLE]. PATRICK WINSTON: Well, let's see. But we have to be careful. So let's suppose that this is 4 things that we get correct. So if we get it correct, then we're going to get the same sign out of H of x and y. We've get a minus sign out there, so we're going to flip the numerator and denominator. So we're going to get the square root of e of t over 1 minus epsilon of t if that's correct. If it's wrong, it'll just be the flip of that. So it'll be the square root of 1 minus the error rate over the error rate. Everybody with me on that? I think that's right. If it's wrong, I'll have to hang myself and wear a paper bag over my head like I did last year. But let's see if we can make this go correctly this time. So now, we've got this guy here, we've got everything plugged in all right, and we know that now this z ought to be selected so that it's equal to the sum of this guy multiplied by these things as appropriate for whether it's correct or not. Because we want, in the end, for all of these w's to add up to 1. So let's see what they add up to without the z there. So what we know is that it must be the case that if we add over the correct ones, we get the square root of the error rate over 1 minus the rate of the Wt plus 1. Plus now we've got the sum of 1 minus the error rate over the error rate, times the sum of the Wi at time t for wrong. So that's what we get if we added all these up without the z. So since everything has to add up to 1, then z ought to be equal to this sum. That looks pretty horrible, until we realize that if we add these guys up over the weights that are wrong, that is the error rate. This is e. So therefore, z is equal the square root of the error rate times 1 minus the error rate. That's the contribution of this term. Now, let's see. What is the sum of the weights over the ones that are correct? Well, that must be 1 minus the error rate. Ah, so this thing gives you the same result as this one. So z is equal to 2 times that. And that's a good thing. Now we are getting somewhere. Because now, it becomes a little bit easier to write some things down. Well, we're way past this, so let's get rid of this. And now we can put some things together. Let me point out what I'm putting together. I've got an expression for z right here. And I've got an expression for the new w's here. So let's put those together and say that w of t plus 1 is equal to w of t. I guess we're going to divide that by 2. And then we've got this square root times that expression. So if we take that correct one, and divide by that one, then the [INAUDIBLE] cancel out, and I get 1 over 1 minus the error rate. That's it. That's correct. And if it's not correct, then it's Wt over 2-- and working through the math-- 1 over epsilon, if wrong. Do we feel like we're making any progress? No. Because we haven't let it sing to us enough yet. So I want to draw your attention to what happens to amateur rock climbers when they're halfway up a difficult cliff. They're usually [INAUDIBLE], sometimes they're not. If they're not, they're scared to death. And every once in a while, as they're just about to fall, they find some little tiny hole to stick a fingernail in, and that keeps them from falling. That's called a thank-god hole. So what I'm about to introduce is the analog of those little places where you can stick your fingernail in. It's the thank-god hole for dealing with boosting problems. So what happens if I add all these [? Wi ?] up for the ones that the classifier where produces a correct answer on? Well, it'll be 1 over 2, and 1 over 1 minus epsilon, times the sum of the Wt for which the answer was correct. What's this sum? Oh! My goddess. 1 minus epsilon. So what I've just discovered is that if I sum new w's over those samples for which I got a correct answer, it's equal to 1/2. And guess what? That means that if I sum them over wrong, it's equal to 1/2 half as well. So that means that I take all of the weight for which I got the right answer with the previous test, and those ways will add up to something. And to get the weights for the next generation, all I have to do is scale them so that they equal half. This was not noticed by the people who developed this stuff. This was noticed by Luis Ortiz, who was a 6.034 instructor a few years ago. The sum of those weights is going to be a scaled version of what they were before. So you take all the weights for which this new classifier-- this one you selected to give you the minimum weight on the re-weighted stuff-- you take the ones that it gives a correct answer for, and you take all of those weights, and you just scale them so they add up to 1/2. So do you have to compute any logarithms? No. Do you have to compute any exponentials? No. Do you have to calculate z? No. Do you have to calculate alpha to get the new weights? No. All you have to do is scale them. And that's a pretty good thank-god hole. So that's thank-god hole number one. Now, for thank-god hole number two, we need to go back and think about the fact that were going to give you problems in probability that involve decision tree stumps. And there are a lot of decision tree stumps that you might have to pick from. So we need a thank-god hole for deciding how to deal with that. Where can I find some room? How about right here. Suppose you've got a space that looks like this. I'm just makings this up at random. So how many-- let's see. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11. How many tests do I have to consider in that dimension? 11. It's 1 plus the number of samples. That would be horrible. I don't know. Do I have actually calculate this one? How could that possibly be better than that one? It's got one more thing wrong. So that one makes sense. The other one doesn't make sense. So in the end, no test that lies between two correctly classified samples will ever be any good. So that one's a good guy, and that one's a good guy. And this one's a bad guy. Bad guy, bad guy bad guy, bad guy. Bad guy, bad guy, bad buy. So the actual number of tests you've got is three. And likewise, in the other dimension-- well, I haven't drawn it so well here, but would this test be a good one? No. That one? No. Actually, I'd better look over here on the right and see what I've got before I draw too many conclusions. Let's look over this, since I don't want to think too hard about what's going on in the other dimension. But the idea is that very few of those tests actually matter. Now, you say to me, there's one last thing. What about overfitting? Because all this does is drape a solution over the samples. And like support vector machines overfit, neural maps overfit, identification trees overfit. Guess what? This doesn't seem to overfit. That's an experimental result for which the literature is confused. It goes back to providing an explanation. So this stuff is tried on all sorts of problems, like handwriting recognition, understanding speech, all sorts of stuff uses boosting. And unlike other methods, for some reason as yet imperfectly understood, it doesn't seem to overfit. But in the end, they leave no stone unturned in 6.034. Every time we do this, we do some additional experiments. So here's a sample that I'll leave you with. Here's a situation in which we have a 10-dimensional space. We've made a fake distribution, and then we put in that boxed outlier. That was just put into the space at random, so it can be viewed as an error point. So now what we're going to do is we're going to see what happens when we run that guy. And sure enough, in 17 steps, it finds a solution. But maybe it's overfit that little guy who's an error. But one thing you can do is you can say, well, all of these classifiers are dividing this space up into chunks, and we can compute the size of the space occupied by any sample. So one thing we can do-- alas, I'll have to get up a new demonstration. One thing we can do, now that this guy's over here, we can switch the volume tab and watch how the volume occupied by that error point evolves as we solve the problem. So look what happens. This is, of course, randomly generated. I'm counting on this working. Never failed before. So it originally starts out as occupying 26% of the total volume. It ends up occupying 1.4 times 10 to the minus 3rd% of the volume. So what tends to happen is that these decision tree stumps tend to wrap themselves so tightly around the error points, there's no room for overfitting, because nothing else will fit in that same volume. So that's why I think that this thing tends to produce solutions which don't overfit. So in conclusion, this is magic. You always want to use it. It'll work with any kind of [? speed ?] of classifiers you want. And you should understand it very thoroughly, because of anything is useful in the subject in dimension learning, this is it. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 5_Search_Optimal_Branch_and_Bound_A.txt | PROFESSOR: It was written about Route 66, which used to be the main highway between Chicago, Illinois and Los Angeles, California. Very famous highway because anybody who wanted to go across country always took route 66 because it was the shortest way to go. And the question is, how do you find the shortest path? Not just any old path or a good path, but how do you find the very shortest path? And that'll be the subject that we're going to discuss today. But route 66, I lament its passing, but it's been largely replaced by the interstate highway system that was created by President Eisenhower. Guess why? Let's see, maybe the ROTC people know. You know why Eisenhower created the interstate highway system? Well, in public affairs, of course, there's always a distinction to be made between the explanation and the reason. The explanation was-- AUDIENCE: To move weapons across the country? PROFESSOR: Well, to move nuclear weapons across the country. Let's put it in, sightly, more benign terms. Eisenhower had observed that the German army was able to move its troops rapidly, even though we bombed their railroads into oblivion because of their auto bond. So Eisenhower conceived that if there were ever an invasion in the United States, we too would want to be able to move our forces around on a highway system. And consequence of that, we have a pretty good highway system and pretty awful railroad system. It's interesting. I'm a beneficiary of that, in a funny way, because I'm from East Peoria, Illinois. And I was surrounded by the factors of Caterpillar Tractor Company, which made all of the tractors that built all of those roads. So my high school spent money like water from that huge tax base of all those factors. Anyway, today we want to find the very best path, instead of just a good path. And like the last time, we'll deal with, both, an example that we can set our program to work on. By the way, can you find the shortest path between S and G? Would you like to bet your life on the shortest path between S and G? Probably, not. With your eye, you can find a good path. But you can't find the best possible path. Today, what we're doing is probably not modeling any obvious property of what we have inside our heads. But being able to find the best path is part of the skill set that anybody who's had a course artificial intelligence would be expected to have. So we're going to look at it, even though it's not like many of the things we do. A model of something that's, probably, going on in your head. So we're going to use, both, this example from Cambridge and our Blackboard example. Let's see, we have to caution ourselves. Tanya, is search about maps? No, it's about what? Starts with a C. And the next letter is H. And it ends up being choice. So we're talking about choice. Not about maps. Even though our examples are drawn from maps because they're convenient, they're visual, and helps understand the concepts behind the algorithms I'm talking about. So let's start off by looking at our classroom example. And I did something today that I neglected to do last time. And that's talk to you about what I meant by heuristic distance. It's those pink lines that I just drew on the map. We're talking about the distance as the crow would fly between two places, even though there's no road that goes between those two places. So in general, and we discussed last time, it's best to get yourself into a place that's close, as the crow flies, to your goal. And of course, that's a heuristic and it can get you in trouble because it's not always true. It would appear that being at node E is a good place to be because it's not very far from G. But in that particular case designed to illustrate the point, being close is, actually, not a good thing because it's a dead end. But in general, it's a good thing to be close. And we talked last time about hill climbing and beam search, being close was the objective of those kinds of searches. And at one point, in a beam search illustration, we had C, B, A, and D. We had paths terminating at all four of those nodes as candidates for the next round of search. And we decided on the basis of these airline distances to keep D and B, and reject A and C because they're further away as the crow flies. Now, I repeat this even though many of you have had this fixed already in your tutorials because we're going to need this concept of heuristic distance today. And I wanted to be sure that that point has been clarified. So now, with this smaller map I imagine you can do, by eye, a determination of what the shortest path is. What is it, Juana? Can you help me out with that? AUDIENCE: S, A, D, G. PROFESSOR: S, A, D, G. And if you add up those distances, the distance is 11 along that path that goes from S, first to A, and then to D, and then from D to G. So Juana asserts that that is the best path. And we're going to treat Juana as an Oracle because we're going to follow, in our initial attempt to understand these algorithms, a very important principle of problem solving. And that is that, if you want to solve a problem, the easiest way is, usually, ask somebody who knows the answer. Or Google, which also, probably, knows the answer. So in this particular case, we believe that Juana knows the answer. And she said that the shortest path is S,A, D, G, and its path length is 11. But we don't trust her because we're applying to the same medical school and she may be trying to screw us. [LAUGHTER] PROFESSOR: So we're going to be very cautious about accepting her answer until we've checked it to make sure that she hasn't attempted to delude us. So how do we go about doing that? Well, one way to do that is to check to be sure that all other possible paths that we could develop end up being, for sure, longer than the one that Juana has told us about. So she's told us about S, A, D, G. And it has a total path length of 11. And now what I'm going to do is I'm just going to develop the rest of this tree-like diagram. But what I'm going to do is, I'm not going to do it in a British Museum or random way. What I'm going to do is, I'm going to look at the choice that corresponds the shortest path that can be extended. So the shortest path that can be extended is this one right here. The one that just has the starting node in it. And I could have gone this other way to B. And if I go that other way to B, then the path length along that side is 5. And likewise, if I look at the path that terminates in A, that has a path length of 3. So now I've got two choices. A and B. I've got choices that extend beyond those two places. So I'm always going to extend the one that has the shorter length. So in this case, that would be the path that goes from S to A. So if I from S to A, I don't have to go to D. I can also go to B. And if I go to B, then the accumulated path length is S, A, B. That's 7. And know that we're talking now about the path, like the accumulated path length, that we've traveled so far. Last time we were talking a lot about distances to the goal. Heuristic estimates of how far we are from the goal. Now we're doing exactly the opposite. We're not considering how far we've got to go. We're only thinking about how far we've gone so far. So now, repeating these steps again. I've got 7 and 5. So I'll go over and consider the choices that go through the B node on the path S, B. And that gives me S, B, A and S, B, C. And what are those path lengths? Well, let's see. S, B, A would be 9. And S, B,C would be 9. And now the shortest path is this one over here. So I extend that. I go S, A, B. S, A, B. The only place I can go is C. That adds another 4. So that's 11. And what do I know about that path? I don't have to take that any further, right? Because the path length, since I've gone on that path already, is equal to the path length that Juana has told me gets me to the goal. So it'll be foolhardy to carry on because, presuming that these lengths are all non-negative, I can't do any better. And I can't even do as well, unless I've got a length that has 0 length. So now that I have that idea, I can quickly finish up by saying, well, let me consider these two paths. S, B, A can only go to D. And if I go to D, that adds 3. 9 plus 3 is 12. Nothing else can happen there because that's 12 and I've got a path of a goal that's 11. C, I can only go to E. It's a dead-end but I don't have to think about that because I know that the accumulated distance along this path is 6 plus 9. That's 15. So all of these need not be extended any further because their length, accumulated so far, is equal to or less than a length of a goal. So I've checked the Oracle. And although we're applying to the same medical school, Juana has told me the truth. So now, unfortunately, Juana's not always around. And I don't always have an Oracle. So I'm going to have to have some way of finding the shortest path without that Oracle that I can check against. Let's see. What can I do? Maybe I can do the same thing I just did. Always extend the shortest path so far and hope that I run into the goal at some point. And then I have to ask myself the question how much extra work did I need to do when I don't have the Oracle? Let's just try it and see what happens. You don't have that path to start. So I just have S. This distance is 0. I can go either to A or B. If I go to A, I've got a distance of 3. Here, I've got a distance of 5. I'll extend the path that goes S, A. That can either got to B or D. Going to B or D gives me 7 that way. S, A, D gives me 6. So looking across all of these and extending the shortest path so far takes me back over to S, B. So I extend those. S, B takes me to A or C. And those, in turn, have total accumulated path lengths of 9 and 9. Now the shortest one is S, A, D. You see the pattern. Now let's see. I haven't found the goal yet. So I can ask myself the question is any of the work that I've done so far wasted? No because all of the paths that I've got so far are shorter than the path of the goal because the goal hasn't shown up. So when I do my oracle checking after I found the goal, none of that work's going to be wasted. So in the end, I don't, actually, need the Oracle. I could just develop this graph by extending the shortest path, so far, until I hit the goal. And then, perhaps, do a little remaining checking to make sure that all the other paths extend with a length that's greater than the path of the goal. So if those words are confusing, let's carry on with the algorithm, and I think it'll be clearer. So let's see. We've got the 7, 6, and two 9s. We're going to extend the one that's 6. That gets this to the goal. Boom, we've got it. And we've got a path length of 11. Note, though, we can't quit because we have to be sure that all other paths are longer than 11. So now we have to carry on with the same algorithm that we started with. The Oracle checking algorithm. And when we do that, we look for this shortest path, so far, that has not been extended. That's B, S, A, B. That goes to C. That's 11. So we're done there. A goes to D. That adds 3. That's 12. C goes to E. That adds 6. That's 15. And sure enough, we're done. OK? Elliot? AUDIENCE: Does it know that there's know that there isn't a chance that you could have a zero distance extension from the [INAUDIBLE]? PROFESSOR: The question is, does it know that there's no zero distance length that's coming up. That's an implementation detail. This guarantees you'll find a path that's as short as any that you can possibly find. But there might be others if they're zero-length lengths. As long as they're non-negative lengths, we're safe. We've got a shortest path. So that was easy. And now we can repeat the exercise with our more complicated map of Cambridge. First of all, let's do depth first just to recall what that looks like. That is, certainly, not a short path. So let's try this idea, which, by the way, bares the label branch inbound. Let's try branch inbound on the same map. And there it goes. Each of those little flickers is trying another path. So you can see it's working it's guts out to find the shortest path. It's almost there but it's almost a pathological case. Or it's almost doing British Museum. There it's finally found the shortest path. Now there are some things we can ask about that. But first of all, before I ask anything about it, I'd like to get the flow chart up on the board because we're going to decorate that flow chart, a little bit, as we go. So the first thing we do is initialize queue. Then we're going to test first path on the queue. Then we might be happy because we might be done. We might have a shortest path to the goal. Actually, that's not quite true, is it? We can't really quit until every other path is it. Well, that's interesting. If the first element on the queue gets us all the way to the goal, and we sorted our queue by path length, are we through as soon as that first element on the queue gets us to the goal? Yeah because every other path must have been sorted beyond it. And therefore, it can't offer us a shorter path to the goal. So if the first path is a path to the goal we're done. Alas, it usually isn't. So we'll extend first path. We're going to put all those extensions back on the queue, and then we're going to sort them. So that's, pretty much, the same as we did last time. We're always going to put the elements back on the queue. We're going to look at the first element the queue and see if it's a winner. If it is we're done. If it's not, we're going to extend it. And then go back in here and try again. Well, sort of. But we noted that this did a awful lot of work because if we look at those statistics up there, it put 1,354 paths onto the queue. That's the N queueing part. And then it extended 835,000 paths that had come to the front of the queue. Now I'd like to give you an aside because it's easy to get confused about N queueing and extending. In all of the searches we did last time, it would have been perfectly reasonable to keep a list of all the paths that we had put onto the queue. An N queueing list. And never add a path to our queue if it terminates in a node that some other path terminate in that has already gone to the queue. What I said last time was let us keep track of the things that have been extended and not extend them again. So you can either keep track of the nodes that have been extended and not extend them again. Or look at the paths with nodes that terminate, and blah, blah, blah and been put on the queue, the queued ones. And not put things back on the queue again. And I think, last time, I may have put a column in there that said N queued. It should have been extended. Even though N queued worked last time, only extended works this time because we want to be sure that anything we extend is a short path. So the N queued idea doesn't work, at all, for these optimal paths. So now I want to come back over here off the side bar and say that we're keeping track of all of the nodes, all of the paths that end in nodes unless they have already been extended beyond that particular place. So we need to decorate our algorithm here and say test first path and extend the first path if not already extended. Because you can see that in the example I had, so far, we did that same silliness that we talked about last time. We extended paths that went through A more than once, like so. Would it ever make sense to extend this path? No because we've already extended a path that got there with less distance. Will it ever make sense to extend this path? No because we've already extended another path that gets to be by a shorter distance. So if we keep an extended list, we can add that to branch inbound to our advantage. So let's see how that would work on the classroom example. And then we'll do Cambridge. So this is bridge inbound, plus an extended list. And I do mean extended. Not in the N queued list. N queued list won't work here. So let's see, I start off the same way as I did before. S goes to either A or B. That's a length of 3. That's a length of 5. So I extend A. That goes to either B or D. But B is as if it wasn't there at all. Oh, sorry. Hang on. B goes there. And those path lengths are 7 and 6. And now I look around on the board, and I say what is the shortest path so far? And it's B. So I extend that to get to A and C with path lengths of 9 and 9. And what's the shortest one next? It's D. And that goes to G. And the path length is 11. And what's the shortest one on the board? The one that has to be extended next. That's this one that gets to B. But I've already extended a path that get to B. So I don't, actually, do that extension. So I've saved some work. But I've got to go over here and do these two now. But wait. I've already extended B. I've already extended A, so I don't have to do that one either. The only one I have to do is the one that goes to C. And that those then to E with a path length of 15. And I'm done. So if you compare this one with a previous one, you can see that there might be vast areas of this tree that are pruned away and don't have to be examined all. So now, just for the sake of illustrating that, I would like to keep track of just one of those statistics. The number of extensions. So for this particular example, case one, the number of extensions was 835. Why don't you see if you can guess to yourself what it would be if I use this concept of an extended list. See, I'm not going to extend anything I've already extended because it's guaranteed to have a longer path length then something that already got to that same place. So it makes no sense to do it. So let me change the type to branch-and-bound with an extended list. I'm going to turn the speed down a little bit so we can watch it. It might take the rest of the hour. Who knows? Still doing a lot of work. Still examining a lot of paths. Well, look at that. Instead of 835 extensions it only did 38. So that's a pretty substantial savings. And you would never not want to do that. So note that that's a layering on top of branching out. That's not a different algorithm. It's an adjustment improvement to the algorithm, and it makes it more efficient. So this whole thing is based on what I call the dead horse principle. As soon as we figure out that a path that goes to a particular place can't possibly be the winning path, we get rid of it, and don't bother extending it. It's a dead horse principle. But if we look at this example, what's the shortest possible length of a path that's already gone from S to B? What do you think, Tanya? Well, first of all, it can't be less than 5 because we've already gone that distance. So when I say what's the shortest length of any path that there could possibly be that goes from S to D. We know it's at least 5. But can we say something more about it? Especially, when we look at these airline distances, and note that this airline distance is 6, and that's a little more than 7, and that's a little more than 7. So what do you think? So it's gone from S to B, and the question is what's the shortest path that could possibly be that had started out going from S to B? 11 right? Because we can't have a path that's shorter than the airline distance. If there were a straight line road from A to G, its length would be 6. But there isn't. So that gives us a lower bound on the distance that we have along that path. So we're using the accumulated distance, plus the airline distance, to give us a lower bound on the path that we've started off on that goes from S to B. Once again, let's solidify a little bit by simulating the search and seeing how it turns out. Not just I did last time, I'm going to forget that I've got an extended list. I don't want to carry both of those things around with me at the same time. So forget that we've got an extended list. We'll bring all those back together a little later. So we're going to forget what we just did there. And instead we're just going to use this concept of an airline distance and see what happens. As before we start with a starting node. We have two choices as always. We can go to A or B. And the accumulated distance, if we go to A, is 3. And then accumulated distance if we go to B is 5. But now we're going to add in the airline distances. So the airline distance from A to G is a little more than 7, which is 10 plus. The airline distance from B to G is exactly 6. So that gives us 11. And following the procedure we've all been using already so far, we're going to extend the path that seems to have the shortest potential. Now it's the shortest potential distance S to G. So that must be this one here. So from A we can go to B or D. The accumulated distance S, A, B, is 7. The airline distance is 6, so that's equal to 11. Standard arithmetic 13. The distance S, A, D. That is 6 plus a little more than 7. So what's the accumulated distance? S, A, D is 3 plus 3 is 6. AUDIENCE: [INAUDIBLE]. PROFESSOR: What? AUDIENCE: The airline distance from D would be 5. PROFESSOR: Would be 5, right. So airline distance, in this case, is the same as the actual distance. So the accumulated distance is 6. The actual distance is 5. So that's equal to 11. So now I've got two 11's on the board. And simulating what we'd ask you do on a quiz, we don't know which of those is going to be better. They've got a tie score. So what we're going to do is we're going to choose the one that's lexically least. So B comes before D. So we'll expand B. And that can go to either A or C. And we have to calculate the best possible distance that goes along those paths. The accumulated distance S, B, A. S, B, A is 9. So that's 9 plus 7 plus. That's 16 plus. This has an accumulated lead of distance of 9. Also plus 7 plus. Also 16 plus. Well, now let's see. Things are shaping up pretty well because this one has the lowest score so far. We extend that to G. And now the accumulated distance is 11. The airline distance is 0, so that's 11. And that's smaller than everybody else. So we've got. So now compare this one with our branch inbound graph. And you see, once again, we've done considerably less work. And that, in many practical cases, means that instead of taking more than the remaining lifetime of the universe to complete the calculation, it can happen in a few seconds. But let's see how it works on the example. So I'm not going to use the extended list. I'm just going to use this idea of using a lower bound on the distance remaining, the airline distance, and see what happens. So this time, the number of extensions is 70. So it didn't do quite as well as working alone as the extended list did working alone. So we immediately conclude that the extended list is more useful than using one of these lower bound heuristics. By the way, this is called an admissible heuristic. If the heuristic estimate is guaranteed to be less than the actual distance, that's called an admissible heuristic. Admissible because you can use it for this kind of purpose. So it looks like the day extended list is a more useful idea than the admissible idea. Right? What do you think about that, Brett? Am I hacking? Am I joking? AUDIENCE: I think you're judging prematurely. PROFESSOR: Why am I judging prematurely? What do you think it might depend on? AUDIENCE: The fact that we're using extensions and the extended list pretty much guarantees you can only extend each node once. PROFESSOR: Well, Brett has sad something unintelligible that I can't think how to repeat. What he meant to say, though, was that-- [LAUGHTER] PROFESSOR: --in these cases, it almost always depends on the problem itself. If you change the problem, you may get a different result. So why don't we change the problem and see if we get a different result? So instead of starting on the extreme left, let's start in the middle and see what happens. So I'll readjust my starting position to be right there. Oops, that's the wrong adjustment. And we might as well start by getting our baseline branch-and-bound without anything. And for that one, maybe, we'll speed it up a little bit. So that gives us 57 extensions. It's an easier problem. So let's try it with the admissible heuristic. That went too fast. Wow, still pretty fast. Six extensions. What do you think this number's going to be? Closer to six or closer to 57? Better than six? Worse than six? Well, let's think. What we're going to do is we're going to just not repeat any movements through the same node again. But it's not going to do something very important for us. It's not going to keep us out of the left side because it has no idea of the remaining airline distance to the goal. So let's see if that's true It sure is. Look at that. It is, foolishly, spending a lot of its' time doing something we would never do. Namely, looking over there on the left side. So this time, the number of extensions is 35. So in case two, the admissible heuristic does very much better. In case one, the extension thing does much better. But wait a minute, would we ever not want to use both at the same time? We wouldn't want to use just one of these, right? They both have the possibility of doing us a lot of good. So maybe if we put them in harness together, we'll get something that's even better. And when we do that-- see here, the extended list is a layer on top of branch and bound. The admissible heuristic is another layer on top of branch and bound. If we put those together, we get something called A star. So A star is just branch and bound, plus an extended list, plus and admissible heuristic. So let's go back to our original problem and try A star on that. We're running this at a pretty slow speed because we're expecting it to be a lot more efficient than the original branch and bound. And sure enough it is. The number of extensions is 27. So look at that. A lot better than either of those working independently. Now I can stick the thing in the center and see what happens then. So in this particular case, the extended list didn't, actually, help us because our admissible heuristic was channeling us so tightly toward the goal it didn't matter. So it all depends on the nature of the space that you're trying to explore. By the way, you know how the whole works, right? So what you want to do is you want to extend the first path and sort. But not just by accumulated distance. Sort by accumulated distance plus admissible heuristic. But what are the theoreticians? You must be complaining. Sort's expensive. Do we actually need to sort? No, we don't actually need to sort. What do we to do? AUDIENCE: We just need to keep track of what's the minimum. PROFESSOR: We just need to keep track of what's the minimum. So we don't need to, actually, do that sort. That's an unnecessary computation. So instead, we can test, not the first path but the shortest path. And now you have it. Now you have the whole of A star. And now you can go home, but I don't think you should because I'm about to show you that this idea of admissibility, actually, leads to certain screw cases that we're very fond of asking about on exams. So it turns out that the admissible heuristic, in certain circumstances, could get you into trouble. It doesn't look like it could because, logically, nothing I've said seems strange or questionable. But that's because I've been working with a map. And it turns out that if you work with a map then admissibility is a perfectly sound way of doing an optimal search. But, Travis, is search just about maps? No, search is not just about maps. So we may have non-Euclidean arrangements that will cause us trouble. So I'd like to illustrate that with the following example. It's not going to be a large map or a large graph. S, then go up here to A or down here to B. Then they merge at C. And then they go out here to the goal, G. And the actual distances are 1, 1, 1, and 10. And over here, we'll make that 100. So it's a kind of oddly constructed map, but it's there because we need a pathological case to illustrate the idea. Now that's the actual distances. And if we did branch and down with an extended list, everything would work just fine. But we're not. We're going to use an admissible heuristic. And we're going to say that this guy has an estimated distance to the goal of 100. This guy is 0. And this guy is 0. Now, 0 is always an underestimate of the actual distance to the goal, right? So I'm always free to use 0. Is that 100 OK? Yeah because the actual distances is 101, so it's less than that the actual distance. So it's OK as an admissible heuristic. So these numbers that I put up here, together, constitute an admissible heuristic set of estimates to the goal. So now, let's just simulate A star and see what happens. So first of all, you start with S, and that can either go to A or B. The actual distance is 1 plus an estimate on the remaining distance. That gives us 100 plus 100. That's equal to 101. If we go to B instead, the actual distance is 1 plus the heuristic's distance is 0, so that's equal to 1. OK, good. So now we know that we always extend the shortest path so far. Did I goof this, or are you asking a question? AUDIENCE: [INAUDIBLE]? PROFESSOR: Yeah, when I say actual, it's the actual distance that you've traveled. AUDIENCE: But that's [INAUDIBLE]. PROFESSOR: So wait a second. If I go from S to A, the actual distance I've traveled is 1. AUDIENCE: I meant like, does the map-- PROFESSOR: So now I'm taking the sum of the actual distance, plus the estimated distance to go. AUDIENCE: All right. I'm just wondering if the original map has to be [INAUDIBLE]. PROFESSOR: See this is not a map. She was asking if the map has to be geometrically accurate. See, this could be a model of something that's not a map. And so, I'm free to put any numbers on those links that I want, including estimates, as long as they're underestimates of the distance along the lengths. So this tells me that my estimated distance here, so far, is 1. So I'll, surely, go down here to C. And if I go to C, then my accumulated distance is 11. And my estimate of the remaining distance is 0. So that's a total of 11. So now I'm following my heuristic again and saying what's the shortest path on a base of the accumulated distance plus the estimated distance? Here, the accumulated distance plus the estimated distance is 101. Here, it's only 11. So plainly, I extend this guy. And that gets me to the goal. And the total accumulated distance is now 111 plus 0 equals 111. And that's not the shortest path, but wait. I still have to do my checking, right? I have to extend A. I when I extend A, I get to B. And now, when I get to B that way, my accumulated distance is 2 plus my-- oh, sorry. S, A, C. My accumulate distance it 2. My estimated distance is 0, so that's equal to 2. So I'm OK because I'm still going to extend to this guy, right? Wrong. I've already extended that guy. So I'm hosed. I won't find the shortest path because I'm going to stop there. And I'm going to stop there because this is an admissible heuristic and that's not good enough unless it's a map. It's not good enough for this particular case because this is not geometric. This cannot be done as a map on a plane. So that's a situation where what I've talked to you about, so far, works with branch-and-bound. Works with branch-and -bound plus an extended list. But doesn't work when we added an admissible heuristic. So if we're going to do this in general, we need something stronger than admissibility, which works only on maps. And so the flourish that I'll tell you about here in the last few seconds of today's lecture is to add a refinement as follows. So far, we've got admissibility. And if we want to write this down in a kind of mathematical notation, we could say that it's admissible if the estimated distance between any node X and the goal is less than or equal to the actual distance between X and the goal. That's the definition of admissible. As long as heuristic does that it's admissible. And A star works if it's a map. But for that kind of situation where it's not a map we need a stronger condition, which is called consistency. And what that says is that the distance between X and the goal minus the distance between some other node in the goal, Y. Take the absolute value of that. That has to be less than or equal to the actual distance between X and Y. So this heuristic satisfy the consistency condition? Well, let's see. Here the guess is 100. Here it's 0. So the absolute difference is 100. But the actual distance is only 2. So it satisfies admissibility, but it doesn't satisfy consistency. And it doesn't work. So you can almost be guaranteed we'll give you a situation where if you use an admissible heuristic you'll lose. And if you use a consistent heuristic, you'll still win. So how can we bring this back into the fold? Well, we can't use that heuristic. It's no good. But if this heuristic estimate of the goal were 2, then we'd be OK because then it would still be admissible. But it would also be consistent. So the bottom line is that you now know something you didn't know when you started out two lectures ago. You now know how MapQuest and all of its' descendents work. Now you can find an optimal path, as well as a heuristically good path. You see that if you don't do anything other than branch and bound it can be extremely expensive. And you can even invent pathological cases where it's exponential and the distance to the goal. So because it can be so computationally horrible, you want to use every advantage you can, which, generally, involves using an extended list. As well as-- no laptops, please. It still holds. No smoking, no drinking, and no laptops. So you're going to use all the muscles you can. And those muscles include using an extended list and an admissible or consistent heuristic, depending on the circumstances. And so, I think we'll conclude there since our time is up. And Elliot, you can ask a question after class. Why don't you come up and ask it now? |
MIT_6034_Artificial_Intelligence_Fall_2010 | 11_Learning_Identification_Trees_Disorder.txt | PATRICK WINSTON: Ladies and gentlemen, the Romanian national anthem. I did not ask you to stand, because I didn't play it as a symbol of Romanian national identity. But rather, to celebrate the end of the Cold War, which occurred about the time that you were born. Before that, no one came to MIT from Eastern Europe. But since that time, we've been blessed by having in our midst Lithuanians, Estonians, Poles, Czecs, Slovaks, Bulgarians, Romanians, Slovenians, Serbs, and all sorts of people from regions of the world formally excluded to us. Believe me, you are all welcome in our house. Almost all, that is to say. Because you may recall that Romania is the traditional home of vampires. And since the end of the Cold War, vampires have had new vectors for emerging from their traditional places and penetrating into the world at large. You may have vampire in your suite, or on your floor. And it's important to know how to recognize them, and take the necessary precautions. So if you have this concern, I would expect that the first thing you would do would be to look at some data concerning the characteristics of vampires. So there's a little database of samples of individuals who have been determined to be vampires and not vampires. And our task today-- and what you'll understand how to do by the end of the hour-- is to use data like this to build a recognition mechanism that would help you to identify whether someone is a vampire or an ordinary person. So this is a little different from the kind of problem we worked with neural nets. Right? So what's the most conspicuous difference between this data set and anything you could think to work on with nearest neighbors, which we studied last time. Katie, do you have any thoughts about why it would be difficult to use nearest neighbors with data like this? The question mark is there because this is MIT, and a lot of people are completely nocturnal. So you can't tell whether they cast a shadow or not. We want to take that into account. So what's different about this from the electrical cover data set? STUDENT: [INAUDIBLE] PATRICK WINSTON: Could you use the nearest neighbor technique to identify vampires with this data? STUDENT: [INAUDIBLE] PATRICK WINSTON: So obviously-- Yes, Lana? STUDENT: [INAUDIBLE] STUDENT: You cannot really quantify-- PATRICK WINSTON: Oh, that's the problem. This is not numerical data. This is symbolic. So we're not saying that your ability to cast a shadow is 0.7. You either cast a shadow, down cast a shadow, or we can't tell. It's a symbolic result. So problem number one we have to face with data of this kind is that it's not numeric. And there are other characteristics, as well. For example, it's not clear that all of these characteristics actually matter. So some characteristics don't matter. And a corollary to that is that some characteristics do matter, but they only matter part of the time. And finally, there's the matter of cost. Some of these tests may be more expensive to perform than others. For example, if you wanted to determine whether someone casts a shadow, you'd have to go to the trouble of getting up during daylight. That might be an expensive operation for you. You'd have to go find some garlic and ask them to eat it. That might be expensive. So some of these tests might be expensive relative to other tests. But once you realize that we are talking in terms of tests, and not a vector of real values, then what you do is clear. You build yourself a little tree of tests. So who knows how this problem will turn out? But you can imagine a situation where you have one test up here which might have three outcomes. And one but only one of those outcomes might require you to perform another test. And only when you've created the tree of tests that look like this are you finished. So given this set of tests and a set of samples, the question becomes, how do you arrange the tests in a tree like that so as to do the identification that you want to do? So since we're talking about identification, it's not surprising that this kind of tree is called an identification tree. And there's a tendency-- and I may slip into it myself-- to call this a decision tree. But a decision tree is a label for something else. This is an identification tree. And the task is to create a good one. So what is a good one versus a not so good one? What characteristic would you like for a decision tree-- for an identification trade to have, if you're going to call it good identification tree? What do you think, Krishna? What would be a good characteristic? STUDENT: Maybe the minimum number of levels? PATRICK WINSTON: Yeah. He said minimum number of levels. What's another way you could say what a good one is? Each test costs something, right? So what's another way of thinking about what a good tree would look like? STUDENT: Minimum cost. PATRICK WINSTON: The minimum cost. And if they all have the same cost, then it's the number of tests. So overall, what you like is a small tree rather than a big one. So you might be able to take your sample data and divide it up, so that at the bottom of the tree, at the leaves, all of the sets that are produced by the tests are uniform, homogeneous. We'd like that tree to be the simplest possible tree you can find, not some big complicated one that also divides up all the data into uniform subsets. By uniform subset-- at the bottom of the tree, you have all of the vampires together, and all the non-vampires together. So you'd like a small tree. So why not just go all the way and do British Museum, and calculate all possible trees? Well, you can do that, but it's one of those NP problems. And as you know, NP problems suck in general. And so you don't want to do that. You want to have some kind of heuristic mechanism for building a small tree. And we want a small tree because-- Why do we want a small tree? Because of the cost. but there's another, more important reason why we want a small tree. Let me give you a hint. It's Occam's Razor. The simplest explanation is often the best explanation. So if you have a big, complicated explanation, that's probably less good than a simple, small explanation. Occam's Razor. Spelled so many ways it doesn't matter how I spell it. And that's good, because I can't spell. So how are we going to go about finding the best possible arrangement of those four tests in a tree like that? Well, step one will be to see what each test does with the data. And by the way, before I go a step further, you know and I know that this is a sample data set that's very small, suitable for classroom manipulation. You'd never bet your life on a data set this small. We use it only for classroom illustration. But imagine that these rows are multiplied by 10. So instead of eight samples, you've got 80. Then you might begin to believe the results that are produced. So I'm just going to pretend that each one of those represents 10 other samples that I haven't bothered to show. But we can work with this one in the classroom, because it's pretty small. And we can say, well, what does this shadow test do? Well, the shadow test divides the sample population into three groups. There's the I Don't Know group of people who are nocturnal. There are the people who do cast the shadow, the Yes people. And the people who do not cast a shadow, the No people. So if I look at those rows up there and see which ones are vampires, it looks to me that if there's no shadow cast-- there's only one that doesn't cast a shadow-- and that is a vampire. So that's a plus over there. Vampire. Now, if we look at the ones who do cast a shadow, all those are not vampires. They're all OK. And now there're 8. Three are vampires. So that means that two of these must be vampires. And I've got three, four, five, six so far. So there must be two left. So that's the way the shadow test divides up the data. Now let's do garlic. Vampires traditionally don't eat garlic. I don't know why. So we look at the garlic test, and we see that all of the Nos-- well, there're three Yeses, and they all produce a No answer. So if somebody eats garlic, they're not vampires. That means the three vampires must be over here. Then there are two left. So that's what the garlic test does. See what we're trying to do? We're trying to look at all these tests to see which one we like best on the basis of how it divides up the data. So now we've got complexion. And there are three choices for this. You can have an average complexion. But a lot of vampires, in my experience, are rather pale. So pale is a possibility. And then the other option is that just after gorging themselves with blood, they tend to get a little red in the face. So we'll have a ruddy over here. Once again, we have to go back to our data set to see how this test divides things up. So there are three ruddies, and one's a No, one's a No, and one's a Yes. So two Nos and a Yes. Two Nos and a Yes. Now we can try for pale complexion people. There are only two of those. A No and a No. That must mean that there are two pluses over here, because there are three vampires altogether. Two, four, six, seven, eight, nine. Eight, sorry. Eight. Only eight. Just one more to go, and that's the accent. Historically, vampires go to great length to protect their accent and not betray their origins. But nevertheless, we can expect that if they've just arrived-- if they're just in from Transylvania, part of Romania-- they may still have an accent. So there's a normal, some still have a heavy accent, and some persist in having odd accents. So let's see. Accent. Four of them, right at the top, have no accent. Two Nos and a Yes. Heavy accent. Three of those. A Yes and two Nos. That means we must have a plus here. 3, 6, plus and a minus. So we can look at this data and say, well, what will be the best test to use? And the best test to use would surely be the one that produces sets here, at the bottom of the branches, that correspond to the outcomes of the test. We're looking for a test that produces homogeneous groups. So just for the sake of illustration, I'm going to suppose that we're going to judge the quality of the test by how many sample individuals it put into a homogeneous set. So ideally, we'd like a test that will put all the vampires in one group and all the ordinary people in another group right off the bat. But there are no such tests. But we can add up the number of sample individuals who are put in to at least homogeneous sets. So when we do that, this guy has 3 in a homogeneous set here. A fourth. But these are not a homogeneous set. So the overall score for this guy will be 4. This one, well, not quite as good. It only puts 3 individuals in a homogeneous set. This one here, 2 individuals into a homogeneous set. Everybody else is all mixed up with some other kind of person. And over here, how many samples are in a homogeneous set? 0. So on the basis of this analysis, you would conclude that the ordering of the test with respect to their quality is left to right. So the best test must be the shadow test. So let's pick the shadow test first, see what we can do with that. If we pick the shadow test first, then we have this arrangement. We have question mark, and we have Yes, casts a shadow, and No, doesn't. We have 3 minuses here. We have a plus here. And unfortunately, over here, we have plus, plus, minus, minus. So we need another test to divide that group up. Yes. STUDENT: How did you get the 4 on the shadow test again? Why was it 4? PATRICK WINSTON: Well, if I look at the data and I see who-- the question is, what about that shadow test? If you look at the shadow test, and you say, well, there are 4 question marks. And if we look and see what kind of people belong to those 4 question marks, there are 2 vampires and 2 non-vampires. That's why it's 2 pluses and 2 minuses. STUDENT: No, I understand that. The question is, how did you get to the score of 4? PATRICK WINSTON: Oh, yeah. The question is how did I get this number 4? It has nothing to do this, because this is a mixed set. In fact, I've got three guys in a homogeneous set here, and one guy in a homogeneous set here, and I'm just adding them up. STUDENT: OK. PATRICK WINSTON: So very simple classroom illustration. Wouldn't work in practice. Yes. STUDENT: How do you adjust this for larger data sets where it's unlikely you're going to have any [INAUDIBLE]? PATRICK WINSTON: The question is, how do I adjust this for larger data sets? You're one step ahead. Trust me, I'll be doing large data sets in a moment. I just want to get the idea across. And I don't want there to be any thought that the method we use for larger data sets has got anything magic about it. OK, so we're off and running. And now we have to pick a test that will divide those four guys up. So we're going to have to work this a little harder, and repeat the analysis we did there. But at least it'll be simpler, because now we're only considering 4 samples, not 8. Just the 4 samples that we still have to divide up that have come down that left branch. So I have the shadow test. It has 3 outcomes. We have the garlic test. It has 2 outcomes. Yes and No. We have the complexion test. There's 3 outcomes. Average, pale, and ruddy. And we have finally the accent test. And that comes out to be either normal, heavy, or odd. And now, it's a little awkward to figure out what the results are for this data set as shown. So let me just strike out. The ones that we're no longer concerned with, and limit our analysis to the samples for which the outcome of the shadow test is a question mark. This is exactly the four people we still need to separate, right? So switching colors, keeping the color the same. We actually don't want to do the shadow test anymore, right? Because we've already done that. There's no point in doing that again. We don't have to look at that. It's already done all the division of data that it can. So the garlic test. Well, let's see. Garlic. 2 Yeses, 2 Nos. The Yeses produce Nos and the Nos produce Yeses. So if the person does eat garlic, they're OK. And if they don't eat garlic, bad news-- they're vampires. Well, that looks like a pretty good test. But just for the sake of working it all out, let's try the others. Complexion. 2 Ruddies, a Yes, and a No. 1 pale, and that's a No. 1 pale, and that's a No. And we must have 1 average, and sure enough, that's a Yes. Now we can do accent, the one on the far right, and look at how that measures up against the people who are still under consideration as samples. Accent. Let's see. 2 Nones, a Yes and a No. No Heavies. 2 Odds, a Yes and a No. All right. So now we can do the same thing we did before, and just say, for sake of classroom illustration, how many individuals are put into a homogeneous sets. And here we have 4. And here we have 2. And here we have 0. So plainly, the garlic test is the test of choice. So we go back over here, and we've completed the work that we needed to do. So that's the garlic test. And that produces 2 pluses. Let's see. Eats garlic, Yes. Eats garlic, No. I guess the pluses go over here like so. And these are the two ordinary people. And we're done with our task. And now you can quickly run off and put this into your PDA, and forever be protected against the possibility that one of those vampires got out in the flood of people that came in from Eastern Europe. Except what do we do a large data set? Well, the trouble is, a large data set's not likely to produce-- if you have a large data set, no test is likely to put together any homogeneous set right off. So you never get started. Everything would be 0. Every test would say, oh it doesn't put anybody into homogeneous sets. So you're screwed. You need some other, more sophisticated way of measuring how disordered this data is. Or how disordered these sets are that you find at the bottom of the tree branches. That's what you need. You need a way of measuring disorder of these sets that you find at the bottom of these branches, so you can find a kind of overall quality to the test based on your measurement of disorder. Now, the first heuristic of a good life is, when you have a problem to solve, ask somebody who knows the answer. It's the least amount of work. It's not even as hard going to Google. So who would you ask about ways of measuring disorder in sets? There are two possible answers. STUDENT: You could just do entropy. PATRICK WINSTON: What? STUDENT: Find the entropy of the set. PATRICK WINSTON: Who studies entropy? STUDENT: Probability. PATRICK WINSTON: What kind of classes? STUDENT: Physics. STUDENT: Thermodynamics. PATRICK WINSTON: Thermodynamics! The thermodynamicists are good at measuring disorder, because that's what thermodynamics is all about. Entropy increasing over time, and all that sort of stuff. There's another equally good answer. STUDENT: Statisticians? PATRICK WINSTON: Statisticians. Perhaps, but it's not the second best answer. It's actually not even the best answer. That's the best answer. What's your name? STUDENT: Leo. PATRICK WINSTON: Oh, yeah. [LAUGHTER] PATRICK WINSTON: Leonardo has got his finger on it. The information theorists are pretty good at measuring disorder, because that's what information is all about, too. So we might as well borrow a mechanism for measuring the disorder of a set from those information theory guys. So what we're going to do is exactly that. Let's put it over here, so we'll have it handy when we want to try to measure those things. The gospel according to information theorists is that the disorder, D, or some set is equal to-- now let's suppose that this is a set of binary values. So we have positives and then we have negatives. Pluses and minuses. But pluses, they don't go very well in an algebraic equation, because they might be confused with adding. So I'm going to say P and N. And then it'll be the total, which is P plus N. We only have two choices, positive and negative. So the disorder of set, according those guys, is equal to minus the number of positives over the total number, times the log to the base 2 of the positives over the total, minus the negatives over the total, times the log 2 of the negatives over the total. Those negatives look a little worrisome, because you think, well, maybe this thing can go negative. But that's not going to be true, right? Because these ratios are all less than 1, and the logarithm of something that's less than 1 is negative. So we're OK. So that's a lovely way of measuring disorder. And then we ought to draw a graph of what that curve looks like. And what we're going to graph it against is the ratio of positives to the total number. So that's going to be an axis where we go from 0 to 1. So let's just find a couple of useful values. And by the way, it pays to pay attention to these curves, because if you pay attention to this stuff, you can work the quiz questions on this very rapidly. Otherwise, we see people getting out their calculators and quickly becoming both lost and screwed. OK so let's see. Let's suppose that the number of positives is equal to the number of negatives. So we've got a completely mixed-up set. It has no bias in either direction. So in that case, if P over T is equal to 1/2, then this is equal to minus 1/2, times the logarithm of 1/2. And I guess, since they're both the same, we can multiply by two. And what's that value? [INAUDIBLE], what does that calculate out to? STUDENT: Minus [INAUDIBLE] PATRICK WINSTON: Minus [INAUDIBLE]. Well, with a minus sign, you just turn the argument upside down, so it's log(2). So what's log(2)? Logarithm of base 2 of 2? 1! So this whole thing is-- STUDENT: 1. PATRICK WINSTON: 1. So [INAUDIBLE], in her soft way, says, well, let's see. 2 times 1/2. That cancels out. The minus, that flips the arguments so it's log to the base 2 of 2, and that's 1. So this whole thing, You work out the algebra, it gives you 1. So that's cool. So right here in the middle where they're equal, we get a value of 1. Next thing we need to do is let's calculate what happens if P over T is equal to 1. That is to say, everything is a positive. Any guesses? Maybe 10, 20, minus 15? Let's work it out. So if P over T equal 1, that would be minus 1 times the log to the base 2 of 1. What's that? STUDENT: [INAUDIBLE] PATRICK WINSTON: A 0? Oh, yeah. Because 2 raise to the 0 is one. So this part is 0. Now, what about this other part? If everything's a P, then nothing's an N. So we've got 0. And we can quit already. Well, not quite. We ought to work it out. Log 2 to the base 2 of 0. What's that? STUDENT: [INAUDIBLE] PATRICK WINSTON: Who? Minus infinity? Uh oh. 0 times minus infinity is What I didn't get that when I was in high school. Finally, 1801 makes a difference. Finally. What's the answer. We're interested in the limit as N over T goes to 0, right? And when you have a deal like this, what do you do? You use that famous rule, that we all mispronounce when we see it written, right? We use the good old El Hospital's rule. OK, it's L'Hopital. L'Hopital's Rule. You have to differentiate the-- I guess we differentiate this guy as a ratio or something, and see what happens when it goes to 0. And what we get when we use L'Hopital's Rule is that, oh thank God, this is still zero. So now we know that we have a point up there and a point down there. So now we've got three points on the curve, and we can draw it. It goes like that. No, it doesn't go like that. It's obviously a Gaussian, right? Because everything in a nature is a Gaussian. Can you put that laptop away, please? Everything in nature is a Gaussian, so it looks like this. That right? No, actually, not everything in nature is a Gaussian. And in particular, this one isn't a Gaussian either. It looks more like one of those metal things they used to call quonset huts. That's what it looks like. Boom, like so. So that is the curve of interest. Now, did God say that using this way of measuring disorder was the best way? No, Got has not indicated any choice here. We use this because it's a convenient mechanism, it seems to make sense, but in contrast to the reason it's used information theory, it's not the result of some elegant mathematics. It's just a borrowing of something that seems to work pretty well. Any of those curves would work just about the same, because all we're doing with it is measuring how disordered a set is. So one thing to note here is that in this situation, where we're dealing with two choices-- P and N, positives and negatives-- we get a curve that maxes out at one. And notice that it kind of gets up there pretty fast. In fact, if you're down here at 2/3, are you're up here, this is about 0.9. So it gives you a large number for quite a bit of that area in the middle. So that, unfortunately, still doesn't tell us everything we need to know. That tells us how to measure a disorder in one of these sets. But we want to know how to measure the quality of the test overall. So we need some mechanism that says, OK, given that this test produces three different sets, and we now have a measure of the disorder in each of these sets, how do we measure the overall quality of the test? Well, you could just add up the disorder. Let's write that down, because that sounds good. So you can say that the quality of a test is equal to some sum over the sets produced. And what we're going to do is we're going to add up the disorder of each of those sets. I'm almost home, except that this means we're going to give equal weight to a branch that has almost nothing down it-- we're going to give the same weight to that as a branch that has almost everything going down it. So that doesn't seem that make sense. So one final flourish is we're going to weight this sum according to the fraction of the samples that end up down that branch. So it's, as usual, easier to write it down than to say it. So we're going to multiply that times the number of samples in the set, divided by the number of samples handled by test. So if half the samples go down a branch, and if that branch has a certain disorder, then we're going to multiply that disorder times 1/2. All right. So now let's see how it works with our sample problem. Well, here is our sample data. And we didn't need anything fancy for it. But let's pretend it was a large data set. Well, let's see. What would we do? Well, go down this way, there are 4 samples down that direction. That's half of the total number of samples. So whatever we find down there, we're going to multiply by 1/2. This one we're going to multiply by 3/8. And this one we're going to multiply by 1/8. Now, what do we actually find at the bottom of these things? Well, here's a homogeneous set. Everything's the same. So we go to that curve and say, what is the disorder of a homogeneous set? It's zero. Let's see, they're all the same. I guess that means it's 0 over there. So the disorder of this set of three samples is zero. The disorder of this set of one sample, all the same, is zero. The disorder of this set-- well, let's see. Half of the samples there are plus, and half are minus, so we go over to our curve, and we say, what's the disorder of something with equal mixture of pluses and minuses? And that's one. So the disorder of this guy is one. So now we've got 1/2 times 1, and 3/8 times 0, 1/8 times 0. So the quality of this particular test, as determined by the disorder of the sets it produces, is 1/5. 0.5. Let's do this one. So we have 3/8 coming down this way, 5/8 coming down this way. 3/8 is multiplied by the disorder of a set of uniform things. That's disorder 0. So this guy over here, let's see. That's 2/5 and 3/5 multiplied-- You know, this is one of those deals where if you look at the curve, you're pretty close to the middle. And that curve goes all the way up to about 0.9 there. So you can kind of just look at this, and eyeball it, and say, well, whatever it is, the overall, this is going to be something multiplied times 5/8. Something like 0.9 times 5/8. So let's just say, for the sake of discussion, that that's going to be about 0.6, which is within a hundredth, I think, of being right. Just kind of guessing. OK, well now we're on a roll. Here, we have 3/8 coming down this branch, 3/8 coming down this branch, 1/4 coming down this branch. This is 0. And this is one of those deals where these two are about 0.9. So it looks like it's going to be 3/8 plus 3/8 is 3/4. Times about 0.9. So that's going to turn out to be about 0.7. So one last go here. 3/8, 3/8, and 1/4. Oh, that's interesting. Because these two are what we got contributed up to that 0.7. This one is 0.4 times-- this is evenly divided, so that's going to have disorder of 1. So that's going to be 0.25 bigger than the number we got over here. So that's going to end up being about 0.95. So thanks god our answer is the same as we got with our simple classroom measurement of disorder. Except this is measuring how disordered stuff is, we want the small number, not the big number. So once again, based on this analysis, you'll be sure to pick the shadow cast, because 0.5 is less than 0.6, which is less than 0.7, which is less than 0.95. So that accent test is really horrible. Don't use it. Just because somebody has a heavy accent doesn't mean they're a vampire. In fact, most vampires have worked very hard on their accent, as I mentioned before. All right, so now we know that we're still going to pick the shadow test as our first go. So that's good. Now, let's see if we can repeat the exercise with our second selection, the one we have to have to pick those guys apart. And this is going to be easier, because there are fewer things to work with. Ooh, wow, look. That's 0. That's 0. That's 1/2. That's 1/2. So the disorder of this guy is 0.0. So this is 1/4, 1/4, 1/2, 0, 0. 1/2 times 1. Ooh, that's 0.5. That was easy. How about this one? Oh, he says 1. Let's see. That's 1. That's 1. That's 1/2. That's 1/2. Yeah, it is one. So sure enough, the answer also comes out to be the same as before, when we did our just simple intuition exercise. So I don't know. Christopher, is this all about using information theory? STUDENT: No. PATRICK WINSTON: No, no, no. See, it's not about the math. It's about the intuition. And the intuition is that you want to build a tree that's as simple as possible. And you can build a tree that's as simple as possible if you look at the data, and say, well, which test does the best job of splitting things up? Which test does the best job of building subsets underneath it that are as homogeneous as possible? So all this information theory, all this entropy stuff, is just a convenient mechanism for doing something that is intuitionally sound. OK? It's not about information theory. It's about a sound intuition. Oh, by the way. Does this kind of stuff ever get used in practice? 10s of thousands of times. This is a winning mechanism that's used over and over again, even when the data is numeric. How would it work if it's numeric data? Well, let's think about that for a little bit. So let's suppose that we have an opportunity. We're an EMT or something, we work in the infirmary. What do they call it these days? Something else. But anyhow, you work in that kind of area, and you have the opportunity to take people's temperature. And so over time, you've accumulated some data on the temperature of people. And maybe you've found that there's a vampire here at about 102. There's a normal person here, about 98.6. But then they're scattered around. Some people have fevers when they come in. So the question is, is there a way of using numerical data-- things that you can put real numbers-- is there a way of using that with this mechanism? And the answer is yes. You just say, is the temperature greater than or less than some threshold? And that gives you a test, a binary test, just like any of these other tests. [? Krishna? ?] Right? But where would I put the threshold? I suppose I could just put it at the average value. But that might not be the place that does the best job of splitting the samples into homogeneous groups. Christopher? STUDENT: So you run this numerical analysis on different places with different thresholds. PATRICK WINSTON: So you try different places, he says. And he's right. Because this is a computer, this is our slave. We don't care how much it works to figure out the right threshold. So what we do is we say, well, maybe the threshold's halfway between those two guys, or halfway between those two guys, or those two guys, or those two guys, or those two guys. So we can try one less threshold than we have samples. And we don't care if there are 10,000 samples, because this is a computer, and we don't care if it works all night. So that's how you find the threshold for a numeric test. By the way, I assured you earlier on you would never use the same test twice. Is that true for this? Yes, you would still never use the same test twice. But what you might do is you might use a different threshold on the same measurement the next time around. So when you start having numerical data, you may find yourself using the same test with the same axis but with a different value. All right. So now that we have this, then we can go back and compare how this method would look when we put it up against the sort of stuff we were talking about last time, with the electrical covers. So with the electrical covers, we had a situation like this. I don't know. We had samples that were places like this, and we had a division of the space that look pretty much like that. Not quite exactly in the right spots, but pretty close. So these are the decision boundaries for the situation where we are using nearest neighbors to divide up the data. What would the decision boundaries look like if these were four different kinds of things, and we were using this kind of mechanism? And maybe there's a lot of samples all clustered around places like that. What would the decision boundaries look like? Would they be the same as this? god, I hope not. Why? Because what we're going to do is we're going to use a threshold on each axis. So therefore, the decision boundaries are going to be parallel to one axis or the other. So we might decide, for example-- Oh, shoot. I think I'll draw it again, because it'll get confused if I draw it over the other one. So it looks like this. And that's how nearest neighbors does it. But a identification tree approach will pick a threshold along one axis or the other. Let's say it's this axis. It's only got one choice there. So it's going to put a line there. And now, what's the next thing it does? Well, it still has these two different kinds of things to separate. We're going to assume we've got four different kinds of things. So it's going to say, oh! I've Come down the negative side, so I need a threshold on the remaining data. And these are the only two things that are now remaining. So my only choice is to put a threshold in there. Now I guarantee this, absolutely guaranteed-- on the quiz, somebody-- presumably somebody who doesn't go to lectures-- will draw that line all the way across. And that's desperately wrong. Because we've already divided this data set in half. Now the choice of what we do over here is governed only by the remaining samples that we see, these two. And so the threshold is going to go in there like that. So that's what happens when you go back. This is used 10s of thousands of times. Always used. What are the virtues of it? Number one, you don't use all the tests. You use only the test that seem to be doing some useful work for you. So that means that you do a better job, because your measurement technique is simpler. And it costs less, because you're not going to the expense of doing all of the testing. So it's a real winner. But you know what? Some classes of people-- not scientists, but I mean people like doctors and stuff. They don't like to look at these tress. They're kind of rule-oriented. So they look a tree like this for determining what kind of thyroid disease you have, and it would have maybe 20 or so tests in it of various kinds of hormones, like thyroxine and this and that. And they say, ah, we can't deal with that. So we have to work with them. So what we do is we convert the tree into a set of rules. How do we convert the tree into a set of rules? Oops, wrong one. Go away, go away. Here's what I want. Yeah, good. How would we convert this tree into a set of rules? It's straightforward. [INAUDIBLE], what do we do? STUDENT: You'd basically just look down each branch-- PATRICK WINSTON: You'd basically just go down each branch to a leaf. So you say, for example, here's one rule. If shadow equals question mark, and garlic equals oh, [INAUDIBLE] want to choose No. Doesn't eat garlic. No. I think I'll say Yes. Yes. That changes the answer. Then if it eats garlic, it's not a vampire, right? That's one of four possible rules, because there are four leaf nodes. Now, almost done. We are done, except for one thing. We can actually take these four rules, and start thinking about how to simplify them. You can ask questions like, if I have a rule that tests both the shadow and the garlic, do I actually need both of those antecedents? And the answer is, in many cases, no. And in particular, in this case, no. Because if we look at our data set, what we discover is that in the event that we're talking about a shadow question mark-- oh, I guess I had a better choice the other way. Oh, no. If you look at the garlic, all the garlics-- Yes, Yes, and Yes-- it turns out that the answer is no, independent of what the shadow condition is. So we can look at the rules, and in some cases, we'll discover that our tree is a little bit more complicated than it needs to be. We can actually get rid of some of the clauses. So in the end, we can develop a very simple mechanism based on good old fashioned rule-based behavior, like you saw almost in the beginning of the subject, that does the job. And now, without any royalty, you're all free to put this in your PDA and use it to protect yourself in the days to com, especially since Halloween's just around the corner. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 19_Architectures_GPS_SOAR_Subsumption_Society_of_Mind.txt | PROFESSOR: [INAUDIBLE]. "Thus Spake Zarathustra" was made famous and popular by 2001. And that is music played at this magic moment when some primate suddenly gets an idea, presumably one of our ancestors. So how do we explain all that? We've got all of the ingredients on the table. And today I want to talk about various ways of putting those ingredients together. So we talked about representations, we've talked about methods, and today we're going to talk about architectures. And by the end of the class you'll know how to put one of those things together. Actually, no one knows how to put one of those things together. But what you will know is about some alternatives for putting those things together so as to make something that is arguably intelligent in the same way we are. So that is our agenda for today. We'll also talk a little bit more about stories. I think it was in 2007 when the Estonians moved a war memorial from the center of Tallinn off to a Russian war cemetery. Prior to that time the Estonians had been building up their national computer networks because they thought that computation was the wave of the future-- networks and all of that. Shortly after the movement of that war memorial, someone brought the Estonian national network down-- a cyber attack. It was widely believed to be the Russians. There's a large Russian ethnic population in Estonia to start with. And the movement of that war memorial irritated the Russians. And so everybody everybody thinks that they did it. But you know what? No computer can understand the story I just told. They can revel through all of the worldwide web finding information that's relevant to that, but no computer can understand the story I just told, except one. You'll see a demonstration of that later on today. So by the way, if you're interested in understanding the nature of intelligence, this is, of course, the most important lecture of the semester. And I should tell you a little bit about what we're going to do on Wednesday. Because for some reason the day before Thanksgiving tends to be a lecture that's lightly populated, except in this class. Because I'm going to talk about the artificial intelligence business and what can be learned from it about how to avoid going broke when you start your company. So for many of you that will be the most important lecture of the semester. It all started back in the dawn age of artificial intelligence. And really, it all started at Carnegie Mellon, sad to say. Because the people at Carnegie Mellon, notably Newell and Simon, were the first to think about sort of a general purpose way of putting things together so as to build a structure or architecture in which particular intelligent systems could be built. So their idea was called the general problem solver. A long name for a simple idea. And the simple idea is that you start your life out in a current state, call it C. And you want to get to some goal state. Call it S. And a way you do that is you measure somehow the symbolic difference between where you are and where you want to be. So that's the difference. We'll call that difference D. And when you observe that difference that's enough, they say, in this general approach to problem solving. For you to select some operation that will move you from your current state to some new state, an intermediate state. Call it I. So I, or that operator, O, is determined by the difference, D. And then, of course, the next thing to do is to measure the difference between that intermediate state and the state you want to be in, and choose some operator that's relevant to reducing that state. So we'll call that D2, and we'll call this O2. And D2 is what leads you to 02, and so it goes. So that's the idea. And that's often called means-ends analysis. Why? Because the end that you want to achieve is being in that final state, S. And the means is that operator, O. So you have some notion of where you want to be and the difference of where you are and where you want to be. And you pick an operator so as to reduce that difference. So this is all very abstract. Let's exercise it in solving a problem that you will all be faced with here in a day or two. That is, for many of you-- most Of you, I hope-- the problem of going home. So here you are. You're at MIT. And where you want to be is over here, at home. So you measure the difference between MIT and home. And for many of you it's further than you can go by car and not so far that you can't go at all. So what you do is, you say, well, the right operator is taking an airplane. So there is the operator, take an airplane. And this is the difference, D. And the difference, D, being sufficiently large, you take the plane. Trouble is, if you happen to be sitting here in [? 10-250 ?] there's no way you can take an airplane, because they don't fit in here. So you've got another problem, and that is to get to the airplane. So the distance between here and Logan is such that the right way to do that is to take the MBTA. And that's determined because you're working on this difference reduction right here, the difference from being at MIT and being at the airport. So that difference dictates that you take the MBTA. So you see, you're re-cursing. But you know there are no MBTA cars in here either. So there's still a difference like so. And that difference dictates that you walk. So you've got D1, D2, and D3. And by the time you've excised the operators relevant to those three differences, you're at Logan. Then you take the airplane, you get over to your hometown, and your faced with the smaller difference of getting from that airport to where you actually want to go. So that's the general problem- solver idea. It was such an exciting idea at the time. It was such an exciting idea at the time because people would say to themselves, ah! This is a general purpose problem solver, so we can set it onto the problem of making itself smarter. And so there was a kind of imagined chain reaction that would take place. And the developers of this architecture warned the public that within 10 years-- that is to say, by about 1970-- computers would be generally as smart as people. And a lot of people made fun of them for that prediction. But it was actually scientists attempting to be responsible. Because they thought something, a quite serious dislocation was coming along, and that people should know that it was coming. And so they felt it was their responsibility in that age of scientific responsibility to warn the public. It didn't turn out that way, because the problem of collecting the differences and finding the operators, that's outside the scope of the architecture. So this is the problem that has to be solved by a human before this architecture can be used. You have to have identified the differences that you might encounter and the operators that you might use, and build this table which relates the two together. So maybe that one, that one, some off-diagonal elements, and so on. But building that table turned out to be a hard job. So not surprisingly, the idea evolved. And eventually the folks at Carnegie who developed the general problem solver-- most notably Newell and his students-- developed a newer, fresher, more elaborate architecture called SOAR. And here's how SOAR works. First of all, what does SOAR mean? It doesn't mean anything. It used to mean State Operator And Result. But for some reason the proponents of the SOAR architecture decided they don't like that acronym, and have asserted that SOAR is merely a label that shouldn't be thought of as an acronym. In any event, SOAR consists of various parts. It has a long-term memory. It has a short-term memory. And it has connections to the outside world, maybe a vision system and an action system. But most of the activity of the SOAR problem-solving architecture takes place in a short-term memory. So you can view the contents of the long-term memory as shuttling in and out of short-term memory. So you can see right away that this mechanism, this architecture, is heavily influenced by certain cognitive psychology experiments having to do with how much you can hold in your short-term memory-- nonsense syllables and all that sort of thing that was popular back in those days. So this was an architecture devised primarily by psychologists. And it had amongst its features a short-term memory and a long-term memory. So that's part 1 of this architecture. So what's in the long-term memory? Well, assertions and rules, AKA productions. A production being the Carnegie vernacular for rule. It's just the rule-based stuff like you saw on almost the first day of class. So the whole thing is a gigantic rule-based system with assertions and rules the shuttle back and forth from long-term memory into short-term memory where processing takes place. The third thing that comes to mind when you think of SOAR architecture is they had an elaborate preference system. You recall that when we talked about rule-based systems there's always a question of what do you do when more than one rule would work? You have to have some way of breaking those ties. The SOAR architecture has an elaborate subsystem for doing that. But I said that these are the first three things you think, and maybe that's not right. Because the next thing you think about is perhaps a better thing to identify with the SOAR architecture. And that's the idea of problem spaces. And that's the idea that if you're going to solve a problem you have to develop a space and do a search through that space. Just like we did when we talked about how we can get from here to home. There's a space of places, that's our problem space. We can do a search through that space to find a way to get from one place to another. That's the sort of thing that SOAR is focused on. Finally, the fifth element that you tend to think about when you think about SOAR is the idea of universal subgoaling. And that's the idea that whenever you can't think of want to do next, that becomes your next problem that deserves it's own problem space and its own set of differences and operators, and rules and assertions. So you start off on a high level, then you have to solve problems at a lower level, just like you did up there with a general problem solver. So if you have these two architectures you can begin to say, well, what are they centered on? And this architecture, this general problem solver, is centered on the idea that everything is about problem solving. Because the problem solving hypothesis-- no one gave it that name. But that's what it was. And this architecture did get its name. And it was said always, by Newell, to be based on what he called the symbol system hypothesis. The hypothesis that what we are as humans is symbol manipulators. And we can uncover how that all works by giving people crypto-arithmetic problems and having them talk out loud, by thinking about what happens when you try to remember nonsense syllables, by all that sort of stuff that was en vogue in terms of psychology experiments in the day when this architecture was first articulated. But when you look at architectures you can sort of see where they come from and what their antecedents are. It has a short-term memory and a long-term memory, because Newell and his associates were cognitive scientists. It has assertions and rules and preferences, because Newell and his associates were also AI people. And it has problem spaces and universal subgoaling because those are ideas that had been work out in a more primitive form already in the general problem-solver architecture. So that's a glimpse of what SOAR looked like in its early days. It's been very highly developed by a lot of smart people. So although it's symbol centered, they've attached to it things having to do with emotion and perception, but generally with the view that the first thing to do when faced with this perception is to get it out of there and get it into a symbolic form. That's sort of the bias that the architecture comes with. So those are two architectures that are heavily biased toward thinking that the important part of what we do is problem solving. But the most important, perhaps-- at least from an MIT perspective-- of these problem-solving oriented ways of thinking about the world, is Marvin Minsky's architecture, which he articulates in his book "The Emotion Machine." And Marvin is not just concerned with problem solving, but also with how problem solving might come in layers. So let me show you an example of the sort of problem what motivates some of Marvin's thinking. So you can read that, it's a short little vignette. You have no trouble understanding it, right? No. It's not difficult for us humans. Awfully tough for a computer. In part, because the thinking you need, your ability to understand that story, requires you to think on many levels at the same time. First of all, there's a sort of, at the bottom, instinctive reaction. You see where there's instinctive reaction? That's the part where she hears a sound and terns her head. That's instinct, right? That's practically built in. But then what she sees is a car. So that's something that we don't have wired in. It would be unlikely that we've evolved in the last 100 years to have an instinctive appreciation of cars barrelling down the road. So the next level in Marvin's architecture is learned reaction. So that's the part about thinking about the car. Now spread throughout there-- well, let's see where is a particularly good example. She decides to sprint across the road. So that's where she's solving a problem. So that's the deliberative thinking level. It doesn't stop there, because later on she reflects on her impulsive decision. So she thinks not only about stuff that's happening out there in the world, but she also thinks about stuff that's going on in here. So that's a level which we can call reflective thinking. Well, you know, it doesn't stop there, because she also considers, in another part of the story, something about being uneasy about arriving late. So she's not only just thinking about events that are going on in her mind right now, but events that are going on right now relative to plans she's made. Some Marvin calls that the self-reflecting layer. But that isn't the whole thing either, because toward the end of the story she starts to worry about what her friends would think of her. So there's a kind of reflective thinking in a more social context. So he calls that self-conscious thinking. So as the Carnegie folks think, the SOAR architecture focuses mostly on problem solving, Minsky's "Emotion Machine" book considers not just thinking, but thinking on many layers. And the blocker to doing any of that can be said to be the development of common sense, which computers, alas, have never had much of. So this could be said to be based on the common sense hypothesis. And the common sense hypothesis holds that in order to do all of that stuff, you have to have common sense like people. And if you have to have common sense like people, you have to think about how much of that is there and how can we go get it? And so this spawned a lot of activity in the media lab amongst people influenced by Marvin, having to do with gathering common sense. The open mind project, the work of Henry Lieberman and others, having to do with the gathering of common sense from the world wide web as a way of populating systems that would lay the foundation for doing this kind of layered thinking. So that is a brief survey of some mechanisms, some older than others, but all but GPS-- GPS too. Let's face it, it's hard to think of solving any problem without means-ends analysis being involved. So GPS isn't wrong, it's just not the only tool you need to think about what to do. So these are early, and late, and still-current. But it's not the only thing there is, because there have been reactions against this problem-solving way of thinking about the development of intelligence. And the most prominent of those counter currents, of those alternative ideas, belongs to Rod Brooks and his subsumption architecture. So along about the early-- along about the years surrounding 1990, Brooks became upset-- subsumption-- because robots couldn't do much. They would turn them on at night, and then the next morning they'd come in the laboratory and they would have moved 25 feet, nicely avoiding a table perhaps. Not doing very much and taking a long time to do it. So he had decided that it's because people were thinking in the wrong way. In those days people thought that the way you build a robot is you build a vision system, and then you build a reasoning system, and then you build an action system. And it can do almost nothing, but it does something. So you improve the vision system, and improve the reasoning system, and improve the actual system. And now you've broken it, because all the stuff you used to be able to do doesn't work anymore. So what's the alternative? Well, the alternative, as articulated by Brooks, is to turn this idea on its side. So instead of having an encapsulated vision system, an encapsulated reasoning system, and an encapsulated action system, what you have is layers that are focused not on the sensing and the reasoning and the action, but layers that are specialized to dealing with the world. So in Brook's way of thinking about things, at the lowest level you might have a system that's capable of-- well, before we get to that, avoiding objects. And maybe the next level up is the wandering layer. And maybe the next level up after that is explore. And maybe the next level up after that is seek. Now in the old days when people took 6001 I had no trouble getting an answer the question, what does this remind you of in 6001? It doesn't remind you of anything in 6001 since you haven't taken it. But it viewed, as a generalization of a programming idea, what is the programming idea? There are only a few powerful ideas in program, and this is a generalization of one of them. What is it? Do you have a name? Yes, Andrew? STUDENT: Layers of abstraction? PROFESSOR: Layers of abstraction, and abstraction barriers. That nails it pretty well. Because each of these guys can have its own vision, action, and reasoning system. And if you think of these as abstraction boundaries, then when you got this thing working you don't screw with it anymore. You build this layer on top. And it may reach down in here from time to time, but it doesn't fundamentally change it. Brooks was inspired in part by the way our brains are constructed. All that old stuff that we share with pigs is down in there deep, and we put the neocortex over it. So it looks layered in a way that would make [? Gerry Sussman ?] proud. So this then is the way that Brooks looks at the world, and it's characterized by a few features just like SOAR is. One of those features is no representation. So this is a detail that's probably right at the level that Brooks was operating, and very questionable when you get above the level that Brooks was operating. But before I go on, let me say what the hypothesis is. The hypothesis is the creature hypothesis. It's the hypothesis that once you can get a machine to act as smart as an insect, then the rest will be easy. Well, how do you get a creature to be smart as an insect? Maybe you don't need representation. We focused on representation in this course, so you can see there's a little stress--- Next thing is, what do you do if you don't have a representation? Let's see. Your representation makes a model possible. Models make it possible to predict, to understand, to explain, and to control. So if you don't have one what can you possibly do? Brooks' answer is, you use the world instead of a model. So everything you do is reactive. You don't have anything in your head that is a map of this room. But maybe I don't need one because I can get around that table by constantly observing it. And we don't have to fill up the memory with that information, I can just react to it. So no representation, use the world instead of a model, and the mechanisms in their purest form are just finite-state machines. So with that, Brooks was able to do things that people were never able to do before. And what's the modern [? instantiation ?] of this architecture? Now, according to Brooks, in use in 5 million homes in the United States? STUDENT: The Roomba? PROFESSOR: It's the Roomba The Roomba robot is, by Brooks' account, approximately the thirteenth business plan of iRobot. And it's the one that made it big, because the Rumba vacuum cleaner has been very successful. Would you like to see a movie of its processor? So this is a film made some time ago that shows, in some sense, the summa of that architecture. What I want you to imagine very briefly is a robot that wanders around in the halls and rooms of the old [? Tech Square ?] clinking the Coke cans. Okay, you all got an image of that in your mind? Because I want you to compare the image you now have of that robot that's wandering around collecting the Coke can, with the actual movie. [VIDEO PLAYBACK] -Herbert, the soda-can collecting mobile robot. He was built at the MIT AI lab in 1989. Work was done by John Cannell under the supervision of Rodney Brooks. Herbert is a robot controlled by subsumption architecture. This is a collection of small behaviors that influence the overall activities of the robot. There are no centralized controllers and no world model. -Herbert navigates by using a number of infrared proximity censors around its body and basically following walls and corridors. It can also look for the can through a laser light striper. Right now it's come out of the door of an office, followed along the wall, and then its laser light striper has seen a can on top of the desk in front of it. When this happens the robots and deploys its arm. You can see the arm going out now. -The arm has a number of censors itself. There are fingertip censors, a break beam in the jaws, and two infrared proximity sensors on the front of the hand. -It grabs cans in a stereotype fashion. First, it lowers down to find a surface somewhere, then it bounces along the surface until it sees the can in front. It uses the hand-based IRs to re-center the arm by rotating the robot's body until the can comes between the jaws of the gripper, at which point the break-beam senses the can. -After acquiring the can, Herbert will have tucked the arm back into its normal traveling configuration and attempt to go home. -Since it has no central [? arm presentation, ?] it doesn't have any map of where it came from. Instead, it has an algorithm which uses a magnetic compass to determine every time it comes through a door, will it be able to find the door? It basically has a policy of always going north every time it exits the door. -So now the can is being tucked away. As the robot turns you'll see a red stripe from the laser range finder. And now it's using the [INAUDIBLE] IR to navigate back, find the door, and go through the door with its prize. [END VIDEO PLAYBACK] PROFESSOR: And there, if you were paying attention, you saw a little glimpse of John Cannell who was the student to develop that system. So that was a tour de force. That was a magic moment. That was when you open the champagne. It's not what you expected, of course, because when I say imagine a robot wandering around in [? Tech Square ?] picking up Coke cans, that leaves open a huge envelope of possible hallucinations. And usually or hallucinations about these things are-- we imagine things to be more fluid, more natural, and more impressive than they actually are. But that was impressive, because no robot came close to doing anything like that before. More to be said about that during the business lecture on Wednesday. So that's the subsumption architecture. By the way, maybe at this point we can say something about how the other architectures relate to what Minsky was talking about. What's this deliberative thinking layer correspond to? That's what SOAR is about, and maybe GPS. So what's subsumption about? It's about stuff down here. It's about instinctive reaction and learned reaction. But shoot, what about Minsky's other layers? If we're going to be building systems that are as smart as those things then we have to worry a little bit about that sort of thing too. So that brings us to the genesis architecture. And now let me give you the standard caution that should be early in the presentation of any academic. I will sometimes say "I," and what I mean is "we." And sometimes I'll say "we," and what I mean is "they." This was a system that was developed mostly by students of mine who persuaded me, after a great deal of time, that they were thinking the right kinds of thoughts. But here's how the genesis architecture works. As no surprise, given recent discussions, it's all centered on language. And the language part of the genesis system has two roles, one of which is to guide, and marshal, and interact with the perceptual systems. And the other is to enable the description of events. That's how it works. So is perception important? I don't know. I might ask you a question like, is there anybody sitting in the front row wearing blue jeans? And it's hard for you to resist, under those circumstances, your eyes from going over there and answering the question. Your eyes answer the question. No symbol processing system is involved, except in so far as my language system has communicated with their language system, which drives your motor system and your vision system to go over there and answer the question for you. But it's not just the real stuff that the language system directs your attention to. It's also the imagined stuff. It's been a long semester. Have I told you the story about my table saw? Probably not. Here's the deal. I bought a table saw. It's a wonderful table saw. I was installing it with a friend of mine who's a cabinet maker. He said, never wear gloves when you operate the saw. "Why?" I said. Before he could answer the question I figured it out. Can you figure out why you never wear gloves when you operate a table saw? You know what a table saw is, right? It's a table with a spinning blade in the middle. And you use it to cut wood. Why should you never wear gloves? Yes? STUDENT: Well-- STUDENT: --Well, you know the answer. Ha, that's not fair. That's old Brett up there. He's heard the story too many times. Yes, Andrew, you got it. STUDENT: I've been told the answer before. PROFESSOR: You've been told the answer. How about somebody who hasn't been told the answer. Yes? STUDENT: Because the gloves might get caught. PROFESSOR: Because the glove might get caught and pull your hand into the blade. And then what happens? It's horrible. You're hand gets mangled and your fingers get cut off, and this happens a lot to professionals. It won't actually happen with that table saw that I bought, because its play detects flesh and stops the blade, which then leads to stopping the blade and having the blade retreat into the table in about two microseconds-- two milliseconds. So, in general though, it's a bad idea, and you always have to suppose that the mechanism isn't working anyway in order to use good safety practice. But here's an example of something that nobody ever told you that he was able to figure out, by imagining what would happen and reading the answers off of the scene that he imagined. So nobody ever says many of the things that we know, but we know them anyway. Here's another example. Imagine running down the street with a full bucket of water. What happens? The water splashes out and gets your leg wet, right? You won't find that in Open Minds database. Nobody ever said that over the web. It's not written down anywhere. But you know it. Because you, we human beings have the capacity to imagine perceptual things and read the answers to questions off of our imaginations with that perceptual apparatus. So that's a very important connection down there. And then if you've got the ability to describe events, then you've got the ability to tell and understand stories. And if you can do that, then you can start to get a handle on culture, both macro and micro. And by macro culture I mean the country you grew up in, the religion you grew up with. And by micro I mean your family and personal experience, and all shades in between. So I don't know, what inspires me and my associates to think in these terms? We talked about a little bit of it last time when I talked about evolution and the apparent flowering of our species about 50,000 years ago, at which time we got something. And I believe that what we got-- and this is the characterization of this particular hypothesis-- what we got is the ability to tell stories and understand them. So if we want to label this representation, it's the label strong story hypothesis. So what's the weak story hypothesis? The weak story hypothesis is, this is important. The strong story hypothesis is, this is all there is. But is there any other evidence of this is really, really, really important? So I've queried Krishna here before the class starts, and he tells me I haven't told you about the following experiment. This, in my way of thinking, is the most important series of experiments ever done in cognitive psychology, developmental psychology, actually. So here's how we get started. There's a rectangular room, if you're a person. If you're a rat, it's a rectangular box. All the walls are painted white. Are you with me so far? So now, in each corner there's a basket, or cloth, or something in which or under which you can put some food. Now, you put the food there while the rat watches you. And then you give the rat a little spin to disorient it. All right? So then, the rest stops and goes for the food. And you can keep track of where the rat goes. And the rat goes with approximate equal probability predominantly to those two corners. So I'd have bet you didn't know that rats were that smart. So they understand the rectangular nature of the room and they don't go to the diagonal corners where the food cannot be. So are these genius rats? Or maybe we're just rats with big brains. Because we do the same thing. So if you repeat this experiment and replace the rat with a small child, and then you put a toy in there instead of food, and the rat-- not the rat. The child is usually held in a parent's arms, usually the child's mother-- usually because they think that if they participate in these experiments up there at Harvard their kid will get into Harvard some day. So the kid goes to a diagonal corner just like a rat. And then the next thing you do is, you try an adult, maybe an MIT student. That way you can use food again. And you get the same result. Who could be surprised? So rats, children, and human adults, pretty much all the same with respect to this experiment, until you paint one wall blue. Rats are not colorblind, in case you're wondering. Then what happens? Well, if you pay one wall blue the rat still goes with equal probability to the two diagonal corners. If you paint one wall blue, the child still goes to the two diagonal corners with approximate equal probability. It's only us genius human adults who go only to that corner. So this invites a couple of questions. One of which is, when does a child become an adult? Any ideas? [INAUDIBLE], what do you think? STUDENT: [INAUDIBLE]. PROFESSOR: You can pick a number greater than 1 and less than 10. [INAUDIBLE], what do you think? STUDENT: Five? PROFESSOR: It's a pretty good guess. Do you have siblings at that age? It's a surprise but, why is it five? Is it because-- what does it relate to? Is there any correlate to the onset of that ability? You might try everything, as [INAUDIBLE] does, because she's extremely careful. So she's tried gender, she's tried the onset of language, the appreciation of music, handedness, and there's only one thing that matters. And that is that the child becomes adult at that time when they start to use the word left and right when they describe the world. Now I said that very carefully because they understand left and right at an earlier age, but they only have started to use the words left and right when they describe the world at the time that they begin to break this symmetry and go to the correct corner. Now for the next element of this I need something to read. Has anyone got a textbook handy? Ah, "China, an Illustrated History." Now I need a volunteer. OK. Andrew, you want to do this? So here's what you're going to do. You can stay there. But you need to stand up. So what I'm going to do is, I'm going to read you a passage from this book. And I want you to say it back to me at the same time I read it. It's as if you're doing simultaneous translation, except it's English to English. This things got words I can't pronounce. OK, are you ready to go? All right. "When overwhelmed by the magnitude of the problems he tackled, he began to suspect that others were plotting against him or secretly ridiculing him." Thank you very much. That's great. So you see, he could do it. Some people can't do it. At least it take a little practice. But he did it. And guess what I've done to him? I've reduced his intelligence to that of a rat. Because if you do this experiment with an adult human who's doing this simultaneous English to English translation, they go with equal probability to the two corners. So what's happened? What's happened is you've jambed their language processor. And when their language processor is jambed they can't put the blue wall together with the rectangular shape. So it seems to be that language is the mediator of exactly the combinators you need in order to build descriptions. Because they can't even put those things together when their language processor is jambed by the simultaneous translation phenomenon. So that brings us to the two gold star ideas of the day. One is, if you want to make yourself smarter you want to do those things-- look, listen, draw, and talk. Because those are the particular mechanisms that surround this area down here, which is the center of what we do-- which is the center of our thinking. So why do you take notes in class? Now because you'll ever look at them before, but because it forces the engagement of your linguistic-- of your linguistic, your motor, and your visual apparatus. And that makes you smarter, because it's exercising that stuff. The second thing you can say, in conclusion, especially from this experiment, is beware of fast talkers. Why do you want to be aware of fast talkers. It's not because they will talk you into anything. It's because that when they talk fast they're jambing your language processor and you can't thing. That's why you want to be aware of fast talkers. Because if they jamb your language processor you won't thinking and you'll buy that car, or you'll buy that drink, or you'll do any manner of things that people who want you to do those things have learned to do by talking to jamb your processor. So that completes what we're going to do today. And I'll give you a demonstration of some of this stuff on another occasion. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 7_Constraints_Interpreting_Line_Drawings.txt | PATRICK WINSTON: Many of you, maybe most of you, will never have to work another search problem by hand in your natural life. Others will want to take another run at it on the final. I've been much criticized for this way of doing grading in the class. But the way I look at it is that the relationship between students and instructors ought to be less adversarial than it used to be in the good old days when I was a student. I especially remember an examination we took-- all of us took-- on Rayleigh scattering. That was 803, I think. And back in the old days, we had to take four physics courses, not just two. And we had to take four math courses, not just two. Back in the days when-- well, I was going to say men were men. But most of us were men in those days. I think we only had 20 women in our class. Anyway, we had a quiz on Rayleigh scattering. And I thought, well, this is pretty hard. And then I got my quiz back. 26. I thought, well, I've been found out. I'm going to flunk out. My father will make me go to law school. I'll never attract anyone to marry. Horrible things will happen. Then the instructor announced the class average was 18. I was two standard deviations above that. They gave us the same exam two weeks later, and accounts vary. Some people say that the class average went down. Anyway, today we're going to study some stuff. We're going to study some stuff that will make it possible for you to understand how you can do that computation in just a couple of seconds, even with the delays introduced by the redrawing. Now this particular program-- I'm not real sure and I don't have a proof, but I think it will take more than the lifetime of the universe to find a legitimate coloring of the continental United States. But by the end of the next class, you'll see how to do that lickety-split in just couple of seconds, even with the re-display delays. Now we could, of course, do this in two ways. One way is I could start off by saying, let C be a constraint satisfaction problem and just give you the result. And anybody can do that. That's great. And you needed to learn some stuff that way. But there are some things that you needed to learn, I think, a different way. And that different way involves my telling you the story of how it all came to be. This is a new field, pretty much, and therefore I know most of the people in it. And I know all of the people who did the work that I'm going to tell you about today. I'm going to tell you a little bit about the struggle for coming up with the ideas that led to one of the most powerful methods you'll learn about in the subject. It all has to do, originally, with an attempt to make a computer capable of seeing. And everybody said to themselves, well, let's start with seeing simple things, like children's blocks. And so Adolfo Guzman, a graduate student of Marvin Minsky's, was charged with the summer project, which led to his Ph.D. thesis, of writing a program that could look at a line drawing and determine how many objects are in the line drawing. So for example, there behind my outline is a line drawing. And you believe instantly that there are two objects there, even though in some deep sense, it's ambiguous. There are all sorts of ways that you could, through trickery, arrange something with even seven objects that look that way. But most people would say there are two objects. So Guzman set about the problem of figuring out how to do that. And then his work was followed by Dave Huffman. And his work was followed up by Dave Waltz. And his work was followed up by Jane [? Froyter ?], who's not listed there quite yet. And I want to tell you a little bit of story how that all happened. So Guzman took a lot of pictures. He went to Boston Baby, the precursor to Toys R Us, purchased a large block set on a government contract, and went about the business of taking a lot of pictures of them so he'd get a feel for the domain. And eventually he decided that he could build a program that could determine that there are two objects here by using the lines as a quanta of evidence about which faces belong together in an object. So he said, after studying these for a long time, that these drawings tend to have a lot of arrow-type junctions and a lot of fork-type junctions. And when you see an arrow, it's a little bit of evidence that the objects on either side of the shaft are the same-- the face on either side of the shaft belong to the same object. And over here, the fork-type junction suggests that pairwise, three pairs of faces seem to belong to the same object. So with that idea, he could come back to a drawing like this and start decorating it with these so-called links, these quanta of evidence that faces belong to the same object. And if I've done it right, you get something that looks like so. It's a little hard to see what's going on on the drawing itself. So let me number these. Now I have an easier-to-deal-with picture. There are two links between 1 and 2 and 1 and 3. One link between 2 and 3. One between 2 and 4. Two here. Two here. And one each from all of these. So Guzman produces this evidence for how the problem ought to be solved and then he begins to think about various ways of using the evidence. So he could, and did, decide that one link is enough to presume that the faces belong to the same object. And you can see that that's a little bit too liberal. Because that says that the whole thing is just one object. So Guzman rejected the one-link theory and went to the two-link theory. And the two-link theory says, oh, well, let's see. If we insist that two links should be there before we are willing to decide that it's the same object, then these three faces are pulled together into one object. These three are pulled together into one object. But alas, 7 doesn't share two links with anything, so it's left dangling out there. So plainly, the two-link theory is too conservative. So that as you would soon discover in any [INAUDIBLE] project, would lead you to a third theory, which is two lengths repeated. So now that we've formed these super regions, we can come back and say we'll merge super regions if they're connected together by two or more links. So nothing new happens up here. But this super region is joined at 7 by two or more links. So it pulls everything together like so. So that worked fine. Well, it didn't work fine. There were lots of examples of situations where it didn't work-- in situations that were considered nonsensical by us humans because it seemed silly, the kind of conclusions that it reached. So when Guzman presented this work in his Ph.D. thesis defense, it's said-- and who knows if it's true-- but it's said that Marvin Minsky thought it was great work and we should make him a professor. It happens that Dave Huffman was also on the committee and said that it was ad hoc and the thesis should be rejected. So we had two polar opposites of impressions there. Dave Huffman, by the way-- you've heard that name before, I imagine. It's the guy who invented Huffman coating as a term paper in a information theory course taught by Bob Fano. He got an A, they say. So Huffman didn't like it. He thought it was a little bit too ad hoc. It was too heuristic. It didn't take advantage of anything we might know about physics. And so people began to say, well, why does it work? And when does it not work? And that's just about the best question you can answer in a situation like this. And by playing with some more of Guzman's pictures and reflecting on them, it turned out that it worked because the world is full of three-face junctions. Or let me say three-face vertexes because they're out there in the world. We'll reserve the word junction for something else. And three-face vertexes generally project into either an arrow or a fork. So Guzman's whole program worked on the weak backward conclusion that if you see one of those, it probably came from one of these. So this is in the drawing. That's in the world. So by a process that's neither deduction or induction, but rather abduction, you see one of those guys. And you say, well, they often come from three-face vertexes in the world, so if you see one, it probably came from a three-face vertex in the world. That's abduction. So once Huffman saw all that, being a mathematician, he began to think about how one might develop a different and better theory. But we have to recognize that all this work started off with the efforts of Guzman, who was an experimentalist. And Huffman was a mathematician. So naturally, they approached the problem differently. So Huffman says, I'm not going to concern myself too much with the actual problem that Guzman was trying to solve. Rather, I'm going to work in a very simple world, which I can deal with mathematically. So he decided that he was going to work in a world which had several characteristics. Characteristic number one was that the world would be presented in general position. That is to say, no screw cases. So if you see a cube, it's going to look like this. And it's not going to look like this. So that's out. And this is in. And the idea here is that that's not general position, because if you perturb your point of view a little bit, you'll change those T junctions into forks and arrows. So we know it had to deal with those kinds of weird kinds of cases. So that's presumption number one. Presumption number two is that we're going to be dealing with a world that's trihedral. That means all vertexes out there are going to be formed from three planes. So you're going to have the intersection of three planes. So how many kinds of junctions can you see if the world is composed that way is the question that Huffman sent himself upon. The next assumption was that there are going to be four kinds of lines. See, Guzman only recognized two kinds of lines-- lines that have a length and lines that don't have a link. And they don't have very much to do with the physical world. So Huffman says, I want to get the constraint out of the physical world. So I'm going to deal with four kinds of lines-- concave, convex, and boundary. So each of those kinds of lines is going to have its own notation. We'll call the convex lines plus, the concave lines minus, and a boundary lines are going to get an arrow on them, depending on which side of the object-- which side you would see the object if you walk along the direction of this little arrow. STUDENT: Question. Which is concave and which is convex? You said something and you wrote the opposite. PATRICK WINSTON: Yeah, sorry. Thank you. So concave, convex, and boundary. Thank you. So this down here, that's a concave line. And that would get a minus label. Oh, I don't know. These lines you're seeing here, if there were stuff behind them instead of me, then all those would be convex lines. Many times you see a boundary line. For example, now you don't see both sides of that line down here at the bottom. So that's a boundary line. And the arrow would point in that direction, so as to keep the stuff of the object on the right as you walk along a kind of mathematical convention. So three kinds of lines, four kinds of labels. And there's some things left out. That's because Huffman wanted to keep his problem simple, something he could manage by hand. What's left out? Cracks are left out. Shadows are left out. There's a presumption that there's nothing of interest here with respect to that. So let's have a little vocabulary before I go on. And I'll try to stick to it. But there's the stuff out there, and that consists of vertexes and edges. And over here, on the blackboard, we'll have junctions and lines. And I'll try to stick to that vocabulary. All right? Yes, Christopher? STUDENT: Didn't you say there were four types of lines? PATRICK WINSTON: Yeah, there are four kinds of-- the question is, didn't I say there were four kinds of lines? There are three kinds of lines, but since we can label a boundary line in either direction, we have four labels. OK? It depends on which side of the object the stuff is on. And that will be clear, I think, as soon as I do an example. So one, two, three assumptions. A little bit of vocabulary. So we have the possibility of making a complete catalog. This is so simple. We have the possibility of making a complete catalog of all the ways that lines can come together to form a junction with respect to these four labels. Now at first you might say, oh my god, that will take a couple of years. But maybe it won't take a couple of years. And in the end, to perhaps your surprise, you discover that there are only 18 ways to arrange the labels around a junction in this world. Everything else is excluded. If you have something that's not in the set we're going to produce now, it can't be built with trihedral vertexes. So I've listed up there six L's, five forks, four T's, and three arrows. Let's see if I can figure out why there are those 18 and nothing else. Well, if we have three vertexes coming together, that means there are eight octants, right? And the stuff of the object may fill 1, 2, 3, 4, 5, 6, 7 or all eight octants. Now of course, if you fill all eight octants, there's no junction. So we don't consider that case. If we don't fill any the octants, there's no junction. There's no vertex. So we don't consider that case. But if just one of the eight octants is filled with stuff, then we can look at it from any of the seven remaining octants. So right now, you're looking at it from one of the seven remaining octants. And if I'm not mistaken, you're going to see a fork-style junction there, right? And you're going to see a fork-style junction in which all of the edges are convex. An unfortunate selection of names, because the linguists tell me that we index all of our words by the first forename. And those have the same first forename, so they're easily confused. So here's a fork-style junction. And we know that one way that's legitimate for the lines to come in is with four pluses. Now that's not the only way you can see that. And here's another way. There's an L-style junction, right? And both of those are boundary lines. And we want to draw the boundary line indicators on there so that if we walk along the arrows, the stuff of the object would be on the right. So I suppose, to make it easier to me do my drawing, I should look at it this way. And then I would say, well, that has to be a legitimate way of labeling a junction. Are there any more? Well, there's seven altogether, but many of them are the same. There is one more that's a little different, though. I can hold this guy up like so. And if I'm holding at the right angle for you, you see an arrow-style junction, right? Two boundaries, and the barb is convex. So in this particular case, I've got that. I've got that. And I've got a plus there. And that's there happen to be for the one octant filled with stuff case. I happen to be able to reverse this, though. And here's the seven octants filled case. So that tells us that it's possible to have a vertex out there in a space that when reduced to a junction on the board deserves three minus labels, because all of these that you're seeing now are concave. So another fork-style junction looks like so. And since there's only one octant to look at from, that completes our analysis of the seven octants filled case. Now we have a couple of other possibilities here. We might have five octants filled with stuff. So that means there are three octants that we can look from. So let's suppose that these little chalk pieces are little people. And they're looking at this junction down there, where this white object is fused with the table. I'm fusing it with the table because I want to consider it to be one object. We can view it from the three objects indicated by those three little chalk pieces. And ask ourselves, well, in the event that we look at it from those three places, what do we see? And if we look at it from the perspective of the piece of purple chalk-- I'll have to walk around and be sure-- looks like an arrow junction with two concaves and a convex. Did I get that right? So I'm looking at it from this perspective. It's an arrow. This is convex and these two are concave because I fused the paper box with the table. Looking at it from the perspective of this blue guy-- let me rotate it so you can have a better understanding of the blue guy-- it looks like a concave line and a boundary. So it's an L. And this one is a boundary. And that's concave. And by a kind of symmetry, we're also going to get that one from the other side, the third of the three octants. Well, we're off and running. But we still have an awful lot to go. And we could manage to deal with it by thinking about this object once again. But instead of this situation out here, to turn it around and look at this vertex. Think about the junctions that it can produce. I think I'll do that for you. Because you really have to play with this and move it around a little bit to see how things are going to look. So let me think about how that's going to work out. I know that one of the possibilities is going to look like so. I might as well not hide that from you. It's going to be what happens when there's a little man looking up at the junction. And this one's going to be minus. And now we've got two more that are just like that. Look like so. And you say, oops. You say, aren't those just a rotational variance of each other? And the answer is sure. I write them all down, because if you get a fork-style junction in space, there are three different ways it could be labeled. Depending on which of the lines you put the minus label on. So that takes care of that. And then there's one more of these fork-style junctions. And that's plus, plus, minus that derives from this case. And there appear to be three more of these L-style junctions. And they look like, let's see, plus, then plus. I'm having to think this through as I go. And then-- and that's it. Well, what about the T's? Well, in this world, the only way you can get a T is if some object is obscuring another object. And if an object is obscuring a line, it can be any line at all. So that's why the four remaining ones are easy beyond description to label. And of course, the cross pieces of the T are all boundary lines, with the obscuring object on the right. Now let's see. We've taken care now of the one, three, five, and seven octants filled cases. What about two, four, and six? Well, it turns out you can have vertexes that are made that way too. But they will have more than three faces. They'll have six faces. They'll be like what happens when an object comes together at a point, like so. Like that. You can play with it a little bit and see that if you have two, four, or six objects filled with stuff, there are more than three faces. So we're going to ignore those. So our constraint is going to be a little more severe than would be suggested by the terminology Huffman uses. They're going to be trihedral all right, but they're also going to be three faces. So we went to a lot of work there. But what have we discovered? We've discovered that in this world, this is a complete, 100% percent, nothing excluded, nothing else can be there, catalog of all possible ways the junctions can have line labels arranged around them. There's nothing else in this world. So that's a very powerful constraint. So now let's see what we can do with it. This example is usually more fun when the Red Sox are doing better, but they're not. Yet we'll use it any way. We're going to start with an object that looks like home plate. And I'll ask you the question. Can you build one of those? I don't know. Let's give it a shot. We're going to assume that this object is hanging, floating in space. So therefore, all of these lines around the boundary are boundary lines, like so. Now that gets us off to a good-- it's just hanging in space, Christopher, all right? You look confused. It's just hanging in space so that all the lines around the edges are places where you see only one side has stuff on it. So that enables us to just quickly run around the periphery and put arrow labels on all those outside lines. Now we have a lot of arrow-style junctions on the boundary. That's commonly the case. So we can run over to our catalog of all possible labels, and we see that if we have an arrow with boundaries on its barbs, there's only one of those. So I know instantly that there must be a plus on the shaft. So we can come back here and take all these arrows here. And label them with plus lines on their shafts. Now a line can't change its nature along its length. So if it's a plus line on one end, it's going to be a plus line on the other end. All right? So what else can we do? Here deep inside is a fork-style junction. It's got convex markers on both of those two lines. So we go over to our catalog and say, what can we say about it, given that there are pluses on two of its lines? Whoop, that means that the third one has to be a plus as well. And now we're done. We've labeled everything. Except-- look at this. What about that guy? There's an L-style junction with pluses on both of its two lines. Is there one of those in my catalog? No. Therefore, I haven't passed a necessary condition for constructability in the world that I've made. You can't make it. You can't construct it. So Huffman's ideas give us a way of testing something to see if it's not possible for it to be in this world. If it passes the test, does that mean it's possible? No. It's a necessary but not sufficient condition. On this one-- blue-collar occupation-- on this one, maybe it will help if we put in another line. For example, we could put a line like so. You feel better about it now? I don't know. Let's see. This has to be a plus for the same argument we used on several other arrows before. That gives us an arrow-style junction here with a plus on everything. Is there such a junction label? No. We lose. It doesn't help. You think you can make it, but you can't. STUDENT: You could actually construct it as a 3-D object, though. PATRICK WINSTON: He thinks you can construct it as a 3-D object. Let me show you the next example, Christopher. Consider this example. Can you make that? Your intuition is yes. So let's label it. Oh, I've already lost. We just boost that up a little bit to make the situation more clear. So already, I've got myself in a situation where I can't label that. But you feel like you can make it. So what's wrong? What's wrong is-- what, Elliott? STUDENT: You have an obscured-- or, we're presuming that we have an obscured [INAUDIBLE] alley from the upper-left corner to the [INAUDIBLE]. PATRICK WINSTON: Putting a little interpretation on what Elliott has said. If you look at this situation back here, you get a four-faced junction there. So you can make it. But not with three faces. So some of these look like you could make it. But they can't be labeled because you need more than four faces at a junction. And we can carry that idea back here. You can make this OK. But this junction, you've got two in the back and two here. So it has four faces. Same idea. So that's Guzman's contribution. That's Huffman's contribution. Huffman was a mathematician. But we wanted to build robots back in those days. And neither one of these guys solved the problem of dealing with natural objects with shadows, with cracks, with more than trihedral vertexes in space, and what to do about that? Well, that was a problem that another graduate student, David Waltz, set about to solve. So Waltz decided that he would not be content unless he added cracks, shadows, non-trihedral vertexes. Uh-- yeah, non-trihedral vertexes. And light. These considerations led Waltz to go from four labels to 50-plus. Because he had to pack a lot of information into each of the labels he put on a line. What kind of light was on the right? What kind of light was on the left? Maybe it's a crack. Maybe it's a crack that-- all sorts of considerations. Here we had 18 ways that lines can come together around junctions. That went to thousands of junctions in Waltz's world. So here's Guzman. He writes a program that sort of works. Down below, we have Huffman. Huffman, who has a theory but solves the wrong problem. So here comes Waltz, and he's trying to solve the right problem with a satisfying theory that has a generalizable principle. So when we get all through this, we'll talk about criteria for success. And we'll conclude that to have a really successful thing, you need a problem, to start with. You need a method that works. And you have to show that it works because of some principal. So Guzman had the problem and something that worked. Huffman had a method which worked on the wrong problem. And it's left to Waltz to bring it all together. So Waltz does all this work. And now he has, instead at 18 labels, he has thousands. Instead of four-line labels, he has more than 50. So naturally, it becomes difficult to work these by hand. We were able to work those Huffman examples by hand. We started with labeling everything on the boundary and worked our way in. There's no particular method there. It was just easy to work out the puzzle. But poor Waltz, he didn't have that luxury. So he might have, in a typical scene, he might have tens or even hundreds of junctions to label and no easy way of dealing with it. So naturally writes himself a depth-first search program. So here is vertex, or rather junction number 1. There are many choices for which label can be used on it. And for each of those choices, whatever he's decided junction number 2 is has its own suite of possibilities. And so it becomes a simple depth-first search problem, right? So in actuality, as soon as Waltz-- he was my office mate at the time. I can tell you this for a fact. As soon as he wrote this program, he kept looking over at the computer-- they were big in those days. They all had lights. So you looked over at the computer to see if the lights were still blinking. Because he'd start this depth-first search program up and nothing would happen. He thought the computer had crashed. Nothing happened. Why did nothing happen? Because the search base is exponential and much too big for an ordinary computer, or maybe even an ordinary universe. So Waltz has to do something else. He has to come up with a new method for using all these labels that he's-- after about a year and a half's worth of hard work, with lots of paper on his desk in little blocks. After year and a half of hard work getting all these junction labels figured out, he then has to come up with a method for figuring out how to use them. And so we don't know whether to think his biggest contribution was that label set or his method. And probably both deserve about equal billing. Oh, I don't know how to explain what Waltz did. Well, one way is to do an example. And I think I will hazard an example. Let's see. Let me find some space. I think I'm reduced to going over here. But that will be convenient, since the line labels are here. Here's my example. And you say, how can I give you just part of a picture? Well, you can assume I'm looking at this through a window. So the edge of the window form boundary lines, and they exert no constraint whatsoever on what's behind them. So this is a legitimate drawing to have to think about. By the way, is this ambiguous? Or do you get a firm sense that there's a unique interpretation of all those lines? I think there's a unique interpretation of all those lines. What I'm going to do is I'm actually going to solve this problem using Huffman's labels but Waltz's method. Because I can't simulate on the blackboard something with 50 line types and thousands of line junctions. So I'm going to use Huffman's set to demonstrate Waltz's algorithm. So Waltz's algorithm involves starting out by plopping on some junction all of the possible labels that the answer has to be drawn from. So let me number these in the order that we're going to visit them. Like so. And so far, I've just put down the three fork options that are resident on that first junction. And I have to take note of exactly what they do with the lines that come out of the junction. So let's see. I'll just copy them down. One possibility is this one. Another possibility is this one. And another possibility is plus, plus, minus. Oops, I've got plus, plus. No, that's right. So that right so far? All I've done is copy the junction labelings from my library. And at this point, Waltz's algorithm says there's nothing else to do but go on to junction number two. And unfortunately, sadly, there are lots of labelings that have to be considered on junction number 2. Six of them. 1, 2, 3, 4, 5, 6. So one of them looks like that. Another one looks like that. One of them is plus here, arrow in. Another one is plus here, arrow out. Another one is minus here, arrow down. And minus here, arrow up. I think I've copied those all right. But now, having copied those down, Waltz's algorithm looks around at the neighboring junctions and says, are any of the things that I just placed on junction two disallowed by what I've already placed on a neighboring junction? So it looks over here in step number two. And it sees that these three arrows require the line that joins junctions 1 and 2 to be either minus or plus. So of the six possibilities, I can only keep the ones that are likewise content to put a plus on that line that joins the two. So that means that the influence flowing from junction 1 eliminates that one, eliminates that one, eliminates that one, and eliminates that one. So half of them are gone. All the ones that try to put a boundary line on that line between 1 and 2 are disallowed. All right? Now likewise, we could say, well, of the remaining ones, do they restrict what I can do at junction 1? So let's see. Here's a minus. And here's a plus. So all these possibilities over here are still alive. So now, continuing on, we have to see what we can do at junction 3. These are arrow labels again. So we have to copy exactly the same labels set as we had before. And now we look down at junction 2 and say, well, what does that tell me about the three that I've just placed? If we look up from junction 2 to see what kind of constraints it puts on here, we have this one alive and this one alive. I guess we've eliminated four of the six. So we have these two alive. And they both but boundary lines-- I think I must have had this boundary line wrong, right? No, that's right. Oh yes, I see. Plus. This one goes-- hang on a second. You let me do something wrong. So plus is out. And that must be one that goes-- this minus goes up. Oh, yes. I'm too hasty and uncertain about what I was doing. So let's see. This guy has a boundary going down. And this guy has a boundary going up. All of the others have been eliminated. So that means that something that tries to put a concave line there is gone. And something that tries to put a plus line there is gone. So the influence flowing up in this direction in the third step eliminates that guy and eliminates that guy, leaving only this guy. But now, the thing I was worried about is you have to also at this point go down to 2 and see if there's any further constraint on what can survive down there, based on what has happened over here at junction 3. Now let's see. This one goes up, which is compatible with a survivor. But this one goes down, which is not compatible with a survivor. So when I bring this down in step three, this guy is eliminated. So now I'm down to just one interpretation for what can be going on at vertex number 2. And one interpretation for vertex number 3. Now let's see. This can propagate. So now that I've made a change on vertex number 2, I have to also see if that causes a change at vertex number 1. So it's propagating through. And now I can see that the only possibility here is a minus, the label that's coming down from our survivor. So that eliminates these two. Whew! It's hard to do this by hand, but I've got three of the four things labeled. And even with just three of the four labeled, I'm down to a single interpretation for all of the junctions and the lines between them. So there's one left. We have to deal with that fork vertex. We better deal with it, because for all we know, this is not a legal drawing in this world. We have five fork vertexes to place. But you know what? I don't have to draw much here, because I know this is forced to be a plus now. And this is forced to be a plus now. And there's only one fork vertex with any pluses on it at all. So now I can come through and say, well, the only possible survivor is this one. These are gone. And now I have an interpretation for all of the junctions. And I see that the winners are this one. And this one. And this one. And this one. So I've got a unique interpretation. This line is convex. This one is concave. This is a boundary. That's a boundary. And this line is convex, and that's convex. Now that's a lot of work. So I better check and make sure I got it right. You'd like to see this demonstrated to make sure I haven't made a mistake. I'm sure of that. Let's see. That it? So each of the places where a line is obscured has four possibilities, labeled E. The arrow junctions are labeled A. The forks-- there are five of them-- at the fork junction 5. So let's just step through here and see what happens. Boom. I've got it. I did do it right. So let's try some more. What do you think will happen with this one? Unique solution? It stopped. Bug in my program? Unthinkable. What's happened? It is genuinely ambiguous. It can be something hanging down from the ceiling. Or it could be something that we can think of as a step going up from left to right. Let's look at something more complicated. You think it'll work? Not enough constraint for us to figure that one out. It's equally ambiguous, but a little bit larger example. What about this one? Yeah, but the stuff is creeping up from the lower left up to the upper right. Yeah, bingo. It worked. It's unambiguous. It's variation on the same theme we had before. But let me, just for fun, take these two lines out. What do you think will happen now? Seems to be doing just fine until it hits the upper right-hand corner and discovers it can't label stuff. So it propagates back down. And what looked OK in the lower left is no good after all. So these results are kind of consistent with what we humans do when we look at these kinds of things. So it's very likely that we, in our heads, do have some constraint propagation apparatus that we use in vision. But putting that aside, we can think about other kinds of intelligence different from human, that might use this kind of mechanism to solve problems that involve a lot of constraint in finding a solution. So here, we saw the constraint propagation activity at work on line drawing analysis. But next time, what we're going do is we're going to see at work in map coloring. And who cares about map coloring? People who do scheduling, because that turns out to be the same problem. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 4_Search_DepthFirst_Hill_Climbing_Beam.txt | PATRICK WINSTON: Today we're going to be talking about Search. I know you're going to turn blue with yet another lecture on Search. Those of you who are taking computer science subjects, you've probably seen in 601. You'll see it again as theory course. But we're going to do it for a little different purpose. I want you to develop some intuition about various kinds of Search work. And I want to talk a little bit about Search as a model of what goes on in our heads. And toward the end, if there's time, I'd like to do a demonstration for you of something never before demonstrated to a 603.4 class, because it was only completed last spring. And some finishing touches were added by me this morning. Always dangerous, but we'll see what happens. There's Cambridge. You all recognize it, of course. You might want to get from some starting position s to some goal position g. So, you'll hire a cab and hope for the best. So, here's what might happen, not too hot. Let's move the starting position over here. I've had cab drivers like this New York. But it's not a very good path. It's the path of a thief. Let's change the way that the search is done to that of a beginner, an honest beginner. Not too bad. Now, let's have a look at how the Search would happen if the cab driver was a Ph.D. in physics after his third post-doc. These are not actually traverse. These are just things that the driver is thinking about, and that is the very best of all possible paths. So, the thief does a horrible job. The beginner does a pretty good job, but not an optimal job. This is the optimal job as produced by the Ph.D. in physics after his third post-doc. So, would you like to understand how those all work? The answer, of course, is yes. I'm going to talk to you about procedures that are different from the way that you just solved this problem. I imagine that if I said to you, please find a path for s to g, you would, within a few seconds, find a pretty good path-- not the optimal one, but a pretty good one-- using your eyes. And we're not going to tell you about how that works, because we don't know how that works. But we do know that problem solving with the eyes is an important part of our total intelligence. And we'll never have a complete theory of human intelligence until we can understand the contributions of the human visual system to solving everyday problems like finding a pretty good path in that map. But, alas, we can't talk about that, because we don't know how to do it. We're working on it. But we don't know how to do it. So, I'm not going to use Cambridge in my illustrations. There's too much there to work through in an hour. So, we're going to use this map over here which has been designed to illustrate a few important points. You, too, can find a path through that graph pretty easily with your eyes. Our programs don't have eyes, and they don't have visually grounded algorithms, so they're going to have to do something else. And the very first kind of search we want to talk about is called the British Museum approach. This is a slur against at least the British Museum, if not the entire nation, because the way you do a British Museum search is you find every possible path. So, it'll be helpful to have a diagram of all possible paths on the board. We're going to start with a British Museum search. From the starting position, it's clear, you can go from my s to either a or b. And already there's an important quiz point. Whenever we have these kinds of problems on a quiz, we ask you to develop the tree associated with a search in lexical order. So, the nodes there under s are listed alphabetically, just to have an orderly way of doing it. So, from a we can go either b or d. And another convention of the subject, another thing you have to keep in mind in quizzes, is it we don't have these searches bite their own tail. So, I could have said that if I'm at a, I can also go back to s. But no path is ever allowed them to bite itself, to go around and enter and get back to a place that's already on the path. Now if I go on to b first, that means that from b I can go to either a or c. This is getting fat pretty fast. But let's see, s, a, b. The only place I can go is c and then to e. s, a, d, without biting my own tail and going back to a, the only place I can go is g. s b, a, I can only go to d and then to g. And finally, s, b, c, I can only go to e. So, that is a complete set of paths as produced by any program that you will feel you'd like to write that finds all possible paths. I haven't been very precise about how to do that, because you don't have to be. You can't save much work by being clever, because you have to find everything. So, that's the British Museum expansion of the tree. So, what have I done? I've been playing around with a map. I showed you an example of a map. And pretty soon you're going to think that Search is about maps. So, before going even another tiny step, I want to emphasize that Search is not equal to maps. Search is about choice. And I happen to illustrate these searches with maps, because they are particularly cogent. But Search is not about maps. It's about the choices you make when you're trying to make decisions. These things I'm going to be talking to you about today are choices you make when you explore the map. You can make other kinds of choices when you're exploring other kinds of things. And, in fact, at the end, if there's time, I'll show you how you do searches when you're solving problems in a humanities class. That's the British Museum algorithm. Search is not about maps. Our first gold star idea, Search is about choice. But for our illustration, Search is about maps. So, the first kind of Search we want to talk about that's real is Depth-first Search. And the idea of Depth-first Search is that you barrel ahead in a single-minded way. So, from s, your choices are a or b. And you always go down the left branch by convention. So, from s, we go to a. From a we have two choices. We can go to either b or d following our lexical convention. After that, we can go to c. And after that we can go to e. And too bad for us, we're stuck. What are we going to do. We've got into a dead end, all is lost. But of course, all isn't lost. Because we have the choice of backing up to the place where we last made a decision and choosing another branch. So, that process is called variously back-up or backtracking. At this point, we would say, ah, dead end. The first place we find when we back up the tree where we made a choice is when we chose b instead of d. So, we go back up there and take the other route. s, a, d now goes to g. And we're done. We're going to make up a little table here of things that we can embellish our basic searches with. And one of the things we can embellish our basic searches with is this backtracking idea. Now, backtrack is not relevant to the British Museum algorithm, because you've got to find everything. You can't quit when you've found one path. But you'd always want to use backtracking with Depth-first Search, because you may plunge on down and miss the path that gets to the goal. Now, you might ask me, is backtracking, therefore, always part of Depth-first Search? And you can read textbooks that do it either way. Count on it. If we give you a Search problem on a quiz, we'll tell you whether or not your Search is supposed to use backtracking. We consider it to be an optional thing. You'd be pretty stupid not to use this optional thing when you're doing Depth-first Search. But we'll separate these ideas out and call it an optional add-on. so, that's Depth-first Search, very simple. Now, the natural companion to Depth-first Search will be Breadth-first Search, Breadth-first. And the way it works is you build up this tree level by level, and at some point, when you scan across a level, you'll find that you've completed a path that goes to the goal. So, level by level, s can go to either a or b. a can go either to b or d. And b can go to either a or c. So, you see what we're doing. We're going level by level. And we haven't hit a level with a goal in it yet, so we've got to keep going. Note that we're building up quite a bit of stuff here, quite a lot of growth in the size of the path set that we're keeping in mind. At the next level, we have b going to c, d going to g, a going to d, and c going to e. And now, when we scan across, we do hit g. So, we found a path with Breadth-first Search, just as we found a path with Depth-first Search. Now, you might say, well, why didn't you just quit when you hit g? Implementation detail. We'll talk about a sample implementation. You can write it in any way you want. But now that we know what these searches are, let's speed things up a little bit here and do a couple searches that now have names. The first type will be Depth-first, boom. That's the one that produces the thief path. And then we can also do a Breadth-first Search, which we haven't tried yet. What do you suppose is going to happen? Is it going to be fast, slow, produce a good path, produce a bad path? I don't know, let's try it. I had to speed it up, you see, because it's doing an awful lot of Search. It's generating an awful lot of paths. Finally, you got a path. Is it the best path? I don't think so. But we're not going to talk about optimal paths today. We're just going to talk about pretty good paths, heuristic paths. Let's move the starting position here in the middle. Do you think Breadth-first Search is going to be stupid? I think it's going to be pretty stupid. Let's see what happens. This Search is a lot to the left, which you would never do with you eye. Let me slow that down just to demonstrate it. It finds a shorter path, because it's right there in the middle. But it spends a lot of its time looking off to the left. It's pretty stupid. But that's how it works. So, now that we've got two examples of searches on the table, I'd like to just write a little flow chart for how the search might work. Because if I do that, then it'll be easier for us to see what kind of small differences there are between the implementations of these various searches. So, what we're going to do is we're going to develop a waiting list, a queue, a line, whatever you'd like to call it. Let's call if a queue. We're going to develop a queue of paths that are under consideration. So, the first step in our algorithm will be to initialize our queue. And I think what I'll do is I'll simulate Depth-first Search on this problem up there on the left using this algorithm. I need to have some way of representing my paths. And what I want to do is I'm going to betray my heritage as a list programmer, because I'm just going to put these up as if there were lisp s-expressions. To begin with, I just have one path. And it has only one node in it, s. That's the whole path. The next thing I do after I initialize the queue is I extend first path on the queue. OK, when I extend s, I get two paths. I get s goes to a, and I get s goes to b. I take the first one off the front of the queue. And I put back the two that are produced by extending that path. Now, after I've extended the first path on the queue, I have to but those extended paths on to the queue. In here there's an explicit step where I've checked to see if that first path is a winner. If it's not, I extend it. And I have to put those paths onto the queue. So, I'll say that what I do is I end queue. Now, I've done one step. And let's let me do another step. I'm going to take this first path off. I'm going to extend that path. And where do I put these new paths on the queue if I'm doing Depth-first Search? Well, I want to work with the path that I've just generated. I'm taking this plunge down deep into the search tree. So, since I want to keep going down into the stuff that I just generated, where then do I want to put these two paths? At the end of the queue? I don't think so, because it'll be a long time getting there. I want to put them on the front of the queue. For Depth-first Search, I want to put them on the front of the queue. And that's why s, a, b goes here, and s, a, d, and then that's s, b. So, s, b is still there. That's still a valid possibility. But now I've stuck two paths in front of it, both of the ones I generated by taking a path off the front of the queue, discovering that it doesn't go to the goal, extending it and putting those back on the queue. I might as well complete this illustration here. While I'm at it, I take the s, a, b off, s, a, b, and I can go only there to c. But, of course, I keep s, a, d and s, b on the queue. Now, I take the front off the queue again, and I get s, a, b, c, e, and not to forget s, a, d and s, b. I take the first one off the queue. It doesn't go to the goal. I try to extend it, but there's nothing there. I've reached a dead end. So, in this operation, all I'm doing is taking the front one off the queue and shortening the queue. We're almost home. I take s, a,d off of queue. And I get s, a, d, c. And, of course, I still have s, b. Now, the next time I visit the situation, buried in that first step, I discover a path that actually does get to goal, and I'm done. So, each time around I visualize the queue. I check to see if I'm done. If not, I take the extensions and put them somewhere on the queue. And then I go back in. And then here there's a varied test which checks to see if we're done. That's how the Depth-first Search algorithm works. And now, would we have to start all over again if we did Breadth-first Search? Nope. Same algorithm. All the code we've got needs one line replaced, one line changed. What do I have to do different in order to get a Breadth-first Search out of this instead of a Depth-first Search? Tanya? TANYA: Change [INAUDIBLE] on the queue. PATRICK WINSTON: And where do I put it on the queue? She says to change it. TANYA: On the back? PATRICK WINSTON: Put it on the back. So, with Breadth-first Search all I have to do is put on the back. Now, if we were content with a inefficient search, and didn't care much about how good our path was, we'd be done. And we could go home. But we are a little concerned about the efficiency of our search. And we would like a pretty good path. So, we're going to have to stick around for a little while. Now, you may have noticed, up there in that the development of the Breadth-first Search, that the algorithm is incredibly stupid. Why is the algorithm incredibly stupid? Ty, what do you think? TY: It can't tell whether it's getting closer or further away from the goal. PATRICK WINSTON: It certainly can't tell whether it's getting closer or further away from the goal. And we're going to deal with that in a minute. But it's even stupider than that. Why is it stupid? What's your name? DYLAN: Dylan. It [? hits ?] the same nodes twice. PATRICK WINSTON: Dylan said it's extending paths that go to the same node more than once. Let's see what Dylan's talking about. Down here, it extends a. But it's already extended a up there. Down here, it extends a path that goes to b. And it's already extended a path that goes to d. Over here, it could extend a path that went through c, but it's already got a path that goes through c. So, all of these paths are duplicated. And we're still going through them. That's incredibly stupid. What we're going to do is we're going to amend our algorithm just a little bit. And we're not going to extend the first path on the queue unless final node never before extended. What we're going to do is we're going to look to see if there-- we've got this path. And we're going to extend it. And it's got a final note. If we've ever extended a path that goes to that final node, and it was a final node on that path, then we're not going to do it again. We got to keep a list of places that have already been the last piece of a path that was extended. Everybody got that? It's a little awkward to say it, because it's the last node we care about. If a path terminates in a node, and if some other path previously terminated in that node and got extended-- we're not going to do it again. Because it's a waste of time. Now, let's see if this actually helps. Now, use the extended list. Let's see, well, gee, we got that place in the center there. Let's just repeat the previous search. Wow, it's taking a long time. But notice it put 103 paths back on the queue. Now, let's add a filter and try again. A lot less. So, let's speed this up, and we'll start way over here. You remember how tedious that search was. And now we'll repeat it with this list, boom, there it is. That's all because we didn't do that silly thing of going back through the final node that's already been gone through. So, you would never not want to do this. We better list this as another option. It doesn't help with a British Museum algorithm, because nothing helps with the British Museum algorithm. Does it help with Depth-first? Yes. Does it help with Breadth-first? Yes. Do we do backtracking with Breadth-first? No, because backtracking can't do us any good. OK, we're almost, except that search that's starting in the middle is still pretty stupid. Both the Breadth-first version and the Depth-first version are going off to the left. And we would never do that with our eyes in any case. The next thing we want to do is we want to have ourselves a slightly more informed search by taking into consideration whether we seem to be getting anywhere. So, in general, it's a good thing to get closer to where we want to go. In general, if we've got a choice of going to a node that's close to the goal or a node that's not so close to the goal, we'll always want to go to the one that's close to the goal. And as soon as we add that to what we're doing, we have another kind of Search, which goes by the name of Hill Climbing. And it's just like Depth-first Search, except instead of using lexical order to break ties, we're going to break ties according to which node is closer to the goal. I went to some trouble to talk to you about this enqueued list. And having gone to that trouble, I'm now going to ignore it. Not because it isn't a good idea, but because trying to keep track of everything in the example is confusing the example. It won't work out right in the small example and all that. Put the queueing thing aside, queued list aside, and think instead just about the value of going in the direction that's getting us closer to the goal. In Hill Climbing Search, just like a Depth-first Search, we have a and b. And we're still going to list them lexically on underneath the parent node. But now which one is so closer to the goal? Now, this time b is closer to the goal than a. So, instead of following the Depth-first course, which would take us down through a, we're going to go to the one that's closest which goes through b. And b can either go to a or c. b is six units away from the goal. a is about seven plus, not drawn exactly to scale. Use the numbers not your eyes. Now where are we? It's symmetric, so a and c are both equally far from the goal. Now we're going to use the lexical order to break the tie. Now from s, b, a, we'll go to d. And now, which is closest to the goal? That's the only choice we have. So, now we have no choice but to go down to the goal. That's the Hill Climbing way of doing the search. And notice that this time there's no backtracking. It's not the optimal path. It's not the best path. But at least there's no backtracking. That's not always true. That's just an artifact of this particular example. Do you think Hill Climbing would produce a faster search? I think so. Let's see what happens when we add these things at one at a time. First, let's turn off our extended list. We turned off our extended list. And we're going to do Depth-first again just for the sake of comparison. It produces a very roundabout path with 48 enqueueings. Now, let's switch over to Hill Climbing. And what do think? Do you think it will produce a straighter path, fewer enqueueings? Boom. You wouldn't not want to do that, would you? If you've got some kind of heuristic that tells you that you're getting close to the goal, you should use it. Now, it's easy to modify my example over there so that getting close to the goal gets you trapped in a blind alley on e. That's easy to do. But that's just an artifact of the example. In general, you want to go along a path that gets you closer to the goal. So, that's 23. I don't know, let's see if using the extended list filter does any good. Yeah, still 23. So, in that particular case the extension list didn't actually do us any good, because we're driving so directly toward the goal. OK, that's that. Now, let's see, is there any analog to-- well, we might say that this is yet another way of distinguishing the searches. And that is, is it an informed search? Is it making use of any kind of heuristic information? Certainly, a British Museum is not, Depth is not, Breadth is not. And now let's consider what we got for Hill Climbing. Do we want to use backtracking? Sure. Do we want to use an enqueued list? Sure. And it is informed, because it's taking advantage of this extra information. It may not be in your problem. It's not often the case you've got this information in a map. Your problem may not have any heuristic measurement of distance to the goal. In which case, you can't do it. But if you've got it, you should use it. Oh, yeah, there's one more. And I've already given it away by having it on my chart. It's called Beam Search. And just as Hill Climbing is an analog of Depth-first Search, Beam Search is a complement or addition of an informing heuristic to Breadth-first Search. What you do is you start off just like Breadth-first Search. But you say I'm going to limit the number of paths I'm going to consider at any level to some small, fixed number, like, in this case, how about two. So, I'm going to say that I have a Beam of two for my Beam Search. Otherwise, I proceed just like Breadth-first Search, b, d, a, g. And now I've got that stupid thing where I'm duplicating my nodes, because I'm forgetting about the enqueued list. But to illustrate Beam Search, what about I'm going to do now is I'm going to take all these paths I've got at the second level, and I'm only going to keep the best two. That's my beam width. And the best two are the two that get closest to the goal. So, those four, b, c, a, and d, which two get closest to the goal? Now, b and d. These guys are trimmed off. I'm only keeping two at every level. Now, going down from b and d, I have, at the next level, c and g. And now I've found the goal. So, I'm done. We could do that here, too. We could choose a Beam Search, not bad. Let's see, let's try this thing from the middle. Let's slow my speed down a little bit. Now, are we going to see anything going off to the left like we did with ordinary Breadth-first Search? No, because it's smart. It doesn't say, I want to go to a place that's further away from my goal. Now, let's see, maybe we can go back to our algorithm now and talk about that enqueueing mechanism and talk about Hill Climbing. Can I use the same basic search mechanism, just change that one line again? Yes. How do I add new paths to the queue this time? Well, it's very much like Hill Climbing, right? I want to add them to the front but with one little flourish. What's the flourish? [? Krishna, ?] what do you think? Remember, I want to use my heuristic information. So, I not only add them to the front, but amongst the ones I'm adding to the front, what do I do? AUDIENCE:Check the distance? PATRICK WINSTON: Check the distance. And how do you arrange them? AUDIENCE:[? You ?] [? keep the ?] minimum [? first. ?] PATRICK WINSTON: Yeah, you can put the minimum first if you like. But let's sort them. We'll sort them, that will keep everything straight. So Hill Climbing is front-sorted. And, finally, how about Beam? What do we do with Beam Search to add them to the queue? Well, it doesn't matter where we add them, because all we're going to do is we're going to keep the w best. So, with Beam, we'll just abbreviate that by saying keep w best. Now, you have some of the basic searches in you're toolkit. There's one more that's sometimes talked about. We've got Depth, Breadth, Best, and Beam, one more is Best, Best-first Search. It's a variant where you say, I've got this tree. It's got a bunch of paths that terminate in leaves. Let me just always work on the leaf node that's closest to the goal. It can skip around a little bit from one place to another. Because as it pursues one path, it may not do very well in some other path quite distant. And the tree will become the best one. We've actually seen an instance of that in then integration program. It's capable of skipping all over the place, because it's always taking the easiest problem in the search tree, in the and/or tree, working on that. That's Best-first Search. You can do these sorts of things in continuous spaces, too. And you've done the mathematics of that in 1802 or something. But in continuous spaces, the Hill Climbing sometimes leads to problems or doesn't do very well. What kind of a problem can you encounter in a continuous space with Hill Climbing? Well, how would you do Hill Climbing in a continuous space? Let's say we're in the mountains, and a big fog has come up. We're trying to get to the top of the hill before we freeze to death. And we take a few steps north, a few steps east, west, and south using our compass. And we check to see which direction seems to be doing the best job of getting us moving upward. And that's our Hill Climbing approach, right? We have explored four directions we can go and pick the best one. And from there, we pick four, try all those, pick the best one, and away we go. We've got ourselves a Hill Climbing algorithm. What's wrong with it? Or what can be wrong with it? Sometimes it works just fine. Yes. SPEAKER 1: You might get stuck in a local maximum. PATRICK WINSTON: We might get stuck in a local maximum. So, problem letter a is that if this is your space, it may look like that. And you may get stuck on a local maximum. Is there any other kind of problem that can come up? Well, it all depends on what the space is like. Here's a problem where the space has local maxima. Now, a lot of people have been killed on Mt. Washington when the fog comes up. And they do freeze to death, why? The reason they freeze to death is the Hill Climbing fails them, and they can't get to the top to the ranger station. And the reason is that there are large lawns on the shoulders of Mt. Washington. It's quite flat. So, it's the telephone pole problem. That space looks like this. Well, this isn't what Mt. Washington looks like. But it's the telephone pole problem. So, when you're wandering around here, the idea of trying a few directions and picking the one that's best doesn't help any, because it's flat. That can be a problem with Hill Climbing. Now, there's one more problem with Hill Climbing that most people don't know about. But it works like this. This is a particularly acute problem in high dimensional spaces. I'll illustrate it here just in two. And I'm going to switch from a regular kind of view to a contour map. So, my contour map is going to betray the presence of a sharp bridge along the 45 degree line. Now you see how you can get in trouble there. You get in trouble, because if you take a step in each direction, every direction takes you downhill. And you think you're at the top. So, suppose you're right here and you go north. That takes you down over a contour line. If you go south, that also takes you down over contour lines. Likewise, going west and east all appear to be taking you down, whereas, in fact, you're climbing a ridge. And that contour line is the highest that I've shown. So, sometimes you can get fooled-- not stuck, but fooled-- into thinking you're at the top when you're actually not. Now, this is a model something. This subject is about modeling intelligence. And this is a kind of algorithm you frequently need in order to build an intelligent system. But do we have any kind of Search happening in our heads? If we're going to model what goes on inside our heads, do we have to model any kind of searching in order to do the kinds of things that we humans do? I suppose so. Anytime we make a plan, we're actually evaluating a bunch of choices and seeing how they work. Let me see if I can illustrate it another way. This is a system that I showed you a little bit of last time. And, shoot, I might as well review one or two things here. I showed you a Macbeth story. This is the story I showed you. And if you had this in a humanities class, the simplest questions that might be asked is why did Macduff kill Macbeth down there at the bottom? Did I demonstrate the answering of questions last time, or just the development of the graph? I can't remember. But we'll do it again, anyway. This is somewhat stylized English. Just so you'll know, it doesn't have to be stylized English. This is English that's made available to the Genesis system by way of something called Story Workbench. There's no free lunch. Either you can use your human resources to rewrite the plot in third grade English. Or you can use your human resources to take a more natural, adult-type version of the story and decorate it with annotations that make it possible to absorb it. Just this summer, in a miracle of summer [? Europe, ?] [? Brit ?] [? van ?] [? Zijp-- ?] one of you-- connected these two systems together. So, we can now work with stories that are expressed in pretty natural English. Everything in our system is expressed in English, including common sense knowledge-- like if somebody kills you, you're dead-- but more importantly, for today's illustration, that reflective level knowledge, that knowledge about what revenge is. Here you are. You're in the humanities class, and someone says, what's really going on in the story? Not the details of who kills whom, but is there a Pyrrhic victory? Does somebody have a success? Is there an act of revenge? These are all kinds of things you might be asked about in some kind of humanities class. So, let me fire up the genesis system. Pray for internet connectivity. Launch the system on a read of that Macbeth story that I showed you just a moment ago. At the moment, it's absorbing information about background knowledge, and about reflective level knowledge, and all that sort of thing. It's building itself this thing we call an elaboration graph. It's not quite there yet. It's still reading background knowledge. Now it's reading Macbeth. It's building it's elaboration graph, the same thing you saw last time, except not quite. Do you see that stuff down at the bottom? Those are higher level concepts that it's managed to find in the Macbeth story. So, its found a revenge. How did it do that? It searched. It had a description of what a revenge is, and it looked to see if that pattern was exhibited in the elaboration graph. So, in a combination of things that were said explicitly and things that were produced by knee-jerk if/then rules, the elaboration graph was sufficiently instantiated that the revenge pattern could be found. That's interesting, Pyrrhic victory is a little harder. You'd probably get an a if you said, oh, there's a Pyrrhic victory in here. There it is. So, I'll blow that up a little bit so you can see what that is. You know what a Pyrrhic victory is. It's a situation where everything seems to be going good at first, and then not so hot. So, Macbeth wants to be King down here. And eventually that leads to becoming King. But too bad for Macbeth, because eventually he gets killed in consequence. So, it's a Pyrrhic victory. All that produced by Search programs who are looking through this graph. Now once you've got the capability of doing that, of course, then you can find all sorts of things. And you can report them in English. But, more interestingly, you can answer questions. Why did Macbeth-- it cares not a hoot about capitalization. ARTIFICIAL INTELLIGENCE: On a common sense level, it looks like Dr. Jekyll thinks Macduff killed Macbeth because Macbeth angered Macduff on a reflective level. It looks like Dr. Jekyll thinks Macduff killed Macbeth as part of acts of mistake, Pyrrhic victory, and revenge. PATRICK WINSTON: Pretty corny speech output. But you see the point. How did it get the stuff on the common sense level? The same way all those programs that build goal trees report, answers the questions. It's just looking locally around in the connections in the goal tree. How did it get the stuff on the reflective level? By reporting on the searches that produced information-- it does that by looking for higher level thoughts about its own thoughts and reporting in which of those higher level thoughts the incident we asked about actually occurs. So, let's see, just for fun, we might be interested in why Macbeth murdered Duncan. Wouldn't this be handy if you hadn't actually read the play, and here it is, you've got to write that paper? ARTIFICIAL INTELLIGENCE: On a common sense level, it looks like-- PATRICK WINSTON: I'll pull the plug on that, because that's just annoying. Yeah, pretty good, Macbeth wants to be King, and Duncan is the King. Let's see, why did Macbeth become King? Oh, it won't answer the question unless I spell it right. I wouldn't be able to show that to you until last spring. In fact, I wouldn't have been able to show you this today until last week with a tweak this morning. Because we've just now connected the language output to, of course, [? Cass's ?] parser system, which is running in reverse, in order to generate that English. So, that's something that has never before been seen by any eyes but me. So, that will conclude what we have to do today. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 12b_Deep_Neural_Nets.txt | The following content is provided under a Creative Commons license. Your support will help MIT OpenCourseWare continue to offer high-quality educational resources for free. To make a donation or to view additional materials from hundreds of MIT courses, visit MIT OpenCourseWare at ocw.mit.edu. [MUSIC PLAYING] PATRICK H. WINSTON: Well, what we're going to do today is climb a pretty big mountain because we're going to go from a neural net with two parameters to discussing the kind of neural nets in which people end up dealing with 60 million parameters. So it's going to be a pretty big jump. Along the way are a couple things I wanted to underscore from our previous discussion. Last time, I tried to develop some intuition for the kinds of formulas that you use to actually do the calculations in a small neural net about how the weights are going to change. And the main thing I tried to emphasize is that when you have a neural net like this one, everything is sort of divided in each column. You can't have the performance based on this output affect some weight change back here without going through this finite number of output variables, the y1s. And by the way, there's no y2 and y4-- there's no y2 and y3. Dealing with this is really a notational nightmare, and I spent a lot of time yesterday trying to clean it up a little bit. But basically, what I'm trying to say has nothing to do with the notation I have used but rather with the fact that there's a limited number of ways in which that can influence this, even though the number of paths through this network can be growing exponential. So those equations underneath are equations that derive from trying to figure out how the output performance depends on some of these weights back here. And what I've calculated is I've calculated the dependence of the performance on w1 going that way, and I've also calculated the dependence of performance on w1 going that way. So that's one of the equations I've got down there. And another one deals with w3, and it involves going both this way and this way. And all I've done in both cases, in all four cases, is just take the partial derivative of performance with respect to those weights and use the chain rule to expand it. And when I do that, this is the stuff I get. And that's just a whole bunch of partial derivatives. But if you look at it and let it sing a little bit to you, what you see is that there's a lot of redundancy in the computation. So for example, this guy here, partial of performance with respect to w1, depends on both paths, of course. But look at the first elements here, these guys right here. And look at the first elements in the expression for calculating the partial derivative of performance with respect to w3, these guys. They're the same. And not only that, if you look inside these expressions and look at this particular piece here, you see that that is an expression that was needed in order to calculate one of the downstream weights, the changes in one of the downstream weights. But it happens to be the same thing as you see over here. And likewise, this piece is the same thing you see over here. So each time you move further and further back from the outputs toward the inputs, you're reusing a lot of computation that you've already done. So I'm trying to find a way to sloganize this, and what I've come up with is what's done is done and cannot be-- no, no. That's not quite right, is it? It's what's computed is computed and need not be recomputed. OK? So that's what's going on here. And that's why this is a calculation that's linear in the depths of the neural net, not exponential. There's another thing I wanted to point out in connection with these neural nets. And that has to do with what happens when we look at a single neuron and note that what we've got is we've got a bunch of weights that you multiply times a bunch of inputs like so. And then those are all summed up in a summing box before they enter some kind of non-linearity, in our case a sigmoid function. But if I ask you to write down the expression for the value we've got there, what is it? Well, it's just the sum of the w's times the x's. What's that? That's the dot product. Remember a few lectures ago I said that some of us believe that the dot product is a fundamental calculation that takes place in our heads? So this is why we think so. If neural nets are doing anything like this, then there's a dot product between some weights and some input values. Now, it's a funny kind of dot product because in the models that we've been using, these input variables are all or none, or 0 or 1. But that's OK. I have it on good authority that there are neurons in our head for which the values that are produced are not exactly all or none but rather have a kind of proportionality to them. So you get a real dot product type of operation out of that. So that's by way of a couple of asides that I wanted to underscore before we get into the center of today's discussion, which will be to talk about the so-called deep nets. Now, let's see, what's a deep net do? Well, from last time, you know that a deep net does that sort of thing, and it's interesting to look at some of the offerings here. By the way, how good was this performance in 2012? Well, it turned out that the fraction of the time that the system had the right answer in its top five choices was about 15%. And the fraction of the time that it got exactly the right answer as its top pick was about 37%-- error, 15% error if you count it as an error if it's-- what am I saying? You got it right if you got it in the top five. An error rate on that calculation, about 15%. If you say you only get it right if it was your top choice, then the error rate was about 37%. So pretty good, especially since some of these things are highly ambiguous even to us. And what kind of a system did that? Well, it wasn't one that looked exactly like that, although that is the essence of it. The system actually looked like that. There's quite a lot of stuff in there. And what I'm going to talk about is not exactly this system, but I'm going to talk about the stuff of which such systems are made because there's nothing particularly special about this. It just happens to be a particular assembly of components that tend to reappear when anyone does this sort of neural net stuff. So let me explain that this way. First thing I need to talk about is the concept of-- well, I don't like the term. It's called convolution. I don't like the term because in the second-best course at the Institute, Signals and Systems, you learn about impulse responses and convolution integrals and stuff like that. And this hints at that, but it's not the same thing because there's no memory involved in what's going on as these signals are processed. But they call it convolutional neural nets anyway. So here you are. You got some kind of image. And even with lots of computing power and GPUs and all that sort of stuff, we're not talking about images with 4 million pixels. We're talking about images that might be 256 on a side. As I say, we're not talking about images that are 1,000 by 1,000 or 4,000 by 4,000 or anything like that. They tend to be kind of compressed into a 256-by-256 image. And now what we do is we run over this with a neuron that is looking only at a 10-by-10 square like so, and that produces an output. And next, we went over that again having shifted this neuron a little bit like so. And then the next thing we do is we shift it again, so we get that output right there. So each of those deployments of a neuron produces an output, and that output is associated with a particular place in the image. This is the process that is called convolution as a term of art. Now, this guy, or this convolution operation, results in a bunch of points over here. And the next thing that we do with those points is we look in local neighborhoods and see what the maximum value is. And then we take that maximum value and construct yet another mapping of the image over here using that maximum value. Then we slide that over like so, and we produce another value. And then we slide that over one more time with a different color, and now we've got yet another value. So this process is called pooling. And because we're taking the maximum, this particular kind of pooling is called max pooling. So now let's see what's next. This is taking a particular neuron and running it across the image. We call that a kernel, again sucking some terminology out of Signals and Systems. But now what we're going to do is we're going to say we could use a whole bunch of kernels. So the thing that I produce with one kernel can now be repeated many times like so. In fact, a typical number is 100 times. So now what we've got is we've got a 256-by-256 image. We've gone over it with a 10-by-10 kernel. We have taken the maximum values that are in the vicinity of each other, and then we repeated that 100 times. So now we can take that, and we can feed all those results into some kind of neural net. And then we can, through perhaps a fully-connected job on the final layers of this, and then in the ultimate output we get some sort of indication of how likely it is that the thing that's being seen is, say, a mite. So that's roughly how these things work. So what have we talked about so far? We've talked about pooling, and we've talked about convolution. And now we can talk about some of the good stuff. But before I get into that, this is what we can do now, and you can compare this with what was done in the old days. It was done in the old days before massive amounts of computing became available is a kind of neural net activity that's a little easier to see. You might, in the old days, only have enough computing power to deal with a small grid of picture elements, or so-called pixels. And then each of these might be a value that is fed as an input into some kind of neuron. And so you might have a column of neurons that are looking at these pixels in your image. And then there might be a small number of columns that follow from that. And finally, something that says this neuron is looking for things that are a number 1, that is to say, something that looks like a number 1 in the image. So this stuff up here is what you can do when you have a massive amount of computation relative to the kind of thing you used to see in the old days. So what's different? Well, what's different is instead of a few hundred parameters, we've got a lot more. Instead of 10 digits, we have 1,000 classes. Instead of a few hundred samples, we have maybe 1,000 examples of each class. So that makes a million samples. And we got 60 million parameters to play with. And the surprising thing is that the net result is we've got a function approximator that astonishes everybody. And no one quite knows why it works, except that when you throw an immense amount of computation into this kind of arrangement, it's possible to get a performance that no one expected would be possible. So that's sort of the bottom line. But now there are a couple of ideas beyond that that I think are especially interesting, and I want to talk about those. First idea that's especially interesting is the idea of autocoding, and here's how the idea of autocoding works. I'm going to run out of board space, so I think I'll do it right here. You have some input values. They go into a layer of neurons, the input layer. Then there is a so-called hidden layer that's much smaller. So maybe in the example, there will be 10 neurons here and just a couple here. And then these expand to an output layer like so. Now we can take the output layer, z1 through zn, and compare it with the desired values, d1 through dn. You following me so far? Now, the trick is to say, well, what are the desired values? Let's let the desired values be the input values. So what we're going to do is we're going to train this net up so that the output's the same as the input. What's the good of that? Well, we're going to force it down through this [? neck-down ?] piece of network. So if this network is going to succeed in taking all the possibilities here and cramming them into this smaller inner layer, the so-called hidden layer, such that it can reproduce the input [? at ?] the output, it must be doing some kind of generalization of the kinds of things it sees on its input. And that's a very clever idea, and it's seen in various forms in a large fraction of the papers that appear on deep neural nets. But now I want to talk about an example so I can show you a demonstration. OK? So we don't have GPUs, and we don't have three days to do this. So I'm going to make up a very simple example that's reminiscent of what goes on here but involves hardly any computation. What I'm going to imagine is we're trying to recognize animals from how tall they are from the shadows that they cast. So we're going to recognize three animals, a cheetah, a zebra, and a giraffe, and they will each cast a shadow on the blackboard like me. No vampire involved here. And what we're going to do is we're going to use the shadow as an input to a neural net. All right? So let's see how that would work. So there is our network. And if I just clicked into one of these test samples, that's the height of the shadow that a cheetah casts on a wall. And there are 10 input neurons corresponding to each level of the shadow. They're rammed through three inner layer neurons, and from that it spreads out and becomes the outer layer values. And we're going to compare those outer layer values to the desired values, but the desired values are the same as the input values. So this column is a column of input values. On the far right, we have our column of desired values. And we haven't trained this neural net yet. All we've got is random values in there. So if we run the test samples through, we get that and that. Yeah, cheetahs are short, zebras are medium height, and giraffes are tall. But our output is just pretty much 0.5 for all of them, for all of those shadow heights, all right, [? with ?] no training so far. So let's run this thing. We're just using simple [? backdrop, ?] just like on our world's simplest neural net. And it's interesting to see what happens. You see all those values changing? Now, I need to mention that when you see a green connection, that means it's a positive weight, and the density of the green indicates how positive it is. And the red ones are negative weights, and the intensity of the red indicates how red it is. So here you can see that we still have from our random inputs a variety of red and green values. We haven't really done much training, so everything correctly looks pretty much random. So let's run this thing. And after only 1,000 iterations going through these examples and trying to make the output the same as the input, we reached a point where the error rate has dropped. In fact, it's dropped so much it's interesting to relook at the test cases. So here's a test case where we have a cheetah. And now the output value is, in fact, very close to the desired value in all the output neurons. So if we look at another one, once again, there's a correspondence in the right two columns. And if we look at the final one, yeah, there's a correspondence in the right two columns. Now, you back up from this and say, well, what's going on here? It turns out that you're not training this thing to classify animals. You're training it to understand the nature of the things that it sees in the environment because all it sees is the height of a shadow. It doesn't know anything about the classifications you're going to try to get out of that. All it sees is that there's a kind of consistency in the kind of data that it sees on the input values. Right? Now, you might say, OK, oh, that's cool, because what must be happening is that that hidden layer, because everything is forced through that narrow pipe, must be doing some kind of generalization. So it ought to be the case that if we click on each of those neurons, we ought to see it specialize to a particular height, because that's the sort of stuff that's presented on the input. Well, let's go see what, in fact, is the maximum stimulation to be seen on the neurons in that hidden layer. So when I click on these guys, what we're going to see is the input values that maximally stimulate that neuron. And by the way, I have no idea how this is going to turn out because the initialization's all random. Well, that's good. That one looks like it's generalized the notion of short. Ugh, that doesn't look like medium. And in fact, the maximum stimulation doesn't involve any stimulation from that lower neuron. Here, look at this one. That doesn't look like tall. So we got one that looks like short and two that just look completely random. So in fact, maybe we better back off the idea that what's going on in that hidden layer is generalization and say that what is going on in there is maybe the encoding of a generalization. It doesn't look like an encoding we can see, but there is a generalization that's-- let me start that over. We don't see the generalization in the stimulating values. What we have instead is we have some kind of encoded generalization. And because we got this stuff encoded, it's what makes these neural nets so extraordinarily difficult to understand. We don't understand what they're doing. We don't understand why they can recognize a cheetah. We don't understand why it can recognize a school bus in some cases, but not in others, because we don't really understand what these neurons are responding to. Well, that's not quite true. There's been a lot of work recently on trying to sort that out, but it's still a lot of mystery in this world. In any event, that's the autocoding idea. It comes in various guises. Sometimes people talk about Boltzmann machines and things of that sort. But it's basically all the same sort of idea. And so what you can do is layer by layer. Once you've trained the input layer, then you can use that layer to train the next layer, and then that can train the next layer after that. And it's only at the very, very end that you say to yourself, well, now I've accumulated a lot of knowledge about the environment and what can be seen in the environment. Maybe it's time to get around to using some samples of particular classes and train on classes. So that's the story on autocoding. Now, the next thing to talk about is that final layer. So let's see what the final layer might look like. Let's see, it might look like this. There's a [? summer. ?] There's a minus 1 up here. No. Let's see, there's a minus 1 up-- [INAUDIBLE]. There's a minus 1 up there. There's a multiplier here. And there's a threshold value there. Now, likewise, there's some other input values here. Let me call this one x, and it gets multiplied by some weight. And then that goes into the [? summer ?] as well. And that, in turn, goes into a sigmoid that looks like so. And finally, you get an output, which we'll z. So it's clear that if you just write out the value of z as it depends on those inputs using the formula that we worked with last time, then what you see is that z is equal to 1 over 1 plus e to the minus w times x minus T-- plus T, I guess. Right? So that's a sigmoid function that depends on the value of that weight and on the value of that threshold. So let's look at how those values might change things. So here we have an ordinary sigmoid. And what happens if we shift it with a threshold value? If we change that threshold value, then it's going to shift the place where that sigmoid comes down. So a change in T could cause this thing to shift over that way. And if we change the value of w, that could change how steep this guy is. So we might think that the performance, since it depends on w and T, should be adjusted in such a way as to make the classification do the right thing. But what's the right thing? Well, that depends on the samples that we've seen. Suppose, for example, that this is our sigmoid function. And we see some examples of a class, some positive examples of a class, that have values that lie at that point and that point and that point. And we have some values that correspond to situations where the class is not one of the things that are associated with this neuron. And in that case, what we see is examples that are over in this vicinity here. So the probability that we would see this particular guy in this world is associated with the value on the sigmoid curve. So you could think of this as the probability of that positive example, and this is the probability of that positive example, and this is the probability of that positive example. What's the probability of this negative example? Well, it's 1 minus the value on that curve. And this one's 1 minus the value on that curve. So we could go through the calculations. And what we would determine is that to maximize the probability of seeing this data, this particular stuff in a set of experiments, to maximize that probability, we would have to adjust T and w so as to get this curve doing the optimal thing. And there's nothing mysterious about it. It's just more partial derivatives and that sort of thing. But the bottom line is that the probability of seeing this data is dependent on the shape of this curve, and the shape of this curve is dependent on those parameters. And if we wanted to maximize the probability that we've seen this data, then we have to adjust those parameters accordingly. Let's have a look at a demonstration. OK. So there's an ordinary sigmoid curve. Here are a couple of positive examples. Here's a negative example. Let's put in some more positive examples over here. And now let's run the good, old gradient ascent algorithm on that. And this is what happens. You've seen how the probability, as we adjust the shape of the curve, the probability of seeing those examples of the class goes up, and the probability of seeing the non-example goes down. So what if we put some more examples in? If we put a negative example there, not much is going to happen. What would happen if we put a positive example right there? Then we're going to start seeing some dramatic shifts in the shape of the curve. So that's probably a noise point. But we can put some more negative examples in there and see how that adjusts the curve. All right. So that's what we're doing. We're viewing this output value as something that's related to the probability of seeing a class. And we're adjusting the parameters on that output layer so as to maximize the probability of the sample data that we've got at hand. Right? Now, there's one more thing. Because see what we've got here is we've got the basic idea of back propagation, which has layers and layers of additional-- let me be flattering and call them ideas layered on top. So here's the next idea that's layered on top. So we've got an output value here. And it's a function after all, and it's got a value. And if we have 1,000 classes, we're going to have 1,000 output neurons, and each is going to be producing some kind of value. And we can think of that value as a probability. But I didn't want to write a probability yet. I just want to say that what we've got for this output neuron is a function of class 1. And then there will be another output neuron, which is a function of class 2. And these values will be presumably higher-- this will be higher if we are, in fact, looking at class 1. And this one down here will be, in fact, higher if we're looking at class m. So what we would like to do is we'd like to not just pick one of these outputs and say, well, you've got the highest value, so you win. What we want to do instead is we want to associate some kind of probability with each of the classes. Because, after all, we want to do things like find the most probable five. So what we do is we say, all right, so the actual probability of class 1 is equal to the output of that sigmoid function divided by the sum over all functions. So that takes all of that entire output vector and converts each output value into a probability. So when we used that sigmoid function, we did it with the view toward thinking about that as a probability. And in fact, we assumed it was a probability when we made this argument. But in the end, there's an output for each of those classes. And so what we get is, in the end, not exactly a probability until we divide by a normalizing factor. So this, by the way, is called-- not on my list of things, but it soon will be. Since we're not talking about taking the maximum and using that to classify the picture, what we're going to do is we're going to use what's called softmax. So we're going to give a range of classifications, and we're going to associate a probability with each. And that's what you saw in all of those samples. You saw, yes, this is [? containership, ?] but maybe it's also this, that, or a third, or fourth, and fifth thing. So that is a pretty good summary of the kinds of things that are involved. But now we've got one more step, because what we can do now is we can take this output layer idea, this softmax idea, and we can put them together with the autocoding idea. So we've trained just a layer up. And now we're going to detach it from the output layer but retain those weights that connect the input to the hidden layer. And when we do that, what we're going to see is something that looks like this. And now we've got a trained first layer but an untrained output layer. We're going to freeze the input layer and train the output layer using the sigmoid curve and see what happens when we do that. Oh, by the way, let's run our test samples through. You can see it's not doing anything, and the output is half for each of the categories even though we've got a trained middle layer. So we have to train the outer layer. Let's see how long it takes. Whoa, that was pretty fast. Now there's an extraordinarily good match between the outputs and the desired outputs. So that's the combination of the autocoding idea and the softmax idea. [? There's ?] just one more idea that's worthy of mention, and that's the idea of dropout. The plague of any neural net is that it gets stuck in some kind of local maximum. So it was discovered that these things train better if, on every iteration, you flip a coin for each neuron. And if the coin ends up tails, you assume it's just died and has no influence on the output. It's called dropping out those neurons. And in our next iteration, you drop out a different set. So what this seems to do is it seems to prevent this thing from going into a frozen local maximum state. So that's deep nets. They should be called, by the way, wide nets because they tend to be enormously wide but rarely more than 10 columns deep. Now, let's see, where to go from here? Maybe what we should do is talk about the awesome curiosity in the current state of the art. And that is that all of [? this ?] sophistication with output layers that are probabilities and training using autocoding or Boltzmann machines, it doesn't seem to help much relative to plain, old back propagation. So back propagation with a convolutional net seems to do just about as good as anything. And while we're on the subject of an ordinary deep net, I'd like to examine a situation here where we have a deep net-- well, it's a classroom deep net. And we'll will put five layers in there, and its job is still to do the same thing. It's to classify an animal as a cheetah, a zebra, or a giraffe based on the height of the shadow it casts. And as before, if it's green, that means positive. If it's red, that means negative. And right at the moment, we have no training. So if we run our test samples through, the output is always a 1/2 no matter what the animal is. All right? So what we're going to do is just going to use ordinary back prop on this, same thing as in that sample that's underneath the blackboard. Only now we've got a lot more parameters. We've got five columns, and each one of them has 9 or 10 neurons in it. So let's let this one run. Now, look at that stuff on the right. It's all turned red. At first I thought this was a bug in my program. But that makes absolute sense. If you don't know what the actual animal is going to be and there are a whole bunch of possibilities, you better just say no for everybody. It's like when a biologist says, we don't know. It's the most probable answer. Well, but eventually, after about 160,000 iterations, it seems to have got it. Let's run the test samples through. Now it's doing great. Let's do it again just to see if this is a fluke. And all red on the right side, and finally, you start seeing some changes go in the final layers there. And if you look at the error rate down at the bottom, you'll see that it kind of falls off a cliff. So nothing happens for a real long time, and then it falls off a cliff. Now, what would happen if this neural net were not quite so wide? Good question. But before we get to that question, what I'm going to do is I'm going to do a funny kind of variation on the theme of dropout. What I'm going to do is I'm going to kill off one neuron in each column, and then see if I can retrain the network to do the right thing. So I'm going to reassign those to some other purpose. So now there's one fewer neuron in the network. If we rerun that, we see that it trains itself up very fast. So we seem to be still close enough to a solution we can do without one of the neurons in each column. Let's do it again. Now it goes up a little bit, but it quickly falls down to a solution. Try again. Quickly falls down to a solution. Oh, my god, how much of this am I going to do? Each time I knock something out and retrain, it finds its solution very fast. Whoa, I got it all the way down to two neurons in each column, and it still has a solution. It's interesting, don't you think? But let's repeat the experiment, but this time we're going to do it a little differently. We're going to take our five layers, and before we do any training I'm going to knock out all but two neurons in each column. Now, I know that with two neurons in each column, I've got a solution. I just showed it. I just showed one. But let's run it this way. It looks like increasingly bad news. What's happened is that this sucker's got itself into a local maximum. So now you can see why there's been a breakthrough in this neural net learning stuff. And it's because when you widen the net, you turn local maxima into saddle points. So now it's got a way of crawling its way through this vast space without getting stuck on a local maximum, as suggested by this. All right. So those are some, I think, interesting things to look at by way of these demonstrations. But now I'd like to go back to my slide set and show you some examples that will address the question of whether these things are seeing like we see. So you can try these examples online. There are a variety of websites that allow you to put in your own picture. And there's a cottage industry of producing papers in journals that fool neural nets. So in this case, a very small number of pixels have been changed. You don't see the difference, but it's enough to take this particular neural net from a high confidence that it's looking at a school bus to thinking that it's not a school bus. Those are some things that it thinks are a school bus. So it appears to be the case that what is triggering this school bus result is that it's seeing enough local evidence that this is not one of the other 999 classes and enough positive evidence from these local looks to conclude that it's a school bus. So do you see any of those things? I don't. And here you can say, OK, well, look at that baseball one. Yeah, that looks like it's got a little bit of baseball texture in it. So maybe what it's doing is looking at texture. These are some examples from a recent and very famous paper by Google using essentially the same ideas to put captions on pictures. So this, by the way, is what has stimulated all this enormous concern about artificial intelligence. Because a naive viewer looks at that picture and says, oh, my god, this thing knows what it's like to play, or be young, or move, or what a Frisbee is. And of course, it knows none of that. It just knows how to label this picture. And to the credit of the people who wrote this paper, they show examples that don't do so well. So yeah, it's a cat, but it's not lying. Oh, it's a little girl, but she's not blowing bubbles. What about this one? [LAUGHTER] So we've been doing our own work in my laboratory on some of this. And the way the following set of pictures was produced was this. You take an image, and you separate it into a bunch of slices, each representing a particular frequency band. And then you go into one of those frequency bands and you knock out a rectangle from the picture, and then you reassemble the thing. And if you hadn't knocked that piece out, when you reassemble it, it would look exactly like it did when you started. So what we're doing is we knock out as much as we can and still retain the neural net's impression that it's the thing that it started out thinking it was. So what do you think this is? It's identified by a neural net as a railroad car because this is the image that it started with. How about this one? That's easy, right? That's a guitar. We weren't able to mutilate that one very much and still retain the guitar-ness of it. How about this one? AUDIENCE: A lamp? PATRICK H. WINSTON: What's that? AUDIENCE: Lamp. PATRICK H. WINSTON: What? AUDIENCE: Lamp. PATRICK H. WINSTON: A lamp. Any other ideas? AUDIENCE: [INAUDIBLE]. AUDIENCE: [INAUDIBLE]. PATRICK H. WINSTON: Ken, what do you think it is? AUDIENCE: A toilet. PATRICK H. WINSTON: See, he's an expert on this subject. [LAUGHTER] It was identified as a barbell. What's that? AUDIENCE: [INAUDIBLE]. PATRICK H. WINSTON: A what? AUDIENCE: Cello. PATRICK H. WINSTON: Cello. You didn't see the little girl or the instructor. How about this one? AUDIENCE: [INAUDIBLE]. PATRICK H. WINSTON: What? AUDIENCE: [INAUDIBLE]. PATRICK H. WINSTON: No. AUDIENCE: [INAUDIBLE]. PATRICK H. WINSTON: It's a grasshopper. What's this? AUDIENCE: A wolf. PATRICK H. WINSTON: Wow, you're good. It's actually not a two-headed wolf. [LAUGHTER] It's two wolves that are close together. AUDIENCE: [INAUDIBLE]. PATRICK H. WINSTON: That's a bird, right? AUDIENCE: [INAUDIBLE]. PATRICK H. WINSTON: Good for you. It's a rabbit. [LAUGHTER] How about that? [? AUDIENCE: Giraffe. ?] PATRICK H. WINSTON: Russian wolfhound. AUDIENCE: [INAUDIBLE]. PATRICK H. WINSTON: If you've been to Venice, you recognize this. AUDIENCE: [INAUDIBLE]. PATRICK H. WINSTON: So bottom line is that these things are an engineering marvel and do great things, but they don't see like we see. |
MIT_6034_Artificial_Intelligence_Fall_2010 | MegaR4_Neural_Nets.txt | The following content is provided under a Creative Commons license. Your support will help MIT OpenCourseWare continue to offer high quality educational resources for free. To make a donation or view additional materials from hundreds of MIT courses, visit MIT OpenCourseWare at ocw.mit.edu. PROFESSOR: Today we are introducing an exciting new pledge in 6034. Anyone who has already looked at any of the neural net problems will have easily been able to see that even though Patrick only has them back up to 2006 now, there's still-- well out of four tests, perhaps two or three different ways that the neural nets were drawn. Our exciting new pledge is we're going to draw them in a particular way this year. And I will show you which way, assuming that this works. Yes. We are going to draw them like the one on the right. The one on the left is the same as the one on the right. At first, not having had to explain the difference between the two of them, you might think you want the one on the left. But you really want the one on the right, and I'll explain why. The 2007 quiz was drawn, roughly similarly, to this. Although if you somehow wind up in tutorial or somewhere else doing one of the older quizzes, a lot of them were drawn exactly like this. In this representation, one thing I really don't like, is that the inputs are called x's, and the outputs are called y's, but there's two x's, so the inputs are not x and y, and then they often correspond to x's of a graph, and then people get confused. Additional issues that many people have are the fact that the summation and the multiplication with the weight is implied. The weights are written on the edges, where outputs and inputs go, and the summation of the two inputs into the node are also implied. But take a look here. This is the same net. These w's here would be the w's that are written onto these lines are here. Actually the better way to draw it would be like so, since each of these can have their own w, which is different. So each of the w's that are down here, are being explicitly set to a multiplier. Where as here, you just had to remember to multiply the weight by the input that was coming by. Here you see an input, comes to a multiplier, you multiply by the weight, then once you multiplied all the inputs by the weight, then you send them to a sum, so the sigma is just a sum, you sum them, add them all together, send the result of that into the sigmoid function, our old buddy, 1 over 1 plus e to the negative whatever our input was, with a weight for an offset, and then we send the result of that into more multipliers with more weights, more sums, more sigmoids. So this is how it's going to look like on the quiz. And this is a conversion guide from version 0.9 data into version 1.0. So if you see something that looks like this, on one of the old quizzes that you're doing, see if you can convert it, and then solve the problem. Chances are if you can convert it, you're probably going to do fine. We'll start off not only with this conversion guide, but also-- I'll leave that up here-- also I'm going to work out the formulas for you guys one more time. These are all the formulae that you're going to need on the quiz. And then we're going to decide what will change in the formulae, if, and this is a very likely if, there seems to be good amount of times that this happens, is that the sigmoid function in those neurons out there was ever changed into some other kind of function. Hint. It's changed into a plus already in the problem we're about to do. People change it all the time into some bizarro function. I've seen arc tangent, I think. So here we go. Let's look at the front of it. First of all, sigmoid. Well our old buddy, sigmoid, I just said it a moment ago, sigmoid is 1 over 1 plus e to the minus x. Also, fun fact about sigmoid, the derivative of sigmoid, is itself-- the derivative of sigmoid is-- let's say that the sigmoid-- we'll just turn sigmoid into like the letter say y. Y is the result, right? So if you say y equals 1 over 1 plus e to the negative x, then the derivative of sigmoid is y times 1 minus y. You can also write out the whole nasty thing, it's 1 over 1 plus e to the negative x times 1 minus 1 over 1 plus e to negative x. So the nice property of sigmoid it's going to be important for us in the very near future, and that future begins now. So now the performance function. This is a function we used to tell neural nets when they inevitably act up and give us really crappy results. At first we tell them just how long they are, with our performance function. The first function can be any sane function that gives you a better score, where better can be decided as lower or higher, if you feel like, that gives you a better score, if your answers are closer to the answer you're looking for. However, in this case, we have, for a very sneaky reason, chosen the performance function to be 1/2 d, which is the desired output, minus o, the actual output squared. So we want a small, well it's negative, So we want a small negative or 0. That would mean we performed well. So why this? Well the main reason is ddx of performance is, the 2 comes down, the o is the variable that we're actually, so maybe I should say ddo, that negative comes out, we get a simple d minus o. And yeah, we're using derivatives here. So those are fine. These are two assumptions. They could be changed on your test. We're going to figure out what happens, if we change them, if we change the performance, if we change the sigmoid, that is if we change the sigmoid to some other function, what's going to happen to the next three functions, which are basically the only things that you need to know to do backpropagation. So let's look at that. First, w prime. This is the formula for a new weight. After one step of backpropagation. A new weight in any of these positions that you can see up here on this beautiful neural net. That w-- each of the w's will have to change step by step. That's, in fact, how you do the hill climbing neural nets. You change the weights incrementally. You step a little bit in the direction towards giving you your desired results until eventually, you hope, you have an intelligent neural net. And maybe you have many different training examples that you run it on, in a cycle, hoping that you don't over fit to your one sample, on a computer. But on the test, we will probably will not do that. So let's take a look at how you calculate the weights for the next level. And then you have the weights for the current level. So first things first. New weight, weight prime equals-- starts with the old weight. That has to go there because otherwise we're just going to jump off somewhere at random. We want to make a little step in some direction, so we want to start where we are, with the weight. And then we're going to add three things. So if we're talking about the weight between some i and some j-- there's some examples of the names of weights. So this is w 1 i, that's the weight between 1 and-- so this is w 1 a, it's the weight between 1 and a. This is w 2 b, which is the weight between 2 and b . Makes sense? Well makes sense so far, but what if it's just called w b, then it's the weight between-- these w's that only have one letter, we'll get to later. They're the bias. They're the offset. They are always attached to a negative 1. So you can pretty much treat them as being a negative 1 here, that is then fed into a multiplier with this w b, if you like. This is implied to be that. All of the offsets are implied to be that. So w plus sum of alpha-- why is this Greek letter? Where does it come from? How do we calculate it? Well alpha is just some value told to you on the quiz. You'll find it somewhere. There's no way you're going to have to calculate alpha. You might be asked to try to give us an alpha, but probably not. Alpha is supposed to give the size of our little steps that we take when we're doing hill climbing. Very large alpha, take a huge step. Very small alpha, take tentative steps. So alpha is there, basically, to change this answer and to make the new value either very close to w, or far from w, depending on our taste. So plus alpha times i, so i is the value coming in into the node. We're changing the weight here. So i is the value, for instance, i sub 1 here, i would be the value of WAC, i would be the value coming output of node a. WBC, i would be the output of node b. i is sometimes as little as i is the input coming in to meet that weight at the multiplier. And then it's multiplied by delta j. Your delta is the delta that belongs to these neural net nodes. What is a delta, you said? Funny you may ask. It is a strange Greek letter. It sort of comes from the fact that we're doing some partial derivatives and stuff, but the main way you're going to figure out what the deltas are are these two formulae that I've not written in yet. So hold off on trying to figure out what the delta is until-- well right now, I'm about to tell you the delta is. So the delta is basically, think of the delta as using partial derivatives to figure out which way you're going to step, when you're doing hill climbing. Because you know when you're doing hill climbing, you look around, you figure out, OK, this is the direction of the highest increase, and then you step off in that direction. So the deltas are telling you which way to step, with the weights. And the way they do that is by taking the partial derivative of-- basically you try to figure out how the weight that you're currently looking at is contributing to the performance of the net. Contributing to, either the good performance of the net, or the bad performance of the net. So when you're dealing with the weights, like WBC, WAC, that pretty much directly feed into the end of the net. They feed into the last node, and it then comes out. It's the output. That's pretty easy. You can tell exactly how much those weights, and the values coming from them, are contributing to the end. And we do that by essentially, remember what the partial derivative, so partial derivative here is, in fact, the way that the final weights are contributing to the performance, is just the performance function. Partial derivative-- I've already figured out the derivative here, it's just d minus o. This is for sort of final weights, the weights in the last level. D minus o, except we're not done yet, because when we do derivatives, remember the chain rule. To get from the end to these weights, we pass through, well it should be a sigmoid, here it's not, we're going to pretend it is for the moment, we pass through a sigmoid, and since we passed through the sigmoid, we had better take the derivative of the sigmoid function. That is, y times 1 minus y. Well what is y? What is the output of the sigmoid? It's up. So that's also multiplied by o times 1 minus o. However, there is a-- let me see, let me see, yes-- sorry, I'm carefully studying this sheet to make sure my nomenclature is exactly right for our new nomenclature, which so new and brave, that we're doing it, that we only knew for sure we're going to do it on Wednesday. So we have d minus o times o times 1 minus o. So you say, that's fine, that can get us these weights here, even this w c, how are we going to get the deltas for the new weights here? Oh, I realize-- yeah, I got it. So the delta-- by the way, this is a delta c, how is neuron c contributing to the output? Well it's directly contributing to the output , and it's got a sigmoid in it. It doesn't really, but we're pretending it does for now. d minus o times 1 minus o. What about inner node? Node d, node a, what are we going to have to do? Well the way they contribute to the output is that they contribute to node c. So we can do this problem recursively. So let's do this recursively. First of all, as you have probably figured out, all of them are going to have an o times 1 minus o factoring from the chain rule, because they're all sigmoid, pretending that they're all sigmoids. We also have a dearth of good problems that are actually sigmoid on the web right now. There's only 2007. But here's o times 1 minus o, what are we going to do for the rest of it? How does it contribute to our final result? Well it contributes to our final result recursively. So we're talking about delta i. I is an inner node. It's not a final node. It's somewhere along the way. So sum over j of w, going from i to j, times delta j. Now sum over all j, j such that i leads to j. I needs to have a direct path into j. So if i, in this instance, was j, everyone, the only possible j in this would be c. That's right. We would not sum over b as one of the j. i does not lead to b, or a does not lead to b, a only leads to c. Also note that c does not need to be here. That's going backwards. So you just-- to figure out which j you're looking at, look directly forwards at the next one. So if there was another d here, or that a does not go to d, a goes to c. You only look at the next level children, and you sum over all those children, the weight between them, multiplied by the child's delta. That makes sense, right? Because the way we affect, if the child's delta is the way the child affects the output, calling these children for a moment, and then if this one directly affects the output, then the way this one affects it is-- it affects it because it affects this, but it's also multiplied by it's weight. So in fact, for instance, if the weight between a and c were 0, then a doesn't affect the output at all, right? Because its weight is 0, and when we do this problem, we go this times 0, and then we try to add it in there, doesn't affect anything. It's weight is very high, it's going to really dominate c, and that is taken into account here, and then multiply by the delta for the right node. So the following question, and since I spent a lot of time with formulae and not that much time starting on the problem, I will not call on someone at random, but rather take a volunteer. If no one volunteers, I'll eventually tell you, which is, we've got some nice formulae on the bottom three. If we change the sigmoid function, what has to change? That's the only thing that changes in this crazy assed problem right here, which by the way, changes the sigmoid functions into adders, is that we take all of the o times 1 minus o in delta f and the delta i, and we change it to a new derivative. We then do the exact same thing that we would've done. Correct. And on a similar note, if you change the performance function, how many of these equations at all have to change out of the bottom three. Yeah. That's right, just one, just delta f. Take the d minus o, make it the new derivative of the new performance function. And in fact, delta i doesn't change at all. Does everyone see that? Because it is very common for something to be replaced, I think three of the four the quizzes that we have, replaced in some-- changed something in some way. All right. Let's go. We're going to do 2008 quiz, because it has a part of the end that screwed up everyone, and so let's make sure we get to that part. That's going to be the part that you probably care about the most at this point. So these are all adders instead of sigmoids. That means that they simply add up everything as normal, for a normal neural net, and then there's no sigmoid threshold. They just give some kind of value. Question? STUDENT: So we talked about those multiplier things, we don't have those in nodes? PROFESSOR: They're not neural net nodes. That is one of the reasons why that other form that you can see over there is elegant. It only has the actual nodes on it. It is very compact. It's one of the front we've used in the previous tests. The question is, do those multipliers count as nodes? However by not putting in the multipliers, we feel it sometimes confuses people of explicitness. The ones that are nodes will always have a label, like a or here, you see there's a sigmoid and an L1. The multipliers are there for your convenience, to remind you to multiply, and also those, if you look those sigmoids that are over there, are there for your convenience to remind you to add. In fact, the only thing that counts as a node in the neural net-- and that's a very good question-- is usually the sigmoids, here it's the adders. We've essentially taken out the sigmoids. These adders are the-- oh, here's the way to tell. If it's got a threshold weight associated with it, then it's one of the actual nodes. A threshold weight. I guess the multipliers look like they have a weight, but this is just the weight that is being multiplied in. This is our witness be multiplied in with the input, but if it has a threshold weight, like wa, wb-- oh, I promised I would tell you guys the difference between the two weights. So let's do that very quickly. The kinds of weights that, say w2b or w1a, our weight the comes between input 1 and a or between a and c, then mentally multiplying the input by this weight, and then eventually that's added together. The threshold weights, they just have like wb, wa, wc. They are essentially to decide the threshold for a success or failure, for a 1 or a 0, or anything in between, at any of the given nodes. So the idea is maybe you at some node want to have a really high cut off, you have to very high value coming in, or else it's a 0. So you put a high threshold. The weight is multiplied by negative 1. And in fact, the threshold weight won't-- one could consider if you wanted to that the threshold weight times negative 1. was also added in it that sum, instead of putting at the same location as the node. If that works better for you, when you're converting it, you can also think of it that way. Because the threshold weight is essentially multiplied by negative 1 and added in at that same sum over there. So that's another way to do it. There's a lot of ways to visualize these neural nets. Just make sure you have a way that makes sense to you, and that you can tell pretty much whatever we write, as long as it looks vaguely like that, how to get it in your mind, into the representation that works for you. Because once you have the representation right for you, you're more than halfway to solving these guys. They aren't that bad. They just look nasty. They don't bite. OK. These are just adders. So if it's just an adder, then that means that, if we take all the x inputs coming in-- let's do x and y for the moment, so we can figure out the derivative-- then what comes out after we just add up the x, what comes out, y equals x, right? We're just adding it up. Adding up all the input, we're not doing anything to it. Y equals x is what this node does. You people see that? So the derivative is just one. So that's pretty easy, because the first problem says, what is the new formula, delta f. So I'll just tell you. You guys probably figured it out. o times 1 minus o. Because we replaced d minus o with 1. OK? Makes sense so far? Please ask questions along the way, because I'm not going to be asking you guys. I'll do it myself. Question? STUDENT: Why to we use d minus o of 1? PROFESSOR: That's a good question. The reason is because I did the wrong thing. So see, it's good that you guys are asking questions. It actually should be replaced with o times 1 minus o with 1. The answer is delta f equals d minus o. So yes, perhaps I did it to trick you. No, I actually messed up. But yes, please ask questions along the way. Again, I don't have time to call on you guys at random to figure out if you guys are following along. So I'll do it myself. We're placing the o times 1 minus o with 1 because of the fact that the sigmoid is gone, and we get just delta f equals d minus o. So great. We now want to know what the equation is for delta i, at the node a. So delta a. Well let's take a look. The o times 1 minus o is gone. Now we just have the sum over j, which you guys already told me is, only c of WAC times delta c. We know that delta c is d minus o. The answer is delta a is just WAC times d minus o. That time, I got it right. I see the answer here. Though it's written in a very different format from the old quiz. Any questions on that? Well that's part a that we finished out of c. Let's go to part b. Part b is doing one step backpropagation. There's almost always going to be one of these in here. So the first thing it asks is to figure out what the output o is for this neural net if all weights are initially 1 except that this guy right here is negative 0.5. All the other ones start off as 1. Let's do a step-- oh, let's see what are the inputs. The inputs are also all 1. Desired output is also 1. And in fact, the rate constant alpha is also 1. This is the only thing that isn't 1, folks. So let's see what happens. 1 times 1 is 1, then this is a negative 1 times 1 is negative 1. That's 0. The exact same thing happens here because it's symmetrical. So these are both 0. 0 times 1 is 0, 0 times 1 is 0. Then this is negative 1 times negative 0.5 is positive 0.5, so 0 plus 0 plus a positive 0.5, the output is positive 0.5. Does everyone see that? If not, you can convince yourself that it is positive 0.5. That would be a good exercise for you, run through one forward run. The output is definitely positive 0.5. First time around. OK? Now we have to do one step of backpropagation. To do that, let's calculate all the delta so that we can calculate all the new weights, the the new weight primes. So delta c. That's easy. You guys can tell me what delta c is. We figured out what the new delta c is going to be. So simple addition or subtraction problem? Everyone, delta c is? STUDENT: 0.5. PROFESSOR: 0.5, one half, yes. All right. We know that delta a and delta b are just WAC times delta c, and WBC times delta c. So they are? STUDENT: One half. PROFESSOR: Also one half, because all the weights were 1. Easy street. OK. We've got all of the deltas are one half. And all but a few of the weights are 1. So let's figure out what the new weights are. New WAC, OK. Yeah, so let's see. What's going to be the new WAC? So the new WAC is going to be old WAC, which is 1, because all of them are 1 except for wc, plus the rate constant which is 1, times the input coming in here, but remember that was 0, so actually it's just going to be the same as the old WAC. This is a metrical problem between b and a, at the moment, this is going to be the same. All right. Somethings are going to change though. What about wc, that was the one that was actually not 1? OK. So new wc, remember, the i for wc, the i that we use in this equation is always negative 1 because it's a threshold. So we have the old wc, which is negative 0.5, plus 1 times negative 1 times delta c, which is one half. So we have negative 0.5 plus negative 0.5 equals negative 1. w 1 a, well we've got w 1 a starts out as 1. Then we also know that w 1 a is going to be equal to 1 plus 1 times the input, which is 1, times delta of a, which is one half, so 1.5. And since it's symmetrical between a and b, then w 2 b is also 1.5. And then finally, wa and wb, the offsets here, well they start at 1 plus 1 times negative 1 times 0.5. So they're both, everyone? STUDENT: One half. PROFESSOR: One half. That's right. That's right. Because negative 1 is their i. Negative 1 times one half plus positive 1 is just one half. That's one full step. Maybe a mite easier than you might be used to seeing, but there's a full step. And it asks what's going to be the output after one step of backpropagation? We can take a look. So we have 1 times the new wa, which is 1.5, you've got 1.5, then the new wa is just 0.5, now is 0.5, that's a 1 coming into an adder. We've got another 1 coming in here because it's symmetrical. So 1 and a 1, 1 times WAC is 1. 1 times WBC is 1. So we have two 1s coming in here, they're added, that's 2. Then this has become negative 1, in fact, at this point. So negative 1 times negative 1, that's 3, and the output is 3. All right. Cool. We've now finished part b, which is over half of everything. Oh no, we've not. One more thing. These are adders. They're not sigmoids. What if we train this entire neural net to try to learn this data, so that it can draw a line on a graph, or draw some lines, or do some kind of learning, to separate off the minuses from all the pluses. You've seen, maybe, and if not, you are about to in a second, because it asks you to do this in detail, than neural nets can usually draw one line on the graph for each of these, sort of, nodes in the net, because each of the nodes has some kind of threshold. And you can do some logic between them like ands or ors. What do you guys think this net is going to draw? Anyone could volunteer, I'm not going to ask anyone to give this answer. That's a little bit tricky, because usually if you had this many nodes, you could easily draw a box and box off the minuses from the pluses. However, it draws this. And it asks what is the error? The error is-- oh yeah, it even tells you the error is 1/8, because why? These are all adders. You can't actually do anything logical. This entire net boils down to just one node, because it just adds up every time. It never takes a threshold at any point. So you can't turn into logical ones and zeroes, because it's basically not digital at all, its analog. It's giving us some very high number. So it all boils down to one cut off. And that's the best one. The one that I drew right here. OK. Did that not make sense to you? That's OK. This problem is much harder. And putting them both on the same quiz, was a bit brutal, but by the time you're done with this, you'll understand what a neural net can do or not. I put these in simplified form because of the fact that we don't care about their values or anything like that. But inside of these little circles is a sigmoid, the multipliers and the summers are implied. I think in the simplified form when we're not actually doing backpropagation is easier to view it, and see how many nodes there are. For the same reason you asked your question about how many there are. So all of those big circles are node. And in those nodes is a sigmoid now, not those crazy adders. We have the following problem. We have to try to match each of a, b, c, d, e, f to 1, 2, 3, 4, 5, 6, using each of them only once. That's important, because some of the more powerful networks in here can do a lot of these. So it's like yes, the powerful networks could do some of the easier problems here, but we want to match each net to a problem it can do, and there is exactly one mapping that will map-- that is one to one, and maps exactly, uses all six of the nets to solve all six of these problems here. So some of you may be going like, what? How am I going to solve these problems? I gave away a hint before, which is that each node in the neural net, each sigmoid node can usually draw one line on the-- it can draw one line into the picture. The line can be diagonal if that nodes receives both of the inputs, which is here, i 1 and i 2. See there is an i 1 and an i 2 axis. Like x- and a y-axis. The node has to be horizontal, or vertical, if-- sorry, the line has to be horizontal or vertical if the node only receives one of the inputs. And then, if you have a deeper level, these secondary level nodes can sort of do a logical, can do some kind of brilliant thing like and or or of the first two, which can help you out. All right. And so let's try to figure it out. So right off the bat, and I hope that people will help and call this out, because I know we don't have enough time that I can force you guys to all get it. But right off the bat, which one of these looks like it's the easiest one? STUDENT: Six. PROFESSOR: Six. That's great. Six is definitely the easiest one. It's a single line. So this is just how I would have solved this problem, is find the easiest one. Which of these is the crappiest net? STUDENT: A. PROFESSOR: A is the crappiest net. But there's no way in hell that A is going to be able to get any of these except for six. So let's, right off the bat, say that six is A. All right. Six is A. That's A. We don't have to worry about A. OK. Cool. Now let's look at some other ones that are very interesting. All the rest of these draw two lines, well these three draw two lines. These three draw three lines. They draw a triangle. So despite the fact that this c is a very powerful node, that indeed, with three whole levels here of sigmoids, it looks like there's only two that's in our little stable of nets that are equipped to handle number one and two. And those are? E and F, because E and F have three nodes at the first level. They can draw three lines. And then they can do something logical about those lines, like for instance, maybe, if it's inside all of those lines. There's a way to do that. You just-- basically you can give negative and positive weights as you so choose to make sure that it's under certain ones, above other ones, and then make the threshold such that it has to follow all three of your rules. So between E and F, which one should be two and which one should be one. Anyone see? Well let's look at two and one. Which one is easier to do? Between two and one. Two. It's got a horizontal and a vertical. One has all three diagonal. And which one of these is a weaker net, between E and F. F. F has one node that can only do a horizontal, and one node that can only do a vertical line. So which one is F going to have to do? Two. And E does what? Good job, guys. Good job, you got this. So now let's look at the last three. Number three is definitely the hardest. It's an exceller. Those of you who've played around with double o 2 kind of stuff, or even just logic, probably know that there is no way to make a sort of simple linear combination in one level of logic to create an x or. x or is very difficult to create. There are some interesting problems involving trying to teach an exceller to a neural net. Because a neural net is not to be able to get the x or, because of the fact that you can tell it, OK, I want this one to be high, and this one to be low. That's fine. You say these both have to be high. That's fine. It's hard to say, it's pretty much impossible to say, this one or this one, but not the other, because need to be high in a single node, because of the fact that if you just play with it, you'll see. You need to set a threshold somewhere, and it's not going to be able to distinguish between, if the threshold is set such that the or is going to work, the whole or is going to work. It's going to accept when both of them are positive as well. So how we can do x or? We need more logic. We need to use some combinations of ands and ors in a two level way. To do that we need the deepest neural net that we have. There's only one that's capable of that. And that is? It's C. There are many different ways to do it. Let's think of a possibility. i 1 and i 2 draw these two lines. Let's call these one, two, three, four, five, node 1 and node 2 draw these two lines. And I'll just sort of draw it here for you guys. Then maybe node 3 gives value to-- yeah, let me see-- node three can give value to perhaps-- let's see-- node 3 can give value to everything that is-- there are a lot of possibilities here. Node 3 can give value to everything that is up here. Actually node 3 can give value to everything except for this bottom part, and then node 4 could give value to say-- doesn't do it yet, but there's a few-- there's a few different ways to do it if you played around. The key idea is that node 3 and node 4 can give value to some combination and or or not, and then node 5 can give value based on being above or below a certain threshold, combination of 3 and 4. You can build an exceller out of the logic gates. I will ponder on that in the back burner for a moment, as we continue onward, but clearly C has to do number three. OK. Now we're left with four and five. I think, interestingly, five looks like it may be more complicated than four, because of the fact that it needs to do both different directions instead of two of the same direction. So however, just the idea of the one with the fewer lines, being a simpler one, may not get us through here. And there's a reason why. Look what we have left to use. We have to use D or B. What is the property of the two lines that D can draw? D being the simpler one. One horizontal, one vertical, that's right. So even though it may look simpler to just have two horizontal lines, it actually requires B. B is the only one that can draw two horizontal lines because D has to draw one horizontal and one vertical. So that leaves us with, B on this, D on this. Excellent, we have a question. I would've thought it would have been possible that we had no questions, or maybe I just explained it the best I ever have. Question. STUDENT: I didn't get why B has to be two horizontal lines. PROFESSOR: All right. So the question is, I don't understand why B to be two horizontal lines. The answer is, it doesn't. B can be anything, but D can't be two horizontal lines. And so by process of elimination, it's B. Well take a look at D, right. So D has three nodes, one, two, three. Node 1 and node 2 can just draw a line anywhere they want, involving the inputs they receive. What input does node 1 receive? Let's go to node 1. So it can only make a cut off based on i 1. So therefore, it can only draw by making the cut off above and below a certain point. Node 1 can only draw vertical lines. Node 2 can only draw a horizontal line, because it can only make a cut off based on where it is an i 2. Therefore they can't both draw a horizontal. That's why this is the trickiest part. This last part, because B is more powerful. B does not only have to do two horizontal lines. It can do two diagonal lines. It can do anything it wants. It just happens that it's stuck doing this somewhat easier problem, because the fact that it is the only one left that has the power to do it. So let's see, we're done, and we'd have aced this part of the quiz that like no one got, well not no one, but very few people got, when we put it on in 2008. The only thing we have left to ask is-- let me see-- yeah, the only thing we have left to ask is what are we going to do here for this? All right. Let's see. For the x or, let's see if I can do this x or. OK. How about this one. Right. I'm an idiot. This is the easiest way. Number one draws this line. Number two draws this line. Number three ends the line, the two lines. Number three says only if both of them are true, will I accept. Number four maps the two lines. And number five ors between three and four. Thank you. No, it's not that hard. I just completely blanked, because there's another way that a lot of people like to do it. It involves drawing in a lot of lines, and then making the clef b 2. But I can't remember it at the moment. Or there any other questions? Because I think if you have a question now, like four other people have it and just aren't raising their hand. So ask any questions about this drawing thing. Question? STUDENT: Why do we do this? PROFESSOR: Why do we do this drawing thing? That's a very good question. The answer is so that you can see what kinds of nets you might need to use in these simple problems, to answer these simple problems. So that if Athena forbid that you have to use a neural net in a job somewhere to do some actual learning, and you see some sort of quality about the problem, you know not to make a net that's too simple, for instance. And you wouldn't want a net that is more complex than it has to be. So you can sort of see what the net's do at each level, and more visibly understand. I think a lot of people who drew problems like this just want to make sure people know, oh yeah, it's not just these numbers that we're mindlessly backpropagating from the other part of the problem to make them higher or lower. This is what we're doing at each level. This is the space that we're looking at. Each node is performing logic on the steps before. So that if you actually have to use a neural net later on, down the road, then you'll be able to figure out what your net's going to need to look like. You'll be able to figure out what it's doing. At least as well as you can figure out what it's doing, for a neural net, since it often will start getting up these really crazy numbers, will have all sorts of nodes in it, and like a real neural net that's being used nowadays, there'll be tons of nodes, and you'll just see the numbers fluctuate wildly, and then suddenly it's going to start working or not. That's a good question. Any other questions? We still have a few minutes. Not many, but a few. Any other questions about any of this stuff? Sorry. STUDENT: Talk about what you just asked. Just because we draw it, does the machine need to learn-- PROFESSOR: You're confused why the machine is run what, by the pictures on the right? Oh OK. Machine does not have to learn by drawing pictures and calling them in. Let me give you some real applications. My friend at the University of Maryland recently actually used neural nets because, yeah, he actually did, because of the fact that he was doing an game plan competition, where the game was not known when you were designing your AI. It had to be able to-- there was some very elegant, general game solver thing that you had be able to hook up into, and then they made up the rules, and you had a little bit of time, and then it started. Some of the AI's, what they did was, they trained, once they found out what the rules were on their own, with the rules, in his case he had a neural net, because it was so generic, you just have a web of random gook. He thought it could learn anything, and then-- he never did tell me how it went, probably didn't go well. But maybe it did. It basically tried to learn some things about the rules. Some of the other people who are more principled game players actually tried to find out fundamental properties of the space of the rules by testing a few different things, so they could view more knowledge is less search so they could do less search when the actual game playing came on. And then when the actual game playing came on, pretty much everyone did some kind of game tree based stuff. He's telling me that a lot of Monte Carlo based game tree stuff that is this very non deterministic as what they're doing nowadays, rather than what determines the alpha beta, although he said it converges to alpha beta, if you've given enough time. That's what he told me, But that someone I know who is using neural nets. I've also in a cognitive science class I took, saw neural nets that tried to attach like qualities to objects, by having just this huge, huge number of nodes in levels in between, and then eventually it was like, a duck flies, and you're like, how's it doing this again? I'm not sure, but it is. So the basic idea is that when-- one of the main reasons that neural nets were used so much back in the day is that people on many different sides of this problem, cognitive science, AI, whatever, were all saying, wait a minute, there's networks of neurons, and they can do stuff, and we're seeing it in different places. And when you've seen it in so many different places at once, must be a genius idea that's going to revolutionize everything. And so then everyone started using them to try to connect all these things together, which I think is a noble endeavor, but unfortunately people just stopped using it. It didn't work as they wanted. It turned out that figuring out our neurons worked in our head was not the way to solve all AI hard problems at once. And they fall into disfavor, although are still used for some reasons, like the sum is like that. So we wouldn't use it just to draw these pictures. The reason why we have these pictures is because we give you simple nets that you can work it out by hand on the quiz. Any net that is really used nowadays would make your head explode, if we tried to make you do something with it on the quiz. It would just be horrible. So I think that's a good question. If there's no other questions, or even if they are, because we have to head out, if there's any other questions, you can see me as I'm walking out. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 18_Representations_Classes_Trajectories_Transitions.txt | PATRICK WINSTON: We might wonder a little bit about the nature of human intelligence, and we might reflect a little bit on the kind of intelligence we've been talking about in the past few weeks. It's been an intelligence of sorts. Those programs, support vector machines, boosting, they can do really smart things. But the peculiar thing about systems that use those methods is that those systems don't have any idea about what they're doing. They don't know anything. So they don't give us very much of an insight into the nature of human intelligence. And, after all, I'd like to have a model of human intelligence because, let's face it, we're the smartest things around. So there are lots of ways to approach that question. And we'll approach that question first from a evolutionary point of view. Some scientists believe, me for instance, that we have a family tree that looks about like this. Too small to see much of that, but the main point is that we haven't been around very long. We humans have been around maybe 200,000 years, and the dinosaurs died out 60 million years ago. So in the blink of an eye, we seem to have, more or less, taken over. When you look at that family tree on a scale where you can see it, one of the characteristics is increasing brain size. There we are on the left, chimpanzee on the right. Clearly, mostly mouth, not to much brain in there. That one down below is a reconstruction of one of those pithecus type bipedal apes from about 4 million years ago or so. So we became bipedal a long time before we had much of a brain. So you might think, well, maybe brain size has a lot to do with it, and I suppose it does. So we can plot brain volume of our ancestors versus time. So the picture I just showed you was from about 3 million years ago, I guess. And then on the upper right hand corner, oh, that's not just us, that's also the Neanderthals. Their brains might have been slightly bigger than ours. So it isn't just brain size. Here's what that guy looks like. That's a Neanderthal, of course, on the left. And that's one of us on the right. Some conspicuous differences, big heads, the rib cage is kind of conical in shape, a large pelvis. People like to make a lot of speculations about how they must have moved around. But one thing is plain, they didn't amount to much. They could make stone tools, but those stone tools didn't change much over tens of thousands of years. And that was pretty much the story with us too, until something happened, probably in southern Africa, probably in a group of individuals, maybe less than 1,000. And what's the evidence for that? The evidence for that is mostly-- comes from DNA studies with a lot of probabilistic assumptions, and Monte Carlo simulations, but it seems among the competing hypothesis for how we came to populate the world, it seems that there was a group of us, homo sapiens, in Southern Africa that got something that nobody else had and the highest probability scenario is that we quickly took over. That population of homo sapiens dominated the rest, went out of Africa, and within the blink of an eye did that sort of stuff. What's that sort of stuff? Those two paintings are from Lascaux about 25,000 years ago. Paleoanthropologists, like Tattersall, take that as plain evidence that there was symbolic thought on the people who were around at that time, us homo sapiens. The head is a carving of a mastodon tusk of a woman 25,000 years ago. Also, plainly symbolic, people are making a lot of jewelry and doing self adornment. The Neanderthals never seemed to do that. That jewelry making seems to have gone back to Southern Africa, maybe 70,000 years ago. People were puncturing seashells and using them as necklaces, apparently. So something happened, and the paleoanthropologists, who write fascinating stuff, don't quite know how to talk about it other than to say it looks like we became somehow symbolic. And somehow that has something to do with language. So if you talk to Noam Chomsky he will say-- let me get this precise, this is near his quotation as I can get. He thinks it was the ability to take two concepts and put them together to form a third concept without disturbing the original concepts and without limit, and each part of that's important. The without limit part is what separates us from species that might be able to do that a little bit, but we can do it without any apparent limit. So that's a linguist speaking. He talks a lot about the merge operation, and combinators and language-- using terms foreign to us. Better not use the term "combinator," it's kind of a computer science term. But whatever it is, it seemed to happen about that time. It didn't happen slowly and proportion to brain size, it seemed to happen all of a sudden in consequence of a brain that had grown big enough to be an enablement but the capability was not what pulled evolution in that direction. So, I believe, that whatever that was, that capability, enabled humans, us humans, to tell and understand stories, and that's what separates us from the other primates. That ability to-- The symbolic ability, whatever it is, enabled storytelling and understanding. And that's what all education is about and that's why our species is special. So what we're going to talk about today is something you might think of as an instantiation of that hypothesis, one way of thinking about it. And it's a way of thinking about what the linguists would call the inner language. It's not the language with which we communicate, it's the language with which we think, which is closely related to the language with which we communicate, but may not be quite the same thing. So many of you are bilingual. Chris is bilingual. Chris, have you ever had the experience of remembering that someone said something to you, but not remembering what language they used? STUDENT: [INAUDIBLE]. PATRICK WINSTON: How about [? you, Sian, ?] have you had that experience? STUDENT: [INAUDIBLE]. PATRICK WINSTON: What? STUDENT: I always [INAUDIBLE]. PATRICK WINSTON: [INAUDIBLE] experience of having-- David, have you ever had that experience of remembering some conversation but not remembering the language in which it was cast? STUDENT: Well, you remember something you usually don't know which language it was in anyway. PATRICK WINSTON: You usually don't remember. That's a common view. You remember something was said, that there was a conversation that had some content, but if it's with a speaker of your own language and your embedded in another place, you often don't remember what language the conversation was in. Is that right, [? Wana, ?] you remember things like that? Not sure? STUDENT: People who [INAUDIBLE] speak in a certain language, so-- PATRICK WINSTON: Sometimes you don't have that confusion, she says, because you always speak to particular people in a particular language. But many people report that they have that experience of not remembering which language something was said in. Well, OK, so what are we going to do? We need an inner language, and maybe we can start just by saying, let's have something that looks sort of familiar to us. We have an object and it's supported by some other objects, so those are support relations. That's one example of what we might call a semantic net. It's a network that's got nodes and links, it's got-- it has some meaning. That's where the word "semantic" comes from. Well, we might have another example that looks like this. There's Macbeth, there's Duncan, and Macbeth murders Duncan, then we also know, somehow, there's a kill involved as a consequence, and then ultimately, Duncan has a property. And that property is the property of being dead. So there's another semantic net recording something that happens in Shakespeare's Macbeth plot. Now, we can decorate that a little bit. So as to get a couple of other concepts in play. First of all, the thing we've got already is we've got combinators. Then combinators. Well, a fancy name for those links that connect the nodes. Another thing we've got is an opportunity for connecting the links themselves. So the murder sort of implies the kill, and the kill leads us to conclude that the victim is dead. So that is treating the links themselves as objects that can be the subject or object of other links. So we call that process "reification." Now, in artificial intelligence, semantic nets we're all over the early work, but if you have a big network that covers the wall you need some way of putting a spotlight on some pieces of it. So Marvin Minsky put a lot of technical content into that idea and created the notion of-- a notion that deserves another color here-- he suggested that we need a localization process, so we have frames, so-called frames or templates. And a frame for this murder action might be that there's a murder frame that has an agent and has a victim, and the agent is Macbeth, and the victim is Duncan. So that's a way of putting a localization layer on top of what we've got so far. Later on, I'll add sequence to that list. So this is where it rested for a long time, and in some sense, still rests there because as soon as you've got combinators, you've got something that's pretty much universal. You can do anything with it. The trouble is it's sort of down at the bit level. It's like assembly code. It doesn't have, as a concept, enough organization to help you go to the next level of achievement. There's also a little problem here that deserves also, some mention, and that is that we have, over this whole thing, the problem of parasitic semantics. Parasitic semantics. A kind of ugliness that surrounds this whole concept because when we look at a diagram like that, and we say, oh, Macbeth murdered Duncan, that means Duncan's the victim. We know there must have been a motive, maybe Macbeth wanted to be king. Well, we know all that stuff, and there's a tendency to project that knowing into the machine. If you're going to play with your telephone, please leave. So if we project meaning into that that's our understanding that's not the machine's understanding. So much of the meaning can be said to be parasitic. We're the parasite, and we're projecting the into that thing. Putting that diagram into some machine form doesn't mean the machine knows anything. It might be able to conclude some things, but it's understanding is not grounded in any kind of contact with the physical world. So we have to worry a lot about that and where philosophers would stop there and go off and write a few books on the subject. But we're not philosophers so we're going to just mention the problem and go barreling ahead. So we need to use this notion of semantic net, and we have to ask ourselves some questions about what elements of the inner language are most useful and yet it might be very complicated, but here's usefulness number one. The notion of classification. So we know about stuff, and we know about, for example, pianos. And we know about tools. And we know about maps. But we know about those things on different levels. So say I'm thinking about a tool, do you have a very good image of what I'm talking about? The answer has to be no because the notion of a tool is very vague, so it's hard to form a picture of what that's all about. On the other hand, if I said I'm thinking about a mac. Well, this is interesting because there's lexical ambiguity there. You don't know if I'm talking about the Apple type Mac or the apple type Mac, or should I say the fruit or the computer? So there's lexical ambiguity there at two levels or more. But let's fill this in a little bit. If I know I'm talking about a piano, you can form a picture of that so that seems to be on a more detailed level where you can do hallucination. At a higher level you have just a musical instrument. And I can give you a tool to think about by writing hammer. And if I'm going to have a mac, it's going to be an apple. And in this case, I want you to think about a fruit. And down here, I can be more specific about these things too. I can add a slight refinement of detail and say, I'm thinking about one of these. Do you know what this is? STUDENT: [INAUDIBLE]. PATRICK WINSTON: No, it's not a mere hammer, it's a ball-peen hammer. In some circles, it's called a ladies hammer. I don't know why. But it's-- What's it for? Most people buy it mostly because it's small and light weight. But, in fact, it's for metal working. It's for taking a piece of sheet metal and pounding it out into an ashtray or something or for seating rivets. It's a metal worker's hammer. So you might not have known about that before but now, at least, you have a word to hang that knowledge on. It's a ball peen hammer. So we have various levels here, going from very specific to very general. And we can even go to a level of specificity for pianos by saying we've got a Bosendorfer. Why is a Bosendorfer special? I mean, is it like a Baldwin? Something's [INAUDIBLE], Yoka-- Lots of piano types, what's special? You see, you don't know because unless you play the piano, and probably unless you're a serious piano player you don't know that a Bosendorfer-- Ariel, you know. STUDENT: I think its supposed to have an extra octave at the bottom, black keys. Pretty cool. PATRICK WINSTON: It's got some extra keys at the bottom. And most people don't know that unless they're serious about the piano. Some professional piano players, when they're confronted with a Bosendorfer have to have someone cover those keys because it screws up their peripheral vision, and hit the wrong key. Because they're not used to having those extra keys at the bottom. So that's a little detail but the Bosendorfer. So you can make a kind of graph, and you can say, let's go from low, very general, to a basic level, to a specific level. So it is the case in human knowledge that that graph has a tendency to look sort of like this. So here's tool, here's hammer, here's ball peen. So that level, where you have a big jump, that's the general to basic level of transition. So that basic level is probably there because that's the level on which we hang a huge amount of our knowledge. We know a lot about pianos, and it all seems to be hanging on that word piano, which gives us power with the concept. So that's example number one of an element of our inner language. The ability to assemble things into hierarchies like that, and hang knowledge about those objects on the elements in that hierarchy. Well, given that you have elements in the hierarchy, how do you talk about them? Well, I like to consider the possibility, just for the sake of illustration, that you're thinking about a car crashing into a wall. So you've got things to think about like the speed of the car, the distance to the wall, and the condition of the car. And you've got the period before the crash, during the crash, and after the crash. So you might want to think about how to talk about those objects in those three time periods. So we can do that with a vocabulary of change and we do that because we believe that most of human thinking is thinking about change causing change. And that flies in the face of what we learned as engineers. Because in engineering, we learn about state, and once you know the state of the system, you know everything you need to know in order to predict the future. The trouble is, in our heads, thinking about everything there is in the world, including the current phase of the moon, is too much stuff. So mostly, our thinking, we think, is hinging on the idea that change leads to change. So that's why we have a vocabulary of change. So in the period before the crash, the speed of the car is not changing. There's a little notation for not change, no delta. The distance to the wall, that's decreasing. The condition of the car, that's not changing. Then, the car hits the wall. So the speed of the car disappears, the distance to the wall disappears, and the condition of the car will change dramatically. Finally, after the crash is over, the speed of the car does not appear. The distance to the wall does not change, and the condition of the car also does not change. So that's hinting at a vocabulary of change, and its use, which will be the second element in our development of a vocabulary of ways in which we might have constructed are inner language. So this a particular idea, that's classification. This is transition, and a system that purports to understand stories with a heavy emphasis on this notion of transition and we believe, that is to say, I think, that that vocabulary has to have decrease, increase, change, appear, and disappear. So there are 10 things you can have in such diagrams. I've done five. That's because, for each of those, there's a not variation on that. So with a vocabulary of 10 things can go a long way toward helping to describe things that are in process of change and making transition. And we have a lot of those words in our vocabulary. We use those words a lot in our vocabulary. They seem heavily connected with vision. Our friend appeared. The cat disappeared. The speed increased. So this is a description of a crash. In terms of those kinds of elements, now, I say to you, how does a camera work? Well, I could say the camera works because a photon crashes into a photo receptor. So when I say a photon crashes into a photo receptors, why am I saying that, and how does it help? I'm saying that and it helps because it's the same pattern of change you already know about when you talk about a car crashing into a wall. How does that work? The speed of the photon, the distance of the photon to the receptor, and the condition of the receptor. So analogies like that are very much of the core of what we think about all the time. Really then, there is representation number two. Number one is class, number two is transition, and now, you're ready for number three, which is trajectory. Linguists, who study sentences, often talk in terms of fundamental patterns that seemed to be in a lot of what we say, and a lot of what we say is about objects moving along trajectories. So we can talk about a trajectory frame. And a trajectory frame will have elements like this. It has an object moving on a trajectory that ends up at a destination. You might start out at a source. It's probably been arranged by some kind of agent, and the agent may assist himself making the motion happen with some kind of instrument. There might be somebody helping out over here, a co-agent. Well, what else can we have? A beneficiary, someone is helped out by the action. Sometimes, the motion is arranged by a conveyance. So these are a lot of slots, finite slots, and descriptions of actions, many of which involve motion on a trajectory. We have a tendency in language to decorate these things in one way or another, depending on the language. And so in many languages, the decoration is by way of position in the sentence. In English, it's often by way of a preposition. It's used to help zero in on a particular role of an object in the trajectory scenario. So if I say, I baked a cake with a friend, there's a with preposition. If I baked a cake for a friend, the friend is the beneficiary. If I baked a cake with an oven, that's an instrument. The object may be moving to a destination from a source, and if I'm going to New York by train, I put a by on top of that. If the agent isn't in subject position, I would say something like, oh, I see all the work was done by a student. So those prepositions have a tendency to help us zoom in on the actual role of particular objects in this whole package, the whole frame. So this is number three. There's a variation on this in which there's no actual trajectory in which case we'll just call that a role frame. Because if there's no trajectory, we can still have things such as an instrument, a co-agent, and a beneficiary. So now, we've got three representations. You might say, well, what good are they? And you can determine what good they are these days because it's easier to go over established corpuses, and say, what fraction of those, of the sentences, in such a corpus involve classification or a transition or a trajectory. The most well known of these is the so-called Wall Street Journal Corpus. It has 50,000 sentences in it, drawn from some period of time, all the syntactical language types work with that corpus a great deal. And we worked with it a little bit too, to see what fraction of the sentences or what the density of trajectories and transitions are in those sentences. So I have to say that a little more carefully, because the finding is that in 100 sentences you'll find about 25 transitions or trajectories. So they're very densely represented. They're often very abstract. Prices rose, the economy went to someplace, but there are still words that denote transition or trajectory. Of course, once you have all this stuff, that you have then have a desire to put it together. So the next thing we need to talk about is story sequences. So a story sequence can be a single sentence, and I want to illustrate that with one of my favorites. Here's the sentences. I think I've chosen a gender neutral name so as not getting in any trouble. So Pat, but I don't call myself Pat because I decided when I was 18 years old that pat is a unit of measure for butter. In any case, with Pat comforted, Mary, do you have an image of what happened? Probably not a very firm image. You know that Pat did something, but you don't know exactly what. Nevertheless, when Pat comforted Chris, you can construct something that looks like a role frame. Because the role frame for that would have an agent, and that would be Pat. There's an action. We're going to put a question mark in there because we don't have a very firm image of what the action is. Then again, we're building a wall frame, like so. The object is Chris. Oh, you know that is the object. Is there anything else we can say? Oh, yes, we can probably say something more. Something else comes to mind when you see Pat comforted Chris. There's a sort of result, and the result is a transition frame. And the transition frame involves an object, which is Chris, and Chris has a mood, which presumably, is improved. It goes up. Did you have something, Elliott? STUDENT: Could you, I guess, analogize the Pat comforted Chris to something like Pat gave comfort to Chris as Chris is the destination? And couldn't this be [INAUDIBLE]? PATRICK WINSTON: Elliott is wandering into a very interesting area having to do with, couldn't you think about this in another way and think of it as something moving along a trajectory. Comfort is moving, if not from Pat, at least to Chris. And that's a very important kind of observation because what it would suggest is there can be a utility in thinking of things in multiple ways, multiple representations. Marvin Minsky has a wonderful aphoristic phrase, which is, if you can only think about something in one way, you have no recourse when you get stuck. So multiple representations mean you have multiple ways of gathering regularity from the world, and collecting it and therefore that'll make you smarter. So yes, you could do that, and that would be a compliment to what I'm doing now. Let me continue what I'm doing now. So what have I done? I've got a role frame and a tradition frame, and the transition frame is the target of the result slot in the role frame. Now, we can modify this a little bit. And maybe want to say, instead of comforted, terrorized. And how would that change things? We don't know exactly what Pat did so the action remains unknown. The agent and the object are the same. But the result here is presumably that the mood went down. With just what we've got so far, we can answer a lot of questions, by the way. Once we've got the sentence understood in these terms, we can say who did the thing? What's this all about? And the answer is Pat. What did Pat do? Comfort, terrorized. Who do they do it too? Who was the object? It's Chris. What was the result? Chris felt better. Chris felt worse. So these representations already give us a question and answering capability that makes for an understanding of the sentence. But still, we haven't been very specific, so our next step takes this same sentence in a more specific direction. So here's a way that goes. Kissed Chris. Now, you begin to get a sense of what's going on, you form a mental image, so you could have a hallucination. And that hallucinations will also be a kind of frame, but, in this case, it'll be a trajectory frame. And the object would be, I don't know, Pat's lips. And the destination will be Chris' lips. I don't know, is that right? It depends. It depends. Have you all formed a picture of what's going on here? So it will be different depending on whether Chris is Pat's girlfriend or if Chris is Pat's daughter or if Chris is a frog and Pat is a prince -ess, I guess, the way the story usually goes. So somehow we have, in our heads, all kinds of libraries that help us to form mental pictures of things when we see things like kissed. So one final one just to show the variety. We could say that Pat stabbed Chris. What changes? Let's see, in the case of kissed, the mood is going up. In the case of stabbed, the mood is going down. You can also probably say that the health is going down. And the destination is Chris' body. And the object that's moving is Pat's knife. We get the same pattern with both of those sentences. So both of them involve a sequence that starts off with the action, moves to a transition and a trajectory and those are all arranged in a line. And that line is something that gives us a lot of power over the situation relative to a semantic net. So I'm going to decorate that one more time, and say that another element we get out of our internal language is sequence. An element we need in order to have anything that looks like an account of an inner language is sequence. Because if you think about things being arrayed in a vast spreading network, it's hard to deal with them. But if you thing about things being arrayed along a line, in a sequence of actions or events, like so, then that imposes enough constraint to get a handle on what's going on. So what we're going to call this is the representation, and I guess I'm up to one, two, three, four. This is a representation of story sequence. so even though that's a kind of micro-story, it's still an example of a story sequence because we get the power out of it by arranging everything in a line. You have a sense, I think, especially if you play a musical instrument, on how dependent we are on sequence. So if you play a musical instrument, you probably know how difficult it is to start replaying a piece of music from the middle of a measure. You have to go back to at least the beginning of the measure, and probably to the beginning of the phrase, if not the beginning of the piece. So our memory seems to be, at least in music, very rooted in the idea of sequences. And that's often true of storytelling too. We have to go back to the beginning of at least a scene, because somehow these things are arranged in sequences that form, somehow, usefulness out of their sequentialness. So there's one more thing we can talk about and that is the idea of not just the idea of how these sequences are constructed and what they're constructed out of, but we can also talk a little bit in terms of libraries of stories. And when we talk about libraries of stories, we can think about kind of the sort of standard stories that we have and how they're arranged, and how we can know a lot about something by what it's super class is. So it's a variation on the theme of learning stuff from the super classes. So here's an event frame. And then, in addition to event frames, there's disaster frames. And then there are party frames. And parties and disasters are both events. And when we talk about disasters, we know, in turn, they break up in a variety of things. We have earthquake disasters, and maybe we have hurricanes. And in the party world we have, I don't know, birthday parties, and we have weddings. And each of these types of frames invites us to fill in particular slots. So if we're reading a newspaper story about a wedding, we know that we're going to be learning the same sorts of things we will learn about any party except there is the additional information that we expect that says something about the bride and the groom. If we have a raw event and we don't know anything more about it than that, there's a time and place. If it's a disaster frame, it gets a grizzly over here, but if there's a disaster frame, we might have the fatalities, and how much it cost. If it's an earthquake frame, we need to know the magnitude of the quake, and the name of the fault. If it's a hurricane, we have the category and the name. So each of these things up there can be viewed, not just an example of something, but as a new frame all by itself. As we mature, we have these first four things as building blocks, and then, we educate ourselves, and we get all those kinds of frames that help us to understand the world. But how to fill those in from a newspaper story can be sometimes, quite a challenge. Actually the worst thing to understand is children's stories because, and this was determined experimentally when people tried to understand children's stories, because it turns out that children's stories are not simpler than the stories we write for adults. In many cases, they're harder. If you read about Shakespearean plots, it's all about intrigue, murder, jealousy, greed, but when you're trying to write a children's story it can be about anything. And worse yet, the children's story often raises problems that you don't see in newspaper stories. Let me illustrate that for you. You want to read that story? You have no trouble figuring it out, of course, but think about a poor machine. It's struggling to understand anything. What's going to be the problem? It's going to have trouble figuring out those pronoun antecedents. Look at them. They're complicated. [MUSIC PLAYING] PATRICK WINSTON: Shut [INAUDIBLE]. Oh, that's the way. [INAUDIBLE]. Here are the pronouns. One of them wanted to buy a kite. He has one, he said. He will make you take it back. So that's pretty hard. Actually, the principal here that I'm driving at is, when you have a new story or any story, if you have an old story but you want to read it-- when you want to understand it quickly, the instantiation of it quickly, you need to be sure that, if you're the storyteller, you don't add to the burden of the understanding any syntactic difficulty. So that's an example of telling a story with additional syntactic difficulty. No newspaper journalist would ever write the story like that. Here's how they would write it. They would even give you a clue that there's certain information you're never going to get, like who told the story. It's that reliable sources business. So this brings us to the final bit that I want to deal with today, and what I'm going to do is, if you came into this class today as a 97 pound writing weakling, you're about to emerge as a 250 pound mountain of a writer. Because I want to tell you a few tricks that will make you usually better as a writer, especially if you're Russian or German. And here's how it works. Rule one. Because they place additional syntactic burden on the understanding of the story by the reader. So if you're telling somebody about some difficult new technical idea, the last thing you want them to do is to burden their syntactic processor with figuring out pronoun antecedents. So don't use pronouns in you're writing, technical writing, at least, will be much clearer. By the way, why does this especially apply to Germans and Russians? Is it because-- is this an ethnic origin slur or is this a fact about their language? It's a fact about their language. Where is the fact? Why can they get away with pronoun usages that we cannot get away with in English? Yes, Andrew? STUDENT: Gender and also-- PATRICK WINSTON: Gender, because if you have all of your nouns decorated with gender, that reduces by three, the potential for ambiguity in the pronoun antecedent. So you'll frequently find that German and Russian writers will have pronouns all over the place that are perfectly clear to them because of gender. They are interpretable by English speakers when translated because we don't have the gender to help us zero in. So these things all have to do with minimizing extra, superfluous, gratuitous, unnecessary burden on the reader. Number two is don't use former or latter. You see those words used frequently in technical writing, and guess what? No human being ever encounters those words without having to stop and go back to figure out what they refer to it. So that's another example of not placing any unnecessary syntactic burden on the reader. And finally, don't call a shovel a spade. There's a habit, probably instilled by well-meaning but misadvised high school teachers, that you shouldn't repeat words, and so people go to great lengths to use some different word. The problem is that the reader doesn't know if the shift in the word is deliberate and attached to some subtle meaning shift, or if it's just adhering to some high school teacher's admonition against using the same word again. So you don't want to say, oh, the right way to dig this particular hole is with a spade and then switch to a shovel because the reader can't tell if it's deliberate, accidental, or a consequence of just the desire not to use the same word. So this is how you, with some very simple mechanisms grounded in AI, you can actually make yourself into a better writer by avoiding those kinds of things that put an unnecessary syntactic burden on the people who are reading your stuff. |
MIT_6034_Artificial_Intelligence_Fall_2010 | MegaR1_RuleBased_Systems.txt | The following content is provided under a Creative Commons license. Your support will help MIT OpenCourseWare continue to offer high quality educational resources for free. To make a donation or view additional materials from hundreds of MIT courses, visit MIT OpenCourseWare at ocw.mit.edu. MARK SEIFTER: Hi, everyone. A few of you may know me from my tutorial, but if you don't I'm Mark Seifter I'm the senior TA here at 6034 and these Friday omega recitations are going to be a little bit different than any of the other things in the class. All right, so, what is omega recitation. Omega recitation is in this lecture style room, but I hope to make it more like a recitation or a tutorial by, unfortunately for some of you, perhaps, calling on you guys to participate during the omega recitation, as we attempt to work with some of the material that Patrick has given us the big picture for in the lectures. And if you've gone to recitation so far you may received a little bit of enrichment into how the algorithms work. Well, here in omega recitation we're going to figure out how you can work the algorithms. And that's very important, because in this class being able to work the algorithms by hand in some of these examples I put up here, or like Patrick sometimes does in lectures to teach you the algorithms, is a crucial way of demonstrating that you understand how the algorithms work, and we use it on a quiz. The way we use it on the quiz is we-- well this is an old quiz problem. It's from last year. It tripped up a lot of people. So we're going to go over it today. I'm going to tell you all the tricks that tripped up a lot of people, in addition to emphasizing tricks that I've picked up over of the three years of TA-ing this course before, so that you will not fall prey to any of these tricks. Hopefully. I say hopefully, because some of these tricks I talked about omega recitation last year and people still got tricked by them. I guess maybe some people didn't come to omega recitation or they were just very tricky tricks. So be sure to come to omega recitation. Pay attention to the tricks. Pay attention to how these things are solved. Ask questions if you're not sure. And when we get done with the hour together I'm hoping that everyone's going to do really well on the rules part of the first quiz. So without further ado, if there are no questions on that, which I don't think there probably will be, let's move on to the problem at hand. As some of you may have noticed, it is a Harry Potter based problem. One of our TAs last year was a big Harry Potter fan. So what we have here a series of rules and a series of assertions. Now if you happen to be real go getter and have already looked at all the past quizzes online, you may have already seen this problem. And if you really, really have attention to detail you may notice something a little bit different in the way I wrote it as compared to the way it was written online. In fact, you may, if you have a lot of times a detail, but you're not really in on Scheme or LISP, you may think the way I wrote it is much easier to understand. And in case you can't understand the other way, and in case one of the TAs decides to write it in a LISP-like way again, let me explain. If you look at one of my rules, say if x is ambitious, and x is a squib, then x has a bad term. That's rule 0 up at the top. That's written in sort of what we call in fixed notation. It's easy, it's simple. All operators are in the order you would expect. 2 plus 3 equals 5. There's another kind of notation that is sometimes used on previous quizzes. And by sometimes, good number of times, actually. I'm not sure why we still use it that way, because of the fact that the class has move on to Python, but we do, sometimes. And that's called prefix notation. In prefix notation, the operator comes first. So for instance, plus 2 3 equals 5. The plus comes first. The function, then, comes first. In this case, the function is plus. So if we were writing this in prefix notation, in fact, the way I received it, when it said, "if and x is ambitious, x is a squib, then x has a bad term. The and went first and there's a parentheses which included everything that was scoped under the and. Does anyone have any questions about prefix notation in case it comes up? Just remember, if things start being written in a really weird way with lots and lots and lots of parentheses, whatever's outside the parentheses, whether it be "plus" or "and" or "or" is an operator that acts on the things that are inside. However, we're not going to deal with that now, because I think the most important thing is to understand is the rules, and the less important thing is to understand prefix notation. And you can do a problem at home that's written in prefix notation to test yourself on that. So what rules do we have here. We've got equals 0 through 5. For some reason they are labeled with P's. They are often labeled with either P's or R's. The secret on why they are labelled with P's is that one year somebody labeled them with P's for like "proposition" or something like that. And then other TAs looked at that test and sometimes also labeled it with P's and it sort of down the line continued to be labeled with P's, and sometimes with R's. So these P's are these six rules. The first rule is if x is ambitious, and x is a squib, then x has a bad term. So what's the deal with these question mark x's? The question mark before an x, or perhaps a y down here, indicate that there's a variable waiting to be bound. We're not assuming there's a Harry Potter character named x, and only if that question mark x, the mystery character, only if mystery x character is ambitious can that person possibly have a bad term. What we're seeing is any character in the Harry Potter universe, or not the Harry Potter universe, maybe a rhinoceros, can fit into that x. So for instance, if a rhinoceros is ambitious, and a rhinoceros is a squib, then a rhinoceros has bad term. That rule saying, for any x this is true. And it's very important how we treat the question mark x, and how we bind question mark x, when we do both back chaining and forward chaining. I'll get back to that, because some people made some very, very small mistakes that really messed up a lot of their forward and backward chaining last year. It was actually very tricky. Jeremy, the TA who wrote this was very clever, and it makes it really great case study for you guys. All right, so Rule 1, if x lives in Gryffindor tower then x is a protagonist. By the way, for conciseness I'm going to be using GT for Gryffindor tower when I write these in later on, and I'll use SD for Slytherin dungeon. Speaking of which, Rule 2, if x lives in Slytherin dungeon, then x is a villain, x is ambitious. Why there are two things here? Well, after the "if" we have what we call antecedent, that's something that needs to be true in order for this rule to match. After the "then" we have what is called the consequent, and in this case there are two consequents. Anyone who lives in Slytherin dungeon is automatically a villain and also ambitious. So you can think of there being sort of an "and therefore" for the purposes of both of those assertions would be added to the knowledge base. Rule 3, if x is the protagonist or x is a villain, and x is ambitious, then x studies a lot. By the way, the scope for this, just to be sure that we're clear, is this. So we need them to be a protagonist or villain and no matter what they have to be ambitious. Rule 4, if x studies a lot, and x is a protagonist, x becomes Hermione's friend. And Rule 5, if x snogs y, and x lives in Gryffindor tower and y lives in Slytherin dungeon, then x has a bad term. So those are our six rules that we can use to understand Jeremy's world of the Harry Potter universe. And we also start off with four assertions. Let me not underestimate the value of always looking to the assertions. It's one of my white star ideas that are up here on the board. Let me see. This. Perfect. Always check the assertions before using a rule. This really tripped people up last year, and you'll see why, because we're doing last year's problem. Our four assertions that we start. With assertion 0, Millicent lives in Slytherin dungeon. Assertion 1, Millicent is ambitious. Assertion 1 is what tripped people up, so remember that Millicent is ambitious. Assertion 2 Seamus lives in Gryffindor tower. And assertion 3, Seamus snogs Millicent. So those those are our four assertions that we've already started with. Now the two things we're going to have to do are backward chaining and forward chaining. Now when you guys learned these two backward chaining and forward chaining, raise your hand if thought forward chaining was harder than backward chaining. I knew it! I can prove whatever point I want because noone wants the raise their hand. I also think backward chaining is higher than forward chaining. Raise your hand if you think backward chaining is higher than forward chaining? First of all, we have a good number of people. Second of all, since no one wants to raise their hand, I could just ask it the other way. That's a pro tip if you're ever with a large group of people. You can prove any point you want by asking the other direction. No one will raise their hand. So I agree that backward chaining is higher than forward chaining. And I disagree with Patrick that we're going to get out early, so let's start with backward chaining first. So that we make sure that we spend the bulk of our time with it. The forward chaining, well, you just go through pretty methodically and add new rules. Backward chaining, you have to draw this crazy tree. There's a lot of places to get lost in the middle of the road. So let's do some backward chaining. And to do that, I'll do it over here on the left side. So when we're doing backward chaining, we have to remember a few things that are written directly on the quiz. So you're not going to have to worry about that. But they're still pretty important. So-- actually, I'll write it on here first. I'm going to read them off. So when working on a hypothesis, the backward chainer tries to find a matching assertion in the list of assertions first. If no matching assertion is found, the backward chainer will try to find a rule with a matching consequence. A rule that has something in the then that can prove the assertion is trying to figure out. So for instance, if I was doing backward chaining on Seamus [? Snaus ?] Millicent, what happens? I'm immediately done because there's an assertion that's assertion three that says Seamus [? Snaus ?] Millicent. I proved it. I'm done. I'm happy. We can leave, we can go home. Unfortunately, that's not what the quiz asks us to do. Now, let's say that instead, we were supposed to say, I don't know, Seamus is a protagonist. Well then, we'd wind up looking through here and we would look at the fact that rule one can prove that someone is a protagonist. And we'd curse and try to prove whatever is in the antecedent of rule one, which is, OK, well, does he live in Gryffinder Tower? Assertion two, he does. So that's some really quick. And if that was too, fast don't worry. We're going to go step by step with the actual problem. But I just wanted to give two really easy problems, really quickly, how it's going to do it. Let's go step by step with the real problem. Let's keep in mind, backward chainer never adds new assertions to the list of assertions. And if you have a tiebreak, you always order based on rule order first, P0 through P5. And if the same rule matches with more than one thing in your list assertions, then you tiebreak based on the order of the assertions. Very important. Disambiguation and tie-breaking are a big place to get messed up. So we're going to try to prove that Millicent becomes Hermione's friend. And so I'm going to abbreviate some of the things, but not for the very first line. So Millicent becomes Hermione's friend. We start drawing a goal tree. Now you guys are going to learn exactly what these mean very soon, in fact, next lecture. But for now, trust me when I say these goal trees are depth first. And I'll explain what that means because some people have messed themselves up and spent more time than they needed to by treating the goal tree in a different way. Now, Millicent becomes Hermione's friend. Let's pretend that we are the backward chainer. We're trying to prove that this is true or disprove and say, well, it's definitely not true with what we have. So let's see. I'm now going to make you guys help. And people in the back think they won't be called on. So I often like to call on people in the back. What do you think is the first thing we should do? Very first thing? Yes, look for a matching assertion. Excellent. Do we have a matching assertion, everyone? No, we don't. That would be the world's easiest quiz problem. We do not have a matching assertion. Great. So, since we don't have a matching assertion, what now? AUDIENCE: We start to look at the rules. MARK SEIFTER: That's right. And do you see any rule that could prove that Millicent becomes Hermione's friend? AUDIENCE: P4. MARK SEIFTER: That's right. You can see in P4, x, which can be anybody. Anyone is capable of being Hermione's friend. Great. So P4 is our rule of the hour. So we're going to use P4. And when we use P4 to prove that Millicent becomes Hermione's friend, we're going to have to add something or another to the goal tree. So let's see. What do we have to add to the goal tree? AUDIENCE: We have to see if Millicent studies a lot [INAUDIBLE]. MARK SEIFTER: That's right. Does everyone see that? In order for her to become Hermione's friend, by rule four, we have to see if she studies a lot and is a protagonist. Question? AUDIENCE: [INAUDIBLE] MARK SEIFTER: OK, that's a very good question. You're going to get screwed up if you look at the antecedent in backward chaining first. It's backwards partially for a reason. You need to look at the consequent. Now why is that? Well, let's say there is a rule six that said if x becomes Hermione's friend, then Hermione feeds x Polyjuice Potion, or something like that. Will that rule help us in back chaining to figure out if Millicent becomes Hermione's friend? AUDIENCE: [INAUDIBLE] MARK SEIFTER: Some people are shaking their head. But think about it, will that rule be able to prove it? No. Now, if they do become Hermione's friend and we want to do some forward chaining, we'll figure out that they're going to transmute into some kind of other thing because of the Polyjuice Potion. But it's not going to help us do the one thing we want to back chain, which is to prove that thing on the top. So we actually need to look for something that has our current goal in its consequent. Then we add on the antecedents. As people I've asked so far-- sorry, I don't know your names-- like Patrick, have correctly given me every time, which is excellent. So notice also that she didn't say x studies a lot needs to be added to the goal tree. She said Millicent studies a lot. Oh, another question. Great. AUDIENCE: So obviously after putting rule 4 there, shouldn't we first check the assertions-- check on the assertions instead of our conditions [INAUDIBLE] MARK SEIFTER: So the question is, once we put rule four in there, should we check to see if those antecedents of rule four are already in the assertions before checking other rules? The answer? You're not only correct, sir, you're exactly one step ahead. So I'm going to assume I called on you for the very next thing. That's exactly what we do once we draw it onto the tree. You're exactly on the way. Further question? Perfect. All right, great. So again, we're putting up Millicent studies a lot. Millicent is a protagonist. We're not putting up x. Some people did that. And of course it's an n node as we heard. We need them both to be true or we're not going to be able to continue onward. So we have-- I'm going to use m for Millicent. Millicent studies a lot. Also over here, Millicent is a protagonist. And we've already heard that we need to search in the assertions before we go into any rules for what we're looking for next. The question is, what are we looking for next and in what order? This is where that thing I told you about depth search is crucial. We are going to look here on the left node. And if it has any children, if we have to keep going down, we are not going to look here yet. There's a few reasons for that. One of which is that we're lazy. The evaluation, if you learned about it, is also-- if you learn about lazy evaluation, is also lazy. And if we can disprove this study's a lot branch, we don't have to do any work with the protagonist branch. We're done. We're out of here. And it's a loss. So we're going to move down the study's a lot branch. And we already heard from the audience, we need to look to see if it's in the assertions. Everyone, is Millicent studies a lot in the assertions? No, it's not. Well then, now we're going to get to that question that I had before about the antecedent or the consequent. Because study lies here in the antecedent of rule four, but we're not going to be able to use rule four here. In fact, that's how we got here in the first place. So the question is-- let's see, the question is, do we have a rule in the consequent that matches Millicent studies a lot. Yeah, that's right, P3. That's right. So P3 would give us Millicent studies a lot. And so therefore what will we add to the goal tree? AUDIENCE: We add both that is Millicent either a protagonist or a villain and [INAUDIBLE]. MARK SEIFTER: Yeah, she. Turns out to be a girl. So that's exactly right. We heard that we need to add-- we need to add an AND node, which also has an OR node at the bottom of it, because this is a little bit of a complicated rule. So we have an AND node. So we have an AND node. And the first thing on the AND node is Millicent is a protagonist or Millicent is a villain. And the second thing on the AND node is Millicent is ambitious. Actually, that is the second thing on the AND node. I hope that's what I said. All right, the second thing on the AND node is Millicent is ambitious. So here's our handy little tree. Now this is where it's important, as I said, we're doing a depth first search. We're going down along the left branch. So where do you think we're going to go next on this tree? AUDIENCE: [INAUDIBLE] MARK SEIFTER: Not quite. That's a good guess. You didn't say we're trying to prove m is a protagonist on the right branch, which is a very common mistake. What do you do? AUDIENCE: You should go down to see if there's a consequence for m is-- Millicent is a protagonist. MARK SEIFTER: Which one of the two Millicent is a protagonist do we want to look at? Yeah, the farthest down left. That's right. So you guys will learn depth for search on Monday, at which point it will become much clearer what I'm saying. But yeah, you always follow the left branch as far down as you can. Then you take the right branch of that same branch that you followed. Great, so we're going to try to find that Millicent is a protagonist. What do we do first? Everyone? Check assertion, yes. Is it in there? No, it's not. OK. So, therefore we're going to try to find a rule as we heard about whether Millicent is a protagonist. So, all the way in the back? All the way in the back, is there a rule that matches? AUDIENCE: [INAUDIBLE] MARK SEIFTER: That matches in its consequent that Millicent is a protagonist. AUDIENCE: Um, there is a rule. MARK SEIFTER: OK. Well, let's give it a try. Which one? AUDIENCE: One. MARK SEIFTER: Rule one, that's right. So therefore, we would have to add what to our goal tree? AUDIENCE: Another [INAUDIBLE]. MARK SEIFTER: Right. And what's on that please? AUDIENCE: [INAUDIBLE] MARK SEIFTER: Yeah, Millicent lives in Gryffindor Tower, perfect. So, our new thing that we're searching for is-- pretend this connects. Actually, I guess it does connect. We're looking for m lives in Gryffindor Tower. Great. And since it's depth first, that's where we go next. Great, we're on a roll. So what do we do first? Do we have that in assertions? Some people say yes, but the answer is no. We have, in fact, Seamus living in Gryffindor Tower. Most of you said "no." You're right. The majority rules. We don't have any assertions. However, do we have a rule with that in the consequent? No. So what do we do? People are saying different things that are all correct. Backtrack, you say. Go to the next. We can't prove it. All correct. Put a big x. This isn't true. Now we'd look up. We're on an OR node. So we're not done yet because either of those can be true. So now we go back up, backtrack, Millicent is a villain. What do we do first? Check assertions. Is it in there? No, it's not. Don't worry, we're getting to one where it will be, where about 40%-- or 30% to 40% of the class lost points, very, very soon. So, we see that Millicent is a villain. And it's not in the assertions. So therefore, is there any rule that has that in its consequent? AUDIENCE: [INAUDIBLE] MARK SEIFTER: We're looking for Millicent is a villain. AUDIENCE: [INAUDIBLE] MARK SEIFTER: Oh, you can't see it. All right, I will move it down slightly. It got a little bit off kilter when I put in the parentheses around the "and." Hold on. There you go. AUDIENCE: If [INAUDIBLE] is a protagonist or is a villain. [INAUDIBLE] MARK SEIFTER: We're trying to prove that she's a villain, though. So do we have any with that in the consequent? [INAUDIBLE] anything that will prove that she's a villain if we fire off that rule. Want to give here a little bit of help? AUDIENCE: P2. If she lives in Slytherin, then she's a villain [INAUDIBLE]. MARK SEIFTER: Ah, see that now? It had two, so it was a little bit harder. Now some people ask, what about that ambitious thing? Can we like add that to the tree or something? No. The backward chainer is single-minded, focused, and stupid. Even though it's later going to need to prove that she's ambitious, it doesn't care. It just sees the villain there. Villain, oh, that's great. As long as we've got the antecedent we are going to be happy. This is actually an important point, not in this particular problem, but in some other problems. So by using P2, we're obviously going to have the antecedent x lives in Slytherin dungeon. And x here is Millicent. So Millicent lives in Slytherin dungeon. All right, so what's the first thing we do when we see this new assertion? We check the assertions. Is it in there? Yes. That's not where everyone lost points, but yes, it is in there. Check. This OR node is happy. Now we move up to the AND node with Millicent is ambitious. So Millicent is ambitious. What do we do now? AUDIENCE: So isn't the consequent of P2 that x is a villain and x is ambitious. I don't know if it's [INAUDIBLE] or whatever, but in our assertions we have that Millicent is ambitious. Is that necessary to let us-- MARK SEIFTER: To let us use rule two. Actually, no. The question is, do we need the x is ambitious to be in our list of assertions in order to use P2. For instance, if we didn't have the assertion that Millicent is ambitious, would P2 have a problem firing? So, it's a very interesting question. It's sort of what actually I was trying to mention earlier when I said it's single-minded, it's focused, it doesn't care about the other one. Because you might say, well, wait a minute, P2 could never have actually fired if it didn't have both of its consequents. But in fact, the backward chainer doesn't care. All that the backward chainer is looking to do is to prove whether or not there's a possibility that Millicent might be a villain. It's not trying to say, oh, all the results of some certain rules are making Millicent a villain and all the other things that are there, like ambitious, are in the database. It's not the forward chainer. It actually doesn't care. This sometimes leads to unnecessary computation. But it's very simple, and it's very simple to code. So it often will probably speed you up in the long run. Like if there's, for instance, 100 thing in the then and you only care about one of them, it would be a waste, almost, to add those 99. There's a follow up over here somewhere? [INAUDIBLE] AUDIENCE: Order m of 99. And so [INAUDIBLE] MARK SEIFTER: You can try. However, then at that point, every time you check the assertions, you'd be attempting to check those 99. Also, you might not have used those rules. The question is, could you make a hash table with all of the 100 things that were in the consequent? And then, after having made that hash table since you were already walking through those anyway when you were checking it, you could add those to the assertions. The problem is that you might not necessarily want-- with the backward chainer-- you might not necessarily want to think that all those things in the consequent are necessarily true. However, you might not care about them. And then you have to make an order and computation every time you check that list of assertions in the upper left. AUDIENCE: [INAUDIBLE] MARK SEIFTER: You don't know what assertions there are. You have to check the entire list of assertions in the upper left every time you get to a new node in the goal tree. So each one is order 1. But I'm not saying to find a specific assertion is order n. I'm saying every time you get a new node in the goal tree, you have to check all the assertions you've added. AUDIENCE: You have to do that anyway. [INAUDIBLE] assertions. MARK SEIFTER: Yeah, you'd have to check for the four. But let's say that there were 10,000 consequents of P2. You would have to check 10,000 instead of four if you added them all after proving P2 was true on the goal tree. Also, another important note is that you might not use every branch of the goal tree, if there's like an OR that's up higher and you wind up using a different branch, in which case you'd probably have to make a separate hash table for every sub-branch that you have and then remember which hash tables were updated based on which sub-branch. And then sub-branches die. It's probably more effort than it's worth. It's certainly an interesting question. It's a great question for a debate in recitation. Also, I'm happy to talk about it later. I think that someone who very intelligently made hash tables and thought the problem through, as Patrick said, more knowledge can mean a better-- well, actually maybe he hasn't said more knowledge can mean a better search because we haven't done that lecture yet, but-- AUDIENCE: You're talking about implementation decision and using hash table might well be a good way to do this. The problem is that we're presuming some things about how the rules are firing and what [INAUDIBLE] they're firing that depend on the order of assertions in the table. So if we used the hash table, we'd lose the order. [INAUDIBLE] MARK SEIFTER: That's true. That's true. But I'm thinking like if someone wanted to make that implementation as like do some research in rules based systems, it's possible you can increase the running speed. In this class, we are not completely as focused on the fastest algorithms. But I still think that a cool thing to try. Questions from people in the crowd? We'll start with you. AUDIENCE: So ignoring the hash table? MARK SEIFTER: Ignore the hash table. That's a good idea because it's a little extra enrichment thing. It's not anything you would possibly need to know for the quiz. AUDIENCE: [INAUDIBLE] So if it became an [INAUDIBLE] are you using this "then" statement because you want the villain. So along with using this you get x is ambitious. [INAUDIBLE] used after that, another "then" statement that we also use that says x is not ambitious. So how would you resolve that? MARK SEIFTER: So the question is, let's say you used rule two somewhere in the tree to get that x is a villain, which then says that x is ambitious. Then later, you have x is not ambitious somewhere in one of the other rules that you then later need. The question is, how do you resolve that? What does it do? Well, first of all-- AUDIENCE: [INAUDIBLE] MARK SEIFTER: First of all, interestingly, if it says x is not ambitious-- literally like that. And I'm not being pedantic here. This is one of the tricks someone played in a previous quiz. You can have that and x is ambitious both in the list because it's a dumb rules-based system it doesn't know that you can't have both of those on the list. So if you have a positive assertion in a consequent, x is not ambitious in a consequent, x is not ambitious, it'll be happy to add Millicent is not ambitious while Millicent ambitious this is already there, because it's stupid. Now if it had delete Millicent is ambitious in the consequent and deleted something that was previously there, that would be an interesting problem. It could cause mistakes. Your back chaining would probably make the mistake of allowing this. However, this is exactly what we do not, I believe as a policy, do not have DELETE statements on the quizzes. At least previous quizzes did not have them, except for in a hypothetical of what happens if we add a DELETE statement here. I don't think we've ever made people work things through with the delete. Delete would possibly cause some issues. Question in the back? AUDIENCE: So, just to check this, we're doing backwards chaining, we don't actually add assertions to your [INAUDIBLE]. MARK SEIFTER: Never add assertions in backward chain to the assertion table. It's dumb. The system is dumb. I'm not saying it's a dumb idea to do this. It's actually pretty fast to do it this way. But the system is dumb. The question is, do you add assertions or you don't add assertions? You do not. You simply go through, checking all the things on your goal tree until the goal tree is either proven or disproven, then you're out of here. So, cool so far, everyone? Good questions from everyone. These are all questions that are things that often trip people up. So, our next thing, Millicent is ambitious. We sort of-- the cat is out of the bag because there is a question mentioning that there is an assertion that says that. But I direct you guys to our favorite rule, rule two. Shouldn't we use rule two to prove that Millicent is ambitious? The answer is if you want to lose four points or however many points, then yes. But if you don't want to lose points, then this is already correct by virtue of assertion one. So I guess I should write, also, that this over here, Millicent studies a lot rule, this rule is rule four. Yeah, that's rule four [INAUDIBLE] and protagonist. Down here to prove that she studies a lot is rule three. And then at the bottom here we have what rules here? We have the protagonist rule, which is one, and the villain rule which is two. This is correct virtue of A0. OK, great. So we've proved this whole AND node. We know that Millicent studies a lot. But now we're going to have to go a little bit more quickly based on time, and so I will do the remaining things. Anyone who has not been called on is off the hook. But that doesn't mean not to pay attention because there's some interesting stuff still to come. So, Millicent is a protagonist. Well, good news, everyone. There's one other reason that I can do this and that's that you've already told me what I need to do. Most is a protagonist, you use rule one, rule one. And check if she lives in Gryffindor tower. Now, will it do it again? Yes it will. It hasn't cached that it's already tried that. Now, hash table not withstanding, caching something you've already tried may be a good idea with backward chaining, but we do not do that in this class. That's an implementation detail. It's another idea of something you can do. We don't catch things that we've already tried. We try them again, dammit. And when we try it again, maybe it'll work this time. No, it never works. It never works this time. This fails. This AND node fails. The whole thing fails. They're not friends at all. They're bitter enemies. Millicent does not become Hermione's friend. So now they ask us a few questions which are pretty easy to answer based on the ordeal we've just been through. So, determine the minimum number of additional assertions that we would need to add for Millicent to become Hermione's friend. But you're not allowed to add an assertion that matches a consequent of a rule. You can only add an assertion-- in other words, you can't add an assertion that will prove some other rules. You have to add an assertion that directly-- yeah, that's basically the rule here. Because there's a lot of choices. But he wanted one particular answer. And so can anyone think of an assertion that doesn't match any consequent of any rule that will make her be Hermione's friend. A lot of people say Millicent lives in Gryffindor. That's correct. So, part three, your solution to part two causes an uncommon situation. What is the uncommon situation? And what should you do to the list of assertions to solve this problem? Does anyone know what the uncommon situation would be if we added that she lives in Gryffindor? Hand is raised up here. What is your answer? [INAUDIBLE] That's true. She lives in Gryffindor and Slytherin. So how would you fix that? [INAUDIBLE] Yes, you could take out Slytherin from the assertions and say that she only lived in Gryffindor. A lot of people gave the answer to this question, which is a perfectly reasonable thing to do in a rules-basis system. They said, well, we can add a rule that says that if you live in x and y is different than x, then you can't live in y. The problem with that, among other things, is that we asked for a way to change the assertions and not the rules. And you gave us a way to change the assertions. That's the right answer. There was another backward chaining problem, this is 2009 quiz one, and I will leave it off for the moment. You guys should take a look at it, in particular with variable binding. Because remember, you always have to bind the variable that's relevant to you. But that doesn't mean that you always have to bind all of the variables. For instance, let's say that we wanted to prove that Millicent has a bad term, which as it turns out, is what we wanted to prove in the other backward chaining. Very quickly and without doing the problem, can anyone who thinks they're very clever at backward chaining tell me exactly what things would be added to the goal tree to prove that Millicent has a bad term? Raise your hand. I won't pick out a victim because I'm asking this to happen pretty quickly. No one? All right. AUDIENCE: You would add the three antecedents to rule-- MARK SEIFTER: To rule five. So what would you add specifically? AUDIENCE: For anyone, variable-wise, you would just have to show that there is at least one person who matches. MARK SEIFTER: So can you just sort of read out, what would the first thing say with the snogs? AUDIENCE: X snogs-- or Millicent snogs anyone. MARK SEIFTER: Yeah, Millicent snogs y. That's crucial. Some people did a lot of different things. Some people put Seamus in already. You can't do that. The system is stupid, it doesn't know that the person is Seamus. Also, some people said it worked, because look, they're snogging each other, but it's the wrong snoggle direction up there. Because we have Seamus snogging Millicent, not Millicent snogging Seamus. But yes, exactly. It would be Millicent snogs y. Millicent lives in Gryffindor tower. y lives in Slytherin dungeon. So, let's give forward chaining all the attention it deserves, about eight minutes, which will be more than enough. OK, we still got all these rules. We've still got all these assertions. Instead of doing that horrible backward chaining-- well, it's not that horrible-- but we'll do forward chaining. It'll be really easy. We'll just add new assertions as they go. Remember the tiebreak order. Rules in order from 0 to 5. And if the same rule could fire off with multiple different assertions, we'll use the assertions in order from 0 to 3. So let's see. We don't have much time, but with our four assertions, let's see, what rules possibly could match with our four assertions? Well, I'll figure out for you. Do we have an assertion about an ambitious person? Yes. But they're not a squib. So not rule zero. Do we have someone that lives in Gryffindor? We absolutely do. Rule one is matches. Do we have someone that lives in Slytherin? We do. Rule two matches. Do we have a protagonist or a villain? We don't have any protagonist or villain. Rule three is not going to match because it's in an "and." do we have someone who studies a lot? We do not have someone who studies a lot. Rule four is not going to match. Do we have some snogging? We do have some snogging. So rule one, two, and five match. So m for matching, we have one, two and five. Everyone, which one wins the tiebreak between one, two, and five? One, because it comes first numerically. Which one is rule one if x lives in Gryffindor tower? The only binding for x is Seamus. Seamus lives in Gryffindor tower. We're firing off rule one. So new assertions, the first assertion that we'll add is Seamus is it a protagonist. Great. Now, there's one other thing, and it's the main thing I wanted to tell you about in forward chaining. About our old friend, rule one, now that we've done this, does rule one still match an assertion in the database? Yes. It does. But what's going to stop us from constantly doing rule one every time because it comes numerically before the other rules? What stops us there is a part of our implementation. And it's a very important part that people sometimes forget. It's what I like to call the no-impotent rules implementation detail. That is, if a rule is completely, 100% impotent, it would do absolutely nothing. Then you do not fire it. You go to the next one. That's pretty important. Let's say that you possibly had a delete, which you won't have on the quiz. That means that if the thing to delete is missing, it's impotent. If the thing to delete is there, it's not impotent because you can delete. Let's say you have 500,000 things that you're going to add to your assertions. 499,999 are already in your database. One of them is missing. Is it an impotent rule? No, it's not. You have to fire it anyway. You can't be like, oh, it's mostly impotent rule. No. You have fire off the rule if it will do anything. But right now rule one won't do anything because we've already added Seamus is a protagonist. So what rules match down? Well, Seamus becoming a protagonist is nice and all, but he's not ambitious, so we still don't match rule three. We still match rules one, two, and five. As I basically just got finished telling you, the one that we're going to fire off now is two. That's right. So, rule two is if x lives in Slytherin dungeon our only [? binding ?] for that is Millicent. And that will show that she's a villain and she's ambitious. So what assertions will we add? Millicent is a villain. And people stopped. That's right. Good job, everyone. You remembered that Millicent is already ambitious from assertion one, so we don't have to add them both. A lot of people did. They were wrong. So Millicent is a villain. The other part of that rule doesn't do anything. But it's not an impotent rule because one of the things is not on there. Great. So, what rules match now? Well, as it turns out, now that Millicent is a villain and ambitious, we match rule three as well. So we match one, two, three, and five. What rule are we going to fire, everyone? Three. Because one and two are impotent rules. We'll fire rule three. Well, Millicent is a villain. She's also ambitious. She's our only match for three. Therefore, what assertion do we add? Millicent studies a lot. That's right. So, Millicent studies a lot. However, she's not a protagonist. We still don't match rule four. Oh well, we figured that out already ourselves from backward chaining that we were never going to match that. So what matches? It's still one, two, three, and five. Seeing as one, two and three are impotent, the only match is five. So the only match we care about, the only one we're going to fire is five. And the result? Let's see. x snogs y. That's Seamus snogs Millicent. x lives in Gryffindor, that's Seamus. y lives in Slytherin, that's Millicent. So what do we add? Seamus has a bad term. Simple, painless, it works out. We're done. And we got 100 on this quiz. Hopefully you guys will, too. Question? The question is, if you have a new assertion which matches a rule but it's really high number, it's all the way down at the bottom of the rules, but it's new and it has new stuff, should we do that first, maybe, because it's new? Oh, you mean that it has a lower number, a lower number? AUDIENCE: [INAUDIBLE] MARK SEIFTER: OK, so the question is, if a rule-- let's say that rule three made Millicent be a squib. Then yes, you'd immediately not only match zero-- zero would immediately fire because numerically it comes first. You'll see that on some of the old quizzes. And you may see that in tutorial. If you guys want to do some more practice problems in tutorials, your TAs will have that available as one opportunity in tutorial. But yeah, that's exactly right. It happened that it went in order this time, but it will definitely hop anywhere it can in order to get to the lowest numbered assertion. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 8_Constraints_Search_Domain_Reduction.txt | PATRICK WINSTON: It's too bad, in a way, that we can't paint everything black, because this map coloring problem sure would be a lot easier. So I don't know what we're going to do about that. How long is this going to take? Here's what we're going to do. We're going to wait till either all the laptops are closed, or this program terminates, whichever comes first. So how long is this going to take? I actually don't know. I think it'll take more than the life time of the universe, but I'm not sure. So let's take a look at a slightly easier map coloring problem. We'll stop this one and change the map to something I call [? simplicia. ?] There's 26 states, one for each letter in the alphabet. And what we're going to do is we're going to do a depth first search for a suitable coloring of this map. We're going to go in order, A, B, C, D, E. And as I've suggested here, at each level we're going to rotate the color choices so we don't over use any one color. So if we launch this particular search, this depth first attempt to color this map. There it goes. And maybe we should wait until it terminates. Or maybe we should just let it run and it'll terminate, perhaps, sometime within the lecture. Or maybe we should just let it run over the weekend. Or how long do you think we would have to wait, if we want to come back and watch it terminate? At roughly 30 frames a second-- I'm calculating it all in my head, I don't want to bet my life on it-- but I think about 5,000 years. And if we want to use as many states as there are in the United States, in this demonstration, you get up to numbers like 10 to the 17th years-- 17th, 18th, 16th, I'm not exactly sure. I did a rough calculation. And, of course, you could do some parallels into that, and it's not as bad as chess, where you need all the atoms in the universe and you still can't do it, and things like that. Acting as computers. But still, it's pretty horrendous. The problem is, well the problem is illustrated by this diagram I put in back of me. If you do a depth first search and you have a problem like Texas. If you label... if you assign a color to Arizona, Oklahoma, Arkansas, and Louisiana first, and then wait around to Texas last, then you get yourself into a bind by your fourth choice, that you don't discover until you're 48th choice. So what happens then, is that you start developing this tree like this, you get to those states surrounding Texas. Texas is last state assigned a color and there's nothing left for it. And that problem was there on the fourth choice, and you don't discover it until your 50th choice. So you develop a horrendous, impossible, impossible search. So you simply can't do it that way. But now, not to worry. I've come equipped with the idea of constraint propagation. So we could just take a country with four states, like this. Each can be labeled red, green, blue and yellow. So just like in a case of line drawings, we'll pile up all the possible things that the value can be-- red, green, blue and yellow. Red, green, blue and yellow. And we start up constraint propagation. So we say for the upper left hand corner state, is there any reason to believe that R is impossible? Well we look at our neighbors and see what kind of constraint flows in from them. And sure, this guy could be green, and this guy could be blue. They don't have to be red, so there's nothing that rules out red. And there's nothing the rules out blue. And there's nothing that rules out yellow. And there's nothing that rules out green. So constraint propagation just sits there with its finger up its nose, doing nothing. So it doesn't look like we can use either depth first search, or constraint propagation. So we could just give up and cry. But maybe there's some other approach that will help. So let me actually work the Texas problem. So with apologies to Houston and Tyler, here's a map of Texas. There we are. And here's, roughly speaking, Arizona is over here somewhere. And we've got Oklahoma in there. And old Bill Clinton's state, Arkansas. And then Louisiana sticks out there a little bit. So there's our map of that particular part of the country. So we've got Arizona here, Oklahoma here, Arkansas here, and Louisiana here, and Texas here. And we have elected to assign colors to these states, in that order. So this is one, two, three, four. And we're going to do that. We're going to rotate our color choices, just so we don't over use any one color. But we're going to also have a look at Texas as we go around. Because Texas is a state that borders on the States they were choosing colors for. So the only possible colors that Texas could be are red, green, blue, and yellow. So as we make our choices around here, we'll say that-- we don't have to adhere to any particular style-- we can say that Arizona is going to get the colored red, R. That's going to rule out R over here for Texas, because no adjacent states can have the same color. Then we go over to Oklahoma, and we're rotating our color choices, so we'll say that can be green. And that's fine, because it's consistent with the red here, but it rules out the possibility that Texas could be green. And then we go over here to Arkansas, red, green, blue. That's fine, that's consistent with the green on Oklahoma, but if we look at its neighbors we know that Texas is forever forbidden to be blue, now. So now we go over to Louisiana, and remember, we're rotating our color choices because we don't want to overuse them. So this means that the first choice we're going to make here, for Louisiana, is yellow. And that's fine because it's consistent with Arkansas, but it's not so fine because it's now ruled out the last possibility for Texas. So even though Texas is going to be the 48th state that we color, we're going to say, at this point, there's no need in going on. We'd better back up. Because there's nothing left for Texas when we get around to coloring it. So that means that this yellow is ruled out here. This yellow reappears. We select the next color in my line for Louisiana, which happens to be red. And now that's consistent with this yellow that's still left for Texas. And it's also consistent with the blue that's up here for Arkansas. So that's cool. I wonder if maybe we could make an algorithm out of that, and solve problems like this. Do you see, sort of, the intuition of what we're doing? We're actually using the martial arts principle, again. Because the whole problem is that local constraints, undiscovered local constraints, are causing downstream problems. So we're going to use the enemy's powers against him, and we're going to look at those local constraints as we go and make sure they're not going downstream, not going to get us later on. So now I'm going to look like I'm getting a little formal, but I'm just getting a little bit more formal. Because I want to have some language that I can use to describe what's going on, so that it's clear what the choices are. So to start off with, we're going to have to have some vocabulary. So let's start up our vocabulary here. We're going to have a notion of a variable v. And that's something that can have an assignment. There's nothing complicated about that. And a value. A value x is something that can be in assignment. It's a little bit circular, but we're all in computer science so you know what I mean. So the next thing is a little slightly less obvious, and that's the notion of a domain, d. And that's going to be a bag of values. OK. So one more thing. A constraint. That's a constraint c, is a limit on-- in our examples it's mostly going to be pairs of variables, pairs of variable values. But in general, it could be variable values. So if we go back here to Texas we could say, OK, how does our vocabulary drape itself over that configuration? And the answer is, the states have the role of variables, the colors have the role of values. And the domains are the remaining color possibilities that we can still use on a particular state. And the constraint, in this case, is the simple map coloring constraint that no states that share a boundary can have the same color. So states are variables, colors are values, domains are bags of colors, and constraints-- there's only one-- adjacent states can't have the same color. So that's how it fits with this vocabulary. So now, what did I actually do here? Well what I actually did here, I'm going to now formalize a little by writing it down in pseudo code. So here we are, we're going to have a look at what we did here with our intuition, and we're going to reduce it to a procedure. And here's the procedure. Remember, we're doing depth first search on this stuff. I did a depth first search. We're going to do a depth first search, and for each depth first search assignment-- OK, so here I am, I'm labeling Arizona, and then Oklahoma, and then Arkansas, and then Louisiana. When I give each one of those a label, a color, I'm going to do this procedure. Every time I make one of those assignments. The last one that caused trouble was coloring Louisiana yellow. Each time I put one of those colors down, each time I make an assignment, I'm going to do this procedure. So for each depth first search assignment, for each variable v, considered. Now you don't know what I mean by considered. But when I put a label, when I put up a value, show something as a color for Louisiana, I thought about Texas. So I was considering the variable, Texas, when I made the assignment for Louisiana. Now I'm going to be a little bit vague about what I mean by considered, right now. Because there are lots of choices about how much stuff you actually consider. So let me just say consider, and then we'll open that up and talk about the options in a moment, so for each variable v considered for-- let's call that variable v sub i-- for each x sub i, for each value in the domain of that variable, consider each of the things that still surviving. For each of those, for each constraint c, that's between x sub i, and some x sub j, where x sub j is an element of the domain of j. Now that sounds awfully fancy, but this just says, in the case of Texas up there, whenever I consider one of the values that's still remaining as a choice for Texas, I want to consider all of the constraints between that variable and the adjacent states. And I want to be sure that anything I leave in the domain is OK for some selection in the other states, some remaining choices in the other states. So that's why we're getting pretty nested here. But we're doing depth first search. We are considering the variables in a certain collection of variables. For each one of those where considering all the values that still remain in the domains of those variables. And then for each of those values, we're checking to see if it satisfies this some constraint, satisfies the constraint that are placed upon it. So for each of these constraints if there does not exist an x sub j, such that, the constraint between x sub i and x sub j is satisfied, well if in that adjacent place there's nothing that is consistent with a value, then we've got to get rid of it. If that's true. If there does not exist some value in an adjacent variable such that that constraint is satisfied, we're hosed. We've got to get rid of that value. So we're going to remove x sub i from d sub i. OK. Now, that's fine. That's sort of what I did with Texas. As soon as I plopped down a value for Louisiana I said, well what are the possible values in Texas? Red, green, blue and yellow. Let's consider red. Let's consider the constraints between Texas and all adjacent states. One of those constraints says it can't be the same color as Arizona. The only color I've got available for Arizona, since I've already made the assignment is red. Red is not consistent with red, so I've got to get rid of it. So it looks complicated, but it's just intuition. So what do we do if we get to a situation where the domain is empty? That means whenever we get around to making assignment to it, there won't be anything left. So if that ever happens, if the domain ever becomes empty, then what do we do? We've got to back up. So the intuition is clear. This is the algorithm. The algorithm when you work through it, think about whether it makes sense and what not. How it fits with Texas. Yeah, it sure does. All we're doing is we're making these death first assignments. And in the neighborhood of those depth first assignments we're looking around to see if the values that are possible include something. And if they don't include anything, we know we made and irrevocable blunder, and we have to back up. So that's the essence of the idea. Now, how well does it work? Well a little bit depends on what we choose for considered. There are lots of choices for what we consider. So let me enumerate some of those choices and then we'll have a look and see what they do. Oh I guess one possibility is to consider nothing. Let's try it out. So our type of search is going to be no checks. What do you think is going to happen? We're not even checking the assignment. That's pretty fast. Unfortunately, since we haven't even check the most recent assignment, we get lots of places where there are states that are adjacent to each other that have the same color. That's no good. So another thing we can do is consider everything everything. That's no good, because that would say, as soon as we color our first state, we check to make sure that all 47 other states can be colored. That seems like it over doing it a little bit. But in any case, at least we want to check the assignment. So if we go back here and check the assignment, let's see what happens. Type assignments, assignments only. Boom. Aw, gees, that's where I got in trouble before. This is the thing is going to run for 17 billion years at nanosecond or something like that. It's only a billion years if you run it at nanosecond speed, so I guess maybe you could do that. Have a fast computer. But this isn't going to work because of our unfortunate choice of Texas as the last state to be considered, and the unfortunate coloring of the four surrounding states right up front. And our unfortunate decision to rotate the color so as to avoid overdoing any one color. So this doesn't work. We know we went to the trouble of working out the business with Texas by hand, using the domain reduction algorithm. Better make a note that this is the domain reduction algorithm. And what we're going to do is we're going to check the neighbors of the assignments. Just like we did here. We checked Texas each time we made one of those four choices, because it's a neighbor of all of the choices of the states that we made. So one thing to do is to check neighbors. This is one, two, three, now let's see what happens. Check neighbors only, go. Shoot, I don't know. It's OK with Texas, right? Because it didn't color the states around Texas with all of the four color choices. But it's still getting into trouble in other places. Like the states like Missouri, Kentucky, Virginia, Tennessee, states with lots of boundaries. So I don't know whether this is going to-- oh, there it finally worked. It went through a lot of effort, though. For the sake of comparison, we might make a note that it ran into 9,139 dead ends. But it did do some good. It didn't take a length of time longer than the remaining part of the universe. But if it's a good idea to check the neighbors, if we make a change to the neighbors, what might that suggests that we do in addition? Well it might suggest that if we make a change to a neighbor, that we check its neighbors, too, make sure they're all right. So another choice is to propagate. So propagate through variables with reduced domains. Let's see how that works. Wait a minute. I must have made a mistake. Let's try that again. Oh, maybe we better slow it down. All that grey stuff is showing the limit of the propagation. Man it's, let's see at four second of 40, that's about 10 seconds. Boom. Not bad. Zero dead ends. And it was a lot faster. I didn't happen to notice how many constraints were checked on that other thing, I think it was around 20,000 or so. This is a lot less. So this looks like a good idea. But why did I label it number five? Well because there's something between this and number three. So number four is, through v with d reduced to one value. So we're not going to propagate through all of the variables which have their domains shrunk a little bit. We're only going to propagate through those that have the greater shrinkage, all the way down to a single value. So let's see how that might work. Anybody want to place any bets on this one? Let's see. We checked 2,623 constraints last time. Let's see what happens this time. You can see that the extent of the gray is less, because it's not propagating so far. And as we breathily await the answer, I'd say we've found our winner. As this does a couple of dead ends, but the number of constraint checked is less than 1,000. So in general, with problems this, you have all of these choices for what you consider. You don't want to consider nothing, because then you're not honoring your constraints. You'll certainly want to consider the things you just made assignments for, because otherwise you'll construct a solution that violates a constraint. You don't want to do everything, because that's excessive work. And so checking the neighbors is a good idea, but it's always better in practice. In practice, inevitably it's the case that it's better to do some propagation through the things that you've changed. How much propagation? It doesn't seem to do much good to propagate through things are just changed. But it does seem to do some good to propagate through the things that have changed and been reduced to a single value. So as soon as you get a neighbor of some assignment you just made that has its domain reduced to a single value, then you check its neighbors, too. So you check the neighbors, of the neighbors, of the neighbors, of the neighbors, on and on and on, as long as you've found a domain being reduced. And not only being reduced, but reduced to a signal value. All right? So that's the demand reduction algorithm. And I guarantee you a problem like that. And I know you don't know how to work those problems yet, because this is a little bit abstract. And to work these problems in the exam setting you need to know a little bit about how to keep track of the variable values that remain in the domain, and that sort of thing. And you'll learn more about that in your recitations, and in this mega recitation, and in the tutorials. So we could go home except that there are few little flourishes to deal with here. And those flourishes, include some dirty, filthy little secrets. For example, I've chosen, as my classroom example, to pick on Texas. And arranged for this situation to be especially ugly. So I could arrange the states in a different way. We have highly constrained states, that have a lot of bordering states around them. And we have other states, like Maine, up there, that only borders on one other state. So I don't know. Will, what do you think? Should we arrange the states for our death first search in the order of least constrained to most constrained, or most constrained to least constrained? In other words, should we start with Missouri, or Tennessee, or Kentucky, or something like that? Or should we start with Maine? What do you think? You have a 50% chance of getting it right, just by [? looking ?] at points. WILL: Start with the most first. PROF. PATRICK WINSTON: He thinks we should start with the most constraint first. Do we have a volunteer who wants to suggest that we start with the least constraint first? That's the way I always work on stuff. I'm working on a book or something, I have 500 things to fix, I'll always choose to work on the easiest stuff first, so that I feel like I'm making the list a lot smaller. Leave the hardest things to last. But we don't have any volunteers who want to bet on that idea of least constraint first? OK. Jason wants to suggest that we should work on least constraint first. Well we have ground truth, because we can just try it out. I guess we'll stick with our shrinking to one value thing, here. But we will reorder things so that we have the least constrained first. So right away, we got a color for Maine. Maybe we ought to speed this up a little bit. Well, that's a good idea. Jason suggested this and we only 1,732 constraints and we had 59 dead ends. So let's try the other way around, and we'll go back to four frames a second. So we're working, kind of from the middle of the country out, with this one. We're going to deal with Maine, I guess, last. Which is better. Too bad, I think it looks like this is better. In fact, let's not be so aggressive with the use of constraint propagation. Let's just check the assignments only. If we go back to an arrangement where we have least constrained first, and we'll crank up the speed. Well actually, we would have to crank it up pretty big, because the states like Missouri, Tennessee, Kentucky, they're going to be like Texas. And so were kind of back to the length of the universe type problem, here. With the least constraint first, and no use of constraints, other than to check the current assignment. So let's stop that, though, and check the most constrained first, assignments only. I don't know how long's this going to take. That's the dirty little secret. If we had arranged our states from most constrained to least constrained, ordinary depth first search with none of this stuff we talked about today would work just fine. All right. So it's a little bit like games. Do you use progressive deepening, or do you use alpha beta? And the answer is both. You use everything you've got to deal with the problems. And depending on the problem, one or another of the things you incorporate into your approach will work just great, if you're lucky. So now, I promise that this is useful not only for people who want to color maps. God, who wants to do that? We know it can be done with four colors. But it's also useful for doing all kinds of resource planning problems. So I want to show you a resource planning problem, and I want you think about-- while I'm doing it-- think about whether it's actually analogous to the map coloring problem. All right? So here's the deal. You have just landed a summer job with the Jet Green, a new airline. And Jet Green is a low cost, no frills, hardly any maintenance type of airline. And they want to fly mostly between Boston and New York. Occasionally they want to fly to Los Angeles. And they're trying to get by with the smallest number of airplanes. So that's why we have a kind of resource allocation problem with Jet Green. So I'm going to write down what their schedule looks like. They have one flight, F1, that goes from Boston to JFK, like so. It's an early in the day flight. Then they want to have another one, F2, that flies from JFK to Boston. And then they want to have another flight a little later in the day that flies from Boston to JFK. And a little later than that, they want to have another flight that goes from JFK to Boston. They're going to start off mostly as a shuttle airline in the beginning. So that's F1, F2, F3, and F4. And they have a fifth flight, F5, that goes from Boston to Los Angeles, that takes a long time. So it looks like this on the schedule. Of course we have time going that way. So that's Boston to LAX. Now your job is to determine if they can fly this schedule with four aircraft? And naturally you don't want to over use any one aircraft, because you would like to have even wear on them. Right? So as you make your choices, you'll rotate the aircraft. So you'll assign to this one to A1, aircraft number one. This one will be A2. This one will be A3. And this one will be A4. And, oops, there's no aircraft left for the flight to Los Angeles, because you only have four. So, it's obvious, right? This is 100%, exactly, the map coloring problem, even down to the four choices. Because, you have the constraint, the no single physical aircraft can fly two flights at the same time. Just like no two adjacent states can be colored the same. So there's a no same time constraint, like so. So if you were assigning aircraft to these flights, you would get down to F4, the fourth flight, and you would say, well, let's see, this guy down here can be A1, A2, A3, or A4. But if I choose A4 for that fourth flight, then there would be nothing left in its domain. So you've thus set the problem up to be identical to the map coloring problem. And, of course, you can enrich it with other kinds of constraints. So, for example, you might have-- this is a not same time constraint-- and these, I mean, this is at JFK. And it flies out of JFK, so maybe you can use the same aircraft for those. But not if they're right up against each other, because you have a minimum ground time rule. So there's a minimum ground time constraint here. And there's a minimum ground time constraint here. There's a minimum ground time constraint here. And if these are at the same city, then you've got to allow enough time for them to fly between the two cities that are involved. So the constraints can get a little bit more complicated, but the idea is the same. So you say to me, I don't trust you, show me. OK. So let me show you. Oh, by the way, there's one more way to make this map coloring problem easier, right? Let's see. The arrangement is going to be alphabetical, the type is going to be assignments only, and we know that's a loser. But not if we use a whole lot of colors. So it's the use of four colors that got us into trouble. Now that's an aside, but it'll be coming back in a moment or two. Scheduling, here's that problem. Boom. You can almost see it working just like the map coloring thing. But that's maybe too easy. Let's do this one. See this is just like [? simplicia. ?] We've kind of got a goofy arrangement here that's guaranteed to lose at the bottom because of choices made at the top. But that's OK. We can that we can stop this, and we can change to check neighbors only. And boom, there it goes. Or alternatively, let me see if I can do it this way. Most constrained first, type will be assignments only. Boom. That worked fine, too. So you might choose to have a slightly harder problem, like this. And we can search away. Actually I don't know if this will complete or not. This is a randomly generated example. But you're going to lose your summer job if you can't figure out whether you can do this, or not. So what are you going to do? [? Elliott, ?] you got any thoughts about how you're going to save your job? So here's the question you've been asked. How many airplanes does Jet Green need? And you decide, well, four seemed to work before, so we'll try four here. You're not sure if it's going to terminate or not, I mean, in your lifetime, let alone in your summer job. [? Elliott, ?] let me give you a hint. Look at the outline. The outline up here, on the board, the last item. [? ELLIOTT: [INAUDIBLE] ?] PROF. PATRICK WINSTON: Yeah, what's that mean? What's the maximum number of airplanes we're going to need? Suppose we've got five flights, what's the maximum number of airplanes we would ever need? Five. What's the minimum number of airplanes we'll need? One. So let's try it with a small number of airplanes and a large number of airplanes. So that showed us very fast that we can't do it with one airplane. That showed us very fast we can do it with ten airplanes. SPEAKER 1: [INAUDIBLE] amount of overlappage from up [INAUDIBLE]. PROF. PATRICK WINSTON: Volunteer? What are we going to do to find the actual number, as fast as possible. And at least give our boss a reasonable answer, even if not necessarily the exact number very fast? It's easy, right? We're going to start up here with one computer, and we're going to start down here with another computer, and see what happens. So let's see if we can do it with nine. Let's see if we can do it with eight. Let's see if we can do it with seven. These take almost zero time, right? Because they're under constraint. Wow, that's good, seven. Let's try six. Actually let's try two. It loses fast. Let's try three. I don't know. Maybe if you let it run one long enough three will work. I doubt it. While we're at it though, we might as well go back here and try it with six. Remember seven worked real fast. Gees, six, that was six, right? Yeah. So let's try it with five. OK. So it runs real fast with five. It terminates real quick with two, so we got three and four left. So we could tell our boss, a la any time algorithm, that you're not real sure, but you know it's going to be either three or four. And then, you got two computers. You can let both run and see if either one terminate. So you have three and four. My guess is that three will eventually give up. But of course, there's another little problem here. We haven't used the most constraint first. If we did that, we might be able to do it even faster. Actually, I don't think I can make that switch without getting another random assignment, but let's see what happens. Maybe so. SPEAKER 2: [INAUDIBLE] PROF. PATRICK WINSTON: Oh, I already had most constraint first? OK. So it didn't help to actually switch. And I think I've got a new schedule to work here. So that's the end of the story. You can do good resource allocation if you do several things are once. Number one, you always want to use most constraint first. Number two you want to propagate through domains produced to a single algorithm. And number three, if you really try to figure out what the minimum number of resources needed is, you do this under over business and you can quickly converge on a narrow range where the search is taking a long time, and be sure that it lies within that narrow range. Because when you over resource, it's fast to complete, and when you under resource, it's fast to terminate. So you can just squeeze it right down into a very small range. And that is the end of the story. Enjoy your holiday on Monday. We'll have two classes next week on Wednesday and Friday, as advertised. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 22_Probabilistic_Inference_II.txt | PROFESSOR PATRICK WINSTON: I was in Washington for most of the week prospecting for gold. Another byproduct of that was that I forgot to arrange a substitute Bob Berwick for the Thursday recitations. I shall probably go to hell for this. In any event, we have many explanations, none of them good. But today we'll try to get back on track and you'll learn something fun. In particular you will learn how a graduate student of mine Mark [? Phillipson ?], together with a summer UROP student, Brett van Zuiden, one of you-- managed to pull off a tour de force and recognize in these two descriptions the pattern that we humans commonly call "revenge." It was discovered. The system didn't have a name for it, of course. It just knew that there was a pattern there and sat waiting for us to give a name to it. That's where we're going to end up. But it'll be a bit of a journey before we get there, because we've got to go through all that stuff on the outline. And in particular, we want to start off by a little tiny bit of review. Because some of the stuff we did last time went by pretty fast. In particular, you may remember they had this wonderful joint probability table, which tells us all we want to know, all we want to know. We can decide what the probability of the police being called is given the this and the that, and all that sort of stuff, by clicking the appropriate boxes. The trouble is, gee, there are only three variables there. And when there are lots of variables it gets pretty hard to make up those numbers or to even collect them. So we're driven to an alternative. And we got to that alternative just at the end of the show a week ago. And we got to the point where we were defining these inference nets, sometimes called "Bayes nets." And the one we worked with looked like this. There's a burglar, a raccoon, the possibility of a dog barking, the police being called, and a trash can being overturned. So more variables than that. That only has three. This has got five. But we're able to do some magic with this because we, as humans, when we define-- when we draw this graph we're making an assertion about how things depend or don't depend on one another. In particular, there's something to break down and memorize to the point where it rolls off your tongue. And that is that any variable on this graph is said by me to be independent of any other non-descendant given its parents. Independent of any non-descendant given its parents. So that means that the probability of the dog barking, given its parents, doesn't depend on T, the trash can being overturned. Because the intuition is all of the causality is flowing through the parents and can't get to this variable D without going through the parents. So that is [? inserted ?] property of the nets that we draw. And we tend to draw them in a way that reflects causality. So it tends to make sense. So somehow this thing is going to be-- we're going to use this thing instead of that thing. But wait. We may need that thing in order to do all the computations we want to perform. So we need to be able to show that we can get to that thing by doing calculations on this thing. So what to do? Well, we're going to use the chain rule. And remember that the chain rule came to us by way of the basic Axioms of Probability plus the definition plus a little colored chalk. So we got to the point last time where we sort of believed this. It's a really magical thing. It says that the probability of all this stuff happening together is given as the product of a bunch of conditional probabilities. And the conditional probabilities in this product are arranged such that this first guy depends on everybody else. The second guy doesn't depend on the first guy but depends on everything else. So that list of dependencies gets smaller and smaller as you go down here until it depends only one thing. There's no conditional at all. So that's going to come to our rescue because it enables us to go from calculations in here to that whole table. But first I have to show you a little bit more slowly how that comes to be. One thing I'm going to do before I think about probability is I'm going to make a linear list of all these variables. And the way I'm going to make it is I'm going to chew away at those variables from the bottom. I've taken advantage of a very important property of these nets. And that is there no loops. You can follow the arrows in any way so as you get back to yourself. So there's always going to be a bottom. So what I'm going to do is I'm going to say, well, there are two bottoms here, there's C and T. So I have a choice. I'm going to choose C. So I'm going to take that off and pretend it's not there anymore. Then I'm going to take this guy. That's now a bottom because there's nothing below it. I've already taken C out. So we'll take that out next. And now I've got this guy, this guy, and this guy. This guy no longer has anything below it. So I can list it next. Now over here I've got raccoon and trashcan. But trashcan is at the bottom. So I've got to take it next because I'm working from the bottom up. I want to ensure that there are no descendants before me in this list. So finally I get to raccoon. So the way I constructed this list like so ensures that this list arranges the elements so that for any particular element, none of its descendants appear to its left. And now that's the magical order for which I want to use the chain rule. So now I can write-- I can pick C to be my variable n. And I can say that the chain rule says that the joint probability of all these variables P of C, D, B, T, and R-- the probability of any particular combination of those things is equal to the probability of C given everybody else. Next in line is D given everybody else. Next in line is T-- next in line is B given everybody else. And next in line is T given everybody else. And finally, just R. So this combination of things has a probability that is given by this chain rule expression. Ah. But first of all, none of those expressions condition any of the variables on anything other than non-descendants, all right? That's just because of the way I've arranged the variables. And I can always do that because are no loops. I can always chew away at the bottom. That ensures that whenever I write a variable, it's going to be conditioned on stuff other than its descendants. So all of these variables in any of these conditional probabilities are non-descendants. Oh wait. When I drew this diagram, I asserted that no variable depends on any non-descendant given its parents. So if I know the parents of a variable I know that the variable is independent of all other non-descendants. All right? Now I can start scratching stuff out. Well, let's see. I know that C, from my diagram, has only one parent, D. So given its parent, it's independent of all other non-descendants. So I can scratch them out. D he has two parents, B and R. But given that, I can scratch out any other non-descendant. B is conditional on T and R. Ah, but B has no parent. So it actually is independent of those two guys. The trashcan, yeah, that's dependent on R. And R over here, the final thing in the chain, that's just a probability. So now I have a way of calculating any entry in that table because any entry in that table is going to be some combination of values for all those variables. Voila. So anything I can do with a table, I can do in principle with this little network. OK? But now the question is, I've got some probabilities I'm going to have to figure out here. So let me draw a slightly different version of it. So up here we've got the a priori probability of B. Well, that's just probability of B. Down here with the dog, I've got a bigger table because I've got probabilities that depend on the values of its parents. The probability of dog barking depends on the condition of the parents, nothing else. So let's see. I've got to have a column for B. I've got to have a column for the burglar and the raccoon. And there are a bunch of possibilities for those guys. But once I get those then I'll be able to calculate the probability of the dog barking. So there are two of these variables. So there are four combinations. There's T T. There's T R, R T, and-- whoa, what am I doing? Wake up! T false. False true. And false false. So what I really want to do is I want to calculate all of these probabilities that give the probability of the dog condition of the burglar and the raccoon. Similarly, I want to calculate the probability of B happening doesn't depend on anything else. So I don't know what to do. Well, what I'm going to actually do is I'm going to do the same thing I had to do up there. I'm going to keep track of-- I'm going to try a bunch of-- I'm going to get myself together a bunch of data. Maybe I do a bunch of experiments. Maybe somebody hands it to me. But I'm going to use that data to construct a bunch of tallies which are going to end up giving me the probabilities for all of those things. So I don't know, let's see. How should we start? Step one, find colored chalk. Step two, I'm going to extend these tables a little bit so I can keep track of the tallies. So this is going to be all the ones that end up in a particular row. And these are going to be the ones for which dog is true. Similarly, I'm going to extend this guy up here in order to keep track of some tallies. This is going to be the ones for which B is true. And this one will be all. So that's my set up. And now suppose that my first experiment comes roaring in. And it's all T's. So I have T T T. That's my first experimental result, my first data item. So let's see. The arrangement here is burglar, raccoon, dog. So burglar as a true. And there's one tally count in there. Likewise, the T T, that's the burglar and the raccoon, that brings me down to this first row. So that gives me one tally in there and dog is true so that gives me a tick mark in that one. All right? Are you with me so far? And now let's suppose that the next thing happens be all false. Well, burglar is false. But there is one experiment. Everybody's false. So we come down here to false false. And that's the row we're going to work on. We get a tally in there. Do we put one in here? No, because that's false. Dog is false. That's what our data element says. So that's cool. Maybe one more. Let's suppose we have T T F. Well in that case, we have a tick mark here and a tick mark here because the burglar element is true. Then we have T T. That brings us to the first row again. So we get a tick mark there. But dog as false, so no tick mark there. That's how it works. I suppose you'd like to see a demonstration, right? Always like to see a demonstration. So here's what it actually looks like. So on the left you see the network as we've constructed it, with a bunch probabilities there. And what I'm going to do now is I'm going to start simulating away so as to accumulate tick marks, tally marks, and see what kinds of probabilities that they indicate for the table. I happen to be using a process for which the model on the left is a correct reflection. So there's one simulation. So the dog barking-- let's see, the burglar is false. The raccoon is true. I get one tick mark. So the probability there is one. Of course, I'm not going to just go with one. I want to put a whole bunch of stuff in there. So I'll just run a bunch more simulations. No [? dice. ?] I don't even have an entry at all yet for T F here. That's because I haven't run enough data. So let me clear it instead of doing it one at a time. Let me run 100 simulations. See, it's still not too good. Because it says this T T probability true. This just because I'm feeding it data, right? And I'm keeping track of what the data elements tell me about how frequently a particular combination appears. Yes, [INAUDIBLE] STUDENT: So when you're doing one simulation, is that [INAUDIBLE] variables? PROFESSOR PATRICK WINSTON: When I'm doing one simulation, I'm just keeping track of that combination in each of these tables. Because it's going to tell me something about the probabilities that I want reflected in those tables. So it's pretty easy to see when I go up here to burglar. If I have a lot of data elements, they're all going to tell me something about the burglar as well as the other variables. So if I just look at that burglar thing, the fraction of time that it turns out true over all the data elements is going to be its probability. So now when I go down to the joint tables, I can still get these probability numbers. But now they're conditioned on reticular condition of its parents. So that's how I get these probabilities. So I didn't do too well here because that T T combination gave me an excessively high probability. So maybe 100 simulations isn't enough. Let's run 10,000. So with that much data running through, the probabilities I get-- let's see, I've got 893 here, instead of 0.9, 807 instead of 0.8, 607 instead of 0.6. And that one's dead-on at 0.01. So if I run enough of these simulations, I get a pretty good idea what the probabilities ought to be given that I've got a correct model. OK, so that takes care of that one. And of course, I didn't draw the other things in here. But by extension, you can see how those would work. Oh. But you know what? I think I will put a little probability of raccoon table in here. Because the next thing I want to do is I want to go the other way. This is recoding tallies from some process so I can develop a model. But once I've got these probabilities, of course, then I can start to simulate what the model would do. All right? How would I do that? Well, do I want to use the same table? I think just to keep things sanitary, what I'll do is I'll go over here and do it again. Here's B. It's got a probability of B. Here's R. Here's a table probability of R. That comes down into a joint table for dog. And it's got four elements. Depending on the burglar condition and the raccoon condition, we get a probability of dog. And now, imagine these have all been filled in. So what do I want to do if I want to simulate this system generating some combination of values for all the variables? Well, I do the opposite of what I did when I was working around with this chain rule showing that I could go from the table to those probabilities. Now I've got the probabilities. I'm going to go the other direction. Instead of chewing away from the bottom, I'm going to chew away from the top. Because when I go into the top and chew way, everything I need to know to do a coin flip is there. So in particular, when I go up in here, I've got the probability of burglar now. So I'm going to use that probability to flip a coin. Say it produces a T. So that takes care of this guy. And I can now scratch it off since it's no longer in consideration. It's no longer a top variable. So now I go over into raccoon and I do the same thing. I take this probability. I do a flip. And say it produces an F. Whatever its probability is, I flip a biased coin and that's what I happen to get. But now, having dealt with these two guys, that uncovers this dog thing. And I've got enough information, because I've done everything above, to make the calculation for whether to dog is going to be barking or not. But wait. I have to know that I've got a T and a T and a T and an F and an F and a T and an F and an F. Because I have to select the right row. So I know that B is T. And I know that R is F. So that takes me into the table into the second row. So now I get this probability. I flip that coin and I get some result, say, T. Voila. I can do that with the other two variables. And I've got myself an experimental trial that is produced in accordance with the probabilities of the table. OK? Of course-- yeah, in fact, how did I get those numbers? Actually what I did is I used the model on the left to generate the samples that were used to compute the probabilities on the right. So you've seen that a demonstration of this already. Now of course-- I don't know, all of this sort of depends on having everything right. I've written a thing to write it one more time. Burglar, raccoon, dog, call the police, trashcan. But somebody else may say, oh, you've got it all wrong. This is what it really looks like. The dog doesn't care about the raccoon at all. So that's a correct model. Now when I do a simulation, I could fill in the tables in either model, right? I'm sure you'd like to see a demonstration. So let me show you a demonstration of that. So there are the two tables. And I can run 10,000 simulations on those guys, too. Now, look. The guy on the left is a pretty good reflection of the probabilities in a model I used to produce the data. But the guy on the right doesn't know any better. it just fills in its own tables, too. So what to do? I say this one's the right model. And you say that one's the right model. Who's right? Maybe we'll never know. And the guy on the left will get rich in the stock market and the guy on the right will go broke. I would be nice if we could actually figure out who's right. So would you to see how to figure out who's right? Yeah, so would I. What we're going to do is we're going to look at naive Bayesian inference. And that's our next chore. So here's how it works. We know, from the definition of conditional probability, we know that the probability of A given B is equal to the probability of A and B divided by the probability of B, right? Equal to by definition. So that means that the probability of A given B times the probability of B-- I'm just multiplying it out-- it equal to that joint probability. Oh, but by symmetry, there's no harm in saying I can turn that around and say that the probability of B given A times the probability of B is also equal to that joint probability, right? I've just expanded it a different and symmetric way. If I've got to write a, b on B, b, a on A. Thank you. Who was complaining? Good work. That would have been a major-league disaster. But now, having written that, I can forget about the middle. Because all I'm really interested in is how I've turned the probabilities around in that conditional. Why would I care about doing that? By the way, we're now talking about the work of the Reverend Bayes. Because we can rewrite this yet again as the probability of A given B is equal to the probability of B given A times the probability of A divided by the probability of B. That's just elementary algebra. But now I'm going to do something magical. I'm going to say I've got a classification problem. I want to know which disease you have. That's a classification problem. Maybe you've got the swine flu. Maybe you've got indigestion. Who knows. But I get all these symptoms. I get all these pieces of evidence. You've got a fever. You're throwing-- oh, well, let's not go into too much detail, there. But what I'm going to do is I'm going to say, well, let's suppose that A is equal to a class that I'm interested in, the disease you've got. And B is equal to the evidence, the symptoms I observe. Voila. I may have a pretty hard time figuring out what the probability of the class is given the evidence. But figuring out the probability of the evidence given the class might not be so hard. Let me get another board in play and show you what I mean. By plugging class and evidence into Bayes' rule, what I get is the probability of some class given the evidence is equal to the probability of the evidence given the class times the probability of the class divided by the probability of the evidence. Now you've got to let that sing to you a little bit. Suppose I've got several classes that I'm trying to decide between. I'm trying to select the best out of that batch of classes. Well, I've got the evidence. And if I know the probability of the evidence given each of those classes, and if I know, a priori, the initial probability the class, then I'm done. Because I've got the two elements in the numerator. Why am I done? Because the denominator is the same for all the classes. It's just the probability of the evidence. And then I could just sum everything up. I know it adds to 1 anyway. So that's cool. But sometimes there's evidence-- actually there's more than one piece of evidence. Let's say that there's some class. some i, and we're trying to figure out if that's the correct class. So we've got c sub i there and c sub i there. And suppose that that evidence is actually a bunch of pieces of evidence. So it could be e sub 1, e sub n, oops, premature right bracket. All that evidence, given the class i times the probability of the class i over some denominator that we don't care about because it's going to be the same for everybody. So we'll just write that as d. Now what if these pieces of evidence are all independent given the class? So if you have the swine flu, the probability you have a fever is independent of the probability you're going to throw up, say. Then can we write this another way? An easier way? Sure. Because when things are independent, the joint probability is equal to the product of the individual probabilities. So that is to say-- it's easier to see it if you write it down than if you just say it-- this probability here from these two elements here is equal to the probability of e sub 1 conditioned on c sub i times the probability of e sub 2 conditioned on c sub i, all the way down to the probability of e sub n conditioned on c sub i divided by some denominator we don't care about. See, what I'm going to try to do is I'm going to go through this for all the ci and see which one's the biggest. STUDENT: That's the [INAUDIBLE] ci, right? PROFESSOR PATRICK WINSTON: This is the probability of-- STUDENT: [INAUDIBLE] right-hand side [INAUDIBLE]. PROFESSOR PATRICK WINSTON: Right here? Oh yes, you're quite right. Oh yeah, thanks. I can't write and think at the same time. Thanks. OK. So I've just figure out which one of these is the biggest. And I've identified the class. Now you say to me, well, I would like to see an example. So-- I don't know, does anyone have any spare change? A nickel, a quarter. This is not because of infinitesimally low raises here at MIT. I just need it for a demonstration. I need two coins. Don't forget to get these back, I tend to be-- Now suppose these two coins are not exactly the same. One of these points is a legitimate, highly-prized American quarter. The other one is a fake. And with this one, the probability of heads, let us say, is 0.8 instead of 0.5. So I mix these all up. And I pick one. And I start flipping it. And I get a head. Then I flip it again. And I get a tail. Which coin did I pick? Well, we're going to use this stuff to figure it out. Here's what happens. Before I forget. Thank you very much. So what we've done is we've selected these things from my hands. And I can't draw hands. So I'll draw a little cup here. And there are two coins in here. And we're going to pick one. And one has a probability of heads equal to 0.8. And this one has a probability of a head of 0.5. So here's the draw. I pick one. Each has a probability of 0.5. This one is the one with the 0.8 as the probability of head. And this one is the one with the probability of 0.5 as a head. OK? So now suppose the first flips as it was is T. Well, that's a piece of evidence. That's here. Probably of evidence given the class. Well in the case of having drawn this biased coin, the probability of coming up with a tail-- ah, let's say a head, just to make my numbers a little easier. Probability of coming out there with a head is equal 0.8 given that it's up here in this choice. The probability given that you have a fair coin is 0.5. So now if we take the next coin and take it to be a tail then the probability of this guy given that evidence is 0.2. And the probability of this guy given that evidence-- it's a fair coin, so it doesn't care. It's still 0.5. So now what's the probability of this class given this evidence? It's the product 0.5 times 0.8 times 0.2. And what's the probability of this guy? It's 05 times 0.5 times 0.5, divided by a denominator which is the same in both cases. So let's forget about this early 0.5 here. Because it's the same in both cases. And we just multiply those numbers together. That gives us 0.8 times 0.2. What's that? 0.16? And this guy, 0.5 times 0.5, that's 0.25. So it looks an awful lot like-- with this combination-- that I've picked the coin that's fair. One more flip? So let's flip it again, and suppose we come up with a head. So that puts a 0.8 in here. And 0.5 in here. When you multiply those out that's 0.125. And this is 0.128. So it's about equal. So you see how that works? All right. So we're using the coin flips as evidence to figure out which class is involved. OK so I don't know, you'd probably like to see a demonstration of this, too, right? You say to me, gosh, just two kinds of coins. That's not very interesting. Let's try five kinds of coins. So what I want to show you is how the probabilities for all these coins-- there are five of them, color-coded-- how the probabilities vary with a series of flips. Let's suppose I've got a head-- the grey line, by the way, is the fraction of heads-- so that's going to be one. Because I'm just doing heads. You see that black line rising? Should look like a rocket. That's the probability that the-- that's the coin which only shows heads, the probability of head is 1. And I'm flipping a whole bunch of heads here. Isn't that cool? Now what happens if I suddenly put in a tail? By the way, you'll no doubt, here one the extreme left-- the initial probability of the P=0 coin was 0.1. As soon as I flipped a head that went to 0. And it will never get off 0, right? That makes sense. Because if the probability that you'll get a head is 1 you should never see a tail. If you ever do, that isn't your coin. What happens now if I interrupt a series of heads and produce a tail? STUDENT: [INAUDIBLE]. PROFESSOR PATRICK WINSTON: What's that? STUDENT: [INAUDIBLE]. PROFESSOR PATRICK WINSTON: The black one will go to 0. What else happens? By the way, the blue one is the one with the highest probability of being a head. [INAUDIBLE] Boom! That blue one shot up. Not going up slowly. It shot up. Because now the preponderance of evidence with all those heads is that I've flipped the coin with a bias of 0.75 towards heads. So let's clear this. Pick any probability you want. 0.25, 0.5, and so on. I don't know, let's pick 0.25 since we've been at the upper end. So orange is 0.25. And sure enough, the probability that I've selected the 0.5 coin is going up and up and up and up after the original irregularity. The Law of Large Numbers is setting in. And a probability that I've got that 0.25 coin in play is pretty close to 1. All right. So that's cool. Now you say to me, that's awfully nice but stop. Awfully nice, but not very real-world-ish. So let me give you another problem. It's well-known that you are, with high probability, of the same political persuasion as your parents. So if I wanted to figure out which party a parent belongs to, I could look at the party that their children belong to, right? So it's just like flipping coins. The particular coin I have chosen corresponds to the parent. Individual flips correspond to the political party that the child belongs to. So let's get up a little bit-- by the way, I wrote all this stuff over the weekend. So who knows if any of it will work. But let's see. A parent party classifier. There it is, Democrats and Republicans. And now the prior for being a Republican given here is 0.5. But I don't know, this is a little bit Democratic state. So let's adjust that down a little bit. Somewhere in there might be about right But let's just, for the sake of a classroom illustration, go down here. So now the meter is showing the prior probability because that's the only thing in the formula so far. I've got no evidence. So now let's suppose that child number one is a Republican. Back to neutral. So I've got a low probability that the parent-- a priori probability that the parent is a Republican and a child who's a Republican. I notice that 0.2 and 0.8, the conditional is 0.8. And the prior is 0.2. That's why it comes out to balance each other, right? So now if we get another Republican in there it goes way up. If I have a Democratic child it goes back down. If I have an equal balance between children then it goes way back down because of that prior probability being low. So if I make that high, even though the children are balanced, I'm still going to have a high probability of being a Republican. Now let's see. If I take that slider there, the conditional probability, and drive it to the left here-- let me make that equally in. And let's make that one thing. I don't know. What am I doing now? If I make the probability less than 0.5, what's that mean? That means you're sore at your parents and you want to belong to a different party. All right, so now, what's next? Oh gosh. What's next? This is what's next. What's next to somewhere? Yeah, this is what's next. This here. We've got two models. Remember when I said we wanted to decide between them? Can we use that Bayesian hack to do that, too? Sure. Because we've got these two models. We've got the probabilities in them. So now I can take my data and calculate the probability of a left model given the data and the probability of the right model given the data, multiply that times their a priori probabilities, which I'll assume are equal. Then I can do a model selection deal much in defiance to what I was hinting at before. so let's try that. Whoa. There are my two models. Yes, there they are. We've already trained them up. And they've got their probabilities. Now what we're going to do is we're going to use the original model to simulate the data. So what we're going to do is we're going to simulate draws, simulate events, similarly combinations of all variables using a model that looks like the one on the left, that is the one on the left except for the slight differences in probabilities, OK? Then we're going to do this Bayesian thing and see where the meter goes. So we'll run one data point. Oops, went the wrong way. Makes me nervous. I just finished this at 9:15. Maybe there's a bug. Oops, two data points, swings to the left. Three data points, back to the right. Of course that's not much data. So let's put some more data in. Yeah. Boom, there it goes. Let's try that again. That was cool. So let's run 1,000 simulations and one data point. It bobbles around a little bit and goes flat over to the left. Because that is the model that reflects the one that the data is generated from. So now we got Bayesian classification, except now the classification has gone one step more and it becomes structure discovery. We've got two choices of structure. And we can use this Bayesian thing to decide which of the two structures is best. Isn't that cool? Well, it's only cool if you could do what? So if you had two choices-- you can select between them and pick the best one-- but there are-- gosh, for this number of variables, there are a whole lot of different networks that satisfy the no looping criteria and don't have very many parents. There's an awful lot of them. In fact, if you strict this network to two parents there are probably thousands and thousands of possible structures. So do I try them all? Probably not. It's too much work when you get 30 variables or something like that. So what do you do? We know what to do, right? We're almost veterans a 6034. We have to search! So what we do is we take the loser and we modified it. And then we modify it again. And we keep modifying it until we drop dead or we get something that we're happy with. So let's see what happens if we change this problem a little bit and do structure discover. We're starting out with nothing linked. And we're going to just start running this guy. So what's going to happen is that the good guy will prevail. And the bad guy will be a copy of the good guy perturbed in some way. So it's a random search. You'll notice that score-- it's too small for you to read. All these things are too small to read. Let me make it a little bigger. Too small to read, but that number on the right there is not the product of the probabilities, actually. It's the sum of the logarithms of the probabilities. They go together, right? And the reason you use this instead of the probabilities is because these numbers get so small that was a 32-bit machine, you eventually lose. So use the log of the probabilities rather than the product of the probabilities. You use the sum of the logs instead of the product of the probabilities. And eventually, you hope that this thing converges on the correct interpretation. But you know what? This thing is so flat as a space and so a large and so telephone pole-like that it's full of local maxima. So what this program is doing is every once in awhile-- I think with probability 1 and 10; I forgot what parameters I used-- every once in awhile, it'll do a total radical rearrangement of the structures. In other words, it's a random restart. It keeps track of the best guy so far. And every once in awhile it does a totally random restart in its effort to search the space. So that's how you go from probabilistic inference to structure discovery. Now when is this stuff actually useful? Well, I hinted at a medical diagnosis, right? That's a situation where you've got some symptoms. And you want to know what the disease is. So as soon as you use the keyword "diagnosis," you've got a problem for which this stuff is a candidate. So what other kinds of diagnosis problems are there? Well, you might be lying to me. So I can put a lie detector on you. And each of those variables that are measured by the lie detector are an independent indication whether you're telling the truth or not. So it's this kind of Bayesian discovery thing. Naive Bayesian Classification. What other kinds of problems speak to the issue of diagnosis? Well, we like to know how well you know the material! So we can use quizzes as pieces of evidence. Thank god we don't use exactly a naive Bayesian classifier, because then we wouldn't be able to do that combination. We have to use a slightly more complex-- what you can think of as a slightly more complex Bayesian net to do that particular kind of diagnosis. You might have a spacecraft or an airplane or other piece of equipment with all sorts of symptoms. You're trying to figure out what to do next, what the cause is. So using the evidence to go backward to the cause. So maybe you've got some program that doesn't work. Happens to me a lot. So I use the evidence from the symptoms of the misbehavior to figure out what the most probable cause is. But now to conclude the day-- last time there weren't any powerful ideas. But if you take the combination of the last lecture and this lecture to be a candidate for gold star ideas, these are the ones I'd like to leave you with. We got here is-- this Bayesian stuff, all these probabilistic calculations are the right thing to do. They're the right way to work when you don't know anything, which would make it sound like you're not very useful, because you think you always-- well, in fact, there are a lot of situations where you either can't know everything, don't have time to know everything, or don't want to take the effort to know everything. So in medical diagnosis all you've got is the symptoms. You can't go in there and figure out in a more precise way exactly what's wrong. So you use the symptoms to determine what the cause is. And then all those other kinds of cases that I mentioned. But now, what other kinds of structure discovery are there? Well, the kind of structure discovery that I hinted at in the beginning will be the subject that we'll begin with during our next and sadly final conversation here in [? 10250 ?] on Wednesday. It will feature not only a discussion of how this stuff can be used to discover patterns and stories, but we'll also talk about what's on the final, what kind of thing you could do next, that sort of thing to finish off the subject. And that's the end of the story for today. |
MIT_6034_Artificial_Intelligence_Fall_2010 | MegaR7_Near_Misses_Arch_Learning.txt | The following content is provided under a Creative Commons license. Your support will help MIT OpenCourseWare continue to offer high quality educational resources for free. To make a donation or view additional materials from hundreds of MIT courses, visit MIT OpenCourseWare at ocw.mit.edu. PROFESSOR: Here we have a near miss learning tree. It's a little bit different but a little bit similar. We've got different types. We're trying to learn about different light sources. So we're concerned about different lamps, flashlights, all sorts of things. So these guys can have had types of support, a base or wires. The base can be flat, clamped, or have legs. The legs can be flat bottomed or wheeled. The light source can be incandescent, fluorescent, or sodium vapor. And the energy source that powers it can be electric, battery, oil, or gas. So there's also some other things. But these are the trees that we might be climbing up with the climb-tree heuristic. Our starting model is we have an incandescent bulb. The height is equal to 24 inches exactly. They have a flat base. There's electricity to power it. And it has a shade. So our starting model is basically like a table lamp that you set on the table. It's got the shade over the top of it. You know, the standard room table lamp is our starting model. So recall in arch learning, which is what we are using here, that there are several heuristics that we might want to use. Let's see, well, the heuristics are-- we've got require-link, forbid-link. We've got climb-tree, extend-set, close-interval. And we also have drop-link-- forgot about that one. That's sort of with the links. I can put it here. OK, so if you want to do well on these questions if they appear on the final, first of all, you better know what those six heuristics do. So basically require-link is a heuristic. It goes into your model as a requirement, where before your model just didn't care. So let's say there was something about color for your light, like for your lamp's base or for the lamp. And right, now it doesn't care about color at all. But let's say that you wound up having an example that was, for instance, blue or white or whatever these colors are. And it wound up getting that blue was OK for-- in its model, it wound up getting, OK, blue and white are OK. Or actually, no, an even more basic example-- let's say that it just kept getting blue lamps as examples, and all of them were blue. And it just didn't care about the fact that they were blue. But they did turn out to be blue. And then eventually it found a red one that was exactly the same as its model, but the red one was a near miss. Then it might require blue, which was already an acceptable, permissible thing in its model. But it could be required. Now, another thing it can do is forbid. Some people ask me, why would you bother saying, forbid red, say, if you can already just require blue? Together they cover the same ground. The answer is because this system, this arch learning system-- remember, it's Patrick's doctoral thesis. It's an old system. Because it was built on old systems and because also it's generally a good idea, it was very parsimonious. Let's say you had a set of colors-- red, blue, yellow, green, pink, purple, orange-- and you found that only orange was giving you problems, you could try to require that it be in this. Or you could even use extend-set, make a giant set of red, blue, yellow, green, purple, pink, but not orange. But if you just forbid orange, you've just saved a whole bunch of space, particularly if there's a large possible series of set. You want to have the ability to forbid. Because it needs to have the smallest model that it can that covers all of its samples. So we also have climb-tree. Climb-tree takes one of the elements in our model and moves up the tree one level. So let's say we said, only incandescent lights are good. But we found that fluorescent lights are also good. Whoop, we move up to light source, climb-tree. Extend-set-- we don't have any sets here. But let's go back to red, blue, orange, yellow. Extend-set is when we say, only red is good as an example. But we see another positive example that's yellow. So we say, OK, I'll extend the set. Red and yellow are both good. All right, close-interval-- close-interval is really obvious when you need to use it. Because not only is it only used for intervals like height equals 24 inches or some kind of numbers, but it is the only thing you can use for when height equals some kind of numbers. Close-interval will let you fuss around with the intervals and say, oh, well I guess one with a 20 inch height is OK, too. So we'll make the whole interval 20 to 24 inches good. Because it doesn't make any sense for it to be like, 20 is good, 24 is good, but nothing in between is good. No, those are awful. So that close-interval covers the entire interval. Last but not least is drop-link. The thing about drop-link is that drop-link is, again, only really used due to the fact that the system wants to be as parsimonious as possible with its model. Drop-link is-- let's say you have a color, and you're like, OK, only red is acceptable. And then you see a blue one. It's OK. And let's say red, blue, and yellow are your only colors. Only red and blue are acceptable. Hey, wait a minute. No, we can just say that only yellow is not acceptable, or something like that, all right? But then you see yellow as a positive example. Actually, you can't even switch to the yellow being not acceptable, because you've only seen positive of red and blue. So you say, only red and blue are acceptable. Question? AUDIENCE: You mentioned that if we had seen incandescent and fluorescent, it's OK that we could climb the tree to light source, because it's more parsimonious. If we then see one that's sodium vapor, say a street lamp, and we reject it, do we have to go back down the tree, or do we just add a forbid-link? PROFESSOR: Then you would add a forbid-link. The question is, we climb incandescent, fluorescent, to light source. When we see sodium vapor, what do we do? The answer is, our system, unlike with lattice learning, in arch learning, our system is memoryless of its previous examples. It's incapable of going back down the tree. And so if you see incandescent and fluorescent, climb up to light source, you're only recourse, if you see that sodium vapor doesn't work, is to forbid-link sodium vapor. That's a good question. It shows good understanding of how we work. So drop-link, as you see, red and blue, they both are OK. You don't know about yellow, so you can't just switch that to a forbid-link to yellow. But then you see yellow is also OK. What do you do? You can drop the link altogether, save yourself space. If all the colors that you have in the entire world are fine, you can just not have part of your model be the color, and just say, any color is fine. Color must not be really part of what defines a lamp. And I'd say it's not. I mean, it's what defines a tacky lamp, I guess. But it's not what defines a lamp. So that might be why color isn't here to begin with. So to make yourself do better on this, more knowledge equals less search. Let's have more knowledge. There's only some of these are used to generalize, to make our model more accepting of new examples. And those are ones we would use after we've seen a positive hit, sort of generalize, learn more things in the model. Some of these are used to specialize and make our model refined, make our model more specific. You'd only use those after you saw a near miss. So let's actually separate those. That way you guys will never make the mistake of using one of these in the wrong situation. And more knowledge, less search. And less search means a faster quiz time. So require-link, what do you guys think? Is that a specializer or a generalizer? AUDIENCE: Specializer. PROFESSOR: Specializer, that's right. We'd only use it when we saw a near miss. Forbid-link, specializer or generalizer? AUDIENCE: Specializer. PROFESSOR: Specializer. We'd only forbid something if we saw a miss. Climb-tree, specializer or generalizer? AUDIENCE: Generalizer PROFESSOR: Generalizer, that's right. We'd only climb up to a more generic thing in the tree if we saw a positive example. Extend-set? AUDIENCE: Generalizer. PROFESSOR: Generalizer. We'd only extend the things in our set if we saw a positive example. Close-interval? AUDIENCE: Generalizer. PROFESSOR: That's a generalizer. We'd only make the interval bigger if we saw a positive example. Someone asks, what do you do if you have like 10 to 30, and then you find a negative example in 20? It's generally pretty annoying for this system to have to deal with that. There's a variety of things you could do based on your implementation. I've never seen us ask it in a quiz. But one thing you could do is forbid 20, exactly just 20, and just have a little hole at exactly 20. So then drop-link-- specializer or generalizer? AUDIENCE: Generalizer. PROFESSOR: Generalizer. This one's an easy one to mess up if you don't understand the system. Because dropping, you think, oh, getting rid of something. That's specializing. But actually no, you're generalizing the entire area saying, we can forget about it. Because they're all good. All right, so given that, we are set to do this problem. Hey, we have a pretty reasonable amount of time considering that we don't have to do our sums. It's great. So our first example is sort of one of those stand reading lights that you sort of have it on a stand, and you can adjust the little light bulb up or down. It is an incandescent bulb. It has a height of 11 inches. And it's got a flat base. It's electric. It's got a shade. And it's a hit. It's a hit, a positive example, a plus. It's good. So right away, we know that we can only use the S's or the G's? The G's. We can only use the G's. Because it's a hit. So we're not going to be requiring. We're not to be forbidding today. We're going to be happy and using a generalizer. So I'll help do the first one. You guys will help me do the next one. Or someone can say, we didn't cover this, and I'll do the first one, and we'll stop. So we've got an incandescent light. That's the same as our model. We've got a height of 11 inches. Oh, our model only covers 24. Flat base-- that's the same as our model. Electricity-- same as our model. Shade-- same as our model. So I say the only difference is the height. I say we need to close the interval, close-interval. And I say that our model is the same as before, except for that height is an element of 11 to 24. Simple enough-- I picked the easy one for myself. Next one. So this one is going to be-- the picture here didn't turn out. But it's basically a lamp that doesn't have a shade. It just sort of has the light shining on you. So it is a positive example. It's incandescent, height equals 11.5, flat base, and electricity. All right, so first of all, which kind would we use, specializing or generalizing? AUDIENCE: Generalizing. PROFESSOR: Generalizing, and particularly, anyone want to take a look at our model, which is the starting model, except for the height can go 11 to 24? Which heuristic do we need to use here? AUDIENCE: Drop-link. PROFESSOR: All right, people are all saying drop-link. Yes, that's right. Someone said extend-set. Well, I suppose we could extend the set to shade or not shade. But that's everything. So drop-link is the answer. We can drop the shade. Shade or not, this is still pretty much a lamp. So that's correct. We drop-link. So the model-- all right, there's been enough changes to it that I think I'll write it out again. The model is incandescent, height equals an element from 11 to 24, and flat base, and electricity. Good, so question? AUDIENCE: So during this all, we're editing our model based on the order that the model is in, right? PROFESSOR: So the question is, we're editing the model based on the order of the examples? AUDIENCE: What happens if there's the same thing except with no shade or something like that afterwards? PROFESSOR: So the question is, what happens if you have the same thing except for that there's one that's the same with no shades and is a near miss or something so that you need to-- or I'm sorry, you're saying, what if there's one that has a shade now that's a near miss? And so it would probably try to forbid-link shades. That's sort of the question? You're right. That's an inconsistency in the data. The system is very fragile. It's very fragile to ordering in particular. And you'd be surprised how awesome this does considering that it was made in, like, the '60s. It's pretty impressive stuff. But it's old. We're not saying that this is a way that you should do all of your learning nowadays. Because it has some serious issues. It's old, but it's pretty damn good for what it did. A newer style of this kind of learning made by me and another friend who's now working at Google is lattice learning, which has its own share of issues, one of which is that since it's trying to act like a little kid, at first it tries to claim that everything is OK until you show it some good negative examples. But very particularly one way gets around this problem is it's not memoryless. It in fact stores all of the examples it's ever seen and compares and contrasts them to what it sees in the new example. And this allows it to say something like-- let's say you're trying to teach it what can fly. In lattice learning, you say, all right, well, a blue jay can fly. And then if you ask it, can a cow fly, sure, a cow jumped over the moon. Everything can fly. So that's its problem. But if you say, actually, that was a nursery rhyme, a cow cannot in fact fly, it will then say, only birds can fly. Because it remembers that the blue jay can fly. And then you can eventually say, well, actually bats can fly, a fruit bat can fly. Well, OK, so chiropterans can fly and birds can fly. But that's it. And you can give it some other examples. But there are other styles of learning. You're absolutely right. In fact, if there's a multiple choice asking you about the strengths and weaknesses of arch learning, one of the weaknesses-- it's very vulnerable to ordering of what you teach it. But think about it. If our first example was one that's exactly the same as this but with a shade, and it was a miss, the system would just yell at you. Like, you're kind of being an asshole to it to give it two things that are inconsistent. You might say in the real world that happens. Well, arch learning, not great with real world messy data. It just goes ballistic. It's very OCD. It wants everything to match up. If it gets two things that are inconsistent, it'll just yell at you. It's like, you're wrong. This can't be true. Because you said it's OK with the shade. So that's a very good question. Another question? AUDIENCE: So if you have a sample that had two different things for the model, and it's a near miss-- PROFESSOR: It's not a near miss. The question is, if you have something with two different things in the model, and it's a near miss-- and I'm sorry to cut you off. Based on timing, I'm just going to say, whatever you said afterwards doesn't apply due to the fact at that point it's not a near miss. It's very important. It's only a near miss if there's only one change. Otherwise, it's just a miss. And if it's a miss with more than one change, you were probably going to say, how do you know what to change? The answer is, you can't change anything. Because any of them could have been what was wrong. That's why for a hit, oh, it can have as many differences as you want. You'll just generalize them all. For a miss, it can only have one thing different. In fact, there might be one here that's not a near miss, and you just say, oh, we don't do anything. We'll keep going. We might see it if we have time. So question? AUDIENCE: Because ordering is an important factor for this bigger thing, and you encounter an item which has two discrepancies in values, but if the ordering were different such that that would be accounted for, do you remember if you turn back to that [INAUDIBLE]? PROFESSOR: So the question is, let's say you have a non-near miss. Can you hold onto it and use it later when it would be a near miss? The answer is, nope. If we're going here, for a system that doesn't remember things, you can put in some kludges. That definitely makes it smarter. It uses all its data. But Patrick in the '60s making this up strove for elegance. And the elegant solution is, let's just be memoryless completely, the idea being, well, little babies can't tell you every experience they had with playing with blocks with an arch. So let's remember nothing. With lattice learning, our idea is that people sort of somewhere in the subconscious maybe do store all the examples, or at least a lot more than you give them credit for. So why not store them? But with arch learning, it's, little babies can't tell you, oh yeah, there's that one time I played with a lamp, and it didn't have a shade, and so it wasn't a lamp, or something like that. They're not going to be able to store it. They're not going to be able to save it for later. Sort of in the style of Turing, who believed a human would be a teacher to a computer, arch learning really focuses on the fact that the human is a kind and good teacher who offers examples that are exactly appropriate at the time. And that puts a lot of pressure on you as the trainer of an arch learning system. Good questions, all. Let's continue working this guy out. So the next example is basically another one of these stand lamps. It's got a light shining down, which is fluorescent. It looks like this. It's fluorescent. Height equals 13 inches, flat base, electric, shade. And yeah, as I put here, it's a hit. So as you guys have told me countless times, so I believe you that you remember, we're going to generalize for the hit. So what should we do here? So some people are saying extend-set. AUDIENCE: Climb-tree. PROFESSOR: The people who are saying climb-tree are correct. Why do we climb-tree instead of extend-set? Because of the fact that it's a tree, not a set. And it is more parsimonious to climb the tree than it is to treat it as a set and extend the set. Sets don't have a hierarchy. It's just like red comma blue comma yellow, or shade comma not shade. Whenever you have a tree, it's a fact that it's more parsimonious to just climb it than it is to treat it as a set and extend it, even though you could do that, I suppose. And arch learning, it's parsimonious. It's elegant. It's simple. And the elegant thing to do-- just climb up. So yes, we're going to climb up the tree. So our model is the same as last time except that this time it has a light source instead of saying, incandescent. It just climbed up to light source, sure. All right, so the next one is pretty cool. It's another one of those fluorescent shine down lamps. But in space, it has a tripod with three legs. And those legs all have wheels. So it is also a hit. So it is fluorescent, height equals 14 inches, wheeled legs, and electric. So obviously we're generalizing. What heuristic will we use this time? AUDIENCE: Climb-tree. PROFESSOR: Climb-tree again, that's right. We're going to climb from flat base to base support itself. Because wheeled legs and flat base, base support is an ancestor of both. So our model is the same as before. But at this point, it's changed enough that I might as well write it out. So we'll say that it is light source, height between 11 and 24 inches, base support, some sort of support, and electricity. Awesome, now, come on parity. Ah, no, back up, you. There we go. So that down there, don't get confused. That down there is our most recent model. We now have-- oh, oops, hehe. OK, I actually missed one of the examples. However, we were actually completely correct, which gives away quite a lot. We had a miss, incandescent bulb-- back in the old days when incandescent was the only kind of light we had. Height equals 60 inches, flat, back in the old days when flat was the only one we had, electric, back in the old days when we had electric. And we also have shade after we've dropped the link for shade. So we know that shade or not shade are both OK. So the question is, what do we do? First of all, do we specialize or generalize with this miss? AUDIENCE: Specialize. PROFESSOR: So obviously we're only to be able to require or forbid. So standardly, traditionally, the answer would be nothing. However, we accept it as sane the possibility that you might put a require onto the height. Some students did that, like say, oh gosh, that's a hard set, 11 to 24. But that's not what you would normally do. So just to show that we tried to understand that there might be multiple ideas in a weird situation, we did accept that that one time. But generally, there's not anything you have to do. This system at the time, and still now, said 11 to 24. This is spookily exactly the correct place where this example would go. But you don't have to do anything to the model. You don't have to use any heuristics. OK, so last step, -- the question is-- oh, question. AUDIENCE: So here, if we did it in this particular order, having it be the next [INAUDIBLE] in our example list, then it only differs by one from our existing model, right? PROFESSOR: No, it differs by a lot. It actually wouldn't be a near miss if we did it out. AUDIENCE: It would be a near miss if we did it out. PROFESSOR: No, incandescent is different than light source. Height is different than height. Flat is different than base support. So it would differ by a huge number, and it would not be a near miss. AUDIENCE: OK, then here's a question. Even though on the tree incandescent is a child branching off of light source, if we instead encountered-- because our current model says light source comma whatnot, if instead we got a negative thing that said, fluorescent, again, would that be considered not a near miss, because fluorescent differs from light source? PROFESSOR: So the question is, would it not be considered a near miss if you had, say, fluorescent rather than light source, because fluorescent differs from light source? So the answer there is-- AUDIENCE: Since it's memoryless, it doesn't know that fluorescent-- PROFESSOR: Well sure, it is memoryless. It doesn't know that fluorescent used to be a positive example before. Unfortunately, let's say we did have an example that was fluorescent, height is an element of 11 to 24, base support, electric. Actually, let's say we have one that was sodium vapor, height is 13. Sodium vapor, height is 13, base support-- for some reason it was called base support and electric. That's possible, right? We haven't seen a positive of sodium vapor. Let's say that sodium vapor was not allowed. If we don't say that that's a near miss and then forbid-link sodium vapor, how are we ever going to get rid of sodium vapor? Unfortunately, because of that, we lose some amount of expressiveness. Actually, what you're saying is very logical and makes sense, that you would want to have the ability to say, OK, this is a subset. This should be OK. But unfortunately, because of the fact that we climbed the tree even when we haven't seen a positive example for all things in the bottom on that tree, we actually lose that ability. It's a trade-off. It is most certainly a weakness of arch learning that you point out here. Lattice learning fixes that, but it has its own problems. So you're right, that you'd want to be able to do that. But you can't. Because it's memoryless. So last but not least, which model would we present in order to teach the system, right now from its final system, that a lamp requires a base support? So it would need to be require-link for the base support. So really fast-- incandescent bulb with a height of 6, a flat base, and electricity. Would that do it? Incandescent bulb with a height of 8 and a battery, incandescent bulb with a height of 12, wire support, and electricity-- oh yeah, by the way, this will have to be a miss to teach a require. Fluorescent bulb, height of 21, clamp base, and electricity, or incandescent bulb, height of 12, flat bottomed legs, and electricity-- so did anyone pick that out over the fastness of me saying it? Because [INAUDIBLE] is going to want to come in. Question or answer? AUDIENCE: [INAUDIBLE]. PROFESSOR: It's going to have a miss. AUDIENCE: If it's on the wire support? PROFESSOR: Yes. So I know it was fast. You can look through quiz 2007 final. Because [INAUDIBLE] is going to come in in a minute. But it would be a miss on the one that was correct on everything except for it was a wire support. Then the question you have is, wait a minute, it says incandescent bulb. That is now a subset of light source. You have his question again. The answer to that is they must have decided here that that was OK. But on the other hand, how would it learn sodium vapor if that wasn't decided to be OK? That's going to have to be an implementation detail. So one thing you can do is try to treat subsets as OK in the logical way. Because questions may do that. And then I guess you've lost the ability to forbid, say, sodium vapor. So you have to lose one or the other. I think you're definitely OK to ask TAs or someone who's proctoring if a situation like that comes up, which implementation detail is chosen. Because you could choose either when you're running arch learning. |
MIT_6034_Artificial_Intelligence_Fall_2010 | 3_Reasoning_Goal_Trees_and_RuleBased_Expert_Systems.txt | PROFESSOR PATRICK WINSTON: Ladies and gentlemen, the engineers drinking song. Back in the day, I've drunk quite a lot to that song. And as drinking songs go, it's not bad. I caution you, however, before playing this song in the presence of small children, audition it first. Some of the verses are sufficiently gross as to make a sailor go beyond blushing. It's an interesting song because there are an infinite number of verses. Here's the mathematical proof. Suppose there were a finite number of verses. Then there would be a last verse. And if there were a last verse, then some drunk alumni would compose a new one. Therefore, there is no last verse, the size is not finite, and there are an infinite number of verses. I play it for you today because I'm an engineer. I like to build stuff. I build stuff out of wood. I build stuff out of metal. I build stuff out of rocks. And I especially like to write programs. I don't know, sometimes people come to me and say, I'm majoring in computer science, but I don't like to write programs. I've always been mystified by that. I mean, if you want to show how tough you are, you can go bungee jumping or drive a nail through your hand or something like that instead. But I've written quite a few programs for demonstrating stuff in this subject. They're all written in Java, principally because I can therefore make them available to you and to the rest of the world by way of Web Start. A few weeks ago, I was mucking around with the system and broke the version on the server, and within 15 minutes, I got an email from somebody in the depths of Anatolia complaining about it and asking me to bring it back up. This particular program is patterned after an early AI classic. And it was the business end of a program written by Terry Winograd, who became, and is, a professor of computer science at Stanford University-- which is on the west coast for those of you-- on the strength of his work on the natural language front end of this program. But the natural language part is not what makes it of interest for us today. It's more the other kinds of stuff. Let's pile these things up. Now, I'm going to ask to do something, maybe put B2 on top of B7. Not bad. How about B6 on B3? This program's kind of clever. Let me do one more. Let's put B7 on B2. OK, now let's see. Maybe B5 on B2? B4 on B3 first, maybe? Oh, I must have clicked the wrong button. Oh, there it goes. OK. Let's put B4 on B1. Agh, my mouse keeps getting out of control. Now, let's put B1 on B2. This is an example I'm actually going to work out on the board. Oh, I see. My touch pad accidentally got activated. B1 on B2. Now, let's ask a question. OK. Well. SPEAKER 2: [SINGING] PROFESSOR PATRICK WINSTON: Stop. SPEAKER 3: [LAUGHTER] PROFESSOR PATRICK WINSTON: Had enough of that. Let's see. Why did you put-- why did you want to get rid of B4? OK, one-- SPEAKER 2: [SINGING] PROFESSOR PATRICK WINSTON: Maybe they think, that's what happens when you use software you write yourself. Why did you want to clear the top of B2? Did I do that? Why did you clear the top of B1? OK. SPEAKER 2: [SINGING] SPEAKER 3: [LAUGHTER] PROFESSOR PATRICK WINSTON: Oh, it's haunting me. Yeah. So the drinking song is easily offended, I guess. But I won't develop that scenario again. What I want to show you is that this program looks like it's kind of smart, and it somehow can answer questions about its own behavior. Have you ever written a program that's answered questions about its own behavior? Probably not. Would you like to learn how to do that? OK. So by the end of the hour, you'll be able to write this program and many more like it that know how to answer questions about their own behavior. There have been tens of thousands of such programs written, but only by people who know the stuff I'm going to tell you about right now, OK? So what I want to do is I want to start by taking this program apart on the board and talking to you about the modules, the subroutines that it contains. So here it is. The first thing we need to think about, here are some blocks. What has to happen if I'm going to put the bottom block on the larger block? Well, first of all, I have to find space for it. Then I have to grasp the lower block. And I have to move it and I have to ungrasp it. So those are four things I need to do in order to achieve what I want to do. So therefore, I know that the put-on method has four pieces. It has to find space on the target block. It has to grasp the block that it's been commanded to move. Then it has to move, and then it has to ungrasp. But taking hints from some of the questions that it did answer before I got haunted by the music, taking our cue from that, we know that in order to grasp something, in this particular world, you can't have anything on top of it. So grasp, therefore, may call clear top in order to get stuff off from the target object. And that may happen in an iterative loop because there may be several things on top. And how do you get rid of stuff? Well, by calling get rid of. And that may go around a loop several times. And then, the way you get rid of stuff is by calling put-on. So that gives us recursion, and it's from the recursion that we get a lot of the apparent complexity of the program's behavior when it solves a problem. Now, in order to find space, you also have to call get rid of. So that's where I meant to put this other iterative loop, not down here. Cleat top has got the iterative loop inside of it. So that's the structure of the program. It's extremely simple. And you might say to me, well, how can you get such apparently complex-looking behavior out of such a simple program? A legitimate question. But before we tackle that one head on, I'd like to do a simulation of this program with a very simple blocks problem. And it's the one I almost showed you, but it goes like this. Here's B1. We'll call this BX because I forgot its name. Here's BY. And here's B2. And the task is to put B1 on B2. And according to our system diagram, that results in four calls to subroutines. We have to find space. We have to grasp B1. We have to move, and then we ungrasp. Now, the way we grasp something is the first thing we have to do is clear off its top. So grasp calls clear top. And clear top in turn calls get rid of. And let me see. Let me keep track of these. This is clearing the top of B1, and this is getting rid of BX. And the way we get rid of BX is by putting BX on the table. And then that in turn causes calls to another find space, another grasp, another move, and another ungrasp. So that's a little trace of the program as it works on this simple problem. So how does it go about answering the questions that I demonstrated to you a moment ago? Let's do that by using this trace. So how, for example, does it answer the question, why did you get rid of BX? [INAUDIBLE], what do you think? How can it answer that question? SPEAKER 4: [INAUDIBLE] PROFESSOR PATRICK WINSTON: So it goes up one level and reports what it sees. So it says, and said in the demonstration, I got rid of BX because I was trying to clear the top of B1. So if I were to say why did you clear the top of B1, it would say because I was trying to grasp it. If I were to say, why did you grasp B1, it would say because I was putting B1 on B2. If I say, why did you put B1 on B2, it would say, slavishly, because you told me to. OK, so that's how it deals with why questions. How about how questions? Timothy, what do you think about that one? How would it go about answering a question about how you did something? Do you have a thought? TIMOTHY: Um, yeah, it would think about what I was trying to accomplish. PROFESSOR PATRICK WINSTON: Yeah, but how would it do that? How would the program do that? We know that answering a why question makes it go up one level. How does it answer a how question? Sebastian? SEBASTIAN: It goes down one level. PROFESSOR PATRICK WINSTON: You go down one level. So you start off all the way up here with a put-on. You will say, oh, well I did these four things. You say, why did you grasp B1? It will say because I was trying to clear its top. Why did you clear its top? Because I was trying to get rid of it. Why were you trying to get rid of it? Because I was trying to put it on the table. So that's how it answers how questions, by going down in this tree and this trace of the program of action so as to see how things are put together. What are these things that are being put together? What's the word I've been avoiding so as to bring this to a crescendo? What are these objectives, these things it wants to do? They're goals. So this thing is leaving a trace, which is a goal tree. Does that sound familiar? Three days ago, we talked about goal trees in connection with integration. So this thing is building a goal tree, also known as an and-or tree. So this must be an and tree. And if this is an and tree, are there any and nodes? Sure, there's one right there. So do you think then that you can answer questions about your own behavior as long as you build an and-or tree? Sure. Does this mean that the integration program could answer questions about its own behavior? Sure. Because they both build goal trees, and wherever you got a goal tree, you can answer certain kinds of questions about your own behavior. So let me see if in fact it really does build itself a goal tree as it solves problems. So this time, we'll put B6 on B3 this time. But watch it develop its goal tree. So in contrast to the simple example I was working on the board, this gets to be a pretty complicated goal tree. But I could still answers questions about behavior. For example, I could say, why did you put B6 on B3? Because you told me to. All right, so the complexity of the behavior is largely a consequence not of the complexity of the program in this particular case, but the building of this giant goal tree as a consequence of the complexity of the problem. This brings us to one of our previous matters-- early on to one of the gold star ideas of today. And this gold star idea goes back to a lecture given in the late '60s by Herb Simon, who was the first Nobel Laureate in the pseudo Nobel Prize for economics. Is that right, Bob? Was he the first? All right, he was the first winner of the Nobel Prize, pseudo Nobel Prize in economics. And in this lecture, which was titled "The Sciences of the Artificial," he said imagine that you're looking on a beach at the path of a ant. And he said, well, you know, the path of the ant looks extremely complicated. And you're tempted to think the ant is some kind of genius or monster brain ant. But in fact, when you take a closer look, what you discover is that there are a bunch of pebbles on the beach, and all the ant is doing is avoiding those pebbles on his way home. So the complexity of the behavior, said Simon, is a consequence of the complexity of the environment, not the complexity of the program. So that's the metaphoric Simon's ant. And what it says is that the complexity of the behavior is the max of the complexity of the program and the complexity of the environment. So that's something we'll see many times during the rest of the semester. Complex behavior, simple program. You think it's going to be complicated. It turns out to be simple because the problem has the complexity, not the program. So that brings us to check box three in today's talk, and there's a little bit of a scene here because now I want to stop talking about goal-centered programming and start talking about rule-based expert systems. The rule-based expert systems were developed in a burst of enthusiasm about the prospects for commercial applications of artificial intelligence in the mid-1980s. At that time, it was supposed lengthy articles are written, but you could account for useful aspects of human intelligence by writing all the knowledge in the form of simple rules. So if this is true, then that's true. If you want to achieve this, then do that. But all the knowledge had to be encapsulated in the form of simple rules. So what might you want to do with this? All sorts of things. Thousands of these systems were written, as I indicated before. But here's an example. I'm going to work out an example having to do with identification. And this example is patterned off of a classic program, strangely also written at Stanford, called MYCIN. It was developed to diagnose bacterial infections in the blood. So you come in. You got some horrible disease, and the doctor gets curious about what antibiotic would be perfect for your disease. He starts asking a lot of questions. So I'm not going to deal with that because that world has all kinds of unpronounceable terms like bacterioides and anaerobic and stuff like that. So it's completely analogous to talk about identifying animals in a little zoo, sort of a small town type of zoo. So I'm going to suggest that we write down on a piece of paper all the things we can observe about an animal, and then we'll try to figure out what the animal is. So I don't know, what can we start with? Has hair. Then there are some characteristics of the following form. Has claws. Sharp teeth. And forward-pointing eyes. And these are all characteristics of carnivores. We happen to have forward-pointing eyes too, but that's more because we used to swing around the trees a lot, and we needed the stereo vision. And we don't have the claws and the sharp teeth that go with it. But anyhow, those are typically characteristics of carnivores, as is eating meat. And this particular little animal we're looking at has also got spots, and it's very fast. What is it? Well, everybody says it's a cheetah. Let's see how our program would figure that out. Well, program might say, let's see if it has hair, then rule one says that that means it must be a mammal. We can imagine another rule that says if you have sharp claws, sharp teeth, and forward-pointing eyes, then you're a carnivore. And I'm using sort of hardware notation here. That's an and gate, right? So that means we have to have all of these characteristics before we will conclude that the animal is a carnivore. Now, this animal has been also observed to eat meat. So that means we've got extra evidence that the animal is carnivorous. And now, because the animal is a mammal and a carnivore and has spots, and it's very fast, then the animal is a cheetah. And I hope all of our African students agree that it must be a cheetah. It's a small zoo-- I mean, a big zoo. Who knows what it is? It's probably got some unpronouncable name-- there's possibilities. But for our small zoo, that will do. So we have group now written down in the form of these and gates. Several rules, R1, R2-- and there needs to be an and gate here-- that's R3 and an R4. All of which indicate that this animal is a cheetah. So we built ourself a little rule-based expert system that can recognize exactly one animal, but you could imagine filling out this system with other rules so that you could recognize giraffes and penguins and all the other sorts of things you find in a small zoo. So when you have a system like this that works as I've indicated, then what we're going to call that, we're going to give that a special name, and we're going to call that a forward-chaining rule-based-- because it uses rules-- expert system. And we're going to put expert in parentheses because when these things were developed, for marketing reasons, they called them expert systems instead of novice systems. But are they really experts in a human sense? Not really, because they have these knee-jerk rules. They're not equipped with anything you might want to call common sense. They don't have an ability to deal with previous cases, like we do when we go to medical school. So they really ought to be called rule-based novice systems because they reason like novices on the basis of rules. But the tradition is to call them rule-based expert systems. And this one works forward from the facts we give it to the conclusion off on the right. That's why it's a forward-chaining system. Can this system answer questions about its own behavior? [INAUDIBLE], what do you think? SPEAKER 5: [INAUDIBLE]. PROFESSOR PATRICK WINSTON: Why? SPEAKER 5: [INAUDIBLE]. PROFESSOR PATRICK WINSTON: Because it looks like a goal tree. Right. This is, in fact, building a goal tree because each of these rules that require several things to be true is creating an and node. And each of these situations here where you have multiple reasons for believing that the thing is a carnivore, that's creating an or node. And we already know that you can answer questions about your own behavior if you leave behind a trace of a goal tree. So look at this. If I say to it, why were you interested in the animal's claws? Because I was trying to see if it was a carnivore. How did you know that the animal is a mammal? Because it has hair. Why did you think it was a cheetah? Because it's a mammal, a carnivore, has spots, and very fast. So by working forward and backward in this goal tree, this too can answer questions about its own behavior. So now you know how, going forward, you can write programs that answer questions about their behavior. Either you write the subroutines so that each one is wrapped around a goal, so you've got goal-centered programming, or you build a so-called expert system using rules, in which case it's easy to make it leave behind a trace of a goal tree, which makes it possible to answer questions about its own behavior, just as this [INAUDIBLE] program did. But now, a little more vocabulary. I'm going to save time by erasing all of these things that I previously drew by way of connections. And I'm going to approach this zoo in a little different way. I'm going to not ask any questions about the animal. Instead, I'm going to say, mommy, is this thing I'm looking at a cheetah? And how would mommy go about figuring it out. In her head, she would say, well, I don't know. If it's going to be a cheetah, then it must be the case that it's a carnivore, and it must be the case that it has spots. And it must be the case that it's very fast. So so far, what we've established is that if it's going to be a cheetah, it has to have the four characteristics that mommy finds behind this rule are four. So instead of working forward from facts, what I'm going to do is work backward from a hypothesis. So here the hypothesis is this thing is a cheetah. How do I go about showing whether that's true or not? Well, I haven't done anything so far because all I know is a cheetah if all these things are true, but are they true? Well, to find out if it's a mammal, I can use rule one. And if I know or can determine that the animal has hair, then that part of it is taken care of. And I can similarly work my way back through carnivore. I say, well, it's a carnivore if it has claws, sharp teeth, and forward-pointing eyes. And then as much as the animal in question does, then I'm through. I know it's a carnivore. I don't have to go through and show that it's a carnivore another way. So I never actually ask questions about whether it eats meat. Finally, the final two conditions are met by just an inspection of the animal. That's to say, it's in the database. I don't have to use any rules to determine that the animal has spots and is very fast. So now, I've got everything in place to say that it's a cheetah, because it's a carnivore, because it has claws, sharp teeth, and forward-pointing eyes, and all the rest of this stuff is similarly determined by going backwards, backwards from the hypothesis toward the facts, instead of from the facts forward to the conclusions. So building a system that works like that, I have a backward-chaining rule-based expert system. But there's a very important characteristic of this system in both backward and forward mode, and that is that this thing is a deduction system. That's because it's working with facts to produce new facts. When you have a deduction system, you can never take anything away. But these rule-based systems are also used in another mode, where it's possible to take something away. See, in fact world, in deduction world, you're talking about proving things. And once you prove something is true, it can't be false. If it is, you've got a contradiction in your system. But if you think of this as a programming language, if you think of using rules as a programming language, then you can think of arranging it so these rules add or subtract from the database. Let me show you an example of a couple of systems. First of all, since I've talked about the MYCIN system, let me show you an example of a MYCIN dialogue. That's a MYCIN dialogue. And you can see the appearance of words you have to go to medical school to learn. And here's a typical MYCIN rule, just like the rules for doing zoo analysis, only a more complicated domain. But here's another example of a system that was written, not in the '80s, but just a couple of years ago by a student in the architecture department, Ph.D. thesis. He was interested in the architecture of a Portuguese architect named Siza. And Siza's done a lot of mass housing stuff. And Siza has the idea that you ought to be able to design your own house. And so Jose Duarte, a Portuguese student, a Ph.D. student in architecture, wrote a rule-based system that was capable of designing Siza-like houses in response to the requirements and recommendations and desires of the people who are going to occupy the houses. So the most compelling part of this thing, of this exercise, was that Duarte took some of the designs of the program, mixed them up with some of the designs of Siza, and put them in front of Siza, and said, which ones did you do? And Siza couldn't tell. So somehow, the rule-based system that was built using this kind of technology was sufficient to confuse even the expert that they were patterned after. But this program is a little complicated. It, too, has its own specialized lingo. So I'm not going to talk about it in detail, but rather talk instead about an analogous problem. And that is a problem that everyone has faced at one point or another, and that is the problem of putting groceries in a bag at a grocery store. It's the same thing, right? Instead of putting rooms in a house, you're putting groceries in a bag. And there must be some rules about how to do that. In fact, maybe some of you have been professional grocery store baggers? [INAUDIBLE] a grocery store professional bagger. You're a-- which one? LISA: [INAUDIBLE] PROFESSOR PATRICK WINSTON: Yeah, what is your name? LISA: Lisa. PROFESSOR PATRICK WINSTON: Lisa. OK, well we got two professional grocery store baggers. And I'm going to be now simulating a highly paid knowledge engineer desirous of building a program that knows how to bag groceries. So I'm going to visit your site, Market Basket, and I'm going to ask Lisa, now fearful of losing her job, if she would tell me about how she bags groceries. So could you suggest a rule? LISA: Sure. Large items in the bottom. PROFESSOR PATRICK WINSTON: Large items in the bottom. You see, that's why I'm a highly paid knowledge engineer, because I translate what she said into an if-then rule. So if large, then bottom. So now I-- SPEAKER 3: [LAUGHTER] PROFESSOR PATRICK WINSTON: So how about you, [INAUDIBLE]? Have you got a suggestion? About how to bag groceries? SPEAKER 6: The small things on top. PROFESSOR PATRICK WINSTON: If small, then on top. Lisa, have you got anything else you could tell me? LISA: Don't put too many heavy things in the same bag. PROFESSOR PATRICK WINSTON: Don't put too many heavy things in the same bag. So if heavy greater than three, then new bag, or something like that. Is that all we're going to be able to-- does anybody else want to volunteer? [INAUDIBLE], have you bagged groceries in Turkey? LISA: So they don't have grocery baggers, so we have to-- PROFESSOR PATRICK WINSTON: So everybody's a professional bagger in Turkey. Yeah. It's outsourced to the customers. SPEAKER 7: So no squishies on the bottom. So if you have-- PROFESSOR PATRICK WINSTON: No squishies on the bottom. SPEAKER 7: If you have tomatoes-- PROFESSOR PATRICK WINSTON: That's good. Tomatoes. SPEAKER 7: You don't want them to get squished. PROFESSOR PATRICK WINSTON: Now, there's a very different thing about squishies and tomatoes because tomato is specific, and squishy isn't. Now, one tendency of MIT students, of course, is that we all tend to generalize. I once knew a professor in the Sloan School who seemed real smart. And-- SPEAKER 3: [LAUGHTER] PROFESSOR PATRICK WINSTON: Then I figured out what he did. If I were to say, I'm thinking about a red apple. They'd sit back and say, oh, I see you're contemplating colored fruit today. They're just taking it up one level of abstraction. Not a genius. He also was able to talk for an hour after he drew a triangle on the board. Amazing people. Anyhow, where were we? Oh, yes, bagging groceries. So we're making some progress, but not as much as I would like. And so in order to really make progress on tasks like this, you have to exercise-- you know about some principles of knowledge engineering. So principle number one, which I've listed over here as part of a gold star idea, is deal with specific cases. So while you're at the site, if all you do is talk to the experts like Lisa and [INAUDIBLE], all you're going to get is vague generalities because they won't think of everything. So what you do is you say, well, let me watch you on the line. And then you'll see that they have to have some way of dealing with the milk. And then you'll see that they have to have some way of dealing with the potato chips. Nobody mentioned potato chips, except insofar as they might be squishy. We don't have a definition for squishy. Nobody talked about the macaroni. And no one talked about the motor oil. This is a convenience store. I don't want that in the same bag with the meat. And then no one talked about canned stuff. Here's a can of olives. So by looking at specific cases, you elicit from people knowledge they otherwise would not have thought to give you. That's knowledge engineering rule number one. And within a very few minutes, you'll have all three knowledge engineering rules and be prepared to be a highly paid knowledge engineer. Heuristic, let's call these heuristics. Heuristic number one, specific cases. Heuristic number two is ask questions about things that appear to be the same, but are actually handled differently. So there's some Birds Eye frozen peas, and here-- ugh, some fresh cut sweet peas. And to me, the person who's never touched a grocery bag in my life-- maybe I'm from Mars-- I can't tell the difference. They're both peas. But I observe that the experts are handling these objects differently. So I say, why did you handle those peas differently from those peas, and what do they say? One's canned, and one's frozen. So what happens? Bingo, I've got some new words in my vocabulary. And those new vocabulary words are going to give me power over the domain because I can now use those words in my rules. And I can write rules like if frozen, then put them all together in a little plastic bag. Actually, that's too complicated, but that's what we end up doing, right? Why do we put them all together in a little plastic bag? SPEAKER 8: [INAUDIBLE] PROFESSOR PATRICK WINSTON: What's that? SPEAKER 8: [INAUDIBLE] PROFESSOR PATRICK WINSTON: Well, there are two explanations. There's the MIT explanation. We know that temperature flow is equal to the fourth power of the temperature difference and the surface area and all that kind of stuff. We want to get them all together in a ball, sphere. The normal explanation is that they're going to melt anyway, so they might as well not get everything else wet. All right. SPEAKER 3: [LAUGHTER] PROFESSOR PATRICK WINSTON: So that's heuristic number two. And actually, there's heuristic number three, that I just want to relate to you for the first time because I have been dealing with it a lot over this past summer. Heuristic number three is you build a system, and you see when it cracks. And when it cracks is when you don't have one of the rules you need in order to execute-- in order to get the program to execute as you want it to execute. So if I were to write a grocery store bagging program and have it bag some groceries, again, eventually it would either make a mistake or come to a grinding halt, and bingo, I know that there's a missing rule. Isn't that what happens when you do a problem set, and you hit an impasse? You're performing an experiment on yourself, and you're discovering that you don't have the whole program. In fact, I've listed this as a gold star idea having to do with engineering yourself because all of these things that you can do for knowledge engineering are things you can do when you learn a new subject yourself. Because essentially, you're making yourself into an expert system when you're learning circuit theory or electromagnetism or something of that sort. You're saying to yourself, well, let's look at some specific cases. Well, what are the vocabulary items here that tell me why this problem is different from that problem? Oh, this is a cylinder instead of a sphere. Or you're working with a problem set, and you discover you can't work with the problem, and you need to get another chunk of knowledge that makes it possible for you to do it. So this sort of thing, which you think of primarily as a mechanism, heuristics for doing knowledge engineering, are also mechanisms for making yourself smarter. So that concludes what I want to talk with you about today. But the bottom line is, that if you build a rule-based expert system, it can answer questions about its own behavior. If you build a program that's centered on goals, it can answer questions about its own behavior. If you build an integration program, it can answer questions about its own behavior. And if you want to build one of these systems, and you need to extract knowledge from an expert, you need to approach it with these kinds of heuristics because the expert won't think what to tell you unless you elicit that information by specific cases, by asking questions about differences, and by ultimately doing some experiments to see where your program is correct. So that really concludes what I had to say, except I want to ask the question, is this all we need to know about human intelligence? Can these things be-- are these things really smart? And the traditional answer is no, they're not really smart because their intelligence is this sort of thin veneer. And when you try to get underneath it, as written, they tend to crack. For example, we talk about a rule, we could talk about a rule that knows that you should put the potato chips on the top of the bag. But a program that knows that would have no idea why you would want to put the potato chips on top of the bag. They wouldn't know that if you put them on the bottom of the bag, they'll get crushed. And it wouldn't know that if they get crushed, the customer will get angry, because people don't like to eat crushed potato chips. So that's what I mean when I say the knowledge of these things tends to be a veneer. So the MYCIN program, during debugging, once prescribed a barrel of penicillin to be administered to a patient for its disease. They don't know, they don't have any common sense. So the question then becomes, well, I don't know. Does rule-based-- do rules have anything to do with common sense? And I'm becoming a little bit agnostic on that subject. Because there are certain indications, there are certain situations, in which rules could be said to play a role in our ordinary understanding of things. Would you like to see a demonstration? What I'm going to show you, when the clip speaks up-- well, before I make any promises, let me see if I'm actually connected to the web. MIT, good. MIT. Guest. Yeah, that's me. Sounds good. OK, I just tested the system, and I've seen it is actually connected to the web. And I'm going to adjust some systems options here. I'll get rid of the text box. And we'll get rid of those changes scale a little bit. What I'm going to do is I'm going to read a little synopsis of the Macbeth plot. You're MIT students. I'm sure you're all classically educated and very familiar with Shakespearean plots. So I'm going to read one. I'm going to read a version of a Macbeth plot. And it's going to go along like this. It's basically reading a rule base so far. And pretty soon, it's going to get beyond the rule base and start reading the Macbeth story. And there it is. It's read the Macbeth story. Let me show you what the Macbeth story looks like as it's actually retained by the system. That's it. Read that. OK, you ran out of time because the machine's already finished. It takes about five seconds to read this story. Now, as you look at this little synopsis of Macbeth, there are a couple things to note. For one thing, it says that Duncan is murdered. Duncan, I hope this doesn't bother you. Duncan is murdered by Macbeth. But at no time does it say that Duncan is dead. But you know Duncan's dead because he was murdered. If murdered, then dead. SPEAKER 3: [LAUGHTER]. PROFESSOR PATRICK WINSTON: So if you look a little further down, what you see is that Macduff kills Macbeth. Fourth line up from the bottom. Why did Macduff kill Macbeth? Doesn't say why in this story, but you have no trouble figuring out that it's because he got angry. And when you get angry, you don't necessarily kill somebody, but it's possible. SPEAKER 3: [LAUGHTER]. PROFESSOR PATRICK WINSTON: So now that you see what's in the story, let me take you back to this display. It's what we call an elaboration graph. And when I blow it up, you can see that there's some familiar looking things in there. For example, up here in the left-hand corner, Macbeth murders Duncan, right over there. And over here, Macduff kills Macbeth. And if you look at what is a consequence of that, it looks like there must be a rule that says if you murder somebody, you harm them. And if you murder somebody, then they're dead. And one reason why you might kill somebody is because they angered you. And if you go the other way, one consequence of killing somebody is that you harm them, and that they die too. And if you harm somebody, they get angry, and their state goes negative. So that suggests that there are some things that we have on our hands that are very compiled and very, strangely enough, very rule-like in their character. Now, to close, I'm just going to read Hamlet. The Hamlet demonstration is much like the Macbeth one. In fact, Hamlet and Macbeth are very alike in their plot. But there's one thing that's well-illustrated by our particular capturing of Hamlet here. And that is that you'll note that the ratio of gray stuff to white stuff is considerable. The gray stuff is stuff that has been deduced by rules. And the reason there's so much gray stuff in this Hamlet story is because everybody's related to everybody else. So when you kill anybody, you irritate everybody else. So look at that. A few white things, those are the things that are explicit in the story, and lots of gray stuff. So what this is suggesting is that when we tell a story, it's mostly a matter of controlled hallucination. I know what rules are in your head, so I could take advantage of that in telling the story and not have to tell you anything I'm sure you're going to know. And so that's why, we've discovered, that storytelling is largely a matter of just controlling how you're going along, a kind of controlled hallucination. |
Child_and_Adolescent_Mental_Health | What_is_Behavioral_Parent_Training_Part_15.txt | [MUSIC PLAYING] ERICA H. LEE: So the gold standard intervention for disruptive behavior disorders is a parenting intervention called behavioral parent training, or BPT. So behavioral parent training really focuses first on strengthening the attachment between caregivers and kids. Because one thing that we know is that when children are disruptive, when they're having a hard time following rules, when they're tantrumming a lot, when parents are hearing a lot of the word "no" versus "yes," this can really fray on the parent-child relationship. So first in behavioral parent training, you want to start by resetting that relationship. I often say to parents, I know you love your kids, but what I'm hearing from you right now is you really don't like them a lot of the time. So how can we create more opportunities for you to just enjoy each other and feel really good about the attachment that you have that's underneath all of the stress and frustration of these disruptive behaviors? And then moving from there, we work with parents to teach them more strategies to reinforce children's positive behaviors. So one thing that happens when kids are disruptive is parents get pulled into this pattern of often being reactive versus proactive. So once their child is acting up, of course they're responding, and they're having to give a lot of attention to that. So that can actually flip the dynamic, where now we're mostly paying attention to negative behavior. And so we're missing really amazing and valuable opportunities to reinforce and positively praise kids who are doing things when they're supposed to be doing them. So we teach parents to motivate kids towards getting attention for positive behaviors versus negative behaviors. And then a third piece is teaching parents, well, what do I do though when my child is acting up, when they break the rules, when they're being unsafe, when they break a house rule that they know is a rule? So then teaching parents strategies to reduce and manage misbehavior. So it's this combination approach that really helps kids and parents reset their relationship to each other and helps kids now be much more motivated to behave the ways that their parents and sometimes their teachers would like, and to be much less motivated to engage in behavior that is disruptive to the family. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | How_Can_Caregivers_Support_Childrens_Mental_Health_Treatment_Part_12.txt | [MUSIC PLAYING] ERICA H. LEE: So caregivers play a really important role when their kids are in therapy. So of course, not only just logistically making sure that kids can get to therapy, that they're connected with a therapist that they feel really positively about. But of course, parents play an enormous role in children and teens' lives. They're their primary caregivers. They're also seeing them all the time. And they're often there when kids are having a hard time. So if a child is feeling depressed or anxious, parents are often the first ones to see it. But they're seeing it on a regular basis. This means that they have a tremendous amount of expertise to not only share with the therapists that their child is seeing, but it means they have really valuable opportunities to participate in their child's treatment. So for example, talking to your child's therapist about the types of skills that they're learning in therapy. They're there to learn all these skills and strategies for, what should I do if my mood is low or my worry feels overwhelming? These are then strategies that we want kids practicing outside of therapy. I might see a child or teen once a week, which means that there's enormous number of hours outside of our therapy session, where they might be having a hard time or they could be practicing and strengthening these skills. So parents have this valuable opportunity to say, hey, I notice you're having a hard time. Let's make sure you're practicing skills in between your weekly therapy sessions. Let's make sure that I'm supporting you and I understand what the strategy is. So if I notice that there's a chance where you might not be realizing, hey, I'm feeling kind of low or worried and I could be using a strategy right now, maybe parents can come in and support you in practicing that. So parents play a really, really crucial role in supporting their children's gains, both in the therapy session and in between sessions. So in psychotherapy, of course, there is confidentiality that comes into play. So what we often explain to children and teens is that what we talk about in our therapy session stays between us. But the exceptions always have to do with safety. So if you told me, for example, that you were harming yourself, that someone was harming you, or that you wanted to harm someone else, we would absolutely want to make sure that caregivers were involved. And maybe even some other important adults to make sure that everyone is safe. But the reason this is also important is because it provides this really safe and sacred space for children and teens, sometimes, to talk about things that feel too hard to talk about with anyone else. This can sometimes be difficult for parents. They're sort of, what's happening in the therapy session? How do I know how to be involved? I don't to ask too many questions, but I'm not quite sure how to participate. So with younger children, I often will say to parents, it's helpful to meet every say a few weeks just to check in, see how things are going, get a parent's perspective on whether or not they think therapy has been effective. And of course, to review any skills or tips that I'm teaching the child to make sure that parents can reinforce those outside of the therapy room. For teens, it's more of a collaborative discussion. What kind of makes sense for your teen? Of course, there's a difference between our younger teens, our 13 and 14-year-olds, and then those that are moving quickly towards young adulthood, our 17 and 18-year-olds. What role do you want your caregivers to play in your treatment, if any? What kind of levels of responsibility do you want to take for sort of managing your appointments, and kind of showing up, and telling me what's going on? And are there ways that parents can be supportive to you that feels both affirming, but also respects your understandable need and growing independence as you're moving towards adulthood? One thing that I think can get underestimated in relation to children and teens mental health are the basics. So what are our children-- a child, or teen's basic routines? Are they getting enough sleep? Are they following a mostly healthy-ish diet? Are they getting regular exercise? Just these three areas of a child or teen's life can go a really long way towards promoting good mental and emotional health. Routines also provide a really important sense of stability and predictability, especially when kids are going through a hard time. So thinking about the last few years, where there's been so much upheaval during the global pandemic, having some sense of, OK, every day, this is kind of the schedule I follow or I've been trying to go to bed at these regular times. Or every Friday, I know that I get to do this fun thing with my friends or my family. That can feel enormously reassuring when we're overwhelmed by all of the uncertainty and unknowns about the future. At the same time, parents can also continue to nurture really strong parent-child relationships with their kids. Especially for our teens, I often hear from parents, well, they kind of want to be more independent. They care much more what their friends think than what I think at this point. And yes, that might be true and that's very developmentally common. And primary caregivers still play an enormous role in children and teens' lives. So nurturing really strong, supportive and healthy relationships with important adults can also be very supportive when kids and teens are having a hard time. Another thing that can be helpful is just normalizing conversations around mental health. So of course, I recognize that this varies depending on family background, cultural background, and norms. But letting kids know, it's normal and understandable that, sometimes, you feel sad, or sometimes you feel lonely, or it's hard to be left out. Or you're worrying about, what's going to happen with college? How am I going to do in soccer today? Do other kids like me? But even just opening up space at home to say, hey, I get it. These are things that happen. I felt this way too when I was a kid. That can go a long way towards helping kids normalize conversation around mental well-being and also open up opportunities to have those discussions with your kids. |
Child_and_Adolescent_Mental_Health | How_Does_the_Mental_Health_of_a_Parent_Affect_Their_Children_Part_19.txt | [MUSIC PLAYING] ERICA H. LEE: We know that the resilience of kids or their ability to overcome adversity and hardship depends on the ability of caregivers to provide consistent safety and support. So that means that parent well-being is critical to the health of child and teen well-being. So during the COVID-19 pandemic, we know that parental well-being has decreased. There's been so much pressure on parents as kids normal daily routines have been completely upended. And we also know that for those families that are already under more stress. So say, for example, families that are experiencing socioeconomic stress, or they've experienced loss because of the pandemic, or they have kids with greater needs, whether that's social, emotional, or physical and medical needs, that these families are under tremendous stress. And one thing that is often overlooked is parental well-being, right? There's this pressure on parents to do more, and more, and more. And it's been really hard for them to manage that with the tremendous amount of stress that they've been experiencing, especially over the last few years. |
Child_and_Adolescent_Mental_Health | What_Types_of_Therapy_are_Effective_for_Youth_Anxiety_Part_11.txt | [MUSIC PLAYING] ERICA H. LEE: So the gold standard evidence-based type of psychotherapy for youth anxiety is called cognitive behavioral therapy or CBT. So CBT focuses on teaching children and teens to understand the connections between their thoughts, their feelings, and their behaviors. So what we know is that the way that we think about a situation or often the way that we interpret a situation has a big impact on the way that we feel in that situation. And then how we think and feel about a situation is a strong driver of how we behave in that situation. So let me give you an example for youth anxiety. Let's say that I'm a child who really worries about tests. We have lots of kids who worry about tests and doing well in school. So I'm really worrying like, oh my gosh, I have to take the science test tomorrow and I think I'm going to fail. I'm not going to do a good job, then I'm going to fail, and I'm going to get kicked out of school. This is just not going to work out for me. If I'm having this thought, I'm going to feel a lot of worry. Maybe I'm also going to feel kind of sad. I might even feel panicked or angry. Why should I even have to take this test? This is so unfair. No one cares about volcanoes. I'm never going to need to know about it in the future. So now, I'm having these negative thoughts, I'm feeling really negatively, and then what happens in terms of my behavior? So maybe one thing that we commonly see is that I procrastinate. So instead of studying for my science test, I keep putting it off and avoiding that feeling of anxiety. Or I can't sleep so now I'm staying up all night worrying about this test and how things are going to turn out. So then what happens? I get up in the morning, I'm really sleepy, I'm not in a good mood, I'm not in a good headspace. I go and I take the test. And I don't do as well as I could have because of my negative thoughts and my negative feelings. What does this do? It reinforces my original belief that I was going to fail and that I'm not good at science tests. So what we're doing with kids is really breaking this down. What are the automatic negative thoughts that you have in situations that scare you or induce worry? How does that connect to your feelings? How does that connect to your behavior? And then we teach kids two things that you can actually change are how you think about a situation and how you behave in that situation. And if we can change those two things or even just one of those things, you can change the other ones. So one thing that we often do in CBT for anxiety is something called exposure therapy. So this is the idea that when we're scared of a situation or an object, we often avoid it. And avoiding it actually just makes our anxiety grow because we never have the opportunity to see, well, maybe the thing I was scared of never is actually going to happen. Or even if it happened, I could cope with it and I could be OK. So in exposure, you make a plan with kids to gradually, safely, and slowly expose them to the thing that they're scared of. So we think of this as approach behavior versus avoidance behavior. At the same time, we're also teaching youth, how do I notice those automatic negative thought patterns? How do I kind of catch those negative thoughts as I'm having them, examine them, and then maybe reframe or change them to be more balanced and realistic for the situation, which is going to make my anxiety calm down. And then lastly, we also teach kids many relaxation strategies. So this idea that when we're anxious, we experience a lot of physiological tension. Kids often say, when I'm nervous, my heart beats really fast. I feel like I can't breathe. I feel kind of sweaty or shaky and then I can't think straight. So what can we do to help you feel more relaxed in your body, to then calm your mind, to feel more in control? Now, I can make really careful decisions about what I want to do next with a clear head. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | When_to_Refer_Patients_to_a_Mental_Health_Specialist_Part_4.txt | [MUSIC PLAYING] ERICA LEE: So the way that we think about a child's level of difficulty throughout the week gives us a really good sense of when we might need a specialist referral. So there are many kids who come in, and they're talking to their pediatrician, and they're reporting sort of low-level symptoms, so what we think of as subclinical or even mild symptoms of, say, depression or anxiety. But when those symptoms start to move into the moderate or severe range, we do want to consider, would it be helpful to get some expert advice on what the next steps would be? So one way to think about this is, what's going on in a child's life? What are the things that they want to be doing? What are the things they need to be doing? And are their symptoms getting in the way? So is there a level of functional impairment at the point where they're really not able to get through the day or it wasn't just a temporary thing? So sometimes kids feel sad or worried for a short period of time, but maybe you're noticing that this is going on a lot longer. Would we be able to get some expert help so we can see if we can figure out a good plan? [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | What_is_Parental_Burnout_Part_20.txt | [MUSIC PLAYING] ERICA LEE: So parental burnout refers to a sense of overwhelming mental, emotional, and physical exhaustion from the demands of being a parent. So it's often associated for parents with these feelings of anxiety, stress, sometimes resentment, and overwhelm just from managing the day-to-day responsibilities, which are enormous, of being a good parent. Parents also often report that they feel really guilty, so they feel this sense of deep shame around I'm failing as a parent or I'm not a good enough parent. We know that parental burnout is associated with worse outcomes for both parents and kids. So for parents, we know that their mental and physical health declines when they're experiencing burnout. We also know that it makes it harder to parent. So when they're experiencing this sense of overwhelm and exhaustion, it makes it harder for them to engage in appropriate discipline. We also know, for example, that parents who are burnt out are more likely to engage in child maltreatment. On the other side, what we know is that when parents are doing better-- so when their well-being is prioritized and they are less burnt out-- we know that children, then, also do better. So their emotional and physical well-being increases, and family functioning improves. On the other side, we know that improved parental well-being-- so less burnout-- is better for kids, better for parents, and also overall for families. So for example, we know that when parents' well-being is prioritized and they're doing better, we see lower rates in children of mental health difficulties. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | What_Behavioral_Therapies_Can_Help_Children_with_ADHD_Part_17.txt | [MUSIC PLAYING] ERICA H. LEE: In regards to behavioral interventions for ADHD, these often focus on behavioral parent training, or teaching parents to increase structure and predictability at home, strengthen the parent-child relationship, motivate kids towards more positive behaviors, and then reduce and manage any misbehavior that arises. You can also intervene behaviorally at schools. So behavioral classroom interventions really help teachers sort of respond, again, more positively to children's positive behaviors and reduce and manage negative behaviors. And there are also behavioral peer interventions that can be used in classrooms where you're teaching other peers to support a child that's having a hard time with ADHD symptoms, to act in a way that's more appropriate for the school setting, and then to get that really positive social reinforcement from other kids when they're able to do things that they need to do and to meet those goals. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | How_Can_Parents_Engage_in_SelfCare_to_Avoid_Burnout_Part_21.txt | [MUSIC PLAYING] SPEAKER: One thing I'll say is that it's really difficult for parents to engage in self-care. Parents put so much pressure on themselves to be really good parents. It's coming from such a loving and good place. But that also means that they can be highly self-critical. Parents are really good at saying, "Here are all the things I'm not doing here, all the ways that I'm failing." But one thing I always say to parents is that if you're not taking care of yourself, you're also not modeling for your kids what it means to engage in healthy coping and how to manage stress. All right. So we always use this analogy from airplanes, right. That you have to put on your oxygen mask before helping your child. Because if you don't have anything to give, where are your kids going to draw from? So I really try to help parents unpack the guilt that comes along with prioritizing themselves and not sacrificing everything for their kids, even though, of course, it's coming from the best intentions. So a few things that parents can do is, one, try to prioritize self-care. And it doesn't have to be huge. Ideally, it would be. It would be a bigger part of your week. But I also recognize parents have so much that's going on. There are so many pressures on them. There's not a time-- enough time in the week, or the day, or even an hour to take a moment. And are there little breaks that you could take in-between in the day? So for example, are there 5, 10, even 15 or 20 if you're lucky, minutes that you could just get to yourself to take some slow, deep breaths, to listen to some music, to connect with a friend or a family member that feels good to you? If you're lucky, even engage in a hobby. But even taking that 15 or 20 minutes, which probably feels very difficult, is going to pay much larger dividends in terms of your own well-being, and then, of course, in terms of your family well-being. Related to that, think about your community, right. So who are the people that support you? And where can you get extra support if you need it? So for many people, that's their family, sometimes it's friends, sometimes it's community or cultural organizations, or a certain type of religious faith. Where are the places where you get to refill your cup, especially when things feel so overwhelming and you just feel like you haven't got a lot more to give? How can you also rely on your community so you don't feel like you're doing it all? And for those of us that have the option, it can also help to think about who can share the load. So are there ways in which your children can engage in activities and with other adults where you can take a little bit of a break? Are there other people in your life, for example, who can help with some of the child rearing responsibilities? And even though it can be hard to ask for help, think of it as me asking for this help is really me asking for help for my family and for my kids if that helps decrease your guilt. Another thing is to go back to this idea of healthy coping skills. We work so much with kids. And I know many of the parents out there are great at saying to your kids, "What are you going to do when you're having a hard time? I want you to use a coping skill. If you're having a hard day, I want you to shift your mindset. I want you to engage in an activity that feels good for you." Parents are not so good at putting this lens on themselves. So saying, "I want my kids to engage in these healthy coping skills. I want to boost their resilience. I want them to feel good about going out and tackling the challenges that they experience in their daily lives." I need to model that too. I need to show kids I can come home from a long day at work. I can be exhausted. And I can take 5 minutes for myself to just enjoy this cup of tea and not be on my phone, not be on my email, and not be thinking about anything else if I don't have to. This helps your kids know too that it's OK to take those breaks for themselves when they're having a hard time. And then the last thing I'll say is that I always encourage parents to take a moment, even if you can, 30 seconds even every day, to pause and say, "What did I do right today? What did I do well? What are the steps that I took forward to show my family that I love them and I care for them?" Your minds are going all day long with all the things you're not doing. "I didn't get to this. The to-do list is too long. I forgot to do this thing for my kid. My kid is mad at me about this." Those things are true. And you have to balance them with what are all the things that you're doing right, and then give yourself credit for those. Because again, parental well-being is essential to child well-being. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | What_Types_of_Therapy_Are_Effective_for_Youth_Depression_Part_10.txt | [MUSIC PLAYING] ERICA H. LEE: So the gold standard evidence based-type of psychotherapy for youth depression is a type of therapy called Cognitive Behavioral Therapy or CBT. So CBT focuses on the connections between our thoughts, feelings, and behaviors. So this idea that the way we think about a situation strongly influences the way we feel about it. And then the way we feel about it has a strong influence on the way that we behave. So here's an example from youth depression. Say that I am a child or a teenager, and I have this repeated thought that no one likes me. Having that thought over and over is going to make me feel sad, lonely, maybe even angry. Why should I need to make people like me? And then because I'm having that thought, and I'm feeling that way, maybe I start to socially isolate. So my thought is like well, if no one likes me, and I'm feeling really lonely, and sad, and frustrated, maybe I'll just stay at home, and I won't text anybody or try to hang out. And then because I'm now isolated, this reinforces my original belief that no one likes me because I don't have opportunities to see that, actually, it's not so bad. Maybe not everyone likes me, but plenty of kids like me. So in CBT, we're really focusing on understanding these thought, feeling, and behavior patterns. And then teaching children and teens that you can change the way that you think, and you can also change the way you behave to change the way that you feel. So some of the strategies that we might treat children and teens from a CBT perspective, include things like behavioral activation. So this is a really well-tested and evidence-based intervention to get kids moving again. So when we're depressed, we tend to isolate, we tend to move less, because we're feeling tired, and sad, and we just want to be alone. But this actually means that kids and teens miss out on really valuable opportunities to feel connected, to get moving, to get outside, and to see that actually there are lots of opportunities to feel a little bit better, even if it's just coming in smaller spurts. We also teach kids and teens in CBT about identifying their automatic negative thought patterns. So when they're in a situation, are they more likely to accidentally either interpret something as being overly negative? Or more negative than it actually is? And then again, that flows into how they feel, and how they behave in that situation. Or are they actually just feeling blue a lot of the time, and so everything feels kind of blah! So can we identify the automatic negative thoughts that we're having? Can we reframe them? Can we try to approach situations in a more balanced, sometimes more realistic, and hopefully even a little bit more positive of a way? And then the third set of interventions that we often use in CBT have to do with relaxation. So how can we teach children and teens to notice? When I'm feeling depressed, when I'm feeling blue, I'm having a lot of physical tension in my body, so what can I do to increase my own physiological self-soothing? So that I feel more in control of how my body feels. And if my body is more relaxed, my mind is more relaxed, and that helps with mood. For our teens with more severe depression symptoms, there's another type of psychotherapy that has been found to be very helpful called Dialectical Behavior Therapy, commonly referred to as DBT. So DBT is actually a form of CBT that was specifically designed for our youth who experience really intense emotions. And then often from that those intense emotions, they experience more emotional or behavioral dysregulation, and then also risky or self-injurious behaviors or even suicidality. And DBT has been found to be very effective at helping teens understand their emotions better, respond to them in a healthier way, and cope with them. So in DBT, we're teaching children and teens how to more effectively move through their daily lives. So often when kids come to DBT, they're experiencing really intense emotions, and that's making things feel really, really hard. Whether that's going to school, being with their friends or family, or doing activities that they really enjoy. And as I mentioned before, these children, teens are often experiencing more severe symptoms of depression. So they may be engaging in self-injurious behavior, thinking about or even attempting suicide. So first, we start with a set of interventions around teaching them mindfulness. This idea of non judgmentally being in the present moment. So often when we're experiencing emotional distress, we're thinking about the future, we're thinking about the past. And usually we're experiencing a lot of negative thoughts about both of those things, which then of course makes us feel negatively. So how do we slow things down, and just live in the present moment? Moment to moment, so we're not so overwhelmed by things in the past that we cannot change, and things in the future that have not happened yet. We're also teaching kids a set of interventions around interpersonal effectiveness. So often kids who have more severe symptoms of depression really struggle with their relationships. Either in getting the support that they need from other people or feeling connected to-- to others in a way that feels really authentic to them, where they can be their authentic selves, and be accepted and loved for who they are. So teaching them when you're having, say, a difficult conversation with someone or even when you're approaching a friend, what are more effective ways to relate to them? To help meet your goals, and to build really healthy and satisfying relationships. There's also a third set of interventions that really focus on emotion regulation and distress tolerance. So this idea that, when I'm experiencing really difficult emotion, it's hard to tolerate them. Many people having this experience. It's hard to just sit with how I feel. It feels so difficult. Sometimes, I don't feel like I can live through it. So how do we help kids slow down in the moment when they're really upset or feeling really intense emotion? How they can get their bodies back on track? Root themselves in the present moment? And then work through those emotions, and find reasons to keep moving forward, and to-- and to stay alive. A third kind of psychotherapy that can be helpful for youth with depression is called Interpersonal Psychotherapy or IPT. So in IPT, you're really focusing on a person's relationships, and how a person's relationships connect to their mood. So thinking together, for example, with an adolescent around, what's happening with your relationships? Are they satisfying? Do they feel healthy? Do they feel reciprocal? Where are the gaps in your relationships with important people? And then, let's think together on a plan for strengthening those relationships, improving your communication with important people in your life, and then tracking. As you're improving your relationships, are you also seeing improvement in your mood, and your ability to go out and do the things that you need to do during the week? [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | What_are_the_Drivers_of_Mental_Health_Challenges_in_Young_People_Part_2.txt | [MUSIC PLAYING] ERICA H. LEE: So for children and teens, life has been tremendously stressful over the past three years. So driven by the effects of the global pandemic, we've seen children and teens' natural routines be completely thrown up in the air. So now we're not able to go to school the way we used to. We're not able to participate in our usual activities. We're not able to safely see our family and friends. So in addition to the social isolation and the change in our daily lives, we're also seeing that kids are tremendously stressed by any number of other major things that have been happening in the country. So we've been seeing big social movements related to racial injustice and LGBTQ+ movements. We've also been seeing a number of huge global events related to school shootings and violence, and things where kids are just feeling really uncertain. Is the world safe for me? What am I going to do to get through this? And are the people around me able to keep me safe? [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | Children_and_Adolescents_at_Risk_for_Depression_and_Anxiety_Part_7.txt | [MUSIC PLAYING] ERICA LEE: The children and teens who are at particular risk for developing anxiety or depression symptoms are those who are coming from underrepresented or minority groups. So these are the children and teens who are at higher risk for things like racism, discrimination, rejection, or victimization. So we're often thinking about our children and teens who come from a racial or ethnic minority background, our LGBTQ+ children, and then, of course, those who have any kind of preexisting condition. So maybe they have a learning disability or they've got a chronic medical condition. And unfortunately, these are our youth who are at higher risk for a number of mental health concerns, and they're often the ones who historically have had a harder time accessing high-quality care. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | How_Do_Children_with_ADHD_Present_Part_16.txt | [MUSIC PLAYING] ERICA H. LEE: So attention deficit/hyperactivity disorder, commonly referred to as ADHD, is a neurodevelopmental disorder where children or teens have difficulty with a number of areas of their lives. So it could be that they're having difficulty paying attention, difficulty sitting still or managing the energy in their bodies, and/or difficulty not behaving in an impulsive manner. So there are three types of ADHD, depending on the child or teen's primary symptoms. So the first type is what we call an inattentive presentation. So these are the kids who are disorganized, forgetful, easily distracted. It's hard for them to sit down and concentrate on one thing at a time. The second type is the hyperactive or impulsive presentation. So these are kids who you notice are just go, go, go. They've got a lot of energy in their body, so it's hard to sit still. Sometimes it's hard to slow down when they're talking and not interrupt other people. But you may also notice that they make a lot of impulsive decisions, so they act quickly without thinking. And then the third kind is something that we call combined presentation, which is exactly as it sounds. These are kids who experience symptoms of both inattention and then also hyperactivity and impulsivity. As for conditions that may look like ADHD, so kids who are having difficulty sitting still, or paying attention, or acting in an impulsive way, but maybe it's not ADHD, there are a number of other things that might drive that. So other diagnoses or categories of disorders that you might want to think about if you're noticing a child has difficulty with inattention, hyperactivity, or impulsivity are things like anxiety, traumatic stress, learning disabilities, and also even autism spectrum disorder. So again, careful screening and talking with the family about what are they noticing, what are you noticing in different settings, not only at home, but maybe also at school, or activities, or with friends, and really carefully thinking about how to put all of that information together can lead you to the best and most appropriate diagnosis. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | How_Does_Depression_Present_Differently_in_Children_and_Adults_Part_6.txt | [MUSIC PLAYING] ERICA H. LEE: So depression can look differently in children and teens than it does in adults. So often for us as adults when we think about the word "depression," we think about being sad, withdrawn, kind of keeping to ourselves and feeling really low. But oftentimes in children and teens, what you're more likely to see is irritability. So you're seeing that they're getting frustrated at small things, that they get annoyed really easily. They just seem really cranky a lot of the time. And also with kids, something that can be surprising to adults is that we often see that they might be happy and really excited and engaged in one setting, so say maybe when they're with their friends. But then they're at home, and parents say they look completely different. They're withdrawn, they're isolating, they're really cranky, and they're keeping to their rooms. So that's also something commonly that we see with children and teens in regards to depressive symptoms. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | What_are_the_Barriers_to_Accessing_Mental_Health_Treatment_Part_5.txt | [MUSIC PLAYING] ERICA H. LEE: So there are a number of barriers to accessing mental health care for children and teens. So first, just starting with the practical barriers, there's always the issue of cost and financial means for families. Not everyone has insurance coverage that covers mental health services. Then, of course, just logistically-- for kids and teens, they're not necessarily able to come to treatment on their own, so they need their parents to have flexibility at work to be able to get them there in terms of transportation. Or also, of course, always thinking about things like language barriers, and whether or not there's stigma. There's a lot of stigma and shame often for many families around the idea that my child needs mental health support, especially from an actual professional. So those can all get in the way. And then another barrier that we're always thinking about, especially these days, that there is a tremendous shortage of child and adolescent mental health professionals in pretty much every part of the country. So even in places, large urban areas, where we have a higher percentage of, say, psychologists, social workers, therapists, it's just not enough to meet the growing demand. And actually, recent estimates have shown that only about half of children and teens in any given year are able to access mental health treatment when they need it. |
Child_and_Adolescent_Mental_Health | What_is_Resilience_and_Why_is_it_Important_to_Mental_Health_Part_18.txt | [MUSIC PLAYING] ERICA LEE: So resilience refers to an individual's ability to overcome hardship and stress that they may experience in their life. Resilience typically results when our protective factors enable us to bounce back from hardship, and they outweigh whatever the stresses that we experience from that hardship. And resilience is especially important in the early childhood years because we know so much learning and growing and development that is critical to children's long-term futures happens when they're young. And so this is when it's most important to have these protective factors that boost a child's resilience. At the same time, while it is true that what happens early in life sort of sets the stage for everything that comes after, resilience is actually something that we can build across our entire lifetime. So it's not something that says, well, just because we can't prevent kids from experiencing hardship or going through sometimes very, very difficult and overwhelming adversity when they're young, means that they cannot then become resilient, say, teenagers or adults down the line. This is something we can continue to grow and to nurture all throughout our lifetimes. When a child or teen is experiencing significant stress, the research tells us that the most important protective factor is having at least one relationship with a supportive, stable, and consistent caregiver. So that's the place to start, right, is first-- and sometimes that's not a caregiver in the traditional sense. Sometimes it's a parent or a grandparent or another family member, but it can also be other important people in a child's life, whether that is a teacher, whether that is a coach, whether that's someone at a local community or cultural organization. But the idea being how do we build really strong and stable relationships? Because this helps kids have a sense of predictability. Even when the world feels upside down, even when I'm going through something, sometimes kids go through things that we can't possibly even imagine. I know that there is one person, a caring adult, who is there for me, who's going to help guide me, and will consistently be there to show up and help me figure out what I should do next. Additional things we can do to help kids build resilience are to maintain healthy routines, so making sure they're getting enough good sleep. Are they following a healthy diet? Are they getting regular exercise? In addition to this caring, supportive relationship that they're hopefully having with at least one adult, how are they doing with peers? Do they feel affirmed? Do they feel seen? Are they able to engage in activities with other people where they feel successful? And do they have areas in their lives where they can nurture their own talents and interests? It's so important for kids, especially teens and kids as they're getting older, to feel that they have areas of their own expertise. What am I doing that I just love? Where are areas where I can feel like I'm competent and strong and I can build that confidence and self-esteem And then lastly, making sure that kids have a nice toolkit of active coping skills. When life feels hard, it's so much easier, and this is just a very human instinct to kind of shut ourselves away, to do less, to isolate. But actually, we find that those only increased risk for things like depression and anxiety, sometimes over the short term, but definitely over the long term. So how, when I'm faced with adversity, can I reach into my toolkit and use some active coping skills so I don't get stuck feeling that way? So all of these strategies can go a long way towards kids building resilience. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | How_are_Youth_Screened_for_Anxiety_Part_9.txt | [MUSIC PLAYING] ERICA H. LEE: So when screening children and teens for anxiety, it helps to start with a well-established and broadband kind of measure. So something like the Pediatric Symptom Checklist, was help assess difficulties in a wide range of areas. So the PSC is great because you can look at our kids and teens reporting difficulties with internalizing problems. So here we're thinking about things like anxiety and depression, externalizing problems, so disruptive behavior problems. And then also potential symptoms of attention deficit or hyperactivity disorder or ADHD. So how are they having difficulty, say with their inattention. And then if you're noticing, hey, now that I've given the pediatric symptom checklist, this child or teen is reporting a good number of anxiety symptoms. You can go back in with a more focused screening measure. So for example, you might use something like the SCARED or the Screen for Child Anxiety Related Disorders, which gives you really specific questions that helps you narrow down. If a child is reporting a clinical level of anxiety, what kind of anxiety is it? So is it more generalized anxiety? Is it separation anxiety? Is it social anxiety? Or school avoidance? So that can be really useful for our youth ages 8 through 18. For our youth ages 12 plus, there's also a shorter screening measure that just focuses on symptoms of generalized anxiety. So this is called the Generalized Anxiety Disorder-7 or the GAD-7. So this can be very helpful at just screening sometimes for our older youth, I don't necessarily want to go as deep as I did on the SCARED to look at all of the different possible anxiety disorders. I just want to check if they're experiencing generalized anxiety disorder symptoms. It's a nice quick measure. And then once you've got this more detailed information about what anxiety symptoms a child or teen is experiencing, you can discuss with the family. What do we think makes sense? You know, based on what you're reporting to me, based on what your child is reporting to me. Let's put all this information together and decide on a plan. So that might be, let's continue to monitor. Let's see how things develop, let's kind of check in. It might be, hey, it would be helpful to talk a little bit with your school counselor to see if we can improve things at school. Or it could be, now would be the time for a referral to a behavioral health specialist. |
Child_and_Adolescent_Mental_Health | How_Common_are_Mental_Health_Issues_in_Children_and_Adolescents_Part_1.txt | [MUSIC PLAYING] ERICA H. LEE: So mental health issues in children and adolescents are, unfortunately, more common than we'd like them to be. So in any given year, about one in five children or teens meets criteria for an emotional or behavioral health disorder. So this would include things like anxiety, depression, disruptive behavior problems, or autism. This means that about 5 million children and teens a year are having difficulty with their daily functioning. They've got a level of symptoms that gets in the way of their ability to do things that they want and need to do at home, at school, or with other kids. So over the last few years, driven by, unfortunately, the COVID-19 pandemic, we've seen a pretty dramatic increase in the rate of behavioral health disorders among children and teens, particularly in the areas of anxiety, depression, eating disorders, and also suicidality. And in response, in 2021, a number of leading pediatric and psychiatry organizations declared a national state of emergency in children's mental health. [MUSIC PLAYING] |
Child_and_Adolescent_Mental_Health | How_are_Disruptive_Behavior_Disorders_Diagnosed_Part_14.txt | [MUSIC PLAYING] ERICA H. LEE: So of course, we want to look at a child's age and stage. So where are they versus where would we expect them to be developmentally. So some of these behaviors like throwing frequent tantrums, getting really angry, saying no or having a hard time following rules or following directions, these are actually pretty common in some form up to about the age of eight. And then after that, you want to look at impairment. How are this child's behaviors getting in the way of what they need to be doing at school, at home, with friends. You can also use standardized rating scales. So you can use a broadband measure like the pediatric symptom checklist which is going to give you information about are there sort of some difficulties around externalizing problems or what we think of as disruptive behavior. You can also use something more focused like the Vanderbilt rating scales. So these are really helpful because there's a parent rating form and there's also a teacher rating form. So you can get a sense of what symptoms is a child experiencing both at home and then also in the school setting. Are we noticing some commonalities? Are we noticing differences? How much is a child pushing back against sort of the rules and boundaries in their daily experience? And that can give you really valuable information about whether or not a child is just kind of having a hard time and maybe a little bit more disruptive on average than you would like or actually moving in the direction of meeting criteria for a disruptive behavior disorder. [MUSIC PLAYING] |
Power_and_Politics_in_Todays_World | Lecture_15_Demise_of_the_Neoconservative_Dream_From_Afghanistan_to_Iraq.txt | - So today, we're starting in part three of the course which I call the end of the end of history. And the end of the end of history harken back to the book, "The End of History" that we talked about earlier. What was this idea that democracy and liberal institutions and market capitalism had come to the world to stay, was embodied in the neoliberal consensus at home and the Washington Consensus abroad. And as I have said several times, I'm calling this the early post Cold War period. And so when I talk about the end of the end of history, when did the end of the end of history happen in the US? And the truth is it didn't happen all at once and it was a bit like turning around a massive ocean liner at sea. I would say that in the full consciousness of mainstream elites, it really ended with a bang with the election of Donald Trump in 2016. Even after Brexit, many people in America didn't think that was heralding something that might be happening here. So I think that, but then when that happened, I think people started to realize we were living in a different world than the comfortable world or the end of history that we had come to identify. But if you want to say when did it really start ending as opposed to when did people realize it was ending, I think that there was a kind of bifurcation. At home, it really starts, I've dated it to the collapse of Lehman Brothers on September 15th of 2008 which was the beginning of the financial crisis which was gonna have huge knock on effects for America's place in the global economy, for the Washington Consensus. There was gonna be this big retreat. Russia and China, as we're gonna talk about this time next week, would fill a big part of the void as the Washington Consensus went home or as the purveyors of the Washington Consensus went home and the consensus itself started to disintegrate. We started hearing about the Beijing consensus, and of alternative development models, and so on. So I think that was the beginning of the collapse at home. Abroad, when we think about geopolitics rather than economics, really it begins with, I've dated it to the day before the invasion of Iraq, so March 19th 2003, because I think that was probably the high point of American global geopolitical influence in the world for reasons that I'll talk about today. So a little taking stock. The Cold War had been won first of all at a surprisingly low cost. The Soviet Union had knelt down and died. There was no big war. It was a fairly peaceful transition and in terms of blood and treasure, much less than people might have anticipated. The 1990s and early 200s as I've just said were the heyday of the Washington Consensus and democracy seemed to be on the march. To remind you of a statistic I told you about earlier, at the turn of the 21st century for the first time in human history, there were more democracies in the world than non-democracies. So there's a lot of confidence in that. 9/11 gave the US huge leeway to act internationally. There was a very broad perception that we had been a victim of an justified terrorist attack, not universal, but extremely broad consensus that the US certainly was going to do what was necessary to chase down the terrorists that had attacked us and nobody was gonna second guess that move that we began in Afghanistan, even if as it turned out later, there were pause not taken and things that might have been done differently. But it was a time of unprecedented self confidence in America's capacity to shape the world. And perhaps this video of George W. Bush on the 1st of May 2003 encapsulates that self confidence about as well as anybody could. (engine rumbling) (faint speaking) (fanfare) - In the outlaw regime that has ties to terrorist groups and seeks or possesses weapons of mass destruction is a great danger to the civilized world and will be confronted. (audience applauding) We've begun the search for hidden chemical and biological weapons, and already know of hundreds of sites that will be investigated. Thanks, we like it. Yeah, fantastic. The transition from dictatorship to democracy will take time but it is worth every effort. Our coalition will stay until our work is done. (faint speaking) (laughing) - So that sums it up and another remarkable statement of this was, this was a comment made by an anonymous White House aide. Turned out later to be Karl Rove to Ron Suskind, the journalist. In October of 2004, he had been complaining about a story that Suskind had written about Karen Hughes who had been the White House Press Secretary. So Rove said to Suskind, "You people seem to think "that we have to start with reality, "but we're an empire now and when we act, "we create our own reality. "And while you're studying that reality, "judiciously as you will, "we'll act again creating other new realities "which you can study too, "and that's how things will sort out. "We are history's actors, "and you, all of you, will be left to study what we do." Right? It's hard to say anything stronger than that about confidence in America's capacity to shape reality. And just again to make the contrast, let's go back and see what George W. Bush was saying about US global strategic ambitions and practices three years earlier when he was running for the presidency. - Are we gonna step up to the plate as a nation the way we did after World War II, the way that generation of heroes said, "Okay, the United States is gonna be the leader." And the world benefited tremendously from the courage that they showed in those post war years. - I'm not so sure the role of the United States is to go around the world and say this is the way it's gotta be 'cause I am worried about overcommitting our military around the world. I want to be judicious in its use. I just don't think it's the role of the United States to walk into a country and say, "We do it this way, so should you." If we're an arrogant nation, they'll resent us. If we're a humble nation but strong, they'll welcome us. And our nation stands alone right now in the world in terms of power, and that's why we've got to be humble. If we're an arrogant nation, they'll view us that way, but if we're a humble nation, they'll respect us. And so I don't our troops oughta be used for what's called nation building. I think what we need to do is convince people who live in the lands they live to build the nations. Maybe I'm missing something here. I mean, we're gonna have kind of a nation building core. I'm concerned that we're over-deployed around the world. I don't want to be the world's policeman. And the exit strategy needs to be well defined. - So what we see between that speech in 2000 or those various clips from 2000 and the speech I showed you last time from Rumsfeld and the mission accomplished speech and Rove's attitude is a more or less complete take over of the Republican administration by the neoconservative wing of the Republican party. Now, I have talked about neoliberalism and the Washington Consensus, neoliberalism being the commitment to the trifecta of low regulation, free trade, and privatization, but I haven't talked about neoconservatism yet in this course and I'm here talking about today the demise of the neoconservative dream, but we need to be clear on well, who were the neoconservatives and what was their dream that demised or came to its demise? So here's some reflections by Irving Kristol who is widely known as the godfather of neoconservatism. - [Interviewer] Irving Kristol, author of "The Autobiography of an Idea: Neoconservatism." When was the first time you ever heard that word uttered? - It must have been some time in the mid 1970s. It was not my term. I did not invent it. I believe it was invented by someone who was criticizing me and I thought that that was a term of a (mumbles). I decided that it was a pretty good description of in fact what I was thinking and feeling, so I ran with it. - [Interviewer] What does it mean? - What it means is that it refers to a constellation of opinions and views that is not traditionally conservative, but is conservative, and is certainly not liberal. Things, particularly intellectual political things, intellectual and political ideas were more in flux in the 1970s than either before or after. That was the decade of transition so far as I was concerned. In 1968, I still voted for Hubert Humphrey, but by 1974, I realized I was going to become a Republican. And by that time, the term neoconservatism had been invented and I decided I was also a neoconservative. - You say that you've been a neo-Marxist, a neo-Trotskyist, a neosocialist, a neoliberal, and finally, a neoconservative. - That pretty much phrases the trajectory of my political beliefs. I've never been comfortable with any of those doctrines because I always saw problems inherent in those doctrines. Those who would 10 years ago have been called neoconservatives these days simply call themselves conservatives. The conservative movement has expanded to include us. - So there it is. Now you know what neoconservatism was. In fact, you can watch that entire interview from which I excerpted three minutes and still not be able to glean. You can glean what it wasn't, but it's very hard to get a sense of what it was. But he nonetheless tells us a couple of important things about neoconservatism that I think give clues to what it was. To start at the end, it was clearly something that had gone mainstream by the time he was talking in 1995. So as he said, people were no longer talking about neoconservatism as different from mainstream conservatism by then. But then the second clue is his own personal history. He started out on the left actually as a kind of Trotskyist critic of the Soviet Union before World War II. As information was coming out on the left, people like the Webbs and others in the UK, there started to be a left critique of the Soviet Union. The people who ultimately became the neoconservatives sometimes were called liberal hawks. And they very much, another one much later actually, people called Christopher Hitchens a neoconservative for his support for the Iraq War. So it had a long lineage and its roots were in the pre World War II criticism of the USSR. And in the '60s and '70s, it was fueled by people, a lot of whom had started out in the sort of Scoop Jackson Democrats who didn't like the peace movement, didn't like the peace nix. They didn't like the new left, didn't like the '60s, and all of that. And they tended to ally themselves with the proponents of rollback in various administrations, so in Republican administrations, and they certainly were a lot closer to Nixon's policies than they were to the Democrats who came subsequently. So they were for aggressive prosecution of the Cold War, staunchly anti-communist and anti-socialist. In the post Cold War period, of course, like everyone else, the neoconservatives were rethinking what their views should be in this new world since their sort of raison d'etre for decades had been virulent anti-communism and communism for all practical purposes had gone away. And this is when the neoconservatives I think started to differentiate themselves from other conservatives, certainly from libertarian conservatives and people who were subsequently critics of theirs come to be called paleoconservatives or to call themselves paleoconservatives 'cause they didn't like being identified with the neoconservative label. And the distinctive feature of the post Cold War neoconservatives were two, I think. One is that they wanted to spread democracy around the world. They were interested in knocking off non-democratic regimes, but they also didn't doubt for a second that democratic regimes would be pro-American. So they really were if you like, we might say today drinking the Kool Aid of the early post Cold War consensus, but they thought that installing democracy in the Middle East or anywhere else would automatically produce regimes that were friendly to the US. And indeed, just to give two illustrations of that. One of the justifications for the invasion of Iraq was that it would be a positive domino process. We would democratize Iraq and we would get a friendly western style regime there and then that would have knock on effects either. And partly because that didn't happen, there was huge pressure to conduct elections in Lebanon even though all the experts were telling them it would actually strengthen the position of the anti-American militias. Or you can see in 2005 and six, they pushed strongly for elections in the West Bank and Gaza and were stunned when Hamas won the elections in Gaza. So there'd been this expectation. Not only is it easy to implant democracy, but democracies will tend to see the US as a benign force and will be pro-American. And they were people like Eliot Abrams, Richard Perle, Paul Wolfowitz, and Cheney who had been the defense secretary in George Herbert Walker Bush's administration and would come back with George W. Bush, and of course, Donald Rumsfeld. And if you want to find a defining example of neoconservative thinking, it's this document published in 1995 called "From Containment to Global Leadership: "America and the World After the Cold War," written by this man Khalizad who is currently by the way Pompeo's envoy to Afghanistan and he served in multiple administrations in various capacities, but he produced this document in 1995 basically in which he said in the post Cold War era, America's got three options geopolitically. It can go back to isolationism. He called it neoisolationism. People like this neo word. I don't know what the difference between neoisolationism and isolationism. One is you withdraw from the world and the other is you withdraw from the world, but. (audience laughing) The other would be to go back to pre World War II, a multi-polarity where it was all about balancing, much of the sort of balancing that had gone in even in the 19th century in Europe. And he made a case that both of those things would be very bad for the US, and instead, we should take a global leadership role, and this is the key claim in the report was or the document was that he advocated that the use of force and even preemptive war if necessary to preserve US global hegemony. And so this is what differentiated the neoconservatives from other conservatives. And it was that document that really became the forerunner for the 2002 and 2006 national security strategies that were embodied in the Bush doctrine that we talked about last time. So that's the sort of main elements of the neoconservative creed and the history or the biography of it, if you like. So intellectuals have ideas, but how did they become important in politics? And as with the neoliberal views, it goes back to think tanks. I mentioned to you when we were talking about the resurgence of the Republican right that after the Nixon administration fell apart in the 1970s, there was this emergence of activist think tanks who were committed not to winning the next election, but rethinking the conservative agenda for the medium term funded by Fulner, Koch, and others in places like The Heritage Foundation, The Cato Institute, and related conservative think tanks who were mostly interested in domestic politics, particularly rolling back the frontiers of the welfare state. And they got their great moment in 1994 when the Republicans took over the house for the first time in four decades and Newt Gingrinch set himself up as a kind of prime minister on Capitol Hill. Had Bill Clinton on the run having lost the midterm elections badly. And that was the opening and we've been through that, the contract with America and all of that. But there was also a geopolitical rethinking among these people we're talking about now, and I'm calling it the neoconservatives. And they were also in think tanks. Some of them in not particularly partisan think tanks like the Rand Corporation which is where that document I just read from came out, and the Council on Foreign Relations which certainly billed themselves as nonpartisan even if many people saw them as sort of right of center. But then places like the American Enterprise Institute which are sort of activist conservative think tanks. People like Perle were there and a number of the other neoconservatives spent a lot of time there. And then perhaps most important was a new think tank created in 1997 by Bill Kristol, that's Irving Kristol's son, and Robert Kagan, the son of our colleague Professor Kagan here in the history department. And the Project for a New American Century was probably the prime incubator of neoconservatism and many people who subsequently wound up in the Bush administration spent time there were passing through these think tanks. And really, 9/11 was their moment. 9/11 was the moment at which they realized that there was a huge vacuum in thinking about US strategic policy and they could go for it. And as I think I mentioned to you last time, by December of 2001, I'm sorry, by December of 2000, there were already conversations going on, I'm sorry, 2001, December of 2001, there were already conversations going on in the White House as to when they could reallocate troops from Afghanistan to Iraq. Now, many people might say, well, why did anybody really want to attack Iraq? It seems something of a puzzle. The neoconservatives had this view about it, but we really have to back into it and look a little bit at the historical background here. So after Saddam Hussein was kicked out of Kuwait in 1991, I described the containment regime that the Bush administration constructed with the no fly zones and oil for food and all of that that was essentially meant to hem Saddam Hussein in. They rejected regime change and so on. They rejected internally externally imposed regime change, but the George Herbert Walker Bush administration was by no means opposed to fomenting the overthrow of Saddam Hussein's regime, just didn't want to commit American troops to that. And indeed, one of the things he was greatly criticized for at the time was encouraging the Shiites in the south and the Kurds in the north to overthrow Saddam Hussein. But when they made those efforts, there was no US backup and support and that led to the slaughter of many of them by the (mumbles) regime. And beyond that, the CIA in the years following spent a lot of time and resources working with opposition groups in Iraq and expatriate groups outside of Iraq to try and engineer the overthrow of that regime even though it wouldn't involve American boots on the ground. And indeed, there was the creation of something called the Iraqi National Congress led by a man called Ahmed Chalabi who was an Iranian expatriate who had been living in London for decades. He hadn't been in Iraq for a very long time. And he worked very closely with the CIA and others to get American indirect support for overthrowing the Iraqi regime. They put together this Iraqi National Congress which was basically an umbrella organization to try and get all the opposition groups in Iraq together and figure out how to knock off the Saddam Hussein regime. This was formed in 1992. By 1994, it was falling apart in acrimony. The two Kurd groups were fighting over the control of different parts of the country, mostly having to do with oil, and some of them were actually working with Saddam Hussein's regime. Chalabi convinced the US nonetheless to persist with the kind of remnants of the Iraqi National Congress which he headed and ultimately became a political party in the new Iraq. And he convinced the US, the Clinton administration in 1995, to sponsor an attempt at overthrowing, sort of Clinton's I guess little under publicized, but Clinton's kind of bay of pigs in that Chalabi predicted and promised that as soon as this got going, the Iraqi military would welcome it and they would overthrow Saddam and they would liberate Iraq. It was quickly put down. 100 people were executed, many more imprisoned, but it didn't end the love affair between the Clinton administration and particularly the neoconservatives in these think tanks and Ahmed Chalabi who constantly worked to raise money and to get support for overthrowing the regime. However, the Clinton administration wasn't that interested in this. It was basically being dragged along by the conservatives who now controlled congress after 1994. Clinton was basically enforcing the containment regime that his predecessor had put in place. So for example, in December of 1998, there were three days of bombing Saddam Hussein which he described as follows. - Three days ago, despite clear warnings from the United States and the international community, Iraqi forces attacked and seized the Kurdish controlled city of Erbil in Northern Iraq. Limited withdrawals announced by Iraq do not change the reality. Saddam Hussein's army today controls Erbil and Iraqi units remain deployed for further attacks. These acts demand a strong response and they have received one. Earlier today, I ordered American forces to strike Iraq. Our missiles sent the following message to Saddam Hussein. When you abuse your own people or threaten your neighbors, you must pay a price. Our objectives are limited, but clear, to make Saddam pay his price for the latest act of brutality, reducing his ability to threaten his neighbors and America's interests. First, we are extending the no fly zone in Southern Iraq. This will deny Saddam control of Iraqi airspace from the Kuwaiti border to the southern suburbs of Baghdad and significantly restrict Iraq's ability to conduct defensive operations in the region. Second, to protect the safety of our aircraft enforcing this no fly zone, our cruise missile struck Saddam's air defense capabilities in Southern Iraq. Saddam Hussein's objectives may change, but his methods are always the same. Violence and aggression against the Kurds, against other ethnic minorities, against Iraq's neighbors. Our answer to that recklessness must be strong and immediate. As President Bush demonstrated in Operation Desert Storm, as we showed two years ago when Iraq mast its forces on Kuwait's border, and as we showed again today, we must make it clear that reckless acts have consequences or those acts will increase. We must reduce Iraq's ability to strike out at its neighbors and we must increase America's ability to contain Iraq over the long run. - So there you see he's speaking the language of containment. They called it aggressive containment. It was the sort of policy he's talking about in that clip. But the neoconservatives weren't happy with aggressive containment. You might think of them as the latter day proponents of rollback. Indeed as mentioned last week, the Bush doctrine was really about rollback of aggression of any sort to anywhere in the world at the time of America's choosing. So the neoconservatives weren't happy with aggressive containment and pushed the passage of something called The Iraq Liberation Act of 1998. The Iraq Liberation Act of 1998, just to give you some sense of how much the democrats here were either on the run or lacking in an alternative conception of their own, a point to which I will return, the act passed 360 to 38 in the house and by unanimous consent in the senate. And that act declared that it was a policy of the United States to support efforts to remove the regime headed by Saddam Hussein. So clear commitment of the United States government to overthrowing the regime. It authorized 87 million dollars to support seven designated groups and it created the opportunity to support more of them with the clear beneficiary that everyone had in mind was The Iraqi National Congress. And this would be cited by the Bush administration in 2002 in justification of the upcoming invasion. But let's just parse the Iraq Liberation Act a little more closely again by watching Bill Clinton that same day, from a little later that same day than the earlier clip I showed you. - This crisis also demonstrates unfortunately once again that Saddam Hussein remains an impediment to the well being of his people and a threat to the peace of his region and the security of the world. We will continue to contain the threat that he poses by working for the elimination of Iraq's weapons of mass destruction capability under (mumbles), enforcing the sanctions in the no fly zone, responding firmly to any Iraqi provocations. However, over the long term, the best way to address that threat is through a government in Baghdad, a new government that is committed to represent and respect its people, not repress them, that is committed to peace in the region. Over the past year, we have deepened our engagement with the forces of change in Iraq, reconciling the two largest Kurdish opposition groups, beginning broadcasts of a radio free Iraq throughout the country. We will intensify that effort, working with congress to implement The Iraq Liberation Act which was recently passed, strengthening our political support to make sure the opposition, or to do what we can to make the opposition a more effective voice for the aspirations of the Iraqi people. - So there you can see he's still talking the language of containment even while addressing The Iraqi Liberation Act. Notice the date. The date's a significant one, December 19th 1998, and it explains the kind of deer caught in the headlights look on Bill Clinton's face because that was the day that the house voted to impeach him. So he was not a happy camper, one might say, on that day while making that speech. So this was the tussle that went on toward the end of the administration. And the Clinton Administration was never ever contemplating actually going in and overthrowing. In fact, of the 87 million dollars, only eight was ever allocated by the end of the administration. So they were largely going through the motions and he was running scared of the Republicans in congress. And so what happens after that is that these ideas are all rolling around in the conservative think tanks and being pushed by Chalabi, but they're not really making their way into, as you see, they didn't make their way into the political campaign of George W. Bush. And when 9/11 came along, they thought this is our moment, and it was their moment. And it was then produced a huge conflict within the administration because on the one hand, you had the Pentagon largely allied with Chalabi who wanted to go in, but the State Department and the CIA were strongly opposed to going in. And indeed, they said that most of the intelligence that they saw, if not all of it, did not show any connection between Saddam Hussein's regime and 9/11, or that there was any active program of building weapons of mass destruction in Iraq which were two of the war aims most commonly cited in the quest to go in to Iraq. In the Pentagon on the other hand is where the neoconservatives were. And they were working with intelligence that was largely provided for them by Chalabi which turned out later to be fake news. And this was the fight that was going on and that once 9/11 occurred, the Pentagon people and the neoconservatives basically got into the driver's seat and took over partly because both Rumsfeld and Cheney were sympathetic to their goals and already in the administration. And so we had this invasion of Iraq that was a colossal catastrophe for America. It was a colossal catastrophe on a number of fronts and I would say that just to give some statistics, first of all, in terms of US casualties, it should be 4,970. I'll fix that slide. Almost 5,000 killed, 32,000 wounded. Between 150,000 and 600,000 Iraqis, I'll fix that too, killed depending on which estimates you go with. They had given as one of the reasons for going into the war was that the cost of containment, but of course, it was nickels and dimes compared to the 1.9 trillion that the CBO was predicting the war would cost in 2007 and now, our colleague Bill Nordhaus has a paper estimating it costed in excess of three trillion dollars. So the financial cost was just enormous and it was a massive display of American weakness. And this is why I mentioned the date before the invasion as the key date if you want to pinpoint the high point of American influence in the region. It was a massive display of American weakness in two senses. One was that the US couldn't in fact, the US couldn't win this war in any sense that would enable them to live up to the mission accomplished that George Bush had announced on May 1st. So what they conflated was the capacity to destroy a country's army with the capacity to actually pacify and govern a country, to win a piece, if you like. And so they put in a barely competent general very early on and he couldn't run the country. They replaced him with Paul Bremer, a neoconservative, and he began by disbanding the Iraqi Army and officer corps in particular because they were Sunni and the majority of the population were Shiite, and disbanding the Iraqi civil service out of fear that they might be loyal to Saddam Hussein. And so it effectively created a failed state that descended into civil war. They discovered that their investment in Ahmed Chalabi had been extremely ill conceived because not only did he give them bad intelligence as he had been doing all the way back to 1995 in the Clinton Administration, he predicted that the Americans would be greeted with flowers in the streets and that we would have no difficulty creating a democracy in Iraq. And interestingly enough, they immediately appointed him president of the transitional governing council in Iraq. But when the 2005 elections were finally held, they had by then appointed him deputy president or deputy prime minister and minister of oil, he didn't even win a seat in the elections and so he had to leave the government. And he subsequently became involved in a variety of corrupt schemes and fell out of favor with the Americans and finally died I believe in 2015. So it was a massive display of American weakness in that sense as well. And it's a lesson that proponents of rollback failed to learn again, and again, and again, and that I mentioned to you last week which is that you can destroy a country's army perhaps, but destroying people's capacity to fight when they have everything at stake and you're fighting an elective war is a very tough proposition. So this massive display of American weakness, the domino theory didn't work. Not only were we not getting democracy of any functioning sort in Iraq, but it was not spreading elsewhere in the Middle East. And you might wonder to yourself why? Why would the neoconservatives believe you could create democracies in these countries that had no history of democracy? And I think that the puzzle there is less difficult to understand if you think back to what they were looking at. What they were looking at was Eastern Europe after the collapse of communism. All of those countries that had been occupied by the Soviets or had been occupied by the Soviets and then run by Soviet puppet regimes, as soon as the Soviets fell apart, democracy sprang up. There wasn't any difficulty in getting democracy installed across Eastern Europe. And so they thought well, that's what happens. You get rid of the oppressive dictator and you create a democracy and everything will be hunky dory. And not only that, those democracies in Eastern Europe were largely pro-American. And so the conflation of creating a democracy and having a pro-American regime you might say was understandable. What they were ignorant of I think were several things and they would have had much better clues if instead of looking at Eastern Europe where they were fairly developed diversified economies everywhere except Albania that had prewar histories of democracy in many of them, they should have looked at the (mumbles) former republics of the Soviet Union to think about what was likely to happen when the Soviet Communist systems disintegrated. And of course, there, there had not been transitions to democracy. And those were countries with much lower per capita income. We know that that's important for democracy. They were undiversifiable, countries like Kazakhstan which is largely just an oil economy. So no diversification. If they had a high per capita income, it was oil curse high per capita income. So it was undiversified economy. We know that's not propitious for democracy either. So for all of these reasons, it was really pie in the sky to believe that you could just get democracy to spring up. And indeed, if you think back to the two most successful cases of externally imposing democracy on a country, we have to go back to the post World War II West Germany and Japan which were very unusual for a number of reasons. One was that the populations themselves blamed their prior regimes for bringing disaster upon them, right? In 2003 in Iraq, the Iraqi population did not blame Saddam Hussein for the American invasion, whereas the Germans and Japanese had started aggressive wars and so they recognized they had themself to blame. They were occupied for decades by either occupying forces or hovering forces. They were protected by the US umbrella during the Cold War. They were supported economically by the Marshall Plan and of course, they had favorable demographics for reasons I talked to you about some lectures ago. None of these conditions were remotely present in the countries in the Middle East that the Bush Administration was hoping for and the neoconservatives were hoping would produce a domino effect of successful pro-Western democracies across the regions. So all of these were costs of the invasion of Iraq, but perhaps the greatest cost of the invasion was a loss of moral high ground on the global stage. - Last month, the US Army announced 17 soldiers in Iraq including a brigadier general had been removed from duty after charges of mistreating Iraqi prisoners. But the details of what happened have been kept secret until now. It turns out photographs surfaced showing American soldiers abusing and humiliating Iraqi prisoners at a prison near Baghdad. Americans did this to an Iraqi prisoner. According to the US Army, the man was told to stand on a box with his head covered with wires attached to his hands. He was told that if he fell of the box, he would be electrocuted. Abu Ghraib Prison under Saddam Hussein was infamous. For decades, many who were taken here never came out. And those prisoners who did make it out told nightmarish tales of torture beyond imagining and executions without reason. Abu Ghraib was the centerpiece of Saddam's empire of fear. We talked about the prison and shared pictures of what Americans did there with two men who have extensive interrogation experience, former US Marine Lieutenant Colonel Bill Collin, and former CIA Bureau Chief Bob Bayer. - I visited Abu Ghraib a couple days after it was liberated. It was the most awful sight I've ever seen. And I said if there's ever a reason we want to get rid of Saddam Hussein, it's 'cause of Abu Ghraib. There are bodies that were eaten by dogs, torture, electrodes coming out of walls, scratches. It was an awful place. - We went into Iraq to stop things like this from happening, and indeed, here they are happening under our tutelage. - It was American soldiers serving as military police as Abu Ghraib who took these pictures. The investigation started when one soldier got them from a friend and gave them to his commanders. We have a dozen of them. There are many, many more. The pictures show Americans, men and women in military uniforms posing with naked Iraqi prisoners. There are shots of the prisoners stacked in a pyramid, one with a slur written on his skin in English. In some, the male prisoners are positioned to simulate sex with each other. And in most of the pictures, the Americans are laughing, posing, pointing, or giving the camera a thumbs up. One of the soldiers facing court marshal is Army Reserve Staff Sergeant Chip Frederick. He is charged with maltreatment for allegedly participating in setting up this photo and for posing in a photograph sitting on top of a detainee. Sergeant Frederick is also charged with an indecent act for observing this scene. He is charged with assault for allegedly striking detainees and ordering detainees to strike each other. Sergeant Frederick's letters and email messages home offer clues to problems inside the prison. Frederick wrote that he was helping the interrogators. Quote, military intelligence has encouraged and told us great job. They usually don't allow others to watch them interrogate, but since they like the way I run the prison, they've made an exception. We help getting them to talk with the way we handle them. We've had a very high rate with our style of getting them to break. They usually end up breaking within hours. According to the Army's own investigation, that's what was happening. The Army found that interrogators asked reservists working in the prison to prepare the Iraqi detainees physically and mentally for questioning. Attorney Gary Meyers and a judge advocate in Iraq are defending Chip Frederick. They say he should never have been charged because of the failure of his commanders to provide proper training and standards. How could such a person described almost universally by people who knew him as a good guy get himself into this kind of jam? - The elixir of power, the elixir of believing that you're helping the CIA for god's sake when you're from a small town in Virginia, it's intoxicating. And so good guys sometimes do things believing that they are being of assistance and helping a just cause and helping people who they view as important. - Sergeant Frederick says he didn't see a copy of the Geneva Convention rules for handling prisoners of war until after he was charged. The Army investigation confirms that soldiers at Abu Ghraib were not trained at all in Geneva Convention rules. Frederick says there were far too few soldiers there for the number of prisoners. About how many prisoners were there? - [Chip] There was, when I left, there was over 900. And there was only five soldiers plus two non-commissioned officers in charge for those over 900 inmates. - That may have been a contributing factor, but at the end of the day, this is probably more about leadership, supervision, setting standards, abiding by the army values, and understanding what's right and having the guts to say what's right. - I see a lot of work going on out here now. We're changing everything. - [Dan] This is Brigadier General Janis Karpinski who ran Abu Ghraib for the Army. She was also in charge of three other Army prison facilities housing thousands of Iraqi inmates. The Army investigation determined that her lack of leadership and clear standards led to problems system wide. General Karpinski talked with 60 Minutes Steve Kroft last October at Abu Ghraib before any of this came out. - And this is international standards. It's the best care available in a prison facility. - But the Army's own investigation found serious problems behind the scenes. The Army has photographs that show a detainee with wires attached to his genitals and another that shows a dog attacking an Iraqi prisoner. Were dogs ever used on the inmates? - [Chip] Yes, sir. They were used for intimidation factors. - [Dan] And then there is this picture of an Iraqi man who appears to be dead and badly beaten. - Dan, I'd like to sit here and say these are the only prisoner abuse cases that we're aware of, but we know that there have been some other ones since we've been here in Iraq. - [Dan] When Saddam Hussein ran this prison, Iraqis were too afraid to come to Abu Ghraib and ask for information on their family members. When we were there last month, hundreds had gathered outside the gates worried about what's going on inside. (woman sobbing) (people chattering) - We will be paid back for this. These people at some point will be let out. Their families are gonna know. Their friends are gonna know. - [Dan] This is a hard story to have to tell when Americans are fighting and dying in Iraq. - So it's hard to overstate the shock that that caused at home, but the biggest cost I think of Abu Ghraib was not at home but it was the loss of the moral high ground that the US had enjoyed until that time across the region. If you look at the polling across much of the Middle East about perceptions of the US during this period, I described it in the reading that I assigned for today, it fell off a cliff. And the notion that the US was in some sense bringing a better world and had different kinds of practices was this sort of flagrant violation of Canon's dictum that we should not behave in the way that our adversaries behave and that the long term strategy of containment is for the hearts of minds of people in the countries that are being contained. And of course, the fact that this was all exposed in, initially, this 60 Minutes report was in April of 2004, but the following month, Seymour Hersh published a huge expose in The New Yorker where he had gotten hold of the Army's report which you saw referred to in that clip which revealed that they had known about what was going on in Abu Ghraib and a number of other prisons for many months. And the story just didn't get any better. What turned out to be the case in terms of prosecutions were minimal. The general you saw there was busted back to a colonel. There was some other disciplinary actions of officers. Nobody above the rank of staff sergeant was convicted of any criminal offense and only I think it was seven or eight people had any kind of criminal convictions, the highest being a 10 year sentence. So the perception was that the US had allowed the same kind of abominable practices that Saddam Hussein had engaged in and as I said, the cost in terms of loss of moral high ground is probably incalculable to this day. What made it even worse was that President Bush announced that the prison would be closed the following year which was not done, and in fact, it wasn't closed until after the Americans left. A decade later, it was closed by the Iraqis. So this combination of a demonstration of American weakness and the loss of any kind of claim to be acting in defense of humanitarian ideals was destroyed. In a geo-strategic sense though, I think the biggest cost and some might in some sense call it a benefit of the Iraq debacle was it was a clear demonstration that the Bush doctrine was unsustainable because the idea that you could go around the world knocking off regimes. This is the neoconservative dream, turns out to be a nightmare because you go around the world knocking off regimes unless you have the propitious circumstances and the kinds of investments that we made in Germany and Japan after World War II. The high probability is you're going to get the sort of outcome that we saw in Iraq. And so the Bush doctrine particularly given that it's being prosecuted in an era of cutting back government spending and downsizing the military as Rumsfeld argued in the clip I showed you at the beginning of last Thursday's lecture and contracting out more and more military activities, you're not gonna be able to go around the world doing the sort of thing we were trying to do in Iraq and Afghanistan and end up with, as Bush says in the clip I showed you earlier, a clear exit strategy. And indeed, the Bush doctrine was a complete repudiation of the so-called Powell Doctrine of you've got to go in with overwhelming force and a clear exit strategy. Instead, it's proxy wars fought on the cheap using expatriates and local militias who have very different agendas than one's own and this is the result that you shouldn't be surprised to see. But if not the Bush doctrine, what? If not the Bush doctrine, what? Because the 1995 document from Rand said what are the alternatives here? The alternatives seem to be neoisolationism as he called it, going back to a multi-polar world where in an era of nuclear proliferation, that seems very dangerous and costly for the US. So you might say well, where were the Democrats? We have a democracy here. We have an opposition party and one would expect that the opposition party would be articulating an alternative vision. And the Democrats though were completely on the defensive. So here, I should give you a personal anecdote. In September of 2008, I was invited to speak at the Yale Club of Tokyo to a group of alumni. We do that all the time. We're asked to go and talk about our research. So I answered with three possible topics of things I was then working on that I would be willing to speak out. And the answer came back that nobody would be interested in any of them. (audience laughing) And what they wanted me to speak about was what the Kerry Administration's foreign policy was going to be. This is September of 2000, I'm sorry, 2004. What the Kerry Administration's foreign policy was going to be. So I ended up giving a talk on why there wasn't gonna be a Kerry Administration and the central thesis of that talk was you can't beat something with nothing. And the reality here, you can see if you now go back over the lecture that I've just given you. You can see the clues that the Democrats really had no alternative. First of all, think about that vote of The Iraq Liberation Act. I said it was only 50 or 60 votes against it in the house, unanimous consent in the senate, and Clinton signs it into law in the midst of his impeachment hearings. The 87 million dollars that I mentioned to you that they authorized for spending on opposition forces in Iraq, that is the famous 87 million that Kerry was later got into trouble for saying well, I voted for the 87 million dollars before I voted against it. And was used in ads against him by President Bush when he was running for re-election that year. And if you look at the votes to authorize the war, Kerry had voted for the authorization of the invasion of Iraq in 2003 as had by the way Hillary Clinton. And so basically what was going on was that the Democrats were playing a game of triangulation. And as we saw when we talked about the anti-tax movement in domestic politics, the trouble with triangulation is that it's good tactics and it's bad strategy. It's good tactics because if you move towards your opponent's position, you can peel off some of their support and people on your flank really have no place to go. That was the theory behind it. But the reason it's bad strategy is that once your opponent figures out that this is what you're doing, their incentive is to move the goal posts. And so in effect, the Republicans and particularly the neoconservatives within the Republican party had the Democrats on the run because they were playing triangulation to catch up to this target that was moving steadily in a neoconservative direction. The same thing happened as happened with the anti-tax movement where the Democrats by 2001 found themselves voting for tax cuts that nobody would have taken seriously even six or seven years before such as the repeal of the estate tax. So the Democrats left any alternative geopolitical strategy to the Bush doctrine. The Bush doctrine filled a void and they just ran with it. And so one might think that once 2008 rolled around and the Democrats were again in control, things might have played out differently. But we'll see next time that that's not how history played out. We'll see you then. (calm music) |
Power_and_Politics_in_Todays_World | Lecture_14_911_and_the_Global_War_on_Terror.txt | - Okay, let's go down memory lane. - The topic today is an adversary that poses a threat, serious threat to the security of the United States of America. This adversary is one of the world's last bastions of central planning, governs by dictating five-year plans. With single capital, it attempts to impose its command across time zones, continents, oceans, and beyond. With brutal consistency, it stifles free thought and crushes new ideas. Perhaps this adversary sounds like the former Soviet Union, but that enemy's gone. The adversary's closer to home. It's the Pentagon bureaucracy. We must transform the way the department works and what it works on. Our challenge is to transform not just the way we deter and defend, but the way we conduct our daily business. We know the adversary, we know the threat, and with the same firmness of purpose that any effort against a determined adversary demands, we must get at it and stay at it. Some might ask, how in the world could the Secretary of Defense attack the Pentagon in front of its people? To them, I reply I have no desire to attack the Pentagon. I want to liberate it. We need to save it from itself. And that means we must recognize another transformation, the revolution in management, technology, and business practices. Successful modern businesses are leaner and less hierarchical than ever before. They reward innovation, and they share information. They have to be nimble in the face of rapid change, or they die. Business enterprises die if they fail to adapt. And the fact that they can fail and die is what provides the incentive to survive. But governments can't die, so we need to find other incentives for bureaucracy to adapt and improve. Already, we've made some progress. We've eliminated some 31 of the 72 acquisition-related advisory boards. We now budget based on realistic estimates. We're improving the acquisition process. We're investing $400 million in public-private partnerships for military housing. Many utility services to military installations will be privatized. We have committed $100 million for financial modernization, and we're establishing a Defense Business Board to tap outside expertise as we move to improve the department's business practices. To transform the department, we must look outside this building as well. Consequently, the senior executive council will scour the department for functions that could be performed better and more cheaply through commercial outsourcing. Like the private sector's best-in-class companies, DOD should aim for excellence in functions that are either directly related to war fighting or must be performed by the department. But in all other cases, we should seek suppliers who can provide these non-core activities efficiently and effectively. We have brought people on board who have driven similar change in the private sector. We intend to do so here. The old adage that you get what you inspect, not what you expect, or put differently, that what you measure improves, is true. It is powerful, and we will be measuring. - So that was Secretary of Defense Donald Rumsfeld in a speech at the Pentagon. As it turned out, on the 10th of September, 2001. (audience laughing) And as you can tell, he's elaborating on the agenda that we've already discussed at length in this course about privatization of military functions without any idea about what's about to hit the country. He's not only talking about outsourcing and privatizing, but he really came into the Pentagon with a mandate to completely reorganize it along management principles that were deployed in the private sector. And he was busily setting about that agenda in the first months and first year of the administration. But, of course, 10 days later, America was very differently preoccupied. - Our enemy is a radical network of terrorists, and every government that supports them. (audience applause) Our war on terror begins with Al-Qaeda, but it does not end there. It will not end until every terrorist group of global reach has been found, stopped, and defeated. (audience applause) These terrorists kill not merely to end lives but to disrupt and end a way of life. With every atrocity, they hope that America grows fearful, retreating from the world and forsaking our friends. They stand against us because we stand in their way. We have seen their kind before. They are the heirs of all the murderous ideologies of the 20th century. By sacrificing human life to serve their radical visions, by abandoning every value except the will to power, they follow in the path of fascism, Nazism, and totalitarianism, and they will follow that path all the way to where it ends, in history's unmarked grave of discarded lives. (audience applause) Americans are asking, how will we fight and win this war? We will direct every resource at our command, every means of diplomacy, every tool of intelligence, every instrument of law enforcement, every financial influence, and every necessary weapon of war to the disruption and to the defeat of the global terror network. Now, this war will not be like the war against Iraq a decade ago with a decisive liberation of territory and a swift conclusion. It will not look like the air war above Kosovo two years ago where no ground troops were used and not a single American was lost in combat. Our response involves far more than instant retaliation and isolated strikes. Americans should not expect one battle but a lengthy campaign unlike any other we have ever seen. It may include dramatic strikes visible on TV and covert operations, secret even in success. We will starve terrorists of funding, turn them one against another, drive them from place to place until there is no refuge or no rest, and we will pursue nations that provide aid or safe haven to terrorism. Every nation in every region now has a decision to make, either you are with us or you are with the terrorists. (audience applause) - That was President Bush's statement to the nation, the 20th of September, nine days after the 9/11 attacks, and it denounced the advent of what we have come since to refer to as the Global War on Terrorism, or the GWOT. And today, this agenda is gonna be to explore that. We're gonna talk about what was the GWOT, and what is it? What has it become? We're gonna ask how radical an innovation was it, and what did it mean for the emerging post-Cold War international order that we talked about in the very first lecture of the course? And finally, I want us to also think about as we go through this material, what were the paths not taken? What might have been done differently that might leave us in a different position than we are today? Let's think back to the 1991 Gulf War that I lectured about at the beginning of the course. And particularly, George Herbert Walker Bush, Bush 41, as some like to call him, what strategy he pursued when Saddam Hussein invaded Kuwait in December of 1990. And this strategy of pursuit had a number of elements. One was that it was fully authorized through the UN Security Council. Resolutions were passed admonishing Saddam Hussein to leave Kuwait, which he ignored and eventually stonewalled but refused to comply with. Secondly, the Bush 41 Administration put together a very large and diverse coalition, which included participants with many different geostrategic interests, including every Arab nation except Jordan participated in that coalition. They stuck to the limited authorized goal of expelling Saddam Hussein from Kuwait and made no effort at regime change in Iraq. This is something for which Bush was heavily criticized by neoconservatives in the US. We'll hear more about neoconservatism next week. And many people thought he was being too much of a wimp and that once you had half a million troops there, that he should have gone all the way to Baghdad and interestingly, Dick Cheney, who was in that administration, came out as a strong defender of not doing that, partly on the grounds that we wouldn't know what, we wouldn't be able to manage the system, the situation thereafter, this philosophy that if you break it, you own it, which, again, we will attend to more next week. But it was this basic idea that Bush was operating on of, that I like to call stop the bully without becoming a bully. Then the notion was that if people engage in aggression across borders, the international community will stop them. But it was consistent with the UN Charter's commitment to the maintenance of international peace and security, which is a line in the UN Charter, which we saw has come to be questioned subsequently with the implementation of responsibility to protect, for example. But at the time, it was seen as a sacrosanct commitment of the UN, essentially to protect the status quo of the division of the world into nations. And so you stop aggression, but you don't engage in regime change, and you don't rearrange the global geography. Over this and the next two lectures, I want people to think about one counterfactual. This is not the only one, but what if that had become the template for facing down international aggression in the post-Cold War world? And that was clearly part of Bush 41's agenda. He wanted to etch these norms into accepted common law practice if you like, of international conduct and also the notion that they should all be mediated through the United Nations. Another way of putting it is that he had a kind of containment sensibility, George Herbert Walker Bush. The idea was to contain aggression, not to solve the world's problems, not to end bad regimes, but to contain aggression, to stop aggression. And of course, the word containment has a long history. It goes back to the Cold War, and I wanna take us back through that history because it's really only by having some sense of that history that we can evaluate the extent of the departure that comes with the Global War on Terror a decade after Bush, Sr. was doing this. If we think back, the idea of containment, of course, was articulated by George Kennan in 1946 in something which was called the long telegram. He was in Moscow at the time as part of the Truman Administration. And it was published subsequently anonymously with the authorship X as a document in "Foreign Affairs" the next year, it called the Sources of Soviet Conduct. And this is the classic statement of the original idea of containment. And it had a number of elements to it, the original doctrine of containment, as conceived by Kennan. Kennan, first and foremost, he was a staunchly anticommunist, very conservative and indeed held views about race and other matters which would not endear him to many people in this room. Be that as it may, this was his view of the Soviet Union. He thought it was, first of all, fundamentally antithetical to Western democracy, and he thought the enterprise of getting involved in public debates with the Soviets about capitalism versus communism and the desirability of different systems, was a total waste of time. There was no possibility of dialogue, and he didn't advocate, he thought it was just pointless and indeed, might give propaganda victories to the Soviets in the developing world during the Cold War. But secondly, he was absolutely unequivocally convinced that the Soviet Union was unsustainable, that it was a flawed economic system for the reasons that we know well, that command economies don't really work, and that eventually it would develop the kinds of problems that it eventually did. But then secondly, and actually in this, he was stating something that has since become conventional wisdom among political economists who study international relations, he was convinced that the Soviet Union's hegemonic ambitions would bankrupt them, that as they sought to become a hegemonic power, they would be spending more and more sustaining that hegemony, and eventually it would drive them to ruin. And as I say, it is conventional wisdom now in the international relations field, made famous by people like Stephen Krasner, Robert Keohane, and others that hegemonic powers become overextended and eventually find it very difficult to sustain the financing of their hegemony. Kennan was contemptuous of all political science and social science. He wouldn't put it in those terms, but his seat-of-the-pants judgment was that this was a system that was just gonna drive itself off a cliff. Then what should the US do? And here, Kennan's view was, well the US should forget about the governments in the Soviet Union and really focus its attention on the populations behind the Iron Curtain, essentially to win the battle for hearts and minds. The thought was that as these systems became more and more dysfunctional, people would be more and more dissatisfied and then it would be important, first of all, for them to see successful models elsewhere. And so he thought it was hugely important to rebuild the American domestic economy at home and to invest in rebuilding Western Europe and Japan, so a huge supporter of the Marshall Plan, no doubt partly because of commitment to the West European countries, but from the point of view of his containment doctrine, the idea was that as the populations behind the Iron Curtain experienced increasing levels of poverty and disaffection from their regimes, they would see on television and learn about in other ways, the thriving democratic systems in the West. The other side of that coin was he very much wanted to admonish, first and foremost, the Truman Administration, but subsequent American administrations and Western governments, not to start to become like the Soviet Union. We needed to not behave in ways that would make us seem to be their moral equal. We should, therefore, maintain civil liberties at home and in our international conduct. We shouldn't engage in adventurism. Essentially, our goal should be about containing Soviet aggression. And he thought the principle means of doing that would be economic and diplomatic. He was opposed to the formation of NATO, for example. He thought creating NATO would just militarize the confrontation between us and the Warsaw Pact, and if we created NATO, the Soviets would do something comparable on the other side, which they did. He opposed the Vietnam War. This is much later, of course, but again, on the theory that this is about something more aggressive than containment. The Vietnam War was motivated by a domino theory that if Vietnam was allowed to fall, then the whole of the rest of Southeast Asia would go communist, none of which turned out to be true. And he was, part of the reason that he was opposed to Vietnam War was quite prescient, which is that if, in his mind, if you are fighting a war of choice against an enemy for whom it's an existential war, unless you can have a very quick, cheap, easy victory, as when Ronald Reagan invaded Grenada or something of that order, basically the cards are stacked against you because you're not gonna be able to sustain support at home for a war of choice. Whereas your adversary isn't fighting for their very existence, and so they will be able to sustain support. So he was very skeptical of conflicts like the Vietnam War, which he thought was a giant mistake. That was classical containment doctrine as articulated by Kennan. And it didn't get very far. Indeed, he rapidly got into conflict with the Truman Administration. And the Truman Doctrine, as I've already intimated, was revolved around militarizing the standoff with the Soviets in ways that Kennan didn't like. And this was laid out in a document called the NSC 68, published in 1950. Many scholars call it the most important US strategic statement of US principles during the Cold War, and it was essentially laid out what's come to be known as the Truman Doctrine. And it supported the creation of NATO, and it supported a much more aggressive stance. Now, some scholars still call it containment, so John Lewis Gaddis in his excellent book "Strategies of Containment", calls this a different kind of containment. And I don't have anything invested in the terminology, but it's a much more militarized version of containment than Kennan originally envisaged. For Kennan, military force should always be a strategy of last resort, not a strategy of first resort. You try everything else first, and you only use it, you use it sparingly to stop aggression when you have to. Even during the Truman Administration, Kennan had all this conflict, and eventually he stormed off in a huff to the Institute for Advanced Study at Princeton where he spent the next several decades, occasionally throwing spitballs at one administration after another. He became a grumpy old man of sorts. But, of course, the Truman Administration was just the beginning of the Cold War. And then we had the Eisenhower Administration, and the Eisenhower Administration took a more militaristic view of US foreign policy than even the militarized version of containment that was embedded in the Truman Doctrine. And so, the Eisenhower comes to power in 1952. The Cold War's ramping up. McCarthyism is ramping up. And proponents of containment get accused of being proponents of appeasement, that this is all about appeasement. And, of course, appeasement is actually giving in to foreign aggression in the way that the British and French did when Hitler was beginning his expansionist policies and then accepting them. Containment is not, it's not that. Containment says you do respond to actual attempts at aggression. But short of that, you don't involve yourself in trying to rewrite global political geography. And so, John Foster Dulles who was the John Bolton of his era, you might say, he was the proponent of rollback. And rollback was to not accept the disposition, particularly of Europe, after the Iron Curtain came down, and that he thought we really needed to push back against the Soviet Union much more aggressively than any notion of containment would suggest. And indeed, Eisenhower ran on rollback in the 1952 election saying that they would take a much more aggressive, anticommunist stance than the Truman Administration had done. Interestingly enough, of course, at that time, the Truman Administration was bogged down in the Korean War, which Eisenhower promptly negotiated a settlement to after not very long in office. But, in any event, they ran on rollback, and, of course, something like rollback was became a much more dominant theme if we think about the Johnson Administration during Vietnam and Nixon. Nixon's strategy became more complicated than just straightforward rollback because he wanted to, actually in true Kennanesque fashion, he wanted to exploit divisions among communist regimes. That's part of what the opening to China was about, but nonetheless, in prosecuting the Vietnam War, at least as he did and in the era before detente, he was something of hawk with respect to the Soviet Union. That was rollback, and interestingly, when we think about, we look back on the Cold War, despite what Eisenhower ran in in 1952 and what Dulles came into the State Department attempting to do, he found a lot of pushback against rollback in the State Department and found that many people were actually more convinced, if not of Kennan's version of containment, at least they were convinced of the Truman Doctrine's version of it. In Europe, the US largely did stick to containment as its policy. They didn't really their to rollback the frontiers of the Soviet empire. And if you look at where they did try to rollback the frontiers of the Soviet empire, such as in Iran where there was an election in 1953 and a pro-Soviet government was elected, we went in and toppled that government and basically implanted the Shah. Didn't turn out that well. We engaged in rollback in Vietnam. Didn't turn out that well. By the end of the Cold War, containment and indeed Kennan was given a lot of credit as a policy that had actually worked. Of course, where that containment would have worked in the absence of NATO is a question to which we'll never have an answer. Whether it was the Kennanesque version of it or the Truman Doctrine version of it, nonetheless, it was given great credit, partly because the early predictions had turned out to be right. The economy was unsustainable, and the Soviet Union did indeed become bankrupt, partly because of Reagan's Star Wars, which spent them, created not just an arms race for space weapons, but a dollars race. And then, of course, very relevant to today's discussion, they got embroiled in an unwinnable war in Afghanistan, which turned out to be the Soviet's Vietnam, interestingly, also motivated by a domino theory. The Soviet belief was that if the Mujahedin succeeded in Afghanistan, this would allow Islamic fundamentalism to start working its way through the Asiatic republics of the Soviet Union, and it had to be stopped. Of course, when the dominoes finally started falling, it was in Western Europe, and it had nothing to do with Afghanistan. George Herbert Walker Bush's policy in 1990, the first Gulf War, is really a version of containment as mediated through international institutions. Kennan was a hard-boiled realist, and so he had no time for the UN and part of the reason he had no time for NATO, by the way, was, again, his hard-boiled realism, he said, "No nation is gonna bind itself by an alliance, "if it's not in their interest to do it." It's either gonna be irrelevant when it's needed or window dressing when it's not. But he also thought no nation was gonna pay any attention to the UN. So that aspect of Kennan's thinking is not part of the post-Cold War containment idea that Bush was working to entrench in the way that he responded to the first Gulf War in 1991. Now, let's come to Bush 43's Global War on Terror and what came to be called the Bush Doctrine very rapidly. And we can see what we think about how it contrasts with what had gone before. The first feature of the Bush Doctrine is that he envisages it as being worldwide in scope. He says, "We will strike anywhere, "and we reserve the right to strike anywhere "if there is a threat to our national security." Now, of course, America's assertion of unilateral rights to strike perceived threats that were not to the American mainland has a long history. In fact, goes back to 1823 when John Quincy Adams articulated what came to known as the Monroe Doctrine; although, it didn't actually get named the Monroe Doctrine until some decades later in 1850. But it was the Monroe Administration asserted that any European interference anywhere in Latin America would be conceived of and responded to as a threat to the national security of the United States. And so, essentially, during the Monroe Doctrine said that we control the hemisphere, and anybody who interferes in the hemisphere is interfering with America's national interests, and we will respond with force. That was aimed at the Europeans in the 19th century, but it is the precedent, if you like, to the extent there is a precedent for George W. Bush asserting, we will act militarily anywhere in the world if our national security is threatened. And indeed, interestingly, he talked about the Axis of Evil, Iran, Iraq, and North Korea. And of course, in those days, Iraq was not an Islamic administration. It was a secular Baathist regime, run by Saddam Hussein. Iran was an Islamic state, and North Korea was communist state. And so he was, by putting those three together as the Axis of Evil, he was doing two things. One was he was saying, and this was the Bush Administration's position until 2006, that is not about Islamic fundamentalism. That terrorism is, and he was very careful not to attack the idea of Islamic fundamentalism. The War on Terror was not just about Islam. It changes that rhetoric after 2006, but also, that we would act anywhere in the world. And if you think about Phnom Penh, it's 12 time zones away from east coast United States, so the only way you could get further away would be by blasting off into space. (audience laughs) Now, we've taken the Monroe Doctrine worldwide. Secondly, we asserted the right to act unilaterally, unconstrained by any existing alliances and unconstrained by what international institution might or might not authorize. And indeed, this is where he coined the phrase, coalitions of the willing. We will put together coalitions on a opportunistic basis if necessary to fight our adversaries. And indeed, as I mentioned to you in an earlier lecture, one of the reasons so many East European countries were added to NATO in the 2004 and 2005 was it was the quid pro quo for their promise to participate in the Iraq invasion that I'll be talking about next week. We can act unilaterally with allies that are willing to form coalitions with us. We're not constrained, either by alliances or by international institutions. Third, he asserted a right to engage in preemptive war. The traditional justification of war was the existence of an imminent threat. Word imminent has a long history. It's partly connected to just war theory, which says that an imminent threat is necessary for war to count as a just war. And if you go back and actually read the NSC 68, the statement of the Truman Doctrine published in 1950, it asserts there that the US will never engage in preemptive war, that it's only gonna act defensively. Very quickly, the Bush Administration said, no. We are not gonna be constrained by the doctrine of imminent threats. And they started talking about things like emerging threats, gathering threats, would be sufficient to justify American military action. And of course, if you think about that doctrine, a doctrine of emerging threats, you then are almost inevitably also gonna get away from the idea that we're only getting involved in wars of existential survival. You're gonna be back in the world of optional wars because there's no clear sense in which emerging or gathering can be neatly calibrated or that you can convince your citizenry of the importance of an emerging or a gathering threat. That's part of what we have with global warming. It's a gathering threat. It's an emerging threat. You can't, much harder to mobilize people for a chronic problem, if you like, than a crisis. Eminent war was thrown out of the window in favor of these notions of gathering threats, emerging threats, and so on. Again, we can engage in preemptive wars against emerging or gathering threats. In that Bush doctrine, as he articulated it, first at West Point commencement address in June of 2002 and then reiterated it in two national security statement doctrines that the Bush Administration published, one later that year and then reaffirmed in 2006. They said, a legitimate war aim was now regime change. A legitimate war aim was to get rid of an oppressive dictatorship. Completely out of the window, the UN notion that international action is legitimate if its goal is to restore the status quo ante of an international borders, but simply if a regime is sufficiently oppressive, vis a vis its own population, this will be a justification for military action. And the reason they articulated this so early was they already had their eye on Iraq. It was very clear, for reasons we'll get to next week, that from the beginning, the administration was committed to removing Saddam Hussein from power even before we had started the battle in Afghanistan. A fifth element of the Bush Doctrine was, again, a very radical departure, which was that there's no possibility of neutrality. As he put it at the end of the clip I showed you, "if you're not with us, you're with the terrorists." And of course, there had been a long tradition of neutrality in international relations. If you go back to the US in the 19th century, we had often asserted neutrality, and if you come even into the 20th century, if you look in the run-up to World War II, Congress was passing neutrality legislation hand over fist to keep FDR out of the war with Germany in the late 1930s. So there was a long tradition of the US asserting the right to be neutral in international conflicts and during the Cold War, the non-aligned nations had declared themselves to be neutral in the conflict between the US and the Soviets. Once we get to the Bush Doctrine, all of that also goes out of the window. Essentially, we're now declaring to the world that if you do not support us, we will feel free to interpret that as support for our adversaries, and we will act accordingly. Finally, the Bush Doctrine, as he indicated in that clip, envisages, if not permanent war, a war that we can never really say has finally been won. We're declaring war on terror. Well, there's been terrorism in the world since the beginning of time. And so the idea that you're now gonna be on a war footing so long as there is terrorism in the world, you're essentially saying the US economy is gonna be on a permanent war footing. We're gonna be engaging in what have since come to be called endless wars into the indefinite future. Some have said that the Bush Doctrine was not that radical of an innovation, but clearly it was. It was much more radical than anything that had gone before. And it, of course, what it did was as it has played out, and we'll see this in the next two lectures, it completely rubbished his father's attempt to re-embed some notion of containment for a post-Cold War era into international politics, and it's one reason his father was very critical of him and of Cheney and of Rumsfeld for, if you like, scuttling his legacy. And this came out in a subsequent biography. In effect, the Bush Doctrine, if you wanted a bumper sticker for it, is rollback of all possible adversaries at any time and in any place. The US now asserts the right to do that. It was about as radical an innovation as one could have come up with. And certainly, the most radical innovation in US national security policy in the history of the republic. Now, let's turn to the Afghan conflict. One question is, well, why was Al-Qaeda based in Afghanistan in the first place? And you already know. Why were they there? They just popped up. Osama bin Laden had a long and antagonistic history of not very good relations with his Saudi government. He was a Saudi Arabian citizen, and he had become radicalized and had, in the 1980s, worked with the Mujahidin in Afghanistan, helping them battle the Soviets. And when the Soviet Union finally collapsed, he had tried to mediate an end to the ongoing war in Afghanistan among the militias, not with much success. He had been bounced around the Middle East and eventually wound up in Sudan. During the invasion of Kuwait by Saddam Hussein that I was talking about some minutes ago, he was strongly of the opinion that the Saudis, first of all, in that war, part of the reason there was so much anxiety about Saddam Hussein's invasion of Kuwait, was people thought Saudi Arabia will be next. And Saudi Arabia is a huge supplier of oil to the West, so there's both the Saudis themselves and others is part the reason there was Arab participation in the coalition to get Saddam out of Kuwait. Bin Laden said that on no account should the Saudis accept Western participation in expelling Saddam Hussein from Kuwait, and if they did, they would be inviting the infidel into the heart of the Middle East. And they ignored him, and then he kept up a steady stream of attacks on the Saudi regime in print and verbally and was increasingly suspective also of even financing terrorist attacks in Saudi Arabia from his bases in Sudan and elsewhere that he moved around to. The Saudis finally, they first banished him from the country and then stripped him of his Saudi citizenship. I think that was in 1994 they had stripped him of his citizenship, and eventually, they started putting pressure on Sudan to kick him out of Sudan because they thought he was, an ambassador to the Americans think he was using this as a base for mounting terrorist attacks, which were ramping up over the 1990s. Finally, the Saudis agreed and everybody agreed that they would leave him alone if he would go to Afghanistan, and there was a sort of tacit acknowledgement that he would not foment terrorism within Saudi Arabia from that base. But as Al-Qaeda became increasingly radicalized, we should remember that their two primary war aims were, first, to overthrow the Saudi regime, and secondly, to get the West out of the Middle East. It was what they, at that time, called a defensive jihad, rather than an offensive jihad, which they had subsequently embraced. That's why he was in Afghanistan. And that's why the 9/11 attacks were plotted from Afghanistan. He had good relations with the Taliban regime of Mullah Omar. They had worked together there in the 1980s. An important thing to say about the Afghan war is we very early on made a decision to pursue regime change. Basically, what happened was we went to the UN, we said, we got resolutions passed saying the Taliban must turn over Osama bin Laden and close down the training camps. They refused, and we essentially said, we're gonna wipe you out. We're gonna take over your regime. But, and here, this is part of the reason I showed you the Rumsfeld clip at the beginning of the lecture, it turns out that when Rumsfeld said, we'll contract out everything except core military capacities, that was not an accurate statement because the Bush Administration was certainly not interested in growing the military. Indeed, Rumsfeld was shrinking the military. And they had no intention of changing that. They wanted to attack Afghanistan, and they wanted to do it largely by contracting out, not to military contractors as we talked about in the privatization lecture, that would come later. But they thought the way to do this since Afghanistan was in the midst of a civil war, the way to do this was to get behind one of the belligerents in the civil war and let them take over the country, and then they will be friendly regime. The belligerent in question was a group called the Northern Alliance, and the Northern Alliance was losing the war. They had been pushed into a small corner of the country, and but for us, involvement with them and support for them, they would almost certainly have been defeated in relatively short order. What we did was we got behind the Northern Alliance and supported them in the Afghan civil war, which led then to the destruction of the regime, which was in two important respects, a setup for failure for two reasons. One is the theoretical expectation, maybe it's dressing it up too much to call it theoretical, but at least, one should have skeptical priors. If you get behind the losing side in a civil war, the odds are its gonna be very difficult for them to govern if you put them in power because one of the reasons they're the losing side in a war is they probably don't have much grassroots support on the ground. And we will see this play out again in Libya and elsewhere in upcoming lectures. There were good reasons for skepticism of the idea that this was a good way to put in a puppet government that would be friendly to the US and be able to govern the country without massive infusions of support. And, of course, nobody anticipated for a second that we would make Marshall Plan type of investments in Afghanistan in 2001 and 2002. The notion we were gonna sustain the rebuilding of this country was not on anybody's scanner, and indeed, as we have subsequently tried to do it, that has been a dramatic failure. There was no postwar planning of any kind. This was in the heat of the weeks after 9/11 and the notion was that this horrific evil has been perpetrated against the United States, and we're gonna go there and take care of business! I mean, that was the dominant rhetoric behind this invasion. One question to ask is, might we have done any of this differently? Was there a path not taken here? Clearly, we, as I said in the chapters I had you read today, the idea of responding to 9/11 through the lens of the criminal justice system was a complete nonstarter given the scale and scope of the attacks, and there was gonna be a military response. And it was also clearly the case that once the Taliban regime refused to be cooperative, there was gonna be military action. Well, there was a path not taken advocated by some at the time. And this is summarized by Lawrence Wilkerson who was Colin Powell's chief of staff. In an interview in "The Cairo Review" five years ago, he said, "The smart strategy, "and there were those who advocated this at the time," and he's correct about that. You can find it in the archival evidence. "Was to go in, pound the hell out of Al-Qaeda, "try to get bin Laden and his leadership. "If you didn't, so be it. "You'd keep hunting. "You'd do it mostly with special ops forces and the CIA. "You would pound the Taliban a little bit, "and as you left you'd tell them, "'Do it again, and we'll do it again.' "This was very formidable, persuasive, "strategic brief, and the president rejected it." Then he says in a rather mean aside, "Well, really the vice president rejected it." Because it was common knowledge that Cheney was really the brains behind and certainly the energy behind the war as it was being fought. What would have been the advantages of such an approach? Essentially you would ignore the Taliban. You would say, we've been attacked by these people, and we're gonna go after them, and we are going to destroy them. And if you interfere with us, good luck with that because we will beat you up, too. And they need not have engaged in regime change, particularly if you're thinking about a world in which you're gonna be acting with a light footprint and wanting to keep your forces available for deployment anywhere, seems like it might be a way to go. Obviously, first of all, much cheaper, vastly cheaper than replacing a regime. You finance your operation, and then when you leave, you leave with your operation. Secondly, it avoids the you break it, you own it problem. If you create a failed state, you are then saddled with a failed state, and you've gotta administer a failed state or at least support some side that's trying to administer a failed state. By toppling the regime, we took on a much bigger set of problems. Now some might say, well, what about the costs in terms of deterrence, that you want other players around the world to know that if they support terrorists or allow terrorists in their country, they're gonna be wiped out, too? That might have been some of the thinking, but I suspect the real reasons it was not done, first of all, the raids after 9/11, the idea that in the weeks and months following that attack, the notion that Mullah Omar's regime was thumbing their nose at the US, was just too much to bear and too much politically to tolerate, as well as hubris about what could be achieved in Afghanistan. And here, we tend to repeat the mistakes of history. You didn't have to go back to the 19th century and Britain's failed attempts to subdue Afghanistan. The Americans had watched in real time for a decade as the Soviets had been ground down in the Afghan mountains. It's about the worse possible terrain in the world for military campaigns. It's not really surprising perhaps that we didn't pursue that path, but certainly, there were people on the ground advocating it, and it's something that we didn't do. What did we do, apart from toppling the regime? What did we do in going after bin Laden? And what we did came to a head in December, in northeastern Syria in what, in northeastern Afghanistan, in these, very high up in the mountains in a remote area called Tora Bora, the Battle of Tora Bora, which was conducted over the middle of December of 2001. And here's some commentary from a key player at the Battle of Tora Bora. - [Scott] Following the strategy of keeping an Afghan face on the war, Fury's Delta Team joined the CIA and Afghan fighters and piled into pickup trucks. They videotaped their journey to a place called Tora Bora. Fury told us his orders were to kill bin Laden, and leave the body with the Afghans. - Right here, you're looking at, basically the battlefield from the last location that we had firm on Osama bin Laden's location. - This ridge line, is it about 14,000 feet? And back this way toward me is Pakistan. - [Dalton] That's right. - On a scale of say, one to 10, 10 being the toughest, how tough a position is this to attack? - (laughs) In my experience, it's a 10. - [Scott] Delta developed an audacious plan to come at bin Laden from the one direction he would never expect. - And we wanna come in on the back door. - You were gonna come up over the tops of the peaks? - That's right, the original plan that we sent up to our higher headquarters, Delta Force wants to come in, over the mountain with oxygen, coming from the Pakistan side, over the mountains and come and get a drop on bin Laden from behind. - Why didn't you do that? - Disapproved at some level above us, whether that was central command, all the way up to the President of the United States, I'm not sure. - [Scott] The next option, Delta wanted to employ was to drop hundreds of land mines in the mountain passes that led to Pakistan, bin Laden's escape route. - First guy blows his leg off. Everybody else stops. That allows aircraft overhead to find them. They see all these heat sources out there. Okay, there's a big large group of Al-Qaeda moving south. They can engage that. - Why didn't you do that? - Disapproved. - [Scott] Why was it not approved? - I have no idea. - How often does Delta come up with a tactical plan that's disapproved by higher headquarters? - In my experience, in my five years at Delta, never before. - The military wouldn't tell us who rejected the plans or why. Fury wasn't happy about it, but he pressed on with the only option he had left, a frontal assault on bin Laden's dug in Al-Qaeda fighters. The Delta Team had only about 50 men, so the mission would depend on the Afghan militia as guides and muscle. Their leader was a warlord, a self-styled general named Ali. Ali, second from the left, met with this CIA officer and accepted millions of dollars in cash from the agency. Dressed like Afghans, the Americans maneuvered up the mountains, calling in airstrikes on Al-Qaeda. By day, they would advance, but at night, they soon discovered that their Afghan allies went home. Well, I have to assume that if you started up the hills of Tora Bora, and the Mujahedin took territory, they didn't abandon that at night. - Oh, yes, they did. - They gave it up to the enemy? - Absolutely, the Mujahedin, would go up, get into a skirmish, a firefight, lose a guy or two, maybe kill an Al-Qaeda guy or two, and then they'd leave. It was almost like it was an agreement, an understanding between the two forces fighting each other, almost, put on a good show, and then leave. - [Scott] In the morning, bin Laden was on the radio. The CIA, Delta, and their Afghan allies were listening. How did the Afghans react when they heard from Osama bin Laden on the radio? - Osama bin Laden is many a Muslim's hero. These guys, in my opinion, were more in awe of Osama bin Laden than they were willing to kill him. When they heard him talking on the radio, they would gather around the individual that handheld that transistor. He would hold it up in the air, almost as if he didn't want the connection to break, and almost like they could see the ridge line Osama bin Laden happened to be talking from, like they could almost see him and feel his presence. And they just stood there with wide eyes and somewhat in awe that here is a leader of the jihad, the leader of Al-Qaeda, and they're actually hearing his voice over the radio. - And these were the men who were supposed to help you capture or kill him? Some allies. And then, something extraordinary happened. Fury's Afghan allies announced that they had negotiated a ceasefire with Al-Qaeda, something the Americans had no interest in. When Fury's team advanced anyway, his Afghan partners drew their weapons on Delta. It took 12 hours to end the bogus ceasefire, precious time for Al-Qaeda to move. You gotta figure he's heading for the valley and to pass into Pakistan. - Our assumption is he's going for the valley at that time. - [Scott] Bin Laden had changed direction and the tone of his radio calls. - Clearly under duress, clearly hurting, clearly caring for his men. - [Scott] Inside this building, the American team listened to bin Laden on the radio. Fury wrote down the translation in a notebook. - Quote, "Our prayers were not answered. "Times are dire and bad. "We did not get support from the apostates "who are our brothers. "I'm sorry for bringing you here. "It is okay to surrender," end quote. - When you heard that, what did you think? - I thought, it's almost over. - So Dalton Fury is a pen name. He has a book, actually several books called "To Kill bin Laden". He was the major in charge of the Delta Forces in the Afghan operation. And you can read his book if you want the detailed account. But basically what happened at Tora Bora has been much in dispute. An important document which gives a lot of the blow by blow is this Senate report, foreign relations Senate report published in 2009, which documents a lot of what Fury said in that clip and that basically that the US, it's not the only thing, by the way. Some might be skeptical of this if you see who was chairman of the committee at that time. And he had failed in his 2004 bid to unseat George W. Bush; though, he was not involved in writing this report. But another, around the same time, an academic, you can read this abstract at your leisure when I post the slides, basically they said there was an alternative path not taken, not just the two strategies that Fury was suggesting, feeding up the command that didn't wanna hear him, but that there were even within the ceiling of 10,000 troops, even within that ceiling, there were battle plans to go into the mountains and at least seal off the escape to Pakistan, and they didn't do it. Instead, there was a lot of hand-wringing about whether, first of all, there were claims that we didn't know he was there for sure, which turned out to be false. The CIA did know that he was there for sure. And then they thought that maybe he had been killed, and they spent some time digging up for bodies. But he hadn't been killed. The strategy had essentially been to rely upon these militias whose loyalty was very much in question. And of course, we shouldn't be surprised by that, not only for the reasons Dalton Fury gave in that video, that they might have had much stronger allegiance, many of those people to bin Laden than to the prospect of killing him, but also because we know from the literature on civil wars, here the state of the art is Stathis Kalyvas's work, used to be at Yale, now at Oxford, but civil wars are never just about ethnic conflicts. There are all kinds of other local interests served. And as we saw in the Host Nation Trucking case we talked about some time ago, people are being paid and making money in these kinds of settings. And so it's very unlikely that you're gonna be able in a setting like this to outsource the fighting against somebody like bin Laden. What explains the American strategy? Well, one is that we wanted to keep the troops to a minimum. As I said, there have been subsequent studies done, both by the Senate Foreign Relations Committee and academic studies which make it believable, at least obviously there's always risk in any military operation, but there's a colorable case that the US could have gone in with troops that were in the region and prevented bin Laden's escape. It seems like a principal reason was that what has become so important in American wars abroad, they wanted to keep the number of US casualties to a minimum. As you saw, there were only some 50 Delta Force people up there in the mountains working with the Afghans. They did not want Americans to start hearing about dead soldiers in Afghanistan. But then, secondly, it is also clear from that report that they were already in December of 2001, President Bush was asking Rumsfeld for a battle plan to go into Iraq and remove Saddam Hussein from power. And they did not want to get a lot of American troops embroiled in any kind of conflict in Afghanistan. So the thought was they were gonna be redeploying whatever assets they had in Afghanistan as soon as possible to Iraq. Now, you might think about, well, what were the costs of this strategy? One was the obvious, immediate cost, that bin Laden and Al-Qaeda escaped through the mountains to Pakistan, which gave them a huge propaganda victory against the US, that the most powerful nation on Earth, in fact, couldn't chase down this terrorist group in the mountains of Afghanistan. But secondly, also because we had such a weak footprint, actually Mullah Omar himself escaped on a motor bicycle, motorcycle from Afghanistan into Pakistan and would come to lead the Taliban insurgency to retake the country from the new regime in Kandahar, so much so that within a year, American generals on the ground were saying up the chain of command, you should negotiate a deal with the Taliban because the new regime's not gonna be able to survive. And here we are 18 years later, and I believe we're on the verge of doing that. But they were saying it as early as 2002. And finally I think that what the Afghan operation did do, though it was gonna take some time for this to become clearer, it made it obvious the that gap between the aspirations for and the capacities behind the Bush Doctrine were becoming, it should say doctrine on that slide, the Bush Doctrine, were coming clearly into view, because if you think about the massive expansive reach that we proclaimed, we'll wage preemptive wars against gathering threats anywhere on the world at any time with any allies, at the same time, as we're shrinking our military capacities, extremely unwilling to get American soldiers killed and wanting to work through proxies, that this is not gonna end well. It didn't end well in Afghanistan, and next Tuesday, we'll see how it ended in Iraq. See you then. (people milling) (peaceful music) |
Power_and_Politics_in_Todays_World | Lecture_16_Denouement_of_Humanitarian_Intervention.txt | - Good morning, everybody. So today's lecture is called The Denouement of Humanitarian Intervention, and it's mainly gonna be focused on Libya, though we'll see, we'll look elsewhere in the Arab Spring as well. Let's start at the end, though. This is Libya late last year. - So that was the situation in late 2018. If anything, it's worse today. There was supposed to be elections held which couldn't be held because of the fighting and Haftar has mounted successive assaults on Tripoli, and so the country is in a continuing stalemate failed state. So that tells you what happened, it didn't tell you anything about why it happened and what we should infer from why it happened, and that's, the way into today's agenda. We're gonna ask how and why it happened, and then we're gonna look at the implications for the prospects for humanitarian intervention efforts going forward, the stability and future shape of the Middle East, and the ability of the US to be a force for good internationally and how that has been impacted by this, what's happened over the last nine years since we begun that action. But let's begin by picking up the thread of responsibility to protect that I lectured about a couple of weeks ago. And you'll remember that this was all in the shadow of Rwanda, and then Kosovo. The NATO forces had gone into Kosovo in 1999 without UN authorization when no NATO country had been threatened, and this has produced a lot of hand-wringing in the international community that led the Prime Minister of Sweden, don't know by what legitimacy he had to do this, but he appointed an international commission which concluded that the NATO action had been illegal but legitimate. We had considerable debate about what that might mean, and they did at the time as well. It was eventually followed in 2005 at the UN by the adoption of something called the UN World Summit Outcome Document where they famously announced that each individual state has a responsibility to protect its populations from genocide, war crimes, ethnic cleansing, and crimes against humanity. They took it upon the UN to be the enforcer of this and said, "We are prepared to take collective action "in a timely and decisive manner "through the Security Council on a case-by-case basis, "and in cooperation with relevant regional organizations "should other means be inadequate." And I pointed out that R2P had a number of notable features, one that it was binding not only on signatories, but on non-signatories as well, unlike the ICC. It's restricted to severe human rights abuses, genocide, crimes against humanity, war crimes, and ethnic cleansing even though others had pushed for more. It was a big departure for the UN because its focus on what was going on within countries seemed to live in considerable tension, to put it mildly, with the UN's hardboiled recognition of the independence of countries and a remit that was only about in maintaining international peace and security and not being concerned with what went on within countries. I pointed out that it was not binding on the Security Council. The responsibility to protect is the responsibility of governments to protect populations under their control, but it's not a responsibility of the social security necessarily to intervene. It gives wide latitude in a case-by-case, case-by-case basis, and that the early opposition of the non-aligned movement and the grumbling in the African Union was a portent of things to come. And the things to come are what we're gonna be talking about today. I also mentioned some early tests in Sudan and Kenya and Guinea, once the idea of R2P had been adopted, and also in Libya, in February of 2011, February 26th in particular, this is after the outbreak of the Arab Spring, and once the conflict in Libya was under way, that first of all unanimously referred Gaddafi to the ICC, the first time a Security Council had done that, somewhat hypocritically as I said because countries like the US don't recognize its jurisdiction over their own nationals, but they were happy to refer Gaddafi there. And they also mentioned the notion of responsibility to protect. And they imposed an arms embargo sanctions and urged that he'd be subject to criminal prosecution, but the first real test was actually the following month, when the UN substituted Security Council Resolution 1970 with Security Council Resolution 1973, which we're gonna spend a considerable bit of time on. But first, it's very important to get some grip of what was actually going on in the Arab Spring, because it was widely misinterpreted at the time particularly in major countries, and the conventional wisdom about what was going on turned out to be wildly wrong. So let me just give you a snippet of the conventional wisdom and then we'll see what was really going on. - While the uprising across the Middle East and North Africa did not take place during a specific calendar season, they were commonly referred to as the Arab Spring, an allusion to the so called Prague Spring, a 1968 democratic revolution in Czechoslovakia. (uptempo drum music) The Arab Spring timeline began in Tunisia, a country that had been under the allegedly corrupt and authoritarian rule of President Zine El Abidine Ben Ali for more than two decades. On December 17th, 2010, a working class vendor named Mohamed Bouazizi was approached by Tunisian authorities about his unlicensed cart. He offered to pay a fine, but instead, his vegetables were confiscated, and he was publicly humiliated by the police. In a short protest, Bouazizi stood in front of the local governor's office and set himself on fire. He died from his injuries on January 4th. (dramatic music) Bouazizi became a martyr who inspired others who were suffering from unemployment at the hands of a corrupt government. His death sparked a Tunisian revolution in which protesters armed themselves not only with signs, but with cell phones, allowing the protests to spread at social media speed. On January 11th, the week after Bouazizi's death, Tunisia's government fell apart, and the disgraced President Ben Ali fled the country. The videos of the successful uprising shared via social media raise global awareness of the protests themselves. State-run news organizations were barely able to keep up. The speed and success of the protests inspired others across the region. Throughout January, protests erupted in Algeria, Jordan and Oman. By January 25th, the movement reached Egypt, followed by Syria, Yemen, Iraq, Libya, and several other countries. On February 11th, the Egyptian President Hosni Mubarak stepped down, and by the end of the year, Yemen's government was overthrown. (people chanting) As was Libya's, ending the dictatorship and the life of their leader Muammar Gaddafi, who was captured and killed by Libyan rebel militia. - So that's the conventional reading of each of the Arab Spring that it erupted in Tunisia and then just basically spread like wildfire across the Middle East and North Africa, and it's one of the reasons there was so much enthusiasm for it in the West. But it was a fundamental misreading, and just to give you one very important illustration that bears on what's coming next, it was a misreading of what was going on in Egypt. Now everybody could remember at the turn of 2011, the eruption of protests in Tahrir Square in Cairo, millions and millions of protesters, supporters. It started out peacefully, and then supporters of the Mubarak regime became involved in conflicts with the opponents of the regime, and they were largely middle class secular people. We tend to, we now with hindsight, we think about Morsi's Muslim Brotherhood government being overthrown two years later, but at the time of the eruption, it was largely a secular opposition movement opposed to Mubarak, calling for regime change. And I'll just show you a couple of videos there that start to give a hint of how the conventional reading, conventional wisdom was was misleading. (people chanting) (men speaking in foreign language) (people speaking in foreign language and whistling) (people chanting) (man singing in foreign language) - So what was unusual there is, and if I played more, you'd see even more of it, is that from the beginning, the military were very supportive of these demonstrations that were going on in Tahrir Square, and indeed as time went on, it became clear that the military was behind the regime change. - [Matt] President Hosni Mubarak will meet the demands of protestors, military and ruling party officials said Thursday. It's the strongest indication yet that Egypt's longtime president may be about to give up power and that the armed forces were seizing control. A military commander told thousand of protestors in Central Tahrir Square, "All your demands will be met today." Some in the crowd held up their hands in V for victory signs, shouting, "God is great." The military's supreme council was meeting Thursday without the Commander in Chief Mubarak, and announced on state TV its support of the legitimate demands of the people. A spokesman read a statement that the council was in permanent session to explore what measures and arrangements could be made to safeguard the nation, its achievements, and the ambitions of its great people. The head of the ruling party told the Associated Press that he expects Mubarak will address the people tonight to respond to protesters' demands. Matt Friedman, The Associated Press. - Either this is the first Arab revolution of the 21st century, either this is the first Arab revolution of the 21st century, or it will be brutally suppressed. - [Reporter] Tense new beginnings for Tunisia, its Arab neighbors nervous of how revolutionary feelings could spread. Mubarak's deposed, Egypt's 18 day revolution defies all expectation. - Rarely can a military takeover have been greeted with such enthusiasm, as it is on the streets of Cairo tonight. - So there again, you see this enormous enthusiasm for this military takeover, which then promised elections. And in the elections, initially, the Muslim Brotherhood was not going to run. They said they would not run because they were the only organized political party in Egypt, and there was a general perception it would be unfair for them to have such a big head start, but they soon changed their mind and the military allowed them to run, and Morsi easily won. But the puzzle is why did the Egyptian military get behind this revolution? It's not what the military typically does. Mubarak had a long history of association with the military. And so this really is a corruption story spelled out in Sarah Chayes' book "Thieves of State: "Why Corruption Threatens Global Security". That book is principally about Afghanistan where she spent 10 years, and for her main thesis in the book, which is sort of orthogonal to what we're doing here today, but I'll just mention it, her main thesis is that insurgent movements are fed by corruption, that people on the ground, citizenrys hate corruption, and if insurgent movements promised to get rid of corruption the people will support them. And she had been trying to convince Petraeus, McChrystal and others that any effective counterinsurgency movement in Afghanistan would have to focus centrally on getting rid of corruption if it was ever gonna get anywhere, and she eventually got, had some headway. But her chapter about Egypt tells a somewhat different and more interesting story in some ways which is that the Egyptian military had a long history of being involved in a certain kind of corruption with a difference. So if you go back to 1970, the 1967 and '73 wars against Israel, the Egyptian military had been badly humiliated by being defeated by the Israelis, and as a result of that, they partly blamed the inept Egyptian state and its inability to deliver the things that the military needed. So they decided to essentially, you might call it, sort of go rogue, and they created their own second economy to create all of their own supply lines for the military. And the thought was that never again was the military gonna be under-resourced and under-prepared in a future war. But as things go, and you get mission creep in all matters in life, they then discovered, many of these generals, that they could actually make a lot of money in this second economy, and so it became a lot of essentially a kind of black market economy through which these generals became rich. However, they were fairly discreet about it, they didn't flaunt their wealth very much and they didn't outrage the population that much. However, what happens in the heyday of neoliberalism and the Washington Consensus is that other Egyptian elites also decide to get rich, and particularly Mubarak's son, Gamal Mubarak was a sort of London playboy with lots of fancy cars, and he and his friends came back to Egypt and they, they did flaunt their wealth, and so they had the Lamborghinis driving around Cairo and the playboy clubs and all the rest of this, and this partly outraged the military and also made the military worried that there was gonna be a reaction against this kind of corruption. So they got behind the coup, behind the Arab Spring because they really, they wanted to get rid of the Mubarak regime. And furthermore, they were very happy for Morsi to be elected because in their view, he would be the agent of cleansing the old regime, and they would be able to keep their hands clean, which he then did because Morsi ran, he played the typical Sarah Chayes playbook, he ran as being a cleaner up of corruption who would get rid of, who would drain the swamp, so to speak. And he set about doing that, going after the old regime, and indeed, Mubarak and his children were convicted of various corruption charges and were in and out, and the sons were in and out of prison for many of the next several years. But once Mubarak had done, once Morsi had done that, the military had no more use for him. And so when people came dissatisfied with him, they had no problem getting rid of him, and General Sisi took over and recreated a military regime, and that's what we've had for the last several years. So there was a popular uprising, but people misunderstood what was really going on, which was essentially the military ridding themselves of liabilities and consolidating a regime over a couple of years. And so that was one form of misreading what was going on in the Arab Spring. Another form is what happened in Libya, as we could also tell much of the same story for Yemen, which is that people widely mistook a civil war for a revolution, a civil war for a popular uprising. So the Security Council Resolution that was adopted in March, 1973, you can see it was voted 10 yes, no no, with five abstentions, and so nobody had opposed it whereas the one three weeks earlier had been unanimous. And it established no-fly zones, this is when violence had erupted. It excluded the introduction of foreign ground forces. It authorized member states acting nationally or through regional organizations or arrangements to take all necessary measures to protect civilians and civilian populations under threat of attack. And it created an exception to the resolution it passed three weeks earlier, so that, which had imposed an arms embargo in order to enable the UN supported forces to arm the rebels. And so that was a signal of what was to come. So here you had, in February 26, you had a resolution passed, including an arms embargo, three weeks later, you have a modification of that and a much more aggressively interventionist but still containment-oriented regime. So one question is, well, what happened? Why? Why did the UN authorize these new actions in three weeks? And here's a clue to that action. Notice the dates on these documents. The Resolution 1970 had been passed on February 26th, on March 10th, so this is a week later, a week before, two weeks later, a week before the new resolution's gonna go ahead, you could read this in the New York Times, "Moving ahead of its allies, France on Thursday "became the first country to recognize "Libya's rebel leadership in the eastern city of Benghazi "and said it would soon exchange ambassadors "with the insurgents. "The move was a victory for the Libyan National Council "in its quest for recognition and a setback for Gaddafi "who had been seeking whatever international support "he could for as NATO members in Brussels "had begun debating the possibility of imposing "a no-fly zone. "The French announcement came as loyalist forces in Libya "claimed new successes against the rebels "west of the capital, while in the east, "loyalist forces renewed ferocious attacks," on several towns actually. So what this paragraph is concealing or not telling you is that when the first resolution had passed, it looked like Gaddafi was basically done. It looked, he had lost control of all of the cities, and as a result of that, it looked like his days were numbered. Most people thought the regime was crumbling. And the French and the British and the Italians had big oil interests and good relations with Gaddafi. Gaddafi, we forget, in the first decade of the 21st century had sort of, internationally speaking, come in from the cold. He had given up the Lockerbie bombers for trial, he had agreed to dismantle his nuclear program, which he subsequently did, he helped with intelligence against Afghanistan in the aftermath of 9/11, and actually also, his domestic atrocities had diminished. If you look at most of the human rights violations he's accused of, they're all before the turn of the century. Didn't mean he'd become a nice guy, but he was a pretty run-of-the-mill Middle Eastern dictator not as nasty as he had been in the 1990s. But now it looked like he was done, and so the question is what was gonna happen to the oil relationships? And so the French decided they better make nice with who's gonna be the new government, and so they started conversations, and these were brokered by a French political philosopher by the name of Bernard-Henri Levy shows you why you should never let political philosophers have too much influence in politics, and he started brokering meetings between the opposition groups and the French government, Sarkozy's government. He was essentially telling them, "This is the government that's taking over," but after the first resolution had passed, the reality on the ground changed, and Gaddafi started winning again, and so they were then in trouble because they had said on March 10th that France recognized the Council as a sole legitimate representative the Libyan people, it put France ahead of other European powers that had been seeking ways of supporting the rebels on toppling Gaddafi, normally European countries don't do that, so their aggressive stance was seen as a way of showing commitment to the popular uprising, so they said, but what they were really doing was showing a commitment to the what they thought was the emerging government of Libya. And you can see here how they spun it, the day that the resolution was passed. The resolution, allegedly just about containment, passed a week after France had already recognized the opposition as the legitimate government of Libya. - [Tim] After being accused of being slow to react to the revolutions in Tunisia and Egypt, the French president has moved quickly to silence his critics and take the lead in reprisals against Colonel Gaddafi. Meeting representatives of the rebels at the Elysee Palace, he promised them immediate diplomatic recognition. (speaking foreign language) - [Interpreter] France has expressed its satisfaction about the creation of the Libyan National Transition Council and after this meeting with President Sarkozy, I can say that France has recognized the Transitional National Council as the legitimate representation of the Libyan people. - The French plan is reported to include proposing limited airstrikes against key Gaddafi targets, such as his headquarters and a military airbase. There was no official confirmation from President Sarkozy's office, and it's a highly controversial proposal that would meet with strong resistance from many European allies, who are also wary of giving the rebel representatives diplomatic recognition at this stage. (speaking foreign language) - [Interpreter] That's a decision by a single government. I think it's better for us to listen to the stance of the entire European community. - [Tim] The Elysee Palace says that Mr. Sarkozy is seeking legal authority to prevent the use of force by Gaddafi. It believes that jamming his communications would be a good start. Tim Friend, Al Jazeera, Paris. - So there was, they were now, they were in a difficult spot because they had recognized the opposition, but Gaddafi was now once again winning the war and he was regaining control of the cities, and he was indeed about to regain control of Benghazi, and so what the French did was that they pushed very hard to get a resolution passed that would allow them to depose Gaddafi because they realized that they had burned their bridges with him, and essentially, that had become their agenda. And right after the resolution passed, they began massive bombing almost immediately, so much so that the Arab League, you can see here, the next day the resolution passed three days later, passed on the 17th, the Arab League was furious as was the African Union who had gone along with this, and African Union countries on the Security Council had even voted for it, and they saw that there was a much more aggressive agenda going on and it was really about getting rid of the regime. And indeed, a lot of reports, and I've given you the jot and tittle of this in the reading for today, a lot of the report, or the reading for two weeks ago, a lot of the reports about imminent slaughter of civilians in Benghazi turned out to be what we call fake news today, and there were very, first of all, very few fatalities in the retaking of these cities, and almost no civilians were killed. Nonetheless, the hype for the intervention was cranked up on the grounds that Gaddafi was about to slaughter hundreds of thousands of civilians in Benghazi. But I said at the outset that this, that everybody in the West who was treating this as an Arab Spring revolution were misreading that this was really a civil war. And the smoking gun comes when you start to look at what was going on in the Obama administration, because after all, Obama had run against, against endless wars. Obama had said that the, yes, Afghanistan had been a war of necessity, but the Iraq invasion had been a war of choice. Unlike Hillary Clinton and unlike Kerry, he had not supported it in 2003, and he said America should not get involved in these sorts of conflicts. And so who in the administration was against it? Well, we'll see that in a second, but here's the smoking gun that it was a civil war, who was leading the revolution? One is that Abdul Jalil was Gaddafi's former Justice Minister, Ali Aziz, Gaddafi's former Ambassador to India, Mahmoud Jebril, his former top economic advisor who was closely linked to the oil multinationals, and the leaders of the so-called Rebel Revolutionary Council included Abdul Younis who, in the battle of Benghazi, had actually been on the government's side initially and switched sides, and Omar Hariri, who had no democratic credentials. Both of them had long histories dating back to 1969 of repressing democratic movements. So what happened here, what happened? As I said first of all, February and March, from February to March, things changed on the ground, and Gaddafi now had the upper hand, and the British and especially the French persuade Hillary Clinton and then Obama to get behind the, to get behind the resolution and go in. But so who was against it? First of all, Robert Gates was against it, then Secretary of Defense, he had been appointed by Bush Obama had kept him on. If you read Gates' autobiography, he says the only reason he didn't resign over this was that he was about, he was three weeks away from retirement or four weeks away from retirement, but he said to Obama, "Why in the world would you do this? "We're already involved in two quagmires in the Middle East, "why get into a third one?" Joe Biden, Vice President, was against it, Mike Mullen the Chairman of the Joint Chiefs was against it, Tom Donilon, the National Security Advisor was against it, the Chief of Staff, Bill Daley was against it, and the Deputy National Security Advisors were against it. Who was in favor of it? Susan Rice, the Ambassador to UN, Samantha Power, a big advocate of responsibility to protect and print and had been appointed as a Special Assistant for Human Rights, Bed Rhodes, a Deputy Advisor in the White House, and Hillary Clinton as well became convinced. And this is how we got to the phrase leading from behind, which Obama was subsequently tied with. As best I can reconstruct, although he did go on to the Jay Leno show later and say he never used the phrase, but basically what happened was that Hillary Clinton had been skeptical, and she was in Paris for the G7, and the French philosopher arranged meetings between her and Jebril, and they persuaded her that they were indeed going to overthrow the regime and install democracy. Even though none of them had the slightest history of democratic behavior, and so there was no reason to believe that they would in fact do that. So she came back from the, I guess it was the G8 at that point, and she said to Obama, "I think we should do this," and he opposed it and they argued about it, and eventually he said, "Well, it's 51-49," and she said, "Well, the French really want to do this, "the Italians are on board," and so on. And he basically, he said, and this is why the French started bombing early, he said, "Well basically, all right, but it's got to be their show. "We're gonna be leading from behind." So what quickly happened though was that NATO expanded its remit, and the evidence of this is that once the African Union and the Arab League saw what was going on, they said, "Now we can, "the crisis in Benghazi is over, "civilians are not being slaughtered, "we can now arrange a ceasefire between Gaddafi's forces "and the opposition and mediate this conflict." And instead, the NATO forces said, "No, no, no, "we're gonna knock off Gaddafi." Now they couldn't put in ground troops, but they massively supplied the rebels until they were successful. It took about eight months until Gaddafi was finally killed and captured. The death toll was somewhere between 2,500 and 25,000, I've seen estimates, probably about 8,000. And they nakedly pursued regime change, and that, but nobody was portraying this at the time as a civil war, they were still arguing that this was a popular uprising and there was a democratic opposition and this was all gonna be great. So again, you can see the hype. - The Libyans I met didn't believe it. They said, "No, it can't be true." Many of them said, "This is something we have been dreaming of." - [Reporter] As he's aged, Muammar Gaddafi's style and his rhetoric had looked more and more anachronistic. His a Theater of the Absurd, punctuated by violent acts from a dangerously eccentric showman, a murderous clown. - [Reporter] The Libyan revolution of 2011 was about one thing above all else: Removing Colonel Gaddafi from power. Today concludes that revolution. - But of course, it was not, it was the beginning of eight years of civil war. So they pursued successful regime change, and peace-building turned out not to be possible. It turned out not to be possible for a number of reasons. First of all, there was nobody that had the capacity to run the country, as we saw. We were not putting in forces, we were prohibited from putting in boots on the ground, and essentially you had rival militias controlling different parts of the country, and as the summary video I showed you at the beginning made clear, every effort to create a national government with legitimacy soon fell apart. But, that's not all. So it was obviously a huge catastrophe for Libya, and the country has just more or less disintegrated over the past eight years, and it doesn't show any signs of turning around, but there was huge fallout across the region. First of all, even as far away as West Africa and Mali, it turned out a lot of Gaddafi's principle guard, his elite troops were from Mali, and they went back to Mali and ended up fomenting conflict in Mali, which led to the downfall of what had been the most stable democratic regime in West Africa, but more important, across the Middle East, first of all, the failed state in Libya created a massive arms bazaar, so weapons and rockets and tanks even started showing up all over the place, feeding the conflicts in other countries and indeed, Libyan rockets started showing up in Gaza in 2014 in the conflict there with Israel. But the most disastrous fallout was in Syria, and the most disastrous fallout was in Syria exemplifying what international relations theorists referred to as the moral hazard of intervention, and this is the moral hazard of intervention. The idea is if people think that if we revolt, somebody will intervene and help us, they will revolt, all right? So if you look at what had been going on in Syria before this episode in Libya, what had been going on in Syria was several years of catastrophic drought starting in 2007, and this had meant that hundreds of thousands of farmers had lost everything, their crops had died, their animals had died, just catastrophic droughts. And so what did they do? They went to the cities, looking for non-existent work. And the Assad regime came to USAID, they came to the US saying, "We need help, we need assistance," and we basically blew them off. Even when it reached the point when it's hard to square with our current perception of Assad, but this is more than a decade ago now, he said he's not gonna be able to maintain control of the country if he doesn't get aid. Nonetheless he didn't get it, and so when the Arab Spring erupted in Tunisia and in Egypt, there were a lot of these protests were going on in Syria of these destitute farmers, but it was completely peaceful. There was no, there was no violent uprising in Syria until the uprising in Libya and the 1973 and the intervention of NATO forces. So that's why, our scholars call it the moral hazard of intervention, they looked at at Libya and said, "Well, if we revolt, "NATO will come and help us too." And so they, so they ramped up the, ramped up the opposition into violent conflict and Assad came down on them with a hammer, of course, and we it turned into, very rapidly the Syrian crisis turned into the worst humanitarian catastrophe of the new century. So Obama was then in a very difficult position because he, you might say well, why didn't he, why didn't he stick to this? This a French-led operation, we were leading from behind. That's not how command and control works in NATO, the American generals run the whole, if NATO does an operation, the US runs it, that's just the way the force structures work. So the Americans were deeply involved in this, and so as things deteriorated in Syria, he was in a very difficult spot, even though particularly when Assad started doing things like using chemical weapons, he began calling for regime change, so you'd see just a few clips here. - President Assad had lost credibility that he attacked his own people, has killed his own people, unleashed a military against innocent civilians, and that the only way to bring stability and peace to Syria is gonna be for Assad to step down and to move forward on a political transition. We're gonna keep increasing the pressure on the Assad regime and working with the Syrian opposition. Assad needs to go. He needs to transfer power to a transitional body, that is the only way that we're gonna resolve this crisis. The only way that the civil war will end and in a way so that the Syrian people can unite against ISIL is an inclusive political transition to a new government without Bashar Assad, a government that serves all Syrians. I discussed this with our Gulf Cooperation Council partners at Camp David, and during my recent call with President Putin. - So, "Assad must go, Assad must go, "and I'm even telling Putin Assad must go," but the truth is, Obama was completely hamstrung in Syria. He was hamstrung, first of all, there was no possibility of Security Council authorization to go into Libya after there was this widespread perception at the UN that NATO had completely violated the containment language and remit of 1973 and basically underwritten the overthrow of Gaddafi. So secondly, Obama's very shrewd political instincts knew that in getting involved in a civil war in Syria would be a domestic political catastrophe for him, so he couldn't possibly actually make it happen, so but then you're in this position of demonstrating weakness, for which the Republicans attacked him relentlessly, drawing lines in the sand, Assad must go, not making it happen. But of course, it was bait for him, as far as they were concerned because they knew it would be a political disaster for him if he got involved. And he realized by then that he'd been suckered into Libya. He'd been suckered, I should have just given you one more wrinkle about how Hillary got turned around on this because it's quite remarkable. The Select Committee on Benghazi that was going after her because of the death of four consular officials in Benghazi, as you might recall, unearthed documents which showed her personal friend and advisor, Sydney Blumenthal, who had no role in the government, sending endless numbers of emails telling Hillary that the US needed to get involved in the Libya, in the Libyan operation, why? It turned out because he, Blumenthal, wanted to get into the, he had interests in Libyan oil and he was, there are emails from him to Hillary saying that the French are grabbing all the oil concessions. And so again, we might think of this about ideals and humanitarian intervention, but there's this material a story of people with real interests at stake pushing this agenda. So I think, and by then, Obama realized that to some extent he had been suckered by her, if not her intentionality, certainly by her gullibility. And so he was very unlikely to succumb to going into Syria. Syria moreover was, Afghanistan and and Libya had been two-sided civil wars, Syria was a three-sided civil war because by then, ISIS had emerged and there was the various opposition groups to the regime, and then there was the Assad regime. And if we couldn't prevail in a two-sided civil war supporting one side, how were we gonna be effective in a three-sided civil war? Because one of the ugly truths of the post Cold War-era is that my enemy's enemy might turn out to be my enemy. In the Cold War, there's two sides and you know who's on which side, and so on and so forth, one of the reasons by the way there's the Libyan thing is ongoing is that the Trump administration is now supporting the other side, because the Saudis are supporting the other side from the UN and this side that the US has traditionally supported. So it was a three-sided, and then former Ambassador Polk has done an exhaustive study of the opposition groups, and there were no, it was again like Libya, there was no Democratic opposition in Syria. And then finally, the Democrats had no alternative to the Bush Doctrine. I said this last time when we talked about the deficiencies of the Bush Doctrine, but what was the democratic alternative? They had no well articulated conception of containment. And indeed, that I mentioned that talk I gave at the University at the Yale Club of Tokyo, that eventually became the little book I wrote about containment, arguing that if the Democrats went back into office without a national security strategy, they were gonna be pushed around, as Obama was indeed pushed around in Libya because not having a strategy and doing things like triangulation leads you to be manipulated by people who have a clear sense of what they want to do. And indeed, the Russians, when they went finally into the Syrian conflict, Obama yelled had screamed and stamped his feet but he couldn't do anything about it, and unlike us, they had a clear strategy. Their strategy was to help Assad win the civil war, which they then did. And so now Assad basically has won the civil war. So we on the other hand have no strategy in the Middle East, and this is continuing to play out today as we are watching in real-time that whatever you say about Trump's abandonment of the Kurds, it completely, his entry into that whole conflict is the correct observation that the US doesn't have a strategy there and there is no democratic regime that's being supported. The Kurds, supporting the Kurds, supporting the Kurds is a bit like supporting the Northern Alliance in Afghanistan. The Kurds are scattered through a number of Middle Eastern countries, the Kurdistan that they would like to build, it would involve going to war with several of them. Turkey, for all of its appalling behavior, is a NATO ally, the Kurds are not. As I said, one's enemy's enemy can turn out very easily to be one's enemy. And so we have throughout the rest of the Obama administration but really starting in 2003 have had no coherent strategy of what it is that we're actually trying to achieve in these conflicts. And so Trump took advantage of that and ran with it. And that's not the only fallout from the Libyan conflict. It's quite a lot, but there's more. - As we return now to the attack on a migrant detention center in Libya, as John yang tells us eight years of conflict and instability there are now enmeshed with the migrant and refugee crisis. - [John] Bags of clothing and abandoned shoes. Remnants of life strewn throughout the blood-soaked debris of death. The early morning air strike hit the Tajoura Migrant Detention Center in Tripoli housing some 600 people, mostly North Africans. Emergency crews struggled to carry away body bags through the wreckage. Survivors said they had no warning and no protection. - [John] Libya's government of national accord recognized by the United Nations blamed the so-called Libyan National Army, or LNA, which is trying to seize Tripoli. The LNA acknowledged carrying out attacks in the area but denied targeting the migrants. The LNA under the command of General Khalifa Haftar is supported by Saudi Arabia, Egypt, and the United Arab Emirates. It controls southern and eastern Libya, but has met heavy resistance from militias defending Tripoli. Yesterday, an LNA commander warned of intensified assaults. This is the latest calamity in eight years of chaos in Libya. Fighting has also made an already desperate journey for migrants trying to reach Europe even more dangerous and deadly. An estimated 6,000 languish in Libyan detention, at least half close to the front lines of conflict. - In the last three months, Tripoli and the neighboring areas around Tripoli have become an active war zone. So now we have migrants who are in abysmal conditions in detention, but on top of that, they can't flee to safer areas of the city. - [John] Many of the migrants were rounded up by Libya's Coast Guard, trained and funded by the European Union, to prevent their crossing the Mediterranean. - Several member states of the EU have effectively begun criminalizing aid agencies who were providing such and rescue services in the Mediterranean Sea. And in the absence of those search and rescue operations, people who continue to try to flee Libya and cross the Mediterranean Sea are intercepted by the Libyan Coast Guard and returned to Libya and are sent directly into the detention centers. - So in addition to everything else, this has fed the refugee problem in Europe, which has fed the rise of anti-system parties in Europe and countries like Germany. People like Angela Merkel were initially open arms for these refugees, but as we know from earlier lectures, that soon turned into political catastrophe for her. And so this ongoing, the externalities of this as an economist would put it are just stunning. So those are the sort of local and then regional and even international costs of this, toppling this regime. So we've done a big, a deep dive into some micro stories here, things are often not what they seem but let's now step back a little and think about what does this mean for the doctrine of responsibility to protect, and what are its implications for other forms of humanitarian intervention? So thinking back to the earlier lecture on responsibility to protect, we now have to recognize that the Kosovo operation was misleadingly benign. It gave people much too much confidence in what can be achieved by forcible intervention in these types of situations. First of all, in 1999, Kosovo was already in the middle a, it was in the middle of a civil war in a failed state, so the idea that one might be creating a failed state by intervening was not on the table. Secondly, unlike the story in Benghazi that was mostly hype and widely discredited at the time, if you go back and look at the news coverage, it was undeniable that there was ongoing ethnic cleansing and gross human rights abuses going on in Kosovo. Third, the intervention was brief and it seemed to work, and that led to enormous hubris about the effectiveness of surgical power in rearranging the furniture in the targeted regime. Another feature of it was that NATO and the US plausibly had clean hands in Kosovo. There was nothing really strategic at stake for them. I said it violated the NATO charter that they got involved because they were, none of them was threatened, but really, they didn't have anything to gain out of this. So the notion that there was some surreptitious agenda, as we've now seen there was surreptitious agenda going on in Libya wasn't there, and it wasn't there for good reasons. Also, the Kosovo operation was before the US-led invasion of Iraq had sullied our international image. Think again to the story about Abu Ghraib that we ended with last time. We were perceived very differently in 1999 across the Middle East than we were perceived after 2003. And just to add to that, the intervention in Kosovo had been on behalf of Muslims, so that muted criticism in the Middle East and the global south, even though there was some, but we were, it was the Americans and NATO intervening to protect Muslims in Kosovo. So for all of those reasons, there was much more confidence in the ability to implement R2P than we now see was warranted, but people like Samantha Power and Susan Rice and others therefore were far too optimistic. And when we take stock, there are really three, three big challenges one has to think about when we're talking about humanitarian intervention. One is big principal-agent challenges. So if you're thinking about somebody going in to intervene in a country, who is the right principal? Who should go in? Who should make the decision to go in, I should say? And then who's the right agent? During the Cold War, the sort of agreement was that the US and the Soviets would not send people in, that's why in the Middle East, you'd get sort of Norwegian blue hats or whatever because if you have an obvious interest, there's a problem. But most importantly, how does the principal control the agent? Once the UN Security Council has authorized NATO to act and they're acting, it's very difficult to reconvene and reverse those decisions because any one of the permanent members can veto efforts, efforts to do that. So Brazil has been pushing something called the Doctrine of Responsibility while protecting, trying to create monitoring ongoing, monitoring device of interventions, but I think good luck with that. Then there are questions about effectiveness. Peace-building is very difficult after an intervention. If you look at the literature that's been published by, the empirical literature on peace-building, some of which I will post if people are interested, it's extremely difficult to do after an intervention. I talked about this some in relation to Germany, West Germany and Japan after World War II, but particularly when you go in during a civil war, again, we went in supporting what turned out to be the losing side in a civil war and then we we basically knocked off the regime that that had very little legitimacy, but neither did anyone else. And so peace-building operations are extremely expensive, extremely slow, and don't work very well. Civil wars are, one of the main factors the scholarly literature shows, one of the main factors that lengthens civil wars is often external involvement. If you look now at the civil war in Yemen, it's becoming a proxy conflict between the Saudis and the Iranians. And so once you start supporting one side in a civil war, the chances are, if I just come in and support other sides, the conflict's just gonna go on, whereas if nobody gets involved from the outside, most civil wars eventually end in about five years or less. But they're gonna also be huge legitimacy challenges. There's no obligation to intervene, I've said this earlier, so it's a case-by-case discretion for the Security Council, but that effectively creates immunity for the Security Council members and their allies. In 2014, when Israel went into Gaza, there were widespread allegations of R2P violations on both sides by Hamas and by the Israeli troops, but the idea that there was gonna be some kind of humanitarian intervention of course is unthinkable when you consider the alliances of people on the Security Council. So this doctrine is gonna be perceived to be an instrument by which the countries that are in the driver's seat in the Security Council are going to protect themselves and not others. So what are the lessons? Well, one is that R2P should remain exceptional. As I said when we were talking about this idea of illegal but legitimate, it's a bit like asking for forgiveness rather than permission, but you can't institutionalize that, so it should be exceptional, it should be very, very difficult to do. You need exceedingly high thresholds for intervention, because if you intervene, if you take on a government on its own territory, you are almost by definition gonna take, you're courting the possibility of creating a failed state. It's important therefore to keep force proportional to the problem. And as as within days of 1973 being adopted by the Security Council, it was clear that the potential humanitarian catastrophe in Benghazi was over, Gaddafi's Air Force had been destroyed as well, and the overtures being made by the African Union and the Arab League should have been listened to, to negotiate some sort of ceasefire. Regime change should always be the last and final resort. Interested players are very bad agents, but the UN doesn't have a military, so who are gonna be the agents? Peace-building, as I said, it's essential, but very difficult to do. You might say a better focus rather than responsibility to protect would be, or responsibility while protecting would be responsibility to prevent. I said to you earlier that the crisis in Syria really starts ramping up in 2007 and eight and nine because of the droughts. That would have been the time to intervene. Look at a country today like Chad. It's a landlocked country surrounded on all sides by failed states. It's oil-dependent, that makes it a prime target, a prime candidate for this sort of collapse and humanitarian catastrophe that we have seen in Libya and Syria. But what is anybody doing about Chad? Very little. So just to sum up, I think that, if you think over the course of these lectures on resisting aggression across borders, it is possible to come up with an integrated account that builds off of Kennan's insights for a post-Cold War world. And I take the three elements of his view to be economize on the use of force, keep it sufficient but proportional. Don't destroy the village in order to save it, the famous line from Vietnam. Always remember that the biggest challenge is and the most important battle to win is the battle for hearts and minds, not the military confrontation, that is not gonna create a governable country. And stop the bully without becoming one. This idea to prevent domination, but without yourselves dominating. Machiavelli says at the outset of the discourses, he says the reason he favors democracy is that the elites want to dominate, and the common people want not to be dominated. So this is the idea of prevent domination without becoming a dominator. And if you wanted to sort of draw a conceptual map here of resisting aggression across borders, you could do something like this. You could say, well think about domination that crosses borders. First, there's self-defense, if a country attacks you, that's the strongest case. Classic containment, you want to prevent, you want to head that off, a country that's in the position to do that. So I think Kennedy during the Cuban Missile Crisis gradually ramping up enough, the embargo and so on, that is the way to, to prevent threats to the homeland, if you like. Make war a strategy of last resort, and ramp it up, ramp up hostility proportionately. If you're going in to defend a third party, as we did in Kuwait, in Iraq, when Iraq went into Kuwait in 1991, then we should be thinking about containment again first of all authorization, there needs to be authorization. That could be invitation from a legitimate government, as was the case in Kuwait, or it could be external authorization through something like the UN. When we're thinking about crossing borders to resist domination, there's got to be a higher threshold, because the stakes are so much higher because once you're a crossing border in order to resist or to prevent domination, that might be better to say than resist there, you're inevitably courting the possibility of creating a failed state. So with responsibility to protect, there must be the imminent danger of human rights violations, there must be external authorization, there must be a diverse coalition, there must be a limited mandate which it should be observed as was done in 1991 in Iraq, but not in 2003, and regime change should be the absolute final resort. What about intervening to install democracy? This is the most ambitious of all. We're not intervening, this is a neoconservative dream, not just to stop particular human rights violations, but actually to produce regime change. And there, first of all, there's no point in trying to install democracy in a country where the plausible economic conditions don't exist, we know what they are, per capita income of about $14,000, a diverse economy being the two most important. There must be an opposition that is predominantly indigenous. Part of the problem with Ahmed Chalabi, he had been living in London for decades, and when things didn't work out in Iraq, he could go back to London. But he did go back and forth to London when things got bad for him. You really want the indigenous opposition because they have, it's like you want the pilot on the plane. You want people for whom if this goes badly, they're gonna pay the price before you go in and help them, or try to help them. You must have some credible conviction that they're capable of governing, and some credible conviction that they're likely to be democratic. So even if and when one side wins in this Libyan civil war, there's no guarantee that it's going to be democratic. Indeed, it seems vanishingly unlikely that that would be the case. So I think, if you want to sort of think conceptually about what should be guiding policy thinking about humanitarian and other forms of intervention, it would look something like that in a post-Cold War, which would ramp up the stakes appropriately as the risks and costs of failure accelerate. But the big problem we now face in addition to all the others is that the catastrophe in Libya has more or less left R2P in tatters internationally. It'd be very difficult to do this again. And America has greatly diminished its capacity to act throughout the Middle East. I'm seeing Russia has expanded to fill the vacuum, and that is the sort of big geopolitical story is that capacity to be influential in the region is much less than it was the day before George W. Bush intervened in Iraq in March of 2003. So it's in that sense a big withdrawal by the US, and Russia filling the void. We will see on Tuesday another kind of withdrawal by the US that has been, had other forms of void being filled, particularly with China's involvement in Africa. We'll see you then. (soft music) |
Power_and_Politics_in_Todays_World | Lecture_13_The_International_Criminal_Court_and_the_Responsibility_to_Protect.txt | - We're gonna continue with the part two of the course on the new or emerging global order after the end of the Cold War. We've spent two lectures on democracy, and particularly the enormous promise that democracy seemed to offer, which was sort of, for which the South African transition was something of a poster child in 1994. This was before the mixed legacy of the South African transition, and particularly the failure to achieve much full transformation in the economy that would play itself out in later years. In 1994, it was really seen as the poster child for this democracy that was sweeping the world, including in places where it had not been anticipated. Today, we're gonna be talking about the International Criminal Court and the Responsibility to Protect, two doctrines that began a fundamental reshaping of the international legal and political orders. So that's going to really consume us for the next several weeks, next several lectures, I should say. Anyone know what this is a picture of? Who are these people? - [Man] Rwanda. - It's Rwanda, yeah, who are these people? - [Man] Tutsi, Hutu? - Hutu. What are they doing? - [Man] Committing genocide. - How do you know? What are they doing? - [Man] There are people with sticks. - They're going along with sticks. And who are these people? - [Man] Peacekeepers. - [Ian] Yeah, these are UN peacekeepers. What are they doing? - Nothing. - Nothing. (audience laughs) They're watching. This is a famous photograph that appeared in The Guardian on the 12th of April, 1994, and did the 1994 equivalent of going viral. Of course the rules of engagement for UN peacekeeping troops was that they could not become involved in ongoing conflicts, and so, in effect, they were not in a position to do anything about it. And so this is, you could call a subtitle for today's lecture, "Rwanda's Shadow," because this was a horrific genocide in the course of which, in about three months, somewhere between 800,000 and a million Tutsi, mostly Tutsi, some moderate Hutu, and some other groups, were slaughtered in the most graphic, much of it appeared in the Western media. You could go up and if you tried to look on YouTube you'll have to sign in confirming your age because a lot of it is so gruesome. So the thought was that the international community had to be able to respond. After World War II we heard these praises like "Never again." This was a clear case of genocide and if the international community couldn't do things to stop it, it was likely to proliferate, genocides were likely to proliferate. This what I had taught you earlier about the sort of frozen-in-place international structure. The Cold War was gone and it was not clear that the major powers would police their surrogates in ways that would prevent this kind of violence. It's really quite a dramatic counterpart when you think that the very same month, 2,300 miles to the south, the South Africans were having their first democratic election, three days of peaceful voting, again, all this euphoria, while at the same time this carnage was going on in Rwanda. Let's listen to President Clinton on this 22nd of July just after it was more or less over. - Good morning. I have just met with my national security team. The flow of refugees across Rwanda's borders has now created what could be the world's worst humanitarian crisis in a generation. It is a disaster born of brutal violence, and according to experts now on site, it is now claiming one life every minute. Today I am announcing an immediate and massive increase in our response. We will provide personnel and equipment to enable these airfields to operate on a 24-hour basis. Our aims is to move food, medicine, and other supplies to those in need as quickly as possible. I've directed the Pentagon to establish a safe water supply and to distribute as much water as possible to those at risk. Safe water is essential to stop the outbreak of cholera and other diseases that threaten the refugees. I've ordered the State Department and our ambassador to the United Nations, who is here with us today, to take immediate action to help create those conditions. The United States will support and urge the immediate deployment of a full contingent of United Nations peacekeepers to Rwanda, to provide security for the return of the refugees. We are making clear to the new leaders of Rwanda that international acceptance, including American recognition, depends upon the establishment of a broad-based government, the rule of law, and efforts at national reconciliation. We're taking action to counteract the propaganda of the extremist Hutu elements, who continue to urge Rwandans to flee. Taken together, these steps will help to relieve the suffering of the Rwandan refugees and create conditions for their return home. - So, President Clinton was heavily criticized after that speech, and the reason he was heavily criticized was not for what he said but for what he didn't say and, indeed, for what the administration had systematically been refusing to say throughout the build up and then the unfolding of the Rwandan crisis. That was the word genocide. It refused to say genocide. And you can now find in the archival documents there was a lot of back and forth within the administration, huge resistance to using the word genocide, on the grounds that that would dictate that we would have to go and do something. And President Clinton was extremely reluctant to do anything to get American troops involved in this conflict because he had had his nose badly bloodied six months earlier in Somalia in the so-called Battle of Mogadishu. And so the US was, this had started under the Bush Senior administration, the US had been the leading player in the UN's peacekeeping operation in Somalia, called the UN Operation in Somalia. And it was an ongoing civil war. And as part of that effort in what was widely seen as badly-botched American operation, two Black Hawk helicopters had been shot down and the so-called Battle of Mogadishu proceeded, in which 18 Americans were eventually killed and 73 wounded. And from the point of view of the administration, what was especially damaging was three dead Americans were dragged, dead or dying Americans, were dragged through the streets of Mogadishu and filmed being so dragged. And this was, again, a 1990s version of going viral. It was seen as a huge incompetence on the part of the American administration. Probably the most damaging thing for Clinton, at least the most damaging thing for an American president, at least since Jimmy Carter's botched operation to free the hostages in Tehran in 1979, where we also had helicopters coming down in the desert. That in fact, the 18 Americans killed in the Battle of Mogadishu was the largest number of US fatalities since the Vietnam War. So it was seen as a really big deal, that would eventually be surpassed in the Second Battle of Fallujah that I talked to you about some time ago, where some multiple of that was killed. I think about 95 American soldiers were killed in that battle. But at the time this was seen as a horrific mess that the US had gotten into, and the Clinton administration had responded by deciding to get out of Somalia, rapidly. And so they pushed the UN essentially to declare the end of the Somalia Operation the following year, and they were in the midst of withdrawing from that. So the last thing the Clinton administration wanted to do was get American troops involved in another civil war where they could have taken a massive domestic political hit. And so it's not surprising, even if it's not impressive, that President Clinton refused and the administration refused to call what was going on in Rwanda genocide and refused to get involved until the conflict was, for all practical purposes, over. So, today's agenda is we're gonna look at nonintervention, this idea that has to be broached if we're gonna start talking about protection of international human rights in a post Cold War world. We'll dig into what the ICC and the doctrine of Responsibility to Protect are, how they came to be adopted, and what force they have in the world. We'll also question how radical a reshaping this is of the international political and legal order. And they're both very much doctrines in evolution and in flux, so to some extent we're looking at moving targets. But we will do the best that we can. And as we go through it, I want us to pay particular attention to our three I words, to the role ideas and ideals, on the one hand, of institutions. We're talking about changing institutions. Secondly and thirdly, the role of interests and how they play out as these ideas and institutions come under pressure to be reshaped. And the place you have to start if you're thinking about the international protection of human rights is the strong presumption against intervention in another country by external players. This hasn't always been that. In fact, if you read my colleague Oona Hathaway and Scott Shapiro's book in the Law School, they will tell you, and they are right about this, that the idea that national sovereignty is completely invaluable doesn't become fully entrenched in international law until the UN Charter is adopted after World War II. And the UN Charter was centrally about preventing the invasion of one country by another and preserving in that sense the nation state status quo in the world after World War II. And so the idea, the norm, that no nation can intervene in the activities of any other nation becomes enshrined in the UN Charter and almost as it's part of its raison d'etre. Because what we need to remember when we talk about the UN, talked about this briefly at the beginning of the course, but the purpose of the UN was not to be a world government. The purpose of the UN, first and foremost, was to prevent another war between the major powers. This is why it was structured in the way it was and this is why the Security Council, which is the closest thing to the executive body of the UN, is structured the way it was, with the five victorious dominant powers at the end of World War II having a veto over anything that the Security Council did, and the notion was that it would be better to force them to cooperate by having to agree before any international action could be taken. And when international action was taken, the default presumption was, the term of art was to restore international peace and security. And the idea of restoring international peace and security was to stop wars, to stop invasions, to stop anybody from trying grab real estate in the world. Of course, it came to take on many other roles, but that was the initial purpose of the UN, it's why it was structured the way it was, and we should also remember that it was basically a creature of a treaty, a multilateral treaty. In this sense, its basic structure is like the basic structure of the EU. Even though we saw with the EU there were efforts around the time of Maastricht to make the European Union, the Parliament elected by the citizens of Europe, sovereign. That didn't work, and by the time we got to the Lisbon Treaty, the Lisbon route, there was basically modification of this treaty-based structure. So the UN is based on a treaty, and that of course means, if you wanna change it, you'd have to negotiate a new treaty, at least in principle. And the treaty created the structure whose initial raison d'etre was to prevent war and certainly to prevent war between the major powers. So, a big part of the question, given that the UN is the sort of presumptive international institution in the world today, is, can the structure of the UN be changed? And if so, how? And what's interesting about the ICC, and particularly with the doctrine of Responsibility to Protect, is we might characterize them as sort of common law efforts, or quasi common law efforts to change the basic structure of the UN, to make it operate as more of an authorizer for intervention, much more radical intervention in the case of R2P than in the ICC. But nonetheless, to make it possible for this idea that nobody intervenes in the activities of any country, to start to be questioned. So, let's talk about the ICC first. If you think about the idea of holding war criminals to account for their crimes, previously, there had been talk going back all the way to 1919 of the idea that there should be some kind of international criminal court. It came up again at the end of World War II. But if you look at what was actually done, we had previous tribunals were all ad hoc tribunals. So we had the International Military Tribunal, commonly known as the Nuremberg Trials after World War II. The International Military Tribunal for the Far East, less well known, but the Tokyo Trials for Japanese war criminals. We had the Eichmann Trial in 1961, I believe it was, in Jerusalem, where Israel captured Adolf Eichmann in Argentine and brought him and tried him and executed him in Jerusalem. And then we had in 1993 the International Criminal Tribunal for the Former Yugoslavia. Again, an ad hoc, this time created by the UN for the first time, tribunal to try war criminals from the Yugoslav conflict. And the the International Criminal Tribunal for Rwanda that was created after the Rwandan genocide to try criminals for the Rwandan during the Rwandan genocide. So they were ad hoc, the last two, you see UN action. Before that it's essentially victor's justice that was often criticized as being, and certainly by people on the other end of it, when the winning powers in conflicts held trials of people from the defeated side. So there was no international body that could engage in these kinds of trials. The International Criminal Court was initiated by the Rome Treaty, which was signed by 120 countries in July of 1998 and finally ratified by 60 of them in 2002. And it has a strange relationship toward the UN. Because the structure of the UN was such, as I said to you, they couldn't create a court without going back to the signatories of the treaty. But they had a strong interest, they thought, or many people at the UN had a strong interest in trying to create some sort of permanent international legal tribunal. And so they commissioned a group of lawyers, called the International Legal Council, International Legal Committee, I can't remember the exact title, but these are experts in international law appointed by the UN General Assembly. And they asked them to draft a statute that could become the basis for an international criminal court, and they did indeed do that. The UN General Assembly sponsored multiple debates of this statute and eventually a conference, which led to the Rome Treaty. And the UN General Assembly was by no means alone. A variety of NGOs were involved in this as well, NGOs pushing for the creation of norms that would change international law so that perpetrators of war crimes could be held to account. It came into operation in July of 2002, when 60 states had signed it. And today there are 129 states are now signatories of the multilateral treaty that creates the ICC. So it's independent of the UN, technically, and it's a separate treaty that is the foundation document, the statute that creates this institution. It can prosecute, originally it was war crimes, crimes against humanity and genocide, and crime of aggression was added in 2010. It has pretty broad jurisdiction in that it can prosecute crimes that occur in any of the signatory countries or crimes that are perpetrated by citizens of any of the signatory countries. Generally speaking, crimes are referred for prosecution by the countries themselves, perhaps after a transition, for example, when there's limited ability of the existing legal order to go after people, perhaps 'cause they are not there anymore, or perhaps 'cause the country is in a civil war, it doesn't have developed judicial capacities. This is not the only way that they get crimes referred to them, but it has 18 judges who are elected by something called the Assembly of State Parties, which are the signatories to the treaty. And they'll sit in three different chambers, groups of three judges, one does pre-trial hearings, one does the trials, and one does the appellate work. I'd say it's a parallel but sort of independent institution in that actually the UN Security Council has begun to refer cases to the ICC. This happened after Darfur when the Security Council referred some cases they had. And once a case is referred, there's a prosecutor for the ICC who does a preliminary investigation to see whether charges should be brought. And then in 2011, in the Darfur referral, the US and China abstained, and this was partly because the US and China are not signatories to the ICC, for reasons I'll get to in a minute. But in 2011 the Security Council unanimously referred Gaddafi for prosecution by the ICC. And as Mark Mazower comments in the chapter that I had you read for today, this is part of what has provoked a lot of commentary about the hypocritical attitude of the US toward the ICC because on the one hand we have not signed the treaty and we refuse to be subject to its jurisdiction, and indeed, in peacekeeping operations, if the Security Council authorizes peacekeeping operations, the Americans require that their soldiers cannot be subject to the jurisdiction of the ICC as conditions for participating in peacekeeping operations. So it's not a legal system to which the US will submit. On the other hand, here they're participating in referring others to the Security Council, most dramatically Gaddafi in 2011. So, how important is this institution? As of 2019, it had handed down 44 indictments, it had issued 36 arrest warrants, it had 23 ongoing proceedings, there were 15 fugitives at large who had had arrest warrants out for them, 21 proceedings had been completed. So this is from 2002 to 2019, so in 17 years six people have been convicted, four had served prison terms and been released, two were still serving them. The rest had been acquitted, charges had been dismissed, charges had been withdrawn or otherwise fallen apart, and four people had died before trial. So, tiny numbers, tiny numbers you might say. How should we then think about it and its legitimacy? Well... One thing to say about it is that it's long been criticized, particularly in the Global South, as heavily biased against African countries. All the convictions to date have been of people in African countries, and virtually all the prosecutions. Not every Wikipedia page is good but there's a very good Wikipedia page that gives a running account of all the prosecutions currently ongoing, and it stays up to date. And the sense that it was anti-African became a source of a lot of criticism of it, and indeed, even today, although prosecutions have been opened, one in Myanmar and one in Georgia, but the vast majority of them are African countries. And in 2016, Burundi, South Africa, and Gambia all said that they were going to leave. There was then an election in Gambia that reversed that decision after a new government came to power. In South Africa, interestingly enough, it was the Zuma government that said they were leaving, but the South African Constitutional Court ruled that it would be illegal for South Africa, it'd be unconstitutional for South Africa to leave, so they rescinded their decision to leave. But the Philippines, after Duterte's anti-drug efforts, where he keep a very violent anti-drug campaign there, were referrals to the ICC for prosecutions in the Philippines, the government announced that they were leaving, and they left, it takes a year to leave, and incidentally does not immunize a country from prosection during the period when they were members of the ICC. So it doesn't mean that, in principle, the crimes committed in the Philippines couldn't still be adjudicated. Israel, Sudan, the US, and Russia have all withdrawn their signatures from the treaty. And the US role over time bears some attention, given that we're now in a unipolar world and the American role is so important. So, the last possible day you could sign the treaty creating the ICC was the last day of the century and the last day of the Clinton, not quite the last day of the Clinton administration, it ended the following January. And the Clinton administration signed it but you can see from what President Clinton said, he said, "We reaffirm our strong support "for international accountability "and for bringing to justice perpetrators of genocide, "war crimes, and crimes against humanity. "We do so as well because we wish to remain engaged "in making the ICC an instrument of impartial "and effective justice in years to come." So they were signaling they didn't like it but they were gonna remain engaged and essentially punt this question to the next administration. They're gonna punt the question to the incoming Bush administration, but they didn't wanna forestall the possibility of the US being involved. The Bush administration came into office, as we all now know. A year shortly after we had 9/11. The Bush administration before that had taken no action to move forward with American involvement in the ICC. After 9/11, in 2002, the Bush administration withdrew America's signature, so that was the end of that as far as the ICC was concerned. When the Obama administration came in in 2009, they took a somewhat softer attitude toward the ICC and began to engage with it and cooperate with it in various investigations and have observer status at meetings about it and of some of its organ bodies. But even the Obama administration showed no interest in actually having the US signed, partly for the same reason that I'd already mentioned, which was the idea that the US involvement in peacekeeping operations would lead Americans to be vulnerable to prosection was not something that the administration thought it could sell to American voters. Of course, since then we've had the Trump administration, and as you might surmise, that hasn't produced any change either. - The United States will provide no support in recognition to the International Criminal Court. As far as America is concerned, the ICC has no jurisdiction, no legitimacy, and no authority. The ICC claims near-universal jurisdiction over the citizens of every country, violating all principles of justice, fairness, and due process. We will never surrender America's sovereignty to an unelected, unaccountable, global bureaucracy. - So, there you have an unsurprising summation of the American attitude and reaffirmation of the idea that international law applied to American citizens is not gonna be acceptable. So, how should we think about whether the ICC is a good or bad thing? Any, any... Any thoughts about that? Anyone thinks it's a good thing? Good thing, somebody who thinks it's a good thing, yeah. - [Man] A metric could be the number of people who are killed in interstate conflict. - So you say it would be whether it's a deterrent or not. - Yes. - Okay, one metric, would it be a deterrent, yeah. And what else, anyone else? So you're just saying that's the metric, you're not sure whether it's a good or bad thing. You think it's a good thing? Somebody who thinks it's a good thing, why is it a good thing? Yes, sir? - [Man] I think it provides an infrastructure for countries who have gone through particular bad situations, war crimes, genocides, to have access to an institution that is judicial in its form when otherwise it wouldn't have it. - Any other thoughts? Anyone else think it's a good thing? Yes, sir? I see one hand here. - [Man] Thank you. So, I think it's not effective in fact, but it's a good symbol in the world. - It's a good symbol. A symbol of what exactly? - [Man] To show the accountability of those-- - That people can be held accountable for what they do, that dictators or people engaged in wars can, there's some chance you might be held accountable. It stands for that idea. So okay, anyone think it's a bad thing? The truth is it's tricky to evaluate. So, the small scale might not matter that much because the question is not whether it's small but perhaps whether it's scalable. And there's a literature in political science for which Martha Finnemore and Kathryn Sikkink are famous where they talk about norm cascades. And the idea is that once a norm starts to get established, provided a critical number, and scholars argue about how many, endorse the norm, others join up as well. And interestingly on that front, in 2016, when South Africa and Gambia and these others were threatening to storm out, there were a lot of predictions that it was all gonna fall apart and that there was gonna be a massive African exodus. That didn't happen. So to the extent that, they also talk about norm entrepreneurs, people who are trying to, this is the realm of ideas now of our three I words, that norm entrepreneurs are trying to get enough countries to start investing in a process that do then reach this tipping point where it will start to become embedded, it might become scalable, and so on. The deterrent effect obviously is very tricky to evaluate, in the sense that at the level of what lawyers call specific deterrents, that's deterring actual individuals from committing crimes, most people, this is sort of parallel to the debates about the death penalty in the US, most people who murder people are not deterred by the death penalty because the point of murdering someone is a point, first of all, nobody's thinking about being apprehended and it's usually seething hatred, passion, et cetera. And so it's not clear it has any specific deterrent effect. Where would it deter others? That's what we call general deterrence, it's hard to say. I would just point out one thing that nobody has really fully thought through about this doctrine, and this is think about about my earlier discussion of Hirschmann in connection to democratic transitions. I said that one of the things that needs to happen to get dictators to give up power is that the cost of exist must be low, relatively low, otherwise they're gonna hang on until the end of time. And so, for instance, Idi Amin got to live out his life, ironically, in Libya, after he was forced out of power. He had a place where he could go and he was not prosecuted for anything. Gaddafi, on the other hand, was referred to the ICC. He was killed before he was ever brought to trial. So, if you're thinking about amnesty as part of a negotiated settlement for a conflict, how that would square with this doctrine is a difficult thing to think through, and there are obviously trade-offs there. So, it's not obvious, it's not obvious how we should think about this doctrine, but it is alive, it sort of wobbled into existence, it is subject to all these criticisms, it's inequitably applied. The US in particular has been rather hypocritical in the way that it's used the doctrine that it won't allow itself to be subjected to. But the norm entrepreneurs would argue this is the way that norms gradually get established and embedded in the international order, and so the goal should be to improve it and to address its defects rather than say that it's just an illegitimate way in which some countries beat up others. Let's turn to the Responsibility to Protect, which is in some ways a much more ambitious rewriting of the international rules. Anyone know what that is? - [Man] Is it (mumbles)? - Pardon? - [Man] A fire in India at the-- - No. Good reasonable guess, yeah. - [Man] Is it (mumbles) genocide and war crimes over (mumbles) in Stockholm? - Well, it's actually a NATO bombing of Kosovo in 1999. And the NATO bombing of Kosovo in 1999 came after many years of civil wars that had basically broken out with the collapse of the Soviet Union and Yugoslavia had come to pieces. And there was ethnic cleansing of Albanian Muslim Kosovars at the hands of Serbian militias going on. And the NATO alliance, notice, the NATO alliance took a decision to intervene. Here's President Clinton talking about it. - The Kosovars have been victims of terrible atrocities. Their only hope was that the world would not turn away in the face of ethnic cleansing and killing, that the world would take a stand. We did, for 78 days. Because we did, the Kosovars will go home. To achieve our aims as an alliance, 19 democratic nations with 780 million people working together in the first sustained military operation in NATO's history. The alliance did stay together. It is now stronger and more united than ever, and I thank my fellow leaders in the alliance for their fidelity and their fortitude. In the past four months we've seen some of the worst in humanity in our lifetime, but we've also seen the bravery of our troops, the resolve of our democracy, the decency of our people, and the courage and determination of the people of Kosovo. We now have a moment of hope thanks to all those qualities, and we have to finish the job and build the peace. - So. What had happened here was that this ongoing civil war had resulted in many human rights violations on all sides. But this was the point at which, when there was this imminent ethnic cleansing in Kosovo, that NATO decided to intervene in mid 1999. And one thing about this intervention was that it was widely judged to have succeeded, that is to say the ethnic cleansing ended and didn't re-initiate. Even though there were no NATO troops on the ground in Kosovo, there were European Union troops massing on the border, some, but they never actually went in. It was very unusual for two reasons. One was that this was a NATO operation and it was in violation of the NATO charter. We've talked about NATO. You know that Article 5 authorizes NATO members to use force and indeed to turn to one another to use force if any one of them was attacked, but no NATO member was threatened during the Kosovo and the former Yugoslavia civil war. There was no threat to any member of NATO. So this was not a legitimate activity for NATO as a defensive alliance to have been engaged in. It was also not authorized by the Security Council. Why was it not authorized by the Security Council? Because the Russians were basically on the Serbian, sympathetic to the Serbian side in this conflict, and clearly the Russians and almost certainly the Chinese would have vetoed any attempt to get authorization for NATO to get involved in this conflict. So there was no Security Council authorization, and it was not anything that one could legitimately say was part and parcel of NATO's mandate to protect the member states. They just did it. They simply did it. Now, one might say, what were Clinton's motives? And some have said his motives were to divert attention from Monica Lewinsky. Christopher Hitchens, among others, claimed that. But even, setting Monica Lewinsky aside, I think part of Clinton's motivation surely was to redeem himself. He had taken so much criticism after Rwanda for the US's failure to act in the Rwandan controversy, and here he was on the verge of leaving the presidency, this is summer of 1999, and he thought, "This is my moment to redeem myself "and to show that the US will in fact act." And indeed, he enunciated what some of his biographers have called the Clinton Doctrine. He said here, shortly after that press conference I showed you, he said, "Never forget if we can do this here "and if we can say to people of the world, "whether you live in Africa or Central Europe "or any other place, "if somebody comes after innocent civilians "and tries to kill them en masse because of their race, "their ethnic background, or their religion, "and it is within our power to stop it, we will stop it." Very different than the stance he took over Rwanda. And the Secretary General of the UN was Kofi Annan, the Ghanaian Secretary General now deceased, he got behind this. He said, "If humanitarian intervention is, indeed, "an unacceptable assault on sovereignty, "how should we respond to a Rwanda, to a Srebrenica, "to gross and systematic violations "that affect every precept of our common humanity?" And his answer was, "The sovereignty of states "must no longer be used as a shield "for gross violations of human rights." This is a dramatic statement in strong tension, to put it mildly, with the charter of the UN. But the question becomes, how can you do that? How does one do that, particularly given a world of nation states whose sovereignty is protected? And so there was a lot of hand-wringing and wrestling with this problem, and, interestingly, the Prime Minister of Sweden took it upon himself to appoint a Commission on Kosovo, which was established in August of 1999. And the Commission reported with the following statement. It said, "The Commission concludes "that the NATO military intervention "was illegal but legitimate." (audience laughs) It was illegal but legitimate. "It was illegal because it did not receive prior approval "from the Security Council. "However, the Commission considers "that the intervention was justified "because all diplomatic avenues had been exhausted "and because the intervention had the effect "of liberating the majority population of Kosovo "from a long period of oppression under Serbian rule." Illegal but legitimate. What can that mean? What do we, can we say something is illegal, yes? You need a microphone, we'll need a dissertation here. Wait a second, get the microphone. They'll never hear you otherwise. - [Man] Yeah, I mean, it just seems like it's the international community trying to reconcile the contradiction of, that you just put, national sovereignty but sometimes intervention's good. This seems like you can't have war or the idea of armed intervention that still respects national sovereignty, and so it seems like this is just a halfway return to the pre-World War II consensus of, well, national sovereignty is a given. - So, it certainly is a different conception of national sovereignty, but we're saying it's illegal but legitimate, we can say we can violate international law when it's legitimate for us to do so. But what legitimates the action? How is it legitimate, what makes it legitimate? - [Man] Perhaps thinking about the ideas that interest institutions? It's a matter moral interest, and especially power interest, than institutions in itself. - One more at the back, yeah? Oh okay, yes, sir? We'll take two more, you and then you, yeah. - [Man] I would just say R2P trumps the international order. - Say that again. - [Man] R2P trumps the international order. - Responsibility, well, we don't yet have a doctrine of R2P at this time, right? So we will have to see, but yeah. Okay, the gentleman at the back, the last one, yeah. - [Man] The ends justify the means. - The ends justify the means. So another way of saying that and perhaps capturing what some others were saying is perhaps sometimes it's better to ask for forgiveness than permission. (audience laughs) Right? Sometimes it's better to ask for forgiveness. If you know some horrific thing is about to happen that you can stop, and you also know that if you ask for permission, it's gonna be denied, maybe you'd better just do it. Maybe you'd better just do it. And moral philosophers have debated this principle, or maybe it's a pragmatic judgment and not a principle, but there's a principle behind it, presumably, which is that some things are so egregiously bad and so frightening that if the legal order will not permit it, it's too bad at some level for the legal order. And so this was I think the thinking behind this idea of illegal but legitimate. But of course, how do you turn a norm like that, how do you turn a dictum like that into a norm, much less a rule, how can you institutionalize the idea that it's okay to ask for forgiveness rather than permission? Not obvious how you can do that because the whole idea is you're acting against the existing institutions because you've concluded that it's the right thing to do. So how might one institutionalize that? And it's important to say that no everybody agreed. Not everybody agreed. We say the international community this, the international community that. The international community didn't agree. Not everyone thought that Kosovo should become a precedent, especially in the Global South. The logic was seen as a potential smokescreen for major powers to pursue their own agendas. Within months, the G77, this is a large number of developing world countries, condemned the so-called humanitarian intervention. South Africa and India amplified this criticism. Nelson Mandela, who had recently stepped down as South Africa's president, said, "When two nations," he wasn't fooled by the idea that it was a large coaliltion, it was basically the US and Britain, "take it upon themselves to police the world "without getting authorization from the UN, "we must condemn that, "because it can lead to another world war." The South African government, reflecting the positions of the Non-Aligned Movement and the OAU, declared, "Unilateral intervention, "no matter how noble the pretext, is not acceptable." India called the Kosovo operation "a flagrant violation of all international norms, "against the provisions of the UN Charter, "and seen as a direct and unprovoked aggression." So, as I said, there wasn't consensus about this by any means across the Global South, particularly because this was a NATO operation led by the US. So that led to the, it was followed by, I shouldn't say led, it wasn't caused, but it was followed by an International Commission on Intervention and State Sovereignty that was convened in London, again, with input from the UN and various NGOs. And it was chaired by Gareth Evans, a former Australian prime minister, Mohamed Sahnoun, who was an Algerian diplomat, and a major player in it was the Canadian academic and sometimes not very successful Canadian politician Michael Ignatieff. And they came up with this idea of the Responsibility to Protect. We'll talk in some depth about it in a minute but I want to note, first of all, that the concept of the Responsibility to Protect is as important for what it doesn't say as for what it does say. It doesn't say that there's any right. It's talking about a responsibility. It doesn't say that there's a right. The responsibility here is not, there's not a responsibility on the UN or on the Security Council but a responsibility, and the phrase Responsibility to Protect refers to governments. Governments have a responsibility to protect not only their citizens but people living within their territories. And it doesn't talk about an obligation to intervene. Even though it suggests there's some kind of imperative for action, it doesn't say when you have certain kinds of rights violations, there's an obligation to intervene. I'll say of this more later. Again, in the Global South, this was not welcomed. And the African Union said, "Well, if the UN is gonna start taking it upon itself "to authorize intervening in other countries, so are we." And so the African Union asserted the right "to intervene in a member state," that's pretty much, I think there's 50 members of the African Union, almost every African country, "pursuant to a decision by the Assembly, "in grave circumstances, namely war crimes, "genocide, and crimes against humanity." And notice that there's no mention of the Security Council despite the fact that the UN Charter reserves to the Security Council a monopoly on declaring non-defensive use of force. So, essentially, the African Union is saying, "Well, if the Security Council or NATO or other groups "are going to start intervening in countries, "then so are we." And so the stability envisaged in the NATO Charter could be seemed to be threatened. In 2005, the UN General Assembly debated this so-called World Summit Outcome Document and adopted it, behaving now as a kind of legislature. That is to say, remember what I said at the beginning of the lecture, that the UN is a creature of a treaty, the General Assembly is not a legislature. But in any case, they adopted this provision and they said, "Each individual state," country, "has the responsibility to protect its populations "from genocide, war crimes, ethnic cleansing, "and crimes against humanity." These all have complex legal definitions which people wrangle over. "This responsibility entails the prevention of such crimes, "including their incitement, "through appropriate and necessary means. "We," this is the UN General Assembly, "accept that responsibility "and will act within accordance of it. "We are prepared to take collective action, "in a timely and decisive manner, "through the Security Council, "on a case-by-case basis and in cooperation "with relevant regional organizations as appropriate, "should peaceful means be inadequate "and national authorities manifestly fail "to protect their populations from genocide, war crimes, "ethnic cleansing, and crimes against humanity." So here you have the UN General Assembly taking it upon itself to say there's a responsibility to protect, it falls on governments, and if they don't live up to that responsibility, the Security Council can authorize regional organizations and other players to intervene. That's hugely important because the UN does not really have much of an army, and UN peacekeepers traditionally have only gone into conflicts once peace is established, as we saw at the very beginning with the Rwanda picture. They don't go in to actually get involved in peacemaking, but only in peacekeeping. So some things to think about with the Responsibility to Protect. First of all, unlike the ICC, it's binding on signatories and non-signatories alike. So they have declared unilaterally that this doctrine is binding on all governments. They restricted it to these four major categories of crimes: genocide, crimes against humanity, war crimes, and ethnic cleansing. Even though the (mumbles) commission had wanted a much more expansive set of things to protect against, but the thinking was that in order to get a majority to accept it, it would have to remain focused on the most egregious violations. The domestic focus here seems that, indeed, does contradict the UN Charter's focus on international peace and security, because now, in effect, what R2P has said is that it is legitimate for the international community, as represented by the United Nations and authorized by the Security Council, to intervene in the domestic affairs of a country, if it's triggered by one of those, genocide, crimes against humanity, war crimes, or ethnic cleansing. Another feature of it that's worth noting, and this, again, goes to my earlier observation that there's no right to demand intervention. It's not binding on the Security Council to intervene. As I said, it gives wide latitude as to when the Security Council will be intervening, because it intervenes on a case-by-case basis. So the Security Council may intervene, but there's nothing in this doctrine which says the Security Council must intervene in any particular conflict, and that will turn out to be important. The Non-Aligned Movement didn't like it. And the Non-Aligned Movement is a group of mostly middle-income countries around the world. So, how should we think about this doctrine? This is obviously, as I said, both the ICC and the Responsibility to Protect are starting to rewrite the international architecture of intervention, but this is a much more ambitious rewriting than the ICC and we're now not talking about prosecuting a few individual perpetrators, we're actually talking about intervening in countries during ongoing conflicts to make peace and to stop gross human rights violations. Is this a good move? Who thinks it's a good move, somebody wanna... Okay, yes, ma'am? - [Woman] More norm setting? - So say a little bit more about the norm setting. - I don't know much more. (audience laughs) It seems like it would be a good norm to have for humanity. - Okay, it's certainly a norm. It seems like it's gone beyond a norm now, right? This is really changing international institutions, right? We now have the Security Council in a position to actually intervene. It is, if you like, a common law, by which I mean a gradual but nonetheless decisive rewriting of the rules governing UN intervention within the affairs of governments. It doesn't mean it's good or bad, but it's very expensive. Any other? Yes, sir? Wait, wait a second. - [Man] I think it's highly contradictory, and any country can, depending on its own utility, can decide whether to consider it as a genocide or not. And I can give two examples. The US was highly quick in setting the war affairs in Syria, while in Ukraine, when the radicals just broke into the Opera House in Odessa, it was kind of okay because-- - So there may be disagreement as to whether the conditions prevail or not? At least in some cases. - [Man] I mean I think it can be still biased, depending on the-- - It can be manipulated. - Yes. - It could be manipulated, but any doctrine can be manipulated. Just to play devil's advocate to your comment for a minute, let's go back to Rwanda in 1994. The fact is no one had any doubt in Rwanda in 1994 that genocide was going on. You could go back now through the archives and it's perfectly plain that all the major powers knew that genocide was imminent and then that it was starting and then it went on for many months. Now, they might not have been able to stop it, it might not have been successful, but the book I put on the syllabus I think is probably the best on this. Certainly the estimates are that of the 800,000 to a million people killed, had there been intervention, at least 200,000 people could've been saved. So you could say, well, there'd still be 600,000 killed. Yes, but 200,000 people's not a trivial number of people to save, even if that's double what really could have been, it still would've been. So that would be, there might be cases where it's manipulable, there might be cases where it's unclear, where it's disputed, and we'll talk about one of those cases in a few classes, three classes from today. So, yeah. Any other thoughts about whether this is a good way for international law of intervention to evolve, yeah? - [Man] Within the constraint of there being a veto, it's an establishment of some kind of world government as a legitimate user of force, and I would support that. - You like the idea that there's a veto? - [Man] No, I like the idea that there's a legitimate use of force in this instance. But it does basically change the charter of the United Nations. But it has the constraint still that there is veto-- - There's a veto in the Security Council, yeah. The permanent members, there are now permanent members and non-permanent members of the Security Council, but the original five still have veto. Okay, so. Anyone think it's a bad idea? We'll defer judgment of that question a little while. So here were some early tests of this doctrine. This is actually before the Outcome had been adopted by the General Assembly in Sudan. The Security Council called for an arms embargo and referred to the idea, it was still being debated then. In Kenya in 2006, after the election, there was an eruption of ethnic violence, 800 people were killed, more than 1/4 of a million were displaced, and the French called for a Responsibility to Protect action to intervene. Kofi Annan headed it off, basically, by going and negotiating a settlement. And some argued that it was the threat of the possibility of intervention that actually helped him negotiate or mediate the conflict in a way that headed off the violence, or damped it down. In Guinea in 2008 there was an attempted coup. The leader Camara's troops rioted against demonstrators, and by then it was Ban Ki-moon who was Secretary General. He set up an inquiry and raised the possibility of crimes against humanity and started again talking about R2P. The US and the EU suspended assistance but it became moot after Camara fled the country. And so, again, they were on the verge of intervening on the grounds that there were gross human rights violations going on, but then they actually didn't. And then in 2011, I mentioned to you earlier, that the Security Council unanimously, including the US, referred Gaddafi to the ICC for prosecution. But in the course of that referral, they talked about the Responsibility to Protect being triggered possibly in Libya. This was in February of 2011. They were hinting already that there might be, this is following the outbreak of the Arab Spring. We've had eruptions in Tunisia and in Egypt and now in Libya. Civil war breaks out and there were claims, disputed claims, about imminent slaughter in Benghazi. Much disputed, and we will talk about the disputes later. In any event, they referred Gaddafi to the ICC and they talked about the Responsibility to Protect and used that to impose an arms embargo on the country and various other sanctions. The first real test of R2P came the following month, when the Security Council adopted UN Resolution 1973 authorizing intervention in the Libyan conflict. And we'll return to that on October 24th and we'll then think about, at the end of that examination we'll see what we think about the future prospects for the doctrine of Responsibility to Protect. But first we need to attend to the Global War on Terror, which fundamentally changed the international legal and political landscape in ways that affected the prospects for doctrines like the Responsibility to Protect, and we'll do that starting next time. (gentle music) |
Power_and_Politics_in_Todays_World | Lecture_19_Crisis_Crash_and_Response.txt | - So, today, we're talking about crisis, crash and response. So just to remind us of where we were in September of 2008. - [Narrator] This is gonna be one of the watershed days in financial markets history. - [Narrator] It was a manic Monday in the financial markets. The Dow tumbled more than 500 points after two pillars of the street tumbled over the weekend, Lehman Brothers, a 158 year old firm, filed for bankruptcy. - I don't think anyone really expected a bank as big as Lehman to, you know be in a position that it's in now. - [Narrator] Brought down by bad mortgage investments. Lehman which has 25,000 employees will be liquidated. - Am starting to another job. - [Narrator] Meanwhile, Merrill Lynch, fearing it could be next, agreed in an active desperation to a shotgun marriage with Bank of America. Merrill, the country's biggest brokerage with 60,000 employees has been battered by nearly $50 billion in mortgage related losses. - It is definitely a very, very difficult time and it's not gonna get better quickly. - [Narrator] So in just six months, three of the five biggest independent firms on Wall Street have now disappeared. Bear Stearns, which collapsed last spring, Lehman Brothers and Merrill Lynch, Treasury secretary Henry Paulson tried to reassure investors today, - The American people can remain confident in the soundness in the resilience of our financial system. - [Narrator] Paulson attempted to broker a deal to sell Lehman over the weekend. But unlike the buyout deal for Bear Stearns, the government would not offer any financial guarantees. - I never once considered that it was appropriate to put taxpayer money on the line in resolving Lehman Brothers, - it would have required an $85 billion subsidy to keep afloat. - [Narrator] Insurance giant AIG is the next name on the list of troubled companies. It's looking for $40 billion in bridge loans. Veteran trader Art Cashin believes there could be more casualties. - This is the fifth time we've seen this movie. And you sit on the edge of your seat and you yell at whichever character it is. Don't go into that woodshed it was they keep going in. - So, he was understating things that of course at the time, anytime the treasury of the secretary comes out on national television to assure you that the financial system is fundamentally sound, you know that people are panicking, and they indeed were. They were speculating at that time, the AIG bailout would cost $40 billion, it ended up costing four times that. And the crisis just stumbled on. And you know, this is one of many pictures one could. This is just the Dow Jones Industrial Average, and you can see there by the end of September when banks were failing faster than people could count them. And Hank Paulson, who you saw there, Henry Paulson, he had been a former head of Goldman, and indeed he had led the group of bankers in the run up to the financial crisis that I told you about some time ago, who came and petition the SEC, under the tutelage at that time of congressman Christopher Cox, on arguing that the biggest banks, those with assets in excess of 5 billion could have relaxed capital requirements because they were so big that they could self insure he was by now Secretary of the Treasury. And the reason this turns out to be a significant date is that he had come to Congress on the 29th of September, asking for 700 billion to bail out failing banks. And they had come with it an extremely brief request for a blank check. And, of course, the congressional committees went crazy. And even though the Bush administration had assured them that it was a done deal, and Congress would vote it up, they didn't vote it up. They voted it down. And, the net effect of that was that the following day, the Dow lost another 700 points and the catastrophe just kept on going. And of course, it wasn't just the Dow. If you wanna get some sense of how rapidly wealth was evaporating, this you can see here from the peak to the trough, basically $16.4 trillion in household wealth disappeared. So people were absolutely stunned. They didn't think this could happen. I mentioned to you in an earlier lecture, Ben Bernanke's famous speech about the great moderation that where he had speculated that the regulators had finally figured out how to tame the business cycle, and that things like this, we're not going to happen again. And just to give you a sense, I should say, everybody prime sufficiently spoke that three days later, Congress did pass a version of the TARP bailout. And, with much more regulatory oversight and a promise, which turned out not to be worth the paper, it was written on that this wouldn't support institutions that were paying bonuses. We'll hear more about that shortly. So Congress had passed, had passed the bailout. And later that month, just to give you a sense of the intellectual shock, that was going through the elites, Alan Greenspan, who had been chairman of the Fed until 2006, and widely a being called the maestro and the greatest genius. This is him, later in that same month, just two, three weeks after Congress had finally adopted the legislation, talking to the House Oversight and Government Reform Committee. [narrator] General with status... - Those of us who have looked to the self interest of lending institutions to protect shareholders equity, myself especially, are in a state of shocked disbelief. Such Counterparty surveillance is a central pillar of our financial markets state of balance. If it fails, as occurred this year, market stability is undermined. - You found a flaw in the reality... - Flaw in the model that I perceived is the critical functioning structure that defines how the world works, so to speak. - In other words, you found that your view of the world your ideology was not right. It was not working. - Precisely. No, that's precisely the reason I was shocked because I've been going for 40 years or more, with very considerable evidence that it was working exceptionally well. - where do you think you made a mistake then? - I made a mistake in presuming that the self interest of organizations, specifically banks and others, was such as that they were best capable of protecting their own shareholders and the equity and the firm's. - So there you have it from the maestro himself confessing how shocked he was, What I wanna do today is look at the 2008 and subsequent choices, first in historical perspective, and then we're gonna bring to bear some of the instruments that political economists used to talk about regulation, in thinking more analytically about the choices. So just to give you a sense of how this was unfolding, of course, this was unfolding over the summer and fall of 2008 when we were in the midst of a presidential campaign. And Barack Obama had been the unlikely winner of the Democratic nomination against Hillary Clinton and was now doing battle against senator John McCain, who was the Republican. And, during the height of the crisis in September of 2008, when actually John McCain had suspended his campaign, but Obama famously said, well, you need a president can do more one thing at a time, so he didn't suspend his campaign, but he came to Washington, and here you can hear him opining on the circumstance. - Late this afternoon, I talked to Senator Obama about the bailout and his concerns. - Well, it's not exactly clear. What happened was when I left Florida, landed, I had heard the news that in fact the deal was there. People are not obviously happy that we got here in the first place and we shouldn't have gotten here in the first place. It's a consequence of speculator greed, but also Washington not paying attention, and regulators being asleep at the switch. So there's resistance. But there had been at least some core principles that were met. I think that from the Democratic perspective, we just wanna make sure that this is helping the small businessman make his payroll, people keeping their money market accounts, that we don't see a cratering of the credit markets that could end up leading to enormous job loss. And obviously, it's pretty frustrating for Democrats, having seen the mismanagement that's been taking place over the last several years, to feel like we've gotta step in and get something done. But that's how I think many of us feel that, that we can't worry about how we got here. Now we've gotta take some serious steps. - We can't worry about how we got here. Now we've got to take some serious steps. And that we've very much reflected the attitude that Obama and the Obama White House brought to this problem. He, and if you can read all the pop books about the early years of the Obama administration, too big to fail is one excellent one run siskins confidence, man is. Another, But basically, they assume that the people who were in the midst of the crisis were the people that knew how to fix it. And I should say right off the bat, they were not wrong about this, in many respects, as we will see, but this meant that Obama took on people who had been greatly implicated in the deregulation of Wall Street over the previous decade and a half people like Larry Summers. And he appointed as his treasury secretary Timothy Geithner. who often would be attacked in congressional hearings as being in bed with Wall Street banks and being accused of being a former investment banker, in fact, it wasn't true of Geithner. Geithner had never been a banker. He had always been a bureaucrat, he had been chairman of the New York Fed until Obama appointed him as Secretary of the Treasury, and so Geithner, took responsibility for fixing the problem. And he did work very closely with Wall Street. And as a result of this, he very much took this attitude that we shouldn't worry about how we got here. What we've got to do is think about how to get out of it. And Obama, bet on him to the extent that produced some considerable grumbling in the administration. And just to give you a sense, the following March now President Obama out there defending Geithner's efforts, particularly after the huge scandal had erupted, that not only was AIG the insurance giant being bailed out, in, as I said, some multiple of what had been speculated a few months earlier, but they were paying bonuses to their executives. So the administration was very much back on their heels and here's President Obama, defending Geithner. - [Narrator] With anger mounting over the bonuses awarded to AIG executives, The President is defending his administration's handling of the fiasco. - Nobody here drafted those contracts. Nobody here was responsible for supervising AIG and allowing themselves to put the economy at risk. - [Narrator] AIG paid out $165 million in bonuses last weekend. The federal government has propped up the failed insurance giant with $170 billion in taxpayer money. And angry president says the buck stops with him. Ultimately, I'm responsible. I'm the president of the United States. Nobody's working harder than this guy. He is making all the right moves, In terms of playing a bad hand. - [Narrator] Despite assurances from the president, Some has speculated that Geithner's job could be in danger. In part because he couldn't persuade AIG to cancel the payouts. - So there you have it. President Obama had embraced the TARP bailout as a candidate, and indeed, the version of it that had been passed on the third of October in 2008, it was when George W. Bush was still president had to be passed up substantially by Democrats. Most Republicans in the House voted against it. And so the incoming administration was very closely associated with the efforts to bail out the banks and to protect the bankers and again, just to give you a sense, a week after that, Obama came out onto the White House lawn and defended Geitner. Here you can see here you can see him being raked over the coals in a congressional committee hearing. - Just last month, AIG paid 343 employees of AIG FP, their financial products division, that created the financial hole that AIG is in. And in turn a multi billion dollar bill for American taxpayers of 56 million dollars in bonuses and are slated to pay an additional $162 million in bonuses to 393 participants in the coming weeks. And there's more. A further bonus payments totaling approximately 230 million are due to 407 participants at AIG Financial Products Division in March 2010. This makes no sense to my constituents. This company claims to be on the brink of disaster, and it's handing out bonuses. - I just wanna point out that compensation practice across the financial services industry over the last years and decades, just got out of whack with basic fundamentals and people were paid for risks that weren't captured in compensation. And part of what we do to make sure this kind of crisis doesn't happen in the future is to change those basic incentives, and there's gonna be a role for government doing that. Now, it's very important that we make sure that we're providing exceptional assistance to these firms, that that assistance is going again to achieve the objectives of these programs. Not to reward the kind of executives that got us in this mess. I'm deeply committed to that objective. And we're gonna be as careful and responsive as we can to the concerns you have, so many Americans have about how these resources have been used. And I just want to say that the judgments made by these boards of directors and senior executives across parts of the finances have just caused a lot of damage to public confidence in the quality of their judgment. And they have a deep responsibility and obligation to make more careful judgments going forward. But we're not gonna depend on them to do it, we're gonna make sure that there's conditions that come with our assistance to assure that. So we're gonna figure out how to apply these new provisions in a way that is carefully designed as possible. Obviously, we want these businesses to achieve the objectives of what we did in the AIG, we want them to be able to run their business and restructure, so we're in a better position going forward. And, that's why this is sort of hard to do. - So that gives you a flavor of him trying to explain why they had to bail these banks out even though they were, in this case, AIG even though they were paying these bonuses and it, in fact had been a big fight in the White House, Obama had been appalled by it and demanded that they find a way to claw them back. But Geithner and Summers and others said these are contractual obligations. We have to enforce contracts, we can't renege on them. And so that view prevailed. And so this was the catechized, if you like, behind the way in which the Obama administration essentially owned the regulatory response to the crisis in ways that would be costly for them and their successes in ways that they couldn't appreciate at the time. Now, I said I wanted to supply some historical perspective on this response, to the crisis. And of course, this is the biggest financial crisis since the Great Depression. So the place to compare it with, or the events with which to compare it and worth a little bit of reflection, is how is the response to the depression. So, in 1932, Franklin Roosevelt won the election, and by the time, by the time he won the election, in fact, the Hoover administration, people tend to run Hoover and Coolidge together, by the way, big mistake. Hoover was a quite a much more enlightened person about macroeconomics, among many other things than his predecessor had been, and if you read the literature on the Hoover administration, one of the things you learn is that by 1930 30, 31 the administration was actually trying to implement much of what FDR would subsequently do. We can call it Keynesian because the General Theory wasn't published until 1936. But essentially, many of the sorts of measures that would subsequently be adopted by the Roosevelt administration were being pushed. American agriculture had been in a recession for over a decade before the crash of the stock market in 1929. They had much higher unemployment through the most of this period, topping 20% and they were looking for ways to spend money and stimulate the economy and after the election, Hoover begged Roosevelt to do what we just saw Obama doing after the 2008 election to endorse in the name of rescuing the American economy to endorse the proposals that the administration was pushing, which he knew. Roosevelt agreed with many off, and Roosevelt flatly refused. He said, this is your government. This is your crisis, you deal with it. And we should also remember that this is in 1932, when the new administration didn't take office until the following March, so we're really talking about a five month interregnum during which Roosevelt left the administration essentially twisting in the wind and made it much more difficult for them to, to see, to implement the sorts of proposals that would have stimulated the economy or at least what most people and certainly in the Hoover White House thought would stimulate the economy. Finally in March of the following year, Roosevelt came in and passed a lot of what we now think of as the first New Deal. And that's I think I said to you, when we were talking about welfare states, the early part of the New Deal had a lot of business acquiescence, if not business support, partly because of a fear of communist influence in American trade unions and a worry that American workers might indeed start to conclude that they had nothing to lose but their chains and so there was some thought among business elites that they should support the New Deal at least at the beginning, by 1935, that was pretty much over. If you look the Glass Steagall Act, which had been passed in 1933, which I'll say a little bit more about shortly, they were already beginning to lobby against it and undermine at the Wagner Act, protecting unions, which was part of the raft of legislation in 1935, immediately drew very hostile reactions from business. And the Social Security Act was also producing, lobbying to cut back its implementation, especially in the south. And so there was a real confrontation by 1935 between the Roosevelt administration and business and far from working with business to try and end the depression. I'll give you a snippet of FDR's famous Madison Square Garden speech in October of 1936. And I should say this is after the administration concreted type, contrary to the usual outcomes had picked up large numbers of seats in the midterm elections in 1934. And so they had big majorities in the House and the Senate. And this is FDR, campaigning, speaking in Madison Square Garden in October of 1936, and a runner to his reelection. - [FDR] We have to struggle with the old enemies of peace, business and financial monopoly, speculations, reckless banking, class antagonism, sectionalism, war profiteering. They had began to consider the government of the United States as a mere appendage to their own affairs. And we know now that government by organized money is just as dangerous as government by organized mob (crowd cheering) Never before in all our history, have these forces been so united against one candidate as they stand today. They are unanimous in their hate for me, and I welcome their hatred. (crowd cheering) - [FDR] I should like to have it said of my first administration, that in it, the causes of selfishness and of lust for power met their match. I should like to have it said, (crowd cheering) - [FDR] Wait a minute, (crowd cheering) - [FDR] I should like to have it said of my second administration that in it, these forces met their master. (crowd cheering) - So there you have it. FDR and he did indeed he coasted to reelection. Again, keeping huge, significant majorities on Capitol Hill. And the contrast with the Obama administration is pretty stark. Again, if you go back and read the popular books about the administration. Actually the bankers were stunned that more was not demanded of them. When they met with various officials from the administration, they were quite taken aback at how how little was in fact, being demanded of them. And particularly in these areas of executive compensation, which were obviously going to be politically toxic in the way they turned out to be. But before we leave the Great Depression and the Roosevelt administration, it is important to say that Contrary to popular belief, among many on the American left, actually, the New Deal did not end the depression. The New Deal policies were not particularly effective. And indeed if you look at another big recession in 1938 What really ended the Depression was World War Two. The war economy ended the depression is now today the pretty widespread consensus among historians. So even though it politically of course what FDR did was absolutely brilliant, it's not clear how effective the new deal was in dealing with the depression. That's not to say it wasn't desirable policy for longer term reasons to build a social safety net and a welfare state. But it was not an, the Roosevelt administration did not have any policies that worked particularly to end depression. And indeed, if you are a hard boiled Keynesian, he was quite haphazard. He had two brains trust one in Columbia that was much more what we would think of as proto-Keynesians and then hard money people in Harvard giving him conflicting advice. And he was something of a push-me-pull-you and he was in 1938, they actually they reduce spending. So it's not as if they had sort of the keys to the policy solutions. And from that point of view, if you could say that the Roosevelt response to the Depression was perhaps brilliant politics and important for the longer run in but was not a particularly effective solution to the depression, maybe nothing would have been but certainly what they did wasn't. So let's now come back to 2008 and the response. And there were several pieces to the response. The first is that the Fed under the direction of Geithner and others really did, they stopped the US and global liquidity meltdown. I underlined global on that slide, in deference to Adam Tooze's excellent book crashed published about a year and a half ago, in which he documents to a degree that had not been fully appreciated before how much the Fed not only bailed out the US economy, but he created global liquidity over a period of a number of years in Europe, in Asia, in particular, trillions and trillions of dollars of liquidity was created. Partly because the Europeans found themselves unable to do it. The, one important takeaway from the financial crisis is that in one respect a lot of mantras about globalization and transnational institutions and so off, turned out to be vastly exaggerated in that the international financial institutions did not have a very big role in fighting the financial crisis and in Europe, the European Central Bank couldn't coordinate on policies. This is basically to Tooze's story. And so it felt to the Fed that could act as a unitary actor to essentially create and maintain global liquidity until the crisis finally subsided. And so as a technical matter, preventing this great recession from turning into a great depression, a lot of the credit does indeed belong to the Fed and the truth is that the real financial crisis in the US system was three or four days when it really looked like the whole place, the whole thing might go to pieces then there were a lot of chronic problems and subsequent aftershocks and so on. But the fact is that the Fed took actions that nobody could have contemplated taking if they had been anybody to take them during the Great Depression. Then a second piece of the response was the piece of legislation called the economic emergency economic Stabilization Act better known as the TARP 700 billion. That was finally enacted, as I told you, at the beginning of October in 2008. And that was largely for bailing out insolvent institutions for which merges could not be arranged or sweetening the pot in order to facilitate mergers of financial institutions that otherwise would have gone bankrupt. Then that the next big piece of the response was the Obama stimulus, the American Recovery and Reinvestment Act, so called, which started out being priced at 787 billion, but ended up costing more like 830 billion, and that was stimulus money, Keynesian stimulus money to try and get the economy going. And then, of course, the last piece of the response was financial services reform. The Dodd Frank law, Wall Street Reform and Consumer Protection Act passed in July of 2010. We will not be able to dig into all of this in today's lecture and the central question I wanna talk about which fits into the arc of the course, is what do these responses to the crisis reveal? What could have been done differently, if anything, and what were the downstream effects of the way, in fact, they did respond. But first, I want to introduce some concepts from political science and political economy and use them as if you like lenses for thinking about regulation and regulatory reform. This is a course after all, in which we're trying to keep political science honest by bringing history to bear but also using the analytical insights of political science to think about historical developments. I wanted talk about the three main perspectives on regulation and then we'll take a closer look particularly at the Dodd Frank legislation from the point of view of these. So the first, which I'm calling technocratic regulation really goes back to Pigou's economics of welfare. He was one of the most famous welfare economists of the early 20th century. And this is when welfare economics was thought of as an explicitly normative enterprise. If you open the economics of welfare and start reading it, the first few paragraphs are all about how to organize the economy, for improving the national welfare. And a lot of Pigou's work was about regulation. But there's a certain innocence to the whole tone of that book, in that he assumes that there are right answers to regulatory conundrums whether or not to regulate it, how to regulate. And that they can be implemented by somebody called a legislator. And indeed, a lot of the early economics work. If you go back read papers from that era, when they talk about regulation they will in their models, they'll be a legislator who is going to do this or not do this, that and the other. And so, the idea is that there are certain kinds of situation where the economist technician can figure out that there's definitely a problem in the market system, and the regulation can fix it. And indeed, there are some examples of regulation, where it's pretty clear what the technocratic answer, in fact is. So one very interesting example in American history is that we used to have heavily regulated airlines. And the federal government justified this, on the grounds that it was interstate commerce, right? Planes fly across state lines. But there was a lot of suspicion that these regulations were mostly serving the interests of the airlines. And then a number of academics had the brilliant idea they found what we might call a natural experiment, which was that there are two states that are sufficiently large, Texas and California, that they have big, big airline routes that don't cross state lines, and so therefore, were exempt from the FAA's regulation. And they did a bunch of empirical studies which showed that the the airline's operating only within Texas and California, were way cheaper and way more efficient than the regulated ones. And so it was a kind of lay down hand demonstration that we shouldn't be regulated these industries at least in the way we were and whatever benefits they were serving, they were not serving the interests of the consumer. And these studies provided a considerable amount of the ballast in arguing for airline deregulation and turning airplanes into instead of play things for the wealthy sort of buses with wings, which is what we have today. Other areas where technocratic regulation gives you the right answer is where there are unambiguous market failures. For example, insurance markets, if you do not require people to have collision insurance, the problem is that the people who would go and want to go and buy collision insurance are the risky drivers with no assets. And they're the people that insurance companies would not want to insure. So if, you know if you're very safe driver, and you can self-insure you wouldn't want to do that. And so if it weren't mandatory for people to buy auto insurance, insurance markets would end up pricing it's so high that most people wouldn't buy it, then the markets would unravel. So this is the this is the idea what sometimes called adverse selection by economist you got adverse selection means that the people the companies don't want to insure are the ones who want the insurance and the people that they would wanna insure don't want it and they certainly not going to buy it at the price that the companies are going to charge. Similar arguments get made about health insurance that only sick people would go and buy a share of insurance voluntarily. And so that would price the market would price it out of the reach of most people. So these are the sorts of areas where you have unambiguous market failures because of adverse selection. And as a result, this case technocratic case for regulation. Another is deposit as insurance in banks. So if you think about how banks make their money, people in an era when banks paid interest, of course, people would deposit their money in the bank and get paid a certain interest rate. The bank would then lend out the money to people who were investing things and the bank would make its money on this spread, right? That's, I'm not talking about a badly managed bank, I'm totally completely well functioning bank. That's what banks do. That's what they're there for, right? But of course, it all assumes that not everybody is going to come along and take out their money at once. Because they don't have the money. They've lent the money, right. So, that is what but what a well functioning bank does. Now if you have a panic, and this is what happened during the Depression, if you have a panic, and people start to believe that the bank is gonna go bankrupt, then everybody's gonna go and take that money out of the bank because they don't want to lose their money, right. And so you're gonna get a run on the banks, right. And this is part of among the things that happened during the Depression. And indeed, part of that we think of Glass Steagall for other reasons, but the financial reforms of 1933 one of the things they did was create deposit as insurance, right? Because then if everybody's told no, if the bank goes bankrupt, you're gonna be covered up to it was until the financial crisis up to $100,000. Then people, they know the government's gonna protect them and they'll leave their money in the bank and you won't get the bank run. Again, so this is there's a clear technocratic case here for regulation, right. These are not hard calls, One might say that. Though I would add that notice sometimes the technocratic solution creates new problems. So if the government is insuring the depositors, then it means the banks can gamble with other people's money because the bank knows the government's gonna bail out their depositors. They're not gonna go bankrupt, so they can be more risk-embracing than they might otherwise be. And so you might think that they have to regulate them. The way in fact that was handled when they created the FDIC was they made the banks finance the bailout funds. So the banks had to essentially, pay a tax that funded the insurance of depositors so that the banks are internalizing the risk, if you like, of providing it. Though, as it turned out, during the financial crisis, it rapidly became obvious that 100,000 wasn't gonna be enough to stop runs on banks. And so the government in the middle of the crisis just increased it to 250,000. Which, among other things made it obvious that if push comes to shove, actually the taxpayers are ensuring the deposit so doesn't solve, if you like, the moral hazard that the banks have to be risky, if they now know that even if the deposit insurance is insufficient, Uncle Sam will come to the rescue. So when we think about even if it's a non ambiguous case for regulation, doesn't mean that there's a solution that's not gonna generate tricky problems of its own. There are also hard cases. So if you imagine yourself to be a senator or a congressman or congress woman sitting in a hearing and in walks Paul Volcker, former Fed chair, extremely distinguished economist and says, we've gotta separate commercial from investment banks. We cannot have investment banks, investing their own money while also being fiduciary for people who deposit funds this had been what had been done in Glass Steagall, and had had been undone in the run up to the financial crisis. And Volcker was firmly of the view that some version of this needed to be reinstituted so that the investment banks would not end up having the sort of incentives that for example, Goldman, turned out to have during the housing crisis, which I will talk about on Thursday, when they unloaded mortgages that they knew were going south on their own clients, conflict of interest between the fiduciary obligation to their clients and their attempt to protect their balance sheets as investors. So you have Volcker coming in saying this has gotta be stopped. And then you have Tim Geithner, you know, the genius from the administration who saved the world economy coming in and saying, this is a terrible idea. We should not do this, it interrupts the way in which markets work because investment banks often have to do market making, they have to facilitate transactions, which sometimes means buying and holding assets from somebody who wants to sell them until you can find a buyer. And he'll give a long speech about why in fact, we shouldn't have what subsidy came to be known as the Volcker Rule. And you had the senators sitting there or the congressman or the congresswoman. Do you think these politicians have the slightest idea who's right? (Audience laughing) They'd have absolutely no idea who's right, right? You know, and you could line up distinguished economists on both sides of this question, and the politicians are going to have to decide what to do. But in fact, these technocratic hard calls that often elude the people who are actually making the law. And then there's In addition, even if you have a good technocratic regulation, will the government actually have the capacity to implement it? Again, if you think when back to as I said in 2004, and 2005, when the big banks persuaded the federal government to ease up on their capital requirements, they said in return, we'll open our books to you. But in fact, Christopher Cox had I think it was four or five regulators or maybe six, whose job was to watch more than $4 trillion in assets, and they simply didn't have the capacity to do it. So, while it is true, that there is sometimes clear cases for regulation on technical grounds, we shouldn't think you can ring the politics out of it or that the process of implementing it doesn't create issues that subsequently will come home to roost in difficult ways. But I think if you wanted the poster child for a technocratic regulator, it would be Tim Geithner. And just to give a give a sense of this, here he is five years after the fact ruminating on the actions that he took. - I think a lot of people still feel that the banks should have been made to pay in some way, - Yeah - Either in terms of restructuring, less executive compensation. Some cases, some people shoulda gone to jail. - I think everybody feels that way. - Do you feel that way? - Well, I feel like it's a complicated thing. You know, there was a terrible amount of abuse, and fraud and bad behavior, and damaging and dumb things that people did in this crisis. And I think Americans definitely deserved a stronger enforcement response, and its completely... - I wanna know what you think, I mean... - and I think - You saw this up close better than anybody. Does it surprise you nobody went to jail? - It doesn't surprise me no one's going to jail. It doesn't surprise me. - [Narrator] What may have been immoral or unethical, he says, was not illegal. Geithner would become a lightning rod and TARP, the $700 billion rescue package he helped push through to save the banks would take much of the heat. Banks have paid back the loans with interest, but the cheap money they were lent allowed them to reap billions in profits as homeowners struggled. - These banks were still taking huge bonus. - That was terrible. That was terrible. It was outrageous and killed us. - [Narrator] But Geithner argues that even though it appeared the arsonists were being rewarded, the banking system had to be secured. - In a financial panic, there is enormous collateral damage. They're deeply unfair and tragic in that context. So the first imperative is to make sure you make the system stable not because you care about the banks, or you wanna protect them from their mistakes, but because if you let the lights go out, the system will collapse around them. - Do you worry about defending your legacy? - I feel, I feel very confident that we made some very good choices and the best choices we could in the circumstances of the time. - So there you have that it's not that he thinks these people were good or that they didn't do terrible things. But that's not his job. His job was to save the system, and this was the way to do it. So the short of it is that technocratic theories of regulations ignore the ways in which power, politics and ideology shape outcomes. So a second perspective on regulation goes back to Chicago economist by the name of George Stigler not to be confused with Stiglitz, who is a famous economist writing today and Stigler pays attention in his work and this is hugely influential to the distribution of power in the economy. And he says, well, what is a regulation? A regulation is actually a valuable thing. And so it's gonna have a price. He thinks like an economist. And what a regulation is, is basically the capacity to use the government's monopoly on power to get something and powerful groups are gonna use it if the government is available. So even though he's a Chicago economist, generally thought to be heavily pro market. Basically, what is his theory of like regulation is that big firms, producer interests are gonna dominate what regulations get made. Big firms that going to, to use terminology we've used earlier in the course, the big firms in an industry have a K group. They can coordinate in order to get the regulations that they want adopted by politicians. They can coordinate, they can lobby for them, they can spend money to get them. And that's what they will do. Whereas for small firms and consumers, it's very hard for them to organize, and it's gonna be unlikely that they're going to get their way. So, as a as a causal matter, we should expect the big firms to dominate the regulatory process. In Stigler's mind this is a reason for avoiding regulation altogether. Which is something one inference one could draw from it. But, there are others, which we'll get to in a minute. So if you want examples of this, it would be airline day regulation in the first place, right? It served the big airlines to get these regulations to protect their markets, and so they regulated to get them. Taxi medallions in New York City. Before the advent of Uber, a New York City taxi medallion was worth a million dollars. Today, it's not even worth $100,000, right. Again, until the changes in the market eroded, eroded the monopoly is just that the taxi medallions had created, they functioned as huge barriers to entry. And indeed in many European countries still big fights that they in Madrid for example, Uber is outlawed. Again you have a taxi monopoly protecting, protecting itself. Also in the I mentioned to you earlier that the erosion of Glass Steagall. So Glass Steagall was passed in 1933. If you read Ron Chernow is the house of Morgan, one of his many brilliant books. You among themes that Chernow recounts in that book is the relentless efforts to get rid of Glass Steagall almost from the moment that it was enacted. And the truth be told when it was finally repealed in 1999. At the end of the Clinton administration, there was virtually nothing left of it. So again, the big financial institutions had used their lobbying power to get regulations that serve their own interests. So very different than the technocratic theory of regulation here's, the argument is the big players shape the regulatory landscape, and if you wanna know how it's going to play out, look at what their interests are because they are the ones who are going to prevail. A third perspective on regulations, takes off from the observation, well politicians need money, but they also need votes. And so they're gonna have to trade off, trade off attending to lobbyists and attending to voters, the two most important pieces here by Samuel Peltzman, from the 1970s, and a more recent one prompted by Dodd Frank by a Columbia economist, by the name of John Coffee and so, so the idea here is that politicians can't completely ignore voters. But so when will they attend to voters because the thing about voters sentiment is that it's sometimes intense, but it's very intermittent. Voters get outraged by something and focus on it for a short time. But then things change, there's something else in the headlines, and it moves on. So Coffee call it called it a kind of sine curve of regulation. So when voter anger is that a high fever pitch, you might be able to, or concern is that a fever pitch, you might be able to enact a regulation that you will not be able to enact at a different time. And most of the time, consumer interests are latent, but they'll become manifest. But of course, if the producer groups, the big players, the Stigler people know that, their motto just needs to be all we have to do is wait, right? And so the examples that that immediately come to mind are gun control legislation. Every time there's a school shooting, this is whipped up, rage by voters, that something has to be done something has to be done and politicians all say something has to be done. But the NRA keeps its head down and a few weeks later, the news cycle has moved on, and you come down the same curve and you don't need to pay attention to voter interest. Support for disaster relief. It's the same kinda thing. In the midst of a disaster, there'll be huge pressure to support, spending money for disaster relief, but once the disaster or the headlines about the disaster are over the support dissipates. This is why Amartya Sen makes the point that you don't get famines in democracies. That's true. But you do get chronic poverty, right, because crises generate voter interest and attention, whereas chronic problems tend not to. So this is the voter orientated model that's gonna tell you politics and ideology gonna matter some of the time. And, but much of the time, you're gonna in fact, have the big producer groups get what they want, regardless of the politics and regardless of the technocratic merits. So if you think about Dodd Frank, it was a major overhaul of banking regulation in the US of capital requirements, how much capital banks have to hold, and of the government's resolution authority, what power the government has to unwind troubled banks. A few things to say about, first of all, it was enacted in the spring and summer of 2010, before the midterms, and that meant actually they enacted it even before the commission that they had appointed to analyze what had gone wrong in the financial crisis had reported. And the reason they did that was that, unlike 1934, the vast majority of the time, the party in the White House loses seats in the midterm elections. They didn't know they would actually lose control, but they knew they were gonna have they're gonna be less well placed to pass legislation after the 2010 midterms. So the whole thing had to be concentrated into a very short period of time. This was not a bill Republicans were going to support, and they knew it. The White House knew it. And so they essentially had to do what to write this bill in over the course of 2010, and get it enacted in the summer of that year, which is what they did. It was subject to intense lobbying before, during and after passage. But it illustrates all three of the models that I've been talking about. And you can see how the different ones give you an insight into different aspects. So one of the nice things about this book by Alan Blinder, who had been on the Federal Reserve as an economist at Princeton, is that he has a four-page table in this book After the Music Stopped, which summarizes the whole Bill and just, I'm not gonna go through all of it, but on the left column is the issue and then the key aspects of what the debate were, then the Treasury proposal, and then what was finally enacted. And the point is that that Treasury proposal very much reflected the interests of the of the industry because the industry has lobbied the Treasury very heavily. And the takeaway from the chart, I'll post all four pages of it for you to look at is that on virtually everything of importance, the Treasury got what it wanted, essentially what the Treasury had negotiated with the banks was what wound up in the law on the too big to fail, on systemic risk, on various other things. And as I said, I will post these so that people can look through them. But on virtually all particulars, the Treasury got what it wanted signaling the Stiglarian theory looked pretty good. When President Obama signed it in July, the industry was unhappy with some pieces of it. They had essentially lobbied to stop the bill for as long as they could, and then to weaken it once they saw a bill was definitely going to pass. And Scott Talbot, the chief lobbyist for the financial services Roundtable, the day the bill was signed, he said, this is halftime. And, so where did they fail? Where did the industry not get what it wanted? So one issue was the issue of proprietary trading, I've already talked about the so called Volcker Rule. And what was interesting about how they wound up with the Volcker rule was the following. They had lobbied to kill it and they had killed it. And you might remember that in 2008, Nine, Senator Kennedy from Massachusetts died. And this was at a time when there was increasing anger in the in the grassroots of the Democratic party with the degree to which the Obama administration was being solicitous of the banks and had let the AIG bailouts go, and so on. And so there was a special election in Massachusetts. The most liberal state in the Union often said to be. Was it Ronald Reagan, who called it the People's Republic of Massachusetts it was some conservative politician called it the People's Republic of Massachusetts. Anyway, shock. The Democrats was so dispirited, there was very low turnout and a Republican, a Republican, a truck driver, since come and gone, actually won the senate seat and Obama was, so the pop book say was sufficiently spooked by this, that they decided they had to do something for the base of the Democratic Party. And they put the Volcker Rule back in essentially outlawing proprietary trading. And so that's how it wound up in the bill. The other area where the industry didn't prevail was in the consumer protection area. There, the feeling was sufficiently strong in the White House that predatory lending had to be stopped. This was Elizabeth Warren who was by the way it was in the White House in 2009 and 10. And this was the central thing she had been pushing and was non negotiable, there was gonna have to be consumer financial protection. The industry hated it. They lobbied and lobbied and lobbied to stop it. And when they couldn't stop it, they essentially, they worked with the folks from the Treasury to create an administrative structure that was almost completely unworkable. Alan Blinder in his book, and he certainly no flaming radical describes it as a Rube Goldberg machine. And it was not given cabinet status. It was contained within the Treasury, it reported to Congress. The first few years of its operation produced huge turnover personnel because they were so demoralized. So there you see, you see the interplay between the Stigler and the Peltzman Coffee views. In fact, the lobbying the lobbying was to just incredible. And one of the interesting things is more than a billion was spent. But, one of the interesting things is, you can see on this chart here is that much of the lobbying will occurred after the bill passed. So the bill passes here, right, and you can see the financial lobbying of the fed, the Treasury and other regulatory organizations actually increases. Here's a New York Times breakdown of money spent lobbying after, after the passage of the legislation. You might think, Well, why? Isn't a game over once the legislation have been passed, and actually not, for example, on the Dodd Frank law. I happen to be friends with somebody who at that time was a partner at Goldman. And the day the bill passed, I said to him, so does this mean Goldman is out of the proprietary trading business for good? He said, it'll be five years before it's clear whether we managed to kill that or not. And so what you look at is, when Congress passes a law, that's just the beginning, then they have to, the regulators have to write rules. And this is where a lot of the lobbying actually occurs, partly because of the Coffee curve of regulation. Voters are on to the next thing now. Nobody's thinking about Dodd Frank. Nobody's thinking about the Volcker Rule. And so you have, they publish a potential regulation, you have a comment period, they got 8000. the vast majority, long detailed comments from the industry. Financial lobbying groups met with regulators 400 times, 78% almost 80% of all the meetings they had. They had to keep pushing back the rule because they couldn't get it written. And so the second time they had another comment period, they got 17,000 comments, again, the vast majority, the vast majority of them from the industry, proposing amendments that would essentially make the rule unadministrable, particularly would make it virtually impossible for a regulator to tell the difference between genuine market making where a bank buys and holds some assets that a client wants to sell in order to find another buyer, from when they buying them for their own purposes and on their own account. And it was finally, is partly repealed for smaller institutions in the Trump reforms of 2018. And just to come full circle on the Stiglarian theory. Once Trump came into office, the Republicans had been campaigning for years on repeal Dodd Frank, repeal Dodd Frank. This is terrible thing that Obama administration has put in, but you would start to see articles like this one in the Wall Street Journal. Guess who's defending Dodd Frank, they said. This is December 2016, a month after Trump's election. There, the Wall Street Journal Op Ed page says there goes another liberal fantasy and progressive law. Giant banks hate Dodd Frank and would like nothing more than to return it to a regulatory Wild West. So why is it been so hard since the Republican election sweep to find a Wall Street CEO Who favors repealing Dodd Frank. JP Morgan's Jamie Dimon, Lloyd Blankfein of Goldman have urged against repeal and other banks CEOs are suggesting policy small ball rather than wholesale reform. Blankfein who will have former colleagues as senior policymakers in the Trump administration, in particular is in a position to encourage significant change. But last year he, this is Blankfein, explain why that's not in Goldman's interest. After describing how regulatory costs have helped raise the barriers to entry in his business higher than at any time in modern history, Blankfein forecast more opportunities for global giants like Goldman, to gain market share as only a handful of players will likely to be effective, will be able to effectively to compete on a global basis. So this is the Stigler theory, up in lights that the Dodd Frank has now become a massive barrier to entry for smaller firms. Likewise, the much hated much despise much criticized Consumer Financial Protection Bureau last week the journal weighed on the presidents of four trade associations representing banks and credit unions also urged deregulatory restraint and pushed for very minor reforms. So when we think about the missed opportunities, what could have been done differently to some extent, and you get this from the popular books? The Obama administration came in there and was sort of drinking out of a fire hydrant, things were happening so fast, they really had very little choice. To some extent they were a victim of weak parties. This is something I'm going to talk much more about in the remaining lectures of the course but one difference if you think back to the 1930s, The parties were much stronger. And in 32 and 34, the Democrats were in a very different position to enact FDR's reforms than was the case in the response to the financial crisis where I told you, the Bush administration had to work with congressional Democrats to get TARP enacted. And the Obama administration knew that it had tiny window of opportunity. And it was it was not gonna be able to do very much and by the way, the Trump administration ran into the same thing when they sought to repeal much of Dodd Frank, they ran into massive resistance to repeal in the senate perhaps from senators who were being supported by some of the people that were being talked about in the Wall Street Journal op-ed piece I was talking about. But I wanna end by just talking about elite complacency. I showed you this picture at the beginning of today's lecture of the collapse of the stock market and the vanishing of 16.4 trillion of net worth. But what I didn't say is if you look, what was the case two years later, two years later and more than half of the wealth had already come back. And if you look at the financial sector, if you look at the stock markets, This is France, Germany, Japan, UK and the US in light blue, you can see that a decade after the collapse, the stock markets had all come back and indeed we I believe we had a record high on the Dow yesterday. So from the point of view of the of people in the financial services industry or upper middle class people, their 401k said come back. For them this crisis was long over for many such people. And I think what that ignored was the increasing divergence between that recovery for people in the financial services sector with significant retirement assets and other assets in stock markets, and the real economy that was playing itself out in Middle America and over much of Europe, where you can see that the jobs deficit was much deeper after this crisis and lasted a lot longer. You can see that long term unemployment that is people who are unemployed for more than 27 continuous weeks was at higher levels and today, to this day, remains at high levels than for more than half a century. And housing, more than 10 million houses would be foreclosed upon in the decade following the housing market, which is the subject I'm gonna talk about next. But this, this divergence between the fortunes of people living off the financial economy and the rest of America and by the way it's not only America. This is the Algo-America welcoming refugees into Germany also greatly underestimating the angst out there outside of the political and economic elites, and that angst is what would be coming back to bite in 2016. Okay, we will talk about the housing crisis on Thursday. (soft music) |
Power_and_Politics_in_Todays_World | Lecture_5_The_Resurgent_Right_in_the_West.txt | - Okay, so today, we're talking about the resurgent right in the West. Let's get fix our intuitions with a little walk down memory lane. - [Ronald Reagan] In Chicago, they found a woman who holds the record. She used 80 names, 30 addresses, 15 telephone numbers to collect food stamps, social security, veteran's benefits for four non-existent deceased veterans husbands as well as welfare. Her tax free cash income alone has been running $150,000 a year. - So that was Ronald Reagan running for the Republican nomination against Gerald Ford in 1976, which he was unsuccessful, and that's his famous welfare queens speech which would become if things had been able to go viral in those days, it did the paper equivalent of going viral and it represented this frontal assault on the welfare state, on what might be more broadly referred to as the postwar consensus that had prevailed in almost all of the capitalist democracies since the 1950s. In the US, it went by the name of the Great Society ushered in by Lyndon Johnson which was not only the civil rights legislation, but the creation of Medicare, considerable expansion of many components of the welfare state. In Britain, it was of course the postwar government of Clement Attlee voted into office in 1945 that put into place the major elements of the British welfare state. And in many of the European social democracies as they came to be called, we got countries that were marked by strong welfare and other forms of social protections such as health insurance, unemployment insurance, pensions, progressive tax systems, a lot of redistribution through the tax system, much more steeply progressive tax systems than in any of those countries today. And it's important to say that the support for this postwar consensus was substantially bipartisan. So for example, in Britain when the Conservatives under Churchill came back into power in 1951, many people said, "Oh, they're gonna roll back "everything that labor has done." Not a bit of it actually. They even expanded it. And through much of the 1950s, '60s, and '70s, you saw in most countries the recognition of a kind of what seemed to be a substantially bipartisan support for this welfare state. In the US, in the Nixon Administration, of course, there was some pressure for lower taxes, but still, much of the welfare state was preserved and in some ways it was even expanded. They got things like, which I'll be talking about later in the course, such as universal basic income. Nixon has supported what he called a negative income tax, a kind of refundable tax credit. We've ended up getting something that's a cousin of that. In Europe in countries like Germany where they have coalition governments, in the '60s and '70s, they were basically dominated by the social democrats. First, Willy Brandt, and then Helmut Schmidt. After in the early '80s, for reasons we'll be talking about next time, the social democrats got weaker and Helmut Kohl's Conservatives took over. But it was still, there was very substantial support for European social democracy across the political spectrum. For example, in the '60s and '70s in France and Italy where Communist parties had been very strong and had traditionally been more Stalinist like Communist parties, we had the advent of Euro-Communism. What was Euro-Communism? Well, one feature of it was that it was denounced by the Soviet Union as revisionist deviationism. And the other was that the standing joke in Italy in the 1970s was how can you tell that which is the Communist politician in parliament? How do you tell the Communist from the other politicians? And the answer was because they have more expensive suits. So the Euro-Communists again stop pushing this idea that social democracy was a way station on the way to Soviet style Communism and instead was a particular form of political and social organization that had come to stay. And it was seen to be as I said, a very broad bipartisan, or in Europe, multipartisan consensus. And this is what starts to come under attack in the late 1970s and early 1980s that Reagan is taking dead aim at the heart of that consensus and the new attack is marked by hostility to trade unions and the welfare state. It's hostility to progressive taxation and hostility to regulation, the sort of forerunner of the neoliberalism that would be coming later. And as in Reagan's famous slogan, government is not the solution to our problems. Government is the problem. We have to weaken the power of government in the economy and let the market run. This is the basic idea. Now, the first thing to say about this is when people like Reagan and Thatcher came into power in the, Reagan got selected in 1980, in the 1980 election. Margaret Thatcher comes to power in 1979. She had beaten Edward Heath for the Conservative Party leadership in 1975 in the first competitive election for a leader in the Conservative Party in its history and people were quite stunned by this and she'd been thought to be a minor, second tier cabinet official without much future in the party. And indeed, I think it was Geoffrey Howe who said she'll be gone in six months. Turned out she was leader of the Conservative Party for 15 years and Prime Minister for 11 and a half years. But I think it's fair to say that most people did not take Reagan and Thatcher seriously. They saw them as real threats, but most people in the establishment, the political establishment, the economic establish, didn't take them seriously as long term threats. Those people tended to see this as a kind of aberrational outcome and that things would revert to normal. Social democracy was about alternation in power between center right and center left parties, and this was a center right moment, but things would revert. There were a few discordant voices. In the US perhaps, the most prescient one was a man by the name of Kevin Philips who wrote a book, published a book in 1969 called "The Emerging Conservative Majority" predicting the rise of the new right in pretty much the form that it took. But he wasn't, he was later on given great credit for having predicted this, but for the most part, for the most part, people did not see this as any kind of sea change or fundamental realignment of politics. And just to give you a sense in Britain. In 1980, the principle intellectual architect of Thatcher's revolution was Sir Keith Joseph and he was her trade and industry minister and he was the source of all her economic ideas about monetarism and cutting spending. And here is Michael Foot who is by then the leader of the Labor Party in opposition commenting on Keith Joseph in the House of Commons just to give you a sense of how seriously they really took the threat from Thatcher. - [Michael] I wouldn't like to miss out the right honorable gentleman, the Secretary of State for Industry who's had such a tremendous effect upon the government and our politics altogether. And as I see the right honorable late gentleman walking around the country looking puzzled, and forlorn, and wondering what's happened, (audience laughing) I've often tried to remember what he reminds me of and the other day, I hit upon it because I recall that in my youth when I used to go in Plymouth, yes, I know, quite a time ago. (audience laughing) You'll come to that one in a moment too. But I used to go when I lived in Plymouth every Saturday night along to the Palace Theater and the favorite I always used to watch there was a magician conjurer and they used to have in the audience dressed up as one of the most prominent alderman in the place a person who was sitting at the back of the audience. And the magician conjurer would come forward and say that he wanted to have from the audience a beautiful watch. And he would go amongst the audience. He'd go up to the alderman and eventually take off him a marvelous gold watch. And he'd bring it right back onto the stage. He'd enfold it in a beautiful red handkerchief. He placed it on the table there in front of us. He took up his mallet and he hit it! Smashed to smithereens! And then on his countenance would come exactly the puzzled look of the right honorable-- (audience laughing) And he would step forward, and he would step forward. He would step forward right 'til he, he'd step forward right to the front of the stage and he'd say, "I'm very sorry. "I'm very sorry. "I've forgotten the rest of the trick." (audience laughing) That's the situation of the government. That's the situation of the government. They've forgotten the rest of the trick. It doesn't work and unless any objector might come along and suggest that it was only a trick, I can assure that at the Palace Theater in Plymouth, the way it was played, he used to come along at the end of the very act and say, "I'm very sorry. "I've still forgotten the trick." That's the situation in this country. - So there was Michael Foot making fun of Keith Joseph as a kind of buffoon who didn't know what he was doing and getting lots of positive reinforcement for that, but of course, the joke really was on Michael Foot. Three years later, he ran unsuccessfully against Margaret Thatcher and his election manifesto was called the longest suicide note in history subsequently by historians. So the joke was really on Foot and he completely underestimated the staying power of what was coming down the pike. So that's our agenda today and we're gonna look at the sources and implications of the resurgence of right wing politics across the Western democracies. So we're gonna look at several sources of this, one being the collapse of Communism, the second being embedded in the logic of distributive politics, and the third being the changing power relations in the economy. And all three of these things have an influence or should have an influence on how we think about the resurgence of the right and its political implications. And when we think about its political implications, I'm gonna talk briefly today but more next time about those implications in two-party systems like Britain and the US versus in multiparty systems where like we have on much of the continent. So, let's start with the effects of the collapse of Communism. When the wall came down and when the Soviet Union started disintegrating, many, many intellectuals and analytical commentators thought that this would be good for European parties on the left. They thought generally that it would be good for two reasons. One, it made it clear that European social democracy or what we think of as a great society kind of political economy in the US is not a way station on the future to some either more statused, if you like, outcome. That once we'd think of it rather as a kind of permanent equilibrium that can be defended on its own terms and is not something that's gonna produce something worse. Another way of putting perhaps the same point is it gets rid of the kind of bogeyman to mobilize against from the right. So you know that if you propose slight increases in progressive taxation or you improve healthcare, the notion of that you're trying to do is create Communism becomes less credible if there's no Communism in the world left to create. And many people thought that with the Communist threat gone, the end of the Cold War, we wouldn't need to mobilize against the Soviet Union anymore, that all of the money that had been spent on the arms race could be plowed back into the Western economies and we were in for a continuation, if you like, of the benign social democracy and maybe the power would shift once more back to the left. Maybe the right wing governments in say in Britain and the US, or the fact that power in Germany had now shifted as I already said from the social democrats to the CDU, maybe it would shift back. And what all of that failed to take account of was that, and this is really the story of Tony Judt's book "Postwar," some of which I've put on the syllabus, is that the social democratic economies and welfare states depended on a very fortuitous arrangement of demographic, economic, and geopolitical circumstances after the war. The demographic one was that you had big working age populations relative to what we call dependent populations. We think of the dependency ratio is the ratio of working age to dependent. And dependent is defined under the age of 15 and above the age of 65. And of course, what really matters from the point of view of the fiscal health of the welfare state is the older crowd because parents mostly internalize the cost of children whereas governments in turn have to bear the cost of elderly people. And for the most part in all of these countries, we had big working populations and relatively small elderly populations. That would change as the baby boomer bulge worked its way through the population. And so there's one big source of fiscal stress inexorably coming at all of these systems as the ratio of working people to retired people would increase. Then secondly, much of the economic health of these countries had been result of the enormous infusion of funds by the Americans, the Marshall Plan in Europe, the rebuilding of the Japanese economy, and the fact that as I mentioned to you earlier, Europe was not spending money on its own defense so it had the resources to build up its domestic economies. And that too would go away or would start to diminish as the Cold War ended. Also, the American interest in sustaining these economies as bulwarks against the Soviet system would start to diminish as the Cold War ended. So the idea that European social democracy was sort of a permanent equilibrium was by no means obvious. It's certainly not in retrospect. A second thing to notice is that while the bogeymen of Communism might not have been there around which the Conservative, the ideological right could organize, there are always other potential bogeymen and particularly as it's turned out since 9/11, Islamic fundamentalism is an alternative bogeyman around which to organize political movements and it has the added advantage that it's not a threat to Democratic capitalism. You know, when Islamist regimes, when Islamists do come to power whether it's in Iran or Libya under Gaddafi, they run capitalist systems badly. There's no alternative. There's no Islamist model of the economy. So the idea that one has to worry about them as a threat to capitalism in a way that Communism had once been perceived is no longer there. And that's important because it diminishes the incentive to buy off working class discontent. So if you think about, go back to the new deal, after the Depression, Communist Party in the US was quite powerful. It's very influential in American unions, particularly in the CIO. We had 20% unemployment. There was real fear in the aftermath of the Depression that Communism might claim hearts and minds of large numbers of American workers and that is partly what motivated business elites and others to at least tacitly support the new deal. The notion was that you want to make sure America's workers do not reach a state of affairs where they think they have nothing to lose but their chains. And so my colleague Peter Swenson has done tremendous scholarship on this, but actually what Swenson shows is that there's never been significant expansion of a welfare state, even the European welfare states by the way, without significant at least tacit and sometimes active support from business. And part of it in the '30s and '40s was Keynesian thinking about revising the economy or proto-Keynesian thinking. The general theory didn't come out until well after the Depression, but there were many people making those kinds of arguments at the time. But it was also political. It was this idea that you want to give the dispossessed people who had been most harmed by the Depression a stake in the existing order. And similar arguments were made in Britain and Europe. For instance, Ralph Miliband, the father of Ed and David Miliband was a political economist who published a book in the early '60s I believe it was called "The Capitalist Industrial State" in which he argued the welfare state is capitalism's best friend. He was a Marxist, well to the left even of Ed Miliband, and he thought that this was a bad thing because it was preventing the workers from becoming revolutionary because they had this stake in the existing order. And so this was the notion that if you don't do this, there's an alternative model out there vying for hearts and minds and we'd better look out for that. Once the alternative model has gone away, once there's no longer a Soviet Union, it's much easier to discount all of that. It's much easier to say well, let's, you know, you might not say it, but think move to the suburbs and build more prisons. It's not obvious that there's any reason of the sort that existed when there was a Communist alternative out there to continue to maintain expensive welfare states that would require substantial levels of taxation and maybe increasing levels of taxation given the changes in the demography. So the short of it is that those who had faith in the idea that the end of the Cold War was gonna be good for left of center parties were wrong and things have played out quite differently since that time. So that's our first thing. But I want now to turn to some more analytical thinking. Let's put on my analytical hat here. Some more analytical thinking about the logic of distributive politics because I think people don't think clearly enough about just what that logic is when they think about what's gonna happen in different political scenarios. So this is a famous picture of a hat from the beginning of "The Little Prince," Antoine de Saint-Exupery. And as you might recall, it's the very beginning of the book. He decides he wants to frighten the adults and he's gonna frighten the adults by drawing a picture of a boa constrictor eating an elephant. So he does. And he's absolutely devastated when the adults say, "That's a nice hat you've drawn." And he's furious. So he has to draw another picture showing the elephant, a sort of X-ray. But he's so despondent that he decides to abandon his career as an artist and instead become a pilot which he duly does. Now, one thing we can know about this story is that the adults he was trying to frighten were not economists because had they been economists, he might have been able to frighten them with this picture because instead of not seeing, instead of seeing it as a hat, they might have seen it as a picture of the income distribution. And in particular, this feature of the income distribution which is that the median income is always below the mean income. Why would that be? Why is the median income always below the mean? Yup. - [Male Speaker] (mumbles) there are some extremely rich people. - Very rich people, right? The very rich people out here. In fact, you know, Gates, and Soros, and these, they're probably on the other side of (mumbles) Street down there actually, right? And so that pulls up the average, right? So that means that in every, you know, obviously there's a cut off point here because there's some level below which people won't survive, but the very rich people pull up the average so there are always more people below the average income. And so this had for a long time led political economists to think that politicians who are trying to win election will be responsive the median voter and there'd sorta be a kind of impetus to redistribute downwardly because politicians need votes, they need to get elected, and so they will do that. This median voter is gonna be below the mean income. And indeed, this is a very old argument. So for example in the 19th century, people like John Stuart Mill were afraid of expanding the franchise for exactly this reason. And they came up with things like giving two votes to university graduates as a way to sort of blunt it. And Karl Marx after he'd been disappointed in 1830 when Communism had not arrived and he'd been disappointed in 1848 when the revolutions had not become Communist and he'd been disappointed in the Paris commune in 1871, he and Engel started talking about the parliamentary road to Socialism. And this was exactly the same idea that maybe the workers would do through the ballot box, what they had not done at the barricades. And the 20th century's ideas the median voter theorem which basically says that politicians in search of the median voter are gonna support downward redistribution. A very famous paper by two people called Meltzer and Richard which I think is somewhere in the recommended readings on the syllabus. If not, it should be. And so the biggest puzzle of 20th century political economy is that this does not in fact happen. That there's lots of redistribution in Democratic capitalist systems, but it's all kinds of distribution. It can be from urban to rural sectors of the economy. It can be from one industry to another. It can be done through food subsidies. It can be done in all kinds of ways, but there's sometimes, there's regressive redistribution in capitalist systems. We'll talk later about the repeal of the estate tax which is an example of using the political system to actually redistribute in ways that produce more inequality. And so that being the case that the sort of standard expectations, the folk expectations that once you've got majority rule with the universal franchise, the poor are gonna soak the rich. Why wouldn't they? They've got less. They're gonna soak the rich, but it doesn't happen. And so people have tried to understand why that doesn't happen and there are basically two approaches to thinking about that. So, one would be to say well, I've got one dimension here. The left wing party wants redistribution. The right wing party doesn't want redistribution and the median voter's gonna want redistribution and so the right wing party is gonna sorta be dragged to the left in search of the median voter or it's gonna lose, right? But what if there's another dimension as well? What if there's a dimension like race? So what if there's a dimension where people have racially exclusionary attitudes versus inclusionary attitudes? And then it might be the case that well, then it depends how much they compare about race versus how much they care about income or wealth, okay? And so this is the story of Richard Nixon's southern strategy, right? So what happens. 1964, you get the Civil Rights Act. 1965, we get the Voting Rights Act. Lyndon Johnson is famously reputed to have said, "Democrats have lost the South for a generation. "I've lost the South for the Democrats for a generation." Historians now debate whether he actually said it, but it was widely attributed to him. And the notion was that the Democrats historically had been the dominant party in the South and now that they had become associated with the Civil Rights Act, the Voting Rights Act, there was gonna be backlash and Nixon and his advisors thought he could capitalize on that backlash by offering to essentially limit the effects of the Voting Rights and Civil Rights Act in the South, stop that, and that would make, this would be the way in which the Republicans could attract voters who had previously supported the Democrats for economic reasons. And so once you put a second dimension into play, if you like, it's no longer the case that the median voter will necessarily want downward redistribution because they may want preservation of their racial status quo more. And so this is one of the ways in which scholars. My colleague in the Political Science Department, John Roemer has developed a number of these models that show this. Once you put a second dimension whether it, I'm just putting race here. It could be xenophobia. It could be women's rights. There might be another dimension. Again, you think about women's rights. Women may think the pro-choice pro-life dimension more important than the economic dimension and so vote differently. So once you put other dimensions into play, maybe the median voter won't vote for redistribution. So that's one possible story about why we have not seen the sorts of developments that one might anticipate. Here's another story. So this is what game theorists refer to as the majority rule divide a dollar game. And here, the intuition is that if you divide a dollar by majority rule, no matter how you divide it, there is some potential majority to upset that division. So if you think if they start out with an equal division, you can see here that Stannis and Arya can decide to completely dispossess Joffrey and keep it all to themselves. But of course, if we think of things going around over time, then what if Joffrey's gonna come along to Arya and say, "I'll offer you a different deal. "You can keep 2/3 of it or 60% of it "and I'll take 40% of it, all right?" And if that were the case, if that became the new status quo, Stannis would come up and propose something else. And this could go on, and on, and on. And what the finding is is that it never ends. So that if you have majority rule dividing up any divisible good over time, there is no distribution that in principle is not susceptible to being upset by some new majority. So notice a couple of things about the majority rule divide a dollar game that are very important for us going forward. The first is you don't need a second dimension here to get instability, right? Maybe there's a second dimension, maybe there isn't, but it's not necessary to have a second dimension to get instability. You can get instability with one dimension, right? Just money. This whole game is just about money. And that carries a further implication which is that interests alone will not produce effective demand for dominant redistribution or for retaining any particular distribution. Now, we're gonna be bringing into play the framework of interests, ideals, and institutions that I mentioned to you last week. And notice here I'm not just talking about interests as selfish. For example, the three people playing the divided dollar game. Maybe one wants to give the money to save the whales. One wants to give the money to soup kitchens and one wants to spend it protecting wolves as my friend Nicoly up there would probably want to do. So they try to maximize their share of the pie, but we're not saying they're selfish people, right? But all we're saying is that if they want to maximize their share of the pie, the interest in maximizing is not gonna be enough to sustain any particular distribution because if it's just that, there will always be someone who will come along and offer you a better deal, right? So interests alone are not gonna be sufficient to get people to support any particular distribution including, of course, distribution from the average to the median voter. So if interests alone are not enough, let's think about the other two concepts. Let's think about ideals. So what we're really saying is in order for the median voter to achieve downward redistribution, all these people have to vote together. All of these people to the left of this line have to vote together, right? Because if the wealthy can buy off the poor with bread and circuses and make the middle class pay for it, you can get a coalition of the rich and the poor against the middle class or you can (mumbles) 2017 tax cuts. You give a small cut to the people at the bottom, a huge cut to the people at the top. People in the middle internalize the cost of that. Again, it's a majority rule, it's a majority coalition. It's not gonna be downward redistributive. So all of these people would have to vote together. So they'd have to some ideals we can think about first to get them to vote together. So what ideals might there be? So one ideal obviously is what people think is fair. People think, you know, what is a fair distribution? Maybe you can motivate them with that. So here's the next question is well, what is fairness? Well, here's one answer to that. - So a final experiment that I want to mention to you is our fairness study. So what we did is we put two Capuchin monkeys side by side. Again, these animals, they live in a group. They know each other. We take them out of the group, put them in a test chamber. And there's a very simple task that they need to do and if you give both of them cucumber for the task, the two monkeys side by side, they're perfectly willing to do this 25 times in a row. So cucumber even though it's really only water in my opinion, but cucumber is perfectly fine for them. Now, if you give the partner grapes. The food preferences of my Capuchin monkeys correspond exactly with the prices in the supermarket. And so if you give them grapes as a far better food, then you create inequity between them. So that's the experiment we did. Recently, we videotaped it with new monkeys who had never done the task thinking that maybe they would have a stronger reaction and that turned out to be right. The one on the left is the monkey who gets cucumber. The one on the right is the one who gets grapes. The one who gets cucumber, note that the first piece of cucumber is perfectly fine, the first piece she eats. Then she sees the other one getting grape and you will see what happens. So she gives a rock to us. That's the task and we give her a piece of cucumber and she eats it. The other one needs to give a rock to us and that's what she does. And she gets a grape and she eats it. The other one sees that. She gives a rock to us now, gets again cucumber. (audience laughing) She tests the rock now against the wall. She needs to give it to us and she gets cucumber again. (audience laughing) So this is basically the Wall Street protest that you see here. (audience laughing) - So what he finds there with the Capuchins is a very, he's studied it in many, many, many social animals and there's plenty of psychology to support the idea that human beings are the same. However, he fundamentally misinterprets the result of his study. Anyone think why do we say he misinterprets the implication, he doesn't see what this implies? Yeah. No? It's not the Wall Street protestor because what infuriates this monkey is that another monkey just like it is getting more. The monkey's not bothered that the researcher has a bowl full of cucumbers and grapes, right? This is a very robust finding in the social psychology literature about people. Their sense of the word welfare is often other regarding, but they make relatively local comparisons. They do not compare themselves to people very different from themselves in the socioeconomic distribution. So where steel workers might compare themselves to auto workers or coal miners might compare themselves to people working in oil fields. And this is true, by the way, up and down the occupational scale. So a professor will be much angrier to discover that he or she earns $5,000 less than the professor in the next office than to discover that he or she earns $500,000 less than the attorney who lives next door. People tend to make relatively local comparisons and later on when we come to populism and all that, we'll discuss the implications of that in the politics that we're seeing right now. And so the point to make here is that people don't make global comparisons. They tend to compare people to people like themselves. Now, you could say. Yeah? - Is it only by occupation? - No, not necessarily. I mean, there's studies of children for example. You give a little boy a half a glass of orange juice and then you give his sister a full glass and he was happy before. Now, he's furious, right? So it depends what your reference group is. Paul Bloom who gave the Devane Lectures about four or five years ago is the local expert on reference groups, but it doesn't have to be in the same industry. It's got to be the kind of people you consider yourself to be like and by whom you want to be given esteem, right? So now, you might say, well, that's not necessarily fatal for solidarity because if you say well, I want to have at least as much as people like me. You could say well, let's get together and make sure we all get it, right? So it's not obviously fatal to solidarity. And that's basically what Marx's theory of consciousness raising was about. But the trouble is that it competes with other ideals of fairness. So let's just consider a couple that are relevant to where we're headed. - Next Tuesday is election day. Next Tuesday, all of you will go to the polls and stand there in the polling place and make a decision. I think when you make that decision, it might be well if you would ask yourself are you better off than you were four years ago? Is it easier for you to go and buy things in the stores than it was four years ago? Is there more or less unemployment in the country than there was four years ago? And if you answer all of those questions yes, why then I think your choice is very obvious as to who you will vote for. - So there, he's saying something quite different. He's saying ask yourself. This is Ronald Reagan in 1984 running for re-election. He's saying four years ago when I came into office, some here can remember what it was like in 1980. Interest rates were at 19%. Inflation was 12 or 13%. The economy was in pretty bad shape and by 1984, things were better. So he's saying you look at your circumstances now compared to four years ago and are you better off or not? This is what we might call a self referential comparison. It has nothing to do with what other people are getting, right? And it's not important from your point of view Reagan is anticipating what other people are getting. And this is what economists call a Pareto improvement, right? Economists are all about absolute gains, not about relative gains. People want to get on as high an indifference curve as they can in standard economic models without reference to the indifference curves that others might be on. Just to take the same logic across the channel, here's another formulation of it. - There is no doubt that the Prime Minister has in many ways achieved substantial success for the economy. (people exclaiming) There is one statistic that I understand is not however challengeable and that is that over her 11 years, the gap between the richest 10% and the poorest 10% in this country has widened substantially. - [Male Speaker] That's right. - How can she say at the end of her chapter of British politics that she can justify many people in the constituency such as mine being relatively much poorer off, much less well housed, and much less well provided than it was in 1979? Surely, she accepts that is not a record that she or any Prime Minister can be proud of. - Mr. Speaker, all levels of income are better off than they were in 1979, but what the honorable member is saying is that he would rather the poor were poorer provided the rich were less rich. Once they start to talk about the gap, they'd rather the gap were that, down here, that. Not that, but that. - [Male Speaker] That's it. - So long as the gap is smaller, so long as the gap is smaller-- - Order! - They'd rather have the poor poorer. - So you know, there it is. Two fundamentally antithetical conceptions of fairness. So how should we think about the contest between them? Which is more powerful politically, do you think? Or maybe that's the wrong way to put it. What are the conditions under which one might become more powerful than the other? Any reflections on this? Yeah. We're gonna give you the microphone, yeah. Do you want to give it to him? Yeah. - [Male Speaker] Can everyone hear me? Yes. - Yeah. - [Male Speaker] It brings back to the mean and average income we saw previously. If you have more people that have less, then politically you're gonna have more people who are going to vote for the income. - So just run with that a little bit. So if you were designing a party platform trying to appeal to more solidarity based idea of getting more people to resist the Reagan and Thatcher logic here, what would your argument be? - [Male Speaker] To resist Thatcher and Reagan? My main argument would be we're all born equal. We need to earn the same thing. There's no reason why other people gain more. - Okay and now, if you had to, if you had to bet in an upcoming election that you were right, how much of your own money would you bet? - [Male Speaker] This is taking the country to my opinions, but I think there's a good chance it will be successful in the next 15 months unfortunately. - Okay. You think the people's basic conception of a solidarity based conception of fairness can actually command people's support despite the appeals to what we might call an economist conception of absolute improvements? Any other arguments that somebody could imagine making in support of that notion? Yes, sir. You need the microphone, yeah. - [Male Speaker] Since the Reagan administration, there has been no real growth in wages of the middle class. - Mm-hmm. - [Male Speaker] And all the additional productivity has gone to the upper class. So I think that is something that you could use as a political argument that would work. - So how would you make the argument? - That life is not better than it was 50 years ago for most of the people. That if anything, your salary is the same and your, and the likelihood of you keeping your salary is less. You're less secure and you're at the same level that you were 30 years ago. - Okay, so I see you oscillating now between two logics here. One is you're saying the world's become more unequal and the other is that you're saying people are becoming more insecure, right? So those two things are not the same and they gotta be importantly different. And what I'm gonna be arguing later in the course is that when people become insecure, there's a very different politics. Insecurity is not the same thing as inequality. People can be afraid of insecurity even if they're relatively well off if they feel threatened. Fear of insecurity produces a very different politics than outrage at inequality. And what's going on here is that if you look at the Reagan and Thatcher arguments, in addition to being completely self referential. They're just saying, they're both saying why should the poor care what the rich have if the poor are better off than they were before or at least as good as? So they're basically appealing to two things. One is absolute improvement's more important than relative changes. And the other is that people, again, they then are appealing for prospects for upward mobility. They're saying you, you know, Reagan famously said I want an America where anybody can become a millionaire, right? People should be able to get ahead. Now, and that was certainly a difference between the politics of the new right in the 1980s and the politics of conservatism pro-market arguments today which revolve more around what psychologists and economists call loss aversion. It's no longer a politics about promising upward mobility. It's actually a politics revolving around as you said wages have stagnated and perhaps in real terms, actually gone down. People are more insecure. Make America great again is a slogan which implies that something's been taken away from you and I'm gonna bring it back, right? It's a very different notion than America's best days lie ahead. Now, it's interesting that Reagan was responsible for both those slogans. Make America great again and America's best days lie ahead. And in 2016, Trump used the first and Hillary used the second. But of course, the context in which Hillary was talking about America's best days lie ahead had been an era when we had had 304 decades of wage stagnation and growing economic insecurity. So it was a very different era than when Reagan and Thatcher were running in the 1980s. But those are more contextual things. I want to just make some more basic points about the logic of distribution here which is what Thatcher's success was keying off of was this wonderful line from the Times obituary of Benjamin Disraeli. He died in 1891. I don't know why it took them two years to write the obituary, but they said, "He discerned the conservative nature of the working man "in the same way as a sculptor "discerns an angel in a block of marble." So what Disraeli, he basically understood this logic of what we might call divide a guinea rather than divide a dollar, but that there was plenty of reasons to think that you could appeal to working class voters to support conservative policies. And one of the, I think, enduring myths on the intellectual left is that when poor people don't support redistributive politics, they're somehow acting against their own interest. Sort of what's the matter with Kansas, right? Thomas Frank. Why are these people voting in ways that are not in their interest? Even setting aside whether there's a second dimension like race which might be part of the story but it might not be part of the story, being promised minor incremental improvements might be sufficient to get people to support a political program without reference to what others are doing. And I think the other thing to say. You could say okay, these are two competing conceptions of fairness. It's not obvious why one will win rather than the other, but the difficulty is that it's much harder to hold a solidaristic coalition together precisely because all you have to do is start picking off some people from that coalition until you've got enough. So even if 40, or 50, or 60% remain in solidarity with this basic conception of fairness, if you could promise upward mobility to others, that might be enough. Or if you could promise at least repairing the losses that they've had, maybe subsidies for the industries in which they work, that might be enough. And again, we'll talk much more about populism since 2016 later, but notice that Donald Trump in no rally, in no campaign speech, in no piece of campaign literature ever said he was gonna reduce the inequality in America. He said, "I'm gonna bring industrial jobs back." He said, "These people are corrupt "and they've been screwing things up," but he did not promise to reduce inequality. He promised to give people absolute improvements or at least prevent them from experiencing further absolute losses period. So it makes it harder, we might say, all else equal, it's harder to sustain solidaristic conceptions of fairness because that's a coalition that can be picked apart in different ways and all it takes is getting enough people to support what you want to do and it starts to disintegrate. And that's what Thatcher did basically. She appealed to lower middle class workers, shop keepers, and working class voters with upwardly mobile aspirations and comparable things happened elsewhere in the 1980s. So the upshot is that even without a second dimension like race, solidarity is always vulnerable to splintering by these ideology, by these other ideologies or people promising to pick off somebody by promising them absolute gains or something better as part of some different coalition. So it sounds like ideals of fairness are not enough. They might help build a coalition and they might motivate people some of the time, but they're probably not gonna be enough. So what about institutions? What institutions could there be that would sustain a solidaristic, what I'm calling here a solidaristic conception that would get people to support single political platforms that might at least prevent regressive change or promote progressive change? Unions, yeah. They seem like the obvious candidate 'cause unions, unions are basically organizations that will pull together the people to the left of the median voter. So let's take a look at what we think about unions. - [Male Speaker] Sam Donaldson, ABC News-- - [Male Speaker] A strike by air traffic controllers. From the President, a strong warning. They are in violation of the law and if they do not report for work within 48 hours, they have forfeited their jobs and will be terminated. - [Male Speaker] From the union, a defiant response. - The controllers out there firmly believe in what they are doing and we will continue the strike. (orchestral music) - [Male Speaker] From ABC, this is World News Tonight with Peter Jennings at the foreign desk in London. And from our desk in Washington, Frank Reynolds. - Good evening. The strike began as scheduled at seven o'clock local time this morning. Its impact was immediate. Most flights delayed, many more canceled, but there was no delay in the reaction of the federal government led by the president himself. We have a series of reports tonight. First, here is White House correspondent Sam Donaldson. - [Sam] Four hours after the strike began, President Reagan described the aides as very angry and very determined to serve notice on the air traffic controllers that they had better return to work quickly or else. - Government cannot close down the assembly line. It has to provide without interruption the protective services which are government's reason for being. It was in recognition of this that the Congress passed a law forbidding strikes by government employees against the public safety. It is for this reason that I must tell those who fail to report for duty this morning, they are in violation of the law and if they do not report for work within 48 hours, they have forfeited their jobs and will be terminated. - [Sam] Strong as the threat of firing may be, the administration moved rapidly to employ other weapons as well. Court proceedings to force the strikers back to work were begun in 19 separate courts. Criminal charges were being prepared against individuals. The first cases would be brought against union leaders, but FBI agents are compiling a complete list of strikers. The union's three and a half million dollar strike fund has been impounded by court order and a petition has been filed asking for decertification of the union meaning to put it out of business permanently. The President was asked why such strong action immediately? Why not some lesser action to begin with? - What lesser action can there be? The law is very explicit. They are violating the law. - The President is said by an aide to be tough as nails in his determination to break this strike. And this morning, Mr. Reagan reminisced in the Oval Office about Calvin Coolidge's similar determination and success in breaking the Boston Police strike of 1919 when Coolidge was governor of Massachusetts. In this case, Mr. Reagan wanted to make the deadline for firing 24 hours at first. He finally settled on 48 in order to make certain that no air traffic controller could say later, he didn't get the word. The President seems convinced that by being tough, he can force many of the union members back to work. He also seems convinced that being tough is what the public wants. Sam Donaldson, ABC News, the White House. - And he did indeed fire those air traffic controllers. Let's go forward three years in the UK. - Mr. Chairman, what we have seen in the past few weeks is not picketing at all. It is an attempt by force to prevent others from doing what they have a right to do. (audience applauding) It is intimidation. It is unlawful assembly. Our duty demands and the national interest requires that we see that violence does not pay and is seen not to pay. (audience applauding) Let violence cease. Let the law be kept. That is the spoken and unspoken hope and wish of millions of our fellow citizens. - So those two confrontations were emblematic of the gloves coming off with the rise of the new right when Ronald Reagan fired the air traffic controllers in 1981 and Margaret Thatcher took on the miners who had been striking through much of the 1970s in Britain and broke the miner's union. And this signaled a start of a very different attitude toward unions on the part of right of center parties. And many people sort of date the big change here to those governments and those particular sort of symbolic confrontations. But the truth is that that was the tip of an iceberg that has a much longer provenance. So if you look at union membership in the US, you can see this starts from, goes from 1973 to 2009, and it basically falls in half. And if you take out public sector workers, it's overall union membership falls even more dramatically. You can see here that private sector unions have all but disappeared. But what do these numbers ignore? What do these graphs not tell us? Most people know that. Most people know that since the 1970s, unionist rates, unions have gotten a lot smaller. Yeah? - [Male Speaker] There are a lot of jobs going away. - Yup, that's true, but what else, what do these numbers not tell us? What they don't tell us is it's not since the 1970s. It's since the 1950s that the high watermark of union membership in the US economy is really in the late 1940s and early 1950s. The Wagner Act protecting unions had been passed in 1935 as part of FDR's new deal legislation and had had very tepid opposition from business initially partly because of the fears of Communism, but that rapidly changed once it became obvious that the US economy was not gonna disappear, was not gonna collapse. And through much of the 1940s, there was a lot of lobbying to either get rid of Wagner or cut it back. And in 1947, Republican Congress passed the Taft-Hartley Act which over President Truman's veto by the way, so it was a very powerful Republican Congress did that, takes 2/3 in both houses, and Taft-Hartley cut back on a lot of union protections. So part of it was the politics, but we've seen as America has transformed from a manufacturing to a services economy, it's become less and less unionized partly for those sorts of political reasons and part because it's much harder to organize service sector than it is manufacturing, and mining, and things like that, and partly because more recently, the advent of free trade. The fact that capital is back to the Hirschman point, capital has many fewer exit costs than labor. Capital is more mobile than labor, so capital can leave at much lower cost than labor. Think about what I said about different exit costs being important when we talked about Hirschman. And those exit costs for capital have come down, and down, and down, and so we saw last time if they can't make enough money in China, they just move to Vietnam. And then they'll go to Cambodia or some place else. So you've got low exit cost for capital, higher exit cost for labor. And so that makes it very difficult to sustain unionization. And this as somebody said when we were talking about, this has a lot to do, the red line there is a share of the income going to the top 10% and the blue line is the decline in unionization since the '50s. And as you can see, as unions get smaller and smaller, the inequality is growing. And somebody else mentioned as this has happened, we've seen a declining share of income going to the middle class. Again, I'll post these slides and you can look at them at your leisure. And this is not only the case in the US. You won't be able to see these probably, but these are all countries in which a different. You'll be able to see them when I post them. They're all countries in which the proportion of the work force has declined over the last several decades. And they're all the major economies. Unions have become smaller and weaker. There are a few exceptions, places like Belgium, Chile, Mexico. It's more or less remained flat. It seems like Finland is about the only country in the world in which unionization rates have actually increased. And so what are the implications of that? The implications of that are that unions which as the gentleman here said are probably the best institutional component for creating solidarity among voters to vote for certain platforms are decreasingly well equipped to do that. And particularly, it has been conventional wisdom that in two-party systems, the unionized workforce is relatively small so it's not gonna be able to get a lot of redistribution. And so this is a picture of inequalities in the US, the UK, Germany, and France over the past century. For all countries, it's this U shape. So we go from the gilded age to where we are now more or less in this U shape with the more equality in the middle of the century. But as you can see, the U is deeper in the US and the UK which have two party systems and it's shallower in France and Germany which have multiparty systems. And this has been a source of the conventional wisdom that two-party systems are less redistributive to the median voter than multi-party systems. This has been conventional wisdom for a very long time. And part of the reason here is as the left of center party is less and less connected to a powerful union movement, there's gonna be less demand on it to deliver those kinds of redistributive changes. And indeed, there's scholarship mostly from 20 or 30 years ago suggesting that multi-party systems are more responsive to the median voter in the policies that they actually enact. Whether that is in fact true anymore is what we'll talk about next time. See you on Tuesday. (calm music) |
Power_and_Politics_in_Todays_World | Lecture_11_Democracys_Fourth_Wave_South_Africa_Northern_Ireland_and_the_Middle_East.txt | - So, in the last several lectures, we've been looking at the combined effects of the disappearance of communism as a viable alternative, both in practice and in the imagination of mainstream political thinking as a serious economic alternative to capitalism on the one hand, and then on the other, the effects of the declining power of organized labor, vis a vis capital, in many of, if not all, of the Western democracies. And the effects of that have been to enhance, in one way or another, the prestige of what was called neo-liberalism at home and the Washington consensus abroad. And we've seen how this played out, particularly in the agendas of privatization, the increasing independence of private money in politics that we talked about last time, and the erosion of what had been the social democratic consensus in Western Europe and the great society in the United States. The next four lectures, we're gonna refocus attention back to the global order and particularly the idea that gained currency in the first decades after the collapse of the Soviet Union and the end of the Cold War, that there really was gonna be a new global order marked by the spread of democracy at home and the, it was the spread of democracy within countries and the reshaping of global international institutions. And that's basically what's gonna preoccupy us. The first two lectures are gonna be focused on the spread of democracy or the expectation that it was going to spread around the world, and then the next two are going to deal with international institutions. If there was going to be a new, global, geopolitical order, what form would people expect it to take? I've called the lecture today Democracy's Fourth Wave, question mark, and I'm gonna focus principally on South Africa as a remarkable transition with some comparisons to the Middle East and Northern Ireland. But I should say something about this language of waves. So, the idea of waves of democracy came into currency, at least in the academic literature, in a book by Samuel Huntington published in 1991 called "The Third Wave", and it had, and I'll talk a little bit more about the waves in detail shortly, but Huntington had argued that a third wave of democratization following the first two, the first wave having been the older democracies that democratized very slowly over the course of the 19th and early 20th centuries, the second wave being mostly the democracies that emerged in the era of decolonization, of Africa and Asia after World War II, and Huntington posited that despite expectations he, among others, had written an article in I think it was in the late '70s, maybe in the early '80s predicting that we wouldn't see much more democracy. By 1991, in "The Third Wave" he said, well, we can really date this new wave of democratization to the Portuguese Revolution of 1974, which led to democratizations in the former Portuguese colonies and elsewhere. He had... By the time he wrote that book, he had also seen the democratizations across Eastern Europe which had largely occurred in 1989, but the changes in the Soviet Union lay in the future and so he didn't have anything to say about that. And, indeed, he had very little to say about the South African transition that we're gonna focus on today, which was also mostly in the future. And the other books that came out around that time, such as Adam Przeworski's "Democracy and the Market", also were basically looking at this period in the last part of the 20th century from 1974 through the redemocratization of Eastern Europe, and they thought, "Is this the third wave?" Since the collapse of the Soviet Union and some of the developments we're gonna be talking about over the next several lectures, there's been a lot of speculation about a fourth wave that would incorporate some of the cases like the South African transition, the Arab Spring that was much heralded in 2011 and 2012. And we'll hold off assessing whether or not we think there's a fourth wave is under way or likely until we've gone through looking at the conditions for the creation and survival of democracy and how we should think about the various efforts, some successful, some not, at democratization in recent times. So, the South African transition to democracy which occurs in the late 1980s and really starts and goes through the first election of 1994, one thing to say about it was a highly improbable outcome. Nobody really expected it. Indeed, if you go back to the 1970s, which is when I started out in political science, it was a very common observation to say that some of the political conflicts in the world were just so difficult they were never gonna be solved. They were, basically, sort of the political basket cases of the world. The conflicts were just too entrenched, the possibility of a successful outcome too elusive, and that we just have to sort of come to grips with the fact that there's some political conflicts that can't be solved. And the three exhibits that were always wheeled out to make this point were the conflict between Protestants and Catholics in Northern Ireland, the Israel/Palestine conflict which had been going on in one form or another since 1948 at least, and South Africa, the struggle between the apartheid regime which had come to power in 1948 but had replaced a previous government that had also been heavily racist. And many times, people said, "Well, the South African conflict "is the most intractable of intractables." Partly because it was based on race and racial exclusion is very difficult to split the difference when you're... One of the terms we use in talking about identity politics is that identity politics tends to involve indivisible goods, not divisible goods. If you think back to our earlier discussion of the divide a dollar game, that presumes that you're dealing with a divisible good, right? So, you can divide it up in different ways and you can get, therefore, you can form different coalitions and so forth. When you're dealing with indivisible goods, those options are unavailable. So, we had an indivisible good, namely race, and then we had a small, white minority, less than 20% of the population, considerably less, and they were in power and had no, they were in control of the most powerful military in South Africa, in Africa until the prospect of a transition became real, they were widely known to have nuclear weapons, which they eventually gave up, and it didn't seem like they would give up with a fight, without a fight and go anywhere. And so most people, even people who thought the system was unstable, which many people did, thought that it would be much more likely to end in a civil war or a bloodbath. And, of course, it was also wired into the Cold War because it was seen as staunchly anti-communist and so supported by many Western governments in the 1980s. And, indeed, in the US there was a huge fight over sanctions against South Africa during the Reagan administration when the Democrats controlled Congress, as you now know, and eventually the sanctions were passed over Ronald Reagan's veto in the mid-1980s, which is a difficult thing to do. You need two-thirds in both houses. So, you can see, by the 1980s, the support for the South African regime was certainly hemorrhaging in the advanced democracies, but, nonetheless, they were wired into the Cold War until the end of it and so didn't come under great pressure from Western governments, certainly not the US government. So, when I left South Africa in 1972, I didn't believe I would ever go back. I thought that the possibilities for a white, young, white person of my, I was 15 years old, the possibilities for making a life there were virtually non-existent. By that time, once you had turned 16, they would not give you a passport if you hadn't been to the military, and by that time, military service increasingly was not sitting around in the desert being ordered about by a lance corporal but doing crowd control in places like Soweto, and I decided it was not for me. And I really did... And my beliefs were not unusual. The vast majority of people in the 1970s would have said this conflict is gonna get worse before it gets even worse, (students laugh) and the outcome is not, certainly a good outcome was not in view. And so, the world was shocked on the second of February in 1990, when Americans turned on the evening news and this is what they heard. - [Announcer] Sitting in tonight, Diane Sawyer. - Good evening. They were bold words, words that millions of people had waited years to hear. Today, in a 45-minute State of the Nation speech, President F.W. de Klerk transformed 30 years of South African politics. He declared the African National Congress, the major black opposition group in South Africa, to be legal. He promised that its most famous leader, Nelson Mandela, would soon be free. Our first report is from ABC's Richard Sergay in Cape Town. (crowd singing joyfully) - [Richard] Nearly 5000 people who gathered near Parliament to protest against Apartheid today wound up celebrating. - It's like a dream come true, that's true. - It's a big step forward, a big step forward. (brass band playing) - [Richard] Opening his first parliamentary session as president, de Klerk made one thing clear right away. The government is ready to begin negotiating South Africa's future with the black majority. - The season of violence is over. The time for reconstruction and reconciliation has arrived. - [Richard] de Klerk told his colleagues he was implementing a series of immediate reforms, partially lifting the state of emergency, suspending executions, and easing restrictions on internal opposition groups. Then came the biggest change in policy. - The prohibition of the African National Congress is being rescinded. (crowd cheers) - [Richard] For the first time in 30 years, the white government's sworn enemy, the ANC, will be allowed to operate openly in South Africa. It wasn't until the end of his speech that de Klerk finally mentioned the name everyone was waiting to hear, Nelson Mandela. - That the government has taken a firm decision to release Mr. Mandela unconditionally. I am serious about bringing this matter to finality without delay. - [Richard] But de Klerk says Mandela won't be released until what the president called his personal circumstances and safety are worked out. de Klerk has met nearly all the conditions that Mandela had set for leaving prison. - I believe that we have to give Mr. de Klerk very, very considerable credit, and we may be seeing history in the making. - [Richard] But leaders of the African National Congress in exile were more cautious. - Quite clearly, an important step forward has been made. But perhaps it's half of a step. - [Richard] Still, President de Klerk's speech confirmed the worst fears of white conservatives. - We were fighting with everything we have to retain the right of self-determination of the white people and to secure that there will be no domination of black people over whites. - [Richard] Growing pressure here and abroad for real change left de Klerk with very little choice. Today, he challenged critics on both sides. "Walk through the open door," he said. "Take your place at the negotiating table." Richard Sergay, ABC News, Cape Town. - So, that was an astonishingly dramatic moment. It's hard to overstate, as I said, how much it surprised many people across the world. And not only that. Everything that de Klerk said, he followed through on. In early... Early in that year, they unbanned all illegal political organizations, they released all political prisoners, including Mandela from his prison, and legalized all political organizations, many of which had been banned. They then, over the course of the next three years, negotiated a settlement and voluntarily gave up power, and since that time, we've had elections in '94, '99, 2004, 2009, 2014, and 2019 which have been judged by independent organizations to be genuine elections, free and fair elections, not rigged elections. It is true that the South Africa has not passed what political scientists sometimes refer to as the two turnover test. The political scientists don't like to call a country a democracy until it's happened twice that a government has lost an election and actually given up power. 'Cause you don't know what they'll do if they lose an election, and we don't know what the ANC would do if it lost an election. So, in that sense, it's not a consolidated or an established democracy, but there are some pretty reasonable signs in a sense that, first of all, we have had turnover in the leadership. We haven't had a sort of situation as with Mugabe in Zimbabwe where he staggered on for decades until he was finally dragged out of there. Instead, we've had hard-fought but nonetheless successful transitions after Mandela stepped down after one term. Some people say his most important, the most important decision of his premiership because it was a norm-setting that people don't stay indefinitely. Followed by our provost yesterday. Then we had transition to Mbeki, to Zuma, and then a very hard-fought transition to Cyril Ramaphosa more recently, but we have seen people giving up power in the leadership of the ANC. And then it's also the case that the ANC has become accustomed to losing elections and giving up some power because in addition to Cape Town, which is controlled by the opposition Democratic Alliance for a long time, they have in more recent elections, they have lost control of the other major cities, Port Elizabeth, Pretoria, and Johannesburg to opposition parties. And so... And local politics is not unimportant in South Africa, and so the notion that when you lose an election you actually give up power, while it's not been embedded at the level of the national government has some traction in the South African experience. There are many problems with corruption in South African politics, much of which will concern us, or the causes of it will concern us on next Tuesday, but by the standards of the evaluation of most democracies, South Africa looks pretty good. They have the other indicators you might be concerned about. They have a relatively free press, freedom of association, an operating legal system, and so on. And so certainly by the standards of many new democracies, it's been quite a remarkable story. So, today's agenda is we're gonna talk first a little bit about getting and keeping democracy, what's required to get democracy, what's required to keep it. This will involve revisiting some of the themes of modernization theory that I talked about earlier when we were talking about China. Then I'm gonna talk about transitions to democracy, and I'm gonna separate out four types of transition and say something about them. Then we're gonna zero in on one of those four types, which are negotiated settlements of the sort that we witnessed in South Africa during the 1990s, and that will lead us to some evaluation of success, failure, and inconclusiveness by looking at that transition in contrast with the continuous failure after failure in the Israel/Palestine conflict and the more inconclusive results in Northern Ireland. Okay, so getting and keeping democracy. There's some historical context to this, some of which I have alluded to. Huntington coined the phrase three waves, and the first wave begins in the late 18th century, really, with the American Revolution and the French Revolution and it extends to the early 20th century with the enfranchisement of women in most countries, except Switzerland when it didn't come for another half century and the enfranchisement, the reducing the voting age often from 21 to 18, which came in many countries during the 1960s. So, it was a very gradual and long transition. And that's not really surprising, if you think about it, because elites were afraid of democracy for reasons we've already talked about. They had some intuitive grasp of the logic behind the median voter theorem, although they didn't use that terminology, but they were worried. They were worried that if they suddenly enfranchised the whole population, it would threaten their position and their property. And so, for instance, in Britain, we got significant expansion of the franchised in 1832, more in 1867, more in 1885, women in the 1920s, and young people, 18 year olds, in the 1960s. So, it's a sort of some baby steps before big steps. In the European countries it happened a little more rapidly. A lot of enfranchisement in countries like Germany in 1870, but still, we see a gradual process of evolution, if you like, toward democracy rather than universal franchise. And that is one thing that differentiates the first wave from the second wave because the second wave is really bound up with anti-colonial struggles against Britain, particularly, in Africa and Asia but also against the other colonial powers, and those were countries that went from very restricted or no franchise to universal franchise more or less overnight. So, it is was a very different kind of set of transitions, and they had, they happened much faster for that reason. And then the third wave, as I said to you already, is retrospectively dated by Huntington, somewhat arbitrarily one might say, to the Portuguese Revolution in 1974 and what subsequently followed in Eastern Europe, in particular, in the late 1980s. Now, this creates this talk about waves of democracy and getting more and more democracies. And indeed by the turn of the 21st century we had more democracies in the world than more non-democracies for the first time in human history, even if it was the case that more people lived in non-democracies than in democracies because some of the non-democracies have huge populations, but, nonetheless, it creates a kind of air of teleology that informed the idea of the end of history that we've already talked about, this idea that the world is steadily becoming more democratic and it just would be a matter of time before these waves would come along, and eventually the whole world would be democratic. That was never... That was never a very good bet. So, if you think about the three waves, each one of those waves has had significant undertow. So, if we think about the first wave, yes, we had the revolution in France in 1789. By 1791, it had turned into the terror and something which nobody would think was recognizably a democracy. In Europe, in 1830, across the continent we had democratic revolutions in just about every country except the UK, which had led Marx and Engels to think this was the beginning and the end of capitalism. And even in the UK we had the Chartist movement, which people were terrified might turn into a revolutionary movement or hopeful that it might turn into a revolutionary movement depending on your point of view. By 1832, those revolutions had collapsed and the monarchies had all been restored across the continent. The same thing happened in 1848. Again, democratic revolutions swept Europe and by 1849, 1850, they pretty much all collapsed, and the non-democracies were back. If we think about the 20th century up through the end of the first wave, of course we had the advent of fascism, of Nazism in Italy and in Germany, and in the US, massive retrenchment after the Civil War of democratic rights through the Jim Crow period, huge disenfranchisement. I talked about the Supreme Court's reading of the Civil War amendments last time to restrict their application and particularly to restrict the application of the 15th amendment, where you couldn't exclude people on the basis of race but you could find surrogates like literacy tests and poll taxes and so forth. So, there was lots of undertow to the first wave of democratization. It's not like you get these things and then you can't ratchet back. If you think about the second wave, again, many, many democracies that were established, when the British left, the first election turned out to be the last election. Lots of theorizing and literature and speculation about why that is the case or why that was the case. Perhaps one of the most plausible accounts is that where the British governed by direct rule, as in India, democracy got better traction because you had whole generations of Indian elites who went to Oxford and Cambridge and came back and embedded democratic norms and values in their institutions before the British left, whereas in South Africa, and much of the rest of Africa, the British governed indirectly, which essentially was much cheaper because you didn't have to set up your own whole, your whole bureaucracy and civil service and institutional capacity and instead found local chiefs or sometimes more or less created local chiefs and said, "Provided you deliver what we want, "you can govern as you please." And so there was much more indirect rule in Africa, unlike the direct rule in India, so some people put it down to that. Wonderful book on this subject by a man called Mahmood Mamdani, who's a professor of political science and anthropology at Columbia, called "Citizen and Subject", about the legacy of direct versus indirect rule. There are other theories, and it's not an unproblematic theory itself because, after all, Pakistan had the same direct rule as India and didn't have such a happy, democratic future as India did. So, it's a complex story, but, nonetheless, there was plenty of undertow. Many, many democracies that had been created, as the imperial powers left, turned out not to survive. And then, if we think about the third wave, again, there's also been significant undertow there. We've seen, I probably should've put Pakistan on the second wave, but much of Latin America. We've seen, we saw collapses of democracy in various countries or at least serious compromising of it in Burma, in Fiji, where coups seem to come and go, Algeria in 1991 we had a coup, and of course the ongoing questions about Russian democracy that have become more acute with time. So, the idea that there's some inevitablism that democracy's just once you get it, you keep it forever, clearly not true. Democracies can fall apart and sometimes they do. And so that puts on the table the question so what are the conditions for democracy to arrive and what are the conditions for democracy to survive? And that provokes some theoretical questions, and there's a huge theoretical literature in political science. I underlined the science because political science aspires to make predictions. It aspires to tell us what the conditions are under which we are likely to get a particular outcome. And so, what can we predict? And on democratization, one distinction that political scientists have made that was not present in the early modernization literature that I was talking about when we discussed China is a distinction between transitions to democracy and the survival of democracy. And, in particular, Adam Przeworski and his co-authors in this book, published at the turn of the century, said we should really think differently about these two things, and the reason is that what produces transitions to democracy is probably inherently unpredictable. It depends on so many contingencies that have to line up that you really can't, you can't predict when it's going to happen. On the other hand, you can say something about, however you get a democracy, how likely is it to survive. And so Przeworski and others said we should really separate out these two things and stop talking about transitions. And so he, his work and the work of many other colleagues represent pretty much a conventional wisdom about survival. So, if I, just to do a little polling here before I tell you what they say, if I said to you, have we got the microphone? What do you think is most important to have for a democracy to survive? Anyone. (mumbles) - [Student] An educated population. - An educated population is one possibility, others. - A strong judicial system. - Strong judicial system. - A healthy economy. - Healthy economy. Couple more here. - [Student] I would say that all actors in the economy or in the political system believe they can win more from the system than from trying to overthrow it. - Okay, yep. - A middle class. - A middle class. Barrington Moore supporter over there, yeah. So, okay, so, it is the economy, stupid, as Clinton is famous for saying. People used to think political culture mattered a lot or that certain kinds of religions were more or less compatible with democracy, but if you look at the data, none of those things is much of a predictor. And so, here are the main things that people say are important for a democracy to survive. One is per capita income of about 14,000 2019 dollars, that if you... This is Przeworski and his... I've converted it into more 2019 dollars, but basically, their data set, which is through the last century not the current one, shows that if a country has, if a country is already a democracy and has per capita income of that amount or higher, its democracy survives indefinitely, and if it falls below that, it becomes more fragile and the further below that it falls, the more fragile it becomes. Notice that doesn't say anything about inequality. Because you can have a very high per capita income in a very unequal country. So, there's per capita income, then a second is the diversification of the economy. This is the whole oil curse literature. We talked about this somewhat in relation to Hirschman, "Exit, Voice, and Loyalty", that if the whole economy is dominated by one sector that's controlled by the government, which is usually the case if it's an oil curse economy, people will very, be very reluctant to give up power if it's the only access to advancement and success, and they'll be very prone to grab power if they can if it's the only access to, provides the only access to advancement and success. And so when Al Gore didn't become president in 2000 in a knife-edge election which he, many people thought he'd actually won, he wasn't happy. Nonetheless, he could go on corporate boards, go around the world making movies about climate change, and I saw his net worth is 250 million dollars. It's not so bad for him. (students laughing) Not so bad for Al. It wouldn't be the case if he lived in an oil curse country. Now, some people would say and some people say oil curse is, generally speaking, bad for democracy. And notice that the countries that have discovered oil after they've been diversified, like Norway or Scotland, don't have this problem. So, it's the diversification which creates multiple opportunities for people to advance in the world other than grabbing the commanding heights of the state. It was important, by the way, in the early American republic that when the capitol was, whether it was in New York, Philadelphia, or Washington, these were not places the elites wanted to be. They wanted to be out making money in Virginia, on their farms, and so the desire to hold onto power was not particularly strong. It wasn't that valuable to them. And Washington had to be persuaded to run again. So did Jefferson have to be persuaded to run for a second term. That's the kind of thing that highlights the fact that political power wasn't that valuable. Alternation in power. I talked about the two turnover test already. So, this is the notion that if you give up power, you're more likely to give up power in the future. Alternation predicts alternation, coups predict coups, right? So... India maybe got lucky, whereas Pakistan didn't and so then you get... There's a big, big... This is, again, a paper by Adam Przeworski saying if you get one or two turnovers, and certainly if you get three turnovers, they go on forever. Now, what's the causal story here is another matter because it may be the conditions that lead to people being given up, being willing to give up power once remain. And so it's not the alternation that's predicting the giving up power but the conditions, like a diversified economy, that's actually making them to keep giving up power into the future. Nonetheless, countries that have histories of alternation tend to alternate, and countries where there's a history of the military jumping into politics when they're not happy, they tend to do it. Another consideration I will just put out there 'cause it's been a theme in the course so far and it will be more of a theme in the future is my intuition, I haven't done systematic research on this, but that, that downward mobility and loss aversion are more important than people give credit for. So, if you look at some of the countries in which democracy has become fragile, often it's not the poor and it's not people complaining about inequality, but it's downwardly mobile people, middle class people in Cairo who find their prices going up, for example, and one could give many other examples. It can be middle class disaffection when people start to experience or expect that they're going to experience downward mobility. But the takeaway is this is largely an interest-based story. It's not so much norms or culture, it seems, but more, more interest-based. Though, as I said, disentangling some of the causation between the interests and the norms, like the norms of giving up power, can be very tricky to do empirically. Let's, though, talk about transitions. Przeworski said it's a mug's game trying to predict transitions. Too much contingency in there. And, indeed, after 1989 he wrote a piece saying now we're gonna see hundreds of trees get wasted by people coming up with theories of transitions, and at the end of the day, they're not gonna be worth a hill of beans because there's too much contingency. Actually, it didn't stop Przeworski from coming up with his own theory of transitions, but be that as it may, I wanna say not so fast. And that while it's true we can't predict transitions, we can learn some interesting things about the conditions that are necessary for them to be successful, and that's what I'm gonna be talking about next. So, what I wanna do first of all is single out some different kinds of transitions from one another. They're basically four ways to get from non-democracy to a democracy. One is intervention from outside. A country goes in and installs democracy somewhere else. And the two famous success stories here, of course, are West Germany and Japan after World War II, and so the notion that you can't install democracy by force is obviously false, at least as a categorical statement. Now, there are many, many reasons that people give about why those installations of democracy were successful, ranging from the fact that they both had histories of democracy in the inter-war period to the fact that the Allies stayed there for decades, that they invested in the Marshall Plan in Germany and economic resuscitation of the Japanese economy to the Cold War and the creation of NATO. But for whatever reason, there have been external installations of democracy. We've seen attempts at that in countries like Afghanistan and Iraq that we'll have more to say about later, where the answer's not so much. Perhaps 'cause all of those conditions were missing or perhaps for other reasons. We'll be digging into that later. The second way in which you get democracy is by a revolution from below, democratic revolution. People are opposed to autocracy. They get together, have a revolution, and create a democracy. Some success stories there, of course, the American Revolution. People tend not to talk about the US when they do comparative policy, but the US is one case in the world and the American Revolution was a successful revolution from below. It was of an elite below in that was an anti-colonial revolution run by an elite. Nonetheless, it was a revolution from below. The Portuguese Revolution, to a large extent, was a revolution from below. Of all the 1989 transitions, the one that comes closest to a revolution from below was the Romanian transition where the Ceausescus were just pretty much wheeled out of a court and shot behind a, behind a barn. One might say that had the Mensheviks been more stronger than the Bolsheviks, Russia might have had a democratic revolution in 1917. The Mensheviks were democrats and might have had a very different history, and it turned out, of course, they were not. The Libyan civil war, ongoing civil war started out with an expectation on the part of NATO and the Western powers that it was gonna be, lead to a democratic transition. Again, not so much. Another subject we're gonna revisit. Now, a third path to democracy is when the elites in power decide to give it up. The military decides to leave power for reasons that, again, people debate. Or even in a non-military regime, an authoritarian regime, the leadership decides to give up power. This happens when Juan Carlos decided to give up power in Spain. It happened in Brazil. Of course, there was pressure from below. It's not, these... Another of way of putting it, these are sort of ideal types. Many transitions are hybrids, to some extent, of them, so I wouldn't wanna say they all fit in one of these boxes but the center of gravity is elite willingness to give up power. Would you look at what happened in Egypt in the beginning of the Arab Spring? The military got strongly behind the transition, for reasons that are very well described, I think, by, in the book "Thieves of State" by, who wrote "Thieves of State", somebody? Sarah Chayes. They got, very much got behind the initial revolution, but of course, they came roaring back into power a year and a half later. So, again, not so successful there. And then, this is... As I said, political scientists are not famous for elegant terminology. This is Huntington's term, transplacement, and the only reason I use that is Przeworski's term is even less elegant. He calls it extrication. It's like going to the dentist. So, what is, how do you get democracy by going to the dentist? This is a negotiated settlement, a negotiated transition from authoritarianism to democracy. And this is, South Africa's a classic case. Of the East European ones, the closest to that was the Polish transition, which really starts long before 1989. It starts in the early 1980s where the Solidarity movement emerges in Gdansk and the regime finds it's necessary to make various concessions to them. And it's almost a decade-long negotiation in Poland which culminates in 1989, but it's basically a negotiated settlement. Israel/Palestine and Northern Ireland, again, not so much. So, again, you can see that they're outside, and from inside above, below, and negotiated. That sort of exhausts the universe of possibilities. Now, most of the interest has been in this last category, among academics, and I think that the reason for that is probably the reason, partly the reason Przeworski gives, that if you think about the other three, the extent to which they really depend on contingencies is very high. When will... Predicting when it'll be the case that an invading country will be willing to make the kind of investments that the US made in Germany and Japan after World War II seems, it seems like a very difficult proposition. Or predicting, if you do have a revolution, whether there'll be divisions among the opposition groups. Predicting, if... As I said, if the Mensheviks had been stronger, maybe they would have won. There were democratic, there were more democratic forces in the French Revolution that lost out. They were less... There were people that wanted to restore some sort of monarchy after the American Revolution. They lost out. But, again, this seems like contingencies that you're very unlikely to be able to come up with any generalizations about. Similarly, with elites deciding to transform things from above, it's probably gonna depend on unpredictable conflicts among military or political elites, the alternatives that they have available to them, and other things that are gonna be, from the point of view of coming up with any kind of serviceable generalizations, accidents of history. So, there is literature, but by far the most attention has been focused on these transplacements. And I think one is that people have thought, actually, they do lend themselves to the tools of social science, even if not predictive tools in the sense of point predictions, that you can use, among other things, the term, the insights from game theory and you can predict at least what makes them more likely to succeed or less likely to succeed. And then I... This was just my observation. It could be wrong, but I think the other reason there's so much attention to negotiated settlements is that it seems normatively good. Democratic paths to democracy seem like a great thing, that we don't have to kill thousands of people, perhaps hundreds of thousands of people. We can find a way to do it peacefully. And so I think the fact that, that this seems to have desirable, normative properties makes people want to understand what are the conditions under which you're likely to get a negotiated settlement to democracy. So, what are the conditions? And I'm gonna give you a sort of stylized picture of these kinds of conflicts and how they tend to get resolved when they do. And we can think about it as the governing elite, government on the one side and then the opposition movement, not a political opposition party because probably the opposition is banned, but there's a government and an opposition but they're not monolithic. They're not monolithic. They're divided into at least two groups. And so, you could think of the standpatters or reactionaries who are gonna fight all change at all times and then reformers who are at least open to change, on the government side. And then on the opposition side, you could think about a division, a mirror image division into moderates and radicals. This is not moderates and reformers about policy, but moderates being people who would, in principle, be open to a negotiated settlement, whereas the radicals think, "No, we've gotta have a revolution. "We've gotta take over and we've gotta impose "a new dispensation on the country." And what creates the possibility of a negotiated settlement is a circumstance in which reformers and moderates, individually, are not powerful enough to force change but if they could get together, they would be. Right, if they could get together and form an alliance, they would be. And so, then, the 64 million dollar question becomes well, what are the conditions that make that possible or likely? And it's not easy. The first thing to say is it's not easy. At least three things have to happen. One is the reformers and moderates have to agree, and if you think of a condition of a decades-long civil war where many people have friends and families who have been killed in this war, very high stakes politics, lots of hatred, lots of resentment and anger, it's not obvious that the reformers and moderates are going to be able to agree, even if there are groups that might be open to change. A second requirement is that the standpatters and radicals have to be convinced, co-opted, or marginalized. If none of the above, they will throw a wrench into the whole process and stop it, right? Inevitably. Or they certainly will try. So, unless they can be convinced, co-opted, or marginalized, they're probably gonna blow it up before it's done. And then the third is that you have to build support for the new dispensation because these groups have grassroots supporters, these groups have grassroots supporters, but they're not grassroots supporters for this. They might potentially be, but they're not gonna drop out of the sky. If you're gonna have a negotiated settlement, people are gonna have to build support for it. So, those three things have to happen, and they have to happen at the right time. And we'll see what happens when they don't happen at the right time shortly. Okay, so that's a lot, right? It's a big menu. And this means there's a, for a negotiated settlement, there's a kind of threading the needle quality to the whole thing. And so when we think about the conditions for successful negotiations, one of them is obviously luck, right? You need luck. And this is what we're gonna, what Przeworski was pointing at when he said you can't generalize about this. And I'm gonna use three word, three things you need. You need luck, you need leadership, and you need legitimacy. The reason you need luck is there's so many more ways for them to fail than to succeed. So many things have to line up in the right way at the right time in order to get a successful negotiated settlement. People often talk about the importance of leadership without defining it. I'm gonna mean three things by leadership. One is people with an unusual appetite for risk. Most politicians are pretty risk-averse. They mostly wanna keep what they've got, get re-elected, and stay out of trouble. Politicians who have the leadership qualities that are gonna be required have to have an appetite for risk, but it's not just any old risk. They... Gambling is risk-taking but it's reckless. You don't want reckless risk-taking. You want people that have good strategic judgment about what are the risks that must be taken and how must they be taken? And to what end? And in particular, you're gonna need people who are willing to take big personal risks for a cause that's larger than their own personal interest and then to have the good strategic judgment to know what is a good calculated risk and what is a reckless move that is not likely to succeed. And the third essential element of good leadership is empathy. Now, empathy is not sympathy, right? Sympathy is feeling sorry for people. Empathy is understanding what people need. It's seeing the world from the point of view of the person you're dealing with. Because if you don't understand what they need, you will not have a successful negotiation with them. So, those are the three important elements of effective leadership. Now, when I say legitimacy, I'm alluding here to the debate about democratization from above and from below. And so the story here is that some people say, "Oh, it's not legitimate "if it doesn't come from below. "The people have to validate it." Others say, "No, it's all about elite pacts," elites making deals with one another. The short answer is that it is both, and the reason is that democracy is a public good in the economist sense. It's a technical public good. Supplying it to some means supplying it to all, right? It's not excluded and it's non-rivalrous in that some people having democracy doesn't exclude others from having democracy. So, it's a public good, and it's notoriously difficult to supply public goods democratically. And so that means there will be moments when elites have to take decisive action, but if they can't build support for the actions that they are taking, then they're not going to be successful. Yes, sir. - [Student] Following the dividing the dollar game and what you mentioned of marginalizing, couldn't you just marginalize elites or marginalize the lower classes and just build a good enough coalition to go over? - The problem there is who is you? Okay, there's no you to do it. So, they have to see it as in their interest. So, what I wanna do now is work through the South African and Middle East cases here from the perspective of these three considerations. So, the luck in South Africa, what was one thing that was very important was a decaying status quo. Whatever the future was gonna be, South African elites understood that it wasn't gonna be the same as the past. Partly because the new generation of black youth was much more militant than its, than their parents. They were burning down their schools, they were refusing to study in the Afrikaans textbooks, and so on and so the costs of repression were going up. That meant the children of the white elites didn't want to join the army, as I indicated the first minute of this lecture, and so the elites felt the sand shifting under their feet. That didn't mean you would get democracy, but it meant that the status quo was gonna change. The bankruptcy that ensued in the 1980s when Citibank refused to roll over South Africa's debt contributed as well. Another factor, a contingent factor that helped a lot was the disappearance of communism as a viable alternative, which made whites much less frightened of a transition for reasons I'll talk about on Tuesday, and for the ANC, the Soviet Union stopped funding them in 1983 because the Soviet Union was going bankrupt in Afghanistan and while it was fighting Star Wars spending with the Reagan administration. And so the ANC was in this situation of facing the most powerful army in Africa and didn't have the resources. They went to the Chinese, the Chinese said no, and that produced a group within the ANC that was willing to think about talking about a negotiated settlement. Leadership. I wanna spend a little bit of time on this. So, I became very consumed with the counterfactual. Here we had the successful negotiated transition in South Africa between Mandela and de Klerk on the one hand, and if you look at Israel/Palestine, it had many structural similarities but it had failed. So, but for example, in the early 1990s, there had been something called the Oslo Accords. People had... There'd been a lot of negotiating between the PLO and the Israeli government, and many people started to believe that there was gonna be a transition to a settlement. And if you went there in 2004 or 2005, people, most people thought this is coming to a successful conclusion. And, indeed, Arafat moved somewhat to the middle. He recognized Israel's right to exist. He took some other of the preconditions off the table that had been so important, and Yitzhak Rabin, they seemed to be converging on a settlement. There was a lot of work had been done to build support for it in both communities. People expected it. Yet when Rabin was assassinated by a right-wing Israeli, it all fell apart. Peres came into power. He temporized, he moved to the right. Arafat was phoning him three times a week saying, "We've gotta get this done," because as he had moved to the middle, that's when Hamas emerged on his flank and seized all the symbols of Palestinian nationalism. And so he realized he had to get a deal. Peres tacked to the right, basically cut off the talks. Netanyahu won an election six months later anyway, and the sort of window of opportunity closed. So, I got interested in what would have happened if Rabin hadn't been assassinated and maybe if de Klerk had been assassinated in 1992. How would things have looked differently today subsequently? And so, I went and interviewed de Klerk about this, and I asked him. And I said, "What would've happened if you "or Mandela had been assassinated during this process?" And the backstory here is that as the negotiations had been getting into trouble and breaking down, violence had been increasing in 1991 and early 1992, and this showed up with increasing white anxiety which expressed itself in the government losing by-elections. They lost three big by-elections in a row. And de Klerk came, responded to this by going on television one night one night and saying, "I'm calling a referendum "on whether I should negotiate a settlement "with the African National Congress." So, that's the backstory here. So, what he said was the following. He said, "Well, "if Mandela had been assassinated in the late 1980s, "it might have scuttled things because "at that time he was having secret talks with the government "and it is not clear there was anyone "willing to do that in the ANC." And he, Mandela, had kept them secret because he thought if they failed, people could denounce him as a doddering old man and it wouldn't damage the ANC's cause. That's what I mean. A willingness to take risks for a cause that's larger than yourself. de Klerk said, "In my case, "if I had been assassinated after the referendum," which he won by a two-thirds majority, he said, "it wouldn't really have mattered "because whoever replaced me would have gone on. "But the reason I went on television "and announced the referendum was "that everyone in my cabinet would have opposed it "and everyone in the national party would have opposed it. "And so I just went in and announced it as a fait accompli. "And if you think of, "if you think of a circumstance in which "I had been shot before I had called the referendum, "there wouldn't have been anyone in a position to do it. "And even if they had wanted to do it, "it's not something a new president can do as their first, "they have to build support in the cabinet and so on." So, then it might well have played out, it might well have played out differently. That was very important because the fact that he won that referendum and essentially took on the far right in his own party meant that the ANC could convince the skeptics there that he was serious, that he was essentially burning bridges with the far right of the national party establishment and making his future depend upon getting a successful negotiation with the ANC. And so that was a level of risk-taking that sometimes what solved what political scientists called the commitment problem. He could... Skeptics in the ANC could be persuaded by Ramaphosa and Mandela, who were for negotiations, that this was a person with whom they could work. I've already, so that was one of the ways in which he took on the opposition. The other thing to say is that it's very important to build support from below. If you look at, fast-forward to 2000 when Bill Clinton is about to leave office, he decides, like many American presidents before him, that the last thing he's gonna do is solve the Middle East conflict on the way out in the summer of 2000, and he brings Arafat and the Israeli leader who by that point was Barak. He drags them to Washington and says, "You need to make this deal that wasn't made five years ago "when Rabin was killed." And they both say no. And he tries to, he thinks... Bill Clinton thought if he just can get people in a room and twist their arms and look them in the eye and tell them that they're the most important in the world, people in the world, he could get it done. And the reason he couldn't get it done was that there was no support in either community in the summer of 2000. The Palestinian support was below 10% and the Israeli support was below 8%. So, even though it was substantially the same deal as five years before, it didn't matter. There was no support for it. There were no constituencies for it, and this is why below matters as well as above. And so all, The Economist was writing about these articles. What's wrong with the Palestinians? They never miss an opportunity to miss an opportunity, and so on and so forth. They didn't understand that the dynamics on the ground were different. And so you have to have the support from below. Right, so, we can't predict transitions, but we can say something about what makes them more and less likely. For one thing, we can predict failure. In 1983, Gerry Adams became head of Shin Fein, the IRA, and people said, "Oh, this is a new kind of politician. "He wears business suits, he talks like Western politician, "maybe there can be a settlement in Northern Ireland." But, of course, in 1983, you had a conservative government in Westminster that was crucially dependent on the Northern Irish for its parliamentary majority. So, there was no question. Far from the status quo being a wasting asset, the sand shifting under their feet, there was no possibility that the Tories would make a deal with the IRA. 1997, when Tony Blair comes in, Labour's not dependent on the Northern Irish vote so then it becomes a possibility, and eventually we get the Good Friday Accords, which semi-unravel later. So, you can think usefully about the conditions here. What is it that's gonna move the elites, and will they think that they can't move? Empathy and strategic judgment, hugely important. One thing de Klerk said to me that I thought was very interesting in that interview. At the end of it, I said to him, "What do you think explains the difference "between the South African "and the Israel/Palestine negotiations?" And he said, "Well, one important difference was "that when we decided to legalize the ANC "and the opposition movements, we had no preconditions, "and when the ANC, three months later, "decided to suspend the armed struggle, "there were no preconditions. "We didn't ask them to recognize our legitimacy. "We wanted them to decommission "but we didn't make it a requirement, no preconditions." Of course, there are implicit preconditions, in the sense that if you don't deliver within a, before the window of opportunity closes, it'll all fall apart. But he said, "The fact that we, "there were no explicit preconditions meant that "we had more room to maneuver." And, of course, it meant the spoilers couldn't violate the preconditions. If you say, "These talks will go on "as long as there are no rockets," there will be rockets, right? So, no pre... That's empathy, it's understanding. It's understanding what people need. Mandela understood that even though the ANC was opposed to power-sharing in the permanent constitution, de Klerk had to get it in the interim constitution or there wouldn't have been a transition, so he agreed to it and later on it went away. In April 1993, a black radical leader, Chris Hani, was murdered, and the violence erupted and it threatened to engulf the whole process. And one of the ways in which it was tamped down was that de Klerk agreed that the election would occur no later than a year after Hani's death. So, it had to be by April 1994, so a definite date that it had to occur, and that enabled Mandela and the ANC to quell the violence and move things forward. So, that's hugely important, that you understand what the other side needs and make sure they can get what they need, otherwise it's gonna fall apart. Exit options, right? People have to find it possible to give up power. In South Africa, they promised an amnesty process. They processed civil service jobs, security for five years. This hearkens back to our earlier discussion of the Soviet Union, that the exit costs were lower because of the way privatization occurred. And when we think about the demonstration effects here of the ways in which Gaddafi and Mubarak met their end, their fates, it doesn't all go well because the leaders in other countries that might be facing opposition movements think about what'll happen to them if they leave. So, good things don't always go together. Let's give de Klerk the last word before we head out. - I think it was an absolute necessity that he had to be released. As his image grew over the years, as he became an icon, as he became a crucial flashpoint in the whole international situation also of South Africa, it became clear to me that, that he has a pivotal role to play and that he was playing it, irrespective of the fact that he was in jail. He was leaner than I thought, he was taller than I thought, and my very first impression was that he was a very dignified man. We were both reserved, but there was immediately a sort of a good, good spark between us. He had a few habits. Firstly, he was a very good listener. He would rarely or almost never interrupt one when one was making a point. On the other hand, he got highly irritated if he was interrupt. I was scoring points in the debate. He leveled the score by, fairly near to the end but before we reached the end, suddenly reaching out and taking my hand. - I am proud to hold your hand, for us to go forward together. (audience applauding) - It was a masterpiece, that taking my hand. He's good at it. Ah, I took his hand, nothing I could do about it. And lifted quietly my, in my mind, my hat to him and said, "Well done." - Okay, we'll see you on Tuesday. (bright ascending notes) |
Power_and_Politics_in_Todays_World | Lecture_23_Building_Blocks_of_Distributive_Politics.txt | - So we're now on the, into the fifth part of the course called, what is to be done? And I want to just begin by talking about the key features of the politics of insecurity, some of which we addressed earlier in the course and some in the last two lectures. But they really shape the landscape on which we have to think about, what can and cannot be done in politics, in the world as we actually find it. And so, first thinking about voters. One thing that we have learned is that local inequalities matter to people much more than global inequalities. Bernie Sanders can talk as much as he likes about the top 1% and it will motivate activists on the left of the Democratic party but most voters really care about much more local inequalities. And we talked about this earlier in the course when we were talking about how capuchin monkey experiment was misinterpreted when the angry monkey was likened to the Wall Street protester but whether the monkey was angry because she or he was not getting something that a similarly situated monkey was getting. The monkey was not troubled that the researcher had a big bowl of grapes and cucumbers. So people tend to compare themselves to similarly situated others. Oil workers might compare themselves to coal miners. Auto workers might compare themselves to steel workers. And this is true up and down the occupational scale. I think I mentioned to you, a professor would be much more troubled to learn that she or he is paid significantly less than say, $10,000 less than a professor in the next office than to learn they're paid half a million dollars less than the attorney next door. So people tend to make local comparisons. And the idea that marked hope for, that people would start to make more global comparisons, is not supported by the research of sociologists and social psychologists. Secondly, think about Rick Santelli's rant, another illustration of this point. This is the famous rant on the Chicago Mercantile Exchange right after the Obama Administration came into office and had started talking about minor mortgage relief for homeowners. Nothing like John Geanakoplos was proposing, but nonetheless he was proposing reducing interest rates at least for a time on peoples loans and it produced a lot of rage that was articulated in that video I showed you. At the end of it, his calling for a tea party and people credit the formation of the tea party with... As being, if you like, catalyzed by Rick Santelli's rant on the Chicago Mercantile Exchange. And you might also recall that one of the things he complained about was, he said, it's a moral hazard. It's a moral hazard. But notice of course, bailing out the banks is also a moral hazard, right? Bailing out the banks is giving banks the incentive to gamble with taxpayer's money. So while it may be a moral hazard to bail out the homeowner's, it's no less, you might say in some ways it's more consequential moral hazard to bail out the banks. Nonetheless, what enraged him was that do you wanna pay for your neighbor's mortgage?, he rhetorically said. So again, local comparisons that matter to people. Think about the difference between the Trump and Sander's campaigns on this question of who people compare themselves with. In many ways Trump and Sanders ran a populist campaigns in the 2016 election. They attacked Wall Street elites for being corrupt. But a difference was that Sanders talked a lot about inequality whereas Trump did not. Trump never promised to reduce inequality. He never, as I said, even said what Ronald Reagan had said, that he wants America to be a country in which everybody can get rich. All he said was that, you people have been screwed. You people have been... Left behind by the... By the policies of foolish elites, he calls everybody idiots. And I'm gonna put America first and so on. And to the extent there was a fairness argument. We talked about this in connection with Arlie Hochschild's narrative of cutting in. Some people have cut in front of me and taken something that I would otherwise have gotten, right? Again, very local comparisons that people are making. They are not thinking about what people, very distant from them in the socioeconomic order are making. Related to that, closely related to that is a loss aversion and in its insecurity, often matter much more to people than inequality. Again, think here of Trump's make America great again. Something's been taken away from you. We're gonna bring it back and do something. We're gonna recreate something that had existed before. I am gonna bring back those jobs. I'm not gonna make you rich. I'm gonna bring back those jobs. That was the essence of his campaign whereas Hilary's was, she took that other Reagan line, American's best days are ahead. Sunny optimism which is a lot harder to sustain during an era of endemic employment insecurity. And again, just driving home the point that it's insecurity and the fear of downward mobility that matters much more to people than their place in the distribution of income and wealth, globally conceived. Remember those data about Trump primary voters. Only a third of whom earned lower than the median income, below $50,000 a head per family. Another, 2nd and 3rd earned between the median income and 100,000, and the third third of primary voters not Republican voters in the general election were Trump supporters. And in those primaries, so, again, middle-class people who feel insecure, and this is, what Dan Markovits' book is about, may be just as anxious and mobilizable by a populist politician as poor people who feel insecure. So the insecurity and loss aversion matter more than inequality and it's local inequalities that matter much more to people than global ones. And then we're living in a world in which almost everywhere organized labor is weaker than it's been in decades. We've seen, in country after country, the decline of union movements, it's way down into single digits, in the US, and the majority of workers that are organized now are in public sector unions which do have some political clout. We will see next week when we talk about education. But for the most part, as a force in politics, unions have seldom been as weak, if ever, in the last 80 or 90 years, if ever, as they are today. This is particularly pronounced in the US, but we saw pretty much everywhere except one or two places like Finland and Iceland, we've seen a decline in the power of organized labor. A concomitant to that, business interests are stronger than they've been for decades. Partly because of the collapse of a serious alternative out there. Communism to the extent it exists politically is now supported by capitalists economies in places like China and Vietnam and so there's no real alternative to capitalism out there, which obviously, greatly, increases the power of capital. Particularly in an era of globalization when it can easily flee, the flying East theory that Christina talked about in the China lecture. And in these, as jobs going increasingly to technology, capital doesn't even need to flee in order to increase its leverage over labor. So we're living in a world in which labor is weaker than its been in living memory and business interests are more powerful than they've been in living memory. And then coming to Thursdays lecture of last week, we're also living in a world in which political parties have become weaker and more fragmented just about everywhere. And this is not something that has gone on for eight decades, but it certainly has gone on over the past four decades as labor has become weaker, parties on the left have fragmented and we saw that as induced fragmentation on the right in multi-party systems and all over the democratic world, this impulse to democratize parties, to get more and more direct democracy in the governance of parties, and in making of decisions has greatly weakened parties in both multi-party systems and even two-party systems and indeed, even the... The platonic form of what used to be two-party systems. The Westminster system between their going for referendum between the changes in their leadership selection rules, between adopting things like fixed parliaments and candidate selection also being decentralized, they have replicated much of the rest of the democratic world in making it much more difficult for parties to present, get elected on programmatic platforms and then implement them as governments. It's a world that's ripe for populist charlatans who exploit insecurity and promise snake oil solutions. So, what is to be done? Now at the beginning of the course when I prefaced some of these themes, in the first few lectures, I also said don't get too depressed. (class laughing) And in many respects, this is a depressing prospect. But what I want to be arguing in these last lectures is that, certainly we shouldn't give up hope and there are ways of thinking constructively about politics going forward. And indeed, in some respects, there are reasons not just for hope but even some optimism. This was a distinction Martin Luther King made towards the end of his life. He said he was no longer optimistic but he hadn't given up hope. I think there are some reasons to think that it might be possible to come up with constructive change and build regressive support for it to happen. And so that's where we're headed in these final lectures. So, a central message of this final part of the course, and if you like, presupposition of much that I'm gonna say to you, power phrasing Immanuel Kant is that, policy without politics is empty, but politics with policy is blind. Policy, politics without policy is empty and politics without policy is blind. So, let me emphasize that it's not just in the real world but in the academic literature, for the most part, people who study policy don't think very much about politics. They think about what policies would be good, what should happen. But they have relatively little to say about how they're gonna get it to happen. Whereas people who study politics tend to explain why what happens happens but have way little to say about what should happen. And so there's this sort of divergence of preoccupations. But I think that that is misguided. That we really need to think about what's desirable in the context of what's feasible and take into account the constraints and possibilities that might be offered for thinking about policy rather than thinking about it in a vacuum. So just to give a couple of illustrations of policy without politics. Perhaps the most influential book of political economy written in the last couple of decades is Piketty's Capital in the 21st Century. This is the book in which he argues that the reason for increasing inequality is that returns to capital exceed returns to the other factors of production. And so, wealth accumulates over time at the top. Economists debate his data and some ways of measuring this but that's not my concern here. The other one I wanna just mention to you as examples of these brilliant economists, there's no question about that, but the things that he says about politics are... Bereft if any serious attention as to how they might be enacted. So for example, in his book published in 2014, in the penultimate chapter of that book, he calls for aggressive tax on global capital. And in the beginning of that chapter he does say, well this might be a utopian idea. But we should still put it out there for thinking about where we want to get to and besides, it could inform more realistic proposals like a Europe-wide tax on capital. Now you might say, well in 2014, that wasn't, you know, it might not have seemed a naive thing to say, as naive a thing to say as it is today, but nonetheless, in 2018 he has pushed this idea further with some other intellectuals in Europe, publishing what they call our Manifesto for the Democratization of Europe, and you can find it and peruse it at your leisure on their website. They're calling for the creation of a new European assembly that would essentially function as a European-wide parliament that would have powers to raise revenue and manage budgets and it would be... The idea is that 80% of the legislators would be elected in the national elections so when you vote for your MP in the UK or Austria, they would also, not only would they go to the British or Austrian parliament, they would also participate in this new European assembly. 80% of them would come there. The others would come from the European parliament by proportional allocation. And they portray this as an alternative to the current EU structures which are so constrained by the treaty-based character of the European union which I lectured to you about some months ago. But what do they ignore in proposing this? Why is it that the Lisbon Treaty ended up with a treaty-based system? It was because the more ambitious proposals envisaged at Maastricht were roundly rejected in referendums in Europe. And so they had to actually cancel a number of the plan referendums because they saw the whole project of a never closer union heading for catastrophe. And so they retreated. The Lisbon Treaty is really a retreat back to the idea that while EU is essentially a system of intergovernmental treaties in its ultimate legitimacy. They also argue that the purpose of this new European parliament is to do things like reduce inequality within the member countries. But why, if the member countries in their own parliaments can't reduce inequality, why would anyone think that sending them to a European parliament would enable them to reduce inequality any more effectively? So, and if you go and read the website that they have there, which I urge you to do. It's certainly an imaginative and interesting proposal in an intellectual sense, you will find the thing I said you should always be careful of. Or should make you watch your wallet. The vast majority of it is written in a passive voice. The vast majority of it said about what should happen. What should occur. What needs to happen. But of course, we can talk until we're blue in the face about what needs to happen without having anything to say about how to make it happen. What will be the political forces that are gonna produce an outcome of this kind? And when you think back to the literature on the European Union, Tony Judt published his book in 2006, post-war, when he warned. He said, the problem with the European Union is that it's been an elite project from beginning until end and as soon as it gets into any trouble the popular resentment of European-wide politics is going to erupt and it's not gonna be pretty to see. He turned out to be right about that. Adam Tooze in Crash gives chapter and verse of the inability of the European governing structures to come up with policies that the populations in the constituent countries will live with. And so, the nationalists of the countries within the European Union is very powerful in the idea that it's going to go away anytime soon, I think is implausible. And I gave you a reason, if you think back to my lecture. Among the reasons is that the Europeans have contracted out their security to NATO over these past several decades whereas in the American case, what built a sense of national purpose when we transitioned from the Confederacy to the Constitution was the creation of a National Defense and the funding of Central National Defense. And then the idea that people would identify with the national project rather than, what we might call the federal project, rather than the states, that was a transition that was set in motion by the centralization of our power over national security in the American system at the time of the Constitution. There's never been an analog of that and so the idea of an ever closer union has largely been a fantasy of elites for which popular support has never been dealt and indeed is less likely to be forthcoming today than it has been since anytime since the financial crisis. So the idea that we can now create a new European parliament that, as they argue in cases of disagreement would actually trump the decisions of the European Union who which are constrained by the treaty-based caricature that strengthens national parliament I think is, it's an example of policy without politics being ultimately empty. What about politics without policy? Well we have examples of that. Just to give one here. (drum music) (crowd chanting) - [All] Occupy Wall Street all day, all week. Occupy Wall Street. - It's our duty as Americans to fight for our country and to keep it, you know, true to serving its people. And when it doesn't do that, it's immoral not to stand up and say something. - I'm here myself as a free individual to humanize the markets and to have true participatory democracy. Bottom up democracy. And to make Wall Street hear the sound of what democracy means. - What kind of power? - [All] People power. - Wall Street, it crashes, and you know, people start, people lose their jobs and things like that. We're very angry at Wall Street. It's a part of capitalism, American capitalism especially, that's why we're here today... At Wall Street. - There's no reason to not be peaceful. We just wanna get a point across. We're just trying to let people know what's going on and why we're here for it. - One in seven children in the United States suffer from hunger, at the same time we're giving billions and almost trillions to Wall Street just for bailouts. Something needs to change. We need an economy for the people and by the people. Not by the rich and for the rich. - I mean, and the government's doing the work for us, all they have to do is cut some more people's insurance, unemployment benefits and it won't be a bunch of 20-year old white college students out here, you know? - What would make today a success? What kinds of changes would you like to see as a result of everyone demonstrating today? - A success, it's already happening. They corral the bull. And that's pretty much a huge symbolic statement. - So there you have it. This is the beginning of the Occupy Wall Street Movement. It was a movement that was famously uninterested in articulating particular policies, in organizing itself behind any particular agenda, but rather its purpose was to express this moral outrage at the bailing out of the financial crisis and the inability of the Obama Administration to do much else. They were to some extent triggered by the Arab Spring Movement which had gotten going that year. But they really didn't advocate any particular policies. And indeed if you think about... If you think about the decision to occupy Wall Street, it was, you know, unlike, at least the Vietnam protesters went to Washington where something might be done about it. Whereas, here they went to New York and then spent the following Fall on many town and village greens around the country but because they lacked organization or leadership or resources, once the cold weather came, they soon faded away. Now, this doesn't mean that they were irrelevant to American politics. They articulated a moral narrative that was, that had a certain kind of coherence to it. You might call it a reactive moral narrative but I'm gonna have some more to say about the importance of moral narratives in effective distributed politics later this morning. But they had nothing else. They had nothing more than a moral narrative. And moral narratives on their own is not enough. Again, this is politics without policy. So my agenda today is to get us to start thinking about how, we can think about how there can be effective policies in light of what we do know about politics. And one phrase that falls to mind is in one of John Rawls's less famous writings, his Law of Peoples it comes from, I believe, the way he first used it, this is the idea of Realistic Utopianism. And so, the thing about being a realist, about politics, in the real politics sense of realist it is the danger of being a realist is it might prevent you from trying to do things that will change reality, right? That you'll be so constrained by one sense of the politic, you know, the politics is the art of the possible. You'll be so constrained by the notion of what's possible that you won't push for policies that might change what's possible, right? So that... I think we actually say in our book on the Death by a Thousand Cuts that real political creativity gets people to think that what they had previously regarded as impossible is in fact possible, right? That's, you know, you wanna use the buzz words. You wanna push the envelope. Think outside the box. You've heard them all before. And if you're very constrained in your thinking about what appears to be feasible in politics as we know it today, you will not push to change the boundaries of what is possible in politics today. And so you run a risk of being, if you like, captive of the inability to think imaginatively about ways to change what is possible. And so I like this phrase realistic utopianism in that, it conjures up the idea of trying to get to a better place but thinking about the steps from here to there. You've gotta think about how you're gonna join the dots. How you're gonna in fact get policies adopted that might change the politics on the ground, which I think is conspicuously missing in the proposal for European parliament that we just talked about. And so to make that case, what I'm gonna do today and then I'm gonna use this framework in our remaining lectures is talk about what I'm gonna call building blocks of effective distributive politics. And there are six of them. And I'm gonna spend a little bit of time on each of them. And I will start with coalitions. Very important in politics to think about coalitions. And the reason it's very important politics to think about coalitions goes back to our discussion of the differences between the median voter story and the majority rule divide the dollar game, right? So just to remind you, the median voter story created the expectation that because the median voter is always below the mean voter, there would be downwardly distribution of wealth because politicians in search of the median voter would advocate policies that the median voter would prefer. And that was the fear of 19th century liberals. It was the hope of Marx in his later years when he started talking about the parliamentary road to socialism and it's embodied in the median voter theorem. The puzzle was that it doesn't happen in any kind of systematic way. And we thought one reason for that might be that there's another dimension such as race that people care about more. And we talked about Nixon's southern strategy in connection with all of that. But then we talked about the majority rule divide. A dollar game where we saw even without the 2nd dimension any distribution of income can be upset by majority rule. All you need is three people and you're off to the races with the potential, right? So another way of putting it is that the median voter theorem assumes that all of these people are gonna vote together. But in fact, you might get a coalition between the rich and the poor against the middle class. So you might get all kinds of coalitions. And that means that no... No distribution of income and wealth is not potentially vulnerable to being upset by majority rule. And an example of this, we spent some time on early in the course, was the campaign to repeal the estate tax which is just about money. It was just one divisible good. And so, that's clearly not the case that a 2nd dimension or anything like that producing the instability. So, coalitions are really important. You have to think about what coalitions are going to support, what it is that you're trying to achieve, and you've gotta think about them, both inside the legislature and outside of the legislature. So if you're thinking about how to actually get a bill enacted in Congress, you have to not only think about, you know, who is gonna buy off who? What deals are gonna be made to get this bill to parliament or through Congress? You're gonna have to think about building support out there in the country for it. Because there are many, many things that politicians are planning to consider but in the American case, for example, the rule of thumb is if you're bill does not get into the top eight, which means it will be known as HR1-2-3-4-5-6-7 or eight, it's not going to actually make it to the floor. It's not going to be voted on. How do you get a bill up there? We saw, again, with the bill to repeal the estate tax that it took building support outside of the beltway. Not just inside of the beltway, to get many politicians to become cosponsors of that bill and see it is in their interest to support it. So you have to think about coalitions inside the legislature and outside the legislature. If you think about the abolition of the slave trade. Again, people like this begins in, it's a story that begins in Britain, people like William Wilberforce build a coalition in the British parliament of so called Dissenting Whigs, the Dissenters. And just by chance, this is when, this is before the hay day of strong parties in Britain. They still had multi-member districts so they had much weaker parties. And it so happened that, it behaved much more like a European system and it so happened that the Dissenting Whigs, for about five consecutive elections, in the mid, between the mid 1830s and the 1860s, the Whigs held the balance of power and were able, since the Dissenters held the balance of power and because they cared more about that than anything else, they made a condition of supporting governments. Even Whig governments. Even governments that supported other things that they wanted. They cared so much about this, that they were able to put it through the legislature. But they also built support for outside the legislature. They threw various anti-slavery societies. And indeed, after William Wilberforce left parliament, he spent the rest of his life pressing the course outside of the legislature. So, you really have to think about building support inside the legislature and outside. A second thing you have to do is not only think about enabling coalitions, but you have to think about the blocking coalitions. If you're gonna try and enact a policy, who's gonna try and stop it? And what are they gonna do? Because, you know, your coalition is not gonna prevail if this ineffective blocking coalition. Again, we saw, in the repeal of the estate tax, very clever use of enabling conditions, and very good strategic anticipation of how to head off potential blocking coalitions. And so we saw, for example, they got the congressional black caucus to support repeal of the estate tax on the grounds that it would protect small business and particularly in African-American communities, and they got people like Bob Johnson to support it. Black Entertainment Television on the grounds that the first generations of wealthy African-Americans shouldn't have their, save their accumulations tax away. They got the support of gay groups on the grounds that estate tax discriminates against gay people. At that time, there was no provision for gay marriage. Of course, it was not the estate tax that discriminated against gay people, it was the law of marriage that discriminated against gay people, but it didn't matter. They got Frank Blethen, a liberal owner of the Seattle Times, to get family owned newspapers to support it, because they were worried about the Gannett and other big multi-national news organizations gobbling them up. They got the pharmacists to support it so that they could get it through the senate. they got small business to support it so they could get it through the house. Very clever use of all of these groups. And then when they thought the potential blocking coalitions, the life insurance industry, estate planning, life insurance, the non-profits benefit from the tax deductions that people make in order to protect themselves from estate taxes. And they thought about unions who generally fight. And they came up with a separate strategy for managing every one of these potential blockers. And so, they managed to thread the needle in that way. If you think about, if you think about the abolition of the slave trade. They very cleverly, in order to head off potential blocking coalition, avoided the question of slavery itself, and began first with the abolition of the slave trade in the North Atlantic. And so here, if you think about the different coalitional possibilities, they change greatly. Because if you're talking about abolishing the slave trade, it did not prevent the cotton manufacturers in Manchester from getting cheap cotton that came from the American South. It did not prevent the Americans supporting after 1808 when it was possible to abolish the slave trade. In fact, part of the reason you couldn't have enforced the abolition of the slave trade without the Americans, and the reasons the Americans supported it are not very morally appealing. There was a racist sense that they didn't want a great increase in the African-American population of the US And in the American South, it was essentially protection. There was a domestic slave trade that they were interested in protecting at the time. So, it was possible to put together a coalition of strange bedfellows, if you like, to get rid of the North Atlantic slave trade. And they left the battling of slavery itself to another day. So, again. It was possible to put together the coalition and head off potential blocking coalitions by thinking in those kinds of strategic ways. A third feature of thinking about coalitions is that, hugely important to pay attention to business interests. This has always been important. When we talked about, we talked about the new deal. While it is true, and I even showed you the I welcome the hatred video on the second new deal that Roosevelt ran on when he was running for reelection. As my colleague, Peter Swenson and others have documented that most significant expansions in welfare states have come about with at least tacit and sometimes active support from business interests. And if you think that was true then, in the world we're living in today that I described to you over the lectures in this course, when business interests are in a position of unparalleled power politically, if you're going to try to enact policies that affect business interests, you better think about how they're going to respond. Now, that doesn't mean you should necessarily kowtow to business interests. But you need to think about, will they be an enabling coalition? Will they be a blocking coalition? Is it possible to divide business interests so that some business interests are working across purposes with other business interests? But in this point of the 21st century, the idea that in any capitalist democracy, anybody is going to enact policies that affect business interests in a serious way without contending with business interests is a fantasy. It's not going to be possible. And so you have to think about how business interests are going to be a factor in whatever it is that you're trying to do or trying to forestall. So, for all of these reasons, some analytical, some empirical, you have to think about coalitions. The median voter is not gonna be enough. Secondly, moral narratives. As I said, when we talked about occupy Wall street, moral narratives matter in politics. Why do they matter? One reason is implicit in what we've already talked about. The very fact that purely interest-based coalitions are inherently unstable. We know that from the majority rule divided all again means that something else has gotta hold them together. If all you're gonna do is say... We'll make a deal where you get the biggest possible slice, somebody else will come along and offer you a bigger slice. And so, you have to think about, well, what is it that will get people to resist when that somebody comes along and tries to recruit you to a blocking coalition? The story I told you when we did the estate tax class was that when the democrats finally figured out what was going on, they came and tried to hive off the farmers. They're said, we'll give an immediate, remember this? We'll give an immediate exemption to farmers. And they had a huge fight during the subsequent meetings of the repeal coalition because the farmers were actually being offered a better deal by the democrats, the permanent repeal, than the bill was gonna give them, which was gonna be this weird 10 year repeal. But they said things when you interviewed them. They sounded like abortion activists. They said, how could I look myself in the mirror? How could I tell my child that I had been bought off? This is a moral crusade, it's an unjust tax, and so on. If you want another example, many years ago I had a call from an academic colleague, a philosopher who said they were starting a new party, it was gonna be called the New Party. (students laughing) And would I support it? And I said, well, okay, what is the New Party all about? And they said, we want it to be an equivalent counterbalance on the left to what the moral majority is on the right. This is the era of the moral majority. Which, moral majority is sort of predecessor, if you like, to the tea party. And I said, well, good luck, goodbye, I hope you do well with it. And of course, there's a reason why nobody here has heard of the New Party. because being a counterbalance to the moral majority is not a compelling moral narrative. The people in the moral majority believe in something. That's why they're there. They're willing to internalize costs. They're willing to stay up. They're willing to go to meetings. They're willing to do something because they believe in the cause. Right? And so, the idea that you can have, that you can have people join a coalition who don't actually believe in the cause isn't gonna be enough when people come along trying to split your coalition. When George W Bush in 2005 sought to privatize social security, I'll have more to say about that in a little while, but one of the things he ran into was, he couldn't split the coalition. They tried by saying the new system will not apply to people... I think they tried 55, and then they tried 50, under the age of 55, under the age of 50. And they found that public opinion among the elderly didn't move a millimeter. So, what did this reflect? It reflected the fact that elderly people didn't just care about social security for themselves, but for their children and grandchildren. And so they weren't sufficiently self-interested to be peeled off in that way. They were committed to the idea that this was gonna be important to be there for future generations. Again, they couldn't be bought off just on the basis of their self-interest. So, an effective movement is almost certainly going to require moral narratives that cause people to resist being picked off by some blocking coalition. And gonna motivate them to do the work. To show up at meetings, to do the leafletting, to do the donkey work of moving a movement forward. Notice it need not be the same moral narrative for everybody. If you look again, go back to the case of the abolition of the slave trade. I keep coming back to the slave trade, it's a very interesting example because an economist, if there had been neoclassical economists around at the beginning of the 19th century, they would have said it'd be impossible to abolish the slave trade. The race to the bottom, all the usual reasons would have been given. So, I think it's a great example of realistic utopianism, that it actually got done. But if you look at the moral narratives, actually people didn't share the same moral narratives. If you look at, for some it was religious moral narrative. For people like David Hume, it was a different moral narrative. There were moral narratives related to driving down wages... In some countries. So, it doesn't necessarily have to be the same moral narrative. But it has to be a moral narrative that will motivate people and prevent them from being picked off. So, a third feature of effective distributive politics, and this is hugely important. In some ways, very difficult sometimes in practice to think about. And this is to pursue proximate goals. So, one way of thinking a proximate goal is, it's a waystation to a better future that will start you on a road to somewhere where you're gonna end up later at a better one. So, getting rid of the slave trade is a proximate goal on the way to getting rid of slavery. Of course, it didn't help in the US they had a civil war about it. But it did help in the British Empire, they abolished slavery in almost the whole of the empire by 1833, including by paying 20 million pounds in compensation in various parts of the empire. And then they spent the next six decades stamping it out, including getting involved in an undeclared war in South America. But pursue proximate goals, move, to use an American analogy, move the ball down the field rather than simply trying to throw hail Mary passes every time. So, why is this important? One is that it's important for sustaining coalitions and motivating support. So... If you have a proximate goal, that's something you can organize your coalition around. You can motivate people to fight for and to achieve. So, if you think about, say, again, repeal of a particular tax. We've gotta repeal the tax, you know, we... We build the coalition, we have the meetings, we push the bill. And then you finally achieve it, you can say, we did it. You can get people to focus on, that can motivate them, and it gets them to focus on the result. Or if you think about the civil rights act, which we spent some time on. The features of the civil rights act that were most effectively adopted involved well defined proximate goals. Like getting rid of restrictive covenants. The more nebulous goals, such as reducing racial inequality, or integrating neighborhoods much harder to achieve. So, proximate goals are important for moving, you know, moving things along in the direction that you want to go. And for making sure that the best does not become the enemy of the good. A couple of weeks ago, we had Gidon Bromberg from EchoPeace here to give the part of the lecture that I had not had time to give when I talked about when we did the Israel-Palestine case. And I think I saw some of you in his lecture. Which is on a fascinating topic of using economic interdependencies around climate to work to move the Israel-Palestine conflict in the direction of a resolution. And in particular, the fact that the Jordan River is drying up, and the two million people in Gaza lack sufficient energy to process their own sewage. Partly because it's being denied by Israel, the energy, but partly just an overwhelming demographic pressure has started causing a situation weighing increases amounts of sewage from Gaza being pumped into the Mediterranean, drifting up an forcing desalination plants in Israel to shut down. It turns out that Jordan has the best available geographical situation and technology at the moment for solar energy. And so EchoPeace has come up with an ingenious plan prepared by solar energy that can produce power that will power the sewage plants in Gaza, which will in turn save Israel money. And they've lined the whole thing up so that the incentives work for everyone. And all of the governments know this. And they've even created, they've raised capital for it. So, there's private sector interest in doing this. And in several rounds of peace negotiations, solving this water problem in this way, which would actually integrate the Palestinian, Israelian, Jordanian economies much more and produce co-dependencies has come up. but it's treated as one of the final status issues. And so the final status issues are these difficult questions to be resolved later. Something that was never allowed to derail the South African negotiations. And so Gidon made the case that it would be much better to solve the water issue now. And he pointed out that every Middle East negotiator that had tried to get a settlement from Clinton to carry, and others in between, subsequently had said, we should have dealt with the water issue. So, this is an example, again, that we've videotaped that lecture. But I strongly urge you to go and see it. An example where it would be much better not to let the best be the enemy of the good, but use the water issue and resolution of the water issue to take it off the table and move the three countries into a situation where they would be more likely to cooperate on other matters in the future. But they're hard cases. They're really difficult cases in thinking about proximate goals. And the hard cases are where if you pursue a proximate goal, it might make it, instead of being a waystation to the future, it might actually make the future more difficult to achieve. So, think about Medicare and social security. Social security, we discussed earlier, was put in place in 1935, and in order to get it through congress, FDR had to exclude the Mexican agricultural workers, because not to do so would have broken up his coalition of Southern democrats who controlled the committees. And so, that was done later. It was expanded to cover those workers later. That's, you know, it's not, it wasn't great for the people at the time, but you could say, well, you know, you take what you can get, and you expand it later. Medicare's a harder case, because Medicare, put in place in 1965, part of the great society by Lindon Johnson creates a world we live in today where it's much more difficult to create universal healthcare because the one group that you would want in your coalition has no reason to be there. Namely the elderly. The elderly, they vote, they're politically active, but they already have universal healthcare. And indeed, if you were gonna go to any of the current plans being proposed, which are being underpriced from Medicare for all, there would be a reasonable fear that Medicare benefits were going to be diluted for the people who already have them. So, that becomes a much harder judgment. It doesn't mean that they shouldn't have gone for Medicare for 65 year olds in 1965, it's just a harder call. It's a harder case, right? Because you're creating a potential to split the coalition that you're gonna need to expand it later. because Medicare funding is seen as more of a zero sum system of funding than the way social security is funded about which I will say more in a minute. Sometimes it should be said that this is shooting in the dark. When Bush passed his 2001 tax cut that included the estate tax repeal, it was done through budget reconciliation. And it was gonna go away 10 years later. Actually, nobody had any idea what that would mean. And you couldn't use the inertia of the system, because the inertia was gonna have it go away. It was gonna be repealed back to the 2001 law if nothing was done. So, they were just gambling. And as it turned out, the Obama Administration made almost all of it permanent as we saw. So, proximate goals, really important, but you need to always recognize that pursuing a proximate goal may not be good in the long run. And so, in that sense, as I said, I think it'd be hard to make the case that if it was Medicare for 65 year olds or nothing, it would have been better to have taken nothing, but it's a hard political call. Entrenching gains. Very important to entrench gains for the long run, because otherwise, they can be undone. The classic example of this, of course, is reconstruction. The gains that were brought about as a result of the civil war, yes we got the abolition of slavery, but then we got the 14th and 15th amendments, and the extension of the franchise, all of which was later undone, because the North was not sufficiently committed to entrenching those changers. If you want an example of what I'm calling here, the gold standard of entrenchment, it's the way that social security was funded. So, what FDR did was quite ingenious. They put in the system where, you know, it's a savings plan. The idea is, you work for your lifetime, you pay in, and then when you retire, you get it out. But, that's not the way it's financed. It's not the way it's financed because they wanted to start paying out immediately. And of course, the first generation of workers in 1935 hadn't paid in. It started paying out in 1937. Those people had not paid in. And so, in reality, the current generation of workers funds the current generation of retirees. And when people say there's an IOU in the social security trust fund, that's what they mean. Right? Now, FDR did that for two reasons. One was to build the coalition. He wanted everybody to support it at the time. So, he wanted existing workers who hadn't paid in to support it. Of course, if you're gonna get it, you know, people about to retire are gonna get it, they're gonna support the coalition. But he also wanted to make it impossible or very difficult to undo. Right? He said, I'm gonna fund this, he said, I'm gonna set this up in a way so that no damn fool politician can ever undo it. And so, in 2005 when George W Bush tried to undo it, he ran into the reality that, you know, what they wanted to do was create private accounts, but instead of the money going to the government, this was two years before the financial crisis. It would have been pretty ugly politics if they'd gotten away with it, but instead of the money going into social security, it would be just like an IRA or like we all have. It would be going to an account that you wouldn't privately manage account. But the problem was, to do this, they would have had to fund both systems for an entire cohort. Right? Because there's no money in the trust fund. Right? And so, this in effect is what made the financing of social security, apart from its popularity, made it bulletproof. So, very clever system of entrenching the program in a way that would make it extremely costly. And as it turned out, eventually the Bush people gave up. But the most important thing, and this should be evident to you from my earlier lecture on bundling is that the way in which you build protection for a policy is to build support for it, and get one of the major parties to bundle it into their major party platform. So, what protected the Civil Rights Act is it's bundled into the democratic platform. What failed to protect reconstruction is, it was not bundled into the things that Northern republicans cared about. And of course, if you can get it bundled into the platforms of both parties, such as social security effectively is, even better. So, entrench proximate gains for the long-term. Resources matter. Resources matter in two senses. Resources for funding the effort. If you look at something like the estate tax repeal coalition, wealthy... Wealthy individuals put up money to fund the organizing effort, to get that done. And we give chapter and verse of that in our book. They paid for the meetings, they paid for the resources to lobby, and so on. And indeed, the whole rise of neoliberalism was funded not through campaign contributions, but spending on think tanks and other areas. So, funding efforts. And then funding the policy. So, what was so brilliant about social security was, it was immediately funded by existing workers. So, again, you have to think about the resources, whereas again, go back to Occupy Wall Street that we began with. One of the reasons the mayors of most cities just ignored them was that they knew Winter would come, they would have no money, they would have no resources, and they would pack up and go back to wherever they came from. So, again, they hadn't any resources. And then finally, leadership matters. Leadership matters partly because all coalitions are unstable. You need people to articulate. You need people to articulate the moral narratives to do the work to keep people together. Again, if you think about the Civil Rights Act. The Kennedy people, they sent civil rights legislation to capital hill in 1963. They had almost no experience with Congress, and it soon got bottled up in the committees. And it wasn't until Lyndon Johnson who had been the master of the Senate, this is all told chapter and verse and volume four of Caro's four volume work on Lyndon Johnson. But Johnson understood, he's widely recognized as one of the most effective legislative politicians of the 20th century. And it was he who figured out how to do the leadership on the hill to get that enacted. So, you need leadership in the legislature, you need grass roots leadership. You need to mobilize support for policies, or they're they not going to get enacted. And of course, you need what's sometimes referred to by consultants as grass tops leadership. You need those Frank Blethens, and you need those Bob Johnsons to push for support if you like to articulate to act as kind of brokers between grass roots support and support in the legislature. But just to finish, effective politics is not enough. Those six elements are not enough. If you go back through, the story I told you about home ownership. There was a great coalition to do it that included politicians from both parties, it included the lobbyists from Fannie and Freddie, it included people who were going to make money out of securitization. They had not one, but two moral narratives. The American dream of home ownership on the one hand, and then racial inclusion on the other. So, they had two moral narratives. They had great proximate goals. They could articulate how many more homes should be owned by minority homeowners by benchmark dates. They had proximate goal after proximate goal. They entrenched the gains indeed. One of the tragedies of it was that it was impossible to unwind when it would have been in the interests of the minority homeowners to unwind it in the ways that Geanokopolis was proposing. So, it became a kind of pathologically entrenched. The resources were brilliantly generated by the securitization, which created a study cycle of liquidity to produce more of the activity until the music stopped, to use Blinder's line. And then there was leadership. There was leadership from both political parties. There was leadership in the country. And there was plenty of support from business leaders who saw that there was money to be made. So, building blocks of effective distributive politics are important, but they're not a substitute for good policy, and you can indeed have bad policy that meets all six of these criteria. And that would be the poster child for that. So, next we're gonna look at good policies that can be enacted by paying attention to these six building blocks. And that's what we'll start with on Thursday. (soft tonal music) |
Power_and_Politics_in_Todays_World | Lecture_12_Business_and_Democratic_Reform_A_Case_Study_of_South_Africa.txt | - So today we're gonna talk about what I advertised on last Thursday as the backstory behind the South African transition. Or at least part of the backstory that involved business. And this is the first of several classes in the second half of the course where we gonna turn the lights up on business. We talked a lot about the declining fortunes of labor in this course, and how that's played out in politics. But we haven't spent much time on business, just generally treating as a black box. Obviously the power, relative power of business has increased in the capitalist world since the collapse of communism. And that's brought new dynamics. It also, as a normative matter tends when capitalism is the only game in town, it tends to make the old Marxist critiques of capitalism a little beside the point. The more interesting questions about business and capital are what are the conditions under which they play a progressive role in politics? What are the conditions under which they play a predatory role in politics? And how might they be shifted from a less predatory role, of the sort we saw in Russia in particular? How might business be induced to play more constructive roles in politics? Even in very difficult politics. So that's one theme that's gonna come back several times in the latter part of the course. But let's begin by going back to Johannesburg in the 1990s. - [Tertius] All of a sudden, I saw the security people running out. And I said, "what's up?" And I said, "they've broken through, and they're coming." (chanting) And my own security man simply grabbed hold of me and ran. (chanting) - [Cyril] Pandemonium just broke out in the entire building. Many people on the ANC side, from the government side, all huddled into the government offices. We thought that the right-wingers would never dare this much into government offices, and start accosting people, or assaulting them. And in fact, we even feared that we would be shot at. (yelling) - [Narrator] The AWB overwhelmed the police who retreated before them. (yelling) - [Mac] Signs of glass, and everything chattering, and everything. And reports coming through that they are looking for us. Particularly ------- - [Narrator] Terrified delegates barricaded themselves in their offices. (chanting) In the conference chamber, the AWB chanted and dubbed slogans demanding a separate white state. (yelling) Soon after the invasion, the negotiators agreed that an election should be held the following April. - I have the demands up here. - [Narrator] With the ANC growing in confidence and looking certain to win the election, the national party had retreated bit by bit on nearly all its power sharing demands. As the conference wound up the ANC triumphant, Ramaphosa celebrated his 41st birthday . - I was able to see Roelf Meyer for the first time dancing on the floor where we were having this party. He was quite happy, he was quite jovial. And I kept wondering whether I would have been as happy as he was if I was in his position, having finally given in in the way that they had. - [Narrator] Key national party negotiators were satisfied they'd struck a workable deal with an adversary they had grown to trust. Now the election could proceed. (cheering) Terre'Blanche threw all his weight into stopping the election. He summoned a rally at the Pretoria Monument to the Africana pioneers. He said if Mandela wanted war, he would get war. (yelling) (drumming) Terre'Blanche then persuaded the folks to boycott and oppose election. - [Terra] (in foreign language). (cheering) (in foreign language) (cheering) (in foreign language) - Pretty frightening stuff you might think. And the reason I put it up there is just to emphasize that when Mandela and de Klerk negotiated their agreement, that was not the end of the matter, it was to quote trachelism; it was the end of the beginning. And there was a lot that had to happen for the transition to actually to be consummated. And it was enormously fragile. And I am just gonna walk you though a bit of a timeline to give you a sense of how fragile it was. We saw last Thursday that on the 2nd of February, 1990, the government announced it was unilaterally freeing all political prisoners, and banning all organizations, and planning to negotiate with the ANC. Three months later, the ANC unilaterally suspended its armed struggle, and began serious negotiations. That led in December 1991, and then May of 1992, to two rounds of round-table negotiations, the so-called CODESA; Conscience on a Democratic South African negotiations, where all of the stakeholders, there were 19 stakeholders initially, all the political parties, all social movements had a seat at the table as a way of trying to negotiate the transition to democracy. As I said to you in the last lecture, democratic path to democracy is difficult because democratic political orders are public good, and it's very difficult to do that to provide them democratically. And indeed, there were people and groups that didn't want to see this transition; the far-right we've talked about. And we'll see it in other groups were not so interested in seeing this transition to a unified democratic South Africa. So they became spoilers and they perpetually interrupted the negotiations, stone-walled the settlement. And in May of 1992, they were sort of reconstituted having been floundering since previous year. But a month later, there was a massacre at a place called Boipatong, which where about 45 people were shot dead apparently, allegedly, by Zulu supporters of the Inkatha Freedom Party, an ethnic separatist party that I'll tell you more about shortly. But it was widely believed that the South African security police were behind the seeds involved in the Boipatong Massacre. The result was CODESA collapsed, and the violence escalated dramatically. So dramatically, that the ANC and the government started having secret talks. But they weren't really secret; everybody knew they were going on. I was in South Africa that year and everyday on the evening news, there would be reports of what had gone on in the secret talks between the ANC and the government. And eventually, they signed a deal; an elite pact, if you like, in September that was called a Record of Understanding about how things would move forward. This then led them to create something called a Multiparty Negotiating Forum. And in the Multiparty Negotiating Forum, basically the ANC and government told the other players what was going to happen to give them time to accommodate themselves and adjust. But they did not really treat it anymore as multiparty negotiation as far as the government and the ANC was concerned. It was a done deal, and people had to get with the program so to speak, or be left behind. Later in that same month, a young, charismatic ANC member, very much on the radical flank of the ANC, was assassinated; Chris Hani. Very extremely charismatic, also in the communist party. And the result of this, he was assassinated by a right, I think it was Polish extraction white South African. And the whole country seemed to be on the verge of exploding. The violence escalated hugely. And the talks were in danger again of falling apart. And I put up on the slide there that the election date was set. And the reason for that was that Mandela and Ramaphosa came to declare conventionally and said that the only way they could temp down the violence, was if it was a non-negotiable agreed that the election would take place within a year of Chris Hani's death. So that the election had to be held in April of 1994 became a hard deadline, if you like, a non-negotiable item, and that turned out to be important later. Realizing that they had been handed a fait accompli, Buthelezi who was the head of this Inkatha Freedom Party, walked out of the talks. And essentially the Multiparty Negotiating Forum collapsed at that point. And true to the divided dollar story we've been talking about earlier in the course, even though he was a leader of an ethnic Zulu party, he joined forces with the right-wing Africans Whites. And also the heads of the traditional homelands, so the Bantustans, who saw themselves as losing their power in the new South Africa, and they announced that they were gonna be boycotting the elections. And that was important because most people who know a little bit about the South African conflict, know the story behind this pie chart. That as you can see, the Black African majority of the country was 79%, the Whites were fewer than 10%, and the rest were mixed-race, or Indians and Asian groups. What is less well-known outside of the South African context, is the ethnic breakdown of the non-White population. And there you can see that Zulu were a plurality of a 28%, but the next race group, almost 20% were the Xhosa. And why this matters is that the Xhosa were more or less a 100% committed to the ANC. The ANC probably won, not 100%, but over 90% of their vote. But the Zulu were deeply divided nationally between the ANC and Inkatha, but in the Eastern part of the country, which was then called Natal, today is called KwaZulu-Natal, they controlled the majority of the Zulu vote, the Inkatha Freedom Party. And so from Buthelezi's point of view, a much better outcome would've been a partition at the end of the day of some sort, where Inkatha could be in a majority, and so actually control what is today KwaZulu-Natal. So he didn't really want a transition to a country in which all the polls told them that they ANC would win a two-thirds overall national majority. And so, Buthelezi here was not just, even though we talk about the deal between the ANC and the government, Buthelezi was the potential spoiler of this settlement. And why did this matter? It mattered because the civil war in Natal would have made the transition impossible. The levels of violence by 1992 in KwaZulu-Natal, had already reached civil war proportions. And the notion that the country could really have a successful election, which was boycotted by Inkatha, was not credible. And so, it was a huge problem going forward. So just going back to the timeline. In June 1993, one of the far right-wing groups torpedoed the negotiations yet again, similar to the clip I showed you, but a different meeting in Kempton Park, they just drove armed vehicles through the plate glass windows and constantly were trying to upset and harass the negotiators. All of the time this election that had to occur in April of 1994, lest the deal between the government and the ANC come unstuck, was approaching. And indeed, a month before the election, in March, it all came to a head in one of the homelands. The homelands had been sometimes called Bantustans, as I said, these were the nominally independent states under apartheid migrant labor system. Violence exploded when the leader of one of these said he was not participating in the election. And some of the right-wing Africanas went ostensibly to support him in the hope that this would precipitate a military coup; that the military would get behind, would get behind the uprising in Bophuthatswana, and there would be a takeover, and the impending election would be stamped out. And so, there were these people went there in power model uniforms, lot of people were killed, including four white Africanas were killed and it was filmed on national television and shown over and over again. And it really seemed, this now we're a month before the election, it seemed like the whole world was going up in smoke. If you think about April 1994, those of you who are old enough to remember, this is when the Rewandan genocide was in full swing to the North. These were the kinds of pictures that South Africans were seeing on television and in the newspapers every night. And so the idea that you could have some horrific bloodbath, it wasn't an academic thought, you know, we're seeing it every night on the evening news. And just literally weeks before the election; three weeks before the election. This is what was happening in Johannesburg. - [Narrator] Johannesburg, the commercial heart of South Africa paralyzed by gunfire, and fear. (explosions) Chaos erupted on the streets as thousands of Zulu supporter of the Inkatha Freedom Party protested against the election. The rattle of AK-47s echoed around the banks of the business district. (gun shots firing) As security forces tried to keep control. (explosions) Tear gas and stun grenades filled the normally quite squares. There were running gun battles as police tried to force back gun-wielding demonstrators using rubber bullets and buckshot. At least eight people were killed by African National Congress guards when the Inkatha crowds reportedly tried to attack the ANC headquarters. (explosions) Bodies littered the streets as mayhem brought central Johannesburg to a standstill. - So it didn't look like we were three weeks from a democratic national election. And at that point, Henry Kissinger, and Lord Carrington of Briton, were brought in to try and mediate a solution. They spent all of two and a half days there, concluded the situation was hopeless, packed up, and left. And so then again, the newspapers were just full of poor taint of failures. This is an article which you can peruse at your leisure. But basically detailing the way in which Zulu militias were preparing to escalate the civil war. And then, miraculously, we turn on the evening news ad this is what we hear. - In South Africa, a breakthrough may be imminent to stem recent political violence. Political leaders appear to have found a way to end the standoff over the country's first all-race elections. Another meeting is set for Tuesday to finalize the deal. CNN's Pete Arnet reports. - [Pete] Following six hours of talk in Pretoria aimed at ending a major political standoff in South Africa, the official line was guidedly optimistic. - Our talks today are going very well. - [Pete] So well, in fact, that sources tell CNN that details of a plan to bring the Zulus into the election had been okayed by negotiators for the government and for the Zulu chief. And a document has been drawn up that must now only be approved by ANC leader Nelson Mandela. - And Mandela did approve the document and the election went forward, and Ikatha did indeed not boycott the election, and it occurred as planned on the 27th of April, 1994. The ANC won almost more than 62.5% of the vote. The National Party came in second, and Inkatha came in with 10.5% of the vote, and enough votes to get control of the regional legislature in the new KwaZulu-Natal, as well as three seats in the cabinet. And so, that's the timeline up through the election. Today's agenda is to talk about business and the back story. You could see just from what I went through, there were many points at which this might have fallen apart. And business had a great deal to do with why it didn't. And in the course of considering the role of business in this transition, we're going to revisit some of our earlier, briefer, glancing visits with modernization theory, and look at what we should think about businesses' attitudes toward authoritarian and democratic systems. And then we'll also, part of our theme is to think about the role not only that business played during this crisis, but also before and after to get a more panoramic historical sense. So, just to be clear as to what I am saying, and what I am not saying. I am not arguing, and it wouldn't be plausible to argue that business caused the transition from apartheid in South Africa. Apartheid was caused by many things, not least the negotiations we talked about last time between leaders, and not least the activities of the liberation movement. Business, however, did have a great deal to do with the fact that the transition was peaceful, substantially peaceful, and that it ended up in a democracy. That is to say it is clear that apartheid's days were numbered. The status quo as it existed in South Africa in the 1980s was not going to survive. But that didn't mean you were gonna end up with a democracy. There could've been a military coup, there could've been an ongoing civil war, there could've been multiple other outcomes. There could've just been a massive crack down of an apartheid authoritarian state. But so the role that business played was in facilitating a peaceful transition to a democracy, rather than some other more bloody denouement. And I am gonna focus on four areas where business had an impact. The first is in getting negotiations started in the first place. By the time Mandela in the late 1980s was secretly talking to de Klerk in messages, various things had been actually been going on for some time that made it propitious for those talks to begin. And I'll say something about that, then I am going to talk about the role business played in actually brokering the deals between the government and the ANC initially, and then between the two of them and Inkartha toward the end. I am going to talk about the role business played in managing potential spoilers. When I put that stylized diagram up of transition negotiations last time, I mentioned that the people on the flanks of the government and the opposition, potential spoilers, and had to be convinced, co-opted, or marginalized if the transition is not to be scuttled. And business played a role in that. And I also mentioned that it's very important in the conversations about from above and from below to build support for the new dispensation, which does not have a constituency in the old order, because it's a deeply polarized situation. So those will be the four areas that we'll focus on. Now, let's go back to the 1970s, this is the, you know, national party had come to power in 1948, it had spend much of the 1960s creating the apartheid order with compulsory segregation, forcing force removals of Africans from the cities into these homelands, where they then worked as migrant laborers. And by the 1970s, it was firmly established that the government had been in power fore more than a decade, there was no meaningful white opposition party, and they seemed to be entrenched. In a setting like that, why might business decide that it should start talking to the leadership of the liberation movement? Liberation movement was illegal, it was underground, after 1963 many of them including Mandela were imprisoned, and Jacob Zuma. Others were like Mbeki, were in exile in London, or elsewhere in Eastern Europe. Why would business leaders, now I am talking about big business, white big business. Why would they wanna start talking to the leadership of the liberation movement? Yeah. - [Student] Thinking about, like, business. - Speak up. (tapping on microphone) - [Student] Hi. No? - Or just yell. (laughter) - [Student] From a business perspective, 80% of the population, if you empower them, 80% of the population have to be democracy. - Okay, so business might have a bigger market if they could do better business. And it was certainly a view that apartheid was produced inefficient labor markets, and so on. We'll come back to that. But we should notice this is quite unusual. Typically in authoritarian systems, business does not get involved in politics. If we look at China, business, when we talked about the rule of law in China and so on, business might go to court over business issues, but typically they keep their heads low in politics in authoritarian systems. You know, they may be interested in the market, but that doesn't mean you have to have a democracy to sell in the South African market. And indeed, Carles Boix, who I put up on an earlier slide where we were talking about modernization theory. His argument about, his modern version of modernization theory, if you go back and look at that slide, you'll see I had him on the slide which said, "modernization comes as a result of pressure from below." And so this might seem inconsistent with what I am about to say now, but it's not, because many people have made this argument, I am just giving him as an illustration, is the thought was that in an authoritarian setting business is gonna resist democratization until they think that they're pretty sure they can leave. It's a Hershman story. So, if I think I can't leave, and I give every body the vote, it's the, you know, the median voter problem. They're gonna tax me, they're gonna expropriate my assets once they have the votes. So I have a strong incentive to resist democratization. But once I can, if I can leave with my capital, and with my assets, then the threat of exit, my threat of exit makes it less likely that they will be willing to risk the economy's future by taxing me. So this variant of modernization, as I said, it assumes the pressure to democratize is coming from below, but the argument is businesses can align up with authoritarian systems so long as there's what economists refer to as high asset specificity . That is, you can't take it with you. You know, if it's a world revolving around human capital, I can just walk out of the door with my human capital in my cranium, and, you know, set up shop elsewhere. But if what I own is a goldmine, I can't take the goldmine with me. Right? I can't take the diamond mines with me. Right? And so this is why it seems counter-intuitive, at least form the standpoint of those traditional theories, it seems counter-intuitive that business would actually favor an opening to the liberation movement, because very high asset specificity, there were big business was invested in big mining, and the diamond mining, and so on. They were also invested elsewhere in the economy, but they didn't have much credible capacity to threaten exit. And so that's initially a passall and we'll come back to it. But it means that there must have been sticks and carrots. There must have been sticks to make business think that they needed to start talking, and there must have also been carrots to make it look appealing. And I wanna start with the sticks. Sticks matter because things, if people don't have to change, they're probably not gonna change, for the most part. Particularly, I said most politicians are risk-averse, and most business people are also risk-averse in that sense, they might wanna take business risks, but they're not gonna get involved in trying to shape the political future unless they start to feel that the status quo is as what an economist would call a wasting asset. That things are changing in ways that are not likely to be good for them. So, what could the sticks have been? Yeah. - [Student 2] To avoid collapse. - Sanctions. Okay, so sanctions, the difficulty with sanctions, and this was the era of boycotts, disinvestment. David Swensen who's managed the Yale endowment since 1985 will tell you a story if you have lunch with him. That he came to Yale from Wall Street in 1985, and he went to the investment office, and he found it occupied by anti-apartheid student demonstrators, and he said, "right then I knew I left "Wall Street." (laugher) There were boycotts, past over Reagan's veto, the truth about sanctions is that they don't work though. Industrial sanctions don't work, and in South Africa they didn't work, because, you know, Kodak left, but AGFA came, Ford left but Toyota came. And indeed, the sanctions led to stimulation of the indigenous capital markets in the growth of the South African economy. So, industrial sanctions didn't really make any difference. I'll come back to financial sanctions in a bit. But a principle stick here was that South Africa had illegal black trade unions. They had outlawed black trade unions under apartheid. And this hadn't mattered that much to business, because there had been labor quiescence through the 1970s, as you can see from that slide. There'd been almost no industrial action. In the early 1970s that starts to change, and although it's child's play compared to what's coming later, as you can see, if you look at what happened in the 1980s, as compared with what they were used to, there were these big outbreaks of wildcat strikes and the mine owners didn't know who was running them, and they wanted to be able to negotiate with them. They wanted to be able to negotiate industrial peace, because they' were trying to run the mines. So in the first instance, they wanted unions to be legalized so that they could have basically a corporatist arranged where the management and labor could negotiate industrial deals, and they could be enforced by a government. The government was strongly apposed to this for the obvious reason that with all political opposition illegal, if you legalized unions, they would become magnets for political activism. The mine owners said, "yes, we know that, "but we should do this, we have to do this. "We can't run the mines." And there was something of a standoff for several years in the mid 1970s, but eventually, the business leaders beat the government into submission. And something called the Wiehahn Commission was appointed that was nominally gonna look at it, but it was run by a man, Wiehahn, who was an academic, who was known to be a supporter of legal trade unions for blacks. Anglo-American, the biggest multi-national in South Africa, and one of the biggest in the world at that time, actually had someone on the Wiehahn Commission. It recommended the legalization of trade unions, and they became, indeed, magnets for political activism, and many people criticized Anglo-America and big business for having supported this. But they did it anyway. And from the point of view of the economy, it worked for them. They did have to deal with very dramatic strikes later, but they did create first the National Mineworkers Unions, then something called COSATU, the Conference of South African Trade Unions; one of the most powerful union movements in the world was finally created, and certainly one of the most powerful union movements in the world today. So, that turned out to be important. It also led to the development of relationships, personal relationships between business leaders, and people who were widely represented in the libration movement; something that had not previously existed. I said that industrial sanctions didn't work. The same was not true of financial sanctions. In the mid 1980s, amidst all the pressure, eventually Chase and Citi Bank refused to roll over South Africa's debt, and they couldn't pay it, so it was a massive run on the banks. And huge capital outflows, and that created a crisis in the financial sector that was not resolved. What the leaders, and you can see this if you go on the case and look at the interview with Sidney Frankel, now deceased, they responded to this by running junkyards, and bringing business people to the country to try and convince them that South Africa was worth investing in, that they shouldn't leave, and they had set up meetings for them with government officials, and took them around plants, and so forth, and weekends in Cougar National Park. And they also had them meet with leaders of the liberation movement, which the government didn't like but they tolerated it. It was complete failure form the point of view of either stopping capital outflows, or bringing capital in. But again, it was a way in which leaders of the South African financial sector developed personal relationships with people in the liberation movement that would turn out to be important later. So, these relationships were built, and they had the effect of breaking down the barriers, at least at the elite level, between the leaders of the liberation movement, and the leaders of both the big mining companies, and South Africa's financial sector. What about carrots? Why might carrots matter? Carrots are important because even if people think the status quo is not an option anymore, they may nonetheless dig in their heels. So for instance in the 1960s, and 70s, in Rhodesia as it then was, the white farmers didn't see any good future, and so they dug in behind the Ian smith government until the very end, until basically the British pulled the plug on him. We're gonna talk more about Israel Palestine in a few classes from now. But I can't tell you how many Israeli businessmen I have interviewed who say, "the status quo is terrible "we hate it, it sucks, and the alternatives are all worse." And they mean it, they believe it, that's how the world looks to them. So in addition to thinking that the status quo is bad, maybe even untenable, people have to be able to see a possible future that has some appeal, at least greater appeal that then status quo. And the important thing here was the collapse of communism. Because the leadership of the ANC and the leadership of the South African communist party had pretty much been the same people; the same actual individuals, people like Joe Slovo. And believe me, these people were not like Euro-communist in expensive Italian suits. Just to give you a sense of what kind of communist party it was, during that coup in Moscow, or failed coup in Moscow, that I talked about in the first class on the collapse of the Soviet Union, those three days when Gorbachev was locked up in his dasher on the Black Sea. Three communist parties in the world recognized the coup, one of which was the good old SACP. So these were hard-lined Stalinist communists, and known to be such. Why are they known to be such? And in fact, the leader of the South African communist party, Bram Fischer, had been an Africana judge, and a son of the state president, and was caught and horribly tortured for the rest of his life in prison. So you know their gloves were off with the communists, and people, and as far as white business elites were concerned, the big fear was that if you had a transition to majority rule, there would be communism. Right? The ANC and the South African communist party were were wanting the same thing from their point of view. Once communism started to go off the table, that reality changed. As one business man I interviewed put it, he said, "I woke up one day and realized we do not "have to be Cuba, we can be Brazil." And when that light bulb went off for him, it was an epiphany, because they suddenly realized that there was a possible future for them in a new order. And particularly once you see the status quo is collapsing, you wanna get to a better future if once you can see that possibility. So, an important fact about South African business is that the ownership of the economy was highly concentrated at this time. Mining had long been dominated by a small number huge companies: Anglo-American and DeBeers being the two most important. Just to give you a sense, in 1985, Anglo controlled more than half of the capitalization of the whole Johannesburg Stock Exchange, and the top five corporate groups controlled 80% of it. Concentration in manufacturing had been increasing since the 70s, and by 1985, the largest 5% of firms contributed at least 50% of output. So, even though the South African economy is not an oil curse kind of a story, it was a pretty diversified economy. It was owned by a small number of large conglomerates. Mike Spica who is the general manager of Anglo that I interviewed in one of those videos, which you can see in the case, he said, "you can get nine people around a table, "would have been nine men, "you can get nine men around the table, "and you basically had the whole South African "economy there." So, this is just, again, you can peruse this at your leisure, but this captures the conglomerate quality of Anglo, all the bits and pieces that it owned, and its subsidiaries. Why did that matter? It mattered because it meant that business could act together. This is a picture of what economists, or political economists, have referred to as a K Group. So if you have a lot of actors who need something together, there's a collective action problem. Right? How do you get them to coordinate and pay the costs of coordinating to provide some good from which they will all benefit. A classic collective action problem, for which man sir Olsen is famous. So the idea of a K Group is, is if there's a sub group within that group that has the quality that if they can solve their own collective action problems within the sub group, and provide the collective good, it's worth it to them to let everybody else freeride. The benefits of providing it for them are such that it's worth it let everybody else freeride. So these big corporate players like Anglo, and DeBeers, and Paula Rand, and the other major companies formed a K Group, and they get and meet at somebody's house and talk about what business was going to do, and that made it possible first to solve their collective action problem, but it also made it possible for them to intimidate the government. If you read Adam Tooze, a former colleague sadly lost to Columbia, he's fantastically good. Economic history of the third Reich; The Wages of Destruction. You know, part of Tooze's story is that when German big business started giving Hitler trouble, he threw a couple of people in prison, and everybody else fell into line. A bit like the story of Putin and Tarkovsky that I talked to you about earlier. I think it would be too much to say that in the 1980s, South African government were country bumpkins. But they basically, their big base of voter support was Africans farmers, they were not very involved in the commercial sector, that changed over the course of the 1980s; Africanas began moving in significant numbers into the corporate sector. But certainly at the beginning of the 1980s, if they had tried to take over Anglo-American, the government wouldn't have had the first clue to what to do with. So business was in a strategically strong position vis-a-vis the government, and they could act together, the big players could set the tone, and maybe the terms of what was going to happen. And, they had a double agenda, they wanted, first of all, not to have a civil war, or some massive repressive outcome that would be horrific to live in, but they also, so they wanted a transition, and they maybe thought some kind of transition was more or less inevitable, but they also wanted to have a big influence on the ANC. Because the ANC, who were were the ANC leaders? After all, they were either living in, they were either imprisoned on Robben Island, Mandela and his friends, or, you know, they were scattered around, some in London, some in East Europe. What they knew about economics, they had learned from reading, you know, summaries of Das Capitale and the Communist Manifesto in East European Guerrilla Training Camps. And, you know, the leaders of South African business knew this. So what they wanted to do was, you know, they like everybody else were reading the Economist every week, and drinking up the Washington Consensus, and all of that, South African Business. They wanted the new South Africa to be part of that new story. And so a big part of their agenda was to influence the ANC. This is why I said dual-agenda; to be handmaidens for a peaceful transition to democracy, and to disabuse the ANC of its traditional positions, which were about nationalizing the command during the heights of the economy. And so they acted first. In 1985, the head of Anglo-American, Gavin Relly, just went to Lusaka and met with the leadership of the ANC in exile himself. And the government as furious, they criticized him, and they attacked him, but they couldn't stop him. When student groups, and church groups, and others said, "we're gonna do it too," they said, "we'll throw you in jail if you do that." But they didn't stop business groups, and this became known as the Lusaka trek. These business elites started going backwards and forwards to Lusaka and to Mozambique, and eventually a whole second set of business leaders started meeting thanks to Kans Gold Consolidated Gold Company in London, with the leadership in exile in London. And so they started having these talks. And they then also decided they had better start talking to the ANC, underground ANC in the country. The ANC was banned, but there was a front organization called The United Democratic Fund, the UDF, and they started having secret meetings with them. In people's living rooms. This is stuff for which you could be thrown in prison without trial, for 90 days, and then later for 180 days. And what they would do, they would throw you in prison for 90 days, or 180 days, without trial, and as you were walking out the door, they would throw you in for another 180 days without trial. There were people who were in there for years, many years, under these detention laws. So these people were taking huge risks meeting with the leaders of the UDF in the mid 1980s. And after a while, the UDF got, you know, irritated with this, they said, "well, you know, it's cool to have nice conversations "in people's living rooms, but we need an organization "to deal with; we need you to form a group." And so eventually they did, they formed this group called the Consulted Business Movement in 1988. And what was hugely important about this group was that they were the first group of white elites to say, "we favor majority rule democracy for South Africa; "one person, one vote." Because all the white parties in government, including the white opposition had been talking about power sharing deals, you know, vetos for minority groups, educational qualifications for the franchise, you name it. But the CBM was the first group to say, "we are for a normal." ANC had kept calling it, "we want a normal democracy "in South Africa." So that solved their credibility problem. They didn't say, "we're neutral about the politics. "If we're committed to a democratic transition, "then we will work to bring it about." And that meant they were the only credible brokers. Church groups were all aligned with different political parties. There was nobody else that could be credible brokers between the government an the ANC. And so, just to give you one pointed illustration of how they broke it, this is a famous Sidney Frankel story. He thought it was important that the two chief negotiators, who would be the two deputies of the ANC and the government, should get to know each other. They turned out to be the two chief negotiators, Ramaphosa and South African deputy president Roelf Meyer. So the took them fishing on this trout farm. He arranged for them to be brought in helicopters. So the story goes, they went fishing and Ramaphosa got a fish hook stuck in this finger, and I am sorry, Meyer got a fish hook stuck in his finger, and Ramaphosa said, "I Know how to fix that, "bring me some pliers, and a bottle of brandy." And they did, and he said, "you drink the brandy, "give me the pliers, but you have to trust me." And so the story goes that this was the bonding moment between, who knows whether it's true, but it stands for the proposition that the business elites were facilitating the personal interaction between the government and the ANC. They formed the Consultative Business Movement, as I've said to you. And then once the CODESA talks were planned, there was no other group that could be trusted to function as a secretariat for the groups, for the CODESA talks. So, the CBM became essentially the convener the round table talks between, among, the 19 parties. And when CODESA I collapsed, it was the CBM people who did all the running around to get CBM reconstituted the following May, CODESA reconstituted the following May, as CODESA II. And when CODESA II collapsed after Boipatong and all that, it was they who became the back channel, walkers in the woods, so to speak, keeping the conversation going. So hugely important role that business played in convening the talks and then picking up the pieces when they fell apart. So, managing spoilers. I talked earlier about the Kissinger Carrington visit. That was also organized at the behest of the CMB, but more importantly, once it fell apart, it was then that the Consultative Business Movement people who decided that they had to act to pick up the pieces. And if you look at the interview with Colin Coleman in the video. Colin Coleman is an interesting man. He was an architecture student at Wits, University of Witwatersrand, came of age, he just graduated in mid 1980s, and was the first full-time employee of the CBM, and they hired him right out of college to work in facilitating the talks with the ANC. Today, he's, by the way, the head of Goldman Sachs for Sub-Saharan Africa. But in those days he was actually a member of the ANC, young ANC member, although he was white was unusual, and so he had a certain amount of street credibility in talking to the ANC. And he was he one after the Kissinger people packed up and left, who went racing around trying to figure out what would get Buthelezi to participate in the election, what would get Inkartha to participate? And they basically, they wanted various guarantees for the Zulu king, Goodwill Zwelithini, they wanted protections of property in KwaZulu-Natal, they wanted a variety assurances that they were gonna have some say in the politics. And so, it was crisis management and shuttle diplomacy that defused the violence that you saw in the final run up to the election. There were also just massive issues of election mechanics. So, one interesting tidbit here is because the election date was non-negotiable, it had to occur in April, or the country would have really descended into a complete civil war. The ballots had already been printed when it was clear that Inkatha was not going to participate. And finally, virtually Coleman and others had gotten Inkatha to agree to virtually to virtually all, gotten the government and the ANC to agree to virtually of Inkatha's demands. But one of the wants, they said, "we've gotta push the election back, "cause we haven't been campaigning." You know, so they wanted to push the election back. And of course, both the government and the ANC knew that that was a non-starter. And so, finally, they were confronted with the problem, again, tiny things, you know. The person who said don't let small things bother you, has never spent the night in a tank with a mosquito. So, the problem was that the national party was on the bottom of the ballot, and they had been campaigning on the slogan: "Vote the bottom line, vote the bottom line." But there was no other place to put a stick on, on the ballots. And these stick-ons had to be printed and, you know, in record time. And this is a developing country, it's not you know, I mean I know we have our issues with butterfly ballots and election interfering and all that, but this was a huge logistical operation. So first, the national party had to be convinced that it was so dangerous not to have Inkatha on the ballot, that they had to not withstand it. Their "vote the bottom line" slogan on all their literature, that had to, there was no other place for the stick-on but to go there. Then the stick-ons had to be printed, they had to be flown around the country. Business played a role in all of that. And so you get the election result. And, the election as I said to you last time, was widely declared to be free and fair. And there is a wrinkle to that though that's always struck me. I've never seen reporting upon on it, which is that the election results took several days to, you know, finally offer the case. And when you got paper ballots being counted by hand, it took three or four days to get the final election results. But at the end of the first day, it was declared that KwaZulu, that Inkatha had won KwaZulu-Natal regional parliament, and that they had enough votes for three seats in the cabinet. And so, you know, I could never prove it, but I have sneaking suspicion that that part of the election results was negotiated in advance of the election. Building civil society support. Again here, business worked with church groups and others to put together hundreds and hundreds of meetings, where they would come and talk about scenarios, they would come and talk about how the economy might unfold in the future. They would do this, part of their effort to influence the ANC. Which, by the way, they succeeded in doing. When the ANC came to power, it still articulated a pretty staunch socialist economic program. They had a manifesto called Reconstruction and Development, and it talked about nationalization. Mandela's first international visit when he came out of prison was to Cuba; heavily criticized by the Americans. Yet, but within months of coming to power, the ANC had largely repudiated its RDP, Reconstruction and Development Plan, and instead embraced something called GIER, Growth In Economic Reconstruction. I forgot what it stood for, but basically it was a shift to the Washington Consensus to a very large extent. It basically put off massive transformation of the economy, the effects of which I'll talk about briefly in a minute. They created an election fund, they helped with the logistics of the election, and ensured that the elections would be free and fair. So, quite a remarkable story in many ways. Business does not normally behave in this way. And there were undeniably notable achievements. The first is that they avoided catastrophe. I think that it is a fair retrospective judgment that but for the things that businesses were able to do, both in creating relationships and networks in serving as through the consultant of business movement convener of talks, and picking up the pieces when they repeatedly failed, helping to manage the spoilers, and getting them to participate. In all of those ways, there were important achievements. They really did help abolish apartheid in a peaceful way. End the country's pariah status. And had a lot to do with the resurgence of the economy; this was the era where people were talking about the BRIC countries, right? Brazil, India, China. Brazil, Russia, India, China. They started talking about the BRICs because South Africa had a big economic resurgence. If you look at sort of macro-economic and social indicators, in the first 12 years or so, the number of people living in poverty went down no in a non-trivial way. And they did create a black middle class. Between 1995 and 2009, half a billion Rand, and that was when the Rand was worth a lot more than it is today, was transferred from white business to black business. So, in all of these respects, you could say there was a significant achievement. And business deserves considerable credit, some of these business leaders for taking big risks and being willing to do what they did. But on the other side of the ledger there, pretty challenging legacies of the way things unfolded. We still have a limited democracy; single party dominant democracy, as I talked about last week. Those tend to be very corrupt, South Africa is not special in this. Go to Russia, go to Mexico, go to India for a long time, single party dominant system, very high levels of corruption come with single party dominant systems. The politically entrenched labor movement has been very costly for the country over time because it protects, and this is the South African variant of the story we were walking about in Germany and in the advance capitalist countries earlier in the course. We have very high wage economy in the formal sector, but we have huge unemployment; we have 30% unemployment, and almost 50% youth unemployment, one of the highest youth unemployment rates in the world. South Africa homages textile jobs to Lo Soto, even to China. So you have the downstream effect of this very powerful entrenched organized labor movement that's wired into the government. By the way, that all exploded in 2012, at a platinum mine in a place called Marikana. Where again, ironically in view of the 1970s, wildcat strikes were organized against the unions and existing managements for sweetheart deals that they believed were excluding them, and about 34, I think it was, people were shot dead by the police, many in the back. Huge blood on Cyril Ramaphosa's character, because it made it much harder for him to win the ANC leadership in 2018 because he was one of the directors of Marikana mine. Since the transition, businesses completely disengaged. While they were very proactive at this time of crisis; it's much harder to get people's attention a chronic problem, than for a crisis. And after the transition, South Africa's problems back from being crisis to chronic: The inequality, the poverty, the lack of racial integration, except for elites. And so if you look, this is a Lorenz Curve, which a Lorenz Curve is just another way of depicting the Gini Coefficient measure of inequality. You can see the dotted line is 2009, the blue line is 1996. So in the first decade and a half after apartheid, the overall structure of inequality in the economy has not changed. And a decade later, it would be much different. This was in the Wall Street Journal, four years ago now. But these numbers wouldn't have substantially different today. On the left-hand column there, you can see, again Gini Coefficients, you can see South Africa is one of the most unequal countries even decades after the end of apartheid in the world. It's actually got the highest Gini Coefficient of countries for which we have data. That doesn't mean the highest in the world, I suspect if we had data for places like Saudi Arabia, it might have a higher Gini Coefficient. But in any event, it's basically 0.7, enormously unequal, even more unequal than, you can see there, than countries like Columbia, Brazil, and Mexico. So, unemployment, I already told you the story. But you break unemployment down by race, and you can see that, you know, a third of the black population is unemployed. Half of the black youth population is unemployed. If you look at other social indicators, if you look at people living in poverty, half the black population is still living in poverty. Adults with high school only education. Again, you've got, you know, 15% of that black population has a high school education. Private health insurance, and the public health insurance is not very good, 10% of the black population. Indoor plumbing, you know, we're living in the richest country in Africa, this is in 2014, but not much different today, a third of the black population does not have indoor plumbing. And so, you know, this is one way of saying that good things don't all go together. So you avoided civil war, the outcome really would have been catastrophic. Very few historians would dispute the proposition that hundred of thousands, probably millions of people would have been killed. And it would have gone on for a long time, and it's not clear, at all, what would have come out of the other end of it. But we are in a world in which, you know, there's very substantial disillusionment as politics now normalizes and revolutionary generation passes from the scene, and we face the run of the mill problems that many developing countries face. Cyril Ramaphosa, you saw him as a young man in that very first video that I showed, where he was a leader of the National Union of Mineworkers, and then COSATU, the chief negotiator for the ANC, and he was worried when in that video talking about being shot at. You know, his story is a very interesting depiction of the interaction between business and politics in South Africa over time. He begins as an ANC activist, he then becomes the head of the National Union of Mineworkers and COSATU, eventually the secretary general of COSATU and chief negotiator for the ANC. He was edged out of the succession to Mandela by Mbeki when he was the heir apparent, so he decided to go off and get rich, he became South Africa's first black billionaire; hugely successful. Then he came back into politics as Zuma's deputy, and in a very bitter fight, finally won the leadership of the ANC. But it would be another entire lecture to go through the slandiness of his grip on the ANC which is partly a result of the way in which it's governed. But that will be a topic for later in the course. Next, we will be talking about the ICC and the Doctrine of Responsibility to Protect. (xylophone music) |
Power_and_Politics_in_Todays_World | Lecture_9_Privatizing_Government_II_Prisons_and_the_Military.txt | - Today we're going to continue talking about Privitization of core government functions. And our central focus is going to be on prisons and the military. I'm gonna start just by revisiting the basic outlook of privatizing government that became so powerful in the early post-Cold War years. We'll then focus first on privatizing the military in general and that'll lead us into a discussion of the Host Nation Trucking example in from Afghanistan about which we put together that case that was on the syllabus. Then I'm gonna give some background on the U.S. prison industry and that'll lead us into a discussion of the privatization of prisons and what we should be thinking about more generally about privatization of core government functions. Just to put my map back up on the screen, this is from last time. We distinguished post-communist privatizations from what I was calling Neoliberal Privatizations part of the small government agenda at home and the Washington Consensus a broad. And distinguished previously public or public sector nationalized industries with things like railways and utilities from the wielding of public authority as we spend quite a bit of time on eminent domain. And then we left for today what I'm calling core government functions; policing, prisons, the military, things that have to do with people's basic survival, the state's monopoly on the legitimate use of coercive force to use Max Weber definition of state, policing, which we're not gonna talk about though, I should perhaps have mentioned that one of the consequences of the privatization of the provision of local services in housing, the common interest developments in many respects, they also have some privatization of policing. There's an old joke about what are the Yale police for by the way? And the answer is, was to protect the students from the New Haven police where they've kind of, one could make of that what one well. But I should have gone a little bit further that in a sense these common interest developments are gated communities, we tend to think of gated communities as communities only for wealthy people. Gated communities around Cape Town where, or around actually any of South Africa's cities, where wealthy people live and they have their private law enforcement but one of their interesting features of the common interest developments in the U.S. is that 60 million people live in them and there are of many different income categories, but they are essentially a kind of gated community even if the gates are often invisible. So today we're gonna go to core government functions. And just to take us back to how people we're thinking about privatizing government, let's go back to December 1994. - Our reinventing government initiative led by Vice President Gore, already has health of shrink bureaucracy and freeing up money to pay down the deficit and invest in our people. All day we pass budgets to reduce the federal government to its smallest size in 30 years and to cut the deficit by $700 billion, that's over $10,000 for every American family. In the next few days, we'll unveil more of our proposals. And I've instructed the Vice President to review every single government department program for further reductions. I know some people just wanna cut the government blindly and I know that's popular now but I won't do it. I wanna linear not leaner government, we can sell off the entire operations the government no longer needs to run and turn dozens of programs over the states and communities that know best how to solve their own problems. My plan will save billions of dollars from the Energy Department, cut down the Transportation Department and shrink 60 programs into four at the Department of Housing and Urban Development. A new government for the new economy creating flexible, high-quality, low-cost, service-oriented just like our most innovative private companies. - So he wasn't talking about shrinking government to the size when it could be drowned in the bathtub as Grover Norquist would subsequently argue, but it was very much, he went on in that speech to say, "We propose to stop doing things that government doesn't do very well and that don't need to be done by government." So there's a sense that we should be looking everywhere we can to privatize what government does. This really had started in the 1980s with Reagan but it accelerates in the 90s with Bush I in the Clinton administration. And we will see particularly under George W. Bush the privatization of core government functions starts to really take off. In 1993, this is the national performance review called that the Clinton administration conducted, they called for aggressive outsourcing of government work. It led to the elimination of almost half a million jobs in the federal civilian workforce. And interestingly, by 2001, this is the beginning of the Bush administration, the Pentagon contract workforce outnumbers civilian Defense Department employees for the first time. Those who can remember back that far, this is where in Donald Rumsfeld was Secretary of Defense. And he made it his business, this is before 9/11 happened, he made it his business to rethink the whole structure of the military much more on private sector principles and to slim it down. And so if you look at the at the U.S. government workforce over the past two decades, you can see that the size of the number of employees working directly for the federal government doesn't really change that much and by some measures it decreased. This Purple Line is what we call Grant employees. Most of that is grant money given by the federal government in Grant programs to state and local governments or for specific projects. And you can see that that had continued to increase but after the financial crisis, it tails off as well when government comes under real fiscal stress. My guess is that this little peak at in the first couple of years after the financial crisis was stimulus spending, things like NIH spending and others which then tailed off. But the notable fact in the beginning decade of the century is this massive increase in the number of contracted workers. This is contracting out of work. And a big chunk of that increase that you can see there in the blue line, also up through 2010, was outsourcing in the military. This is courtesy of Rumsfeld and his successes. But the idea that we would rely much more heavily on the private sector in all of the operations of the military's a big part of that surge that you see up through 2010. And even then when it falls, it goes nowhere near where it had been before. And I suspect for reasons I will get to later that if you have these numbers through 2019, this will increase. Okay, privatizing the military. So the idea that there are private contractors involved in the conduct of war is not new. If you go back to the Revolutionary War, you can see that actually in every single war, we have had private contractors in military personnel involved all the way back to the Revolutionary War. They are the yellow bars rather than the blue bars who are irregular military. And you can see that they've had a rather small presence in most of those wars, the largest proportionately before the Iraq war being in the Civil War and in the Second World War. But the Iraq war is notable because for the first time, you can see that we have more or less the same number of private military contractors. This is by 2008 in Iraq. And you can see basically we have more or less equal numbers of private military contractors unlike the first Gulf War which we talked about in the very first lecture where the coalition forces were almost half a million people even though we didn't engage in regime change, we had a much more ambitious agenda in the 2003 Iraq war even though we used many few of our own troops and we relied much more on these private military contractors. The Afghanistan war takes it to a whole new level, this again is 2008, you can see, this is, We were drawing down troops in Iraq after that and this is the surge in Afghanistan that starts at the end of the Bush administration continues under Obama and the private military contractors become less of a proportion of the total during that surge although we will see later that in a decade since then, the picture is somewhat different. So you might be asking yourself, well, who are these private military contractors and what do they actually do? Is this just a euphemism for mercenaries? Mercenaries, they're as old as the hills after all. And at least as they are employed in the U.S. or by the U.S., they're strictly not supposed to engage in front-line fighting, the offensive fighting, but they can do just about any other function, and they do a huge variety of things. So just to give you some sense just this as a flavor of some of these companies of their size and what they do. Here's one called G4S, it has about 625,000 employees mostly involving itself in Iraq and Afghanistan. And it you can see it does a whole gamut of things; routine security, banks, airports and prisons but more heavily armed security, mind clearance, military intelligence and training. Another one, Erinyes. Mostly in Africa, although it's had a lot to do in Iraq as well, it guards oil pipelines. It has about 16,000 guards, so it's considerably smaller but working or it has worked almost everywhere in Iraq often protecting commercial ventures such as iron or oil and gas projects in the Republic of Congo, The Asia Security Group, a smaller one about 600 guards but it's been very active in Afghanistan particularly security for the high government officials. It's had millions of dollars in U.S. contracts and protected supply convoys in Afghanistan South about which I'll say more in a little while as well. DynCorp or DynCorp, not sure how they pronounce themselves, pretty big sized one, 10,000 employees, 3.4 billion in revenue. Again operating in many parts of the world; Iraq, Africa, Latin America, Eastern Europe. It become involved in, you might think of, as policing missions, anti drug interdictions, disarmament of fighters in stalemated civil wars, a whole variety of activities. Triple Canopy, have interesting names, some of these companies. About 1,800 people, mostly from Uganda and Peru. And they're making contracts worth about 1.5 a year. Involved as you can see is countries as different as guarding the US Embassy in Haiti, protecting US personnel in Israel and many other things. Aegis Defense Services, a very big one, 5,000 staff, UN missions, peacekeeping missions sometimes and it's been involved in emergency response for governments risk assessment activities, so the government gamut of activities. And so this is a particularly notable military contractor, changed its names several times, it used to be called Blackwater and it was one of the most notorious of these, particularly it had was famous for a very aggressive tactics, it's big, it's had some 20,000 people working for it at different times as you can tell many parts of the world. They got notorious mainly because of Iraq in two different episodes in Iraq. One was that it precipitated the first battle of Fallujah in 2004. The army had not intended to take Fallujah but four Blackwater security personnel got themselves killed and it went viral, their bodies hang, their burned bodies hanging from bridges in Fallujah basically forced the U.S. to redirect its plans and go and take Fallujah which they had not intended to do. And this is one of the ways in which military contractors can actually affect the primary missions. But then several years after that in Baghdad, Blackwater shot and killed 17 civilians including a nine-year-old child and much of this also went viral as had the Fallujah episode. And eventually four of them were were convicted of murder and one got a life sentence, the other two got very strong I think 30-year criminal sentences. So Blackwater became synonymous with these freewheeling uncontrollable military contractors which actually had an effect in thinking about how they should be restrained and governed. There was something adopted called the Montreux document, the following year by 52 countries including the U.S. that listed a whole series of good practices and said among other things that the government's employing them would be accountable for what they did. Of course like many of these kinds of protocols whether they can be enforced is another matter. But so Blackwater then has renamed itself twice; first to G services and now is called Academy, or Academia, I don't know how they like to pronounce themselves, but most people who have heard of any of these companies have heard of them. So that's the kind of thing that they are, as I said they do, they'll do virtually anything, they will fight for any government if the price is right and they will largely do whatever they are asked to do. The personnel in them have changed a lot. For instance in the 1990s, a lot of them were former South African military after the settlement in South Africa. Those people now are pretty long in the tooth, probably too old for this kind of activity. And so they come from all over the world. But in certain situations and particularly will see in Afghanistan, there's very heavy reliance on local populations as a source of employees by these companies. So let's talk a little bit about Host Nation Trucking in Afghanistan. These, the convoys, you can see that they were employed to guard. And it's important to say little bit about why this came to be the case. So this is a map of Afghanistan. As you can see, first thing to notice is it's a landlocked country, I haven't got Kabul marked on it but Kabul is right there pretty much. And for reasons we're gonna talk about when we get into the the class about the Global War on Terror, the U.S. went into Afghanistan after the 2000, the 2001 attacks with a very light footprint, we went with very few troops on the ground. In fact, the way we fought the war in Afghanistan was to get behind one side in an existing Civil War. There had been an ongoing civil war in Afghanistan between the ruling Taliban and a group called the Northern Alliance. And the Northern Alliance was all but defeated in 2001. We wanted to do Afghanistan on the cheap for reasons I will talk about later and we made the judgment that the way to do Afghanistan on the cheap was to get behind the Northern Alliance. And the idea was to help them win the civil war so that they would then become a government that we could work with. And that is exactly what we did. But the thing we didn't think about or certainly we didn't draw the right conclusions, if we did think about it, is that if you get behind the losing side in a civil war, it's probably gonna be the case that that government is going to be having a hard time governing because they're probably reasons why it's been the losing side in a civil war. It might not have a lot of support, for instance, among the population or it might not be able to create on its own a Weberian monopoly of the use of legitimate power precisely because it doesn't have that capacity. And so indeed we did help the Northern Alliance when we defeated the Taliban and we put them in power, we may have seen them as a government we could work with but of course a government that comes to power that way, it is inevitably gonna be seen as an American puppet government on the ground which tended to be what happened. And so basically people, you know, you'd read article after article and if you would interview people, you would hear statements to the effect that the government in Kabul for years didn't... You know, you could once you're five miles outside Kandahar, really the government had no real control of the country. And so this was the reality in which this ongoing war was being prosecuted. And indeed as we will see when we dig in the lecture on the global war on terror, by 2003, many of the commanders on the ground in Afghanistan were advising up the chain of command that the Bush administration should be trying to make a deal with the Taliban perhaps to create a government of national unity but the judgment on the ground was that there was no way that the government we had put in place was ever going to, given the resources we were spending and and so on, was ever gonna become a legitimate and effective government. That argument was systematically either ignored or rejected and where it's now of course the longest war in American history and we are on the verge of doing a deal which will return the Taliban to power. So this was the world in which we were operating for much of the first two decades, we were basically supporting a government that doesn't control its own territory. And we had a very light footprint. And so the challenge was how to move supplies around Afghanistan which was essentially not secured in particular this Highway I which you can see goes basically around the country. And we have to be able to move personnel and particularly equipment through this hostile territory. And that was not gonna be easy to do and so you can see what happened at the start into, not at the start, but in 2007, we had about 25,000 troops there and more or less equal number of military contractors. You can trace these numbers up through the surge but by, you know, six years later, we'd brought these troops from Iraq but then drawn them down somewhat. But reliance on military contractors continued, so they vastly now in 2013 and this has continued since, we've drawn down troops further. So we've been heavily reliant on these military contractors. And basically what we've done is use them to guard convoys. The reason we wouldn't... why wouldn't we wanna use American troops to guard these convoys? Any, why would we wanna contract out the guarding of these convoys? (man speaks off the microphone) Pardon? (man speaks off the microphone) What kind of cost? The financial cost, there's some financial, yeah? (background noise drowns other sounds) We don't have a draft so that again we have to use professional soldiers that can be expensive, yeah. (background noise drowns other sounds) They understand that terrain better than the U.S. troops. That's also true but it comes with the sting in the tail which we'll talk about a little in a few minutes. But the real reason goes back to the lack of a draft that this is really hostile territory and if we had started using a lot of American soldiers to guard these convoys, we would have started having a lot of American casualties. And having a lot of American casualties would have made this war much more difficult to fight because people would be calling their Congress people, their senators, you know, you don't have to think that far back to see how this played out during the Vietnam War. It was when large numbers of Americans were being killed that the support for fighting the Vietnam War went away. So the idea was not to expose politicians to the political cost to, not to expose them to the political cost of having a lot of soldiers there. But the difficulty with relying on people who know the terrain a lot better is that they also know how to take advantage of you, right. And so we had a lot of contradictory imperatives because we were trying to pacify the country, we were trying to help the government get control of the country. So for instance among other things, we put in place rules which said that private contractors guarding millitary convoys couldn't have weapons more powerful than Ak-47s. The problem was that these convoys were being attacked with rocket-propelled grenades. So how are you going to actually protect the convoys if you don't have the relevant firepower? Big challenge, what's the answer? (man speaks off the microphone) Pardon? (man speaks off the microphone) And how do you do that? - Through the terrain. (audience laughing) The answer was money, right. The answer was money. You basically had to pay the people who might otherwise be attacking you not to attack you, right. So there were huge shakedowns. And these are from the case, I put them on the slides, it's not to read them out to you now but you can look peruse them at your leisure, but they're basically whistle blowers and others reporting what the going cost was to make sure that a convoy didn't get attacked. And so the Taliban was charging $500 per truck from Kandahar to Herat. And the different prices listed there. This is a slide from a presentation to a congressional committee which basically said, in order for the Host Nation Trucking contractors to be able to work in the Sharana region, they had to basically pay $150,000 a month. So, and that this became scandalous because it turned out as you, it shouldn't take you too long to realize that we were sawing off the branch we were sitting on in that we were actually funding the guerrilla movements that we were supposed to be fighting. And the reason was that we had to essentially pay in order to be able to prosecute this war with reduced cost, economic cost, that when you're not sending professional well-paid U.S. military in the hundred tens and hundreds of thousands and avoiding the political cost that would have come with doing it with our own troops. So that became a huge scandal and there were a lot of hearings about it. And some of the problems were fixed but it remains something of an ongoing problem. But, well, the question I want us to puzzle over is assuming that the problems could be managed, and we'll come back to whether they really can be in a little while, but assuming the problems can be managed and it really is more efficient to fight wars this way, it really does save money and it saves lives and you know, it's a market solution, you're letting people who want to spend their lives taking these kinds of risks take those risks. It means we could fight what would be otherwise unpopular wars. Is that a good thing? - [Student] No. - [Man] It's anti-democratic. - Why is it anti-democratic? - [Man] Because the war is unpopular, it's a war that could not (mumbles) - Well, let me be the devil's advocate there literally. He's saying it's anti-democratic because if the war is unpopular, we shouldn't be fighting it. But when we fight the war this way, it's not sufficient, it's not unpopular, right. We're contracting out to people who wanna fight the war and we're getting to stay home and not have to send us, ourselves or our sons and daughters to die. So it would be an unpopular war if we were sending them but we're not sending them. We're only sending small numbers of professional soldiers. Yes sir? - [Man] (mumbles) perverse incentives to fight more wars. - It creates a perverse incentive to fight more wars. Yeah, if something is cheaper to do, you're more likely to be able to do it. What's so wrong, what's so bad about that? - [Man] You create the for (mumbles) specifically. - So there if you go back and read the Federalist Papers, you find some of the same sentiments as we're hearing from the floor here. There was a great nervousness, this huge debates about whether we should be having standing armies at all because standing armies wanna fight wars. And maybe it should be really difficult to fight wars, and we shouldn't fight wars unless there's a lot of support for fighting the wars. Yeah? (man speaks off the microphone) You gotta yell or get a mic, I forgot to get the mic. Yell, you just gotta yell. (background noise drowns other sounds) - [Man] And that army could even (mumbles) to carry out some of the missions that they carry out. - So you're worried about these armies going rogue and doing their own thing, yeah? Okay, yeah? - [Man] I would say that the problem is not so much in the war in itself but what is the oversight, so we've seen that in the Iraq and Afghanistan wars rather than the Congress declaring a war declaration that's been done through giving power to the President to the war. So if there's no oversight of how the war is going on then we can have this broke armies that we're talking about. So it's sort of undemocratic because there are no checks and balances on the war itself. - Okay, so we'll come back to the oversight issues but the argument of the Federalist Papers and the resistance to having professional armies in the first place was the idea that, you know, war is not a great thing and we should only, I think part of the idea was, we should only be fighting wars that we really need to fight. And if you really need to fight a war, you'll be able to mobilize the citizenry. And if you can't mobilize the citizenry, maybe it's a war that doesn't need to be fought at all. And we will consider later whether there was another option in Afghanistan after 9/11, a path not taken. But the other thing to think about is there's a literature in political science and political philosophy that goes all the way back to Immanuel Kant who first observed that what he called Liberal Regimes of what we think of us today is more democratic countries tend not to fight one another. They tend not to fight one another. And this turns out Bruce Russert who's now an emeritus professor, he is very famous for having studied this empirically at great length, Michael Doyle at Columbia, a political theorist who's spend a lot of time on this, but many people have studied this so-called Democratic Peace. And it seems empirically to be the case that democracies tend not to fight one another. And also at least until recently that democracies tend only to fight wars that they're going to win which again depended upon the idea that it's gonna get people, it's gonna be very difficult to get people to fight unless the war is really important to them in which case they'll be motivated to win. And so if you start to make it very cheap for politicians to wage wars as we have done increasingly by reliance on first of professional military and now military contractors and increasingly we're gonna be relying on things like drones which can be also for without very many soldiers and perhaps at very low cost. And we are funding these wars on debt, we're not actually making people who live and vote today to pay taxes to fight these wars. It's gonna make us more warlike because politicians will not have the incentive to avoid war. And so the finding, the empirical finding in the Democratic peace literature might start to go away. Let's shift focus to prisons and then we will come back and see what these two have in common. So some summary points about prisons and the main big one is that we are a huge outlier in the world. So here you can see incarceration rates 400,000 of population and the U.S. beats everybody, hands down. This was in 2012 but the picture doesn't look substantially different comparatively, we have over 700 per 100,000 citizens. And as you can see, the Czech Republic which comes in there has 200,000. If you look at the top 10 incarcerators in the world, you can see what kind of company we're in even though we dominate them all, Russia, Belarus, countries of that kind. South Africa comes in 10th but we're all, you know, we're not quite double but we're there. And if you wanna look at it over time, you can see that it's really has accelerated in the last four, since the 1970s, it's accelerated by a phenomenal rate particularly for males and we'll see particularly males who are not white. Now some of this has to do with developments in the criminal justice system but not all of it. So just to provide some larger context here, in the 1960s and early 70s, there were big advances in the treatment of psychiatric disorders, so-called mood disorders, things like depression, started to be treated with drugs like lithium and it's cognates and thought disorders like schizophrenia started to be treated with drugs like Thorazine and its derivatives. And that meant that people who had previously spent their whole lives in mental hospitals started to be released. So if you go up to Connecticut Valley Hospital in Middletown or Creedmoor Hospital, at State Hospital in New York, the first thing that will strike you is they'll be, and it's as true in State Hospital, after State Mental Hospital around the country, there'll be six or seven building, the majority of which are boarded up. And the reason is that these people will all be institutionalized in very large numbers. People who might otherwise have spent their lifetimes there. But it also happened to occur during a massive fiscal crisis for a lot of state governments. And this is the dark side of what we sometimes call fiscal federalism. And so as these state governments were saving huge amounts of money by de-institutionalizing mental patients, where were they going? They were going to in Connecticut, they were going to, you know, to Hartford to New Haven to Bridgeport and the state governments we're not giving those resources to the cities 'cause they were so, the state governments were so strapped. And a lot of those people wound up in the criminal population and wound up in prisons. So I'm not saying it's a causal argument or anything like that but there are part of this, part of this population was fed by the fiscal federalism and the fiscal crises of the states that led to an increase in the population that was likely to be very vulnerable to winding up in prison. So they went from one total institution to another. But as you can see here the numbers are staggering and that's only a small piece of the story. So here are the bigger pieces of it. What drove this to big developments? One was a war on drugs which begins in the Nixon administration in 1971. And then secondly in the 1980's, the massive increase in mandatory sentencing which shifts power from judges to prosecutors because the vast majority of prison sentences are negotiated plea agreements with prosecutors, and led to things like the three strikes and you're out rule. So if you're convicted of a third felony, you get a life sentence even... And they're all these famous cases of the third felony being a bounced cheque or something of this sort might not be a violent crime at all. And so these two developments were largely responsible for the much more punitive turn in criminal sentencing and the massive expansion of our prison population to over 200, over 2,000,000 where it is today. Big racial component to this, particularly in the war on drugs, the crimes that were more heavily punished and criminalized, the drugs that were, the drugs for which people were convicted at higher rates tended to be the drugs used by minorities but there were other reasons about which one could teach an entire course about why minorities are disproportionately locked up in the criminal justice system. You can see here, this is 2009, the percent of males, adult males, incarcerated and the African-Americans and Hispanics are much more widely represented. This of course has a political dimension because we have so-called felon disenfranchisement laws. Many states have laws, again you can peruse this slide at your leisure but the reddest are the most punitive. So for instance in Kentucky and Virginia, you're permanently disenfranchised if you've had a felony conviction. And then these other colors them as they get lighter, it's easier to get your voting rights back once you have served your time. And if you look at this contributes to the disenfranchisement of minority populations because here the reddest states are the ones in which African Americans are most heavily disenfranchised as a result of felon disenfranchisement laws. So we've become much more punitive mostly because of the war on drugs and the much more punitive sentencing policies particularly the move to strong mandatory sentencing. But the big paradox is violent crime has actually been falling. Violent crime has been falling. This is from 1993 to 2017 and this is breaking it, the first two graphs are breaking it down first by people and then by age and the second to property crimes for 100,000 people or per 100,000 households. And you can see that the number, the proportion of convictions for violent crime in all of these categories has come down substantially. So we're locking up more people. This is from 1990, the red line shows the number, the prisoners per population but violent crime is coming down. So why might that be? Why would we be locking up more people, why would be seeing locking up more people and violent crime coming down? - [Man] Drug crimes are largely non-violent. - Pardon? - [Man] Drug crimes are largely non-violent. - Okay, but we're locking up... Okay, so we're locking up people for nonviolent crime, that could be one of the reasons. Why else might we be locking up more people with violent crime is going down? Yep? (man speaks off the microphone) Pardon? (man speaks off the microphone) That criminals are in prison. So some would say, well, that's great the policy is working, right, we're locking them up. That's why the crimes going down. Some people would say that. Any other reason somebody might come up with? Yes sir. - [Man] Contractual agreements with private prisons. - Contractual agreements with private prisons. We'll talk about private prisons, they are significant but they wouldn't be significant enough to explain this development. And the private sector prison increase is a relatively recent. Yeah? - [Man] It could be the Police Department are changing their tactics and (mumbles). - More community policing. Well, we've seen we have some experts on community policing in this room but I believe the short answer is we had a rise and fall in community policing. Even though community policing is more effective, it's gone up and down. Maybe it's having a comeback now. Among other reasons, one is, one reason violent crime is going down is demography. The vast majority... One of the best predictors of violent crime is males between the age of 18 and 24 in a population. And as the baby boom bulge has moved through, we have fewer, we have relatively fewer people of that age. So some of it is just demography. There's a theory that has been very controversial by the people who wrote Freakonomics which purports to show that Roe versus Wade is responsible for the decline in violent crime on the hypothesis that those likely to commit violent crimes are not being born because of being aborted. Very controversial, I think. The data was questioned, much criticized and in the last couple of years, they've done a whole series of new empirical studies purporting to defend their hypothesis. But if you read Steven Pinker's book, The Better Angels of Our Nature which traces the decline in violence in Western countries over the last seven centuries. He puts a lot of stock in education and labor force participation of women and argue that as that goes up, violent crime decreases. So the point of the slide being why violent crime has decreased is not a subject about which this much consensus. And there are these and you could probably put up a number of other possible contributors to that. But more interestingly from a political point of view is that even though violent crime is going down, most people don't know that. So the dark blue line is percentage of people in polls who believe that violent crime is increasing. As you can see, the light blue line is the rate at which violent crime is actually occurring. And you can see this pretty big disjunction, an increasing disjunction, between what people believe and what's actually going on. And that's quite remarkable. Various hypotheses about that, a political scientist by the name of Stuart Scheingold who's worked on this, a book called Politics of Law and Order argues that it's being tough on crime is cheap talk for politicians because the politicians who can run and get elected on it are often not the politicians who have to justify paying to lock up the felon to voters. So for instance people running for federal office can run on being tough on crime but the vast majority of prisoners are actually in state prisons that are paid for out of state budgets. So coming back to this slide, one thing I didn't point out earlier is that you can see that around the turn of the century, this all started to tail off. That in fact we started to see declines in incarceration rates and part of that was cost, part of it was perhaps recognizing the inefficiency of punitiveness rates but if you look at private sector prisons, they are increasing. So you can see that they went from being about 6.3 in 2000 to 8% of the population in 2009 and they're well above 10% now, so perhaps even more than that. So the private sector prison industry is getting an increasing proportion of a declining population. And so that's, you know, violent crime is going down, the number of people locked up are going down and yet we are seeing this growth and flourishing of the private sector prison industry. - Last week there was a terrific article, The Financial Times, about how U.S. states are looking to privatize their prisons in order to close their deeping budget deficit. It turns out sell your prisoners to private contractors is a great way for cash-strapped States and even a federal gun to raise money. First off, there's the money upfront. Private firms will pay as much as $10 to $30 million per prison. Last month Ohio sold one for 73 million bucks and Arizona recently finished a series of hearings on a play to add thousands of privately operated prison debts. But the real money comes from the long term cost savings. Since private companies cannot operate prisons at a much lower cost than state governance, the reason, it's not just the magic of capitalism at work, it's really because people who work for private companies don't have huge guaranteed benefits like many state employees. And that's why numerous state governments looking to achieve some kind of long-term fiscal sanity are thinking about privatizing their prisons. In Florida, 29 state prisons are said to be privatized by the end of the year. I'm hoping it can save the state 22 million bucks a year. It's good for the state, it's good for the prison operators and it might even be good for the inmates since public prisons are so overcrowded. And the worse the economy gets, the more desperate the states would be to be raising money by selling their valuable prison real estate. And that's why private incarceration game works so well during recession. This industry is basically duopoly. It's a duopoly between Corrections Corporation of America and GEO group, yeah. Those are the two. For you home gamers, these are the only two significant public (mumbles) firms. Well, companies have a lot going for them right now but the thing I like most is their track record during the Great Recession because if we give it another series flutter, you know you can count on these guys to profit from. Both of these companies have consistently growing earnings every year since 2007. Neither one of them sell, anything you feel and even recession, nothing just didn't do it. These stocks give new meaning to the term recession proof. You can't just bust people out of prison when your state runs out of money, they gotta keep paying for inmates. And of course about $25,000 for your inmate per year and of course the national prison population is sadly one of the strongest secular trends out there. You know kind of like a political but we're not about politicians here by right, we only care about the profits and there's no doubt that this business is lucrative as all get-out. It's only going to get better from here. Right now, only about 10% of prisons in the U.S. are privatized but it's clear the country is moving that direction. So there's a lot of room for both Corrections and GEO to keep on growing plus the ability of governments to build new prisons is simply not keeping pace with the need for more prison space. It's simple supply and demand. The economic and situation dictate that our new prisons will be private prisons, that is good for both CXW and GEO. There's no escape from the notion that this is a fantastic business to be in. - So Jim Cramer knows we're off, he speaks. As you can see that the private sector prison population through 2016 has increased from under 100,000 to about 130,000. If you wanna get a sense of where they are, the darker the blue heading for deep deep blue is the heaviest concentration of these prisons. And if you wanna compare them by state and federal, you can see the federal line is red and the state line is green. Most of these prisoners are in private, in state prisons and indeed you can see here, the federal, this red line started going down basically because the Obama administration decided to phase out private sector federal prisons. However, the Trump administration has brought them back and particularly all the interdictions on the southern border. Almost all of that is massive business for private detective prisons. So I'm sure when we get data for the next few years, this red line is gonna take off to the northeast. While some of you might have noticed a couple of weeks ago, the California State Legislature just passed a bill. I don't think that Gavin Newsom has signed it, he was deterring about signing it but facing him out in California which would be, I think, it'd take a significant, may take a significant chunk away. So what do we think about private sector prisons? Again, there's plenty of room for abuse, we all know the stories about the judges who were bribed to send children to private juvenile facilities and so on. But assume they are more efficient. So assume by whatever metric you think prisons should be judged, recidivism rates or conditions in the prison. Let's just assume for the purposes of discussion, the private sector prison is better. Who would still be against it? Why? - [Woman] I feel (mumbles) and if we decided something that a person does is illegal then it's our responsibility to directly hold them responsible for their actions that (mumbles). - Okay, that might be one reason. Yeah? - [Man] The California private prisons was lobbying to keep the (mumbles). So they are as well political groups to maintain the incarceration. - Okay, any other reasons people might be against them? Assuming they are more efficient, what if they do have lower recidivism rates, what if they had better conditions? A lot of them don't but why would you still be against them? Yes sir. - [Man] It doesn't really adjust the problem that if public prisons are overcrowded and private prisons would have been better, why are public prisons (mumbles)? - Okay, so this goes to the lobbying question perhaps, is that implicit in your question or what you're saying-- - [Man] Yes, it just doesn't address the problem. - It doesn't address... Yeah, so Jim Cramer says the demand is there. You're saying well, we don't know why the demand is there, violent crime is decreasing, most people seem not to know that, yeah. - [Man] The for profits logic will prevail over the good for the prisoners in the long run. What about rehabilitation, private prisons want to keep them. - Well, you could say well, you know, just to be the devil's advocate, if you said, the metric by which they're gonna be judged is how well they rehabilitate prisoners. What if they rehabilitate prisoners better? - [Man] The question is rehabilitation of the prisoners that they don't have. - Right, so let's say, your next renewal of your contract is gonna be conditioned on the recidivism rates from your prison. - [Man] I just find it interesting that we're little more bothered by having private prisons and yet we don't look to other sectors the government looks like health or education, what may not seem to be more prone to be okay with privatization of forth profit behind these sectors-- - I think that is a common response but I'm asking you, why? - [Man] Right, so I think because the case of prisons you're taking away someone's liberties, it affects you at the core much more. And to think that you're putting value to someone else's enter to a private probation hands, it doesn't judge well with that idea of I'm restricting someone's freedom and I'm giving it over to private sector handling. - Yep, you're really gonna have to yell. - [Man] It is (mumbles) to compensation around why we have this social problems, the injustices and the high incarceration rates. - Right, it doesn't address that, it takes for granted, Cramer takes for granted, there's the demand and he has nothing to say about why, you know, as we saw earlier, maybe it's the war on drugs, maybe it's the structural hostility to minorities built into a lot of the differential sentencing and enforcement, maybe it's the people who should be getting mental health care and not being locked up at all. So by focusing on this, we're not addressing the underlying issues. But I think coming back to pull the two parts of the lecture together now, when I talked at the beginning about core state functions, I mentioned Webers definition of a state as having a monopoly on the legitimate use of force in a given territory, that's how Weber defines a state. And the truth is that neither of these activities, the military, privatizing the military or privatizing prisons is really privatization, right. What it really is, is contracting out a government monopoly. We're not actually turning this over into fully private, so that is the problem. And as I've put up on the slide, that creates what economists call principal agent problems of a particularly difficult sort because the idea of a principal agent problem is that the principle contracts out to the agent to do things in this, either the private military contractor or the private sector prison. But it rapidly becomes the case that the agent has more information that the principal needs to monitor the agent then the principal has. And so that creates a lot of difficult problems. So if you think about, you know, in a schematic view of a democracy, we have the voter then we have the politician, the politician is the voters agent then we have the regulator who's the politicians agent and then we have the contractor. And so the chain from here to here has many many points at which... It's like a series of nested principal agent problems that have to be managed. And of course principal agent problems come up all over the place. You say well what is the solution to principal agent problems? One is to have competition. There's not a lot of competition in any of these industries. So as Cramer noted in that clip, basically the private prison industry is completely dominated by two corporations. And this, by the way, they're now multinationals, they operate in Australia, they operate in Europe, they operate in the UK, they operate in lots of places. There's a huge cost to entry into the private sector prison business. You've got to build a prison, you've got to commit yourself for long periods of time, not very competitive industry. Military contractors somewhat more competitive but again you develop, a government's develop relationships with these contractors, they're not gonna suddenly switch to other contractors very easily, say, contractors who might have been fighting for the other side. So it's actually difficult to get a lot of competition in people, instead stick to their relationships that they have. A second way of managing principal-agent problems is to try a better align the interests of the agent with the interests of the principal. Because if they have the same interests then you don't have the agent going off and in a rogue fashion doing something that's not in the interest of the principal. The trouble in this area too, it's very difficult to do that as people pointed out with respect to the prisons, the industries themselves have very different incentives from like reducing crime, you know, having more prisoners is the business in which they traffic. And so indeed, what you find in the prison industry is a lot of lobbying. It grew dramatically, this is up through 2010. If I had a slide since 2010, that number would keep going up. And you can see here they give to at the federal level to candidates from both parties. The light blue is Republican and the dark blue is Democrats. See this is particularly when Democrats are in control of the federal government. As happened here, it's gonna go up, they're gonna give money to incumbents. So they give money to both sides. It won't shock you to know what they lobby for? More lockup quotas, stiffer penalties as somebody pointed out and immigration enforcement because these are big industries. So if good public policy is to reduce the number of people in prison, it's gonna be almost impossible to align the incentives between the principal and the agent. Similarly, and we think about fighting wars, these, you know, we might say good public policies to have fewer wars and certainly fewer unnecessary wars, the military contractors have very different interest. So for them, if the war in Afghanistan goes on for another 10 years, it's just more business. The final way in which people try to manage principal-agent problems is monitoring. But that is very difficult in these types of circumstance. Here's the hearing. - So today we're considering our oversight on the United States government contracting on conflicts overseas. And we're gonna ask the important questions, who's getting the United States taxpayer money and how are they using those funds once they get it? Last week, the subcommittee held a hearing that examined Asia at the Host Nation Trucking contract in Afghanistan. That investigation uncovered distressing details about the United States taxpayer money is funding warlordism and corruption in Afghanistan and how the contract is undermining United States counterinsurgency strategy. Equally troubling is to finding that the United States officials charged with overseeing this contract had no visibility into the actual operations of the contractors or subcontractors. In most cases, official did not know who the subcontractors were let alone who they employed, how they functioned and where they spent their money. To give one example, seven of the eight prime subcontractors from the Host Nation Trucking contractor employed either directly or indirectly, a man by the name of a commander Ruhullah, and he provided security for the supply convoys. Commander Ruhullah claims to spend $1.5 million per month on the ammunition and has reportedly attacked convoys that do not use his security services. Still no United States military officials have ever met with commander Ruhullah. And despite the fact that he receives billions of dollars of taxpayer money, there have been no attempts to enforce the United States laws that govern his U.S. funded contractual relationship. With $2.16 billion of taxpayer funds at stake, it's unconscionable that the military does not have tide of control over Host Nation Trucking subcontractors. But the Host Nation Trucking contract is not the only problem. This week's Economist reports that 570 natal contacts worth millions of dollars were issued in southern Afghanistan but nobody is quite sure to whom? In January, the Special Inspector General for Iraq Reconstruction, one of our witnesses here today, issued a report about the State Department contract with DynCorp which noted that, and I quote, "Over $2.5 billion in U.S. funds were vulnerable to waste and fraud." Close quote. In May, the Inspector General for the United States Agency for International Development issued an audit of his private security contractors in Afghanistan which highlighted significant problems with USAID contracts. It found that USAID does not have, and I quote again, "Reasonable assurance that private security contractors are reporting all serious security incidents are suitably qualified in their authorized to operate in Afghanistan." Close quote. Audits from the Department of State, USAID and others have found problems with subcontractor management that areas as diverse as embassy construction, fuel delivery and educational outreach programs. The Government Accountability Office, another of our witnesses here today has reported that the agencies are not even able to accurately report the number of contractor or subcontractor personnel working on United States contracts. And just yesterday, The Wall Street Journal reported that over $3 billion dollars in cash has been flown out of Afghanistan in the last three years. This $3 billion dollars of cash on a plane flying out of Afghanistan. Officials believe that at least some of that money has been skimmed from the United States contracts at these projects. - So, in this sort of hearing you can find them, you know, dime-a-dozen on YouTube. And it's extremely difficult if you think about the kind of combat situations we're discussing for the principles to have the kind of information that they need to monitor the agents particularly problematic in Afghanistan even when there were only 14,000 security contractors. You can see unlike in Iraq, almost all of them were local Afghans for reasons we've already talked about which creates even bigger information asymmetry because the agents have so much better knowledge than do their principals. And so that the actual, this is, I took this out of the case and you can work through it at your leisure but this is just the Defense Department's chain of monitoring in Afghanistan and down to the Host Nation Trucking. And you can see both the managing and the reviewing of the contract has multiple steps. And then of course some of these are controlled also by the State Department which had a whole different set as well. So suffice to say monitoring is very difficult, very similar story in the prison industry that the information you need think, it's all very well to say, well, we could use recidivism rates but they don't show up for very long periods of time. And you know when a prisoner comes to a parole hearing and the parole officer says, has the prisoner have been behaving, if the guard who's there knows that if that prisoner leaves, the cell is gonna be empty and it's gonna affect the bottom line, very difficult to monitor that sort of problem. And so you could just multiply those sorts of problems by a huge number and you can see that monitoring is difficult in that area as well. But we're out of time and we will talk in on Tuesday about money in politics. (bright music) |
Power_and_Politics_in_Todays_World | Lecture_20_Fallout_The_Housing_Crisis_and_its_Aftermath.txt | - Okay, good morning, everybody. Today, we are going to talk about the housing crisis and its aftermath, which could've been subtitled the tragedy of errors. And we're gonna work our way through the various errors and the tragedy of it over the next hour or so. I want to start with President Clinton speaking in the East Room of the White House at a session with his then Secretary of Housing and Urban Development, Henry Cisneros in 1995. - One of the great successes of the United States in this century has been the partnership forged by the national government and the private sector to steadily expand the dream of home ownership to all Americans. In 1934, President Roosevelt created the Federal Housing Administration and made home ownership available to millions of Americans who couldn't afford it before that. 51 years ago, just this month, Harry Truman rewarded servicemen and women with the G.I. Bill of Rights, which created the VA home loan guarantee program. That extended the dream of home ownership to a whole new generation of Americans. For four decades after that, in the greatest period of expansion of middle class dreams any country has ever seen anywhere in human history, home ownership expanded as incomes rose, jobs increased, the educational level of the American people improved. But in the 1980s, as the vice president said, that dream began to slip away. I ran for president in large measure because I wanted to restore that dream to grow the middle class, shrink the underclass, promote the mainstream values of work and responsibility, family and community, and reform government in a way that would enhance opportunity and shrink bureaucracy. We made good progress, but we have to do a lot more. I ask all of you just one more time to look at that chart, and I wish I had a lot of other chart to show you that would reinforce that. Home ownership declines then stabilizes at a lower level. At the same time, more and more American families working harder for the same or lower wages every year under new and difficult stresses. It seems to me that we have a serious, serious unmet obligation to try to reverse these trend. - So there you have President Clinton in 1995 bemoaning the decline in home ownership that has set in over the previous several decades and pledging his administration to increase it. And we will see that Clinton was not alone that the desire to expand home ownership, which is closely associated, as you can tell, from his comments there, with mainstream accounts of the American dream has been bipartisan and pursued by administrations over a number of decades. Our agenda today is to think about what produced the subprime mortgage crisis that played out in the financial crisis that we talked about on Tuesday. Think about the results of the subprime mortgage crisis. And what are the lessons that we should draw, given the larger narrative and concerns in our course. And you really, in order to get an understanding of the housing crisis and how it played, has played itself out, we really have to go back and understand the place of housing in fights about justice and particularly racial justice in America back to the 1930s. And what I'm calling here the transition from hard to soft apartheid, or if you like, from legalistic to market-based apartheid in America, is something that really occurred over the course of a number of decades after the 1930s. If you go back to the 1930s, it would not have been unusual to see signs like this in neighborhoods, particularly, but not exclusively in the south. And this is what I, this is the height of Jim Crow and the era in which explicit racial segregation was the order of the day. When people bought and sold houses, they would often contain so-called restrictive covenants. Here's a passage from a deed transferring some land. It says: None of the said lands, interests therein or improvements thereon shall be sold, resold, conveyed, leased, rented, or in any way used, occupied, or acquired by any person of Negro blood or to any person of the Semitic race, blood, or origin, which racial description shall be deemed to include Armenians, Jews, Hebrews, Persians, or Syrians. So this is a notion of restrictive covenants and they were, by the way, perfectly legal until 1948, famous Supreme Court decision Shelley versus Kraemer, a five to zero decision with three justices abstaining because they all own houses that had restrictive covenants, and what the court said in that opinion was that while it was a violation of the Equal Protection Clause of the 14th Amendment for states to enforce these restrictive covenants, it was no violation of the Constitution for people to put them in the deeds of houses and voluntary abide by them. So if you were, wanted to signal to buyers that this is a Whites only neighborhood and we shouldn't allow non-White to come in, that was not illegal. And so, and that of course goes back to something I mentioned to you in an earlier lecture, namely after Reconstruction, we had very restrictive readings of the civil right, Civil War Amendments by the Supreme Court and in particular they said of the 14th Amendment that it didn't apply to private action but only state action. So Shelley versus Kraemer in 1948 said these restrictive covenants could no longer be enforced, but that didn't mean that they went away and they didn't for some time thereafter. This is something called the 8 Mile Wall in Detroit. Anyone here from Detroit? Anyone know what the 8 Mile Wall was? It was actually half a mile long, it's not eight miles long, it's half a mile long, but it's on something called 8 Mile Road, and 8 Mile Road was a physical barrier between White and non-White neighborhoods. So the idea was to keep African American, presumably principally African American children from going into White neighborhoods and it since, you can see, has had been covered in graffiti, but it's still there as a kind of monument to the physical segregation, what I'm calling hard apartheid of the Jim Crow era. That was actually, the 8 Mile Wall was built in 1941. Anyone know what this is? It's a map, yeah, (students mumbling) what's all the scribbling on it? - Redlining. - This is so-called redlining, and you can see this saying, this is an OK place to write mortgages, this is OK, this is super, this is the hood, and this is don't even think about writing mortgages in this neighborhood. And so actually the history of redlining goes all the way back to the 1930s and initially the government was in fact behind it. In 1934, we got the National Housing Act which created the Federal Housing Administration, and the following year, they commissioned the Federal Loan Bank Board to run studies in 239 cities in the service of underwriting. They wanted to help underwriters know where these risky loans actually were. And so the idea was that mortgage underwriters would know what to charge and whether, and what circumstances underwriting would be warranted. But of course, it soon became a proxy for denying loans, mortgages in minority neighborhoods and to minority lenders. So redlining became used for that purpose. And as you can see, just if you Google up redlining on Google Images, you can find maps of virtually any American, and by the way, also Canadian cities from this period. This is one from Seattle, and you can see there from the color code, the green is best, the blue is still desirable, the yellow is the declining neighborhoods, and the red, hence redlining, was hazardous. And so that's one from Seattle, I've just put a few up here, one from San Francisco, Buffalo, New York, Austin, Texas, and I could've put many, many more, many more of these. And so this practice of redlining reinforced, of course, racial and class-based segregation of neighborhoods, because banks would only want to lend in the areas that were least risky and most desirable. In 1968, the Fair Housing Act prohibited discrimination in lending based on race and also by neighborhood. This was reinforced by even more stringent legislation a decade later, the 1977 Community Re-investment Act, reiterated these bans, but they're extremely difficult to actually enforce. This is what I'm calling the era of soft apartheid, you don't necessarily need restrictive covenants and we'll see this play out in four-part harmony and in the rest of today's lecture. But even though redlining and loans based on discrimination either against particular neighborhoods or by race has been illegal since the 1960s, as recently as 2015, there were huge settlements with banks who had been caught in this practice. For instance, Hudson City Savings Bank, which services New York, New Jersey, and Pennsylvania reached the largest settlement ever with the Department of Justice, $33 million settlement for denying loans to African American and Latino borrowers and were required as part of that settlement to open branches in minority neighborhoods. And in that same year, there was actually a $200 million settlement with HUD, the Department of Housing and Urban Development, between, for a bank called Associated Bank for doing the same thing in Chicago and Milwaukee. And again, they were required to open branches. So even though redlining is no longer legal, that doesn't mean it doesn't occur. And so this is one of the ways in which it's become possible for markets and market practices essentially to underwrite the segregation of neighborhoods. So I asked you to spend a little bit of time with the Parable of the Polygons, the Schelling, which is a user-friendly version of Schelling neighborhood tipping game. Who did, did you spend some time with it? What did you take away from it? - [Student] I found it a little depressing when I got toward the end. - What's depressing about it? - [Student] Well, that in most of the scenarios, you end up with very segregated neighborhoods and you have-- - But how do we get to segregated neighborhood? - [Student] Well, there was, a lot of the rules were about, you're only happy if, they came down with the 1/3 rule, so if it went down-- - Just back up a little bit and explain how it works for anybody who didn't-- - [Student] All right, so there's two types of polygons. There's the triangles and the squares. And basically the point of the game was to show you at what point, by looking at the map, when all the triangles and all the polygons were happy based on who their neighbors were. And what you discovered was that the more you played around with it, you needed to have segregated neighborhoods for people to be happy or generally happy with their environment. - Okay, anyone want to add a point that was taken away to what he said? Yeah? - [Student] Yeah, I was just really surprised by how little personal bias it takes in order to get-- - Okay, sorry, elaborate on that. - Collective bias. Because even though one square or triangle might only need to be surrounded by like 1/3 of people who are like them to be happy, the way that that ends up playing out is that in order for everyone to be happy it reconfigures so that you're actually really surrounded by people who are mostly like you. It doesn't actually factor out to that 1/3 in like a collective way. - Exactly, so he's, just to summarize what you're saying, very mild levels of prejudice, where mild is understood to mean, I just want some of my neighbors to be like me, 20% to 1/3 is enough to, of course, entire neighborhoods to segregate completely. So very low levels of prejudice will segregate neighborhoods simply by the voluntary choices that people make. You don't need restrictive covenants, you don't need all this other stuff to get it, that's what I'm calling soft apartheid. And not only that, it's very difficult to undo, it's very difficult to reverse, and less people have an active desire to integrate. That we have to change their preferences, otherwise, with very low levels of discomfort at having large numbers of your neighbors be unlike you is gonna be enough to segregate neighborhoods. And so this is a very famous, Schelling was a brilliant economist, most of his works is on arms races, but this neighborhood tipping game tells you that if you start off with a neighborhood that looks like that and people have these very low levels of desire to be surrounded by some people like themselves, it's gonna wind up in pretty short order like that. And of course, we don't start off with something like that in most American cities, so it's even tougher. What we really start off is something much more like this, this is a map of Chicago with neighborhoods are already heavily segregated. And what would be required would be to desegregate them and that's taking Schelling to heart is not going to happen unless you have pretty radical changes in the population. And so in some ways, it might be said that actually soft apartheid is harder to combat than hard apartheid or Jim Crow. I've given you one reason already, first of all, you would really have to have active changes in people's preferences, positive desires to desegregate the neighborhoods, not just absence of resistance to the desires of others. Any other reasons why soft apartheid might actually be harder to fight against than hard apartheid? By hard apartheid, I'm talking about de jure, restrictive covenants, laws outlawing people living in certain neighborhoods and so on. Yeah? - [Student] Is it more unconscious, like unconscious bias? - It could be, there could be issues of unconscious bias, that might be part of it. Anything else, what were you gonna, yeah? - [Student] There's an economic factor now, housing prices have gone so much that it's difficult to actually, even though you enforce it, to have people move in if they have lower incomes. - So people are gonna self-segregate by socioeconomic status, yep, okay. Anything else? - It's much harder to prove. - Pardon? - It's much harder to prove. - It's harder to prove, and it's harder to identify. If you think about hard apartheid, it provides targets. If you wanna get rid of restrictive covenants, let's say, we can have a campaign to get rid of restrictive covenants, gives you a kind of a proximate goal to organize around and to get rid of and I'm gonna come back more and talk about the importance of proximate goals. There's also a literature in psychology about so-called boundary permeability. Academics never say in words of one syllable what can be said in words of five syllables. But here, this is the basic idea. If there's a law which says no Blacks and Jews can join a New Haven Lawn Club, there will be a campaign to get rid of that rule. There will be a campaign to have Blacks and Jews allowed to join the New Haven Lawn Club. If there's no rule, but one Black family and one Jewish family is allowed to join every 15 years, there won't be a campaign. Similarly, if you have a rule which says no junior faculty will get tenure. Then you'll get a junior faculty union. If you don't have a rule, but in fact, only one in eight of the junior faculty gets tenure, which has been a Yale equilibrium for many decades, then you don't get a junior faculty union, why? It's not entirely clear, but it may well be, and this is claim in the boundary permeability literature, everybody thinks I'm gonna be the one, I'm gonna be the one, so if you have norms or practices that function so as to produce exclusion without actually having formal exclusion, it becomes much more difficult to get people to organize, to change them. And so this is, if you look at something like the Civil Rights Act, and you say this is 1964 and the Voting Rights Act of 1965, what has been, what's been achieved? It's mostly that the de jure, rather that the de facto forms of racial discrimination that have been stamped out. It's much harder to stamped out the de facto. So we have inherited a world in which the hard apartheid of the Jim Crow era has been replaced by the much more nebulous, but in some ways, more difficult to deal with, soft apartheid, that gets generated as a byproduct of the lack of proximate goals, the tipping phenomenon that Schelling describes, and the very big obstacles to organizing collective action in circumstances where you have high levels of boundary permeability. So that's something to bear in mind and we'll come back as we see how this played out. People attempting to combat the effects of soft apartheid played out in the housing crisis. Now, if you said, what was the cause of the housing crisis, there's no single answer, it was really a perfect storm. A number of factors came together to produce this result. And I wanna mention four. The first one we've already heard from President Clinton's mouth, and here I, this he wrote the year before that speech, but here you can see, "More Americans "should own their own homes, "for reasons that are economic and tangible, "and reasons that are emotional and intangible, "but go to the heart of what it means to harbor, "to nourish, and to expand the American dream." Home ownership is part of what people who want to realize the American dream aspire to, and a big part of the housing crisis grew out of the agenda of extending home ownership to previously excluded racial and ethnic minorities, exactly the people who are being discriminated against as a result of soft apartheid. And this graph sort of reflects the numbers that Clinton was talking about, you can see that in March of the, starting actually at the beginning of the post-war period, you see big increases in home ownership, and then it levels off and successive administrations do take actions to increase it. Now we're gonna talk about till it finally reaches a peak in 2008. Second is securitization of the subprime mortgage market. So just to be clear, a subprime mortgage is, we're talking about the subprime mortgage crisis, what is a subprime mortgage? A subprime mortgage is a mortgage that is risky to hold. It is risky because it is being given to somebody who either has not a very good credit history or it's being given with a very low down payment so that the home owner doesn't have a lot of incentive to keep paying the mortgage, we're looking at more about that later. It's a mortgage that it's risky for a lender to hold. It's subprime and so normally, in a market system, subprime mortgages charge higher interest rates than regular mortgages precisely because they have to bear the cost of that risk. So that's what a subprime mortgage is. Now, securitization of subprime mortgages, what securitization of subprime mortgages refers to is the following, (coughs) if I had wanted to buy a house, say, in the 1970s, I would have gone down to the local bank, they would've said, "You want a mortgage? "Well, give us your credit history," I would've had an interview with the bank manager and I would have had to put 20% down on the house, the bank would have, probably the bank manager would have known me, it would have been, I lived in Guilford, that would be in the Guilford Savings Bank as it in fact was, but they would run their credit history, they would become comfortable with lending me the money, I would take out a 30-year mortgage and they would then hold that mortgage and service that mortgage and I would spend the next several decades paying off that mortgage back to the Guilford Savings Bank, that is basically how mortgages used to work. Securitization is the process whereby banks not only sell, once they give the mortgage, they sell it to another bank or another financial entity and those mortgages then become chopped up and sold off in pieces and traded. So they start, they become securities. And it's important to understand that the securitization of mortgages actually begins with the federal government. Fannie Mae, Fannie Mae is so-called government sponsored entity. It was established in basically the 1938 as part of the second new deal in the wake of the depression, more than 25% of America's home owners had their homes foreclosed on in the depression, and it was extremely difficult for people to get mortgages, to buy homes. And so the Roosevelt administration wanted to make it easier. And so what the Freddie, Fannie Mae quickly got involved in doing was essentially two things. One was buying up mortgages from banks, and secondly, creating liquidity, because if you think about, if Guilford Savings Bank is holding that mortgage for 30 years, they can't use that, they can't give another mortgage with that same money, but if they sell that mortgage to Fannie Mae, they can then use the money to give out more mortgages. So Fannie, starting way back in the 30s, but much more in recent decades began buying up mortgages to create liquidity in the mortgage market. And then they also wanted more money with which to do this. So they began the process of securitization. Fannie Mae actually started chopping these mortgages up and selling them and then they would become traded as securities. And it's, it was a government entity, but eventually it became partially privatized. So you can trade Fannie Mae stocks on the stock exchange now, it's a public-private partnership, if you like, and that is part of the source of its problems. But the securitization of mortgages essentially got going because Fannie started selling them off and then they would be resold and repackaged and resold and repackaged, and more and more finely sliced. And they got the great book about this is "The Big Short" which some of you might have read, but people had very little understanding about internal composition of these buckets of mortgages that were being traded. Another thing that nobody was aware of at the time but became obvious in hindsight was people had generally thought of real estate as a hedge against the stock market, because you don't wanna have all your eggs in one basket. But the more you had the securitization and trading of these mortgages on Wall Street, the more actually they were operating in tandem, and so when the real estate markets crashed, so did the stock market. The government started to do this, but then private banks also started to do this, and the big investment banks got into their own securitization and indeed after 2002, 2003, 2004, when regulators were finally getting nervous and started pulling back the government's activity in securitization, that's when the private sector securitization ramped up. And so the good news was that more and more liquidity was being created, it was getting easier and easier to write these mortgages, the bad news was that there was more and more vulnerability in the system and people started to lose the incentive to conduct the sort of due diligence that you would otherwise conduct in giving out a loan. If the Guilford Savings Bank is gonna give me a loan, they need to be confident that I'm gonna be able to pay that loan back. If the Guilford Savings Bank is gonna sell the loan they give me to somebody else the next day, then it doesn't matter to them whether two years from now I'm not gonna be able to pay my mortgage. So they don't have the same incentive to do the due diligence in giving out the loans. And it's not just in giving out the loans. If you think about appraisals, if the Guilford Savings Bank is giving me a loan, they wanna be sure that when they're told the house is worth $150,000, it is worth $150,000. If they're gonna sell that loan, it doesn't really matter, and so you started to get not only the phenomenon of drive-by appraisals, but what the appraisal company soon discovered was that if they didn't give the appraisal, that the borrower and lender wanted in order to facilitate the transaction, they would just go to another appraisal company. And so the appraisal company's also were complicit, and indeed, I can remember refinancing a mortgage in about 2006, 2007, I wanted to borrow out the equity to do this and that, and I needed the appraisal to be some number, I've forgotten what it was, $850,000, and we got the drive-by appraisal, it said it was worth $900,000, and I remember sitting there in signing the papers looking around the room, the guy from the bank, from the attorney, and so on, I thought, nobody in this room believes this house is worth $850,000. And indeed, it turned out later it was not. (students laughing) But there it was, nobody had the incentive to do the kind of due diligence. So the securitization of subprime mortgages is, and we'll see why we got more and more subprime mortgages in a minute, but there were a lot of these very high risk mortgages that were being securitized and being traded and there was nobody with enough so-called skin in the game to have the incentive to do their own due diligence. The third component of the crisis was what we talked about last time, the behavior of the banks, the relentless drive to deregulate them and make it easier for investment banks in particular to get involved in this activity. But the fourth and most important in many ways piece that I wanna emphasize was politics. And so the idea that the government should get involved in addressing soft apartheid goes back at least to the 1980s. This is Jack Kemp, quarterback for the Buffalo Bills. He was actually in 1996, Bob Dole's running mate in his failed attempt to unseat President Clinton, but before that, sorry, after that, before that, 1989, he was Secretary of Housing and Human Development in the George Herbert Walker Bush administration and he called himself a bleeding heart conservative, and he came up with this idea of HOPE programs, Home Ownership for People Everywhere is what HOPE stood for, Home Ownership for People Everywhere. And he had this idea that the government should sell off, this was like Margaret Thatcher's privatization of council houses in London. The government should sell off all its public housing to the owners and he was beating the drum to do this, he actually ran into huge opposition within the Bush administration from people like Darman who's managing the budget, and Congress was not very excited about funding it and it turned out that just to get one of these government owned houses sold was gonna cost close to $100,000, so it kind of crashed and burn but it didn't really work as a policy. But Bush did it about whether to get behind it, then we have already seen that in the Clinton administration, there was a lot of effort to get behind this. And it didn't just come from Clinton, it came from the left of the Democratic Party and from minorities in the Democratic Party. This is Barney Frank, as in Dodd-Frank Frank. He was considered a very liberal congressman and he was a huge champion for saying to Fannie Mae that you have to put pressure on these private loan originators to give mortgages to low income neighborhoods, to minority neighborhoods. Maxine Waters, congresswoman from California was part of this same group of people, and so essentially, what happened was that Congress started putting pressure on Fannie and Freddie Mac, which is a, Freddie Mac is a younger and smaller Fannie Mae created in 1970 to have some competition in this market, but essentially, Congress, under pressure from these kinds of people in Congress was saying that you have only to make, you have only to make Fannie Mae funds available in subprime mortgage markets. That is what bankers would call subprime mortgage markets in minority neighborhoods, in poorer neighborhoods, and then came relentless pressure to make the subprime more and more sub, so reducing down payments from 20% to 10% to 5%, eventually to next to nothing. And again, using non-traditional criteria for granting people mortgages in terms of evaluations of their credit and so on. And when people say, oh, so it's actually the government that caused the financial crisis, which conservative sometimes argue, this is what they're talking about. Because they're essentially saying that these practices by Fannie and Freddie distorted the whole mortgage market because they, after all, even though they are publicly traded, are underwritten by the taxpayer. So they can take more risk than they would if they were not underwritten by the taxpayer. But that means that private sector lenders have to do the same thing if they're gonna be competitive. So the argument gets made that these practices distorted the entire mortgage market and create and even as the government started to get out of the business of securitizing subprime mortgages in the new millennium, nonetheless, they had already created this world in which the only way in which lenders could compete was by giving more and more risky loans and further because of the process of securitization, everybody underprice the risk because everybody was passing them on. It's one of the reasons I think Alan Blinder called his book that I was referring to last time "After the Music Stopped," you don't find out who doesn't have a chair to sit on until the music actually stops. And so that is what they, that became the game. And there was enormous pressure in the Clinton administration and it continued into the George W. Bush administration as well. Again, pressure to reduce underwriting standards and so on. And it was heralded just as Clinton had heralded it and Bush had also heralded it as expanding the American dream into minority communities and making home ownership available on a more equitable basis. So let's talk about the aftermath when it all came crashing down for a minute. And I wanna have you listen to a colleague of mine in economics department for a few minutes by the name of John Geanakoplos. He is one of the most brilliant people at Yale. He's one of those people, if he's giving a paper, you just go. You know you'll learn something. He also happens to, he happens to have a hedge fund that invests in mortgage backed securities, so he knows whereof he speaks, and this is his post-mortem on the subprime mortgage crisis. - Some errors. So what we did right, I guess, is we averted a complete collapse of the banking sector, but that is also the source of what we did wrong, because we were so obsessed with the banks that we forgot what caused the crisis, which was the subprime mortgage market and the mortgage market in general, home owners and what would happen to home owners. So we've lost about four million or five million households to the crisis who've been thrown out of their houses, and another couple million more that are on the way out that still haven't finished with the foreclosure process. So altogether, they're gonna be something like six million or more, maybe even will get to seven million households that will have lost their houses. If you're thinking three people per household, maybe some people had multiple houses, so it might be a slight exaggeration, but we're talking about 20 million people lost their homes and who would've thought that America would ever have to deal with 20 million people losing their homes? And who would have had to deal with this long period of stagnation and growth, which I think is related to the housing crisis. So I think the biggest mistake we made was concentrating too much on the banks and recapitalizing the banks both during the crisis and still now and too little on helping the home owners, who I think were the heart of the crisis. So to give you an example, we have a lot of data now on what happen to people who lost their homes and what happen to the lenders. So a typical subprime loan might have been for say $160,000 and the house originally was worth a little more than $160,000. But during the crisis, housing prices fell quite dramatically and that kind of house conceivably could have fallen to $100,000. So think of yourself as a subprime borrower with a bad credit rating to begin with, that's why he is or she is a subprime borrower, who owes 160,000 on a house that's only worth 100,000 in a time period where the future looks very bleak for earning money, for example. It would be almost crazy for such a person to keep paying. And in fact, most of the people didn't continue paying for a while in 2008 and 2009. Six or, people now position six or 7% a month, not a year, six or 7% a month were defaulting on their mortgages. So every month, there'd be a new and different set of 6% or so who defaulted on their mortgages. So what we did in most cases was tell them they'd stop paying. The Obama administration might have lept in and said, "Oh, it must be because you don't have a job, "we'll try to reduce your interest "payments for a little while." But so, imagine again you're that person who still owes 160,000, instead of paying 8% a month, you now pay 4% a month. Your house is still worth only 100,000. So what happened was 50 to 80% of those people after about a year re-defaulted. So the plan of helping the homeowners by reducing their interest payments for a while till they got over the bad times was completely unsuccessful. They defaulted. And what happened to the lenders after the default? The homeowner didn't just walk away from his house, he sat there until he was thrown out, which took on average almost three years, during which time the homeowner didn't pay his mortgage, after he found he was gonna lose his house, he didn't pay his mortgage, he didn't pay his taxes, he didn't fix the house, and on the way out of the house, maybe some things got taken or the house got ruined on the way out. So we have the data now, the recovery in such a situation was in fact, on all the subprime loans was under 25%. So the lender would've gotten back 40,000 out of the 160. Had we forgiven some of the debt, had the lenders gotten together and reduce the debt from 160 to 90 and maybe even slightly subtle, more subtly said, "We'll reduce your debt to 90, "but in case the house goes back up in price, "we'll raise the debt, 50% of the increase in the house "we'll add to the debt again." Had they done that, they probably would've gotten their 90 because the home owner now could've seen that his house was worth more than the debt. At worse, the home owner could've sold the house and pay back the 90 and kept 10. And so the home owner would've been better off, the lenders would've been better off, because they would've gotten 90 instead of 40, and the economy would've been immeasurably better off, and by the way, the banks, would've been better off, because that's where they were losing all their money in those kinds of loans. So I think our biggest mistake was that we focused on the banking sector and trying to recapitalize it and find enough money for the banks, and we didn't focus on the home owners whose debts were actually owned by the banks and who were gonna determine the fate of the banks at the end anyway. And we would've done much better concentrating on the home owners. I think that is the major mistake, major policy blunder that the Obama administration made. And they made it not because it never occurred to them, because people like me, and I wasn't alone, went and talk to them and said, "This is what you should do." I even testified in Congress a few times. And it's not just academics like me who were saying things like that, it was the business people who were making the investments and who effectively owned the loans who also wanted to see the debt forgiven. This isn't some pie in the sky, academic idea, there were a number of hedge funds, these mortgages, they're not done individually. A bank may give the mortgage, but the banks then sells the mortgage into a pool, and it's the shareholders of the pool who are effectively lending the money, the bank is just the middle man. So a bunch of, but these bondholders, shareholders of the pools, they don't know each other, so it's very difficult for them to coordinate any activity. That's why we need the government to coordinate things that are difficult to coordinate otherwise. So a bunch of those hedge funds and other investors also testified in Congress that they would like to see the debt forgiven not because they were good guys but because it would help them as well as helping the economy. And we just weren't willing to be bold enough to listen to those voices and we didn't take, I think, the most important step we could have to make the ensuing recession less dramatic. - So why didn't the politicians take Geanakoplos's advice? He's saying it would've been an all around win. The banks, the home owners, the investors would all have been better off. Why didn't the government do that? Any, any thoughts about why? Yeah? - [Student] I just feel that everyone thought that they had a good thing going at that point, and they didn't want to put a monkey wrench in there. - Well, but this is when we were having mortgages defaulting at a very alarming rate. So that can't be the story. Any, yeah? - [Student] For me, this is akin to getting away with murder. So if you know that you can taken a loan and they'll cut it to half, then why pay the loan at all, 'cause you know you can get away with it. - So you're worried about the moral hazard problem. Well, let's hear elaboration of the moral hazard problem here. - I tell you what, I have an idea. The new administration's big on computers and technology, how about this, President, new administration, why don't you put up a website to have people vote on the internet as a referendum to see if we really wanna subsidize the loser's mortgages or would we like to at least buy cars and buy houses in foreclosure and give 'em to people that might have a chance to actually prosper down the road and reward people that could carry the water instead of drink the water. - Hey Rick, it's a novel idea. - Hey, Rick-- (others mumbling) They're like putty in your hands. Did you hear-- - No, they're not, Joe. They're not like putty in our hands. This is America, how many of you people wanna pay for your neighbor's mortgage that has an extra bathroom and can't pay their bills, raise their hand, hello, we all. (others booing) President Obama, are you listening? - How about we all stop paying our mortgage? It's a moral hazard. - Hey, Rick, how about the notion that Will reported that you can go down to 2% on the mortgage-- - You could go down to minus 2%, they can't afford the house. - And still have 40% not be able to do it, so why are they in the house? Why are we trying to keep 'em in the house? - I know Mr. Summers is a great economist, but boy, I love the answer to that one. - Jason, Jason-- - You're thinking of having a Chicago Tea Party in July, all you capitalists that wanna show up at the Lake Michigan, I'm gonna start organizing. (someone whistling) - What are you jumping-- - So that is so-called Rick Santelli's rant. He is a financial commentator, that was a morning on the Chicago Mercantile Exchange in February of 2009, the Obama administration had just taken office and it's a famous rant because you've heard him calling for a tea party in July and that is often credited with galvanizing what became the Tea Party movement, Rick Santelli's rant. Although some people dispute and say his influence has been exaggerated, but it's become emblematic of the scorched earth resistance to doing anything remotely like that. And so one argument he's making there is the moral hazard argument. Now, people like Geanakoplos and others, they weren't entirely ignorant of such things, and there were ways to address it, namely first of all, to point out to people that if you didn't write down these loans, say of your neighbor's home, your house has become even worse, even, it's gonna lose value, because it's gonna be next to a decrepit house. And secondly, the people who did get mortgage assistance, for example, what the government, as John Geanakoplos said, they would give them a lower interest rate for a while, it didn't work because of the underlying incentives didn't line up, but they were put through humiliating hell. It wasn't like you just filled out a form and said, now I get my mortgage reduced, or my interest rate reduced, you went through endless interviews with people that made you prove your need, this is like the sorts of humiliating things people have to go through to get out of government assistance. So the argument was that there are ways to manage the moral hazard, if you like, even if you can't make it entirely go away. But the Rick Santelli's rant, I think is instructive for other reasons, which is, go back to, remember our discussion of the Capuchin monkeys when I said that the researcher misinterpreted the result of his experiment when he said that the angry monkey was like a Wall Street protester, and it's really people make much more local comparisons, and it's the idea even if you're gonna be worse off because your house is gonna get less valuable being next to a dilapidated house, you make the local comparison and that becomes the source of your resentment. And so for this reason, the kind of proposal that Geanakoplos and others were pushing is dead on arrival. Even though people are gonna have to actually take an absolute hit for refusing this kind of assistance, it's not gonna possible to mobilize them behind that. And it might not surprise you to know that, you maybe can guess from his name that John Geanakoplos has over the last number of years be in acting for the Greek government in its dealings with the IMF and the European banks where he's made exactly the same argument. He said the Greeks cannot pay their debts and you should write the principal down because all we're doing by, they just keep rescheduling it into the future, they're not solving the problem. What they should be doing is saying for every euro you pay, we will forgive the principal you owe by two euros, or something like that. And it won't surprise you to know that he's got no further defending the Greek government than he got defending the home owners during the mortgage crisis. And that is because brilliant economist as he is, he's not paying enough attention to the politics of the situation and the resentment that would get generated if the government did do this on a significant scale. So even though it makes pristine economic sense and Geanakoplos is right that everybody would have been better off, it was something that was politically was dead on arrival. And that's scarcely surprising. It should also just be said that the banks were divided about this. They were not as united as that clip suggests. Some of them actually preferred the wholesale buyouts, bailouts of the banks rather than worrying about the home owners, partly because there would've been quite substantial transaction cost to doing it they would have had, because the mortgages are owned by many people and so how you're actually coordinate the writing down. At the end of the day, the really simple way to do it would be even harder politically, which namely the government would've just given the money to the home owners, written them a check, but again, I think, for reasons we're gonna go into next week, for government to give people checks is politically very difficult. There's also, was also some sunk cost fallacy, so this is related to loss aversion that I've talked to you about, which is if a lender writes down a debt, you're admitting you made a bad loan. Whereas if you hope against hope that the market's gonna come back, you will not have to own the fact that you made a bad loan. So you see this all the time, an economist will tell you if you're trying to decide whether or not to sell a stock, the only important question is what is it gonna be worth in the future if I don't sell it today. But nine of 10 people will focus on what did I pay for this stock. It's the wrong question. For an economist, it's a completely irrelevant question what I paid for the stock, but Kahneman inference, a psychologist who understands the human psychic cost of internalizing the fact that you've lost something means that you will irrationally ignore the data that you should be paying attention to, which is not what did I pay for it, but whether, what, if I don't sell it today, what will I be able to sell it for tomorrow. So it wasn't only the issue that Geanakoplos was speaking about, but certainly, that was a significant part of it. So what do we to make of all of this. So here, we've had this long history, as I said, the transition from hard to soft apartheid did not get rid of class-based and race-based discrimination in housing markets, it just made it more difficult to deal with. Here we had decades where success of administrations, Democratic, Republican, Democratic, Republican, sought to address this problem by first outlawing discrimination based on neighborhood and discrimination in who you would give mortgages to. This created all this pressure on Fannie Mae from the federal government to restructure mortgage markets to make them more friendly to minorities, but when we look at how this all played out, it's not a happy story, if you look here, the run up to the financial crisis here, you can see that African Americans and Hispanics received a great majority of the increase in subprime mortgages. And if you look into the aftermath of the crisis, when everything came crashing down, Blacks were 47% more likely to be facing foreclosure than Whites, Latinos were 45% more likely to be facing foreclosure. If you look at the loss of property, you see from 2007 to nine, nearly 8% of both African Americans and Latinos lost their homes to foreclosures, basically double the rate of Whites. African Americans were 80% more likely to lose their homes compared to similarly situated Whites and Latinos was 70% more, and if you look, more likely. And if you look at the recovery after the, that I ended with last time, you can see again that this is household wealth. By 2009, it was starting already to turn around for Whites but not so for African Americans. If you look at changes in home equity, you can see that in the run-up to the crisis, Whites represented there by the blue bar, their increase in equity slowed relative to Blacks, but in the crash, the Whites lost less and turned around more quickly, turned around more quickly than it did for African Americans. So not entirely a straightforward story. In fact, in 2016, right at the end of the Obama administration, they did start doing a little bit of what Geanakoplos had been advocating. Fannie and Freddie approved a one-time plan to reduce mortgage balances with unpaid principal balances of 250,000 or less, but 33,000 home owners were expected to qualify, many fewer than that actually took advantage of it. They also, in their settlements with banks, the government gave the banks some incentives to write down principal. In 2013, they gave JP Morgan Chase $1.15 credit towards its huge settlement for every dollar of loan forgiveness that they offered home owners. So they did, and they did a similar thing with Goldman. So they did do a little bit of this at the margin, but nothing on the scale that would have made a dent in the problem. And if you look at projections of the effect that this has had on the creation of wealth into the future, you can again see that this big wallop that African Americans have taken during the collapse is gonna cost them and their children for many decades into the future, I'm not sure projections this far out worth that much and so I wouldn't put too much stock by the absolute numbers, but the general trend is clear. If you look, this is the peak to the trough. You can see African Americans and Hispanics lost more value in their homes, and then the trough to the peak, they gained back less. And we see, if you look at home ownership rights by race, you can see that African Americans, again, the pink line on that graph have fallen to levels that hadn't existed for many decades. And this has produced a new rounds of outrage in people, you could be saying, here we go again. These are some headlines from newspapers last year complaining that Black home ownership is as low as it was when housing discrimination was legal and Black home ownership rates are where they were 50 years ago. The Urban Institute, a liberal think tank says, "We've made important progress, "but we can't claim to have vanished housing discrimination "and its pernicious effects." And so here we are again, we are starting to see pressure to make home ownership available to less advantaged groups. Now you might say, well why is home ownership so important, why is it centrally associated with the American dream? And anyone wanna take a shot at that? Why do we, why do people care about home ownership so much? Why is it important that people own a home? - [Student] Financial security. - Financial security is one answer. We'll see how much financial security it really gives them. - [Student] It's a major avenue for wealth creation. - Pardon? - Wealth creation. - Okay, so let's start with that. People think that home, it's a way of creating wealth because most people believe that the value of homes always goes up. People believe that, that's one reason, but Shiller's book "Irrational Exuberance", this is Robert Shiller, our colleague in the Economic Department, a Nobel Prize winner, he wrote a book, this book "Irrational Exuberance" in 2000 right before the dot com bubble burst and predicted the dot com bubble would burst. He then wrote a second edition in 2006 about the mortgage market saying they're gonna collapse, and they did. You might say, "Why don't people listen to Shiller?" I did hear one economist opine once the trouble with Shiller is that he's predicted nine of the last three recessions. (students laughing) But the most famous, the most famous graph in Shiller's book is this one where he points out that if you deflate nominal prices from, you control for inflation and you, basically the story is that the real value of homes does not change, they've been flagged for over a century. So people think it's gonna go up, you might say why would people think they're gonna go up, the price of nothing else just goes up, and people would have theories like, well, populations are growing and so on, but as Shiller points out, you've also got a factory and cost of building comes down, all sorts of things, but so it's just not true. It's just not true that assets keep, that homes, that real estate keeps appreciating. Goes up and down like other things and the actual deflated value by inflation is more or less flat over the last century. Another reason you might think it would lead to building up security is that people are forced to save. Americans are notoriously bad savers, and if you spend, every month, you're sending in your mortgage payment, you're paying, you're building up that nest egg, it's forced saving, so-called the third leg in the three legged stool that you're gonna need for retirement. The other being social security and your private pension. So you build up the nest egg. The trouble with this theory is that it's no longer true. And it's no longer true principally because of the 1986 Tax Reform. The 1986 Tax Reform was, tax simplification was the big issue in the 1986 Tax Reform law. So they were trying to, this is the Reagan administration, they were trying to greatly simplify the code and among other things, they sought to get rid of the home, deductions off your taxes for paying interest on debts. And so they were gonna get rid of all interest deductions. They rapidly ran into the buzzsaw, which is that the mortgage, the banking lobby is extremely powerful and the real estate industry's extremely powerful, and they were strongly opposed to getting rid of the deductibility of home interest mortgage deduction. But the '86 Act did get rid of credit card interest deduction. Used to be until '86, you could have, whatever interest you were paying on your Visa card was deductible, just like your mortgage interest is deductible. And so the '86 Act got rid of deduction for credit cards and not for home owners, for mortgages. Well, enter the HELOC, HELOC, you could all remember HELOCs, right. The home equity loans. So what people would then do is essentially borrow out the equity in their house, pay off their credit cards with it or spend it on vacations in the south of France or whatever it was, and then the HELOC would be deductible. And so what people started to do was to, if you like, the government gave them a big incentive to raid the nest egg, and so the net result of this was, after '86, people would regularly refinance their mortgages, borrow out the equity, and so in effect, there wasn't gonna be a, there isn't gonna be a nest egg there whenever you finally sell the house, people essentially were financing this. And of course, it should be said, going back to what I said right at the beginning of the course, and forward to what I'm gonna be talking about next week, that as wages stagnated and people were going from one income to two incomes to keep the same amount of money coming in the door, the pressure to finance current consumption out of one's home equity would get ever stronger. And this is the source of Rosner, he teamed up with Gretchen Morgenson to write that book "Reckless Endangerment" in 2011, that's the best political book on the politics of the financial crisis, I think, and certainly the best book on the mortgage crisis. He had predicted in that piece I posted in 2001 that this market was gonna become unsustainable. In that piece, the subtitle of which is called "Housing in the New Millennium" and the subtitle is "A Home Without Equity "Is Just a Mortgage with Debt." so it doesn't give you any appreciating asset. A source of security. Not in an era of permanent employment insecurity. So just, let's just listen to John Travolta for a minute. - How many of you work jobs that just pay the rent no matter how many hours you put in? I see. My momma worked jobs like that after my daddy died. I remember her coming home from work just bummed, weary, you know what I mean. And I know she wanted to play with me and ask me about school, but sometimes you're just too tired to do anything but heat up a TV dinner, blob out in front of tube. - You got that one right. - There you go. And I don't have to tell you how hard it is to be looking for work. Hey, I don't have to tell you anything about hard times. So you know what I'm gonna do? I'm gonna do something really outrageous. I'm gonna tell the truth. (audience applauding) I know, I know what you're thinking. You're thinking he must really be desperate to do that, but if you had to swallow enough garbage-- - You can say shit, we're X-rated. (audience laughing) - Yeah, me too, if you believe what you read in the paper. (audience laughing and applauding) All right, here's the truth. No politician can reopen this factory or bring back the shipyard jobs or make your union strong again. No politician can make it be the way it used to be because we're living in a new world now, a world without economic borders. A guy can push a button in New York and move a billion dollars in Tokyo in the blink of an eye, and in that world, muscle jobs go where muscle labor is cheap and that is not here. So if you wanna compete, you're gonna have to exercise a different set of muscles, the one between your ears. - So that is "Primary Colors," an anonymous book, it turned out later was written by Joe Klein about the Clinton candidacy. And the reason I put that up there is because of the contrast between Clinton, and he did actually say that, I couldn't find an actual clip of the New Hampshire Primary, so I took it from the movie, but it's a radical contradiction between his claim in the clip I showed you from 1995 saying he wants to expand home ownership to his argument in the New Hampshire Primary three years earlier that the era of long term permanent employment is gone. And if you look, this is data released earlier this year by the Bureau of Labor Statistics looking at baby boomers and the takeaway point here is that they change jobs 12 times during their lifetimes. And only the first five or so could be explained by summer college employment or something like that. So people are looking at long term employment insecurity, and if you're not gonna know whether and when you're gonna be in a position to service a mortgage, then a home without equity is indeed nothing more than a rental with debt. Okay, so next we will talk about backlash, 2016 and beyond. (slow music) |
Power_and_Politics_in_Todays_World | Lecture_7_Shifting_Goalposts_The_AntiTax_Movement.txt | - Let's start with The Daily Show. (fanfare) (audience applauding) - You're here! - Good to be with you. - The evil Grover Norquist. - Sweetie pie. - There you go, "Debacle." So listen, you've come up with the pledge. Here's where you seem to have stepped in it. You came up with this pledge that you don't want people to raise taxes ever and you got a bunch of Republicans to sign it, mostly Republicans to sign. It strikes me as people can come up with any pledge they want. The people seem like at fault for this if you don't feel like this is a good thing are the people who sign it. I can come up with any pledge. - There have been over time quite a number of questionnaires and pledges and things that somebody says you should make all these commitments. The reason why the pledge has been successful is that it's simple. They're not 12 moving parts. Remember some group in Iowa had like a 14 part pledge including how you spent your evenings and they wanted candidates to sign it. - That's my group. The how you spend your evenings pledge. - They couldn't get a lot of signers. The pledge is look, as long as I'm a congressman, or a senator, or president, I'm not gonna raise taxes. When I see a problem, I'm gonna reform government, I'm gonna spend less, I'm not gonna raise taxes in the pledge ever. - Now, what about the idea though that because it's so simple and so absolutist, that's it's almost impossible? Like if I had a pledge that said America can never go to war. - Mm-hmm. - And I got a bunch of people to sign it. People would think, well, that's unrealistic. It's too simplistic for this complex world. Do you worry that your pledge? I mean, Reagan raised taxes, I don't know, seven times, eight times? - Yeah, and George Washington lost the Battle of New York. That was not on purpose. I mean, Reagan didn't want to raise taxes. - Wait, Reagan didn't want to raise taxes? - No. - But he did it eight times? - No. - He had a congress. He had a congress with whom he raised taxes a number of times. When I talked to him about it, he thought the big tax increase was the '82 tax increase. - I talked to him about it too, Grover. You're not the only one who talked to him. He told me that he did it eight times. - The '82 tax increase which was the large one, he said was the biggest mistake of his presidency because, of course, remember he was promised, oh, if you raise taxes, we'll give you $3 of spending cuts. - But then he kept raising them. '82, '83, '85. I have a whole list here. - Yeah. - So would you have run somebody against him? - No. The good news is here. - Why not? - Well one, he hadn't signed the pledge, and we didn't have a base of support. - So that's Grover Norquist who turns out to be a pivotal player in the class today. We're gonna look at the anti-tax movement and our agenda is gonna be first to go back and look at the history of it. I think it is one of the two most successful social movements of the 20th century, the other being the civil rights movement in that it completely transformed the political economy and the landscape of politics perhaps irreversibly we'll see as we go along. So our agenda is we're gonna go back and look at the origins of the anti-tax movement. Then I'm gonna take a step back to bringing a little, again, some of the machinery of political science into our conversation and talk a little bit about referendum politics because the origins of the anti-tax movement have a great deal to do with that as we will see. We'll then look at 1994 as a pivot point in the anti-tax movement and in the transformation of American politics. We'll then zero in on a case study of the estate tax. The estate tax is the most progressive tax in the US tax code. It was put there in 1916 to fund World War I and it was repealed as part of the Bush tax cut of 2001 which we'll talk about in some detail. And it was repealled quite remarkably with significant support from Democrats in the House and the Senate despite the fact that all of it is paid by the top 2% of income earners and more than half of it is paid, at that time, was paid by the top half of 1%, people with estates in excess of 20 million dollars. And so we'll do something of a whodunit about how the estate tax was actually repealed in 2001. We'll then talk more about tax cuts and the republican coalition as it has evolved. This intersects somewhat with the lecture I gave a week ago on the resurgence of conservative parties across the West, but it has a distinctive American flavor. And then we'll end by considering how far the goal posts have actually shifted as a result of this social movement. But let's go back to the beginning and the origins of the anti-tax movement I think really go back to the early 1970s and particularly after the Nixon Administration had collapsed as a result of Watergate, there was a lot of soul searching among conservatives. The democrats had been in charge of Congress for decades and there were a number of so-called activist think tanks that either emerged or grew very dramatically, places like The Heritage Foundation, the Cato Institute, the American Enterprise Institute, people, wealthy individuals and organizations put huge amounts of money into these think tanks and the idea was not to come up with a strategy for winning the next election. The idea was rather to rethink the republican agenda for the medium run. And they put their money into think tanks quite self-consciously, people like Feulner and others, the Coke family. Their reasoning was that they really needed to change the ideological terrain. So this was three categories of interests, institutions, and ideas, and they really felt that the new deal great society consensus had to be tackled at the realm of the ideological justifications for it out there in the American polity. And they thought that trying to do this in the universities was hopeless because the universities were massively dominated by people to the left of center, by liberals, and so the thought was a much more efficient way to do this would be to do it through think tanks. And so they invested a lot of money into these activist think tanks that would help design what we're talking about today and their general revival of pro-market politics that we were talking about last week at home, and then also the aggressive agenda abroad that we're gonna be talking about later when we see what happened after 9/11. But they were really concerned with taking head on the consensus that had informed the new deal, the great society for decades. And there was a lot of research done in these think tanks and a lot of activism that came out of these think tanks. They were frankly, if you go to The Heritage Foundation in Washington, DC, and you look at the signs on the elevator, it's very clear what it is that their agenda is about. Small government, shrinking the welfare state, and fundamentally strengthening market forces. So that was one piece of it. In my book with Michael Graetz that I have on the syllabus, we have several chapters that go through that history. But then secondly, the other more specific source of the anti-tax movement goes back to Proposition 13 in California. Many people, this is again, be taking some people down memory lane and introducing others to some history about which they may not be fully informed. So let's go back to June of 1978 in California. (fanfare) - Good evening. Here in California in the primary tomorrow, people have the rare and no doubt pleasing opportunity to vote their taxes down, to tell the politicians that they will pay this much and no more. Proposition 13, it's called, and it's an absolute flat ban on how much property tax they can charge homeowners. It is thought to be about as much a vote against big, expensive, wasteful government as it is against the property tax. And it seems that nobody loves government anymore. Here's Don Harris. - By far the most controversial of the tax reform measures is Proposition 13. Simply put, this amendment would roll back the assessments on all property to the 1975, '76 levels. It will allow assessments to increase at the rate of only 2% per year and it would limit property taxes to 1% of that assessed value. In terms of taxes paid, that means the owner of a $60,000 home in Los Angeles would find his taxes reduced from about $1,400 a year to about $600 a year. Proposition 13 applies to all property, businesses as well as homes. Big companies would get big tax reductions. Pacific Telephone, for example, will receive a tax cut of about 130 million dollars if Proposition 13 passes. - So Proposition 13 promises huge tax cuts to all California tax payers. This is a so called ballot initiative which we think of as something like a referendum. You vote only on this question and the result was indeed, a landslide. Here is Governor Jerry Brown. Those of you who are younger here and remember Governor Jerry Brown, you don't know perhaps that he had an earlier tour as governor when he looked a little different than the Jerry Brown you might know. So here he is. - [Jerry] Over four million of our fellow citizens have sent a message to city hall, to Sacramento, and to all of us. The message is that the property tax must be sharply curtailed and that government spending, wherever it is, must be held in check. We must look forward to lean and frugal budgets. It's a great challenge and we will meet it. We must do everything possible to minimize the human hardship and maximize the total number of state jobs created in our economy. Proposition 13 takes place on July 1st. We have only three weeks to act, three weeks to decide multi-billion dollars of fiscal questions, to set a new direction for the 5,000 units of government throughout our state. It's time to put aside partisan differences, the vote represented democrats or republicans, people from the north and the south, old and young, and all parts of our wonderful state. We must follow three basic principles. No new state taxes. Voters have told us they want a tax cut. They don't want a shell gain. Number two, the state must share the burden. We must adopt a thoughtful, austere budget. Already, I have imposed a hiring freeze. There will be no new hiring and when someone leaves the state service, he or she will not be replaced. I will propose budget cuts of at least 300 million dollars. I appeal to you to rise above the partisan temptations to work together as a body to fashion a bill that keeps faith with our fiscal realities and with the mood and the philosophy of the people which we serve. As for business, business will reap savings on the order of three billion dollars. Many individual businesses will save tens of millions of dollars. These people, these corporate presidents, have a moral obligation to invest that money in California. We have a very monumental task. There is not much time. The people are waiting for us to act. - So there it was, actually the legislature had tried to head this off when they saw how powerful the movement was. They came up with an alternative called Proposition 8 which would have cut taxes somewhat more modestly, but the victory was overwhelming. It was a 2/3 vote and Jerry Brown, although he was a democrat, felt as you can see, compelled to go with the result and so it wasn't challenged in court or anything like that. They went forward. So the entrepreneur for this cause was actually an Orange County businessman by the name of Howard Jarvis whose picture you saw up there. And just to get a flavor of his worldview, here was an interview with him that was done the following day. - (mumbles) is filled with moochers and loafers right up to their ears. And they have a great idea. The object of the lot of them is to get the job and sit there until they get a paycheck. And in the meantime, they don't move in any direction. And if you think that isn't so, just go over there this morning to city council's office and you walk through 15 city councilmen's office. You'll see more people asleep and reading Playboy than you do in a hotel. You've got so many consumer protection agency in this state, and city, and county, that they're like a bunch of rabbits in a field and they multiply about as fast and they all go into different directions and the whole thing, it's a whole flop. They know about as much about consumer protection as I do about swimming. This is gonna be introduced pretty soon in Michigan, Oklahoma, New Hampshire, Connecticut, Massachusetts, and Pennsylvania that I know about. I might say that I get calls from the BBC. I might say that I got an invitation to go to Canada. They want to start it there. And the thing is, it's just sweeping the country and it's beyond my capacity to really comprehend it if you want to know the truth about it. - So there he was and he encapsulates this idea that Ronald Reagan would later articulate that government is the problem and we have to wage a frontal assault on it. And as he intimated with the invitations he was getting and the notion that this was being picked up in other states, this was indeed really the starting point of this anti-tax crusade that played itself out over the next several decades and in which we're still in the midst of. But I want to take a little step back now and think a little bit more analytically about the logic of what we call these initiatives, referendums, or voter initiatives on single issues. And just to go somewhere else for a minute, let's talk a little bit about Brexit because I think Brexit, there's a dimension of Brexit that really encapsulates the nature of referendum politics. So here's a paradox about Brexit. Britain had an election in 2015 and elected two parties to the House of Parliament, two of the two major parties. They were overwhelmingly pro-remain. 2/3 of both the Labor party elected in 2015 and the Conservative party, maybe not quite 2/3, but a very comfortable majority of the Conservative party were pro remaining in the European Union. David Cameron had made this promise to have a referendum because he was worried about UKIP, but nobody really believed that the referendum was gonna come out as it ultimately did. 2016, we had this referendum and the British public votes by 52 to 48 to leave the European Union and the Brexit process begins. But here's the more puzzling question. In 2017, exactly a year off the referendum, Britain had another general election and it elected a parliament that was actually more pro-remain than the 2015 parliament. And not only that, Theresa May, who by then had replaced David Cameron as prime minister who had voted for remain but was now tasked with negotiating Britain's departure from the European Union, her cabinet of 24 people, 17 of them were pro-remainers. So how can it be, how can it be that the British electorate on one year before this referendum and one year after this referendum elect a parliament and a government that's overwhelmingly pro-remain, but nonetheless, in the referendum, they vote to leave. Any thoughts about that? Are they schizophrenic? Anyone got a? We need the microphone. I've been seeing the videos of these lectures and nobody can hear the questioners, so yeah. Yes, sir. Take the microphone and talk. - [Male Speaker] They vote against their own self interest. - Well, but how does that explain? I'm not asking whether it was in their interest to leave or stay. Why would they vote one way in 2015, a different way in 2016, and a third way in 2017? Yeah? - [Male Speaker] I think when you look at a referendum-- - Take the microphone, yeah. - [Male Speaker] When we're looking at a referendum, it's always a single issue. It's easier to vote in a single issue and not look at the total ramifications of that decision, whereas in an electoral campaign, there are tons of issues that are affecting one another and that are on the ballot. Whereas here, I also compare to what happened in Quebec with the sovereignty issue. It's much easier to put on an easy one liner question and not think of the ramifications of that decision. - Why is it easier? - [Male Speaker] Because you're only taking into account facts that are directly linked to that one question and you're not taking into account the possible larger consequences of the yes or the no answer. - Okay, I think you've nailed it. So that's exactly what goes on in that when you think about what parties do when they're putting together platforms, they have to bundle lots of issues, right? And they have to bundle them in ways that will try and appeal to a broad swath of voters. And so in doing that, they will discount everything they propose by everything else they propose to try and create what they think is the best overall bundle. If you're only voting on one issue, you don't have to do that discounting, right? So just to give another example which I think will help us get back to the tax issue today. If you ask American voters in any poll would you like to get rid of a tax? 2/3 will say yes. Even the estate tax that almost nobody pays. At least 2/3 will say yes. But if you say, and some of the subsequent polling on the estate tax confirms this, would you like to get rid of the estate tax if doing that also meant getting rid of prescription drug benefits for senior citizens? Then they say no. Why? Because in the second case, they're discounting their preference for a tax cut by their preference for losing prescription drug benefits for seniors, right? And so what referendums with single issue politics does is it enables people to vote on questions without doing that discounting. So it's a little bit like having a child eat as much candy as they want without thinking about the stomach aches and the doctors bills coming later. So referendum politics, single issue voting, it sounds terrifically democratic, let the people decide every issue, but in fact, it's an invitation to people to engage in choosing things without considering the costs. So in the case of California, there would be downstream costs on local government. That actually when we talk about next time there were gonna be downstream costs on the California school systems and so on. Now, you might say, well, the activists, people like Jarvis don't care about that because their idea is starve the beast, starve the beast, but most of the people who are voting are instead thinking about one issue at a time. This is what I mentioned Kahneman and Tvwersky in connection with loss aversion last time. Another one of their great contributions is the idea of framing effects, that how you frame an issue has a lot to do with what people will say about it. So if you frame an issue saying would you like a tax cut? Should we get rid of the estate tax? Should we get rid of the death tax? And everybody says, 69% say yes. You can go and say 69% of the American people want to get rid of the death tax. But if you ask a question with prescription drug benefits, then a majority says no. So what public opinion actually is, it's not that clear, right? It depends how you frame the choices. It depends how you frame the issue. And so when you turn tax cutting into single issue politics, you're gonna be able to pick up the ball and run with it in a way that the anti-tax movement subsequently did. And as you saw in the Norquist interview, this would become the test for republican politicians that he created this group called Americans For Tax Reform that we're gonna hear more about in a few minutes that became enormously powerful in Washington and had meetings every Wednesday morning and everyone who was anyone in the lobbying world in Washington had to go to those meetings. And among the things that he did was create this pledge that he was, you could see him ambivalent about not wanting to be seen to criticize Ronald Reagan on The Daily Show, but in the years subsequent to Reagan, it became non-negotiable for republicans to support tax cuts. And so you can see here. - And I'm the one who will not raise taxes. My opponent now says he'll raise them as a last resort or a third resort, but when a politician talks like that, you know that's one resort he'll be checking into. (audience applauding) My opponent, my opponent won't rule out raising taxes. But I will and the congress will push me to raise taxes and I'll say no. And they'll push and I'll say no. And they'll push again and I'll say to them, "Read my lips, no new taxes." (audience applauding) - And that became the instrument of his undoing in 1991 when off the Gulf War, he was at the 99% popularity rating. The country was in a pretty significant fiscal crisis and he did indeed raise taxes. And that caused the anti-tax movement to turn on him led by Newt Gingrich who was at that time, the leader of the republican minority in congress. But it was Gingrich on this anti-tax platform who was to orchestrate the republican take over of the House of Representatives in 1994. And Gingrich declared open war on George Herbert Walker Bush as a traitor to the anti-tax cause and is widely believed by consultants to have been a major factor in his defeat by Bill Clinton in 1992, that he was so badly damaged. And Gingrich who would then become the house speaker had something called a contract with America which was essentially the manifesto of the anti new deal coalition on taxes. It was going to require 2/3 votes in both houses to raise taxes. Every republican was now required to sign the anti-tax pledge, pledging that you wouldn't raise taxes. And it really was quite stunning when you think about the sweep of American history that from 1932 until 1994, congress had been controlled by, the house had been controlled by democrats for all but four of those 62 years, 58 years, and the senate had been controlled by democrats for all but 10 of those 62 years. So the democrats believed they were the natural party of government and seemed to be the natural party of government. You see the presidents are in the middle there. So you can see from time to time, there were republican presidents, but they basically had to govern with democratic congresses. Now, if you look at what 1994 ushers in, it's a completely different world. The republicans control the house for much of that time, all but six years from then until now and they control the senate also for the bulk of the time. So this really is, 1994 is a pivot point. It is a sea change. And the glue that held this crusade together was the anti-tax movement. And I want to now zero in on the repeal of the estate tax because as I said at the beginning, this is a tax that almost nobody paid. It was the most progressive tax in the code and yet it was repealed with substantial bipartisan support. So here's candidate Bush running for election and listen to him on the estate tax. - I will use this moment of opportunity to bring common sense and fairness to the tax code and I will act on principle, on principle, every family, every farmer and small business person should be free to pass on their life's work to those they love, so we will abolish the death tax. (audience applauding) - And interestingly, when you interview the people who ran that campaign, they said they were actually surprised that he got these vast applause lines for saying he'll abolish the death tax. Now, calling it the death tax, that was one of Frank Luntz's ideas actually. It's sometimes attributed to Norquist, but the truth is calling it the death tax increases the support from mid 60's to around 70% saying yes, get rid of it. It's not that big of a difference, but obviously, they decided that it had a certain ring to it and that why should you stop off at the tax man on the way to the undertaker? And much of the argument was arguments about small businesses, the vast majority of which didn't pay estate taxes. We'll come back to that. And then there were other arguments made like it's double taxation. People have already earned the money and paid taxes on it. Why should they pay taxes on it again? But of course, the tax code is filled with double, triple, quadruple, and quintuple taxation. Sales taxes are taxes on, you're paying taxes again on money that you paid taxes on before. So no more double taxation than many other taxes in the code, but they were surprised at how much resonance this got. Indeed, if you look at the contract form with America, at that time, the threshold below which you didn't have to pay estate taxes was $675,000. And all they proposed was putting it up a little bit, putting the threshold up a little bit. They didn't think that they could really get rid of this tax because they thought how would you get the coalition to do it? Nonetheless, in June of 2001, George W. Bush signed the Economic Growth and Tax Relief Reconciliation Act. I'm gonna talk a little bit about why it was called reconciliation in a little while. And it was one of the principle components of that act that it was gonna get rid of the estate tax. And just to give a sense of proportionality here, this was a bill that was gonna cost 1.3 trillion dollars over the next decade in terms of lost revenues. And as you can see, the bulk of it came from almost half of a billion dollars of income taxes, but then the other major pieces were a big increase in child tax credits and the estate tax and getting rid of the so-called marriage penalty. So it's gonna cost the federal government. It's not chump change in 2001 dollars. It was gonna cost the federal government 138 billion dollars over 10 years to repeal this tax. And a couple of other things worth pointing out about the repeal. In the American system, in order to pass a bill in the senate, you need 60 votes because the minority can filibuster with 100 senators but the minority can filibuster a bill and you can only end a filibuster if you have 60 votes. And the republicans didn't have 60 votes in the senate. And although they had some democrats who were willing to support this, in the end, they had eight senate democrats ended up supporting this, they didn't have 60 votes and they knew they weren't gonna have 60 votes. And so the way that you got around this, this is why it was called the word reconciliation appears in the title, was the so-called procedure of budget reconciliation which President Trump used again in 2017 to get his tax bill through. And the good news about reconciliation if you're trying to pass a bill is you only need a majority. So you only need 50 votes because the vice president breaks ties in the senate. So you get 50 votes, you've got 51 out of 100. That's the good news. The bad news is that if you do something through the budget reconciliation process, it has to balance over a 10 year period. And so the problem was how to do a massive tax cut and have the budget balance over 10 years. And in those days, the congressional budget office that scores this legislation, it was as it still is, it's manned by career professionals who use their own methods for scoring bills. And there was a big fight which has since been lost by the CBO as to whether or not to use so-called dynamic scoring of bills. So dynamic scoring of bills is informed by the idea, well, what effect will these tax cuts have on the economy? Because it's a standard belief among republicans that if you cut taxes, you'll get more growth, you'll get more revenue, and so there will be a dynamic effect. And these theories are always controversial and we'll see the predictions of the 2017, the dynamic predictions behind the 2017 act have gone the way of other dynamic predictions. But in any event, the CBO refused to do this. They refused to do at that time dynamic scoring. This is now changed. It tells you something about how far the goal posts have shifted. CBO will do dynamic scoring of bills. So the numbers, you couldn't do fuzzy math as George W. Bush used to call it. The CBO did the math and it had to balance. And so the way they got around this as you can see from here is that they phased it in. So they phased it in over 10 years and so what they did was they gradually raised the exemption. As I said, it had been $675,000 to a million dollars in 2002. You have to double all these numbers for a married couple, so two million for a married couple. And they gradually increased the exemption to 2009 to 3,500. They had to repeal. They repealed the bill, the tax completely in 2010, and then worse if it was not extended, it was gonna come back in 2011 as it would have gone back to 675,000. So some people call this bill the push granny off the bus in 2010 bill because if you died in 2010 as George Steinbrenner who was the owner of the New York Yankees had the good fortune to die in 2010, or his children had the good fortune that he died in 2010. There was no estate tax. But if you died in 2011, there was gonna be an estate tax again at a 56% rate with a $675,000 threshold. So you don't have to worry about the rest of it, but so you can see they gradually brought the tax rate down but it would have gone back up. So it's a kind of bizarre piece of legislation because nevermind throwing granny off the bus, people don't know when they're gonna die so how do you prepare? How do you do tax planning? Should you be buying insurance with respect to covering what you may or may not have to pay out of your estate? So they were basically banking on the idea that these tax cuts would be extended. And by the way, this was true. The whole bill. I'm just focusing now on the estate tax piece of it, but the whole bill would have to be extended because if it had been (mumbles) by reconciliation. So essentially, it was a tax cut with a sunset clause where the default of doing nothing would be to repeal the cuts in 10 years down the road. In politics, one year is 1,000 lifetimes. 10 years, who knows what's gonna be going on 10 years later? So that was the estate tax repeal and they felt they had to get a full repeal because that was part of their rallying cry. Get rid of the death tax even if it was only guaranteed initially for the year 2010. So how was it done? As I said, almost nobody paid this tax. How did they put together a coalition to do it? And it had to be a diverse coalition. So let's just listen again to our friend Grover. - [Grover] When Reagan came to Washington, he was almost the only Reaganite in town. When George W. Bush came in, there was a town full of Reaganites at the airport greeting him. And when he says, "I think we oughta do X," there are institutions and newsletters, and magazines, and think tanks, and whole trade associations that go, "Boy, yes," and we have that written up. - [Host] The people at Grover Norquist's Wednesday meetings all have different interests, but the glue that holds this coalition together is tax cuts. Norquist has been toiling away for tax cuts for a long time. He invented the no new taxes pledge that's now considered a political requirement for candidates in many parts of the country. He helped mastermind Newt Gingrich's contract with America and he helped popularize the republican's new name for the estate tax, the death tax. - [Stephen] When you think about Grover's influence on the conservative movement, you would really have to say that Grover Norquist is the Ralph Nader of the right. - [Host] Stephen Moore is the President of the Club for Growth which supports conservative republican candidates in primary challenges against other republicans it considers too moderate. - [Stephen] Right often fails because we tend to find ourselves in circular firing squads. And what Grover says quite correctly is look folks, at least if we're gonna be in a circular firing squad, let's put Dick Gephardt in the middle. And that's the kind of tactician that he is. - [Grover] One of the signs of maturity on the right is that you don't ask a president to win votes that's not there. My job is to get the death tax ready for repeal, to have the votes in the house and senate for repeal, and to make the world safe for a president to cheerfully sign it so that he's proud and happy to sign it. It's a political winner for him to sign it. The central question in American politics is the size of government. It's gonna get bigger or smaller. Of the resources that you and your family put together, more of it go to the state or more of it stay with you. - [Host] When the tax cut is enacted, Grover Norquist says, there will be much less money available to the government, fewer resources that the state can use to bother the members of what Norquist calls the leave us alone coalition. Grover Norquist's long term goal is simple and very radical. He wants to cut the size of government in half over the next 25 years. That's a serious and reasonable goal, he says, and one he's willing to advance with as blunt a metaphor as possible. - [Grover] I don't want to abolish government. I simply want to reduce it to the size where I could drag it into the bathroom and drown it in the bathtub. - Shrink government to the size where you can drown it in the bathtub, and this became a famous phrase of his. So he saw it as his job to work with the groups outside and put together the coalition to do this. And one of the things you can do with single issue politics is you can get all kinds of people who wouldn't agree on anything else, but will agree on this. And so this is what was done to create the coalition to get rid of the estate tax. Just to talk about a few people who were important players. This is Frank Blethen, the owner of The Seattle Times, a political liberal, and he poured a huge amount of effort and energy into this campaign to get rid of the estate tax. Why? Because he saw that the family owned newspapers were being gobbled up by the big corporate conglomerates, Gannett, and so forth. And he thought that if family owned newspapers couldn't be passed on to his next generation, it was gonna accelerate this process of getting rid of family owned newspapers. Bob Johnson, founder of Black Entertainment Television. He was on the liberal democrat. Gave a lot of money to liberal democrats and interestingly when the momentum was building to repeal the estate tax, Bill Gates and a few other millionaires published a letter in The New York Times and The Wall Street Journal and The Washington Post saying, "No, we should keep the estate tax." And they had about 20 or 30 wealthy individuals who signed up for this. And Johnson was so angry that he ran a counter ad and he got a counter ad calling for the repeal of the estate tax. And I went and interviewed him about that and I said, "Why did you do that?" And he said, "Well, doesn't matter to me. "I'm gonna be dead, "but for the first time in American history, "we have African American millionaires "and some African American billionaires. "If they have to sell their businesses "to pay the estate tax, "do you think that money "is gonna be reinvested in the Black community? "I don't." Then he said, you know, if you could exempt African Americans from paying the tax for 50 years, I might think differently, but when we finally get some wealthy African Americans, this is the time when we're gonna get rid of the estate tax? I don't think so. And then he said, "And besides, "Gates didn't call me. "He didn't call Oprah. "Who does he think he is?" So part of it was personal. But basically, he had a story that he told himself about why this was actually bad for African Americans. Congressman Albert Wynn, one of the strange things about this was the White House, not the White House, the democratic leadership on the hill were stunned that the Congressional Black Caucus came out in favor of getting rid of the estate tax. And when I interviewed Albert Wynn about why did you get, he was Maryland's fourth district, somewhat centrist, but certainly not a conservative democrat. Why did you support it? He said, "Well, we have lots of small business in Maryland. "My district has got the 10th largest number "of small businesses in America "and I was getting a lot of pressure "from small business constituents to support this." He didn't see any cost. And I think the interview with him was very revealing on this point about bundling because what the coalition to get this done would do is they would run all of these single issue polls, getting polls to get people to say they want to repeal the death tax, and then they would go into congress congressional offices and say to the congressmen, or the congresswomen, or the senator, "Look, this huge support for this. "You're not gonna lose your seat over this. "This is not a voting issue "for most people in a general election, "so it's a free vote. "It's a free vote to vote for it, "but it might not be a free vote to vote against it "because if you vote against it, "we could make your life difficult." So for instance, Senator Blanche Lincoln of Arkansas, a centrist democrat, she voted for this repeal and you interview her staff basically. They said she hated this, but here it is, it's a free vote if you do it. If you don't, who lives in Arkansas? Well, the Walton family live in Arkansas of Walmart. And the Walton family can make any sitting senator from Arkansas' life very difficult if they want to. And so why infuriate one of the most, the wealthiest, probably the wealthiest family in your state over a vote that's not gonna cost you anything electorally if you go the other way. So a lot of these people felt that sort of pressure on the one side and didn't see any cost on the other. Neil Abercrombie, the only congressman from Hawaii at that time. He's since been a three term governor. A liberal democrat, he supported it. A great puzzle to me. A very, very liberal democrat he was, at least at that time. When I went to interview him, I got a big part of the answer which was that his office was right next door to the office of the principle supporter of the bill and her name was Jennifer Dunn, a very smart republican congresswoman also from the west and since deceased, but her office was next door to his and she went to work on him and found out what he needed in his district. And so it did things for him and it was personal. He didn't see why he shouldn't support her. Again, there was not gonna be any electoral cost. So there were these kind of people and then there were arguments like we didn't have gay marriage at that time, as you know. This is 2001. And so gay groups came out in support of the repeal on the grounds that the estate tax discriminates against gay people because if you're a heterosexual married couple, you could double the exemption, but gay people couldn't double the exemption. So that's unfair to gay people. So gay groups came out in support of it. So it's a very cleverly orchestrated coalition of very strange bedfellows that nonetheless could be brought either to support this or not to oppose it. And then when you think about the opposition, the opposition to this bill. First of all, one thing to say about it is that they really didn't take it seriously until it was a long way toward happening. They didn't take it seriously because they thought it would be electorally unpopular and the republicans had been voting this up in the house in the last years of the Clinton Administration and the senate and there was always the security that President Clinton would veto it anyway if it passed, but they didn't take it that seriously until it was done. But where was the opposition? Where was organized labor? Well, we know from the last lecture, organized labor by this point is so far back on its heels that they don't really care about tax issues. I interviewed the lobbyist for the AFL-CIO in Washington and basically what he said was, "We have four lobbyists in Washington now. "One of them spends a quarter of his time "on all tax issues." He couldn't even remember what had been at stake. So it's a kind of thing that three or four decades earlier, organized labor would have been very active against, but they were much more involved in other issues at this time like getting the NLRB to be more friendly to card check and systems of organizing that might help reverse their decline. This just was not an issue for them. Non-profit groups, non-profit groups, Yale gets hundreds of millions of dollars every year partly because people want to avoid estate taxes. They don't want to give it to the government, so they give it to Yale. But the difficulty is it's very difficult to say, well, we should have a tax so that you'll have an incentive to avoid the tax in order to give it to us. It's not an easy argument to make. Insurance industry. There were at least four big insurance companies that stood to lose a huge amount from this because there are people who buy tax planning insurance from them. I interviewed one of their lobbyists and they said, "Well, "we're a very republican industry. "It's very difficult for us to oppose any tax cut, "and besides, we knew President Clinton would veto this." And I said, "Well, but there was an election coming up," and they said, "Well, we thought Gore would win." And I said, "And you sell insurance?" (audience laughing) So the insurance industry was largely mute. And then what about liberal democrats? At best, they were split. So you had this campaign which was extremely well organized, well resourced. It was funded by a tax planner called Patricia Saldarno. She was a tax planner who worked for very wealthy individuals and she came to Washington once a month and spent a week meeting with the coalition groups. It had very smart strategy and leadership. And particularly in managing conflicts within the coalition because there were conflicts of interest. You can tell that there must have been conflicts from the divided dollar game that we have talked about already. For example, the rates versus the threshold. Huge conflict because if you're a small business, you're gonna have an estate worth seven or eight million dollars, whether the threshold is two million dollars or four million dollars is huge, right, because if you're married, the four million dollar threshold is gonna be eight million dollars. You might not pay anything at all. So for small businesses, it's all about the threshold. If you're a billionaire, whether the threshold is two million or four million is chump change. What matters to you is the rate, right? If it's 56% or 40%, there's hundreds of millions of dollars at stake for you, right? So there are plenty of sources of potential conflict within the coalition. Another one is farmers versus small businesses. So they needed the small businesses to get it through the house and they needed to farmers to get it through the senate because the senate, farm states are massively overrepresented in the senate. And of course, when the democrats finally realized, when the democrats finally realized that this was serious, they tried to split the coalition. They tried to do exactly what you would do if you understand the divided dollar game. They came and offered the farmers an immediate permanent exemption which is better than they were gonna get in the bill 'cause the bill, we saw there's a phase in and then it sunsets in 2010 and who knows what's gonna be the case after that? So the democrats offered the farmers a better deal. You have a permanent exemption from the estate tax. Now, they were trying to split them off the coalition and some of the most difficult meetings of the coalition were over these sorts of fights. I interviewed a number of the members of this coalition and they gave two answers as to how they held it together and it was very interesting. One was they said, I must have interviewed maybe 15 or 20 people who had been in this coalition over a period of three or four months. And so at least five or six of them said, well, there were all these possibilities, but we realized that if we got off the idea of total repeal for at least one year, total repeal, which would then generate all the math to squash in the rates and the thresholds to get to total repeal within the 10 year window and meet the budget reconciliation target, if we got off the idea of total repeal, we wouldn't know what we could agree on. We knew we wouldn't be able to agree on anything else. And so it was as if they intuited the logic of the divided dollar game and that total repeal, we can all get behind. We can rally it. We've gotta repeal the death tax. We've gotta repeal the death tax. Once we get into other kinds of formulations of the problem, they sensed the whole thing would fall apart. So the rallying cry of total repeal was hugely important. But it wasn't just that, but it was also that they had a kind of moral narrative. So we're now in the realm of ideas, ideologies, right? Not just interests, ideologies. They had this idea that this is a morally repugnant tax. You're taxing people for dying and they've been taxed before and it's wrong. And multiple people I interviewed including I interviewed some of these farmers who were in the coalition. I said here, you were actually being offered a better deal, a predictable immediate exemption from the estate tax, permanent. Why wouldn't you take it? And they said things like this is a crusade. This is a moral cause. How could I look at myself in the mirror and know I had been bought off? How could I tell my children that I had sold out on this movement because we had been offered a better deal? And so there was this moral narrative that they used as a kind of way of heading off the splintering potential of divide a dollar. And this narrative, by the way, was not just used in the coalition. They used it in congress. They brought people to testify with horrific stories about how a family that had died and had had to pay the estate tax had had to sell things. And they brought famously a tree farmer from Mississippi called Chester Thigpen to testify before the House Ways and Means Committee. And Chester, you can read his testimony there. He says, "It took us half a century, "but Rosett and I have managed to turn our land "into a working tree farm "that has been a source of pride "and income for my entire family. "Our tree farm made it possible "to put our five children through college. "It made us possible to leave a legacy "that makes me very proud "because forests and ponds "that can live on for many, many years "after my wife and I pass on. "We also want to leave the tree farm in our family. "Right now, people tell me my tree farm "could be worth more than a million dollars. "All that value's tied up in land or trees. "We're not rich people. "My son and I do almost all the work on our land ourselves, "but my children might have to break up the tree farm "or sell off timber to pay the estate taxes when I die." He actually, he died the next year and he didn't owe any estate tax, but Chester Thigpen, there was Chester, you know? And so this was the moral face of the repeal and the ideas were important in sustaining this coalition as I said of heading off the impulse to break it. So let's just now zoom out a little bit and talk about tax cuts and the republican coalition and why this social movement as I'm saying is so important. And it's one of the recurring phrases you hear from consultants and academics alike is that tax cuts are the one issue that unites the entire republican party. And if you think about this, I'm gonna just put a little flesh on this. I said that when we think about the median voter and so forth, one reason why people don't vote as the median voter says they should could be that there's a second dimension like race, right? Remember when I put the, I had the second dimension up. So the tax cutting, it might seem not to be about money, but listen to Lee Atwater on this topic of tax cuts. Lee Atwater was a republican strategist who worked in the Reagan White House. (static buzzing) - [Lee] Here's how I would approach that issue as a statistician for political science, or no, as a psychologist, which I'm not, is how abstract you handle the race thing. In other words, you start out and now y'all are gonna quote me on this. All right, I won't do it. You start out in 1954, by saying, "Nigger, nigger, nigger." By 1968, you can't say nigger. That hurts you. It backfires, so you say stuff like forced busing, state's rights and all that stuff. And you're getting so abstract now, you're talking about cutting taxes and all of these things you're talking about are totally economic things. And the byproduct driving the media is blacks get hurt worse than whites. And subconsciously, maybe that is part of it. I'm not saying that, but I'm saying that if it is getting that abstract and that coded that we're doing away with the racial problem one way for the other. You following me? 'Cause obviously sitting around saying we want to cut taxes, we want to cut this, is much more abstract than even the busing thing and a hell of a lot more abstract than nigger, nigger, you know? So any way you look at it, race is coming on the back burner. - So that's a pretty candid admission of the proposition which the Edsel and Edsel book on the syllabus which became actually a kind of almost a manifesto for the 1992 Clinton campaign makes explicitly, namely that the tax cutting agenda is linked to the racial agenda because the notion that's pushed as in Ronald Reagan's welfare queens speech, the notion that is pushed is that most of this money is spent on the undeserving poor and probably undeserving poor who are not white. And we'll come back to this later when we talk about the kind of resentment that people like (mumbles) found in Louisiana when she spent a lot of time with poor white voters there. So the tax cutting agenda is not an alternative to the southern strategy, but actually becomes an expression of it. And then the second thing about the tax cutting agenda is that it heads off the divide a dollar game by giving something to everybody. So what's interesting about the Bush tax cut as you can see, you can look through these at your leisure when I post the slides, but the take away point is that everybody got a tax cut. Every rate was cut. So that is quite remarkable. It's a way of saying we don't have to worry about who's gonna split the coalition. Everybody's getting a tax cut, right? When you framed the issue that way, it tends to take the divided dollar problem away. So how far have the goal posts shifted? Well, in 2010, President Obama was now president and the country was still in recession after the financial crisis and the democrats controlled the whole of congress at that time, but it was very, very broad consensus that if you put in a tax increase in the middle of this recession, it was gonna be bad. So they extended the tax cuts for two years to 2012, by which time, of course, the democrats had lost the midterms and so there was republican control of Capitol Hill and there was a long negotiation, the result of which, basically you can see that it turns out to being a pretty good bet way back in 2001 because the bill that Obama signed in 2012 kept 82% of the Bush tax cuts that they're now permanent. And if you dig into the legislation, all of the income tax cut rates were kept except that the incomes above 400,000 go back up to the 39.6% rate that had prevailed before 2001. So the vast majority of the Bush tax cuts were kept and if you look at the estate tax, you can see that here in 2010, it's repealed. In 2011 and '12, it's put back at around five million. And then in the new legislation, it increases rapidly. So now, in other words, today, the exemption is 22.8 million for a married couple and their rate is a 40% rate as opposed to 55% rate which they got rid of in the 2001 bill. So it turned out not to have been a bad bet to do this wonky legislation through the budget reconciliation process. We might end by asking how effective has it been with respect to Grover Norquist's agenda of shrinking the size of government. The answer is not very. If you measure government by the number of employees, it looks like federal employees have remained flat. They've actually fallen some, but that's misleading for two reasons. One is that if you look at state and local government, there's been a lot of devolution of government to stay at the local levels. By employees, the government has grown. Not only that. As we will be talking about in the next lecture, a lot of what used to be done by the government has now been privatized and contracted out and doesn't show up as federal workers. But if you look at federal spending per capita, you can see comparing receipts to spending that the tax cuts have not starved the beast in any form that has led to a shrinking of government spending. Here, you can look at government spending. That's from 1969 to 2029, not 1069, but you can see that what you're seeing here is the trillion dollar deficits that are projected by the CBO for the next year. So what's happening is the republican's context, they're continuing to increase spending because that's what their constituents want them to do and that's the cost of unbundling. These pictures, you can look at at your leisure as I post them, but they basically tell the same story that whether you look at deficits or debt, there's a secular increases now going on and it's not partisan. So you can see here and in the Reagan years, the deficits go up, and in the Clinton years, they come down. In the Bush years, this is the Iraq War and so forth, they go up. And then this is the financial crisis. They've been coming down since the financial crisis, but now, Trump's agenda is driving them back up. These is just either pictures of the same thing with some details. that you can peruse at your leisure. And again, here you can see there's no partisan story. If you take the long look over American history since World War II which was a very high point of public debt, it came down under democratic control and then it's bounced around and it's now going up substantially under republican control. And it's not just that republicans spend money on guns and democrats spend money on butter. If you look at Medicare Part D which is prescription drugs, when republicans controlled all three branches of government in 2003, they put in Medicare Part D and it's a program which had not cost as much as the CBO projected at the time, but it's now costing a billion to a 100 billion dollars a year and will cost therefore a trillion over the next 10, over the next decade. And as for the Trump tax cut, as you can see, Steve Mnuchin predicted it would reduce the deficit by a trillion dollars, when in fact because of growth, it had a they reckon about a 0.07% effect on growth for a year and now it's gonna add 1.9 trillion dollars to the deficit over the next decade. And much of that in the red on the slides which again you can see is basically Trump's additional new spending which is approaching half a trillion dollars per year. So the idea that cutting off the supply of money to the government would be a way of shrinking government has not played out even though it's politically so potent. Instead, we've moved from a world of funding government through taxes to funding government through debt that will be paid by people who have yet to be born but they don't vote now. Okay, next we will start talking about privatizing government on next Tuesday. (calm electronic music) |
Power_and_Politics_in_Todays_World | Lecture_3_Advent_of_a_Unipolar_World_NATO_and_EU_Expansion.txt | - Today, we're gonna be talking about the international architecture of what I'm calling the early post-Cold War world, and by early, I'm really focusing on the time from the collapse of communism until the financial crisis of 2008 because we're gonna see that that ushered in pretty big structural changes, both within a lot of the countries we're gonna be talking about and in international political economy and in international relations. Our agenda, we're gonna talk about NATO expansion after the Cold War, we're gonna talk about the Washington Consensus and then we're gonna talk about the European Union, its enlargement and its challenges. So it's a pretty big menu. I do wanna take a minute to say something about three lenses for thinking about politics. One of the problems with political science is that people tend not to say in words of one syllable what they can say in words of five syllables. And so there's an awful lot of terminology and jargon, and it's my ambition to use as little of this as possible. But I do want to alert you to three different ways of thinking about politics, that, to some extent, compete and, to some extent, are complementary. The first is one that focuses on people's interests. You might think of Marxism as focusing on people's economic interests but also former rational choice models of politics that use economic approaches to politic, focus on people's individual interests. It's all interest-based. And we think about just common folk wisdom about politics, people expect people to do things that are in their interest. When we move into the realm of international relations, this interest-based way of looking at the world sometimes traffics under the title of realism, that, realists in international relations are people who say that countries follow their individual interests. Now, not every interest-based theory of international politics is realist because there are some who say, well, yeah but you've gotta look at the domestic politics of countries and how that influences what they do internationally. So they might still be interest-based but they're not gonna take countries as their basic unit of analysis. But the idea behind realism but trying to, all of these different schools I've just mentioned, is that if you want to understand what's gonna happen in politics, look at the interests of the relevant actors. Might have different theories about that but it's all an interest-based. This second basic lens that people bring to bear on politics is about institutions. And here, in domestic politics, it might be people who think independent courts and the separation of powers are important, that they structure what happens. Others think, no, it's the kinds of political parties, but institutional arrangements are important. When we think about the international system, institutionalists travel under various labels, again, they call them liberal institutionalists, some of them, but they have, sometimes they just call themselves institutionalists, plain and simple, but they look at international institutions, things like the United Nations. NATO is a kind of institution, it's an alliance, it's not an international institution, it's an alliance, but it it has an institutional presence. So this second lens focuses on institutional arrangements which may or may not be consistent with the way people's interests line up, right? So for instance, George Kennan, who I mentioned to you last time, thought the United Nations was a waste of time because countries always behave in their interests and if the UN told them to do something that wasn't in their interest, they would ignore it. So there, you can see possible tensions between institutionalists and interest-based accounts. Rosa Luxemburg is famous for saying, the rivers of history run through the most finely meshed statute. So again, it's in a Marxist view that interests are gonna prevail and institutional stuff is irrelevant. And then the third, I'm putting under the heading of ideals, and this can be ideas, culture, norms, things other than interests and institutions that affect what actually happens in politics. And again, people disagree a lot about, some hard boiled realist will say, norms and institutions are all beside the point. Others think norms are very important and that they actually structure what happens. Again, here, there's lots of fancy terminology to capture this notion, there's something called constructivism in international relations theory, it's basically a theory of normative behavior, norms shaping outcomes. And so, a lot of the squabbling in the academic journals and so on is among people who focus on interests, on institutions, or on ideas. And my own view, just putting my cards on the table before we dig in to today, is that, to say, which is the right one, is the wrong question. And it's better to try and understand what are the conditions under which interests tend to prevail or institutions tend to prevail or norms can re-structure things. And those conditions change. And there's often much more play, there's sometimes much more play for norms, say, than at other times, like when the Cold War collapses and the institutional architecture's up for grabs, then it might matter a lot what ideas are out there or not out there but in the middle of the Cold War, when everybody's locked into positions that are highly, highly rigid, then ideas are probably largely gonna be beside the point. So that's the way I will tend to use these notions and you will see them coming up, obviously, today but throughout the course. Okay. So let's think about the first post-Cold War international security crisis which was prompted by Saddam Hussein's invasion of Kuwait in late 1990. - Just two hours ago, allied air forces began an attack on military targets in Iraq and Kuwait. These attacks continue as I speak. Ground forces are not engaged. This conflict started August 2nd, when then dictator of Iraq invaded a small and helpless neighbor. Kuwait, a member of the Arab League, and a member of the United Nations was crushed, its people brutalized. Five months ago, Saddam Hussein started this cruel war against Kuwait. Tonight, the battle has been joined. This military action, taken in accord with United Nations resolutions and with the consent of the United States Congress, follows months of constant and virtually endless diplomatic activity on the part of the United Nations, the United States, and many, many other countries. Arab leaders sought what became known as an Arab solution only to conclude that Saddam Hussein was unwilling to leave Kuwait. Others traveled to Baghdad in a variety of efforts to restore peace and justice. Our Secretary of State, James Baker, held an historic meeting in Geneva only to be totally rebuffed. This past weekend, in a last ditch effort, the Secretary General of the United Nations went to the Middle East with peace in his heart, his second such mission, and he came back from Baghdad with no progress at all in getting Saddam Hussein to withdraw from Kuwait. Now, the 28 countries with forces in the Gulf area have exhausted all reasonable efforts to reach a peaceful resolution, have no choice but to drive Saddam from Kuwait by force. We will not fail. As I report to you, air attacks are underway against military targets in Iraq. We are determined to knock out Saddam Hussein's nuclear bomb potential. We will also destroy his chemical weapons facilities. Our objectives are clear; Saddam Hussein's forces will leave Kuwait, the legitimate government of Kuwait will be restored to its rightful place, and Kuwait will once again be free. Iraq will eventually comply with all relevant United Nations resolutions and then, when peace is restored, it is our hope that Iraq will live as a peaceful and cooperative member of the family of nations. I had hoped that when the United States Congress, in historic debate, took its resolute action, Saddam would realize he could not prevail and would move out of Kuwait in accord with the United Nation resolutions. He did not do that. Instead, he remained intransigent, certain that time was on his side. Saddam was warned over and over again to comply with the will of the United Nations, leave Kuwait or be driven out. Saddam has arrogantly rejected all warnings. Instead, he tried to make this a dispute between Iraq and the United States of America. Well, he failed. Tonight, 28 nations, countries from five continents, Europe and Asia, Africa and the Arab League, have forces in the Gulf area, standing shoulder to shoulder against Saddam Hussein. These countries had hoped the use of force could be avoided. Regrettably, we now believe that only force will make him leave. - So that was the first Gulf War, so-called Operation Desert Storm, as President Bush said there, it was it was authorized. This is resolution 678 which I will post and you can peruse at your leisure. I do wanna notice that, just notice that it was authorized by the Security Council with Cuba and Yemen voting against. They were temporary members of the Security Council. The way the security council works is, any one of the five permanent members can veto it, and China chose to abstain but it didn't veto it, and the Soviet Union supported it. So... I wanna, I start with this 'cause I wanna point out several things about what Bush did. He certainly was, the US was certainly not blameless in this whole affair, and I don't want to sugar coat every aspect of what President Bush did in this regard. He was heavily criticized later for having apparently signaled to Saddam Hussein that it would be okay with the US if they went into Kuwait, he was also criticized for encouraging a Shiite uprising in southern Kuwait and then not supporting it after we left, so there there are grounds for criticizing what we did but the things I wanna point out is, first of all, it was an action of last resort, every effort, as he said in the video, had been made to end this without an invasion. Secondly, it was proportional, stop the bully without becoming one. Bush took a lot of criticism for this. People said, you should go to Baghdad, you should knock off this regime, and he steadfastly refused to do it, partly because of the fact that he was not behaving unilaterally. As he indicated, there was genuine coalition including every Arab country except Jordan, was actually participating in this, and he knew that if he did more than the UN mandate, the Security Council resolution mandated, the coalition would collapse. So it was a broad-based coalition with strong regional support to stop the bully without becoming a bully, that is to say, this aggression will not tolerated. And this, as I said, was the first international security crisis of the post-Cold War world. And had that become the template going forward, we would be in a very different place today. Well, that's one of the things I'm going to be arguing to you later in the course because, as it turns out, it was one of the paths not taken to treat the way in which Saddam Hussein had been expelled from Kuwait as perhaps norm-setting for the future handling of international security crises, it was a path not taken. Okay, let's talk about NATO. What, when, and why? So NATO is a creature of the Cold War. And I've already mentioned the United Nations several times today, and it's important to think about NATO and the origins of NATO alongside the creation of the United Nations because they were created both in the 1940's. After World War I, Woodrow Wilson had wanted to create a League of Nations and that had failed, largely because the American Congress wouldn't go along with it and FDR had, during World War II, greatly invested in the idea that there must be an institution of this general sort created after World War II to prevent nations from going to war. And when President Truman came into office, he made it his business to make sure that that happened. And so here, you can see President Truman addressing the conference in San Francisco in April of 1945 as they were drawing up the charter of the UN. - There were many who doubted that agreement could ever be reached by these fifty countries differing so much in race and religion, in language and culture. But these differences were all forgotten in one unshakable unity of determination, to find a way to end wars. (audience applauding) If we had had this charter a few years ago, and above all, the will to use it, millions now dead would be alive. If we should falter in the future in our will to use it, millions now living will surely die. Well, there is a time for making plans, and there is a time for action. The time for action is here now! - And indeed, they did act, they did create the institutions, and just to elaborate what he said later on in that speech, he said, "The essence of our problem here "is to provide sensible machinery for the settlement "of disputes among nations. "Without this, peace cannot exist. "We can no longer permit any nation, or group of nations, "to attempt to settle their arguments "with bombs and bayonets." So he was foursquare behind the creation as the US was, unlike after World War I, the creation of the United Nations. But time moves on and four years later, we were looking at a very different reality. (triumphant music) - For us, war is not inevitable. We do not believe that there are blind tides of history which sweep men one way or another. In our own time, we've seen brave men overcome obstacles that seemed insurmountable and forces that seemed overwhelming. Men with courage and vision can still determine their own destiny. They can choose slavery our freedom, war or peace. I have no doubt which they will choose. The treaty we are signing here today is evidence of the path they will follow. If there is anything certain today, if there is anything inevitable in the future, it is the will of the people of the world for freedom and for peace. - So there was the creation of NATO, an alliance to face down what was seen as the Soviet threat. The previous month, just after the text of the proposed treaty had been released to the public, Secretary of State, Dean Acheson went on radio. And I tried to, I was gonna play his radio speech for you but the sound was too bad. But here, you can see him saying that, "The best deterrent to aggression is the certainty that immediate and effective countermeasures will be taken against those who violate the peace." And then probably the most quoted line he ever uttered was that, "If the free nations do not stand together, "they will fall one by one." So here, the US had moved from backing an international institution for solving conflicts to forming an alliance. And just what was this alliance? The most famous part of the NATO Charter is Article 5 which says that the parties agree that an armed attack on one of them in Europe or North America shall be considered an attack against all of them and consequently, that if such an armed attack occurs, each of them, in the exercise of the right of individual collective self-defense, recognized by Article 51 of the UN Charter, will assist the party or parties so attacked by taking forthwith, individually and in concert with other parties, such action as it deems necessary, including the use of armed force to restore and maintain the security of the North Atlantic area. And any such armed attack and all measures taken as a result thereof shall immediately be reported to the Security Council. And, this is an important kicker, such measures shall be terminated when the Security Council has taken the measures necessary to restore and maintain international peace and security. So here is an alliance created that says an attack on one is an attack on all. We are entitled to protect any member of that alliance as though we had been attacked, we will report it to the Security Council but we will not cease and desist until the Security Council has taken the measures necessary, essentially, to eliminate the threat. So not only had the US created this new alliance, but it clearly said that the UN, a realist might say this is exactly what you should expect, a realist would say that the UN is subordinate to the interest of NATO. Now, it tells you how hard boiled the realist, Kennan, was that he was also against NATO for two reasons. One is that he thought it would unnecessarily militarize the standoff with the Soviets, they would create a similar alliance which they indeed did a few years later, that's where the Warsaw Pact came from, was the Soviet response to NATO, but he also said it's stupid, because when the chips are down, countries follow their own interests, they're no more gonna be guided by membership of an alliance then they are gonna be guided by UN resolutions. Nonetheless, the US created NATO. The first thing to say about NATO is it's a historically unprecedented, highly unusual alliance. If you think about George Washington's Farewell Address to Congress in 1796, he said, and this was seen as a warning for the future, "It is our true policy to steer clear "of permanent alliance with any portion "of the foreign world." If you Google up things like international encumbrances, you can find American president after American presidents saying we will not commit ourselves to any encumbrance on a long term or permanent basis. And they all echo this philosophy that was once attributed to Lord Palmerston from a famous speech he made in the House of Commons in 1846 when he said, "We have no eternal allies, "we have no perpetual enemies. "Our interests," another hard boiled realist, "are eternal and perpetual, "and those interest it is our duty to follow." So American presidents have been guided, American administrations had ever formed any kind of permanent alliance, that essentially followed the Palmerston Washington view. And here, we have the creation of an alliance among the western powers, dedicated to protecting one another in perpetuity with armed force, if necessary, and subordinating international institutions to that purpose. Now, a lotta debate about whether NATO was successful during the Cold War. You could say, well, the US won the Cold War and some people say, you could say that we wouldn't have won it without NATO. That's one of the, we can't run the counterfactual, so it really is something of an imponderable, whether we could have won the Cold War without NATO. It is also worth noticing that NATO, during the Cold War, never actually went into battle. The first time a NATO operation occurred, which I'm gonna talk more about later in the course, was in Kosovo in 1999, after the Cold War over. And by the way, when Article 5 was not created because no NATO country was threatened. And that same thing is true with subsequent NATO actions in the post-Cold War period, such as the invasion of Libya in 2012, about which we'll also be spending time later. The partial exception was 9/11, although even then, the US, essentially, went to war more or less immediately after the attacks and NATO allies participated but NATO didn't assume full operational control in Afghanistan until several years later. So NATO after the Cold War. Some people such as French President Francois Mitterrand said, well, it's done its job, we should get rid of it, and was there to protect these countries against the Soviet Union, and it's now an alliance without a purpose. That was a difficult thing to do right at the end of the Cold War because of East Germany. And the thing about East Germany was it rapidly became clear after the wall came down that Germany was gonna be reunited, and it was just, unstoppable force to reunite Germany. And reuniting Germany made a lot of people nervous. Many people remembered World War II and some, not that few, also remembered World War I, and the idea of an independent Germany, outside of NATO, made a lot of Europeans nervous. And so the impetus to say the reunified Germany would be part of NATO came from the other European powers. And the Russians didn't like it. They, too, said, what's going on here, why is NATO starting to expand? And George Herbert Walker Bush and Helmut Kohl, who was then the German Chancellor, promised Gorbachev that NATO would not expand beyond including Germany and that'd been an artifact of the reunification. Gorbachev was less than impressed by their reasoning. (speaking in foreign language) - [Translator] And I'm not persuaded by the assurances that we hear, that Russia has nothing to worry about. You may not humiliate a nation, a people, and think that it'll have no consequences. So my question is, is this a new strategy? I feel that if the same kind of games continue to be played, if one country plays some card against the other country, then all of those problems, all of those issues that we've been mentioning today will be very difficult to resolve. Glasnost, free elections, political pluralism, the problems of concern to you. - So he was skeptical. He thought that this is not likely to be how things are gonna play out but he did, he saw this, as he says in that clip, he saw this as a humiliation of Russia, that was still the Soviet Union at that time. He saw this as a humiliation of the Soviet Union and was adamantly opposed to it. But history played out differently. So let me just give you a sense of what was coming now. (camera shutters clicking) (applauding) - Today, we welcome Hungary, Poland, the Czech Republic, finally erasing the boundary line the Cold War artificially imposed on the continent of Europe. Strengthening an alliance that now, clearly, is better preserved the keep the peace and preserve our security into the 21st century. For the 16 of us already in NATO, enlarging our alliance, our goal is to help to build a Europe that is undivided, free, democratic, at peace, and secure. - So there you have it. The first three east European countries included in NATO in 1999, and that, as you can see, is the first edition of countries besides, Germany's not on this list, since the end of the Cold War. So this didn't go down very well in Russia. As Gorbachev had predicted, it produced a sense of outrage and humiliation. So here, again, just to give you some, a couple of, a little flavor of it. (chanting in foreign language) (speaking in foreign language) - So a couple of comments about that clip. So one is, on the signs, one of the things that they're demonstrating about is that president Clinton had bombed Iraq for violating the no-fly zones, and they're saying, basically, only an idiot or worse would be bombing Iraq, the no-fly zones that had been set up following Saddam Hussein's invasion from Kuwait. This gentleman at the end of the clip is somebody by the name of Sergey Baburin. He was a Soviet Russian politician and subsequently became one of the leaders of a far right populist movement, and what he's saying at the end of that clip, is he's basically saying, we made a mistake once but the next generation are not gonna be making the same mistakes. And it's interesting that he said this in response to them demonstrating against the American bombing of Iraq, partly because, as I noticed in 1991 when I was in Moscow and I just told you about in an early lecture, the US Desert Storm Operation was hugely popular among Russians in 1991. In 1991, the US could do no wrong. In fact, in Moscow in 1991, in March of 1991, there were actually commercials for selling Desert Storm condoms. (audience laughing) Go figure. (professor laughing) I never quite could understand what it was trying to communicate but. There had been no resistance to the idea that Bush Sr. had gone in but, you know, seven years later, when these east European countries were starting to join in, to join NATO, things played out very differently. And it didn't stop there, it continued. (applauding) - [Woman] Yeah! - Thank you, all. - [Man] Thank you Mr. President! - You're welcome. Thank you all. Thank you all, good afternoon, and welcome to the White House. Today we proudly welcome Bulgaria, (audience cheering) Estonia, (audience cheering) Latvia, (audience cheering) Lithuania, (audience cheering) Romania, (audience cheering) Slovakia, and Slovenia. (audience cheering) We welcome them into the ranks of the North Atlantic Treaty Organization. (applauding) When NATO was founded, the people of these seven nations were captives to an empire. They endured bitter tyranny. They struggled for independence. They earned their freedom through courage and perseverance. And today, they stand with us as full and equal partners in this great alliance. (applauding) (pen thudding) (laughing) (applauding) This is a special moment in the hopeful story of human liberty as America formally declares its support for Albania and Croatia's entry into NATO. We strengthen America's partnership with nations that once found themselves in the shackles of communism. We rejoice in taking a major step toward welcoming the people of Albania and Croatia into the greatest alliance for freedom the world has ever known. - And I want to reaffirm as strongly as I can, the United States' commitment to honor Article 5 of the NATO treaty. No ally or adversary should ever question our determination on this point. It is the bedrock of the alliance, and an obligation that time will not erode. The NATO membership process, which requires applicants to make reforms across their political, economic, and defense sectors, has helped create the stable democratic Europe we see today. We were glad to see the alliance welcome Albania and Croatia last year, and there can be no question that NATO will continue to keep its door open to new members. - Montenegro's accession is good for Montenegro, it's good for NATO, it's good for the stability of the western Balkans, and it's good for international peace and security. Today is a historic day. So, President Vujanovic, welcome so much to NATO, it's a great honor to have you here, welcome. - With NATO membership, our future will be stable, secure and prosperous, and we will make decisions about the most important issues within the strongest, the most organized, and most efficient alliance in the history of mankind. - So as you can see, this had been a bipartisan story on the US side. Democrats and Republicans alike have led the expansion of NATO to include all of the former members of the Warsaw Pact except for Russia. And this actually doesn't tell the whole story because there were others who wanted to join NATO as well. So in 2008, for example, NATO announced that it would welcome the addition of Georgia and Ukraine to NATO. And that is part of what prompted the Russian incursion into Georgia in 2008, and also had a big impact in Ukrainian politics that contributed to the way things would play out in 2014. A couple of other things to notice about this accession, you might wonder why, in 2004, we added all of these countries to NATO. Why do you think, anyone have a thought? Why would the Bush administration, many east European countries wanted to join NATO from very early on but there had been resistance. I mean, after all, you know, you're committing yourself to go to war if one of these countries is attacked. That's not a light thing, right? You're actually, it's not like saying, well, let them join the European Union, if they don't like it, we'll kick 'em out, or they can leave. This is really committing you to defend these countries. Why would the US, anyone? (student mumbling) Pardon? - Deflection. (mumbles) Suppression in the Middle East. - It was related to what he was doing in the Middle East but it was not deflection. What had we just done in the Middle East in 2004? Yeah. - [Student] We had just invaded Iraq and then the former (mumbling) were not supportive (mumbling). - Bingo, exactly right. So what had happened was, after, and this is something we will return to later in the course, but once the Bush Jr., George W. Bush, had decided to invade Iraq in 2003, unlike what his father had done, he couldn't get Security Council resolution to authorize it and he had a lot of trouble putting together a coalition because all the Middle Eastern countries were not in favor of it and he needed the fig leaf of an international coalition to create, he thought, legitimacy for this action. And so if you look at all of these countries, except Slovenia, every one of these countries joined the coalition, nominally at least, and sent troops to Iraq as part of the US-led coalition of the willing. They joined the coalition of the willing and the quid for the quo was that he supported their accession to NATO the following year. Interestingly though, they all got rid of the draft in the meantime. Clearly, they were worried about the domestic politics of committing their youth to fighting in wars to defend NATO countries. So here, you see that geopolitics playing itself out in this way, that this is very much an interest-based story, that Bush is looking to get legitimation for his invasion, and so that became the impetus for getting these countries to join. It's also, one might think it's something of an anomaly, given the Trump administration's hostility to NATO in general, why they allowed the Montenegro thing to go forward. Of course, it'd been started much earlier but I think, as far as Trump is concerned, it doesn't really matter who's in NATO. Nonetheless, the question that Trump asks about NATO is not a crazy question, right? And it's the same question Francois Mitterrand was asking, what is this alliance there actually for? So there's some enduring questions about post-Cold War NATO which are gonna come back later in the course. One is, was Mitterrand right? Was there a missed opportunity at the end of the Cold War? Was Gorbachev right? Did NATO expansion make it more likely that someone like Putin would come to power in Russia? Gorbachev had said, humiliate people and there will be consequences. And certainly, the steady inclusion of NATO into the former, not only Eastern Europe, but as I said, they were running up trial balloons about incorporating the Ukraine and Georgia in 2008. There's a colorable case that this might have bolstered the nationalism of a figure like Putin. We'll come back to that, too. And what are the advantages and costs of having a permanent military alliance that now lacks any clear motivating purpose? It's led by the most powerful military on Earth, it's not clear what its goal is or is not. We will see later in the course that, when it has become active, it has had nothing to do with Article 5 of the NATO treaty, but it's rather taken on new missions and new rationales, mostly ad hoc, dreamed up on the fly, to suit the interests of different players at different times. Let's turn to the Washington Consensus, or what I'll say, taking neo-liberalism global. And I'll talk somewhat briefly about this 'cause I want to spend most of the rest of the time talking about the European Union. So the Washington Consensus, again, and what I call the Washington Consensus is essentially, a global version of what often gets called neo-liberalism. And I think of it, again, distilling it down and getting rid of all the jargon that plagues these discussions, as having three main features. One is deregulation, the second is free trade or trade agreements, and the third is privatization of state assets, previously held, state-held assets. Within countries, that tends to be called neo-liberalism. When the World Bank or the IMF or the Americans try to get other countries, as they did particularly up until the financial crisis, when the Washington Consensus, we'll see, started to lose some of its ideological power in the world, the core elements of the diet on which they would insist for countries to get US aid or aid from institutions controlled by the World Bank or the IMF. And it's difficult to overstate the confidence, the hubris, we might say in retrospect, the confidence with which this was viewed as the one size fits all approach to economic development. In 2004, in a speech to the the New York Fed, I think it was to the New York Fed, Ben Bernanke, who was then Chairman of the Federal Reserve, made a speech that you could, if you Google up his great moderation so-called speech, it's a speech in which he says, essentially, the technocrats running the fed are now sufficiently competent that we've, well, he didn't actually say this but this is what it's come to be interpreted as his having said. What he actually said was somewhat more nuanced but the takeaway lines from the speech are, to the effect that the fed has basically managed to, if not abolish the business cycle, manage it so successfully that the peaks and troughs could not be too damaging and we can, largely, have no inflation, full employment, and let the technocrats run it. As I said, what he actually said, if you go and read the whole speech, was a little more nuanced than that. But it became emblematic of the idea that this diet of economic deregulation, free trade, and privatization, informing all public policy, was the way to go, and that the world would be hunky-dory thereafter. Deregulation of Wall Street after the repeal of Glass-Steagal in 1999. So here, the Glass-Steagal Act had been enacted in the 1930's in the wake of the Depression to stop trading, to stop investment banks from engaging in investments on their own accounts with depositor's money. You either had to be an investment bank or a commercial bank. The banks didn't like it, they lobbied against it for decades and actually, by the time it was repealed in 1999, there wasn't much of it left. Read Ron Chernow's "The House of Morgan" if you want the blow-by-blow of the lobbying about that. Brilliant book among his many brilliant books. And so, this great debate now, about how much actually getting rid of Glass-Steagal contributed to the financial crisis, some say not at all, some say it was really all about real estate markets and had nothing to do with that, and some say a lot. But my point in mentioning it here is not whether repealing Glass-Steagal contributed to the financial crisis, I tend to think probably not, but it was emblematic of and ushered in a whole other set of steps of deregulation. So, for instance, in 2004 Hank Paulson, who was then Chairman of Goldman Sachs, led a group of the five biggest banks to Washington to lobby Christopher Cox, who was then the head of the SEC, not the sharpest knife in congress, to lobby him on the grounds that if a bank was sufficiently big, if it had five billion dollars in assets, it should be exempted from capital requirements or have reduced capital requirements because they were big enough to self-insure. The thought that because they were so big is maybe why they should have the capital requirements didn't seem to occur to him. And so the ones that were too big to fail were the ones that were subjected to less in the way of capital requirements. And we might say, go figure, how could anybody be so foolish? The quid for the quo, by the way, they said, we'll open our books to you. But the SEC at that time had a tiny office with, I think it was seven people working in it, ostensibly, to keep track of four trillion dollars worth of assets. So they weren't gonna be able to do this in any serious way, monitor the books of these companies. And indeed, months before the financial crisis, literally, two or three months before the financial crisis, Cox was reassuring congress that the big banks were fundamentally sound and then, turned around and Lehman Brothers went belly-up. But I'm not giving you as story about the causes of the financial crisis, but I'm giving you a story about the confidence people had in this deregulatory enterprise, that, just leave it all to the Fed, get rid of all the regulations, and everything will be fine. And so this was the new-liberal game at home and it was the Washington Consensus taken abroad. There was much more resistance to it in what we might think if as the Global South. This is just one of endless numbers of cartoons one could pick up put out, indicating that it was perceived as a kind of imperialist venture from the point of view of, particularly, developing countries that were having this diet imposed upon them. But in the second and first worlds, the hegemony of this idea was, it's hard to overstate. And if you, I always talking about Russia last time, if you look at US aid to Russia, it gives you some sense also, I think, as a kind of proxy for American influence over Russia. You can see that at the collapse of the Soviet Union in '92, '93, that you get huge amounts of US assistance to Russia. And then it goes down and then, in 1998, when they default, this is run Russia defaulted on its debt, I talked about that last time, and just bailed out by the IMF, you see it increase again and then, in the early 2000's, it goes down after the crisis wanes and then, here, and this is the period when Putin is basically implementing the Washington Consensus diet. This is when he's reforming the tax system, it's when he bringing tax rates down to 13% flat rate, balancing the books, building up a massive sovereign wealth fund or a rainy day fund. And so he's essentially getting rewarded for following what many would have regarded as pretty orthodox, pretty orthodox Washington Consensus policies and the best book about this is by a man by the name of Christopher Miller which I put on the suggested reading list, called "Putinomics". But then you can see, and this is for later in the course, what happens after the financial crisis. Not only does US aid go south, but so does US influence, and we'll talk more about that later. So let's talk about the European Union and its challenges. It's been around since early 1950's. The European Coal and Steel Community came into being as a byproduct of the Treaty of Paris signed by the original six; France, Italy, the Netherlands, West Germany, Belgium and Luxembourg. It became the European Economic Community as a result of the Rome Treaty of 1957. A much more ambitious treaty was enacted in 1992, the so-called Maastricht Treaty. And the most recent piece of legal architecture, which I 'll talk a little bit more about in a minute, was the the Lisbon Treaty of 2007. I'm not gonna give you the blow-by-blow of how all the 28 countries, how they got from six to 28, but you can Google it up for yourself and see when the different ones joined. What I want us to do is focus more on, what was the point of this? Why was a European Union thought to be necessary? What purpose would it serve? Yeah. - [Student] Economic and political union between European countries could stave off, like, war in the future. - To stave off war. So there were multiple motives but it's good that you're clairvoyant and got the first item on my slide 'cause what if you had come up with the third item on my slide? I would've been crossed off but you came out with the first item, so that's fantastic. So part of the impetus was, particularly people were worried about France and Germany going to war again. Not only the French and the Germans, Churchill, for example, was a big fan of the European Coal and Steel Community and he didn't want Britain to join but he wanted the Europeans to join, partly for that reason. So that was a beginning part of the initial impetus. This is for the political philosophers in the room, this is Montesquieu's idea of (speaks in foreign language), that if you have commercial interchange with a country, it's gonna be peace-generating, right? So that was one. Other reasons why people might have favored the creation? Yeah. - To reduce the pressure from the Soviet Union and (mumbles). - To reduce the pressure from the Soviet Union and? - America. - Okay so let's start, let's take the second, that this is less about the Soviet Union and more about US 'cause I want to keep you on my slide, here. (audience laughing) This was the big French motive. de Gaulle, for example, was a big proponent of the idea that what the Europeans needed to do was to band together to be a comparable force in the world to the US, right? Now, de Gaulle, I'm sure, as you all know, was a staunch nationalist, right? Staunch French nationalist, so he did not want a political union that could in any way erode French national sovereignty. He would be, I think, chuckling at the Brexit fight that's going on at the moment. But the way in which de Gaulle thought you could both have a strong, united Europe and protect the national sovereignty of France, which they guarded very jealously, they kicked NATO troops out of France, they bought their own, independent nuclear deterrent, they had a very jealous conception of their independence and sovereignty and they were not gonna be pushed around by the Americans. So they came up with the idea that any important changes in the union would have to be resolved by unanimity rule, and of course unanimity rule is a dictatorship of one, right? If you have unanimity rule, if you don't like the proposed change, you can stop it and block it. So de Gaulle was a strong proponent of unanimity rule and so this was this was the idea of a big block that could be of comparable economic clout in the world to the US. Any other reasons why people might have thought it important to have a big economic union of countries, European countries? - [Student] Trade. - Trade. You guys are amazing. (audience laughing) Wow, okay. Yes, so but here's the thing, it's complicated because it's, and this is part of, again, Trump administration's beef about the EU, it's an internal free-trade zone but it's externally protectionist, right, in many respects. You go and talk to African farmers about the tariffs that they experience trying to sell their goods in Europe. So in this sense, and this is one of the reasons Britain had such a difficult time before Britain finally joined under Edward Heath in the 1970's, they had been rejected twice because Britain wanted to preserve its imperial preference before they realized there was no empire left to prefer. But they had essentially had, they had had trading arrangements with members of the empire and then, when they created the Commonwealth, they wanted to have preferential treatment for trade from the Commonwealth and the European Union said, no, they couldn't have that and so the French, actually, de Gaulle was flatly against Britain joining the European Union and vetoed it and they didn't join until after he left power in 1969. So it was an internally free trade but externally, it was a protectionist trading block and that incidentally, we can fast forward, it's the story of TPP, that what the Trump administration failed to understand or at least to articulate if they did understand it, was that the TPP was not just a free-trade deal, it was about containing China from developing its influence in the Far East and it was largely motivated, actually, by a political agenda because China, for reasons we're gonna talk about later, had become much more muscular in the South China Sea and elsewhere during the course of the Obama administration. So the EU, we could say, is internal free-trade zone, externally protectionist. But actually, that's too simple because the sense in which it was internally a free-trade area is complicated because, as I put up here, there was a first mover advantage. And so if you think back to what I just said about de Gaulle insisting on unanimity rule, that meant that the things you could put in place early on would be very difficult to change because you would have a veto power. And so the French didn't agree to the British joining until they had put in the Common Agricultural Policy and the Common Fisheries Policies. And particularly, the Common Agricultural Policy was massive protection for French farmers, and was operated to the great disadvantage of British farmers. And so it was a free-trade zone with a difference. And an interesting irony of history here is that, when Britain joined the European Union in the '70's, the British parties were a polar opposite of what they are today about Europe. At that time, the Labor Party was badly divided over Europe and the Conservatives were largely in favor of Europe because Europe, the Tories thought, would be an instrument for disempowering British unions and deregulating the economy and making it function more efficiently. Margaret Thatcher, a lot of people don't know this, Margaret Thatcher was one of Ted Heath's most aggressive lieutenants in getting Britain into the European Union. She only became a Euro-skeptic later. Not only that, later, about halfway through her administration, Margaret Thatcher was the person who pushed the European Union toward qualified majority voting on things like reducing non-tariff barriers to trade because she could see. She had, earlier than before she did that, she had negotiated big rebates for Britain because of the unfairness of the Common Agricultural Policy and so on, so Britain was getting rebates but she wanted, in her privatizing agenda, which we're gonna be concerned with in a couple of weeks from now, she wanted more, what we call, neo-liberal policies, she wanted the reduction of non-tariff barriers to trade in Europe. It was completely obvious you couldn't get that with unanimity rule in the union because whoever stood to lose by getting rid of these barriers to trade would veto. So they went to qualified majority voting to make decisions in the European Union at the behest of Margaret Thatcher. So that's another reason that the European Union was created, there was this internal free-trade zone but with these wrinkles to it. A fourth reason, which is much less intuitively obvious, is that saw people saw this as a counterweight to some of the pressures of their own domestic politics. And I can give you a dramatic example of this that I ran into at a conference in Portugal about three years ago, when the rolling Euro crisis was going on and the southern European countries were in extremis. And a left wing German intellectual, sort of neo-Marxist by the name of Klaus Hoffa stood up, this was in Portugal, and he gave a speech in which he said, "You know, we Germans are really screwing "you southern Europeans. "You should insist on being "let out of the Euro because we're just forcing, "you should be able to devalue your currencies "and re-equilibrate your economies to get yourselves out "of this crisis and you can't because of us nasty Germans." So he thought he would get huge applause. And this Portuguese politician, I don't speak portuguese, she stood up and screamed at him. And evidently, he understood enough Portuguese to get the sense that she was angry with him, but not what he was saying. So he said, "I think you've misunderstood what I'm saying." And he said it again and she screamed at him even more. So then, we went for lunch and I was with a bunch of economists and lawyers and I said, "What was that about?" And the economists all said, well, the Portuguese government knows that, but for these constraints, the Portuguese government would be printing money hand over fist and would have massive hyperinflation. And the lawyers all said, you don't seem to understand. Portugal has never been able to be a democracy outside of the European Union, right? So these people saw it, actually, the fact that this was a constraint on Portugal was a good thing. And in a less dramatic fashion, even the Germans like some of the constraints. And then finally, of course, there was the idea of a federal European state. This is the idea most famously associated with Jacques Delors, of an ever closer union, it actually comes from a declaration in Stuttgart, signed in 1983, including by the British, by the way. The heads of state or government on the basis of an awareness of the common destiny and the wish to affirm the European identity, confirm their commitment to progress toward an ever closer union among the peoples and member states of the community. So five motivations there. What do they all share in common? What's common, I mean, it's a disparate list but there's one thing in common. Yeah. - [Student] They have focus on Europe. - Their focus on Europe. What's another thing in common? - [Student] A power in numbers? - Power in numbers, what's another thing? So they're all examples, there all from above, they're an elite project, if you think about it. They are the views of what elites think should happen in Europe and a symptom of that happens in 2005 when the Maastricht had created a constitution that was gonna integrate Europe even more, a much more close, ever closer union, and as you can see, France and the Netherlands held referendums in which it was voted down by the populations, the European constitution that had come out of Maastricht. And that quickly led the Czechs, the Danes, the Irish, the Polish, the Portuguese, and the UK governments all to cancel their referendums. And actually, the Lisbon Treaty is a step back from Maastricht because Maastricht was really trying to create a federal state and the Lisbon Treaty steps back to making it an intergovernment arrangement. Very much a top-down enterprise. And so the implications of this are one, and this is, the chapter and verse of this is in Tony Judt's book, "Postwar", which he published in 2005. He died shortly thereafter so he had a ready finished it before these referendums but the basic message of that book about Europe is look out when the first crisis hits because there's no grassroots support for this institution. And what happened when the first crisis did hit? We'll talk about later but the chapter and verse of that is in Adam Tooze's book, "Crashed", which is mainly a book about the Eurozone crisis and why the European governments have been unable to resolve it. And again, it's to do with the fact that this was an elite project, although he plays out the conflicts of interest within the elites. And then there's one final thing I will mention before we close, and to bring why have I been talking about NATO and the European Union. So NATO is actually a source of the obstacle to a Pan-European identity because they contract out their national security. If you think back to the beginning of the US, that we almost lost the War of Independence because the states wouldn't ante up soldiers and funds to fight the war. And the reason we created the Federal Constitution was to enable Washington to raise an army so that he could defend itself and pay its national debts. Because Europe contracts out, essentially, its national security to NATO there's no European army, there's no sense that Europeans might have to go and die for Europe in a way that, as a result of the transition from the Confederation to the United States of America, we have created a national entity with people actually identify. So there's an irony there that, actually, the strength of NATO can be an obstacle to the formation of a Pan-European identity. Okay, we're out of time and next, we will be talking about fusing capitalist economics with communist politics in China and Vietnam. (gentle music) |
Power_and_Politics_in_Todays_World | Lecture_25_Tough_Nuts_Education_and_Health_Insurance.txt | - So today we're going to talk about two what I'm calling tough nuts, education and healthcare, but I want to just begin by taking stock of where we were before we went on the break. - Right now, the human worker does $50,000 worth of work in a factory. That income is taxed. If the robot comes in to do the same thing, you'd think that we'd tax the robot at a similar level. (energetic music) What the world wants is to take this opportunity to make all the goods and services we have today and free up labor. Let us do a better job of reaching out to the elderly, having smaller class size, helping kids with special needs. All of those things where human empathy and understanding are still very, very unique and we still deal with an immense shortage of people to help out there. And so, if you can take the labor that used to do the thing automation replaces, and both financially and training-wise and fulfillment-wise, have that person go off and do these other things, you're net ahead. But you can't just give up that income tax, because that's part of how you've been funding that level of human workers. Some of it can come on the profits that are generated by the labor saving efficiency there. Some of it can come directly in some type of robot tax. You know, I don't think the robot companies are going to be outraged that there might be a tax. It's okay. - So, that was a fairly agreeable version of a tax proposal, which I'll come back to later. But this is part five of the course, focused on what is to be done, and the bumper sticker that, with apologies to Immanuel Kant, was that politics without, policy without politics is empty, but politics without policy is blind. And when we talked about policy without viable politics, I gave as examples Piketty's proposal for global wealth tax with no identified enforcement mechanisms and his more recent proposals for a European Parliament that would engage in redistributive politics and have trumping power over the existing European structures, even though the attempt to create such a parliament had failed because of popular hostility to it, and had instead resulted in the Lisbon Treaty. When we talked about politics without viable policies, politics without viable policies, I gave us the bumper sticker for that, the subprime mortgage crisis, where extremely politically effective coalitions had been put together across the aisle and with lobbyists and industry advocates, not to mention the advocates for housing for the poor, but it produced a train wreck of a policy. We then focused on the subject of insecurity, arguing that insecurity is in many respects, more important than inequality and certainly than global inequalities for people, and made the case that universal basic income, despite the apparent support for it across various ideological groups, that doesn't withstand scrutiny when you see that the conservative case for it would largely result in the abolition of what we think of as the welfare state, Social Security, Medicare and so on, so they would easily be blocking coalitions for that, and indeed when you look in Europe at, where they've held referendums on this, such as in Switzerland, there's not much support for it. We talked some about minimum wage and the difficulties of that, particularly the race to the bottom or the race out of the country problem that accompanies minimum wage politics, and instead argued for an expanded earned income tax credit, which goes back to the Nixon administration's notion of a negative income tax, and could become the basis for a more effective wage subsidy, and has been greatly expanded in recent decades under Republican and Democrat administrations. And there are also state-based earned income tax credits, which have the advantage of creating a kind of race to the top incentive in that they become magnets, states can make themselves magnets for employment because of the employment subsidy. Economists will still complain that this distorts labor markets, which it does, but any policy that attempts to affect employment at some level is going to affect labor markets, and we argue that this is a better way to do it. We then went on to the idea of universal adjustment assistance and I worked through the sort of tragic history of trade investment and insurance adjustment assistance, where labor had initially supported it as a quid pro quo for the 1962 Trade Act which reduced tariffs, but then it was largely not implemented by the Johnson administration that came into office in 1963. Nothing was paid out. In fact, over the next decade and by the time business got around to supporting it, labor had bailed out, and it had garnered a lot of opposition because by only supporting people who lose their jobs from trade, this seemed arbitrary from from the point of view of people who lost their jobs for other reasons, and indeed when the Reagan administration effectively killed off the program they made that same argument, so it was made both on the left and the right, and we argue, particularly given the sort of dynamic that Bill Gates is referring to, they have jobs going in any case increasingly to technology rather than to trade. There should be a Danish style, much more universal system of adjustment assistance that would provide for relocation, for retraining across a lifetime. And when we think about how might that sort of adjustment assistance be funded, as Gates, Gates perhaps is a little naive to think corporations will not abolish it, but in fact this is in the interest of business because one thing that economists will say, but it's not obvious to lay people, is that public education of any sort is actually a subsidy to business. It's a government subsidy to business by creating trained workers that businesses can then hire, and so government subsidies of retraining programs, again, is at the end of the day, a subsidy to business, and we make the case in in my forthcoming book with Michael Glass that business leaders should be getting behind proposed, doing what Gates is doing there, getting behind such proposals, because they will sort of, if you like, Tocqueville's self interest rightly understood, it really is in businesses' interest to support the creation of a perpetually re-educatable workforce that's going to be flexible in ways that are going to be required by the demands of the 21st century. So, today's agenda is to talk about two other challenges: K through 12 education and universal health insurance, and they're very vexed subjects in America, and they're vexed subjects partly because there are two other features of the American political order, which make it very extremely difficult to enact a lot of social policies and these are worse here or more challenging here than in many other countries. The first is fiscal federalism. We've referred to this earlier in the course, you might recall, when I talked about the deinstitutionalization of psychiatric patients in connection with rising prison populations and with the discovery of antidepressants and anti-psychotropic drugs to repress things like schizophrenia, when in the 1960s and 70s, this led to many people who had previously lived their whole lives in psychiatric hospitals to be released, but it was also a time of great fiscal crisis in many of the states, and so those savings were put to the bottom line of state budgets and instead those people wound up in American cities where there was not, there were no resources for them, and consequently, many of them then wound up back in the prison system. And so this is the kind of the dark side of fiscal federalism that we have a very decentralized federal system, unlike say the German federal system, where there are very clear demarcations, because the whole federal system was designed at once, of what's the risk, so there are clear demarcations of what is the responsibility of which level of government, whereas we have a kind of bottom up federalism that was created piecemeal, and power is not explicitly delegated to the federal government, reside with the states, some things are the province of cities and not the responsibilities of states, and so it's very easy for things to fall through the cracks and we will see as we talk about education and healthcare that this decentralized character of American federalism makes it extremely difficult to enact policies that in other countries would be done as a matter of national policy or what we would call federal policy here. The other is that the that the American system is much more subject to what some social scientists refer to as path dependence, and the idea of path dependence is if, as the name implies, as you start heading down one path, and maybe you take a left fork, then you're confronted with some other forks, but you no longer have the option of the right fork you would have taken, perhaps if you'd seen, later on, what were the consequences of going on the left fork. And so this is the idea of path dependence, things that are done earlier shaped the choices of things that are coming later. Now if you have a very centralized political system, that may not matter very much, so say in Britain where they have strong parties, I'll talk more about that on Thursday, they can do things like nationalized and denationalized and renationalized British rail. It's relatively simple because most political power resides in the House of Commons, so they can make decisions that are decisive, and they can revert. If the incentives change or new information arises, they can make a different policy. Our system is designed to be extremely difficult to change, and so there are multiple veto players in the system, as we've discussed in earlier classes with the bicameralism, with the separation of powers, with the presidentialism, with the federalism, with the Senate filibuster rules and with the over representation of small states in the Senate, with the court. All of these veto players mean that the choices made, once they're embedded, are very difficult to change, and we talked about this in connection also with embedding proximate goals, that you want to make your approximate goals difficult to change. And so FDR, in the case of Social Security, found a way to use the inertia in the system to protect Social Security. But the inertia is there, and it means that path dependence is a much bigger deal in the US system than in other systems. And so as we go through these two examples of K through 12 education and healthcare, you're going to see the fiscal federalism and the inertia that privileges path dependent popping up again and again and again and making, creating substantially bigger challenges for these sorts of policies in the US than we see elsewhere. So I asked you to read the Dale Russakoff's "The Prize" about New Jersey educational reform and let's just get our minds around that with a look at this video. - Republican Governor Chris Christie, Democratic Mayor of Newark, Cory Booker and the founder of Facebook, Mark Zuckerberg, are putting politics aside to help turn around the failing public schools in Newark, New Jersey. This is the first time all three of you really come together to talk about this big idea. So, Mayor Booker, for those who don't know, what's the big news? - Well, we've been talking for quite some time about creating a bold new paradigm for educational excellence in the country to show the way, to put the people of the city of Newark really in the driver's seat and the focal point, and to work to get all the assets and resources we need to give to them to succeed. - So, Governor Christie, what are you committing to? What are you committing to? - What I'm committed to is changing the schools in the city where I was born, I spent the first years of my life, and Mayor Booker is going to be the point person, our lead guy in Newark in helping to develop this entirely new plan of how to reform the education system in Newark and create a national model. I'm empowering him to do that. I'm in charge of the public schools in the city of Newark as Governor. I'm going to empower Mayor Booker to develop that plan and to implement it with a superintendent of schools that we're going to pick together. - I think that is so fantastic. (audience applauds) I think that is so fantastic. In this age of red states and blue states and everybody being so partisan about everything, the fact that you all would come together, did that ever enter into it, the politics of you're a Republican and he's a Democrat? - No, it's about the children. I mean, you go to Newark. (audience applauds) Go to Newark and you watch, as you saw in the movie, and there's other movies that display this about parents and children sitting in a gymnasium waiting to see if their number gets picked out of a barrel and watching parents sob, and I met a woman who got her son into one of the best charter schools in Newark, the Robert Treat Academy, and her son was now in the third grade at that point, and I said, what was it like the night you got him in it? She said when I was waiting that gym, I knew that whether his number got picked out was the difference between him going to college or going to jail. - [Oprah] Ah, yeah. - No mother in America should have to sit there and by chance decide that, and when I heard that, that's when I was US Attorney, when I heard that story, that turns my entire attitude towards putting this first, and he feels that way too and now we've got another partner with us to make it happen. - If we as a nation keep pulling left and right, we will never move forward, and. (audience applauds) Yeah. Most of us understand that. Most of us understand that. - So, Mr. Zuckerberg, what role are you playing in all of this? (room laughs) Are the rumors true? Will there be a check offered at some point? Yes. - Well yeah, I've committed to starting the Startup: Education Foundation, whose first project will be a $100 million challenge grant. - $100 million. (audience applauds) Yo yo yo yo yo. (audience cheers) (Oprah laughs) So okay, so why education and why Newark? - Why education? Because every child deserves a good education and right now that's not happening. (audience applauds) I mean, I've had a lot of opportunities in my life and a lot of that comes from, you know, having gone to really good schools and I just want to do what I can to make sure that everyone has those same opportunities. - And why Newark, though because, you know, what's your relationship with Newark? - Well, Newark is really just because I believe in these guys, right? (audience applauds) Running a company, the main thing that I have to do is find people who are going to be really great leaders and invest in them, and that's what we're doing here. We're setting up a $100 million challenge grant so that Mayor Booker and Governor Christie can have the flexibility that they need to implement new programs in Newark and really make a difference and turn Newark into a symbol of educational excellence for the whole nation. - So, that is not a good way to make social policy. It was a predictable train wreck and I asked you to read it for two reasons. One is the brilliant depiction of the extent of the problem in Newark. I mean, it is very, we we live increasingly in segmented worlds where we call it, we talked about soft apartheid earlier, but it's difficult to overstate the devastated neighborhoods in the school system, which had been taken over by the state some years before. The dropout rates and failure rates that she describes are just mind blowing. So, it is a very sobering description of the extent of the problem of public education in America's inner cities and I think Newark is certainly, in that respect, quite typical. But I also wanted to have you read it because it is so clearly designed to, if it had been designed to fail, this is how one would have designed it. It wasn't actually 100 million, it was 200 million, because it was a challenge grant and other foundations put up additional money. And there were variety of reasons that it was almost destined to fail. One was the top down and mandate quality of the whole effort. It was later revealed huge amounts of money were wasted on consultants, who paid themselves to write long reports. There was almost no buy-in from the parents of the children in the schools concerned, and the neighborhood-based funding model resists this kind of injection. $200 million sounds like a lot of money. The Newark schools' budget, of course, is billions and billions, so at most, these could be some kind of demonstration projects, they're not going to have a big impact. Very little of the money went into classrooms. Much of it went into administrative reform. And the other thing that they did was they brought in a school superintendent, who we'll hear more from, who was confronted with the difficulty that by creating impetus to increase charter schools, which are schools that are funded with taxpayer money, but are not subject to the same kinds of regulation, the issue becomes twofold. One, are the charter schools better? Which generates a lot of controversy, some are some are not, and on average, it's not fair that they are, but then secondly, what happens to the children who are left behind? And this is the ones who do not go into the charter schools. And so what you get is essentially an accentuation of differential options due to fewer and fewer possibilities of upward mobility. The emphasis on charter schools, Cami Anderson was the, she was, she had the added disadvantage that the population is overwhelmingly African American and she was the first white school superintendent brought in, but she, she correctly identified this as what she called a lifeboat theory of public education, that is you take some children out of the schools, the failing schools and you send them to the charter schools and it's a lifeboat for that child but then you have to think of, if it's a good charter school, but then you have to think about what is the likely effect on the children who remain behind? This is very comparable to the school voucher movement that gained a lot of traction in the 1990s and has now actually come back, we'll see, with Betsy DeVos and the Trump administration, but so again the notion here was that it's a Hirschman story that if you think about America's school system we, first of all we have private schools, so people who can afford private schools can leave, and that that means that, their relatively low exit cost mean that they don't have the incentive to stay and exercise voice to improve the system for the rest. And so, like the charter schools movement, the vouchers movement pressed by these two academics, John Chubb and Terry Moe, the idea was that the government should give vouchers to parents, and that they would create competition, first competition within the public schools, and so the competition, if you like, would would give poor people, and this was a sort of had a populist flavor to it, poor people the capacity to move to better schools and that would put pressure on the worst school to improve themselves or go out of business. And so this was the public schools' variant of it, and then increasingly there was pressure, also, to let these vouchers be used in private and parochial schools, and indeed, that was challenged in the courts, because if you, a lot of these schools were Catholic schools, religious schools, but in 2002 by a five to four decision, the court said it was okay. It was Rehnquist, I believe, writing the majority opinion. It was okay to use taxpayer money for vouchers, even if they were going to be spent in religious parochial schools, so that in effect it's violating the so-called wall of separation between church and state, so this was another way to try and diminish the costs being experienced by people in the system by introducing some kind of competition and better options for children in the public schools. But as Cami Anderson says at one point, she says she didn't come here to rearrange the deck chairs on the Titanic. And so basically what happened in Newark was that about 40% of the children wound up going to these newly improved charter schools chosen by lotteries, and the remaining 60% of the kids were in even more underfunded schools in the rest of the public system that's bred so much of the backlash, and what we have seen in the wake of the failure of that initiative is quite a substantial backlash against the school reform movement, both backlash against charters and against vouchers, both of which, predictably, are strongly opposed by teachers' unions and teachers' unions turn out to be one of the few relatively potent union movements left in America. So in Newark in 2014, Ras Baraka became Mayor after defeating another pro-charter candidate who went on to become president of the pro-charter Democrats for Education Reform, and he was reelected with over 77% of the vote four years later. In New York, this was a central plank of de Blasio's victory with 73% of the vote in 2013, a strident anti-charter campaign ushering in an era of feuding with the various proponents, and he was also reelected by a huge majority. Rahm Emanuel in Chicago, longtime former Chief of Staff of Barack Obama, longtime mayor of Chicago, also has run into massive backlash in Chicago over the Chicago charter schools. 2016, Hillary Clinton was, who had also been a longtime advocate of charter schools, was forced to back away and reverse her long standing support. And, indeed, in January of this year, there was a huge teacher strike in Los Angeles, and it was just when the candidates were announcing, you know, declaring interest in running for President, and all of the wannabe Democratic candidates quickly came out in strong support of this strike, a central pillar of which was to try and stop the expansion of charter schools in Los Angeles, and Cory Booker, who you saw in that video, had actually long been an advocate of charter schools, was very quick to announce his support for that strike. So the fight over vouchers continues in 2018. The Trump administration cut over three and a half billion out of the Department of Education's budget, and also devoted a billion to be spent on private school vouchers and other school choice programs, so there's an ongoing political battle, but it's extremely difficult to see a path forward for this kind of educational reform, which tends to be piecemeal, city by city, you know, high visibility initiative by high visibility initiative, but doesn't get at the systemic challenges of the decentralized system, the. By the way, efforts to get reform through the courts of the neighborhood-based funding system have also been rejected by the Supreme Court, so you can't go that route either, and it makes K through 12 education a real challenge and you can see that it's gotten worse. This is the divergence of educational opportunity over the 40 years, up to 2010. You can see, if you look at the top income quartile, number of children getting college degrees has doubled, but if you go to the next quartile, between 50 and 75%, it's grown, it's also doubled, but to a lower number, but in the bottom two quartiles, there's basically been no expansion of educational opportunity measured as getting a college degree, so no upward mobility for the bottom half of the population, despite the increase in upward mobility for people at the top. So, these local efforts largely fail. One counterexample example to this, by the way, is New Orleans where, after Hurricane Katrina pretty much wiped out the whole school system, they put in a system that is almost exclusively charter schools. And there are, again, contending debates about the merits of these schools. There's a, I think, a general perception that they're an improvement, but that is a very unusual situation where you have your entire school system wiped out by an exogenous shock and can redesign from scratch, and thereby defy some of the path dependencies that have been built into the system. Now, so this is the sort of pragmatic piecemeal solutions at the local level. We should talk a little bit about what efforts have been made at the national level where, as in most countries educational policy is decided, and there have been efforts. Probably the most notable one in recent decades was the 1965 Elementary and Secondary Education Act that was put in place by the Johnson administration, and they were able to do that first of all because of the huge landslide that they won in the 1964 elections, and it also gave Johnson his mandate for other elements of the Great Society and the Civil Rights and Voting Rights Acts that we have talked about. He also engaged in in very clever coalition building politics, thinking back to our lecture on building blocks. So for example, there had long been opposition from groups, parochial groups, Catholics and so on, for the federal government's giving money to local school districts because it wouldn't benefit Catholic students, and so he instead devised a plan where the money would actually go to the students and not to the schools, and so the argument was that the federal government was not supporting religious schools. This is before the Supreme Court decided, as it had in the voucher cases, that would come later. But then secondly, civil rights advocates and advocates of minority groups had also opposed giving money to segregated schools, because if you gave money to segregated schools, it was federal money to segregated schools, gonna reinforce probably, or so it was believed, the advantages achieved by children in predominantly white schools. With the Civil Rights Act and the Voting Rights Act passing, but particularly with the Civil Rights Act passing, those urban liberal groups got turned around and now became supporters. So, the Johnson administration was, it was a pretty big policy. It was a billion dollars which would be eight billion today, rising to 4.2 billion later in the decade, which would be $31 billion today, and literally a tenfold increase on what the federal government had been spending a decade earlier, so it was a really substantial federal program, but it soon came under Republican attack by 1968. Of course the Nixon administration had come on and it was very inefficiently administered. Much of it was spent on local school administration and not on classrooms, but also it was basically remedial education, and they ran into the problem that Russakoff also identifies in the New Jersey schools, that essentially trying to tackle underachieving children in the classroom is, it's too far downstream from the causes of the problem and so you have these programs that improve kids' performance over the summer, let's say, but over the school year, but then, over the summer, they tend to, most of the benefits dissipate and when they come back the following year, many of the apparent advantages have gone away and this had actually been identified by a sociologist name of Coleman in 1966 that remedial programs don't really work, and this is the same message that actually Zuckerberg took away from the failure in Newark that if you're really going to address the problems of inner city schools, you have to start with the homes and the neighborhoods and what people are eating and where they are sleeping and whether they're sleeping and the basic capacities that need to be developed cannot just be developed in classrooms. So, that didn't go well. The next big federal effort was in the 2001 reauthorization of the SCA. This was the reauthorization of the legislation that had been passed in 1965 and this was the so-called No Child Left Behind program put in place by the George W. Bush administration. I think I got the wrong date on there. It was actually 2002. The idea here was to identify certain benchmarks that had to be met in order to get federal funding. It was largely attacked. It was roundly attacked from all sides, partly because it was seen as much too top down, even though there were there were concessions like allowing each district or each state, at any rate, to identify the benchmarks that had to be met, but then this produced critiques to the effect that they were manipulating the benchmarks in order to show the results of improvement on the one hand, and that other districts, which were actually achieving absolute gains, but didn't meet the benchmarks, weren't getting the funds. And so it was again attacked from all sides, strongly resisted by teachers unions as teaching to the test, not providing genuine education and manipulation of the test and so on. Finally, what was left of it was pretty much put out of its misery by the Obama administration in 2015, where they turned over what was left of the program to the states. So, it's very difficult to see how you put together a national coalition to address K through 12 education, partly for all the reasons I've already identified, and the one final thing to say is that the coalition to demand it is not there because at the very high end, of course, the American system is extremely good, so people in private schools, people in wealthy neighborhoods where the public schools are very good have no real incentive in this highly decentralized, locally funded system. They have no incentive to get behind any kind of comprehensive reform. And, you know, I think I mentioned to you earlier in the course I used to live in Guilford, Connecticut, and I watched over a course of decades as the proportion of people living in the town with children decreased because of the graying of that population of that town, it became increasingly difficult to pass the town budget because there's a lot of genteel poverty in those towns, there's a lot of resentment of school teachers who have good unions and good jobs, and so people who maybe inherited a house but can't pay the taxes on it, they show up and vote down the town budget because half of it is the schools, and so you have a big divided dollar problem. In that case, eventually the Connecticut State Supreme Court ordered the raising of the mill rate on the grounds that there's a right to an education, Connecticut state constitutional right to an education, so they limited the damage in that way, But the combination of the fiscal federalism and the Hirschmanesque problem that if you decrease exit costs, which sounds good to give people other opportunities, you increase voice costs, and you make it less likely that the residual system is going to be improved. So, that is the rather depressing story of K through 12 education in America. Let's talk about health insurance. And here again it's in many respects, a depressing story that exhibits infirmities of the education system, though not as as depressing as the K through 12 education story and actually we have a proposal that might actually work. So thinking about the healthcare system, the US system is unusual in two respects for any advanced democracy. One is that there is not universal coverage, and the other is that we have these employer-based systems. Most people get their insurance through their work. 180 million Americans, at least, get their insurance through their work. Now you might say well, why did we get this system? And here's a good example of the path dependency issue that I was talking about. In 1942 during World War II, the government instituted wage controls. Employers were not allowed to give pay raises, but there were also great labor shortages, and so employers were looking for ways to compete for labor, for workers. And so the following year, the War Labor Board said, well, benefits would not count as wages, and so what happened was that employers started giving out health insurance as a way of increasing the real wage they were paying to workers without violating the stricture that the wage freeze that was the case had been put in place for the duration of the war. And so this meant massive increase by the end of the war in this this employer provided health insurance, and then of course, once you have it, the people who have it don't want to give it up, right? And so this is a very sort of micro but poignant illustration of how the path dependency for a system can get going and then accelerate. So we don't have a universal system partly for that reason, but there are other reasons as well. One is the employer-based I've already mentioned, but the other is Medicare, as I said to you a little while ago that Medicare was a very major achievement in 1965, single payer healthcare for people over the age of 65. The problem with doing that is that it creates a potential fissure in the coalition of people to demand universal health insurance, because the elderly now have what they need and the employed have what they need, and so those groups of people don't have much of an incentive to support expansion, particularly if you think about the elderly, because when we look at these Democratic plans on offer from the candidate, you know, Elizabeth Warren initially costed her Medicare For All at $20 trillion. Most of the independent evaluations are closer to double that and even she is now being forced to walk away from that. But if you think of the available pool of resources for health insurance as finite, expanding the coverage is going to come at the probably the cost of coverage for people who already have it. So, this patchwork of partial provision makes it harder to generate the demand for a universal system, unlike Britain for example, which put in a universal system in one fell swoop for everybody in 1947 under the Attlee government. So those are among the obstacles. And then Medicaid is stigmatized as a poor persons' program, and so not viewed as a desirable thing for people to want to have. And it's also caught up in fiscal federalism because it's funded with certain amounts of money that come from the federal governments to the states, but then they have to raise the difference in local taxation and so that creates huge amounts of pressure on the states to limit what Medicaid is going to cover because, particularly in recent decades as the federal government has reduced what it provides and has gone to systems of block grants, which again are not sensitive to changes in the populations that are actually being served. So that's another set of obstacles, and it's in this context that we have to think about the Obama legislation. And it's much pilloried and it's got many defects to it, but it's worth pausing just to note what it did in fact achieve, given the place we're starting from. It did increase coverage for about 20 million people. There are still about 32 million uninsured and 44 million what are called underinsured, and underinsured people are people who have partial coverage, but the copays or the conditions that are not covered put extreme fiscal financial stress on them, and so if you look at this chart you can see that the number of uninsured actually came down, not insignificantly as a result of the Obama law, and it should also be said that Obama was the fifth administration since the war to have tried to do universal healthcare. So many administrations have tried before, and the only success had been the partial creation of Medicare in 1965, so it's not an insubstantial achievement. But the way in which it was enacted tells you a lot about the pathologies of the American political system, and that's what I want to focus the rest of our time on. So, if you think about Obama coming into office in January of 2009, and this is one of the top issues on the agenda. After Dodd-Frank, this is the number one issue, much to the chagrin of environmentalists who wanted environmental stuff to be the number two issue, but Obama decided universal healthcare, he had campaigned on it, and he wanted to do it. And so, one issue that was central to trying to get this done was controlling costs. Now, medical costs are high for two reasons in America. If you plot the rate of inflation against the rate of medical inflation in the US, what you find is that the rate of medical inflation is sometimes three and often at least two times as the rate of regular inflation, so medical costs are going up faster than the rate of inflation. And you might say well, why is that? So one reason is the lack of universal coverage itself, because what it means is that people without coverage essentially use emergency room care for their basic medical provisions and that's about the most expensive possible way to give people healthcare, because the overhead involved in running an emergency room is enormous, and so then hospitals tend to pass on those costs to everybody else in the form of higher prices. So that's a big part of the reason that medical inflation is so high, and then a second piece of it is that the cost of drugs to the federal government is extremely expensive, because when Medicare Part D was created by the George W. Bush administration in the early part of the century, this is free prescription drug benefits for seniors, people on Medicare. They forewent the opportunity to negotiate drug prices for the federal government directly with the drug companies. If the federal government's going to be this big purchaser, they obviously have a lot of leverage in price negotiations. But they forewent that opportunity and you can see in the early years when when when the Medicare Part D was going through this, as you can see, big increases in lobbying, as that bill was going through Congress, to stop the proposal that the Democrats were pushing at the time, and even many Republicans, to negotiate drug prices with the pharmaceutical companies. And so, when Obama came in in 2009, everybody who had studied the problem, all the health policy experts said, you have to deal with the cost question first, because if you're going to expand coverage, before you deal with the cost question, then what's going to happen is, you know, the sticker shock of this for the American voter is going to be out of sight, and you're going to make the whole program vulnerable. So that would mean taking on Big Pharma, because as you can see, as the Obama legislation was going through Congress, that the lobbying also skyrocketed because they were extremely concerned about this issue. Obama had run on it in the campaign, and he was determined to address it, so that's Big Pharma. But then the other problem they had was that if you create universal coverage, particularly if you do it by including what was called a public option, which meant that for people who couldn't find private insurance, they would be able to buy medical insurance from the government. If they included in a public option the thought there was seen as several advantages to this. One was it would make sure that everybody would have the possibility of buying insurance, even in communities where, there was no other insurers. Secondly, it would force competition into the system and particularly in areas where there was only one private insurer, they would now have to compete with the government with the public option. And thirdly, and perhaps most importantly from this perspective of thinking about the pathologies of path dependency in the American system, it would be a path to eventually creating Medicare For All, because more and more people could gradually move into the public option over time, if it was providing better coverage. And so it would least create the possibility of expanding the American system into something more recognizably like what we see in other countries. But, adding the public option meant that not only would you have to take on Big Pharma, you would also have to take on the insurance industries. Big Pharma or drugs, insurance, a different part of the, of the healthcare industry, because for all the reasons I've just given, they didn't want the public option, right, and so the big debate in the administration, if you read the pop books on the first years of the Obama administration is, so, how to square this problem, because if you don't solve the cost problem and you expand the coverage, you're going to get the sticker shock, but if you try to do both of these, you've got to take on two of the most effective lobbying groups in America, and as the lobbying went on, it became clear that, or it became believed in the White House, I don't know if it was clear or not, but it became believed that they couldn't fight both of these battles at once. And so, against the advice of the policy wonks, Obama punted on the drug controls, essentially, gave up on a public option in return for some window dressing, apparent cost savings that, as we described in our book, and I won't go through in detail, were largely chimerical. And so effectively they gave up and abandoned the idea to negotiate drug prices in order to get the Big Pharma off the back of the administration and the thought was, it's more important at the end of the day to increase coverage than it is to fight this to battle on two fronts, which we're probably going to lose. And so that took Big Pharma out of the picture and actually they started to support the bill. But the universal coverage problem was a much more difficult one, and here you can see as Obama is still pressing to try and get the public option adopted, you can hear him making the case. - An additional step we can take to keep insurance companies honest is by making a not-for-profit public option available in the insurance exchange. (room applauds) Let me be clear. (room applauds) Let me be clear: it would only be an option for those who don't have insurance. No one would be forced to choose it, and it would not impact those of you who already have insurance. In fact,-based on Congressional Budget Office estimates, we believe that less than 5% of Americans would sign up. Despite all this, insurance companies and their allies don't like this idea. They argue that these private companies can't fairly compete with the government, and they'd be right if taxpayers were subsidizing this public insurance option, but they won't be. I've insisted that like any private insurance company, the public insurance option would have to be self-sufficient, and rely on the premiums it collects, but by avoiding some of the overhead that gets eaten up by private companies, by profits, excessive administrative costs and executive salaries, it could provide a good deal for consumers and would also keep pressure on private insurers to keep their policies affordable and treat their customers better. It is only one part of my plan and shouldn't be used as a handy excuse for the usual Washington ideological battles, but I will not back down on the basic principle that if Americans can't find affordable coverage, we will provide you with a choice. (room applauds) - So the reason there was all that applause from the Democrats in the House was that the previous month, he had been getting cold feet about the public option and in fact had made a couple of speeches saying that the public option was not essential to the bill and he wasn't going to fight to the death to keep it and this had produced big backlash on the left of the party and so he came out there and made a strong speech in favor keeping the public option in the bill, and seemed to be suggesting that it was non-negotiable. But the insurance industry had smell blood in the water by then and were going to work to try and get rid of the public option. And again we run into some of the pathologies of the American system. As you all now know, you need 60 votes in the Senate to pass a bill, unless it's budget reconciliation, when you only need 50 votes, as we talked about in the class on taxes. So you needed 60 votes, and they had had 59 votes for a long time, and finally when Al Franken's recount was resolved and he got his seat, they were up to 60 votes but the 16th vote was from Connecticut Senator Joseph Lieberman, and he had indicated support for the legislation all along. But, Connecticut is one of the insurance capitals of America, and so you can see here, you can get some sense of how this was going to play out. - With us now Senator Joe Lieberman of Connecticut. You just heard the White House top advisor on such things saying that it is likely that both the House and Senate will finally come up with some kind of healthcare bill that has some sort of public option in it. What's your take? - I hope not. In other words, I'm all for healthcare reform. We have a system that needs fixing, but we've got some more urgent problems than that, and the first most urgent is to fix our economy, to get it creating jobs again, and I think that a public option will actually hurt the economic recovery and our long term fiscal situation because it will end up causing the government to raise taxes, will probably raise premiums, or it will put us further into debt. I want to be able to say yes. There's so many good things we can do to make health insurance more affordable and to extend it to people who don't have it now, but I feel so strongly about the creation of another government health insurance entitlement, the government going into the health insurance business. I think it's such a mistake that I would use the power I have as a single Senator to stop a final vote. - But what does that mean that way that you might wind up with nothing instead of something? - Well, the truth is that nothing is better than that because I think we ought to follow, if I may, the doctor's oath here in Congress as we deal with healthcare reform: do no harm. - I'm gonna ask you this question because I want to give you a chance to respond to it. Some of your critics say that the reason that you are so dead set against the public option is because there are so many insurance companies headquartered in your home state in Connecticut and they've been some of your biggest supporters. What have they giving you this year, $400,000, something like that? Does that have anything to do with your position on the public option? - No. You know, I wish people would come out and debate me on the public option instead of questioning my motives. That's not the reason, but I will say this: this recommendation of a public option, a government health insurance company, takes our government down a road that we've never gone down before. In other words, we believe in a market economy. It's what's created the great American middle class, but it doesn't have a conscience. When it behaves badly, we regulate companies, we sue them. I've been angry at oil companies. I never had the idea that the government should go into the oil business to make oil companies behave better. I think this would be a terrible mistake. - So, as you can see, that is Lieberman making his case as to why he's opposed to the public option. Of course oil markets are completely unlike health insurance markets because they're not subject to the problems of adverse selection that we discussed earlier, where the people who, if you don't have a universal system, the people that most need the insurance are going to be the ones that the company's gonna least want to insure, and so it's not a very convincing argument. However, it was it was a political winner, and 10 days before the bill was going to go through the Senate, he was the 60th vote, and he insisted that unless the public option was taken out of the bill, he would not vote for it, and so it was taken out of the bill. And so, having made the decision not to take on Big Pharma over prices because of the difficulty of fighting two wars at the same time, they didn't in the end get the universal coverage either. And so you can see the pathologies of the American system playing themselves out with this. What happened subsequently with Obamacare? Well, it was challenged in the courts and finally made its way to the Supreme Court, and many people thought that the court was going to strike down the entire bill, particularly because of some drafting errors in provisions about the individual mandate. And there were two surprises in the Supreme Court's decision written by John Roberts in NFIB. I should say that NFIB is the National Federation of Independent Businesses, and they tend to represent to represent small business rather than big business and they're very conservative and very opposed to healthcare insurance, and so they brought this case. And so, Roberts, and this was the surprise that gets a lot of press, said well actually, the individual mandate, which requires you to have insurance, so to get around the problem of adverse selection, or pay a penalty, that that penalty was a tax. And so it was the taxing power of the federal government saved Obamacare. But a different feature, and in another surprise, it's a great surprise to everybody in the healthcare industry and people who study this was that there had been another provision in the Obama bill to try and move it toward more universal coverage with respect to Medicaid coverage for poor people, which is that they gave a lot of funds to expand Medicaid in the states to improve it, to get away from the niggardly system of block grants that diminished Medicaid funding over such a long period of time, and what they did, almost all of the money for this was coming from the federal government, so states would not have to raise taxes, but they also conditioned all their Medicaid funding, all of the federal government's Medicaid funding, on signing up for this. And of course, many of the states had conservative Republican legislatures or governors who didn't want to expand Medicaid, and so, they were, they were arguing that this was this was an unfair coercion and the court bought this argument. And so this, in effect, while it it saved part of Obamacare, it saved the individual mandate, it made it impossible for the federal government to force the states, to even condition the federal aid to the states on their expansion of Medicaid. And so, that greatly reduced the reach and reduced the expansion of coverage that Obamacare would have provided to poor people. And playing out the path dependency in 2017, when the Trump administration comes into office, because the individual mandate has now been declared to be a tax, they could repeal it as part of the 2017 tax law, for which they only needed 50 votes because it was going through as part of the budget reconciliation, so which they duly did. And so the individual mandate was taken out of the bill and that, of course, is the stop gap against the unraveling of insurance markets in the face of the adverse selection problem, and they've also damaged it in other ways that are being fought out in the courts by allowing consortiums of small businesses to clump together to buy insurance, which would again enable better heeled groups to bail out of the system and leaving less well heeled groups in it. Again, another version of the Hirschman problem. So what might be done? So one one argument that we propose in "The Wolf at the Door" is to think of bottom up Medicare expansion and what what we, the idea here is the following: that currently, most people can cover their children up to age 26 on their health insurance. This is through existing regulation. If you said that people in year one from age 26 to 30, could buy Medicare, this would be quite cheap, because it's not young people who get sick, it's old people who get sick, right? And so you could move towards Medicare For All step by step, but if you started with the younger people and then they could keep it as they as they grew older, it would take, you know, it would take a generation to achieve that gain. But eventually, but you would be getting there, and it secondly takes account one of the problems, one of the things that all of the Democratic candidates are running into, well certainly the ones on the left of the party, is that most people, it's said over and over again 190 million people have health insurance from their employers and they're satisfied with it. So, but two things are happening over time. One is they're getting less satisfied with it because the cost to employers are going up, and so they're getting less and less generous plans, or they're, like, you know, just as at Yale now, or you can get the health plan or you can get ETDA and then you can get a more expensive ETDA. They're creating tiered systems. But more and more people are finding that the coverage they're getting from their employers isn't that great. And then secondly, think about what we've talked about over the course of the last few weeks. More and more people are going to be losing their jobs over and over again over the course of their careers. So even if they have employer-based healthcare, they're probably not going to be able to keep it. And every time they change jobs, the question is going to be well, what kind of health benefits come? And so many of those people who might today not think there's anything wrong with their health insurance that the government should be fixing might start to have a very different view over time, given the fiscal pressures on firms to reduce these costs, and the fact that people are going to be changing employment and might have to go from a job that has coverage to a job that hasn't got coverage, or at least might have no coverage while they're looking for another job. And so even though current polling suggests most people are satisfied with their employer-based policies that number is likely to change, and if you had a way of voluntarily buying into Medicare, the group for that would grow. And then of course, this builds the political coalition over time, because you have more and more people in the coalition who want to be able to keep what they're getting, and so this is the kind of policy that we argue can work its way through our building blocks of distributed politics and actually become a credible political, a viable political path to a serviceable policy. And so, it's a, as I say, we didn't have much of a positive thing to say about K through 12 education, but this is a kind of path forward which wouldn't have the sort of sticker shock that Medicare For All has today, and if you think about it as a 10 or 15 year project, actually has a chance of viability as a proposal for which support can be built. Okay, we will talk about proposals for political reform in the final lecture on Thursday. (serene music) |
Power_and_Politics_in_Todays_World | Lecture_21_Backlash_2016_and_Beyond.txt | - Today we're launching into pod four of the course, The Politics of Insecurity. And that's gonna preoccupy us in the next two lectures. Let's start by going back a few years, before the real backlash starts in. This is 2011. - Today is the Tea Party movement's biggest day. It's the day people are most frustrated with their government, when they're sitting down and writing those checks to the IRS and thinking, why am I paying all this money? Today really highlights the Tea Party's message that politicians here in Washington are just out of control. - They are fed up with Washington, and sick of runaway spending. (crowd clapping and cheering) And at Tea Party rallies across the country this weekend, there were plenty of Republicans fighting for their support. - Let's send them this message. Don't tread on me. - Real solidarity means coming together for the common good! This Tea Party is real solidarity! (crowd yelling) - Some, like former Alaska Governor Sarah Palin, and Minnesota Congresswoman Michelle Bachmann are Tea Party favorites. - They've been very exciting, very encouraged, about my potential candidacy and I'm excited to be here with them. - And not to be left out, billionaire businessman Donald Trump made his first appearance at a Tea Party rally. - We have a man right now, that almost certainly will go down as the worst president in the history of the United States. - No one knows if Trump is serious. But some early polls have him on top for the Republican nomination. He's gotten attention with his questions about whether President Obama was born in the United States. A widely discredited issue other Republican contenders have discarded. But in playing to the Tea Party, the potential candidates also will have a challenge. Recent polls show 47% of Americans have an unfavorable view of the movement. So candidates looking for Tea Party votes have to be careful not to alienate moderates. - That's what the White House is sure hoping, that all of these Republican presidential candidates are out there, and sort of waving the flag of the Tea Party. And they believe that that will turn off independent voters. - Now this movement, of course, started just two years ago. So this'll be the first presidential campaign these Tea Party voters can try to influence. And we of course, don't know what kind of impact they will have in the fight over the Republican nomination. But some of these candidates are assuming that they're going to need these Tea Party votes to win. - So that was 2011, two years after the advent of the Tea Party. People still didn't quite know what to make of it. It had been quite a successful lobbying group in Washington in producing major changes in the Obamacare bill that had been passed the previous year. But it still was very unclear what sort of electoral force it was going to be. Today's agenda is to really explore the economic demographic and cultural sources of the 2016 backlash. And the larger puzzle of resurgent identity politics. The political sources of these developments are gonna be deferred until Thursday's lecture. And today we're really going to be thinking in terms of my ideas, interests and institutions. We're really gonna be thinking about the tensions and complementarities between interest and ideas in the evolution, I should say in the origins and evolution of the backlash. So I want to begin by digging a little more deeply into the Tea Party ideology that you saw being appealed to there by going back to the very start of the movement. And this is an anti, tax rally in 2009, September of 2009. (bright piccolo music) - Most of those there would have called themselves patriots. That don't tread on me flag was an early symbol of the American Revolution, who argue that their government today is betraying traditional principles. Steve Butler, a physician from Indiana, was handing out copies of the Constitution. - If you read the quotes of Thomas Jefferson, these guys were conservatives. And they said that the control should be with the people and not with the big government. - Iliana Johnson came to America from Romania some 30 years ago. - I find myself now every morning when I wake up, what kind of freedom have we lost today? - You could also find plenty of signs of something else. A rage that identifies President Obama with Hitler, or Stalin, that questions his citizenship. That seems to celebrate the death of a famous liberal. And among the main currents of the protestors here, a conviction that the media, Fox News and Talk Radio accepted, are deliberately concealing the truth. - Yeah, they're blocking it out. I mean we know it, everybody knows it. - You see these people back here? They all know there's something happening, but the mainstream media, they don't tell it. - It's your freedom! (crowd yelling) - But perhaps what most united these protestors was a broader discontent. A sense that they are not being heard. That their interests and the nation's interest are in the hands of a few. Dawn Newman came up from Texas. - It's all a good 'old boy network. You know they don't care what we think. Everybody's in everybody's bed, you know? And everybody's pocket. And it's who's got the most money. It's all about greed and power. - So there you have, I think in a very tightly distilled form, the four major elements of the Tea Party outlook or ideology. One is this very staunch, anti-elitism. The sense that the political establishment and the economic establishment, are reflected in both political parties is fundamentally opposed to the interests of people like themselves. Secondly, lots of conspiracy theories floating around this movement and you saw some of them there. Conspiracies among the media to conceal what's going on, being central among them. Third, hostility to taxes and to government, especially to the federal government. And this sort of harkens back I think to the Edsel and Edsel that I talked to you about some weeks ago. The connection between the anti-tax movement, and typically for ethnic minorities. And you see these things go together there as well with the racism, questioning President Obama's citizenship, and so forth. And we're gonna come back to the view of the world that these four elements constitute later in the lecture. But let's first talk about the economic sources of the backlash. And so, this is a little bit of review of things I've told you before. But then some addition to it. Last week I pointed out that one of the sources of the anti-elitism in the backlash was the divergence that on the one hand, after the financial crisis, people connected to the financial sector, or with significant financial assets, had their fortunes turn around pretty quickly. And by 2016, 2017 all the major stock markets were basically back and much of the wealth that had evaporated, or had appeared to evaporate during the financial crisis, had pretty much been restored for those people. But that there was very different impact on the real economy. And when I showed you these slides first about unemployment that more people were unemployed for longer, both in terms of short term unemployment, and longer term unemployment. And that this didn't show signs of ending any time soon. Indeed the long term unemployment rates were the highest at least since the 1950s. I talked about the impact of the housing crisis, the very high level of foreclosures. And then in a separate lecture, I talked about the disparate racial impact of the housing crisis. So just building on some of that. If you look at what happened to real wages in many of the advanced countries. But we're gonna focus today on the US. You can see that this is real wages, changes in real wages, between 2008 and 2017. And you can see it's very, very modest. So real incomes in the US grew by over that period by about 6% total. Not very much at all. And, if you look at who benefited from the revival of the economy, even though the growth was pretty slow, we had big increases in productivity growth. That's the top line over there. But if you look at hourly compensation with wages there's two different measures of it, some technical differences that need not concern us here. You can see that the returns to the increases in productivity were not going to hourly wage workers. And this is part of the Pacetti story that returns to capital greatly exceed the returns to labor over time. So it's a period of wage stagnation for people earning hourly wages. This decade, or eight years if you like after the financial crisis. Then, if we zoom out a little bit further into the larger context of long term employment insecurity this has been building for decades. I talked in an earlier lecture about the decline of union membership, which you can see beings pretty much right after World War II. That's the blue line there, declining union membership, where as the share of income going to the top 1% is flat through about the 1970s. But then starts to take off pretty dramatically. If you look at union membership falling and middle class incomes, they pretty much fall in tandem. So with the middle class share of income, as workers get less and less leverage, also has been falling for decades. This is another slide I showed you relatively recently, this is the job insecurity of the late baby boom workers, workers born between 1957 and 1964, by age 52 have changed jobs 12.3 times on average. Very big levels of employment insecurity. And again, zooming out even further, this is again the story that's highlighted by Pacetti, his co-author Siaz and many others that we have had this very high levels of inequality that were in the Gilded Age. And people by the end of World War II, thought were going away, in fact have come back. And we're now at the same levels of inequality that haven't been see for more than a century. And this is of course what partly motivated the Occupy Wall Street movement, which I will have more to say about in a future lecture. And the Bernie Sanders insurgency in the 2015-'16 primaries. And to give a sense of what the Sanders campaign was all about, let's go look at him, or listen to him for a minute a couple of years before he decided to run for President. - Today, the wealthiest 400 individuals in America own more wealth than the bottom half of America, 150 million people. 400 and 150 million. Today, and this is really quite amazing, the six heirs to the Walmart fortune, the Walmart company of course started by Sam Walton. His children, one family, now own more wealth than do the bottom 30% of the American people. One family, owns more wealth than the bottom 30%, 90 million Americans. And I know we have some of my colleagues coming up here and saying, "Look not everybody in America's paying taxes. "Got millions of people are not paying "any taxes," no kidding! They don't have any money, 'cause all of the money is on the top. - So, that was the economic story and the outrage that Sanders is articulating there. He, on an interest based story one would anticipate that this would resonate with large numbers of voters. And we've seen in this that the damage that has been done, was not just to white voters, but to African Americans disproportionately and Hispanics during the housing crisis. And on an interest based story, one would expect that this was an opportunity to build a coalition that one would have anticipated would do something about that. But, life has more imagination than us. What this story missed is what I'm gonna spend most of today's lecture talking about, namely the demographic and cultural sources of the backlash, which have also been building in different ways for a very long time. But before we dig into that, I want to first make one point that I think is a symptom of the fact that this story is perhaps a little bit too simple. At least as I've told it this far. And I'm referring here to the fact that employment insecurity is not just about working class voters. This very important new book by our colleague in the law school, Dan Markovits called, "The Meritocracy Trap." A big part of Markovits' message in that book is about the so-called hollowing out of the middle class that many occupations that used to be middle class occupations of upwardly mobile people are disappearing either offshore. Or more importantly to technology. So, if you think about we talked about bankers giving mortgages awhile ago. If you wanted a mortgage, the person giving you the mortgage was a local bank manager who had worked their way up to that position. It was a stable job. You would go down and have meetings with that person. They might be a neighbor of yours in a middle class community. Today when you apply for a mortgage it's to Quicken Loans, you do it over the phone to somebody who probably has a high school education. Maybe not even, on the other end of the phone, might not even be in the United States. They type your characteristics and economic data into a software program, an off the shelf software program, and you get the answer as to whether or not you're eligible for a mortgage in a matter of minutes. The bank manager who would have interviewed you about your mortgage has long since disappeared. Those jobs have gone away. So it's not just the bank tellers, it's the bank managers. And it's creeping, this disappearance of employment is creeping up the occupational ladder. I, for the last seven years now, have been teaching at the Yale School of Management. Seven years ago, one of the most attractive things to do for people who were graduates of business schools, was to go into investment banking. It's not a favored activity anymore, why? Because a lot of those jobs are going to big data. There are algorithms now that make is much less appealing to go there, but then the question arises what are these people going to do instead? So we should when we think about employment insecurity we shouldn't just focus on the disappearance of industrial jobs that force people into low paid service sector work. And if you want another data point that tends to support this. If you look at primary voters for Donald Trump in 2016, I'm not talking about the general election when you would expect most Republicans to vote for whoever the candidate was. But primary voters, when they had many other choices. They don't fit the stereotype of Joe lunch bucket. So if you look 1/3 of the Trump primary voters earn less than the median income. That's less than $50,000 is the median income for a family of four in the United States in 2016. Another third earns between $50,000 and $100,000. And the third third earned more than $100,000. So 2/3, by most definitions of working class, 2/3 of Trump's primary supporters were not working class voters. So, there's obviously something much more complex going on here than the stereotypical image of who the Trump voter was. And it was something that I think the sort of Bernie Sanders kind of analysis, didn't capture at all. And I think that that's what we have to try and understand with some other tools. Now, two very interesting scholars, one a sociologist by the name of Arlie Hochschild, and another political scientist by the name of Kathryn Cramer, had spent the five years leading up to 2016, engaged in ethnographic studies of angry white voters. We'll talk more about Hochschild in a minute, she studied and lived with oil workers in Louisiana. Cramer studied rural workers in Wisconsin, which she was trying to understand the rise of Scott Walker. Cramer's book is called "The Politics of Resentment." And what was interesting about these, both these books is that they, the research for them was really done before the advent of the backlash we're really studying now in the years leading up to 2016. But they shared a lot of light on it. So, Hochschild's book, "Strangers in their Own Land," is based on this five years she spent in Louisiana with oil workers who had been downwardly mobile. Many of whom had lost their jobs. Their communities had often been savaged by the environmental effects of the industry within which they were working. Lakes which they had fished in as children were now so polluted that the fish in them were all long since dead. There were very high levels of cancer and other diseases that were widely attributed to the pollution in the community. Large numbers of people had been laid off, because of mechanization of the industry. And yet, the paradox that she tried to understand was what that the anger of most of these people was directed not at the oil companies, but at the government. And particularly at the federal government. And she tried to understand this through participant observation. Here's a summary that she gives of the central finding in her book. - I stuck upon something I call the deep story. Which is a story that feels true. You take the facts out, you take the moral judgements out and have a picture, a metaphor for what goes on. In this, people are waiting in line as in a pilgrimage to the top of a hill where there's the American dream. People have been waiting in line a long time. They feel deserving, their feet are tired. The line hasn't moved in many years. And then they see, some people cutting in line, who's that? That would be blacks, who now have access to jobs that used to be reserved for whites. That's women now have access to jobs reserved formally for men. Immigrants and refugees, and now endangered animals all seem to be getting ahead of you, having cut in line. Then, people notice Barack Obama, a representative of the federal government, seeming to wave to the line cutters. He's sponsored them. Indeed the federal government comes to seem like an instrument of their marginalization. It has indirectly put them back. You're now at the back of the line, and no one's really paying attention to you. And then someone ahead of the line turns around and says, "Oh you're a Southern, ill-educated "Bible thumping redneck." And then you feel estranged. There's a moment which you feel like now a stranger in your own land. And you're kind of stuck until one day someone comes along, a magisterial figure, powerful, who sees and recognizes your estrangement and says, "I'm gonna give you your country back again." - So, this is this narrative of cutting in. There's some notion that people are getting ahead in line in front of you. And therefore preventing you from advancing. If you read the whole book it is a story of felt humiliation and rage at this phenomenon. And so, when you think about the reasons for their dislike of the government, first that they perceive the government of an agent of this cutting in, as the principal agent and facilitator of this cutting in. And then, if they need help from the government, because they lose their jobs, they have to go and be interviewed and be humiliated by people who have job security, who have government pensions, who might well not be white. And so, it reinforces this sense of victimization. That feeds this narrative of cutting in and what Cramer calls the politics of resentment. I think it's important, this was a study done in Louisiana, but Cramer's is in Wisconsin. And there's since been five or six other books. So this is not just a phenomenon about the South. And I think that's important to understand. It's really a much larger phenomenon. The story that Cramer tells is very analogous. That rural voters in Wisconsin have many of the same resentments, and many of the same causal stories that they tell themselves, about the sources of their failure to advance, and the sources of their alienation. And one other thing in the Hochschild comments that I just want to highlight and I'm gonna come back to. Is that she noticed when she described the cutting in there she said, "Cutting in, African Americans are now "getting jobs that were formally reserved for whites. "And women are now getting jobs that were formally "reserved for men." And so this raises the whole issue people's implicit baselines. When poor whites of this kind think about something like affirmative action, they see this as somebody being given an unfair advantage, a leg up. But they can not perceive that this is a remedial response to a kind of advantage that they had in the past. So a very important book on this subject by Ira Katznelson, at the Columbia Political Science Department called, "When Affirmative Action was White." About the advantages that whites previously enjoyed that now have been eroded or taken away. As I said the question becomes, what is one's baseline? And this feeds if you like what I'm gonna call irreconcilable narratives about the nature of what's going on the country. Because the people from whom the traditional advantages have been taken away can't see that in the sense that the fish can't see the water it's swimming in. And so the narrative that they will tell is in some important sense incommensurable with the narratives that others will tell. So I want us to try and understand the sources of these irreconcilable conflicting narratives that I think have become in many ways weaponized since 2016. And I think there's no way to do this except by going back into American history and seeing how contentious and tangled these perceptions of who has had what advantage and when. And what is the appropriate thing to do about it. Let's reflect on some landmarks in the battles over race in America. And particularly the battles over separation of the races and integration of the races that have been here literally since the country's founding. I'm not gonna go back all the way to the founding, but let's go back to before the Civil War. In 1857, in the Dred Scott case. This was a fugitive slave had escaped to a non-slave state. The question was whether they could sue in order to prevent themselves from being returned. And the court held, among other things, this person had no standing to sue, because black Americans were not citizens. And so this was essentially an affirmation of the idea that there has to be a fundamental separation of blacks and whites, because blacks are not truly citizens. And this, Dred Scott was one many decisions, and this was in the court, but of course, things that were done in the federal government that eventually led to the Civil War debates about extending slavery to the territories and so on. But then we get the Civil War Amendments in 1865, 1868 and 1870, which at least on their face, are pressing for more racial inclusion. Undoing if you like, the Dred Scott decision. It outlawed slavery, they created equal protection, forbid it being denied on the basis of race, and the 15th Amendment granted voting rights. It said that voting rights could not be withheld on the basis of race, or previous condition of servitude. So there you see an attempt to get more inclusive polity. That then runs into the reality that citizens in the South, white citizens in the South had many, had not accepted the outcome of the Civil War as legitimate. And huge backlash and attempts to undo the effect of the Civil War amendments. I talked in an earlier lecture that the Supreme Court was to some extent complicit in that by reading the amendments very narrowly. So as to limit their effect. And so we had the rise of Jim Crow, first hard to partite, and then soft to partite. 1896, the court upheld the idea of separation of the races, again in the educational context, and separate but equal was a legitimate way to organize K through 12 education. And we see this pull back from the idea of an integrated polity. 1935, the Social Security Act is passed. Another wonderful book by Ira Katznelson called "Fear Itself," deals with the passage of the New Deal. And in particular, the reality that the only way in which Roosevelt could get the New Deal through congress, which was controlled by, the big committees were controlled by Democrats from the South who were strongly opposed to all forms of racial integration. Was to make essentially exceptions that would get them to support it. And in particular the Social Security Act excluded domestic and agricultural workers, most of whom were African American. And so, that was if you like, the quid for the quo. We're talking about quid pro quo these days. That was the quid for the quo, to get enough Southern support to pass the Social Security Act. Again a deprivation of a racial group and the continuing if you like, soft de partite in that case, because an proxy rather than race was used to achieve the exclusion. Then we get the Warren Cord, and we get Brown versus Board of Education in 1954 and 1955. Overturning Plesy and saying no separation is not acceptable, there's no possibility of having equal education that is based on racial separation. That it's again bound to be a form of discrimination and perceived as such, and experienced as such by black children. And so the court ordered integration of the schools, which turned out to be much more difficult to achieve for reasons that we have talked about earlier, than people might have anticipated. And so again you see this tussle over what is the ideal? You know the public ideal of America is this notion of a melting pot, that is a nation of immigrants, people come from all places, and Italians, and Irish and others have all eventually become part of the melting pot. But, with African Americans it seems to be fundamentally different, because of this battle over inclusion. And of course we shouldn't think that this is only a battle among white voters. A lot of the conversation about inclusion versus exclusion tends to treat African Americans as incapable of using agency of their own on these matters. Which of course is not the case. If you look at the history of the Civil Rights Movement, Martin Luther King was very much for holding up America to its own ideology. And saying you're not living up to the ideals which you profess. This is what Michael Walzer calls eminent criticism in his little book, "Interpretation and Social Criticism." You take the melting pot ideal and the immigrant story and you say it's being violated with respect to African Americans, you're being hypocritical. We need to have the kind of inclusion the American dream should be equally available to all. But not everybody in the African American community agreed. And a rift opened up, the young... A young activist by the name of Malcolm X took a very different view of the matter. So let's listen to Malcolm X for a minute or two. - What is your attitude on the relations between the black and white races, what do you seek to accomplish? - The best way to solve the problem that black people are involved in this country, is to let us separate from whites and solve our own problems. Do for ourselves that which whites do for themselves. Stop baiting the white man and stand on our own feet and solve our own problems. - Is it far to say as a generality and as a succinct way to put it that you believe in segregation of the races? - Segregation is that which is forced upon inferiors by superiors. Separation is done voluntary by two equals. You never refer to the oriental community in which orientals live exclusively as a segregated community. Because they live there voluntarily. Everything there is controlled by them. The economy, the politics, the civic organizations, but the negro community is referred to a segregated community. Because Negros are forced to live in that community contrary to their will. And they don't control the businesses of their community. They don't control the politics of their community, nor their social life. We do believe in separation, standing on our own feet, among our own kind, and solving our own problems. And that's the only way you'll get a solution to the vital race problem in this country. Our people should unite among ourselves and try to solve our own problems, instead of trying to force ourselves upon whites and blame them for our plight. - Could we press for a very simple answer to one question? Do you hate all white people? - I don't think it's a fair question. White man doesn't even come into my attitude. Mr. Mohammad teaches us to love our own kind, and let the white man take care of himself. For a white man today, sir, after kidnapping millions of black people from Africa, stripping them of all human characteristics, and relegating them to the role of chattel, or cattle, or animals, commodity, merchandise that can be bought and sold at will. And then 100 years since the Emancipation Proclamation, using every type of deceptive method to further us into slavery, called second class citizenship. I think that it would take a whole lot of nerve for white people today to ask Negros do they hate all of the injustices that black people are crying out against in this country, are being inflicted upon them by whites. It is the whites who are depriving our people of civil rights. It is whites who are depriving our people of implementing what the Supreme Court decision said should be implemented. Wherever black people turn, they are facing the hostility of whites. Martin Luther King teaches Negros to love all white people no matter what they do to you. And the same whites whom he teaches Negros to love, sic dogs on him, sic dogs on their children, dogs on their women, and dogs on their babies. So, I think it's very hypocritical on the part of whites to accuse Mr. Mohammad of a hate teacher, because he says separation. And then sic dogs on Martin Luther King who is teaching integration. It means that no matter what the Negro does he is not going to get along with whites. So I think that Mr. Mohammad's whole philosophy is more intelligent than Mr. King's. - So that was this, he was a representative of the black nationalist idea that trying to get integration into the dominant culture was a fool's errand. It hadn't worked in 100 years, and it wasn't gonna work any time soon. And so, there should be this separatist agenda that he had gotten behind. It's important to say that he subsequently changed his mind. He had a falling out with Elijah Mohammad who he talks about there. And he spent some time traveling in Saudi Arabia and elsewhere and saw that there were Muslims who of many different colors, who actually did get along and work together. In the following year he actually separated from the nation of Islam, and began to support the civil rights agenda, although he remained a black nationalist. His falling out with the nation of Islam was so grave that they actually, they assassinated him, as I'm sure some of you know. The following year he was murdered. But interestingly, there's a curious kind of historical convergence because, Malcolm X became more in favor of integration and the civil rights agenda. Whereas actually Martin Luther King, in the last years of his life would become less sanguine about the possibilities of integration. And by the time he was assassinated four years later, he said why he had not given up hope, he was no longer optimistic having seen how little had been achieved by that time. So, but there you have it. You have this rift up in the African American community over integration and separation. Which would subsequently become transformed into debates about multiculturalism versus the integration American dream. Should we try and get everybody to approximate to some national norm? Or, should we respect the differences among different groups? And this would become a central motif in battles about conflicting narratives going forward. So, if we then think about the two great achievements of the 1960s from the point of view of race. And perhaps the two greatest achievements including the decades since then, were the Civil Rights Act of 1964, and the Voting Rights Act of 1965. And these were seen as great achievements for the integrationist view, that people were no longer gonna be denied just the formal incidents of citizenship, but the substantive incidence of citizenship. And, that this was gonna be the path going forward. It should be said the reason there were two separate pieces of legislation was that in 1964. The Johnson Administration was not in favor of voting rights legislation, because they thought it would be potentially too explosive. And battles over voting rights had also beset only during the New Deal. The Roosevelt Administration in their battles with Southerners who controlled congress. After the passage of the Civil Rights Act was when the first big wave of racial violence exploded in the US. There wasn't racial violence before the passage of the Civil Rights Act, it came later. And that that was what eventually prompted the recognition that it was gonna be essential to have voting rights legislation. So that's why we had two pieces of legislation. Another important landmark I think I the evolving narratives was the publication of what came to be known as the Moynihan Report, named for Patrick Moynihan, a famous senator who brought a very different set of assumptions and tools to thinking about racial disparities in the United States. The Moynihan Report was called The Negro Family, the Case for National Action. And it was, I think, an early indication of how the views of white elites, who think of themselves as progressive harbingers of change, can be perceived very differently by the people about whom they are talking. And what concerned Moynihan was the high levels of out of wedlock births in the black community. So, you can see on the slide the Moynihan Report is released in 1965. And at that time, African Americans had about 28% of children were born to mothers who were not married. And the numbers were, and the percentages were much lower among Hispanics, and whites and all other groups. So, this produced another narrative that would become hugely contentious about the social pathologies of the black underclass, so called. And that it was really the disintegration of the family that was at the roots of the difficulties being experienced in the black community. Again, this was widely trumpeted by some as an important thing that needed to be dealt with. But also widely vilified by others as again denying the agency of African Americans and using language and terminology of medical, medicalized terminology of pathologies and so on. And interestingly of course, the numbers were hotly disputed at the time. And if you look at how they've evolved subsequently, you can see that the number of children born out of wedlock in white families today is just as high as the number of children born out of wedlock in 1965 when Moynihan was writing. And moreover, and I think this is the reason we get an irreconcilable narratives about this, if you ask the question how many, you know where are the out of wedlock children born? Of course, because the white community is much larger than the African American community, 2/3 of out of wedlock children then and now are actually born to white parents. And so, you can find if you like, if you want to find facts to support a different narrative there are plenty of conflicting stories that can be told. You know another piece of this is that people would point out that when upper middle class white people decide not to get married, it's seen as a choice of lifestyle. But when a poor black person does it, it's seen as a social pathology. And so again, you find that these conflicting moral narratives are irreconcilable. People glum onto different facts, and use them as the decisive ones to illustrate what it is that they are talking about. So, I think that the Moynihan Report foregrounded a lot of these debates about African Americans in ways that made people increasingly speak past one another. Nixon's Southern strategy in 1968, we've talked about this in an earlier lecture. This is the backlash against the Civil Rights Act, and the Voting Rights Act. He takes advantage of that by recognizing that there's a second dimension, when we talked about the Downsian Models. That there's a second dimension in that resistance to racial integration matters more to many Southern whites than economic benefits from the welfare state. And so it's possible to, if you like, divide Southern whites by appealing to race. And that was Nixon's Southern strategy. And again I think brought racial divisions back into the foreground of American politics. And then the next really important step, I think comes in 1971. And so this is the McGovern-Fraser report on the structure of the Democratic Party. The Democratic Party had been in a major crisis in the late 1960s. Not only because of the assassinations of Robert Kennedy and Martin Luther King, who had both been strong supporters of civil rights, but the 1968 election had been explosive for the Democratic Party. But what had happened was not withstanding Lyndon Johnson's massive victory in 1964. A huge landslide he was fortunate to be running against Barry Goldwater at that time. Goldwater was sufficiently far from the mainstream of American politics, that Johnson could prevail. And he had used those majorities to enact the civil rights legislation that I've already talked about. But he also, of course, massively escalated the increasingly unpopular war in Vietnam. And by the time 1968 rolled around, it was clear that Johnson couldn't run for reelection. He wasn't going to win. But this was at a time when parties would govern much more top down. I will talk more about that on Thursday. And there was a sense that the anti-war sentiment in the Democratic Party had to be contained. But the trouble was once Johnson was not running, all of the candidates who were running were anti-war candidates. So, Vice President Hubert Humphrey was sort of parachuted into the race. He didn't run in any primaries, he was named as the candidate. And this produced an explosion at the Democratic National Convention in Chicago. And there was violence in the streets. And there was a strong sense of crisis in the Democratic Party. And this produced a movement for reform. No more smoke filled rooms. We are going to democratize the Democratic Party. And so, a political scientist, a very distinguished political scientist by the name of Austin Ranney chaired this commission with some others that became known as the Mcgovern-Fraser Commission. And it recommended, among other things, much greater use of primaries. And I'll talk more about that on Thursday. But also democratization of the power structures of the Democratic Party. And in particular, the use of quotas for various minority for women, and various minority groups in the decisions about platform committees, the governance of the party, the selection of delegates. Much more what we've since come to call descriptive representation. Much, a kind of that different groups must all be represented in the governance of the Democratic Party. And that indeed is what transpired. And this is part and parcel of the acceleration and attention to racial and ethnic characteristics in thinking about the distribution of offices and other benefits. That same decade, in 1978, the Supreme Court finally took up the question of affirmative action in this Bakke decision in University of California Davis Medical School. Which had set aside some positions for minority candidates. And Bakke was somebody who scored less well, I'm sorry, scored better on the test than the minority candidates who were admitted in the set aside places and he sued. And he prevailed. But, what was interesting in the decision was that the court said you can't have quotas for minorities, but you can take race into account. And this then became the moment at which the hugely fought subsequent history of battles about affirmative action got started. And particularly with words like diversity. We're now going through this in all the elite universities of the government looking at whether diversity's being used as a smoke screen for ethnic and racial quotas. You can have diversity, you can take race into account, among other characteristics, but you can't have quotas. And so, this whole debate about affirmative action has become another one of these conflicting moral narratives. In 1984, in the Democratic race, we're now in the middle of the Reagan Administration and the Democrats are running, basically. Walter Mondale is gonna likely be the candidate. And Jesse Jackson, a young civil rights leader at that time, has mounted a campaign against him, did surprisingly well in a number of primaries. But in the end had to support the ticket. And he makes a speech at the Democratic National Convention in 1984, where he tries to put together these very disparate ideas of melting pot versus multiculturalism. And he uses the metaphor of a rainbow coalition. So he's trying to square the circle, if you like. And this is how he goes about it. - Our flag is red, white and blue. But our nation is rainbow, red, yellow, brown, black and white, we're all precious in God's sight! (crowds cheering) (crowds yelling, applauding and cheering) America's not like a blanket, one piece of unbroken cloth, the same color, the same texture, the same size. America's more like a quilt, many patches, many pieces, many colors, many sizes. All woven and held together by a common thread. The white, the Hispanic, the black, the Arab, the Jew, the woman, the Native American, the small farmer, the business person, the environmentalists, the peace activists, the young, the old, the lesbian, the gay and the disabled, make up the American quilt. (crowd cheering and applauding) Even in our fractured state, all of us count and fit somewhere. We have proven that we can survive without each other. But we have not proven that we can win and make progress without each other, we must come together. (crowd cheers) - So there you have the idea of a rainbow coalition. And it's interesting that when you actually go back and find the speech, it's very clear that he intends everyone to be included. But again, this is an area where I think you get divergent narratives, because the... In the commentary on the history of identity politics, and racial identity politics in the US, this is interpreted as the moment at which the Democratic Party starts to think of itself as a patchwork of minorities. We talked about the patchwork metaphor there. But a patchwork of minorities, a strong deference to multiculturalism. All the different minorities should form an alliance to change the status quo. Whereas in the actual speech he's very clear that he's also referring to whites. But that's not how it came to be interpreted in the dominant lexicon. And indeed you can hear people say, "Well the thing about a rainbow, "it contains every color except white." And so the notion that this patchwork of minorities, if that becomes the dominant ethos of the Democratic Party, there's gonna be no place in it for the sort of people that Arlie Hochschild is talking about when she's talking about cutting in. There we have the idea of the rainbow coalition. And it's important to say that while the idea of descriptive representation in politics starts in the Democratic Party with the Mcgovern-Fraser reforms. It also migrates into the political system more broadly with the advent of majority minority districts. And the fight over majority minority districts. This was the idea that was developed in order to try and get African Americans elected to congress. Particularly from the South, where it seemed almost impossible to do that. And this is partly because of the way districting is done in the United States. It's done by whoever, at least at this time, it's now a more complicated story, but at this time the districts were always drawn by whoever controlled the state legislature after the decennial census. So you'd have a new census and then, it's still the case in 2/3 of the states, that some states now have independent commissions to do it. But basically, most states still have this system. And what was done was the districts were drawn, so as to dilute African American representation. And when the Voting Rights Act was reauthorized in 1982, stronger requirements were put in that essentially said that the section two of the Voting Rights Act required that drawing of districts in certain circumstances that would lead to the possibility of electing African Americans. And it was, you might think somewhat surprising that legislation reauthorized in 1982 would do this. Because this is during the Reagan Administration which one would not have thought would be particularly friendly to this idea. But although the White House was not in favor of this legislation. In fact, a number of Republicans on the hill, and Republican strategists were in favor of it. Because they saw it as a possibility of actually reducing Democratic strength in the South. So the thought would be, well if you create a majority minority district you might increase the number of African Americans in congress at the expense of reducing the number of Democrats in congress. And so some people have spoken about an unholy alliance between the congressional black caucus and congressional Republicans. And at some points even the Justice Department to support this idea, or at least not to oppose it very strongly, so as to, as I said, people with very different motives, the congressional black caucus wanted to increase its size. But Republicans who supported this idea saw it as a way of strengthening themselves in the house. In any case in 1986 the Supreme Court took up this matter and said that it's okay to have majority minority districts, provided the minority group is large and compact enough to constitute a single district. And provided the group is politically cohesive and can demonstrate that. Provided the minority group can show that it votes sufficiently as a group to defeat the minority group's preferred candidate. So the idea was if you can show that you're voting together as a group, and you conceive yourself as a group, they can draw districts to increase minorities. And so as of 2015 I believe there were 122 majority minority districts in the United States, which is roughly a 1/4 of the districts in the house. So again, you have appeal to the idea of separation in drawing districts, creating districts which will send a minority candidate to congress. Which they indeed did. The number of minority candidates in congress increased as a result of this. And it should be said that in state legislatures the same thing was done. Again, this was as a response to previous forms of districting, that had excluded African Americans from the possibility of electing African Americans to office. 2009 Reece versus Destefano, the reason I bring this case up is I think again, it really speaks to the conflicting, irreconcilably conflicting narratives that have beset these debates about inclusion and exclusion. So this was a Connecticut case, Mayor John Destefano, mayor of New Haven. This was a firefighter's union case. And what had happened was the city of New Haven had given an exam for promotion to officer. No African Americans had done well enough on the exam to be promoted, so the city fairing litigation from African Americans about a discriminatory test. At least a test that had a discriminatory impact, disregarded the results of the test, and didn't use it. This then produced a lawsuit. There was one Hispanic, and 19 white plaintiffs. And the court upheld their objection. The decision to get rid of the test was held to violate the rights of Reece and his colleagues. So the reason I bring this up in this context is not about the result, but rather it really underlines one of the realities about affirmative action. Which is that we tend to think about affirmative action very differently from the way many people experience it on the ground. This was summed up in 1995 in a book by a man called Michael Lynn, "The Next American Nation, "The New Nationalism and the Fourth American Revolution." He had a discussion of what he called a white over class. Again, this is the kind of liberal elites that would turn out to be objects of rage and ire from those voters that Arlie Hochschild was gonna be interviewing later. Lynn made the observation he said, "The white over class," and he was thinking here of sort of people we might today characterize as limousine liberals can, live right and think left. And so the thought here was that affirmative action has no real effect on the kind of people who live in Scarsdale. They can get advantages for their sons and daughters. We're seeing one of the scandals about how they do it in college admissions playing out right now. Lynn was saying this in 1995, very interesting book. But the real front line, the real battles that effect affirmative action are exactly in places like promotions in the police department and the fire department. And that's where people on the short end of affirmative action say, "I never had slaves, I never did anything "to discriminate against African Americans. "Why am I not getting that promotion?" So this is, if you like, Lynn in 1995 already identifying the beginning of this cutting in narrative. That feeds the sense of grievance that drives the anger and resentment that Hochschild and Cramer have documented. And, just to bring us right up to date, if you look at the Democratic primaries, you know among the things which they are talking about and fighting about are questions about reparations. Again, reparations is an idea which produces very divergent narratives. Some people see them as absolutely morally compelling. Others see them including African American left wing intellectuals such as Adolf Reed, Jr., sees debates about reparations as hugely destructive of the possibility of building the sort of alliances that could produce progressive change. Again, you're back in this world of irreducibly conflicting narratives that are important to people in understanding themselves and their identities. And that motivate them in politics, even to the point where perhaps they'll take positions that are self defeating from a materialist point of view, or from an intraspace point of view. And that really is I think the takeaway when we delve into this history and we see the ways in which race has played out in American politics, it constantly feeds these mutually incompatible narratives. And this is an important thing to know, because it's one of the ways in which modernization theory got politics very badly wrong. To this point in this course I've talked about modernization theory and democracy. That there were these predictions that countries would become more democratic as they developed economically. But there was another important prediction of modernization theory, which in some ways underlay the prediction about democracy. And that was the prediction that identity politics would over time be displaced by interest based politics. That things like religion and race, and so on, would atrophy as sources as political mobilization. This is something about which they couldn't have been more wrong. It's partly the prediction that democracy would arise is that when you think about divisible goods, which is important for the divided dollar game, and for democracy to work, that is things like money and income. Once you start talking about identity politics, it's not about divisible goods, it's about a very different kind of politics. It's about recognition, it's about status. It's about telling moral narratives that either validate you or humiliate you. And so we have reached a world in which we have these conflicting narratives that are not easily reconciled, even when they're about the same topic, such as affirmative action, or majority minority districts, or, racial set asides of other sorts. You get these mutually incompatible narratives. And what happened after 2016 is that one of these became weaponized, essentially. Trump found a way to weaponize the narrative of cutting in, that Hochschild had uncovered. And that is the story of the socioeconomic sources of populism as it's played out. And we'll turn to the political sources of it on Thursday. (bright digital bell tones music) |
Power_and_Politics_in_Todays_World | Lecture_22_Political_Sources_of_Populism_Misdiagnosing_Democracys_Ills.txt | - Good morning, everybody. (students chattering) So we talked about the economic demographic and social sources of populism last time. And today, we're gonna zero in on the political sources that interact with them and produce the kind of politics that we have all been experiencing since 2016. - The leaders to the US, India, China, Australia, every single one of our allies, the Bank of England, the IFS, the IMF, the CBI, five former NATO secretary generals, the chief exec of the NHS, and most of the leaders of the trade unions in Britain all say that you, Boris, and Nigel are wrong. Why should the public trust you over them? - I'm not asking the public to trust me. I'm asking the public to trust themselves. I'm asking the British public to take back control of our destiny from those organizations which are distant, unaccountable, elitist, and have their own interests at heart. - Elitist, elitist? - Absolutely, because-- - The Lord High Chancellor conspiracy of elites. It sounds like something out of (mumbles). - Well, I haven't seen (mumbles), but the one thing that I would say is that the people who are backing the remain campaign are people who have done very well, thank you, out of the European Union, and the people increasingly-- (audience applauding) Absolutely. (audience applauding) And the people who are arguing that we should get out are concerned to ensure that the working people of this country at last get a fair deal. I think the people in this country have had enough of experts with organizations from acronyms-- - The people of this country have had enough of experts? What do you mean by that? - From organizations with acronyms saying that they know what is best and getting it consistently wrong. These people are the same ones who've got consistently wrong. - [Faisal] This is proper Trump politics, isn't it? - No, it's actually a faith in the-- - It's Oxbridge Trump. - It's a faith, Faisal, in the British people to make the right decision. - [Faisal] Blind faith. - I don't think it is because one of the striking things about this debate is that those who are arguing that we should remain have a vested financial. The majority of people in this country are suffering as a result of our membership of the European Union. - What's your factual basis saying that the majority of people in this country are suffering from EU membership? - Well, I know myself-- - You don't have it, do you? - I know myself from my own background. I know that the European Union depresses employment and destroys jobs. - [Faisal] We haven't got a majority. - My father had a fishing business in Aberdeen destroyed by the European Union and the common fisheries policy. - That's one person. - The European Union has hollowed out communities across this country and it has also contributed to lower salaries for working people and it has also ensured that young people in this country don't have the opportunities to get the entry level job. You're on the side, Faisal, of the elites. I'm on the side of the people. (audience applauding) - You said a majority, you, the Lord High Chancellor said a majority of the British public were suffering from the EU. You have no factual evidence for that whatsoever. - Let me give you a fact. - [Faisal] No, you don't have any evidence for it. - Yes, I do. Because every year, we give billions of pounds to the European Union, billions of pounds that we should be spending here. - Well, let's have a fact. There's been a fact on the side of your bus in one meter tall letters, and let me quote, "We send 350 million pounds to Brussels every week. "Let's fund the NHS instead." Which week was it that we sent 350 million pounds to Brussels? Was it this week, last week, was it Christmas week, Easter week? Which week? - Every week, we send actually more than 350 million pounds to Brussels. - [Faisal] It comes out of a bank account in Britain and goes to Brussels. - Yes it does. - Does it literally leave the country? Does it leave the country? - And then we get some of it back and we do get some of it back. - You know it doesn't leave the country. And you (mumbles) to me, UK Stats Authority say the figure's misleading. The Treasury Select Committee containing Brexiteers says it's deeply problematic. One of your own backbench colleagues, a conservative MP who's campaigning to leave, is a respected doctor, refuses to hand out leaflets because he say you're treating the public as fools. (audience applauding) - So that was a video that went viral because of Michael Gove's declaration that the British people are tired of experts. It was shortly before the referendum that was held later that month. I chose to start with that because like the Tea Party videos I started with on Tuesday, it contains all the four elements of in this case, British populism, but astonishingly similar, you might realize on a little reflection to the four elements of American populism that we were talking about last time, namely the anti-elitism, the conspiracy theories, the hostility to taxes and government, and though not racist, this hostility to foreigners, this kind of xenophobic declaration that foreigners are hollowing out the labor market, taking jobs, destroying the British economy. And so I start with this just to make the point that while we spent all of our focus on the US last Tuesday, we're talking about a much more general phenomenon and that's what's gonna concern us today. We'll have much less to say about the US, although the three remaining lectures are gonna be very centrally focused on the US. Today's agenda is to answer this question: what is it about democratic political systems that has fostered the resurgence of populist politics? And our answer's gonna differ from much of what you read in that I'm gonna shift the focus away from exclusive attention to voters and voting behavior and instead also include attention to the incentives and motivations of politicians. If you wanted a bumper sticker for this lecture. I talked about the Tea Party narrative that Arlie Hochschild so cleverly portrayed with the cutting in comments that she made in that clip. I said that that had become weaponized unlike many of the other narratives in around the questions of race in American politics that I gave you some flavor of. But any time a social scientist starts talking in the passive voice, you should get nervous. Become weaponized? How did it become weaponized? Just as people constantly talk about our politics has become polarized, but again, weaponization of, and populism, and polarized politics don't fall out of the sky and we need to think about who it is that is mobilizing people to vote in a particular way. It's not as if 63 million Americans were sitting around saying, "If only somebody would come along "and propose building a wall." That's not how mobilization works in politics. Politicians see opportunities to mobilize voters in certain ways rather than others and then use them. So this lecture's gonna be very much more focused on the incentives confronting politicians, how those incentives have changed, and how that has shaped the politics that we're experiencing. I just am gonna be very centrally focused on the political system today, but we shouldn't lose sight of the broader content that we've talked about several times in the course. The impact of globalization, technology, and slow growth on middle class incomes and working class incomes. The aging populations in all of the developed countries that put more and more pressure on fiscal states as the working age population that's supporting retired populations is getting relatively smaller and smaller, and the collapse of the corporatist consensus between business and labor that underpinned the social democratic welfare states in the early postwar period largely as a result of the collapse of organized labor. These are all themes from earlier in the course that we should have in the back of our minds. What I want to say about them today is that all of those things make it much harder to govern and it leads voters often to demand political reforms and to the rise of anti-system parties or anti-system groups within political parties. And that in turn creates the risk of what I'm gonna call the political equivalent of bloodletting, namely fixing things that were not broken in ways that actually make the situation worse. And we will see particularly how this has played out in the United Kingdom since the 1970s. The pressure to reform the British political system ramped up in the 1970s. A very important book by Anthony King, unfortunately now deceased, called "The British Constitution" points out that apart from the war decades, the 1970s were one of the most difficult decades in British economics. It was the three day week, it was the oil shock. Britain was actually forced even to go to the IMF and be bailed out. And that caused a lot of people to demand political reforms even though actually the problems that Britain was experiencing had nothing to do with the political system. But they reformed the system in ways that have hamstrung its ability to confront many of Britain's problems. So the central questions I'm gonna zero in on this lecture are how do the number and nature of political parties affect politics and there, I'm gonna be interested in when I talk about number two party systems versus multiparty systems. And when I talk about the nature of political parties, I'm gonna be talking about weak parties versus strong parties and I'll say more about what I mean by weak and strong as we move along. And then the second question is I'm gonna look at what are the implications of this for the economy and society? How does it affect? How do these phenomena, the number of parties and the strength of parties affect labor markets, corruption, the propensity for corruption, and growth, and also how do they affect the likelihood that we're gonna see populist politics. That's out agenda. So just for people who like tables, here's a two by two. If we think of two party systems, the classic strong two party system would be the UK system before they started messing it up in the 1970s and subsequently. So I'll have a lot to say about that and then weak two party systems would be the US system. We have very weak political parties in ways that I will elaborate on. Western Europe has traditionally had strong multiparty systems although they've also been weakened in recent years. And in much of Eastern Europe, in France and much of Latin America, we have multiple parties that are also weak parties. Israel also fits into that box. To some extent, we'll see these if you like ideal types and many actual political systems are hybrids, but nonetheless, the center of gravity is what I'll be talking about when I think about these different possibilities. Now, I'm talking about political parties. Well, what is a political party? How do we know one when we trip over it? So here's one famous classical definition that Burke came up with in the 18th century. He said a political party is a body of men, as they then were, united for promoting by their joint endeavors the national interest upon some particular principle in which they're all agreed. It's kind of a verbose mouthful. More simply put, you can say that parties promote what I'm gonna call partisan conceptions of the public interest. They're partisan in that they come from a point of view, but they are declared to be conceptions of the public interest, not of a private interest or a sectional interest. So if you remember when Mitt Romney got caught on a microphone I guess in what he thought was a private meeting saying that he represented only, he could only aspire to represent 53% because 47% were the takers and we're the makers. It was a politically disastrous thing for him because you can't say that. Politicians are gonna say the American people want, the American people need. I'm gonna represent all of the American people. So this is the idea of a conception of the public interest even though he comes from a partisan point of view and I'll unpack what I mean by partisan. More of that later. But let's start by two party systems versus multiparty systems. Now, what is it that determines whether you get two parties or lots of parties? It's the electoral system and there's something that political scientists call Duverger's Law. And so Duverger's Law, this is a rather impenetrable mumbo jumbo. It says single member plurality systems, SMP. That's like we have and like they have in the UK, systems in which all you need is the most votes in a district to win the seat. They will tend to produce two parties. And indeed, his formula he says is district magnitude plus one. So district magnitude means the number of people getting elected from a district. So district magnitude in the US or the UK is one and plus one, you're gonna get a two party system. All else equal, but need to unpack all else equal. That assumes that most districts are gonna be like most other districts. Or if you'd like to put it in more technical terms, that the median voter in a district is pretty much gonna be the median voter in the country. So every district looks more or less like every other district in its political composition. If that's not true as in India where there's big regional variation, then you'll get more parties. So even though in India, they have single member district systems, they nonetheless have more parties. Or in the UK for instance, you have regional variations so you can get the Scottish nationalists in Scotland. But for the most part, if you have big diverse districts that are more or less like one another, you're gonna get two parties out of single member plurality systems. Whereas if you have proportional representation and this is if in the limiting case, in pure PR as say in Israel, if you get 60% of the vote, you get 60% of the seats. Your party gets 30% of the vote, it gets 30% of the seats. Then you will get lots of parties, other things being equal again. So for instance, you can reduce the number of parties in a PR system by putting in a threshold. So in Germany, there's a 5% threshold. Parties that fall below 5% don't get any seats in the parliament. So that's one way in which you can make a multiparty system function more like a two party system. Another way in which you could make a two party system function more like a multiparty system is by having multiple rounds. So in France, they have a first round and a second round. So even though they have single member district system at the moment. The French change their system so often that it's hard to keep track of it. But at the moment, they have single member district system. They're beaten only by the Italians for changing their electoral system every five minutes. So in France currently, they have single member districts, but because they have two rounds, the small parties stay in in the first round in order to then essentially extract rents or benefits for supporting bigger parties in the latter round. So the two rounds keeps the small parties alive. There's these variations around the edges, but basically, the big point is if you have PR, you're gonna get part proliferation of parties, and if you have single member district systems like we have, like the British have, like they have in Canada, although again, regional variation produces the different outcome with the Quebecois, but basically, single member district's gonna produce two party systems. Okay. Now, there's one big debate about well, which is better? Which is fairer? Which is more representative? And people often say well, multiparty systems are much more representative in the sense that if you think about a multiparty system, the greens get their party in parliament, and the social democrats get their party in parliament, the libertarians get their party in parliament, the anti-immigrant people get their party in parliament, everybody gets a seat at the table and so it's more representative. And in that sense, multiparty systems are often defended as fairer because they are more representative. And the British liberal party endlessly complains that they get very few seats and if Britain had PR, they would do much better, and they've tried to at least. They only took this view once they lost power at the beginning of the 20th century. Before that, they were strongly opposed to PR, but now, they see that it would benefit them. One important caution though is we shouldn't conflate the electoral stage with the government formation stage in a political system. So if you think about, we think about the 2015 elections in Greece, that was the outcome of the election. All the different parties were represented, but that was the government that formed with a far left party and a far right party. So the people who voted for all these other parties were unrepresented in the government, right? And so you could say it's more representative at the electoral stage, but not more represented at the government formation stage. Three years later, the Italian electrons produced a comparable result. Again, you've got a far left party and a far right party. And as I mentioned to you in our very first lecture, in Germany in 2017, they had had a grand coalition between the SPD and the CDU prior to that. The election result was such that the social democrats said, "We're sick of being in this coalition "which we get damaged by "'cause of all the concessions we have to make "to be in the government." And so they didn't want to join and Merkel spent six months trying to put together a coalition between the free democrats who were sort of like libertarians, sort of Rand Paul type libertarians, and the greens who are for environmental regulation. Not surprisingly, she failed because they want no regulation and they want green regulations so there's not much of an intersecting thet. And during those six months, all the polling showed that the alternative for Deutschland was gonna do even better if they went to another election. So finally, they put back together the grand coalition. So when we're thinking about, this is really something of a trade off between increasing accountability and representativeness at the electoral stage which might come at the expense of accountability of actual governments, and that, I'm gonna dig into a little more deeply. Now, one of the objections that people make to two party systems is the Tweedledee Tweedledum problem. And the Tweedledee Tweedledum problem or as I put it here, the Tweedledum Tweedledee problem which is the mirror image of the Tweedledee Tweedledum problem, (students laughing) is basically that politicians in a two party system are chasing the median voter. So if you think about it this way, if the right of center party went over here where I've got that B on the slide, the center to the left, the party to the left would move toward it to pick up a lot of votes, right? So if you think about 1964 in the US, Barry Goldwater has a very conservative platform. And the result is Lyndon Johnson wins a massive victory. So you can't do that if you want to win. The parties are gonna head for the middle. But then if they do that, both parties are gonna be offering the same policies, right? This just follows almost definitionally. They might get it wrong, right? So Goldwater maybe thought he was close to the median voter and when Ronald Reagan ran in 1980 on more or less than same program, everybody said, "What an idiot," "we saw what happened to Barry Goldwater." But whether it was Reagan's brilliant strategy or blind staggering luck or some mix, the median voter had moved and so Reagan could prevail. But generally speaking, all of the incentives for politicians which is what I'm focusing on here are to head for the middle in a two party system. And so then people say, "But voters are not gonna get a choice." If two parties are basically, especially if they have very sophisticated polling and they learn about where the median voter is. It's much better now than it was in 1964. It's gonna be a problem from the point of view of giving voters choice. So that is true up to a point. So if you think about Britain, National Health Service put in by Clement Attlee in the postwar labor government in 1946. Comes into operation over the next couple of years. Hugely popular in British politics. It's bulletproof. If you look at polling on the most popular institutions in Britain, it's more popular than the second most popular institution which is the Queen. And this means that the National Health Service really is bulletproof. People want to keep it. And so even in Margaret Thatcher's rolling back the welfare state in the 1980s, she couldn't dismantle the National Health Service. Of course, there'll be squabbling at the margins about funding, and coverage, and this sort of thing, but basically, it's there to stay because the median voter has an intense attachment to it. Not everything's like that. So if you think about nationalizing and denationalizing British Rail, that has occurred three times and if Jeremy Corbyn comes into office, they might renationalize British Rail. Similarly, in the US, think about social security brought in by FDR, I talked about this last time, in 1935. And Medicare insurance for people over 65 brought in by Lyndon Johnson as part of the great society in 1965, these are also bulletproof, hugely popular with the median voter, basically bulletproof. But again, not everything is like that. So if you think about how to respond in a recession, your left of center parties are gonna generally push for deficit spending. They have Keynesian outlooks and think that the problem is insufficient demand and there's gotta be a demand-led return to revive the economy. Whereas supply side is on the right. They're gonna say no, what we need in recessions are tax cuts. And so you're gonna get very divergent responses to recessions. Or if you think about protections for labor, the Wagner Act I've mentioned before put in place by FDR in 1935 was the strongest protections for unions in American history. By 1947 when the republicans were back in control of congress, they repealed the guts of it over Harry Truman's veto. That means they had 2/3's in both houses to do that, greatly weakening the Wagner Act. Or more recently if you think about the Dodd-Frank law which we've discussed a little bit, the Consumer Financial Protection Bureau put in as part of that law has since been savaged, and weakened, and might be completely dismantled by the Trump Administration and the republicans in congress. So that's what I mean by a partisan conception of the public interest. Go back to the beginning of the lecture. So this is one partisan conception of the public interest and this is another partisan conception of the public interest. There will be some overlap, so some Tweedledee's Tweedledum, but there will be a lot of areas where in fact there's divergence, partly for ideological reasons, partly because people don't actually know what the best way is to respond to a recession, and partly perhaps for other reasons as well. And so while their Tweedledee Tweedledum is an issue to some extent, we should notice that for the most part, you do get significant divergence. Now, some people say, "Well, "the electoral system doesn't really matter that much." All the stuff I'm going on about, there's less to this than meets the eye because if you think about a two party system, it's about big tent parties. They've got lots of different interests in. Basically, they've put together a coalition before the election. Multiparty systems put together the coalition after the election. What's the big difference? They're more or less gonna turn out to be the same. So we shouldn't really get this hot under the collar about the differences. I think that that is fundamentally misguided and there are five reasons and I'm gonna work my way through them why particularly when we focus on the incentives created for politicians, the differences matter and they matter in ways that are relevant to our larger concern. And the differences all come down to the fact that in two party systems, it's winner take all and loser lose all. And that creates a fundamentally different dynamic for politicians. So I'm gonna start by again, this is mumbo jumbo, but I will unpack it into common sense terminology in a minute. It's that in two party systems, politicians have an incentive to internalize the costs of the deals that they make. So let me explain what I mean by this. Let's suppose I in a two party systems am in charge of strategy for the party that's to the left of center in the diagram I just did. And who's a right wing? Steve Wisener over there is a right winger. He's in charge of strategy for the right of center party, okay? So Steve and I both know that if I lose, he wins, if he loses, I win, right? And everything, can gain control of everything. And so it's very important to both of us not to leave any voters on the table that we might be able to get because that one last voter you leave on the table might be the difference between winning everything and losing everything, okay? So now think about in a multiparty system. After an election, maybe the big like Germany grand coalition which often happens. The big labor party and the big business party are forming a coalition. What does business want? Business wants industrial peace, predictability, no strikes. What does labor want? They want protection of worker's wages, protection of their members. So they might make an agreement that gives both parties those things while externalizing the costs of that agreement onto others like higher prices for consumers, right? Or if you think about an agrarian party, thinking of joining a coalition. What does the agrarian party wants? It wants protectionism, it wants price supports for agriculture, right? And they make that the condition of joining a coalition, right? But that might produce higher food prices for the rest of the population. But the election's already over. Nobody's worried about that, so they'll externalize the costs of the deals that they're making. When Steve and I are confronting one another before an election, we have to worry that if we are seen as responsible for inflation or advocating policies that are gonna produce inflation, then the other party's gonna win. So we both have incentives to make the deals we have to make to put together our party without externalizing cost any more than we have to. We have to worry about what Americans like to call swing voters. We will satisfy our core constituencies in ways that imposes few costs as possible on swing voters. That's the basic incentive. So that's an important distinction. Then secondly, when we do that, we're gonna have to worry about I put it here and I mentioned this earlier when we talked about referendums during the lecture on the anti-tax movement, but I'll just come back to the analytical core of the point there, is that when we put together a platform, we have to bundle a lot of policies together, right? Policies on abortion, policies on education, policies on spending. We have to bundle all of these things because voters are getting a choice. Essentially, you're forcing it down to one dimension and you've gotta push everything together. And when you do that, you have to discount everything you propose by everything else you propose. The example I gave you earlier just to remind you is if you ask Americans, "Would you like to get rid of the estate tax "which almost nobody pays?" You can get 70% to say yes, but if you say, "Would you like to get rid of the estate tax "if that also meant getting rid "of prescription drug benefits for seniors?" then you get majority saying, "No." So in the second case, they're bundling their preference or discounting their preference for a tax cut by their preference for keeping prescription drug benefits for seniors. So that's what parties do. They bundle together lots of issues, but they've go to do the discounting. Steve and I both have to worry about doing the discounting of all these issues against each other in ways that will alienate as few voters as possible because of the winner take all character of it. In a multiparty system, it doesn't work like that. If you're a representative of the green party, you're just gonna say, "I'm gonna go hell for leather "for green legislation if I get elected, "if our party gets elected," right? You're essentially only appealing to your core voters and the more fragmented a system gets, the more parties you get, the more they're gonna be like that. All your voters are sort of gonna be like primary voters in the US. I have more to say about primary voters later. And often, you'll have parties that only have a single raison d'etre like the green party has a single raison d'etre. Anti-immigrant parties have a single raison d'etre. The religious parties in Israel have a single raison d'etre. So they're just going to try to, if you like, turn out their base because they know they're not really gonna appeal to many other constituencies. And so they are not gonna do this kind of bundling and discounting of the things that are important to them with the things that are important to others born in mind. And you can see this by taking a third analytical comparison that I put up here of two types of arbitration and industrial disputes. So if you have industrial disputes, management and labor can't agree on a contract, sometimes there'll be a provision, okay, we'll go to arbitration. We'll let a third party decide. And when you do that, typically what happens is management takes their position, the unions take their position, the arbitrator comes in, does fact finding, and picks the solution, usually some form of compromise. Of course, everybody knows that's gonna happen and so because they know that that's gonna happen, they behave strategically. So the management might take a much more extreme position than it really believes in. As well, the union. Just like a store putting up their prices later to put them down and say they're having a sale so that when the compromise is made, they hope it'll get somewhere closer to what they really want. There's a different kind of arbitration called last best offer arbitration. And last better offer arbitration works like this. Management takes their position, the union takes their position, the arbitrator has to pick one. There's no fact finding, no compromise. The arbitrator has to pick one. And you won't have to think about this for very long to realize that in last best offer arbitration, people tend not to behave strategically and tend not to take extreme positions because of the fear that if you do that, the arbitrator will pick the other side. So people tend to be reasonable, they tend to converge, they head toward their sincere preferences, if you like, and they head toward the middle. Well, two party competition is much more like last best offer arbitration because the voter is essentially put in the position of being that last best offer arbitrator, right? So Steve has to put together a platform and I have to put together a platform which we think that last best offer, that median voter in the position of political analog of the last best offer arbitrator will choose. Whereas if you think about multiparty systems, again, the big incentive is to create a surplus to give away during the negotiations. So you can say, "Well, I've made all these compromises," and so you're gonna get a lot more strategic behavior and you're gonna have parties making very different kinds of deals based upon the positions that they take. And so if you want parties to put together programs that appeal to as many voters as possible, what we sometimes call programmatic policies, you want policies that are gonna serve the interests of the largest number of voters, you're more likely to get that out of a two party system than out of a multiparty system where coalitions are gonna be put together in the short run to protect the members of the coalition for that government, but nobody, as we'll see, is thinking much about the future. Another difference that's important to notice has to do with opposition. If Steve gets it right and I get it wrong, I go into opposition. And then I will spend my time attacking his government and telling voters, "Vote them out. "Put us in and you'll know what you'll get." And we put together the alternative program. In a multiparty system, you have no idea what will happen if you kick out the government because you don't know X and T what coalition is gonna be formed next time around, right? Just think back to the Germany example I started with. Had Merkel managed to put together the coalition with the greens and the free democrats, then the SPD, you'd have a centrist opposition. And now that you've got a grand coalition, you've got splinted opposition of very disparate parties lobbing their different grenades at the government, but nobody knows what will happen after the next election and what sort of government might be formed. So voters have no idea what the alternative to the current government actually is. And so again, it's a feature of multiparty government. Or to put it another way, in a two party system, parties are sort of like marriages. Of course, sometimes people get divorced, but people don't marry intending to get divorced. They intend to stay together and to be seen as a family, and to have certain characteristics, and to plan to exist for the long run. Parties are just the same. Occasionally, they fall apart. The wigs fell apart in the 1850s in the US. The liberal party collapsed in the first decade of the 20th century in the UK. But for the most part, parties expect to be together over the long run to think therefore about policies for the long run and so on. Multiparty governments are more like hook ups, right? This is cool for me right now, but who knows? All bets are off about the future. The future will take care of itself. And so again, if you want governments that are going to invest in policies for the long run, it's more likely that you're gonna get such governments when you have two party systems rather than multiparty systems. So for all of these reasons, it's mistaken to think that there's nothing at stake here. And if you want to think about what will make it more likely that politicians will develop policies for the longer run, you're more likely to get that out of two party systems than out of multiparty systems. And the more fragmented that the multiparty systems become, the more difficult or the more pronounced these disadvantages are going to be. So let's talk a little bit about strong versus weak parties. So this is the second dimension of the two by two with which I began. Now, an important feature of political parties that often goes unnoticed, and I haven't mentioned yet, but I want to bring it to your attention is that they're very unusual institutional creatures in that nobody owns them. The closest thing there is is actually a private university in the US. So think about Yale. Who owns Yale? The truth is nobody owns Yale. If you sold off all of Yale's assets, as with a firm, if you sell off all the assets, there's a residual claimant. It's the shareholders. So if you sold off all the assets, pay the debts, you'd give the residual claimant is the shareholder. There is no residual claimant at Yale. So if you sold off all of Yale's assets and paid off all of its debts, it's not clear who would be entitled to the residual. And that has the implication that governance is always controversial. The faculty think they should govern Yale. The students think they should govern Yale. The alumni think they should govern Yale. The corporation think they should govern Yale. The donors think they should govern Yale. Nobody agrees on who should actually be governing, right? Because it's an unknown entity. There's no analog to the shareholders. Political parties have the same feature. There is no equivalent of a shareholder or a residual claimant in a political party. And so for that reason, the governance of political parties is endemically controversial. There's always gonna be fighting over who should govern political parties. And a lot of what we're seeing in our current politics is because the fighting over who should govern political parties has tended to come out a certain way. And so that's what I want to zero in on now. So, definitions are important. What is a strong party? Well, here's one definition from an article in World Politics last year. A unified, centralized, stable, organizationally complex, and tied to long standing constituencies. As I said, political scientists never saying words of one syllable what can be said in words of five syllables. I like simple things, so I would say a strong party is a party in which everybody's on the same team. Everybody is trying to achieve the same goal. Everybody's pulling in the same direction. That's what I would think of as a strong party. And one feature of that would be back benches choose front benches, back benches being what we call members in the US, choose front benches, the people running the parties, who choose back benches who choose front benches. And so let me explain why that's important. First of all, back benches will give front benches a lot of leeway, what I call to discipline them. I call it Ulysses and the Sirens discipline in the sense that you know the famous story of Ulysses and the Sirens is that Ulysses is tempted to do things that won't be good for him, so he wants to be disciplined, right? So the intuition here is that going back to remembering Steve and I facing winner take all, loser lose all, if our back benches, if we're the leaders of the party, they know we both have incentives to design a platform and pick candidates for running in the different constituencies that can both win in their districts and support a national winning platform, right? Whereas if you just focus on the individual back benches or the individual members, they're just trying to survive in their districts. So they might want a very big margins. If their idiosyncratic preference is in their district that might be unpopular nationally, they're gonna focus on those things. So the idea is back benches will give a lot of leeway to front benches to discipline them, to determine what the platform's gonna be, to have a big role in candidate selection, but they only have enough rope to hang themselves. One indicator is that we have very weak political parties is that Nancy Pelosi led the congressional democrats to four successive defeats and still stayed there, right? Just as with a sports team, the quarterback or the coach has a lot of authority, but it's conditioned on delivering victories. If they don't deliver victories, they're gonna be gone very quickly, right? So this is the idea. The back benches have a big interest in delegating authority to the front benches conditional on their delivering victories. And if they don't, they won't survive for very long because the reason for delegating the authority will have gone away. Weak parties are parties in which everyone's in it for herself or himself. Lots of people worried about their own survival in their own district. Lots of retail campaigning, not wholesale campaigning. This comes back, by the way, to our discussion of money in politics. Politicians in weak parties need much more money 'cause they're all campaigning for their own survival. You'll see a lot of deal cutting. It'll make an American system look more like a European system. I'll support the unnecessary military base in your district if you'll support the unnecessary military base in my district, what political scientists call pork barrel spending. You will not see individual members worried about the national program. Candidates and platforms will tend to be controlled from below or from outside by the people who are funding the candidates. And so you'll see a situation where leaders in the legislature find it very difficult to set the agenda and to whip the back benches to support their agenda. Again, another symptom of the fact that we have very weak parties in the US is that when the republicans were in opposition, they voted 71 times, I believe, to repeal Obamacare, but when they came into office, they couldn't do it. The leadership couldn't whip the back benches to support the program of repeal because they knew they would pay too high of a price for it in their own constituencies. Now, the recent trend across the democratic world has been to weaken parties. And this has taken a number of forms in a number of systems, but it's all tended to produce the same result. So one way in which this has happened is in moving away from back benches selecting front benches, members of a parliamentary party picking its leadership, to instead having leaders elected at large. So here's a pretty dramatic instance. - Jeremy Corbyn elected as leader of the Labour Party. - [Presenter] The veteran left winger has come to be reinforced his position as head of Britain's main opposition party. Corbyn was backed by nearly 62% of half a million party members who voted in the leadership contest after he'd been challenged by many Laboor MPs. He immediately called for unity. - We have much more in common than that which divides us. As far as I'm concerned, let's wipe that slate clean from today and get on with the work we've got to do as a party together. - [Presenter] The immediate question is whether more centrist Labour MPs who'd opposed Corbyn will rally behind him. Beyond that lies the task of mounting a serious challenge for power. Critics say the in fighting has left a vacuum in opposition that's been filled by the Eurosceptic hard right. - So Corbyn's call for unity there is a little bit alike we might call it the British equivalent of whistling Dixie in that. But here's the interesting set of facts that led up to that. It used to be the case that, as I said, the parliamentary labor parties elected their own leader. Starting in the early 1980s, they put in a series of reforms to weaken the role of the parliamentary party and strengthen the role of larger constituencies out there of unions and others. But in an any event, when Ed Miliband was leader, they put in a system that any member of the Labour Party could vote in the leadership election. It costs three pounds to join the Labour Party and it has as you saw 450,000 members. So in 2015 after Miliband was forced to step down because they lost the election. As I said, in systems like this, if you don't deliver a victory, you're gone. Corbyn won. And Corbyn was a card carrying Marxist way to the left nevermind of the median voter in the Labour Party but of the median British voter. And the parliamentary labor party quickly discovered they couldn't work with him and so there were mass defections from the shadow cabinet and eventually a vote of no confidence in him of 172 to 40. That's a pretty big majority. And then three months later as you just saw, he was reelected by 62% of the membership. So if you now think about Corbyn coming into government which might happen in December of this year, you'll have a situation where he will not be able to govern with the members of the parliamentary Labour Party, many of whom could not survive in their constituencies if they supported his program. So this is something that has gone on in country, after country, after country. Many, many political parties have moved toward more direct election of their leaders. And the difficulty is that the political activists who tend to join parties and participate are on the fringes with the parties and tend not to be representative of the people who turn out in the general election and elect people to parliament. And so you get that problem. Secondly, increasing use, I talked a little bit about this when we did the anti-tax class. Increasing use of ballot initiatives, referenda, and plebiscites. Again, since we're talking about England, let's stay there. - Their verdict has been given by a vote and by a majority bigger than that achieved by any government in any general election in the history of our democracy. No one in Britain, in Europe, or the wider world should have any doubt about its meaning. It was a free vote without constraint following a free democratic campaign conducted constructively and without rancor. It means that 14 years of national argument are over. It means that all those who have had reservations about Britain's commitment should now join whole heartedly without stint in the task of overcoming the economic problems that assail us as a nation. And work whole heartedly with our partners in Europe and our friends everywhere to meet the challenges confronting the whole world. - So this is what many people don't know when they think about Brexit referendum that in fact the first British referendum ever was in 1975 called by Harold Wilson. At that time, the Labour Party was split over Europe and tended to be more anti-Europe because British workers had stronger protections than European workers, whereas the Tories tended to be more pro-European. So it was a flip, the mirror image of what it is today. And so Harold Wilson was confronting a very divided party and rather than do the hard work of the discounting and so on to force a resolution within the party, he had this idea, let's just have a referendum. And so they had their first referendum and as he said, 2/3's voted to stay in the EU. I couldn't resist putting this on the slide. He crowed later. He said, "It was a matter of some satisfaction "that an issue "which had threatened several times over 13 years "to tear the Labor movement apart "had been resolved fairly and finally. "All that had divided us in that great controversy "was put behind us." Of course, five years later, they split anyway when the left wing had gotten control. After the first set of reforms to the leadership election process, the left wing had gotten control and so called Gang of Four, Roy Jenkins, David Owen, Shirley Williams, and Bill Rodgers left and formed the social democratic party which would eventually unite with the liberal democrats. Same thing with Cameron in 2016. He was trying to duck resolving the conflict within the conservative party and so he thought he would do it with a referendum. It wasn't any more successful than Harold Wilson's effort. Interestingly, just to go back to Margaret Thatcher, in 1983, she was then a sitting prime minister. She had won in 1979 the 1983 election. She was pro-European in those days, but she said, "Britain's getting a raw deal "from the common market "and so we need to renegotiate." And particularly the, as I mentioned in an earlier lecture, the agricultural policy and the fisheries policy that Gove was complaining about in that video worked to Britain's disadvantage. And so she had negotiated a rebate for Britain. It came to about 40 billion pounds a year. That would be an example of trying to figure out a way to redo the bundling, right? So yes, there's a Eurosceptic wing of the tory party. They've got these complaints, so we'll do something to address them and redo the bundling and go on into the future. And it was pretty successful. Not only did she win in 1983 promising to do that. After she did it, she won the next election as well. So this was a very different approach than the use of ballot initiatives, referenda, and so on. Fixed parliaments act. Again, we've seen this playing out in real time. Weakens the power of the prime minister to whip the back benches to support a program. So in this case, if they had not had put in the fixed parliament act which was put in in 2015 again to limit the unilateral power of the leadership, you now have to get a 2/3 vote in parliament to call an election, but for that, Theresa May could have forced the tories to support her deal that she had negotiated. No longer has that power. Lots of calls to strengthen presidents partly because legislatures have become so dysfunctional and so difficult to form governments that can govern. People say oh, we need to give more authority to executives to punch through all of the bickering in parliament. And so you get especially in Eastern Europe, in Latin America, and increasingly here, claims that it's important to concentrate executive power because you can't otherwise govern. And this is not just Trump, but in series one president after another relying increasingly on executive power. Growing reliance on primaries and caucuses. Again, primaries and caucuses are seen as ways of democratizing the control and governance of political parties. They've been around for over a century in the legislature, but become more and more important because we have more and more safe seats. And if you have more and more safe seats, the primary election becomes the only election that matters. At the presidential level, primaries didn't become important until the McGovern-Fraser reforms that I talked about last time. But here's the thing about primaries and caucuses is you're thinking about the primary, the left party's gotta go to the median voter you might think of their party, and the right party of the median voter of their party, but that's not what tends to happen because the people who turn out in primaries and caucuses tend to be on the fringes of the parties and so not even these people, but these people. It's extremely low turn out in these primaries and caucuses, so you can get Alexandria Ocasio-Cortez elected in New York with an 11% turn out, Jim Jordan in Ohio or Mark Meadows again selected in primaries with 15 or 16% turn out. And these people are not representative of the voters in their district and nevermine representative the voters in the country. And so what tends to happen is they control who gets selected and then when the people go to Washington, either they do the kind of compromising that's needed to pass legislation in which case they get accused of having sold out by the people who selected them, or they don't, in which case you get gridlock and alienation of other voters. And so either way, the remedy does not address the voter alienation that people so often complain about. And just if you want a very recent illustration of this, I saw the other day on Morning Joe Mike Murphy, a republican strategist, was reporting on interviews of senate republicans and had concluded that if the vote on removal from office of Trump were a secret ballot, 30 republican senators would support it. But they know that if they're known to have done it, they'll face brutal primary challenges and almost certainly not survive. So that's the story about weakening parties. Now, some people say well, maybe the best solution is the middle, is some mix. We talk about Labradors being, a Labradoodle is a dog with a nice personality that doesn't shed, right? And so people have often said the Germans got it all right. They have a mixed system, partly of single member districts, partly PR, and they've been very successful for much of the postwar period, so maybe what we really want is a system that gets the both of all worlds. And Germany was always held out as the Labradoodle, strong parliamentarism unlike Weimar which gave a lot of power to the president. The president is very much answerable to the legislature. The single member districts that they have, they only have 299. They're very big and they're very diverse. They're three times the size of Britain's and that tends to make it function more like a two party system. You tend to get predictable governments. Lots of party discipline until relatively recently in candidate and leadership selection. No longer true in the SPD. And then there are various external constraints, budget rules, constraints from the EU, anti-protectionism requirements, and so on that hemmed in the deals that governments could make. But it was all underpinned by this postwar social democratic consensus that I have mentioned to you several times. And even in Germany now, the truth is we saw it took Merkel six months to form a government after 2017. The decentralized control of the SPD which meant that half a million almost members had to vote on every provision of the coalition deal meant that the negotiations were very tortured. Merkel had to give up six ministries including the finance ministry and produced a government over which she has very little control. The grand coalition splinters the opposition as we've seen. And it's been very bad for both the parties in the coalition. In the regional elections last year, they both hemorrhaged support and this has happened more recently again in the local elections. So what's going on in Germany is replicating the pattern elsewhere. I just remind you of the slides I showed you. The number of parties has gone up steadily since the 1960s. Particularly the fragmentation of parties on the left has been dramatic and it's paralleled the decline in unions. Less dramatic on the right, but still if you look at the number of right wing parties as compared with the 1970s, they are much higher, many more of them than there used to be. And so the upshot is that you get weakened parties who have diminished capacities to form governments and govern which produces more alienation of voters that then feeds demands for further reform. And so think about the parallels with the 1930s with which we began this course. So in the 1930s in Britain, there were also fascists. They were called the Mosleyites. And in 1936, the Mosleyites actually had larger membership than the Nazi party in Germany, but they didn't get any seats in parliament because of the two party system, right? Whereas in Germany, in Italy, in these other countries, you got, I showed you this picture in the very first lecture. You had a highly fragmented system, government impossible, and then the Nazi party just sweeping through and eventually taking power. And people worry about whether this isn't where we are headed again. And if you think about the last decade or so, again in Britain, we've had the rise of UKIP. We've had the British national party which is even further to the right than UKIP. But because they don't have PR, these parties don't get anywhere in the political system. Instead, as we've just seen now with Nigel Farage, he's basically having to have influence through the tories if he really wants to affect the outcome if the tories win. Otherwise, we'll be looking at a hung parliament and it'll make it look more like a European system. And so he's essentially pulling Johnson to the right to take some of his concerns into account, but not being able unilaterally to determine the outcome. Whereas in all the European systems, these very anti-Europe parties, anti-system parties are doing much better. So again, we worry that this might not be some kind of portent for the future. So what are the implications more broadly for the economy and society? The take aways of today is that in two party systems, we should expect comparatively weak protections for organized labor because unions are not a majority of the party and comparatively low spending and redistribution. That incidentally is why people on the left used to prefer multiparty systems because they were thought to be more redistributive. But in multiparty systems today, you should expect organized labor to be decreasingly effective at protecting a shrinking labor aristocracy and that's what we have seen playing itself out in all of these European systems. We should expect the emergence of populist parties with anti-immigrant and protectionist agendas. And where parties are weak, we should expect primary challenges or presidential candidates running on anti-immigrant and protectionist agendas. And we should expect that where you have fewer stronger parties, it'll be easier to resist this. When we think about what's called clientelism, this is when you instead of pursuing programmatic policies in the interest of the polity as a whole. Rather, you reward sectional interests. We should expect that you're most likely to get programmatic politics out of strong two party systems when Steve and I are worrying about not leaving any voters on the table over the long run, you are gonna get what I call wholesale clientelism, giving the farmers their high prices or giving whatever you need to give, give the religious parties in Israel whatever you need to give them to form a government that's wholesale clientelism when you have multiparty system. Where you have weak parties, you should see a lot of retail clientelism, bridges to nowhere, special deals for firms in your district 'cause that's what people are worried about. And when you have lots of weak parties, we should expect lots of wholesale and lots of retail corruption. Cleintelism, lots of corruption. If you look at the statistics on economic performance, again, this is that paper from last year that I mentioned to you, the data also supports the idea that you get better economic performance when parties are strong. You can peruse this slide and that paper which I strongly recommend at your leisure. So reform agendas, they're hard because of the very widespread misdiagnosis. People feel out of control. They want more democratization of political parties, more control from below, more unbundling, more referendums, more direction of leadership, and they don't understand that that's making things worse. It's hard also because the beneficiaries tend to not want to change the system that has put them in place. So when Renzi tried to reverse some of this in a referendum in 2016 in Italian politics, he had his head handed to them. Because starry-eyed progressives tend to think that more direct democracy is gonna help them, but in fact, the agenda soon gets hijacked. It should be remembered that primaries were brought to us by the progressives at the beginning of the last century. Hard because of the misguided faith in PR, people think it's more representative. And it's hard because the direction that reform needs to go in is really the opposite of what most people think. And what to do in that difficult situation will be the subject of my lecture on December 5th. See you next week when we will talk about the building blocks of distributive politics. (calm music) |
Power_and_Politics_in_Todays_World | Lecture_6_Reorienting_the_Left_New_Democrats_New_Labour_and_Europes_Social_Democrats.txt | - I started the lecture last Thursday by playing a clip of Michael Foot making fun of Sir Keith Joseph who had been the intellectual architect of Thatcherism. With his conjurer's trick, and said at the end of the day that the joke was on Michael Foot, because he thought Thatcher would be a flash in the pan, and she went on to be Prime Minister for 11 1/2 years and went on fundamentally to restructure the British economy and political landscape as we saw. I want to start today with a clip from his, his comrade in arms Tony Benn, who had been his staunchest supporter on the left of the Labor Party throughout the 1980s. And actually ran for the deputy leadership when Michael Foot became leader In 1981 and lost narrowly to Healy. But this is Tony Benn on the day after Thatcher has announced she was stepping down, evaluating the Thatcher legacy. - [Tony] Despite the fact we've been told we're an entrepreneurial society, this is a country today that has an utter contempt for skill. You talk to people who dig coal, run trains, doctors, nurses, dentists, toolmakers, nobody in Britain is interested in them. The whole of the so-called entrepreneurial society has focused on the city news we get in every bulletin telling us what's happened to the pound sterling to three points of decimals against a basket of European currency. Skill is what built this country, strength and it is treated with contempt. I must confess the auctioning off of public assets, particularly the latest disgusting Frankenstein advertisements which told me more about the mentality of the Minister who devised the scheme than it did about the sale itself, these are assets built up by the labor of those who worked in electricity and by the taxpayer who put the equipment in, now to be auctioned off at half their price to make a profit from the tax cut for the rich before the next election comes. If these were local councilors they would be before the courts for wilful misconduct, and because they are ministers and then some of them later go on the boards of the company's they privatized, they are treated as businessmen who know better how to handle it as members of the board of directors than allegedly they did as ministers responsible. The undermining of the trade unions with less rights in Britain than they have in Eastern Europe. The tax cuts for the rich and benefit cuts for the poor, the censorship of the media, the abuse by security services the restriction of civil liberties. And when we look back in the 1980s we will see many victims of market forces. I do not share the general view that market forces are the basis for political liberty, every time I see a person in a cardboard box in London I say that person is a victim of market forces. Every time I see a pensioner who can't manage a victim of market forces. The sick who are waiting for medical treatment that they could accelerate by private insurance, they are the victims of market forces. And with the disappearance of the Prime Minister who's a great ideologue, I mean her strength was that she understood a certain view of life, and when she goes, and she's gone, there will be a great ideological vacuum. It's no good saying we will run the market forces better than she did, because her whole philosophy was that you measured the price of everything and the value of nothing. And I had one experience the other day which confirmed me in my view that she hasn't really changed the thinking or culture of the British people. I don't know how many people as I do travel on trains, but I go regularly on the trains and I see all the little businessmen with their calculators working out their cash flow frowning at people, looking and glaring at each other. Thatcherite trains, the train of the competitive society. Coming back from Chesterfield the other day the train broke down and it wholly changed, somebody came in and said have a cup of tea from my Thermos. Then people looked after each other's children and a young couple talked to me and I said after about half an hour, "How long have you been married?" "We met on the train", they said. And a woman said, "Will you get off at Derby "and ring my son in Swansea because he'll be worried." By the time we got to London we were a socialist train. Because you can't change human nature. There is good and bad in everybody, and for 10 years it is the bad. And the good that has been denounced as lunatic, out of touch, cloud cuckoo land, extremist and militant. And that's what the party opposite have done, they don't quite yet know, they think it's the retirement of a popular headmistress under circumstances some might regret. Actually, they killed the source of their own philosophy and opened the way for quite different ideas. - So there you have again a leader of the Labor Party, deputy leader of the Labor Party who's extremely eloquent, but in contrast to Foot a decade earlier, very bitter about the legacy of Thatcherism. And more important for our purposes today, despite his story about the socialist train and human nature, he was actually in denial, because Michael Foot and Tony Benn continued to believe that this ghastly interregnum was now over and there would be a resurgence of the traditional left, a traditional Labor Party. By then of course Labor had been in the wilderness since 1979 and it would actually continue to be in the wilderness for the next several elections and not come back into power until 1997. And when they did finally come back into the power it was under the leadership of a very different Prime Minister with very different priorities and views about the nature of the Labor Party and how it could govern. - Across the nation, across class, across political boundaries the Labor Party is once again able to represent all the British people. We are the mainstream voice in politics today. A believe in society, working together, solidarity, cooperation, partnership. These are our words, this is my socialism. (audience applauds) It's not the socialism of Marx or state control, it is rooted in a straightforward view of society. In the understanding that the individual does best in a strong and decent community of people with principles and standards and common aims and values. A new politics, a politics of courage and honesty and trust. Now it means telling it as it is, it means not opposing everything every other party does for the sake of it. Politicians are looked to by the people for leadership, and leadership is about having the courage to say no as well as yes. Now even this week I heard people say that a Labor government must repeal all the Tory trade union laws. Now there's not a single person in this country who believes that we shall actually do it, no one believes strike balance should be abandoned. So why do we say it? We shouldn't, and I won't. I have said and I mean I am committed to the goal of full employment we will develop the plans to achieve it. But I won't ever pretend that I can deliver it overnight. It requires a modern constitution, that says what we are in terms the public cannot misunderstand and the Tories cannot misrepresent. (audience applauds) The next election will offer us the chance to change our country. Not just to promise change, but to achieve it, the historic goal of another Labor government, our party New Labour, our mission new Britain, New Labour, new Britain. (audience applauds) - So that was Tony Blair speaking to the Labor Party Conference in 1995. And the reason that was such an important speech, once that was the speech in which, when he speaks about the new Constitution for the Labor Party, he's talking about getting rid of clause four of the Labor Party's Constitution which called for the nationalization of the means of production, distribution and exchange, and it hearkened back to the original founding of the Labor Party in the early part of the 20th century as a Marxist Socialist party. And there had been a battle for the soul of the Labor Party that had gone on since Thatcher's years, actually it was before Blair, it was Neil Kinnock, and after Kinnock it was John Smith, but Blair finally won this battle, he got rid of clause four and reoriented the Labor Party to become a party that they thought could be competitive in the new world that I was describing to you last time. Now similar things were going on elsewhere, including back here. The Democrats felt that they had been pulled to the left too much in the decade since the McGovern Frazier reforms of the Democratic Party which had greatly empowered the grass roots. I'll talk more about that later, and had produced candlelight dates like George McGovern or subjected Jimmy Carter to a debilitating primary challenge from Jimmy Carter which many thought had contributed to his defeat by Ronald Reagan in 1980. And there was a similar sense within the Labor Party, within the Democrats that just as something like New Labour which Tony Blair was bringing into existence needed to happen in the US as well to the Democratic Party, and there was something formed called the Democratic Leadership Council and Bill Clinton was one of its first leaders. And the agenda of the Democratic leadership Council was very much similar to the agenda of New Labour, just to give you a flavor, this is Bill Clinton actually speaking in 1988, before he ran for the nomination four years later. - Take this fight about civil rights, the Republicans have set up so that if you're for the civil rights bill you've got to be for quotas, so that if you're not for quotas we have to say you're for discrimination. It's a bogus debate, and the White House ought to be ashamed of itself for breaking up the honest. (audience applauds) For breaking up the honest attempt of the Business Roundtable and the civil rights groups in this country to have a new choice you can have economic growth, small-business vitality, you don't wake up every day being scared to death of a lawsuit, but we protect women and minorities and people who deserve it from unfair discrimination on the job, which we all know still exists in this country. (audience applauds) Take the debate about poor children, the way the Republican's set the debate up, they say the Democrats are for throwing more money at these problems and we know you can't throw money at them. We just said that. And we are family values. Let me tell you something, family values won't feed a hungry child, but you can't raise that hungry child very well without them. We need both. Do you really believe that if we permit these conditions to go on for 10, 20 or 30 years And we permit national politics to continue in its present relevant track for 10 or 20 or 30 years, that America will lead the world we have made, that you can keep the American dream alive for the next generation of Americans? I want my child to grow up in the America I did, I don't want her to be part of the first generation of Americans to do worse than their parents did. I don't want her to be a part of the country that's coming apart instead of coming together. Over 25 years ago I had a professor of Western civilization who told me our country was the greatest country in human history because our people had always believed in two simple things, one is that the present doesn't have to be as good as the future, the future can always be better. And two, that every one of us has a personal moral responsibility to make it so. That is what the new choice is all about, that is what we are here in Cleveland to do, we're not here to save the Democratic Party, we're here to save the United states of America. Thank you very much and God bless you. (audience applauds) - So there was Bill Clinton on the cusp of his leadership role in the Democratic Party and he would subsequently go on to become President. And so today's agenda is to understand this reorientation of the parties of the left and then think about their place in the broader scheme that we are developing as the course goes a long. Our agenda is gonna have three parts to it, I'm gonna start by revisiting the comments I made about distributive politics or the psychology of distributive politics in a little more depth than I did last time. And the reason is I've had a number of questions about it, and it's really important. It's gonna be important later in the course and it's gonna be important for today's lecture. So I'm gonna reprise what I said about that and extend it a little bit. Then we're gonna come to the heart of the matter and talk about unions, left of center parties and the median voter as they played out in this period. And then finally we're gonna talk more generally about distributive politics in two-party systems such as Britain and the US, versus multi-party systems as prevail in much of Europe. So let's go back to the comments I made about absolute and relative gains. I said to you that absolute gains are self-referential, as in the Ronald Reagan video that I put up saying ask yourself and I better off than I was four years ago. Relative gains are the sorts of things that Margaret Thatcher was lambasting in her speech in the House of Commons when she was saying that the left would rather that the poor will be poorer so long as the rich were less rich. So just for people who like diagrams, some people like diagrams and some don't, and there's nothing I'm gonna do with a diagram that you can't explain without a diagram. But for those who like diagrams, imagine you're gonna ask yourself which is better for me from the point of view of you representing the red dot. And we're thinking about my income on the vertical axis, and her income on the horizontal axis. So just to be clear about what we're saying, with absolute gains anything above that line is better. I will be better off than I was four years ago if I can move above that line. When we're talking about relative gains, or other referential comparisons, then anything to the north-east of that, to the left side of that blue dotted line, I will be better off relative to her. And so we can think about four quadrants here, we can think about up here I'm better off either way, over here I'm worse off either way, here I am absolutely better off but relatively worse off, and here I'm relatively better off and absolutely worse off. So this is the difference. Now economists like to think in terms of absolute gains because they think in terms of win-win. So this is a close cousin of the Pareto principle, those that have taken Econ 101 will know, if we move from the status quo up here both people are better off, so that's the realm of market transactions, they'll go there voluntarily. If we move from here, maybe to here, maybe because the government taxes both of us and sends the money to a country we both despise then we're worse off. And these two quadrants are where one person gets better off and the other person gets worse off. And much of economic theory is about welfare economics is about whether and under what conditions we can say that the one who gains gains more or less than the one who loses. So the reason economists tend not to like relative comparisons is that there's no win-win with relative, there is no circumstance with relative comparisons where both people can be better off. And so I think that's a lot of the reason for the distaste for relative comparisons among neoclassical economists. But, this is not a course in neoclassical economics this is a course about power and politics in today's world. And when we think about power and politics we have to recognize that although economists tend to prefer absolute gains, relative gains are often much more potent in politics, people make relative comparisons whether our models tell us a good idea or not. But they don't just make any relative comparisons and this was the point about many people not caring what multimillionaires make. Local reference groups, people relatively similar to me are the sorts of people to whom I tend to compare myself, that was the upshot of the capuchin monkey story and why I said that the experimenter misinterpreted his own results. The example I gave is a college professor will be much more upset to learn that his salary is $10,000 less than a similar professor down the hallway than to learn that it's half $1 million less than the attorney who lives next door. So people tend to make relatively local comparisons. But a corollary of that that I didn't emphasize enough is that these local comparisons can be either invidious or solidaristic. So when we talk about the majority rule, divide a dollar game, I said if people agree upon a conception of fairness for example, that everybody should get an equal share because it's fair, that might breed solidarity among them to vote together if they're all below the median income. But the difficulty with that is that it's vulnerable to that very logic of the divide a dollar game that I talked about, namely that you can split people off of the coalition by offering them some other benefit, unless there is some institution to actually hold together and reinforce that solidarity among a particular group, it's gonna be too easy to pick it apart, and we will in fact see in four-part harmony on Thursday that efforts to pick apart a distributive coalition in accordance with the divide a dollar game when we talk about the anti-tax movement. And so the reason we got to talking about unions was that the thought was that unions might be an instrument for supplying some if you like institutional backbone to the solidarity of people below the median income. And I'm gonna spend a lot more time on that topic today when we're thinking about the reorientation of left parties. But before I do that I wanted to make one more point about the psychology of distribution and the way that it plays out in politics. And I want you to engage with me in a thought experiment. Let's say that Steve gets a phone call one day and he's told that he's won the lottery. He's won the lottery and he's gonna get $2 million. The next day however, he gets a follow-up phone call and they say well actually the news is good but it's not quite that good, we didn't know there was a second ticket that had been sold, so actually you're not getting $2 million but you're getting $1 million. Still, it's $1 million Steve didn't have $1 million before he had been called. Now Art on the other hand, he gets a phone call and he says, "Art you've won the lottery, "you've won half $1 million congratulations." And he's delighted. But the next day he gets another call and he says, "The reason we told you "that you had won half $1 million "we thought there was another ticket sold "but it turns out there wasn't another ticket sold. "So actually Art there is even better news, "You've won $1 million." Now notice that both Art and Steve are $1 million richer than they were two days ago. But who's happier? Who's happier? Why? Pardon? (people mutter off microphone) Art, you tell us why are you happy? Yeah, so what you're all saying is exactly right, that Steve somehow thinks he's lost something. And this is the concept of loss aversion which economists sometime refer to as the endowment effect. This is the idea that if you have something taken away from you it bothers you more than you become happier if something is given to you. And this is the idea for which Danny Kahneman won a Nobel prize, it should really have gone to him and Amos Tersky, but they don't give it to dead people and Tersky had died. And again for people who like diagrams this is the diagram they called it Prospect Theory and the idea is the curve is always the shape. The idea is to capture this notion that the prospect of a gain does not bring as much happiness as the prospect of an equivalent loss brings unhappiness. And so this is why if you want to motivate people to do something it's better to talk to them about undoing a loss or avoiding a loss then the prospect of a gain. So we will come back to this idea in the last part of the course because I think it has a lot to do With the psychology behind the politics of populism that we'll be talking about later. So this is the notion of loss aversion. Now some people wonder about well why is it? Because at some level it's irrational to weigh potential loss more heavily in terms of futility than weighing a potential gain, particularly if the things are equivalent, the endowment effect idea is if you own something you won't take the price you would take if you didn't own it to give it up. Again just this idea that it's mine, I don't want to give it up is so important. And there's speculation about why people are like this. One of the leading theories is that for so long in human history people lived so close to the survival that even a small loss might have been catastrophic from the point of view of survival and so people have built in this idea of avoiding losses as being more consequential than getting gains. It's one theory, there are others. So we're just keeping all that in our analytical bag of tricks as we work our way through the rest of today's lecture and some of the other materials in the course. So, the basic challenge for left of center parties in this new world, is that unions as institutions that can reinforce solidaristic ideals among voters below the median income are becoming less effective. This is the basic challenge. And the reason is why? Because unions are becoming, unions protect their members and are becoming smaller and weaker. And that means two things, they're gonna be less effective at protecting their members, and the externalities of protecting their members when they do, the spin-offs are gonna be less, less dissent. So that's what we are gonna be exploring, and I put this picture up last time, when we were talking about the right to make the point that in the US people think about the decline of the unions since the 70s, but really the high point of union membership in the American economy was in the 1950s, and it's been falling ever since. So the UK is a comparable story, just later. You can see here union membership in Britain peaks at around 1980, ironically just as Michael Foot is becoming leader of the Labor Party and we'll have more to say about that when we talk about parties later. And it's been going down ever since, so if you look here at trade union density, that means the proportion of a sector of the economy that is unionized. Since 1995 in the whole economy that's the blue line, it's gone from about 33% to about 25%. And if you look at the private sector, which is increasingly important because the public sector is shrinking, it's gone from above 20% down to about 15%. So not as extreme as story as here, but nonetheless the size of the unionized sectors of the economy has been falling. Now the upshot of this is that unions are decreasingly effective institutions to enforce solidarity below median income voters. And this requires quite a bit of unpacking. The reason is that where unions remain powerful within left of center parties they will pull them away from the median voter. So you saw this for example in the 1970s and 80s in the UK. Where unions were very powerful within the Labor Party, they controlled the leadership selection process for example, that's how Michael Foot became leader in 1980. And the TUC, the trade union organization, had, it has less now, but it had at that time enormous power in setting Labor policy. But if you think about union membership declining, that means that if they support unions they are gonna be pulling the Labor Party away from the median voter most of whom are not unionized. And indeed one way of talking about this story of British politics in the 1970s and 80s, certainly until Thatcher in the late 60s and 70s is this divergence of interest. So both Conservative and Labor governments alike used to have things that they called incomes policies. And incomes policies were basically agreements that they worked out, it sometimes went under the name of liberal corporatism among academics. You might think liberal corporatism is a contradiction in terms, when we say corporatism we think of Italian fascism. The idea of liberal corporatism was the idea that the government would negotiate a deal among business and the unions, and then turn it into law and enforce it. And so the Heath government, the Wilson, Heath was a Conservative, the Wilson and Callaghan Labor governments were always trying to negotiate these incomes policies with the unions. And they would agree that wage increases are gonna be capped at 6%. And then what they would find was that the unions say in the steel industry or the coal industry figure that they could actually negotiate a higher settlement with their employers, and would bust the agreement. Even under Labor governments, Wilson and Callaghan would call in the union leadership and say, why are you doing this, you agreed to hold the line on wage increases, as part of our incomes policy. And the answer was we are here to represent our members, we are not here to run the British economy. And our first obligation is to our members and if we are in a position where we can negotiate better raises for our members why should we not do that? They are the ones who pay our dues they are the ones who are for us. So to the degree that unions controlled the Labor Party, they're gonna pull the interests of the party in the direction of the unions. And so this was a reason why we had such hectic politics in Britain in the 1970s many strikes. I lived there in those years, three day week, garbage in the streets. Eventually who governs Britain became the slogan in the 1979 election, which Thatcher won. Who governs Britain the unions or the government? Governments both Conservative and Labor had been unable to govern effectively. If you think about another pretty extreme case, in South Africa we have an extremely powerful union movement The reasons for its power will concern us in a subsequent lecture, they trace all the way back to the transition in the 1970s, but it was extremely powerful union movement That is part and parcel of the governing African National Congress Alliance. And so unions protect their members and so would you have by African standards high wages in the formal sector of the economy which is protected by the unions, but you have 30% unemployment and 50% almost youth unemployment, third highest in the world and the rest of the economy. You have a circumstance in which textile jobs are hemorrhaging to Lesotho, and even at one point they were hemorrhaging to China. So unions if they can will protect their members, but that may come at the price of employment in the rest of the economy. At the other end of the continuum, a system that prevailed in Germany particularly in the 1980s. And still prevails in some sectors today, the system in Germany was that if the unions negotiate a bargaining agreement in a particular sector of the economy, even the non-unionized plants or companies would abide by that agreement. So that's a system where unions, the interests of unions and the interests of all workers, become much closer to aligning. If even people in the non-unionized sectors are getting the same deal as the unions the big unions negotiating with the powerful companies are negotiating for their members, then you can see that unions can function better as a spine behind the solidaristic commitments of voters below the median incomes. So to the extent that the economy is more like the German economy in the 1980s and less like the British economy in the 1980s or the South African economy today, unions can in the German kind of case be a source of policy that's gonna be appealing to the median voter, or more likely to be appealing to the median voter. And so one of the things that we have to think our way through is that without reinforcing institutions solidaristic ideals among people below the median voter will be very hard to sustain. So this leads to the logic of what we call sometimes triangulation. So just to give you a couple of examples here. - They're a new generation of Democrats, Bill Clinton and Al Gore. And they don't think the way the old Democratic Party did, they've called for an end to welfare as we know it, so welfare can be a second chance, not a way of life. They have sent a strong signal to criminals by supporting the death penalty, and they rejected the old tax and spend policies. Clinton has balanced 12 budgets and they have proposed a new plan investing in people detailing $140 billion in spending cuts they would make right now. Clinton Gore, for people, for a change. - So there you see the Clinton campaign moving very far to the right. To capture the middle ground because they no longer think that they can get the solidaristic support of people on the left. And indeed, Clinton felt that he actually had to attack the left of the Democratic Party to get his bona fides as someone who could one in this new world. And this is his famous speech about Sista Souljah. - You had a rap singer here last night named Sista Souljah, I defend her right to express herself through music but her comments before and after Los Angeles were filled with a kind of hatred that you do not honor today and tonight. Just listen to this, what she said. She told the Washington Post about a month ago, and I quote, if black people kill black people every day, why not have a week and kill white people. So you're a gang member and you normally kill somebody, why not kill a white person? Last year she said you can't call me or any black person anywhere in the world a racist, we don't have the power to do to white people what white people have done to us, and even if we did, we don't have that low-down dirty nature. If they are any good white people, I haven't met them, where are they? Right here in this room. (audience applauds) That's where they are. I know she is a young person but she has a big influence on a lot of people. And when people say that, if you took the words white and black and you reversed them you might think David Duke was giving that speech. - So that was very dishonest in its presentation, because she had not actually said that, she had talked about a rap singer singing that. It wasn't her music, she was talking about a rap song that made those claims and was talking about the claim. She wasn't herself taking those views. But it didn't matter, this went viral, or the 1992 equivalent of viral. Because it was Clinton distancing himself from the African-American affirmative active agenda as it was portrayed by his critics as a way to try and move to the center. And so the thought here is, the motivating thought is in a two-party system if you move to the middle, the people on your flank have no place to go. So when Trump was running in 2016 and he said to African-Americans, vote for me, what have you got to lose? He was exactly saying that they don't get any attention on the left of the Democratic Party, so the Democratic Party can take them for granted and ignored them. And so this becomes the argument for the politics of triangulation, which I'll have more to say about on Thursday. But the idea is that you move to the middle to peel off support from the other side, and enough support to win. And you might depress turnout a little bit on the left of your own party if you're a Democrat who is triangulating, but at the end of the day they have no place to go, and so it seems like a winning strategy. And so this idea of triangulation that was dreamt up by people like James Carville and others and very much also informed Tony Blair's approach in Britain. So there the race issue is not salient in the way it is here, but as you saw when he was talking at the 1995 Labor Party Conference, he said, we are not gonna undo the Thatcher changes to labor laws and in fact he did not undo the Thatcher privatizations, actually they went into additional privatizations, and the did not bring back regulation. And indeed, part of the bitterness against Blair on the left of the Labor Party would turn out to be that they thought he was more Thatcherite than Thatcher. And so if you have this situation where the traditional source of solidarity on the left of a left party has now dissipated because of unions are too weak to supply it, or they are too segmented in their agenda is that they are gonna follow. The path of least resistance from the point of view of getting elected would be to move to the center, try and peel off support from the other side, and know that the people who are on the left of your party might be angry, but at the end of the day, they have no place to go. So that is in a nutshell the story about why in two-party systems we saw this logic of triangulation most dramatically with Clinton and Tony Blair. And you can understand why the desire to win on the actions would drive politicians to do that. But what about multi-party systems? Multi-party systems are different in their basic functioning and in their basic logic. So the conventional wisdom about multi-party systems is the following. Multi-party systems, people like them, they say they are more representative. They are more representative because in a two-party system we see the parties are heading for the median voter. But what if you're not the median voter, you feel unrepresented. So in a multi-party system, the Greens have their people in parliament, the anti-immigrant crowd have their people in parliament, the Free Democrats in Germany, the libertarian people have their people in parliament. Everybody has a seat at the table, and so that seems much more representative how many people think it's more representative? How many people think it's not? So the more representative have it for the moment. This is a topic we're gonna revisit in considerable depth later in the course. Because the argument on the other side is it's more representative at the electoral stage, but that doesn't mean it's more representative at the governing stage when governments have to be formed and coalition governments have to be put together. So we'll come back to that whole subject. But for a long time the conventional wisdom was that, it's put up here, Bingham Powell's book, "Elections as Instruments of Democracy". But the conventional wisdom was that multi-party systems over time the produce policies that are more responsive to the median voter, and that therefore they are more distributive. So if we go back to what we were talking about, the median voter theorem a couple of lectures ago, the claim was, and the evidence seemed to support the proposition that multi-party systems are more redistributive because of what you would expect if you think that the people below the median voter would want more redistribution. As you can see, Bingham Powell's book was published in 2000. And that means that most of his data is at least several decades old now, it's all from the last century. But this picture of inequality that I showed you a while ago captures that intuition and that inequality went, this is the gilded age, this is today, we all know about the U-shaped curve that we went to much less inequality in the middle part of the century, and then since the 1970s it's been climbing. But then if you look by national, by country you can see that the U-shaped curve is flat in Germany and France than in the US and in the UK. So that's consistent with the suggestion that these systems are less inegalitarian, or less susceptible to inegalitarian outcomes then the two-party systems like Britain and the US. Of course for me to say that this graph supports that idea would be to commit to selecting on a dependent variable that I was talking about last time, so I'm not making a causal argument here. But this is just describing what was the conventional wisdom. And that's indeed why many people on the left prefer the European systems, they think that they are more likely to be egalitarian, more likely to be redistributive. So is it the case? And the thing that we have to start thinking about is that, Unions have also been declining in those systems. Now why might that matter? I put up a lot of graphs that would be too small to see, but basically you can see them when I post them. In virtually every single major European economy union membership has been declining in the last several decades, just as it has been declining in Britain and the US. Some at different rates. There are a couple of outliers, I mentioned Finland. Even in Finland in the last five years, union membership has fallen by 5%. And another outlier is Iceland, Iceland is a very tiny economy and it's not exactly clear why they have such high rates of unionization, but even in Iceland there's been a kind of backlash in one of the biggest unions just in the last couple of years, actually there's been a insurgency to take over control of the union because they feel it's too much of a company union, too close to the interests of the company's. So in any case, Iceland and Finland to one side, there's decline of unions everywhere for the reasons that we've talked about before. Namely, globalization of capital, that makes it easier for capital to move around. The exit costs for capital are low, the exit costs for labor are high, the Hershman story. And of course increasing, increasing jobs going to technology, that indeed may be more important than globalization going forward. A McDonald's restaurant that used to have 20 people working in it that you now see eight people working in it, five or eight years from now may have one person working in watching a robot making burgers. So we've seen this long term decline of union membership that's been going on everywhere pretty much. And that has two implications that I want to now dig into a little more deeply. The one implication is that left of center parties are gonna be less effective in protecting employed workers. So why might this be the case? I said unions protect their members. But the fewer members there are in an industry the less leverage the union has in negotiating settlements. You see this is going on right now, this isn't in a multi-party system, but you see this right now going on in Detroit. Why are those workers on strike, they are on strike or threatening to go on strike because the GM cannot compete with the foreign car makers working in southern states of the US, the so-called right to work states that are not unionized that are paying 30 bucks an hour, whereas the union agreement is more like 40 bucks an hour. So these unions are less powerful in negotiating with employers, GM could choose not to produce electric cars in Detroit and start producing them elsewhere. And so you would expect unions would be less able to negotiate strong deals. Whether or not it's part of the corporatist arrangement subsequently enforced by the government. So Germany, I said Germany was in the 80s was the heyday of what political economists were calling liberal corporatism, these big deals done between the big unions and big business and governments would get behind them. But in the early part of this century, interestingly when Gerhard Schroder as the last social Democratic chancellor was in a governing alliance with who? The Greens. Not a grand alliance with the CDU, but with the Green party. They implemented the so-called Hartz reforms, and the Hartz reforms were not that different from what was done by New Labour in the UK, and not that different from what was done under the Democrats in the US. The new Democrats under the US, there were market friendly reforms designed to reduce unemployment, to protect unemployment, reduce unemployment rather than protect workers wages. More market flexibility. Hartz, the author of the Hartz reforms had been head of Volkswagen, eventually actually resigned in disgrace a couple of years later for unrelated reasons, but he had designed these very business friendly reforms that were implemented over the first few years. There were four stages of them. In Germany their form of moving to the middle, so that the Social Democrats even when their only alliance partner was the Greens couldn't resist this tide. They were no longer in a position to force more generous settlements to their workers as they had been in the past. And by the way, even though Schroder implemented the Hartz reforms, it still didn't stop him from losing to Angela Merkel in the 2005 election, and she's been Chancellor of Germany ever since. So the move to the center, even in a multi-party system with a labor oriented party was in the driver's seat of the government was unable to do any better because of the declining power of the unions. And the other implication is that to the extent you can protect unionized workers, that may impose costs on others below the median income. Service sector workers who were notoriously difficult to organize even though these economies are all increasingly becoming service second economies, and the long-term unemployed. Because if you think about the union it's going to protect the short-term unemployed, namely their members who might recently have become unemployed and might be going back into the workforce, but they don't have an interest in protecting the long-term unemployed. On the contrary, there may be a trade-off between creating employment for the long-term unemployed and protecting the wages of unionized workers. So it's not surprising that even in multi-party systems you might think that unionized workers as represented by traditional left of center parties Are going to be less well protected and less effective as instruments of solidarity Among others whose income is below the median voter. So, how might this play out in a multi-party system? In a two-party system as I've said, the path of least resistance is to head for the middle and let the people on the flanks of your party stew. Multi-party systems are different because there are plenty of people to form coalitions with. And so this is, I should say that the slides that I'm gonna show you now are very much from work in progress that I'm doing with Francis Rosenberg and some economists, postdocs, that is somewhat tentative but there are some pretty dramatic conclusions I think on some matters. So you can see on here, this is from 1960 to 2020. 26 OECD countries, the vast majority of them multi-party systems. The number of parties represented in the legislature, the number of parties differs from the number of different effective partners, only in that effective parties weight them by their size. But what you can see is the number of parties in the legislature has gone up. We see fragmentation of political parties. And if you look at the left parties in these 26 OECD countries, what you can see is that while the voting share, which is the black line is pretty much flat, it's gone down a little bit since 2010. The number of parties has gone up dramatically. Why might that be? Why might we think, given what I've been talking about here, why would one think that we would start to see more parties, fragmentation just means more parties? Yeah, you're gonna have to yell, we forgot to get the microphone today. You'll have to scream. - [Man] If you move to the center then somebody can move to the left and take your left flank? - Bingo, that's exactly right. So just to give an example, the Social Democrats moved somewhat to the center. And alienated people on their left and those people then were up for grabs by other parties. It's not the case that they have no place to go. And so one possible cause of this is just another way of saying, what was said at the back of the room, that as industrial jobs have gone down. We can see here industrial jobs falling over this period quite dramatically, the number of parties has gone up. And so as these parties become weaker and represent fewer workers and therefore move in search of new coalition partners, other parties are gonna emerge because they can get some seats and they can get a place in the legislature. And it's interesting that if you look at fragmentation of right-wing parties there's no systematic shape, at least not since 1990s. If you try to draw a line through that it would be pretty much flat. So it's really more fragmentation on the left. And if you look here, where has the SPD's vote gone to? It's gone all over the place. So this is comparing, this is the 2013 election, the 2017 election in Germany. And you can see they have lost votes to the Alternative for Deutschland, they've lost votes to the Free Democrats, they've lost votes to the Greens, and they've also lost votes to people who stopped voting down here. So as the effects of the Hartz reforms and this move in the neoliberal direction in the SPD has accelerated, by 2013, by the way they were in a governing grand coalition with the CDU. So it was a very different kind of government. And many of their traditional supporters indeed thought that they had been selling out, giving away too much to Merkel to be in the coalition. And as I said to you in the very first lecture of the class, that in 2017 when they got hammered, they said we're not going back into another grand coalition because all we do is erode our base because of the agreements we have to make to be in a grand coalition with the CDU. And it was only after seven months of Merkel failing to make other coalition arrangements while the AFD was increasing its popularity in the polls, frightening everybody at the prospect of another election that the SPD finally went back into a grand coalition. It didn't help by the way, I mentioned to you in that lecture that the following year in the regional elections in Hesse and Bavaria that both the SPD and the CDU continued hemorrhaging support. Mostly the SPD to the Greens and mostly the CDU to the Alternative for Deutschland. Since then we've had local elections where the AFD has again cleaned up. So that is not gonna abate any time soon. So just again to, I don't want to talk about a lot of this but this again preliminary results. One of the things you see in that circle is that what might be, again this is not causally established, but what we're observing in this data. What might be the case that when left governments are in power as part of coalitions, unemployment actually has not gone up in these 26 OECD countries by about a percent. So again, this is not a causal argument, but it's consistent with the thought that these left parties are less able to protect the people below the median voter. That they're protecting a shrinking labor aristocracy less and less effectively and with fewer and fewer positive spin-offs for people around and close to them in the income distribution. So the upshot of this is that unions in multi-party systems are protecting shrinking groups of industrial workers less well than they used to do. And party fragmentation makes solidaristic ideologies among voters below the median voter more difficult to sustain. And this is why we see the hemorrhaging of all of these voters on the slide I put up earlier. So just to finish out the German story, what happens is that as the SPD starts to lose voters, Merkel thinks oh, I could pick up some of these voters, so she also moved toward the middle. That then opened up space on her flank for the emergence of the Alternative for Deutschland party, and so to some extent fragmentation on the left will produce fragmentation on the right in response to that. And make it more difficult for parties on the right also to form predictable compilations. And so the sorts of outcomes we have been seeing in European elections in the last few years are not that surprising. Namely you get inconclusive results in many of these elections, or you get situations where you need three or four parties to form a government. And the traditional in Germany since we we're talking about Germany, traditionally in Germany the two biggest parties were the CDU and the SPD, and they were, it wasn't that different from a two-party system, it was very predictable that they were gonna dominate center-left and then center right, and then center-left were gonna dominate coalitions over time, so they operated more or less like two-party systems. Now all bets are off as to who will be in that coalition next time. Mostly because of the decline of center-left parties. And you can look in country after country after country where the left of center party traditionally always would come in first or second. It doesn't matter if it's Germany, it doesn't matter if it's Holland, it doesn't matter if it's Israel. The Labor Party in Israel always came in first or second. Now these parties come in fourth or fifth reflecting their much weakened state as representatives of organized labor within the electorate. So party fragmentation means that to the extent unions and the political parties that are answerable to unions are sources of solidarity of voters below the median voter, that has really gone away. And that means that even though it's a sort of commonly repeated mantra that PR systems are more redistributive, PR systems are more responsive to the median voter, the truth in today's world is that PR systems are simply vastly less predictable than they used to be, because of the proliferation of parties, nobody has any idea who is gonna be forming the government. And sometimes, so we have a grand coalition in Germany right now, but it could be parties from the two extremes as happened in the Greek elections of 2015. When you have a far left Populist party making a coalition with the far right populist party. What they shared in common was their antipathy for Europe, but very different alternatives to your membership being envisaged by them if they will actually in the position to implement their policies. So the traditional advantages of PR, which produces multi-party systems seems very much in question. So the last implication of this discussion is that the proliferation of far right parties in these systems is worrying, we talked about this in the very first lecture, but at the end of the day it's not that surprising because it's been driven by the logic of competitive politics in this new world that has emerged since the Cold War. Next we will talk about shifting the goalposts, the anti-tax movement in the United States. See you on Thursday. (gentle music) |
Power_and_Politics_in_Todays_World | Lecture_2_From_Soviet_Communism_to_Russian_Gangster_Capitalism.txt | - So, in the interests of, again, both taking some people down memory lane and recreating where we're coming from for those who are too young to have experienced it, let's go back to August 19 of 1991 in Moscow. - [Reporter] Mingling with the rush hour traffic, Red Army armored personnel carriers on the streets of Moscow this morning, heading to the Kremlin. They first moved in at four a.m., the first sign of the coup d'etat that removed Mikhail Gorbachev from power. With tanks in Red Square, the official word from the new government calling itself the National Emergency Committee, was that the architect of glasnost and perestroika was too ill to continue in office. A short statement reminiscent of the Soviet Union's past was broadcast by state television. - [Reporter] By mid-morning APCs were ringing the Defense Ministry and most government buildings. For many people on the streets, the reaction was one of sheer surprise and resignation. - So that was a pretty dramatic moment in August of 1991. And the coup of course was being prosecuted against Mikhail Gorbachev, who had come to power as the head of the Soviet Union in 1985. Leonid Brezhnev had been General Secretary of the Communist Party for a number of decades after Khrushchev, but he died in 1982 and it was an aging Politburo, and he was replaced in quick succession by two transitional figures, one of whom died after 14 months, and the other who died after 11 months. And so when Gorbachev came to power in February of 1985 there was a lot of head-scratching going on in the West as to what would become of the Soviet Union. And he rapidly transfixed people both at home in Russia, as the Soviet Union as it then was, and abroad. He seemed to be a different kind of politician, he was extremely charismatic. He was much younger. He could talk and behave like a Western politician and he quickly went on a kind of charm offensive around the world. He developed a strong rapport with Ronald Reagan which led to arms reduction talks in Reykjavik and he began the processes of what he called perestroika and glasnost. Perestroika is Russian for restructuring, and he planned fundamental restructuring of the Soviet economy. He had become convinced that the old command system didn't work and had to be reformed. Glasnost was basically free speech, that they began opening up the state controlled media, they started creating much less, much more access to different media outlets. They allowed criticism of the regime that had been unthinkable until that time. And it seemed like, it seemed like the Soviet Union was really beginning to change, and then of course in 1989, when the revolutions happened across Eastern Europe that I talked about last time, the Soviet Union did nothing. They didn't intervene as they had previously intervened in Hungary, they did not do anything to stop the departure of the East European countries from the Soviet Bloc, and the transitions that were by then would soon be getting underway. And Gorbachev was seen as this radical reformer. He was a reformer, he was not a revolutionary, and he thought that he could adjust the Soviet system in a gradual way to some new order. But as he did that, of course he ran into a lot of push-back from vested interests, from communist ideologues who didn't want change. And he was fighting a battle, both against them, and as we'll see in a moment, against reformers who thought that he was going too slowly. In any event, in the summer of 1999, 1991, the hardliners clearly had the upper hand and they locked him up in his dacha on the evening of August 18th, and that was the coup that appeared to be well underway by the next day. But things were not so simple, and history was gonna play out a little differently. So this is what happened a few minutes later. - [Reporter] Tanks were also positioned outside the Russian parliament building, Boris Yeltsin's headquarters. - [Crowd] Yeltsin, Yeltsin, Yeltsin. - [Reporter] The democratically elected President of Russia was soon striding out of the building to address a crowd of supporters. His own radio and television stations by now occupied and forced off the air, he climbed aboard one of the Red Army's own tanks, and said the coup leaders had disgraced the Soviet Union. (crowd applauding) - [Reporter] And he went further, calling for civic resistance to the hard-line grab of power. (crowd shouting and applauding) (crowd cheering and applauding) - [Reporter] Yeltsin's supporters reacted with their bare hands, building makeshift barricades with whatever piece of disused machinery or brickwork was available. And one crane driver joined the effort, shifting concrete blocks to the delight of the crowd, helping build the obstacles designed to prevent any army attempt to seize the parliament building. No attempt was made. One army commander loyal to Boris Yeltsin, Colonel Konstantin Kobets, tried to reassure the crowd. (crowd shouting and applauding) - And that of course was a decisive moment because it indicated that the military was not supporting the coup. And that instead they were gonna be loyal to Boris Yeltsin. Now Yeltsin had himself been a traditional Soviet politician in the 1970s and 1980s and had initially been a supporter of Gorbachev's reforms. But he actually was on the more radical side of the Communist Party, at least at that time, in thinking that the reforms had to go faster and had to go, be more extensive. And he in fact got so frustrated that he did something that at that time in 1987 was unprecedented in the history of the Soviet Union, he resigned from the Politburo of the Communist Party. And nobody had ever done that before, there was consternation and it was sort of covered up in that in the end Gorbachev fired him and there was then some pretty significant conflict between them. It is said that in the midst of all of this Yeltsin actually tried to commit suicide at one point. And in any event he left the Communist Party and started to reconstitute himself as a Russian politician and eventually would run for the, to be a member of the Russian parliament. He had done that and in fact in June, in June of 1991, so we're talking about three months before the coup attempt he had been elected President of the Russian parliament. And where they were in this video outside the White House, that was the building of the Russian parliament, as distinct from the Soviet government buildings, which we all know about, the Kremlin opposite Red Square. And so it was Yeltsin who stood up and faced down the coup plotters while Gorbachev was in his dacha locked up on the Black Sea. And then the military decided to get behind him, and the next day you can see Gorbachev returning. - [Reporter] A president returns to power. Mikhail Gorbachev, flying into Moscow after being detained for three days on the orders of some of the men he appointed to senior positions within his own government. Their attempted coup d'etat crushed, he said tonight it was business as usually at the Kremlin, and despite the treachery of many of its senior figures, he wouldn't be leaving the Communist Party. - Gorbachev said tonight he'd made mistakes by putting his trust in men who would try to topple him. Many here now believe he's making another mistake by staying within a Communist Party that's so widely discredited after the events of this week. For World Monitor, I'm Simon Marks in Moscow. - So this was in fact the moment when the real power passed from Gorbachev to Yeltsin, as you'll see in a minute because Gorbachev couldn't detach himself from the idea that there had to be a gradual guided transition under the auspices of the Communist Party of the Soviet Union. And Yeltsin seized the moment and this is how it played out. (crowd shouting and chanting) - [Reporter] A president gets the cheers. He used the occasion to call for increased power for the Russian government and thanked his people for staying loyal. (audience applauding) - [Reporter] Present at the rally, many of the men who once served with Mikhail Gorbachev. Economic advisor Stanislav Shatalin, and former Foreign Minster Eduard Shevardnadze. The men who helped mold glasnost now siding with the man of the hour. The Mayor of Moscow, Gavriil Popov, said to regain the support of the democrats, Gorbachev had to break with the Communist Party and recognize its seven decades in power were at an end. Joining the celebrations, Gorbachev's former press spokesman Sergei Grigoriev and his wife. After three nights on the barricades, he says his old boss faces a fight to take the political high ground. - But I think that he will have to define his attitude and to elaborate his position very quickly, otherwise, his days in power might be counted. You can see judging by the rally and by the attitude of the people, that the democratic movement is actually gaining strength. - [Reporter] At the end of the rally, the Russian flag was hoisted above the parliament building in place of the hammer and sickle. Ships on the River Moskva blew an audible salute. (crowd chanting in foreign language) Within an hour the crowd was heading to the seat of Soviet power, Red Square and the Kremlin. A tide of people bringing with them the banner of independence, a huge Russian flag, and shots of down with the Communist Party. (crowd chanting) The young guards on duty at Lenin's mausoleum looked as if they were about to be mobbed. When the time came to change the guard the normally solemn ceremony was derided by the crowd, a Communist shrine now openly mocked. (crowd whistling) (crowd shouting) From Red Square they moved on. Next stop, KGB headquarters and the statue in front of it of Felix Dzerzhinsky, the man who founded the secret police here in the 1920s. Yeltsin's spokesman promised them a crane would be brought in to bring the monument to the ground. Former Gorbachev supporters say the Russian president now has a better feel for the popular sentiment than the leader in the Kremlin. - I think that with his democratic instincts and his feel what the people want, and what the people need, he is really turning himself into man number one in the Soviet Union in terms of popularity and in terms of appeal to the mass. (crowd cheering and applauding) - So that was the dramatic four days that were really the denouement of the old Soviet order. Today, our agenda is gonna be five-fold. We're gonna talk about why the Soviet Union collapsed, but only somewhat briefly. We're then gonna talk about the more interesting puzzle which is why it collapsed so peacefully. Why basically one of the two superpowers in the world folded up and just died. We're then gonna focus on the rise of gangster capitalism, the title of today's lecture. We're going to turn to the transition from Yeltsin, who governed throughout the 1990s, he gave up power to his hand-picked deputy Putin on the first day of the new millennium. He resigned effectively December 31st, 1999 and Vladimir Putin has been in control of Russia in one role or another ever since. And then lastly we're gonna focus on the question why is Russia so corrupt, and what lessons can we draw from the fact of its corruption. So that's what we'll do. Just to review the events that led up to those four days in August. In May 12th, the Baltic states, echoing an old agreement from 1934, having watched six months earlier the wall come down in Germany, signed an entente that indicated they were planning to secede from the Soviet Union. In June of that year the Congress of People's Deputies in Russia, that is where Yeltsin was run initially as a deputy and then become president, declared that it was sovereign over Russia's territory. In March of 1991 a referendum was held on whether or not to keep the Soviet Union together, and there was high turnout and a strong majority saying yes, although there were boycotts from the Baltics and four or five other Soviet republics that didn't participate in the referendum. June of that year, exactly a year after the Russian parliament, Congress of People's Deputies had declared itself sovereign, Boris Yeltsin had been elected as its president. We then had the four day attempted coup in August as we've just witnessed. In the midst of that coup I should say, the day before it was finally over, Estonia and Latvia had already declared their independence. The coup was then defeated. Ukraine held a referendum and voted to secede, and then Yeltsin met with the Ukrainian president secretly and agreed that the Soviet Union no longer existed. Gorbachev was still chasing the horse out of the barn that had long since left him behind, he declared the agreement illegal and dangerous. But in December leaders of the 11 remaining republic met in Alma Ata, now know as Almaty, in Kazakhstan to essentially declare the Soviet Union dead and constitute the CIS, the Confederation of Independent States. At that point Gorbachev realized he had no choice, and he resigned on what we think of as Christmas Day, but not Christmas Day in Russia, which comes later, and at the end of the month the Soviet Union ceased to exist. So why did the Soviet Union collapse? Any thoughts? What's the first thing that comes to somebody's mind when we ask, why did the Soviet Union collapse? - [Student] Arms race. - There was an arms race. And why would that have, we'd had plenty of arms races. We had arms races all through the Cold War, so it certainly contributed, but why else, yeah. - [Student] My parents remember how they were waiting to buy sausages, like in the market. You couldn't buy jeans. - The shops were empty, you couldn't buy anything, right, you couldn't buy anything, the GUM department store in Moscow, which is the biggest department store right off Red Square. Sort of six or seven store department store, was empty. There was nothing in it. You could buy things, there was a black market, you could buy things to some extent on the black market, but there's a hugely inefficient command system that had basically ceased to function. And the Soviet Union was dependent on this system but the great difficulty with a command economy, and we'll talk about this more next week in relation to some of the other former communist systems, is that the people making the decisions have very bad access to reliable information. Prices are set by decree, not in response to supply and demand. There are very poor incentives to work. And so the unsustainable economy had become steadily more unsustainable. So that's one part of it, but then I want to come back to the gentleman who mentioned the arms race. The other reason is that containment worked. Now containment had been the policy dreamed up by George Kennan, who had been a US Ambassador to Russia and eventually became the first head of the Policy Planning Staff, it was an office that was created for him in the Truman administration. And he was one of the first people, this is in the 1940s, he had been one of the first people to say the Soviet system is unsustainable, it's got a dysfunctional economy at home, and its international ambitions are such that it's going to pursue, it's gonna pursue foreign adventures that eventually it's not gonna be able to finance. And so all we have to do is contain it. We don't need to go to war with them, we just have to contain them, prevent their expansion, and wait. And he had, there were various disagreements about what containment meant, we'll talk next week, for instance he was opposed to forming NATO even, which he thought would unnecessarily militarize the conflict with the Soviet Union and eventually he resigned from the government and went off in a huff to Princeton. But very largely, the US in Europe did follow the policy of containment. What changed in the 1980s was two things, which increased the fiscal pressure on the already pressured, highly pressured Soviet system. One was that they got involved in a quagmire war in Afghanistan. They invaded Afghanistan in 1979, having failed to learn from the examples of the British and others who had tried to invade Afghanistan in the past, a very difficult country to invade given the terrain. They spent the entire 1980s embroiled in this war in Afghanistan. It was their fear of the rise of Islamic fundamentalism, which they thought would infect the other Asiatic, would start to infect the Asiatic republics of the Soviet Union, and so they were very nervous about it. And it was a huge fiscal drain, a failed war. It was the Soviet Union's version of our Vietnam War. And then secondly, when Ronald Reagan came into office he announced Star Wars, which increased the cost of the arms race to the Soviets. By an enormous amount. So these were the reasons why the Soviet Union eventually just creaked and fell apart. We could spend the entire lecture unpacking them, but we're not, this is not a course about the Cold War. What I want to put our attention on is why was the collapse so peaceful. You know, you would think if somebody shoots an elephant it's gonna stumble around and do an awful lot of damage before it finally collapses. This elephant basically just knelt down and died. And that's quite remarkable. Any thoughts or intuitions about why it ended with a whimper rather than a bang? Yep. - [Student] There was a sense of desperation throughout the entire population, whether it was waiting in lines for any store, any product, or being overly bunched up in housing. The lack of jobs and the lack of innovation made the people very desperate and depressed. - People were desperate and afraid, but my real question is about the elites. Why did the elites give up power that they had guarded so jealously for such a long time. And after all there are many authoritarian systems in the world where there are angry, desperate populations and the elite clings to power and just becomes more and more repressive. Yeah. - [Student] I think they had prepared themselves for this having looked at what happened in Berlin, they had already started making sure that if this came that their interests would be protected. - Who would be protected? - [Student] That the elites-- - Okay, that's part of it. So that the elites, just to repeat the comment in case some didn't hear it. The elites were looking for and seem to have found ways to protect themselves through the collapse. So what I'm gonna do here is introduce an analytical schema from the social sciences, this is actually from a famous economist, a development economist by the name of Albert Hirschman who published this book "Exit, Voice, and Loyalty" in 1970. And he developed a framework for thinking about what happens when, as you can tell from the subtitle, firms, organizations, states, decline. How do people respond to decline. And the framework he developed I think is, I think this is, if you read no other book in the social sciences before you graduate, or this is certainly one of the two or three books you should definitely read. And we will use Hirschman's framework at various points in the course. And so what is, he said, well if you think about yourself if a declining organization, there are three things you can do. One is, you can leave, exit. Two is you can complain and try and get it changed, voice. Three, you can try to fix it yourself. You believe in the cause and you want to make it better. And a lot of Hirschman's analysis says that what mix of exit, voice, and loyalty people engage in is determined by the costs of those things. So for instance, if you think about a public company. If you're a shareholder in a public company and you don't like the way it's being run, you can just sell your shares and buy shares in a different company that you like better, right? So you don't have much incentive to go to shareholders meetings and trying to get the company, you maybe think it's got too much debt or something, you just buy shares in a company that has less debt. That's all you need to do, just exit, very low cost of exit. So people are not gonna engage in voice. Now of course you might care about the company, it might be ostensibly a green company and you might be holding the shares in the company because of your commitment to green energy. Then you might feel loyalty to the company and you will want to try and get it changed. Think about yourself as a worker, working for a company. If you don't like the way the company is being run unlike a shareholder you can't just pick up and leave, you've got, you know, you live in this town, you've got your family, your kids are in the schools and so forth. The costs of exit for you are very high. So you're more likely to engage in voice, you're gonna go to meetings, you're gonna complain, you might form a union. You're gonna do things, or join a union. You're gonna do things to get the company to change its ways. So his book is an analysis of the ways in which exit, voice, and loyalty might be complementary in some circumstances, and might be competitors in some circumstances for the ways in which people respond to the decline of organizations. So the first thing to say about the Soviet Union in the 1980s and '90s is there was a substantial decline, almost disappearance of loyalty among both the elites and the citizens. And I got a firsthand view of this in March of 1991. There was a Yale delegate invited by the Communist Party of the Soviet Union to a conference in Moscow, actually the week before the referendum that I mentioned to you earlier. And we had been impressed by Gorbachev and so forth, but we didn't really have much knowledge about, I certainly was no expert on the Soviet Union, and most of the other people in the delegation were not, and we had this stereotypical picture of the KGB shadowing us around and telling us what to do. And we needed to be careful and so on. And the first thing we noticed was widespread evidence of brazen corruption. I'll give you just two examples. We are all afraid of what the KGB people might be doing to us, the first KGB person I ran into was trying to sell me currency at black market exchange rates. And then secondly I went to a reception, and somebody came up to me with a translator and the translator said this gentleman wants to speak to you and he said, "Professor Shapiro, "I understand you're a South African." This is 1991, okay, it's before the collapse of apartheid, it's the beginning of the collapse of apartheid in South Africa, Mandela is being released from prison and so forth. But the apartheid regime was still there and it was not clear how it was gonna play out. And so this person speaks in Russian, which I don't, I speak a little bit now but I spoke none then, and basically the translator says, this gentleman wants to know how to make, if you can facilitate this making a connection with the government of South Africa because he has access to weapons which he would like to sell to them. And we regard apartheid as a domestic matter, we don't care what's going on in the country, but I can sell all kinds of weapons, including nuclear weapons. Which might have been nonsense, but this is quite, actually South Africa's one of the few countries that had actually given up its nuclear arsenal as the, that's another story which we'll get to later, but so I was stunned by this and I said, "I think maybe you mis-translated what he was saying to me, "could you ask him to say it again?" And he said it again, exactly the same thing. And so I said very nice to meet you and goodbye. So that was, this was, the levels of brazen, you know, it wasn't, nobody was trying to hide what they were doing, it was completely public, this is in March of 1991. We then had a meeting with Vadin Zagladin, a politburo member and he was at that time essentially the secretary of state for Europe. And he gave a speech, and he said we have concluded that the Soviet system is unsustainable, it's got to be replaced from top to bottom. And we were stunned by this. That a leading official in the Politburo of the Communist Party of the Soviet Union was essentially saying the system is bankrupt. He gave a speech and we went around the room and it was clear it was the kind of situation where you're only gonna get one question, so you better think about what the right question is. So when it came to my turn, I said, "Mr. Zagladin, "when did you conclude that the Soviet Union "was unsustainable?" And he didn't blink, he said, "1978". He said, "I was in a Young Communist League "actually with Gorbachev and there was a whole cohort of us "who concluded that the Soviet Union was unsustainable." So here you had a country led by a cadre of elites who didn't believe in the public ideology of the country. And so of course they had no loyalty to the public ideology of the country, it simply had dissipated. And then some of our hosts invited us to meet with various radical, more radical factions of the Communist Party of the Soviet Union who were more like activists and so on, and by then we'd been there for about a week and we thought, you know, we were getting some grip on this situation, so I met with this group of, they were members of the Communist Party, political activists and I had the temerity to say, well wouldn't it maybe a good idea to transform this economy in an ordered and gradual fashion. And they started screaming at me, calling me a Stalinist and a reactionary, and so it was just, these are just three pieces of evidence but everywhere you went, there was nobody who was defending the idea that, you know, whatever the sorts of, nobody had the sort of attitude that Gorbachev was expressing in that final video where he said, I am a loyal member of the Communist Party and we have to fix it, even though it had turned on him. So in terms of loyalty, Hirschman's category of loyalty, by 1991, loyalty was pretty thin on the ground. But let's think about exit. Exit in some ways turns out to be even more interesting. So as I said with the example of a company you don't like, you can just sell the stock. If the costs of exit are low and you don't have any loyalty why waste your time trying to change an organization? And so what was it that reduced the costs of exit? So there were some fortuitous political things. So if you think about I already mentioned, Boris Yeltsin had been a member of the Soviet hierarchy for much of his career, but he had the opportunity to go off and be in the Russian Federation and eventually become President of the Russian Federation. Eduard Shevardnadze who had been the Foreign Minister of the Soviet Union could go off and become President of Georgia, which he did. So there were places for these people to go, and these of course are the high visibility people, but many of the lower visibility people in their entourages and so on, could go with them. So there were opportunities for political exit for the leading figures and subordinate figures in many of the figures of the Soviet hierarchy. But what I want to spend more time is that the economic costs of leaving, of giving up, had reduced, had come down. And this leads us into the discussion of the rise of the oligarchs. The rise of gangster capitalism. So where did the oligarchs come from? They came from a variety of places. Just to mention some. One was the dynamics of the collapse itself. This is described in Karen Dewisha's book, the first chapter of Karen Dewisha's book, which I put on the syllabus. During the late 1980s as things were getting more and more extreme in terms of the fiscal pressure on the Soviet government, various ministries including the KGB started shifting a lot of money offshore and putting hard currency in bank accounts in Western banks. And some of the motivation for this was from loyalists actually. They thought Soviet Union's gonna collapse, we want to be able to live to fight another day, they sort of saw themselves as like, you know, the latter-day version of the French Resistance if you like, and they were gonna need resources. But others maybe have had different motives. But what happened in the August 1991 coup attempt was that the Soviet bureaucracy disappeared. And suddenly there were all kinds of people who found themselves with the control over these bank accounts, and they were now completely unsupervised. And so one relatively small, but not inconsequential, source of money from former Soviet officials was the money in these offshore bank accounts. And some people decided to, to make off with it. And that money resurfaced later, at later times. But by far the bigger source of oligarchs was essentially the theft of state assets. And so here for example, a gentleman by the name of Rem Vyakhirev, who had also been a Soviet apparatchik, he'd been the Deputy Minister for oil and gas in the last years of the Soviet Union. He was put in charge of Gazprom, which is a large state venture controlling gas in the Soviet Union. But what did he do with this power that he had? He essentially started giving away large amounts of its assets to friends and relatives through various schemes. So they would, they would essentially let friends and relatives buy gas at below market prices and then sell it and pocket difference. And they even did this with large numbers of gas fields. And so we started to see the emergence of people making huge amounts of money from these dirty transactions. And essentially the looting of state assets by people who were put in charge of them. So somebody like Vyakhirev was able to do this because he was strategically placed in order to be able to facilitate it. But a second, a third if you like, source of oligarchical power was derived from the fact that Russia had inherited a weak and bankrupt state. And I want to say a little bit about what I mean by a weak and bankrupt state. So much of the 1990s had been a period of fiscal crisis for the old Soviet Union. There had been massive debt racked up by the government. They were heavily dependent on oil revenues, which I'll talk more about in a little while, and when the price of oil gyrated, they would suddenly lose massive amounts of income. And this had all come to a head in 1998 when Russia eventually, when the Soviet Union, I'm sorry, when Russia by then, defaulted on its debt and had to be bailed out by the IMF. And so throughout the 1990s the Russian government was fiscally strapped. And so what happened was that people took advantage of this fact. Mikhail Khodorkovsky, who would later become famous for his ownership of Yukos, the private oil company, which I'll also have a little bit more to say about, he started a bank called Menatep Bank. And what he basically did was he, he speculated against the ruble, he would get the government to deposit large numbers of rubles in his bank, he would then, he would exchange them into hard currency and he would have a commitment from the government not to demand them back for four months, or six months, or eight months, and so essentially could speculate against the ruble, making large amounts of money. And he would pay it back later in less valuable funds. So there were people doing that kind of thing. And then there was the so-called loans for shares scheme, which is how he got control of Yukos and this was in the mid-1990s, once the sort of bloom was off the rose of the revolution and people were starting to get angry. Things weren't improving fast enough. Population was hungry. And it didn't look like, it didn't look good for Yeltsin's reelection. There was looming elections in 1996 and there had been something of a communist resurgence. They did extremely well in the 1995 legislative elections. And so the oligarchs who had started to emerge and who'd get control of these various assets realized that for them it would be a disaster if the communists came back into power. And so they decided to get massively behind Yeltsin, and they did two things. One was the so-called loans for shares, where they essentially lent the government money in return for shares in state companies, and Khodorkovsky did this with respect to oil, and of course later on the government couldn't repay the loans and so they would get to keep the shares. And so this was another way in which a lot of these state assets were looted. But then they also just poured money into Yeltsin's reelection campaign in 1996 and he was handily reelected against the communist candidate in the second round. So basically the oligarchs managed to emerge by taking advantage of the fact that the former Soviet Union was broke, and Russia in particular was broke for the vast majority of the 1990s. Then at the end of the 1990s, as I said, by then even the Communist Party had sort of gotten religion about capitalism. Yeltsin had reappointed many former communists to his government, but they got behind the reforms anyway, and so we saw that capitalism was not going away in Russia any time soon, even though the country was staggering from debt crisis to debt crisis pretty much for the bulk of that decade. In comes Putin, and now we start to see a very different politics and economics emerging. And so what I want to do now is focus on how Yeltsin, how Yeltsin's oligarchs if you like, were rather different from what was coming down the pike. So what I've said so far is that the Soviet Union at the end of its years and certainly Russia in its first decade as a sovereign country, had a very weak state. The party had been strong and the state had been weak in the Soviet Union, but the party had essentially disintegrated. But I haven't defined my terms. So I probably should have done this earlier, but what do we mean when we talk about a state being a weak state, what makes a state weak? Yeah, you're gonna have to yell. - [Student] Lack of a rule of law? - Lack of the rule of law. That would be one symptom certainly of a weak state. What else would be a symptom of, yeah. (muffled speaking) You gotta yell. - [Student] No control over violence within its borders or with it's borders themselves. - That would be another sign. Any other signs of a weak state, yeah. Somebody's point, yeah. - [Student] Finances. - So one sign, yeah, of a weak state is that, an inability to tax. A low capacity to raise revenue. And the Russia that Putin inherited had a very, very bad tax system, very low enforcement. So for example, in the oil industry, which is the most important industry in the Soviet Union, they taxed output and exports, not profits because they didn't have the auditing capacity to see whether people, you know, profits are revenue minus costs, and they didn't have the capacity to audit, to stop companies from hiding profits. And so they would tax in a very inefficient way because that creates very bad incentives as well to innovate. So they didn't have the capacity to raise revenue very well, and indeed throughout the 1990s the oligarchs, led by Khodorkovsky by the way, had waged war on efforts by the Russian government to improve the capacity to tax. They had constantly lobbied for, and been successful in getting, the parliament not to improve the capacity to tax, and when Yeltsin came into power, in January of 2000, his first speech was mostly about taxation. He said that he was determined to reform the finances of the Soviet Union and he was going to completely reform the tax system, which he did. They had a very complex, multi-tiered system with huge deductions. He brought in a 13% flat tax, kept taxes low across the board with no exceptions, and he greatly simplified corporate taxation as well. So one of his big missions was to put Russia's fiscal house in order. And indeed, he partly got lucky because the price of oil went up, but they used a lot of that money, they created a sovereign wealth fund for a rainy day, knowing about the volatility of oil prices. And indeed, when the financial crisis finally hit, later in the decade, Russia was pretty well placed to weather that storm. So that's one thing that Putin did in the first decade of this century to try, and people thought at the time that he was strengthening the Russian state. But the other side of a weak state is the haphazard use of power. And by this I'm referring to the idea that when a government's capacity to enforce things is low, it has to compensate by being unpredictable. So for example, in a famous article about 18th century England, there were no police in 18th century England, they didn't get police until the 19th century. 1834, I believe, or 1832, somewhere around there. But if you look over the course of the 18th century in England, the number of capital offenses went up massively. You could be hanged for being a pickpocket. But the number of executions remained constant. They didn't actually execute more people. So basically what would happen is that the assizes would come into town, they would string up a few people, and then they'd go off to the next parish. Because that was the only capacity that they had, so there's a kind of trade off. But one symptom of, one symptom of a weak state, is the haphazard and often dramatic use of power in this way. And anyone who's read Foucault's "Discipline and Punish" the opening pages of a gory execution of somebody on the scaffold is again an illustration not of a strong state, but of a weak state. Again, when you see the Egyptian government sentencing hundreds of people to death on one day, almost none of whom end up being executed, it's a sign of a weak state, not a strong state. So haphazard use of power. And you look at Putin's early moves, and they underscore this. The most important thing he did was he took down Khodorkovsky. Khodorkovsky by then was the most wealthy Russian oligarch. Owner of Yukos, and Putin threw him in prison. He was charged with various forms of fraud and eventually, when he'd already been in prison for about a year he was finally sentenced to nine years of imprisonment. And this got the attention of many of the other oligarchs. Nobody thought that Putin would have the temerity or the capacity to do this, but because of his KGB background, he had been a KGB officer in East Germany before he came back to the new Russia, he was able to do that. And then not only did he throw, not only did he throw Khodorkovsky in prison, he basically seized back through a complex transaction, which I'm not gonna describe, much of Yukos, and claimed it for Rosneft, the state oil company. And then he took down Vyakhirev, who, as I said to you earlier, he had essentially been giving away gas fields and profits and so on. And William Browder, about who I'm about to tell you a little bit more, had exposed this. William Browder was the head of the Hermitage hedge fund and he had done a lot of research on the way in which these gas fields were being given away and published them in the New York Times and elsewhere. And Wall Street Journal, Financial Times, and so on. And everybody thought this guy is crazy. He's gonna get himself killed. But Putin responded instead by firing his chief of Gazprom and again reclaiming many of the assets that had essentially been given away at bargain basement prices. So Browder's worth pausing to think about for a minute. He was actually the grandson of Earl Browder, who had run as the communist representative against FDR for the US presidency in 1940. But he didn't get along with his family traditions. His parents had also been very much on the political left, and he didn't get along with his parents, and he thought, how can I annoy them the most and the way I can annoy them the most is by becoming a capitalist. And so he had gone to Stanford business school and worked for Boston Consulting Group and got himself to be in on the ground floor for the first privatizations in 1989 and 1990, and then he had gone to Russia. And he became involved essentially in, he built up the biggest, at that time by far the biggest hedge fund in the Soviet Union, it soon became worth well over a billion dollars. And the way he did it was by taking on Russian corruption in firms. So the first company he bought, he bought an oil company and he had some shares in this company, and, the management of the company decided to respond to this by diluting his shares, essentially they were gonna offer a whole block of new shares, but not allow him to buy any. So effectually diluting the value of his shares. And he filed a shareholder's lawsuit, again people said you're crazy, you're gonna be killed. And he prevailed. And the share dilution scheme was stopped. And again, when he went after Gazprom and exposed what they were doing, he was convinced that he had figured out what he could do and that he was immune because he was a foreigner. He's a famous person today for reasons that will be obvious from this video clip. - Joining us first on CNBC to react, the man in question Bill Browder, Hermitage Capital Management. Bill, good to have you here, I bet you're not surprised to hear Vladimir Putin refer to you directly. - It's not the first time and it won't be the last time. Vladimir Putin is very mad at me. He's mad at me because of the Magnitsky Act. The Magnitsky Act is a piece of legislation named after my murdered lawyer Sergei Magnitsky. He was murdered after uncovering a $230 million Russian corruption scandal. Some of that $230 million went to Vladimir Putin himself. And so Putin feels personally exposed and potentially at risk of having his assets seized in the West, and so every chance he gets he tries to get foreign countries to come after me. This is just one of many of his attempts. - He leveled an accusation at you, that you funneled, and helped funnel lots of money to Hillary Clinton, did you? - I did not. I did not, I'm not a US citizen, I don't live in the United States, I've been living in Britain for 29 years. I make no campaign contributions. And so that's not true, I should also point out that Vladimir Putin and his regime have accused me of serial killing, of being a CIA, MI6 agent, and about a thousand other things, and so they just kind of, he's kind of unhinged in these accusations. - You said before, Bill, that this is not the first time that President Putin has mentioned you, has called you out. What is your reaction in terms of your personal security after such a news conference, do you increase it, do you look behind you when you're walking down the street a little bit more often? - Well my personal security has been at risk for a very long time. This is not, this struggle that I'm having with Putin has been going on for 10 years. I mean, just a month ago on May 30th in Madrid, I was arrested on a Russian Interpol arrest warrant by the Spanish police. Thankfully I was, Interpol contacted them two hours after I was arrested and made sure I was released, but the Russians have been coming after me for a long time and so I'm not surprised, it doesn't change anything. I'm always on high alert. And there's a lot of people around the world, governments, people, who are looking after me. - So what had happened with Browder was that he believed in these early years of the Putin regime, he believed that it was all due to his own acumen as an investor and his status as, at that time he was a US citizen, that he was able to go after these oligarchs. What he didn't understand, and that he learned later, was that he was pushing against an open door because he was going after Yeltsin's oligarchs which Putin also wanted to get rid of. But once he started going after Putin's oligarchs he ran into a brick wall. His accountant and lawyer, Magnitsky, was killed by the Russian police in a horrific scheme that's described at great length in his book "Red Notice" and he then went on this crusade to get the Magnitsky Act passed, which enables the US government to seize the assets of oligarchs who are involved in illegal activities, and which was used after, after the invasion of Ukraine in 2014. And that's what the famous 2016 Trump Tower meeting was about, that the Putin regime was trying to get Trump to promise that they would repeal the Magnitsky Act. And so that saga is ongoing. So Russia, one of the most corrupt countries in the world, you can see here it's up there with countries like Nigeria. This is one standard measure of corruption from Transparency International. You can see this is subjective perceptions of corruption in Russia that people think, the vast majority of people think the vast majority of the Russian state is highly corrupt. And one question is well why is it so corrupt? One obvious answer is a kind of path dependence story that we've talked about, that the fact that all these assets could be seized by former Soviet bureaucrats on the way out, or could be extracted from this very weak state, particularly in the 1990s, is a big part of the story. Secondly, there's this notion of a resource curse, which is often called the oil curse. And here, this is the political economist's story. Now the Russian economy at first glance doesn't look like it's so completely dominated by the oil sector. You can see here at least half of the Russian economy is services, but if you start to look a little bit more closely you see that while there are, you know, many different sectors in the Russian economy, one way or another the vast majority of them are connected to the hydrocarbon industry, three quarters of the Russian economy. So it's very oil dependent. You can see, and this is even more important, that the government's revenue is massively dependent on oil. The only countries that are more oil dependent for government revenue are places like Saudi Arabia, Iraq, Nigeria, Venezuela. The Economist magazine in 2016 estimated that 40% of the Russian government's revenue is from oil. And this of course makes for a lot of corruption because if the only way to get rich in Russia is to have access to the oil sector, then people who control that access can extract huge premiums for letting you participate. This is what makes Russian finances so dependent on the price of oil. And as you can, this is just another picture, which I will post these and you can peruse them at your leisure, but you can see that if oil prices fall, government revenues also fall with them. Which raises the question, why don't they diversify their economy? Why wouldn't they diversify the economy if it makes, they're so hostage to the price of oil. Quick suggestion, anybody? Why don't they? - [Student] They can't. - Pardon? - They can't. They don't have the ability to move like that. - Why don't they? - [Student] Because their structure is just so incomplete. - So the beneficiaries, I would use beneficiaries where you use structure, the beneficiaries like it this way, they're making huge amounts of money out of it, right? The government needs it because it's very difficult for them to raise money in any other way. As I've said, the capacity, the institutional capacity of the Russian state to raise money in other ways is very constrained. But also, from the point of view of the government it's a source of geopolitical power. If you think about where does all of this oil go. Think about Western Europe, you can see here that three of the biggest companies selling oil to Western Europe are Russian companies. And if you look at where European countries get their oil, just look at Germany, it's the top importer of Russian crude oil in the world. And this is why the Russian invasion of Ukraine played out the way it did because not only oil, you can see here, Germany is the biggest consumer of Russian natural gas in the world. So if you think about Russia going into NATO and the NATO alliance, which we're gonna be talking about on Tuesday, putting together and speaking with one voice about responding to the Russian invasion of Ukraine, it's not gonna be something that's high on Angela Merkel's agenda. She needs Russian oil and gas, she can't afford to antagonize the Russian regime. And so it was not surprising that the response to NATO was a lot of hand wringing and strong statements, but nothing more than that. So it's because it's a source of geopolitical power for a weak state that it's very desirable from the point of view of Putin to retain this powerful sector. Okay, we will see you next week. (chiming music) |
Power_and_Politics_in_Todays_World | Lecture_10_Money_in_Politics.txt | - So, today we're gonna talk about money in politics. - [Obama] Last week the Supreme Court reversed a century of law that I believe will open the floodgates for special interests, including foreign corporations, to spend without limit in our elections. (audience applauds) Including foreign corporations, to spend without limits. I don't think American elections should be bankrolled by America's most powerful interests. Or worse, by foreign entities. They should be decided by the American people. And I'd urge Democrats and Republicans to pass a bill that helps correct some of these problems. - So that was Barack Obama's 2010 "State of the Union" in late January of 2010, a couple of weeks after the Supreme Court had handed down its Citizens United decision. And the clip went viral because of that playback of Associate Justice Samuel Alito who lip readers were debating whether he was saying, "You lie," or, "Not true," in the middle of Obama's speech. It was an unusual speech because normally presidents do not criticize the Court during something like a "State of the Union" speech and nor do members of the Court respond to what the president says. And so it became a sort of emblematic of the tussles between the Obama administration and the much more conservative judiciary. Many people thought and think that Citizens United was a horrific decision that, as Obama said, opened the floodgates to corporate money in politics, making even more problematic the sort of things identified in the book about dark money which I've put on the syllabus for today. I'm gonna give you a somewhat different argument. I think that Citizens United is a symptom of a much different and a much bigger problem, and so that reversing Citizens United would really be sort of taking a bandaid to a much bigger problem. So our agenda is we're gonna start off by noticing that there's a difference between the supply of money in politics and the demand for money. Almost everything you read about money in politics is about the supply. It's about the lobbyists, it's about the fat cats, it's about the people who spend money in order to work their way in politics, and that's why they're supplying it. Almost nobody talks about the demand side. Why is there so much desire for money in American politics? Much more in American politics than in many other systems. And so in the last part of the lecture I'm gonna come to the demand side. For most of the lecture I'm gonna focus on the supply side, as almost everybody does, but we should remember that it's only one piece of the equation. I'm gonna begin by talking before I get into the guts of the money in politics subject I'm gonna talk more generally about courts in American politics and separate out some myths from realities. Many people think that they idea of an independent judiciary with powers to review the constitutionality of legislation that comes out of the political branches is a great thing. If you look at the third and fourth waves of democracy around the world, which I'm gonna be talking about on Thursday, the idea of a separation of powers is a big American export. It's often the feature of the US Constitution that is most often copied in constitutions around the world. So I'm gonna start out by spending some time on myths and realities about the role of courts in American politics and more generally. Then we'll go into the history of the First Amendment since the 1970s sixth decision, Buckley v. Valeo, where the modern world of money and speech really gets started. I'll pay some attention to the changing media context that all of this is going on in as we've evolved from the world I was describing in the last lecture, the world in which everybody watched the same evening news at 6:30 to the world of Twitter and blogging and people getting most of their news from places like Facebook. And then we will conclude by attending to the demand side, the much-neglected demand side, which will become an important feature of my argument later in the course. So let's talk about courts in American politics, as my subtitle here is "Myths and Realities." This is a picture of the Warren Court. And I put the Warren Court up there for a very particular reason, which is that the Warren Court was a massive outlier in American history. It was a massive outlier because it was an extremely progressive, to use the term that liberalism has been reinvented as in the last decade or so. Possibly the most progressive Court in American history. And it was remarkable, first and foremost, because Earl Warren, who was appointed by President Eisenhower in 1953, Eisenhower called it "the worst damned-fool mistake I ever made." He had been a conservative governor, a Republican governor of California. He had been responsible for the internment of Japanese prisoners during World War II. He didn't have any signs that he was gonna be a much different person than he turned out to be on the Court, and so he presided over this era known as the Warren Court. And as you can see, actually, interesting, if you go down to the bottom of the slide there you can see another one that Eisenhower pretty much got wrong. The Eisenhower appointees are in blue there. William Brennan also turned out to be a very liberal justice and didn't make Eisenhower very happy. But you can see that there were four holdovers in beige appointed by FDR, and so it turned out to be the case that for much of this period the Court was extremely liberal. And it's also the case, and maybe not so surprising when I make this point, that if you look at the academics who are champions of judicial review, the Laurence Tribe's, the Ronald Dworkin's, the Owen Fiss's of this world, they are people who came of age during the Warren Court. And many people thought that this was a sort of secular development that the Court, which had struck down, this Warren Court which had struck down a lot of racist, racist legislation and legislation that had been harmful to poor people, was sort of a new normal, a wave of the future, if you'd like. And after Warren was replaced by Warren Burger and we started to get more conservative drifts of the Court, you could often read, for instance, in "The New York Review of Books," you could read articles by people like Dworkin attacking what was happening to the Court as though they were undoing the new normal, and that this is a terrible thing and we just needed to get over this and things would snap back. So there was in what retrospect is quite a lot of denial about what was actually happening. But just to give a sense of how liberal the Warren Court was, this is the Court that was responsible for the decisions in Brown v. Board of Education, 1954 and then the follow-up decision the following year that said separate but equal cannot stand. This is the Court that in Baker v. Carr in 1962 and a subsequent case two years later said the Court can get involved in redistricting cases, that this is not a political question that they have to keep out of. And then they ruled that districts have to be more or less equal in size, population size, which was a way of limiting the disenfranchisement of urban voters. This is the Court which in Mapp v. Ohio said that unreasonable searches and seizures would produce evidence that wouldn't be able to be used in court. This is the Court that in 1963, Gideon v. Wainwright, said that indigent defendants would have a right to an attorney appointed by the government if they couldn't pay for one themselves. This is the Court that in New York Times v. Sullivan said that if a public figure wants to accuse you of defamation there's a very high standard. They have to find actual malice on your part. It's an extremely difficult thing to prove so it was seen as a huge boon for the freedom of the press to act as the fourth branch of government, if you'd like, in criticizing public officials. This is the Court that in 1965 adopted Griswold v. Connecticut, a decision that said that outlawing contraception in the State of Connecticut was illegal and would become very important later because it became the building blog for Roe v. Wade, that was actually at the beginning of the Burger Court. But the logic came from Griswold v. Connecticut, which was the idea there was a concept of privacy that the Court affirmed in Griswold that would become the basis of the decision in Roe v. Wade in the following decade. This is the Court that adopted Miranda v. Arizona. You've all heard about Miranda warnings. You have a right to remain silent. You have a right to have an attorney. And if the police violate those Miranda warnings, convictions will not stand. This is the Court that adopted Loving v. Virginia, which outlawed anti-miscegenation laws, laws against interracial marriage, many that had then prevailed in many states. So this is by no means an exhaustive list. Just to give you a flavor of the Court. It was really seen as the vanguard of progressive change by many liberals and led, as I said, to that generation of legal and constitutional theorists to think that courts were tremendously important engines of social change. But it was an outlier. Typically, presidents don't make the kind of damned-fool mistake that Eisenhower accused himself of having made with the appointment of Earl Warren. Typically, presidents get pretty much the courts that they want. And typically, judges are fairly conservative figures that don't wanna rock the boat, that tend to affirm the status quo. And so the Court that we've actually had since the Warren Court, first the Burger Court, then the Rehnquist Court, and then the Roberts Court, those courts are much more typical of how courts have behaved throughout American history. So if we go back and just, again, just to give you a flavor of how courts have tended to behave through most of American history, not the Warren era, Dred Scott v. Stanford was the case in which the Court held that blacks were not citizens within the meaning of the Constitution, they're not being contemplated to include them as citizens and so a fugitive slave did not have standing to go to court to get protection. This is the Slaughter-House Cases in 1873. This was in the aftermath. This is the years when the Reconstruction was beginning to fail and the Court did its bit to make sure that it would be reigned in. The Slaughter-House Cases limited the application of the Fourteenth Amendment's Privileges and Immunities Clause to the federal government and refused to apply it to the states. The case of United States v. Reese in 1876 said that the Fifteenth Amendment, which said that the franchised could no be denied on the basis of race, only denied it on the basis of race. It didn't deny it on the basis of what came to be used for proxies of race, like poll, taxes, or literacy requirements, or the other instruments that were there wheeled out across the American South to disenfranchise blacks. The same year, they adopted US v. Cruikshank, which said that the Bill of Rights does not apply to private actors or state governments. Again, limiting the Civil War amendments in their range and force. The Civil Rights Cases the following decade said that the Thirteenth and Fourteenth Amendments didn't apply to private individuals. Lochner v. New York in the early part of this century became emblematic of the Court that would strike down much new deal legislation in subsequent decades, eventually leading to FDR's attempt to pack the Court, which blew up in his face as they were striking down so much new deal legislation. But in fact the Court, even though FDR, which he had threatened to appoint five new justices and produced a lot of opposition, but after that the Court did back down somewhat. This is the Court that, this Court in 1944 adopted Korematsu v. the United States which allowed for the internship of Japanese-Americans during the war. 2013, Shelby County v. Holder. This was a decision that cut back on really important parts of the Voting Rights Act which had required states with a history of suppressing votes or making it difficult for African-Americans to get to vote to have pre-clearance from the Justice Department that there was no longer a problem. And as I think it was Justice Ginsburg said in her dissent, it was a bit like saying, "We no longer need, "we can throw away our umbrella "even though the reason we're dry "is that we're holding it up." So big, big reductions on the Voting Rights Act. And then of course the Court that the topics we're gonna be talking about in more depth today, the courts about money and speech, from Buckley to Citizens United and subsequently. So again, this is just a flavor as well, but the big point there is that these kinds of decisions are much more typical of how constitutional courts have behaved across the range of American history. And so we need to come to terms with that, that the Warren Court was a remarkable outlier and it should give us pause in thinking about how important it is to empower courts to behave in the ways that they typically have done. There's a lot of scholarship on this subject. I wanna just alert you to some of it. Perhaps the most famous article is one piece by Robert Dahl published in 1957. He was a longtime professor at Yale and this is a landmark paper in which he shows quantitatively that if you look at the data there's no evidence that Supreme Court decisions or judiciary review has actually contributed anything more to the protection of minorities' rights or vulnerable people's rights than not having the Court there. This is a book by a professor now at the University of Toronto, Ran Hirschl, it was a dissertation in this department, where he looks at countries that have gone from not having judiciary review to having it, countries like New Zealand, Israel, and elsewhere, and shows you can't see any significant difference in the protection of individual rights when they go from not having to having it. Mark Tushnet, a legal scholar now I believe at Harvard, has written an important book arguing about why giving more and more control over the Constitution is not a good idea from the point of view of protecting the sorts of rights we want protected in a democracy. His Australian counterpart, another legal scholar, Jeremy Waldron, his famous article here in the "Yale Law Journal" is part of an ongoing debate between him and Ronald Dworkin. Dworkin was advocating a Bill of Rights for Britain, and Waldron was saying don't export this aspect of the American system. And finally, another important book that started out as a PhD dissertation at Yale by Gerald Rosenberg, a professor at Chicago, about Brown v. Board of Education and subsequent decisions which says that if you look at the real progress that has actually been made in school disaggregation, where it's been made, has come from legislatures and not courts. So all of this is just a salutary reminder that we should think about the reality that is not only in American history but if you look comparatively the most important variable from the standpoint of protecting individual rights and vulnerable minorities is getting democracy, about which I'll have more to say in coming lectures, and adding judiciary review on top of that doesn't do very much, if anything at all. And indeed, sometimes it's inimical to the functioning of democracy and the decisions that are made by democratic legislatures. So, all of that is byway of throat clearing preamble to our subject for today. And this exhibit A, if you'd like. This is the main message of this lecture. It's exhibit A in observing that relying on courts as ways of protecting democratic institutions might not be such a great idea. So. Where we have to start, because we have a written Constitution and a constitutional system, is with the First Amendment, and particularly the Speech Clause of the First Amendment which says that "Congress shall make no law "abridging freedom of speech, or of the press; "or the right of people peaceable to assemble, "and to petition the government "for a redress of grievances." This is the text which we're all working from. And what I'm going to do now is take you through basically the 40-year history since Buckley v. Valeo of how the Court has interpreted the Speech Clause of the First Amendment as it pertains to money in politics. So the background to this is in 1974, when Congress had enacted very radical changes to the Federal Election Campaign of 1971. They did this over Gerald Ford's veto and it was the first comprehensive effort by the federal government to regulate campaign contributions and spending. Money was starting to become a really important issue in American politics and they felt the need to do something about it. I would just note and pick up this point later, this is the period at which American political parties were starting to become weaker. And I think there's an important connection between weak parties and the need on the part of politicians for money that I'm going to spend more time on later. So the key provisions of the 1974 legislation were to limit contributions to candidates for federal office, limit how much can be given to campaigns. It required the disclosure of political contributions, no secret contributions. It limited expenditures by candidates and associated committees, what we today would call PACs. And it limits independent expenditures to $1,000. And then finally it put limits on what candidates could spend from their own personal funds in order to get elected. So the goal was really to rein in the use of money. And you'll see two terms there that the distinction between which becomes important. The two terms: contributions and expenditures. And they make a distinction between contributions and expenditures which drives a lot of what they were doing. They also provided for public financing of presidential elections. Two years later, we got Buckley v. Valeo. And Buckley struck down most of that legislation as unconstitutional violation of the Speech Clause of the First Amendment. It held the following. It said it's all right for Congress to limit contributions to campaigns, and the argument there was because there's a risk of corruption. If you have people making contributions, they can get a quid pro quo, and that's not cool. But it's not okay to limit expenditures. Here the First Amendment rules. So the idea of an expenditure is if I wanna go out and spend money saying I think that Cristina should be president, I'm just spending money, I'm not giving her money. I'm expressing my First Amendment views. And that is not legitimate for Congress to limit expenditures. It's also not legitimate to stop Cristina from using her personal funds to get elected because there's no risk of corruption there either. We saw Donald Trump making this argument in the 2016 campaign, saying, "I don't need money from special interests." So there's no risk of a quid pro quo, so the argument went. And most important they said the idea of leveling the playing field in elections, it might be a desirable thing for governments to do but it's not that important that it should rise to the level of interfering with the First Amendment. Because the First Amendment protects what the Court in its interpretative dicta calls a fundamental right. And a fundamental right stands unless there's a compelling government interest, that they subject possible violations of fundamental rights to the highest or the most stringent levels of judicial scrutiny. And they said it'd be nice to level the playing field but there's no constitutional requirement of a leveled playing field in elections. There's a constitutional protection of free speech. And so the opinion said that the First Amendment "simply cannot tolerate" any interference with a candidate's freedom "to speak "without legislative limit on behalf of his own candidacy." This was the phrase about self-funded elections. And then more generally, this is a nice metaphor. This is actually in a footnote to the decision where the Court says, "Being free to engage "in unlimited political expression "subject to a ceiling on expenditures "is like being free to drive an automobile "as far and as often as one desires "on a single tank of gasoline." Who is the government to tell me how much gasoline I can put in my car? They might wanna level some playing field but this is about a constitutional right protected by the First Amendment. So, that was the decision. There's some important assumptions I want to just flag. One is that money is speech. In a dissent, Byron White said, "Money isn't speech, money is money." And he might have also said, "Money is a megaphone for speech." So the second is the one I've already mentioned, that the government interest in leveling the playing should say field is not compelling and so it must give way to fundamental rights protected by the Constitution. And the third is that the only corruption worth worrying about is quid pro quo corruption. Of course we're hearing a lot about quid pro quo corruption at the moment, and unless you can actually show the smoking gun of quid pro quo corruption, I give you something and get something in return, then you can't show, there's nothing to worry about from the standpoint of the corruption of the system. So that was what was launched in 1976 by the Court. Now, this is vastly more important than Citizens United in the sense that, once this decision was made, the changes that have led us down this 40-year path to Citizens United are more or less, nothing's inevitable in politics, but exceedingly likely to follow. I'll just mention a few other decisions that the Court made. So in 1990, when the drift to the right on the Court was not as far as it has subsequently gone, the Court did say that corporations didn't have the same First Amendment rights as individuals. So it was okay for the Michigan legislature to stop corporations from making independent expenditures to support candidates on the grounds that "corporate wealth can unduly influence elections." So they were saying corporations are not protected in the same way at least as individuals by the Speech Clause of the First Amendment. They don't have the same First Amendment rights. And then, in that case, the government interest in the conduct of elections and not in leveling the playing field, if you'd like, not having undo corporate interest trumps. The reason I'm putting the Michigan case there is because that was what was overruled in 2010 in Citizens United. Citizens United was founded by Floyd Brown, a DC political consultant, in 1988. And it was an organization, you can see here its big motive. It's "an organization dedicated "to restoring our government to citizens' control. "Through a combination of education, "advocacy, and grassroots organization, "Citizens United seeks to reassert "the traditional American values of limited government, "freedom of enterprise, strong families, "and national sovereignty and security. "Citizens United's goal is to restore "the Founding Fathers' vision of a free nation "guided by the honesty, common sense, "and good will of its citizens." So American as apple pie. So what was at issue in this case is that they have planned to advertise a film that was essentially an anti Hilary Clinton commercial shortly before the 2008 Democratic primaries when she was still running for the nomination. And at that time, federal law had prohibited corporations and unions from making what they called "electioneering communications" 30 days prior to an election, the primary, or 60 days prior to an election, or it's making any expenditures advocating the election or defeat of a candidate at any time. So there was some limit on the content and some limit on the timing, and the argument had been that under the then prevailing Michigan law that these were the kinds of constraints that had been legitimate. And in Citizens United, the Court said no. The Court held instead that Congress cannot limit corporate expenditures, and so overruled its prior decision in the Michigan case, and said the First Amendment rules in the same way with respect to corporate expenditures as it does to individuals. There's no risk of corruption here so there's no, again, quid pro quo is what counts. And the idea that the Court should be involved in protecting the shareholders of corporations cannot stand. By that time, it was a long-established, SOM students here will know, a long-established rule called the business judgment rule which said that while it's good, perhaps, for corporations to maximize shareholder returns, there's no requirement of that. And the business judgment rule says that basically if the shareholders don't like the way a corporation is managing the company, they can either sell their shares, low-exist cost, or they can get new management. So it was not the business of the Court to protect shareholders from the political uses of corporate money. So the assumption's now that the corporations have the same free speech rights as individuals, and so any government interested in leveling the playing field isn't compelling and so must fall. And again, reaffirming this idea that the only corruption worth worrying about is quid pro quo. Getting something in return for giving a politician some money. Now, there's a subsequent couple of decisions that are less often talked about than Citizens United that I just wanna mention. So the first one was the same year. SpeechNow v. the Federal Election Commission by the DC Circuit Court of Appeals. This is in the wake of Citizens United. Where they said contribution limits, whether to independent advocacy groups cannot be, to whether to individuals advocacy groups or other organizations cannot be limited. So the idea here is that you can make a contribution to another organization and that can't be limited. There's no risk of corruption in the First Amendment rule. They said, though, that Congress can require disclosure. So if I'm gonna give money to a political action committee or something like that, I might be, it's reasonable for Congress to require at least disclosure that people will know where the money came from. The public has an interest in knowing who's speaking about a candidate while making an argument in politics. Again, they're affirming this idea, now become a mantra, that the only corruption worth worrying about is quid pro quo. But the second assumption they made is that this idea of shining light through publicity would be enforceable, that it would be possible to say that corporations could give money to political action committees and, as long as they disclosed it, there was no risk, if you'd like, to the political system. It was certainly not a serious enough risk to justify interfering with the First Amendment rights to make political expenditures as they chose. As I say, good luck with that. - Here to help me make my move to secrecy and obfuscation completely transparent, please welcome former general counsel to the McCain campaign and my personal lawyer, Mr. Trevor Potter. Trevor, nice to see you. (audience cheers) Do you want some? Can I carve you some? Do you want some Ham Rove? - I'll wait 'til later. - Yeah, well, he's not kosher. (audience laughs) All right. Now, Trevor, I've got all these people down at the bottom of the screen who've been giving me money, individual Americans, but I haven't gotten any of the big corporate money, that's why I have a Super PAC. Why wouldn't a corporation give money? - Well, they'd been nervous about giving in a way that their name is publicly disclosed. People might object to what they've done. Their shareholders, their customers. - Okay, so that's where a c4 comes in. A corporation or an individual can give to a c4 and nobody gets to know that they did it, right? - That's right. - Okay, so, how do I get one? - And that money can be used for politics. - Oh great, that's good, too. - So we need to get you one. - As long as it goes through me, it can go to anything it wants. (audience laughs) So how do I gets me one, Trevor? (audience laughs) - Well. Lawyers often form Delaware corporations, which we call shell corporations, that just sit there until they're needed. - So just some anonymous shell corporation? - Right, and I happen to have one here in my briefcase. - Let's see it. Okay, what's it called? - It's called Anonymous Shell Corporation. (audience laughs) - That's got a real ring to it, Trev. - Registered in Delaware. - Now, I don't have to go to Delaware, do I? - [Trevor] No, it's already been done for you. - (whistles) Okay. (audience laughs) Okay, (mumbles) Anonymous Shell Corporation filed in Delaware, I've got this. So now I have a c4? - Now we need to turn it into your shell corporation, your anonymous one, and we do that by having normally a board of directors meeting. - And who's on the board of directors? - Well, just you. - Sounds like a nice group of people. (audience laughs) All right, let's do it. Let's call, (gavel thuds) okay. (Trevor and audience laugh) (gavel thuds) And I've shattered my champaign glass. (audience laughs) I hope there's no sensitive electronic equipment down there. All right, (gavel thuds) call to order, let's do this thing. - All right, so. This says that you are the sole director of the corporation. - [Stephen] I am. - And that you are now electing yourself president, secretary, and treasurer. - It sounds like a great board. (audience laughs) - And you are authorizing the corporation to file the papers with the IRS in May 2013. - So I could get money for my c4, use that for political purposes, and nobody knows anything about it 'til six months after the election? - That's right. And even then, they wouldn't know who your donors are. - That's my kind of campaign finance registration. (audience laughs) Okay, okay, so now I've signed it. I have a c4? - You have a c4, it's up and going. - So without this, I am transparent. With this, I am opaque. (audience laughs) - That's it. - Without it, you get to know. With it, you go to hell. (Trevor and audience laugh) Without it, here's who gave me my money. With it, you know what? Your mother gave me my money. (Trevor and audience laugh) Well, I like that, Trev. (audience cheers) (audience applauds) Okay. Okay, so now I can get corporate, individual donations of unlimited amount for my c4. What can I do with that money? - Well, that c4 could take out political ads and attack candidates or promote your favorite ones, as long as it's not the principle purpose for spending its money. - No, my principle purpose is an educational entity. - There you go. - I wanna educate public that gay people cause earthquakes. (audience laughs) - There are probably some c4s doing that. - Can I take this c4 money and then donate it to my Super PAC? - You can. (audience exclaims) (audience laughs) - But wait, wait. Super PACs are transparent. - Right-- - And the c4 is secret. So I can take secret donations off my c4 and give it to my supposedly transparent Super PAC. - And it'll say, "Given by your c4." - What is the difference between that and money laundering? (audience laughs) - It's hard to say. (audience laughs) - Well, Trevor, thank you so much for setting me up (audience cheers) (audience applauds) in this brave new world. (gavel thuds) Dismissed. Trevor Potter, everybody. - So, he called it right. What we see after SpeechNow is enter the Super PAC, and you have political action committees that can raise and spend unlimited amounts of money from corporations, unions, individuals, and associations, that is, from anyone. And so just people who are interested in the details of this, Super PACs can choose either to be what had been known as a 527 or a 501c4 organization. So the 527s, these had been created by the IRS to allow organizations and committees to raise and spend money specifically to influence politics, but they had to report donors and contributions. So they accepted donor lists would be made public in exchange for greater freedom to directly influence election initiatives, if you'd like. The 501c4, as you saw, that's what was being described in the video, works a little differently. It's a nonprofit. It's just a little different from a 501c3, which is what Yale University is. It was created in 1913 with the intention of creating tax exemptions for civic and commercial organizations that focus exclusively on social welfare. And it can also collect unlimited funds, and here's what they were keying off of, it's 501c4s do not have to disclose their donor lists. Basically, they forfeit their ability to engage in politics as a primary function but they gain the ability to keep the sources of their funding anonymous. Somewhere along the line, however, the interpretation of the 501c4 statute started to evolve. First of all, the word exclusively educational came to mean primarily educational. And, of course, once you have a word like primarily, you need lawyers to interpret what does primarily mean. And somewhere along the line, it seems, in the 1960s, the agencies started to interpret primarily to mean 51%. It's not clear why, I've done some digging into this. Perhaps there's a lobbying story there, I don't know. But so what it basically means is that you've got to spend 51% on your social welfare purpose, that gay people cause earthquakes, and you can then spend 49% directly on political activity. And this is what gave rise to the Lois Lerner scandal in the Obama administration, which some of you might remember. She was a mid-level operative in the IRS. They were getting floods and floods of these new 501c4s. This was in the wake of the Tea Party movement. And so she was tasked, or tasked herself, with investigating whether they were doing the 49% or not. And she then got accused of disproportionately investigating conservative 501c4s, and that blew up into something of a scandal that went viral. Well, that was all generated by this law. So, in effect, what it's done is it's said that you can engage in unlimited expenditures without disclosure, the slide is misleading, I will fix it before I post it, while optional disclosure, I guess is sufficient to say, without disclosure for 51%. Essentially paying a 51% quasi-tax. It's not really even a tax because you can define your social welfare purpose to suit yourself. And so if you wanna see what effect this had, this is Super PAC spending by disclosure. You can see that in the 2010 election cycle, before all of this became operative, most of the money was going to companies that, whoops, to 527s that were disclosing and not very much was going to others. But after 2012, you can see a lot more money started to go to the new 501c4 Super PACs, at least half of the spending in that election cycle. So we've seen this is a move into more dark money than we had had before operating in politics. The upshot is that contributions to candidates or campaigns can still be limited by the federal government, but independent expenditures can't be limited and contributions to organizations that engage in expenditures. So contributions to candidates can be limited but contributions to organizations that engage in expenditures cannot be limited, and reporting rules would require a massive overhaul to be meaningful. So that has been the story of campaign finance reform in the American courts through 2010. I should mention one additional decision, and this is 2014 McCutcheon v. the Federal Election Commission. This is the only subsequent time the Court has revisited campaign finance. And there, there had been a law which had limited the number of candidates that you could give money to. I think it was to five candidates in any given election cycle, and the amount you could give to a party was also limited. And so that meant the total amount you could give in any two-year election cycle came up to $117,200. You couldn't give more than that. And the Court said no, why should you be limited to only giving to five candidates? The justification for this combined restriction was unwarranted. It's not tailored to any particular purpose that the government is entitled to engage in. Even though, actually, Buckley had upheld aggregation limits, they said, again, there's no quid pro quo here, you're not giving money to get anything in return. What's the difference between giving money to five candidates and giving money to 30 or 40 candidates? Again, assuming quid pro quo is the only corruption worth worrying about. And a massive increase in inequality of contributing power now. Instead of 117,200, you could give 3.5 million in contributions in any election cycle. So, again, much more inequality of influence in the world of contributions. Now some people say this isn't so terrible because it may attract more money into contributions to candidates and less money into dark money. Rick Pildes at the NYU Law School, for example, has made this argument. So the idea is that if you can give more money to the candidates, you won't spend it on dark money, and so we might actually get less darkness. At this point, I think that's speculative. The jury is out. So that is the jot and tittle of the Court's involvement in campaign finance reform efforts. Before I turn to the demand side, I just wanted to say something about the changing media context of all of this because we're living in a very different world from the world that existed in 1976. Apart from what we make of this equation of money with speech that Byron White and many others have subsequently criticized. In 1976 there were basically three TV channels. And because of the scarcity of TV channels, the federal government came up with, this was not done through the courts or even by Congress, it was actually done in the executive branch, they came up with, the FEC came up with something, it wasn't the SEC but it was, I forget exactly which agency, but they came up with this, Federal Communications Commission, the FCC, came up with something called the fairness doctrine. And the notion there was that because there's a limited number of channels, that if a TV channel puts on one point of view, they must put on a competing point of view from the other side. So there was this notion of achieving some kinds of political balance. And this was never actually challenged in the courts so the Court never ruled on it, and we had the fairness doctrine until the 1980s during the Reagan administration. And during the Reagan administration, we also had the advent of cable television. And the FCC said, well, now there's an unlimited number of channels, so there's no reason to have the fairness doctrine. And then we got satellite, so instead of 80 or 90 channels we have hundreds of channels. So there's no reason to have the fairness doctrine because the logic for the fairness doctrine was the scarcity of channels. Now, there's no scarcity. In fact, more or less unlimited number of channels. And so the Reagan administration abandoned the fairness doctrine and that started us on the political consequences of increasingly segmented media markets. The world that we moved to being a world in which liberals get their news from MSNBC, conservatives from Fox News or Breitbart, and then of course onto the world of social media, the world of people getting news from Facebook and so on. And so we're now in a world in which it's almost impossible to have any requirement that different points of view confront one another in the sense that, say, John Stuart Mill talked about in the competition for ideas, the idea that if you think about prime minister's questions in the UK Parliament where the prime minister has to stand there and be questioned and defend his or her views to the opposition. We saw earlier Margaret Thatcher defending her views against her Labour critics when she was leaving office. We never have that now. Political speech is not about institutionalized argument over policies, it's rather about mobilizing your own supporters. It's about you speak to the people who already agree with you. We know from Kahneman and Sunstein's research that I referred to last time, that tends to make people more extreme. And so political speech is really about mobilization, mobilization of people within your world of like-minded people, and not about argument over policy. And that is a development that is underpinned by other changes in the political system that I'm going to get to briefly today but more fully later on. So let's look for the last few minutes at the demand side. We've been talking this entire lecture about the supply side. We've been talking about why people wanna spend money on politics and what kind of influence they think they can get from spending money on politics. Whether it corrupts the system or not and what sort of definition of corruption we should have. But why is it that politicians want money? Why do they need so much money? We spend much more money in politics in this country than in almost any other country. Why do politicians need money, yeah? - [Man] They attempt to put out their message. They attempt to buy advertisement from one side (mumbles). - Okay, so they want to put out their message. But British and German and Dutch and Australian politicians wanna put out their messages too and they don't have so much need for money in politics. So any, why? Who else, okay, yeah? - [Man] There are no primaries in the UK and the election cycle takes exactly six weeks. - So frequent elections and a staggered cycle, as well as primaries. So we have essentially permanent campaigning and campaigns are much longer. In most parliamentary systems, you call an election three or four weeks before the election. There's a very short election cycle, so that's one reason. And of course, if you have primaries, you need money for the primaries. So those are among the reasons people need so much money. Any other reasons? Yeah? - [Man] It's seen as a measure of support for the candidate. - Seen as a measure of support for a candidate. So we see how do you get to be on the stage in the current democratic debates? You've got to not only show that you've risen to a certain level in the polls, but you've also gotta have raised enough money from some specified number of independent contributors. So it's a way of demonstrating that you can get support. Again, it's not obvious why that differentiates American politics from other countries. Yeah? - [Woman] Well, for presidential elections, this is a huge country and people need to have teams all over the place. And unlike a parliamentary system where the local elections then govern who comes in power and so forth, here, you need to have a completely national campaign, which becomes very, very expensive. - Well, I mean, you certainly, in many countries you have to have national campaigns, and first and foremost, you have to win in your district, you need the money to win in your district here. But I think you're onto something nonetheless about thinking about the difference between national campaigning and local campaigning. There's a hand over here. - [Woman] I'm thinking about investments. So, like the stock market. Almost like maybe they'll take in the finances and then invest in bigger corporations and look for those corporations to put their name back out. It's almost like a partnership. - Well, I think focusing on people investing is also part of it, and political investing. Here's a way to think about it, that we have very weak parties in the United States, and I'll say more about what makes a party a weak party. Part of it is primaries. But a weak party is a party that is very decentralized in its control. So in a country with stronger political parties, such as Britain, the party has a big say in who the candidates are gonna be. They don't have primaries, as somebody pointed out, we have primaries. So another way of putting it is, in our system all campaigning has to be retail. And retail campaigning is much more expensive than wholesale campaigning because every person is running on their own for their own seat, and they need a lot of money to do that, and they can be threaded in their own individual seat. Some of you probably heard on the news this morning reports that there's some 35 Republicans in the Senate who privately would support convicting Donald Trump if the impeachment resolution goes through the House, but publicly will not. And the reason, and people come out and say it, they say, "If you did, you would immediately face "a primary challenge in your district." And so everybody has to, if you'd like, in retail campaigning, it's every person for themself. In wholesale campaigning, you're part of a team that is trying to win as a team. And so in our system, politicians need vast amounts of money because they have to get elected retail. All politics is local. It's a famous American aphorism, but that means all politics is retail. So there might be benefits of politics being local but the cost is that politics is very expensive. And of course once you have a lot of retail campaigning, it also turns raising money into something like an arms race, because if you don't raise as much as your opponent, you're not gonna be able, you're not gonna be able to have enough money perhaps to get on the debate stage or to get enough media time in order to prevail. And so this is why public funding limits are so inadequate, because what happens is, no matter where they're set, people soon discover that they can raise more money if they don't take the public funding. The idea is if you take the public funds, then you've gotta limit your fundraising. But the last time the capital was increased was before 2008 and President Obama, who was the first person to use crowdsource fundraising, realized very quickly that he could raise a lot more money than the limit. I forgot what it was then, it was several hundred million dollars, but his was the first million-dollar presidential campaign. So public funding limits don't work in a world in which people can raise as much private money as possible and in a world in which you're basically trying to communicate with your base voters, you're trying to communicate to the people who mostly agree with you. And so that's one of the reasons why we find it so difficult to limit the role of money in politics. Some other reasons why there's so much demand for money is that, I said to you earlier when I was talking about the resurgent right and the anti-tax movement that a lot of those bodies invested heavily in think tanks because they thought the universities were dominated by the left and it'd be much efficient to spend the money in places like Heritage, the American Enterprise Institute, the Cato Institute, and so on. Now, a lot of money, particularly in response to these recent Court decisions, is moving from the think tanks. The think tank gap between Republicans and Democratic think tanks has closed somewhat and the money it seems is going into these 501c4s, and that, again, means, from the point of view of the arms race of politics, that your opponents are gonna have more money spent on them and so you therefore need to get it spent on yourself. And so, again, we find that the remedies sometimes make the problem worse. Another point just about the arms race. I think that this is an area that's actually right for reform because politicians hate it. They hate, if you do any interviewing of candidates, you'll see this, politicians hate it because they have to spend all their time raising money and because they're then controlled by the people who give them the money, and they don't like it. And the structure of an arms race is you can't stop. Underneath it is something called the prisoner's dilemma. Even though it would be in the interest of both of you to stop raising money, if you could agree to do that, the logic of the situation keeps driving people to raise more and more and more money. And so actually politicians, the vast majority of them would like to get something to stop the arms race, which is why they passed the campaign bill in 1974 that we began with over Gerald Ford's veto. And the reason is that they were trying to put a constraint on the arms race so that they wouldn't have to raise this money. So but for the Court, this problem wouldn't exist. And I think it's, as I said, a salutary reminder of the cost of having independent judiciaries. One final point is that there's been a lot of talk about crowdsource funding. Maybe this is the antidote, this is the antidote to big money in politics. You know, lots of millions and millions of little contributions, isn't that great? It's like democracy is in the streets. People can give one or $2 and, as I said, actually Barack Obama was the first person to use crowdsource funding in a big way. Bernie Sanders is very effective at crowdsource funding. And so a lot of people have said, well, maybe there'll be a technical fix here that we will get to counterweight the role of big money. But what the research shows is that even people who give small amounts tend to not be representatives of their districts. They tend to be activists on the fringes of parties, they tend to be people funding groups like the Tea Party or groups on the far left of the Democratic Party. So the idea that you're gonna get less extreme views funded, particularly people who give any money outside of their own district tend to have very extreme views. Almost nobody gives money outside their own district who doesn't have, is not on extremes of their parties. So the idea that this is gonna be a way to sort of restore the influence of the median voter in politics is not something we should get too excited about. On Thursday we're gonna turn to part two of the course and we'll look at democracy's fourth wave, question mark, as it emerged in the wake of the Cold War. (gentle music) |
Power_and_Politics_in_Todays_World | Lecture_8_Privatizing_Government_I_Utilities_Eminent_Domain_and_Local_Government.txt | - Okay, let's get to work. - We want to turn now to the issue of eminent domain, which is being debated right here in New Hampshire. And Josh McElveen is the political director and the anchor of WMURTV. Josh? - Thank you, David. And good evening, candidates. Mr. Trump, you have said, quote, "I love eminent domain", which is the seizure of private property for the sake of the greater good, theoretically. You've tried to use the measure in business endeavors. You've said you support its use for the Keystone Pipeline Project. Here in New Hampshire, a project known as the Northern Pass would bring hydroelectric power from Canada into the northeastern grid. Do you see eminent domain as an appropriate tool to get that project done? - Well, let me just tell you about eminent domain because almost all of these people, actually, Chris hasn't. But so many people have hit me with commercials and other things about eminent domain. Eminent domain is an absolute necessity for a country. For our country. Without it, you wouldn't have roads, you wouldn't have hospitals, you wouldn't have anything. You wouldn't have schools, you wouldn't have bridges. You need eminent domain. And a lot of the big Conservatives that tell me how Conservative they are, think I'm more than they are, they tell me, oh, well they want the Keystone Pipeline. The Keystone Pipeline without eminent domain wouldn't go 10 feet, okay? You need eminent domain. And eminent domain is a good thing, not a bad thing. And what a lot of people don't know, because they were all saying, oh, you're gonna take their property, when somebody, when eminent domain is used on somebody's property, that person gets a fortune. They get at least fair market value, and if they're smart, they'll get two or three times the value of their property. But without eminent domain, you don't have roads, highways, schools, bridges, or anything. So, eminent domain, it's not that I love it. But eminent domain is absolutely, it's a necessity for a country. And certainly, it's a necessity for our country. - The difference between eminent domain for public purpose, as Donald said, roads and infrastructure, pipelines and all that, that's for public purpose. But what Donald Trump did was use eminent domain to try to take the property of an elderly woman on the strip in Atlantic City. That is not public purpose. That is downright wrong. And here's the problem with that. To turn this into a limousine parking lot for his casinos is not about public use. And in Florida, based on what we did, we made that impossible. It is part of our Constitution. That's the better approach. That is the Conservative approach. - So, eminent domain will be a big part of our discussion today. This class and Thursday, we're gonna be concerned with privatization at home and abroad. Today we're gonna focus on what is privatization, first of all. Then we're gonna revisit neoliberalism and the Washington Consensus, both of which feature privatization centrally in their recommendations. And that'll be some conceptual table setting for what's coming later. Then we will talk in some detail about the politics, law and economics of eminent domain, followed by privatization in the realm of utilities and local government, with some links that might not be immediately apparent to you yet, to last Thursday's lecture on Proposition 13 and the anti-tax movement. And then on Thursday, we will deal with the trend towards private sector prisons and privatization of the military, both of which have been, both of which have been led by the US, but they're going on in many other parts of the world. So, that's our agenda for the next couple of sessions. Privatization means a lot of different things. And I want to just start out by distinguishing some of them with a kind of conceptual map of privatization. So when we think back to the early lectures of this course, we talked about the post-communist privatizations. This was the era after 1989 of creating market systems de novo, which meant creating private property rights, the enforcement of private contracts, systems of adjudication, and so on. And as we, this is just reviewing what was involved in those debates about creating a private economy. The basic disagreement in privatizing things like oil, mining, heavy industries, retail, services, basic property rights, the basic disagreement was between the shock therapists, on the one hand, and the gradualists on the other. And we worked through those debates when we talked about China, Vietnam, and modernization theory and all of that. And particularly, the rather simplistic view that if you democratize a country before you had completed the transformation of the economy, the transformation would never occur. And so, it was important to really to do the economy first and do the democracy later. And we saw that the upshot of those debates, 20 years later, they washed out in Eastern Europe, so that countries like Czechoslovakia, which had been gradual in their economic transformation, certainly didn't seem to be performing any worse than countries like Poland that had gone for shock therapy. And indeed, by some measures, including poverty and inequality, you might say the Czech economy is more appealing. But then with a focus on China and Vietnam, the question might be also that if you wait to democratize, it might be waiting for Godot. It might never actually arise. So now we're gonna be focusing on different privatizations. This is what sometimes gets called neoliberal privatization. And that too includes a lot of quite disparate things. So, one is privatization as in the Thatcher and Reagan and as we saw even in the Blair and Clinton eras of previously public sector or nationalized industries. Things like transportation, British rail, which was nationalized in 1948, was denationalized in 1984. It was re-privatized. Many countries have privatized their mail systems in whole or in part. And then what we're gonna talk about today, privatization of utilities, privatization of many of the instrumentalities of local government. But there have been other kinds of privatizations. Much more central to the traditional functions of government. So, under the heading that I'm gonna call privatization of public authority, we've had things like much more use of private arbitration in solving government disputes. We're gonna talk about eminent domain. The use of eminent domain, as Donald Trump was being accused of in that video to facilitate the private transactions of individuals or individual corporations. Another form that I'm not gonna spend time on but just want to put on the list there is the increasing role of lobbyists in actually writing legislation. More and more legislation in Washington is not written by staffers. It's actually written by lobbyists. So, the private sector's getting more centrally involved in the wielding of public authority through the legislature. And then we might say, even further into the traditional conception of government than that, the privatization of core state functions. And here, we think of things like policing, prisons, and the military. And those are gonna be our topic on Thursday. So as I said, this is a word with many meanings. And that's fine. They bear a family resemblance to one another, but they are all different. And it's important to keep those differences distinct. So, let's revisit the larger political and ideological framework within which these debates have been going on. And I talked earlier about what was what I called neoliberalism at home and the Washington Consensus abroad. Pretty much the same basket of policies. But let me ask a question. When I say the word neoliberalism, how many people had never heard of it before you had entered this classroom? Okay, so most people have heard of it, okay. Those who had heard it, how many think it's a pejorative term? Pejorative term. Not pejorative term. Not pejorative, not sure. (laughing) Okay. So the not sure's have it. Okay, what about Washington Consensus? Pejorative term? Fewer think it's pejorative than the neoliberalism. Not pejorative? Not sure? (laughing) Okay. So that's helpful. (laughing) And I think what we saw in the room is reflective of general opinion. I think that more people tend to think of neoliberalism as a pejorative term than the Washington Consensus as a pejorative term. Though it's interesting, and it might be a reflection of our own sort of naval gazing myopia, that they're pretty much the same policies. It's just when you call it the Washington Consensus, you're saying other people should abide by these policies. When you say neoliberalism, you're saying we should be abiding by these policies. They come with a certain amount of freight. They are descriptive, but they are normatively loaded. And the three components that I've mentioned several times to you, one is the idea of getting rid of regulations, cutting back on government regulation of the economy, the second is privatization, that we're talking about today, and the third is free trade. And these were the main components of the neoliberal agendas of the Reagans and the Thatchers and the Blairs and the Clintons. And they were also the main components of what came to be called the Washington Consensus. And they were part of my message here, is that they were hegemonic until the financial crisis. After which, they began to fray. And we have mostly been in this class so far talking about the run up to the financial crisis, when neoliberalism at home and the Washington Consensus abroad did enjoy virtual hegemony. We'll see as the course goes on, that they fray, that the component pieces actually fray at different rates. That after the financial crisis, the first on the chopping block, so to speak, was deregulation and we got significantly new legislation to regulate the economy, the Dodd Frank bill. He has similar legislation in the UK and in many other countries. So there was a certain, there was a backlash against the deregulatory agenda of the early post Cold War period. Right after the financial crisis. Privatization has always been messier and more contentious as we will see today. But the backlash against privatization, and we're gonna see some of it setting in actually even before the financial crisis, in the rest of today's discussion. But free trade doesn't really get its comeuppance until 2016 when you get the resurgence of protection. Frankly, avowedly protectionist economics. There are many countries which say they're for free trade and surreptitiously have protections, but the idea that governments should really protect their own infant industries, at least, and perhaps more, gains real currency much later. So those are some of the upcoming topics. But let's start by observing what's been implicit in what I've just been saying. That even when an ideology is hegemonic, such as neoliberalism or the Washington Consensus, hegemony is never complete. The most subtle philosophical reflection on this subject is Michael Walzer's Tanner lectures, which he gave at Harvard in 1984, and appeared as this excellent little book, Interpretation and Social Criticism, published by Harvard University Press three years later, in which he makes the case that however hegemonic a system of ideas seems to be, there are always internal resources within it that enable criticism of it and the possibility of changing it into something different. His phrase is imminent criticism. And there was always gonna be room for imminent criticism of the Washington Consensus. So, another thing to say is that, and this is, I think, reflected in different reactions to the ideological valence of neoliberalism and the Washington Consensus, when I said we tend to think of the Washington Consensus as what other people should do. And of course, we don't know how those other people necessarily think of it. So let's get a bit of a grip on some of that. (horns beeping) - [Narrator] When Bolivia sought to refinance the public water service of its third largest city, the World Bank required that it be privatized. Which is how the Bechtel Corporation of San Francisco gained control over all of Cochabamba's water. Even that which fell from the sky. (speaking in foreign language) (yelling) (speaking in foreign language) (chanting) The press this beleaguered country paid for World Bank loans was the privatization of the state oil industry and its airline, railroad, electric, and phone companies. But the government failed to convince Bolivians that water is a commodity like any other. (speaking in foreign language) (yelling) (speaking in foreign language) Bolivia was determined to defend the corporation's right to charge families living on $2 a day as much as one quarter of their income for water. The greater the popular resistance to the water privatization scheme, the more violent became the standoff. (yelling) - So, this was privatization of the water in Cochabamba, Bolivia, in attempted, at the end of the last century, by this corporation, Bechtel, out of San Francisco. And this is just to give you a sense of how squarely in the mainstream of Washington Consensus privatizations this was. That water privatization had been occurring in much of the developing world since the early 1990s. And as was said in the video, it was one of the standard things that was urged upon, which urged upon developing country governments as conditions for getting aid in addition to other kinds of privatization. But you can see that even in countries like Britain and France, we've seen increased privatization of things like water supply during this period. And interestingly, there were very considerable efficiency gains from this privatization. In Cochabamba, there had been a lot of people who were without water, or had water one or two days a week. And the Bechtel Corporation put together a big consortium. They bid for this contract. It might have been wise for them to become nervous when they discovered nobody else was bidding for this contract. But they didn't, they went ahead and did it. And as you saw, what basically happened was as the water came on stream, that produced more demand for the water, and the part of the deal with the Bolivian government was that they were going to earn back their revenue from selling the water. And of course, the price of the water then went up because there was increased demand for the water. And this, Oscar Olivera, who you saw in the video, he didn't look so charismatic in that video, but in fact, he is an extremely charismatic and dynamic labor leader. He pretty much single-handedly led the fight against Bechtel. As you saw, the government started out trying to protect the company. And this led to massive confrontations with these farmers and peasants and others who thought they, as was said in the video, they thought, we have a right to this water. The notion that people can make money out of selling us our own water became an incredibly effective rally and cry. To the point where the government finally gave up and backed down and renationalized the water supply. And what happened with Bechtel was the Bolivian government just called them up one day and said, "We really don't think we can protect your people "and your property anymore. "So you better act accordingly." And they packed up and left. They then spent the next five or six years in litigation over 50 million dollars, which they were hoping to get back to compensate them for their sunk cost. Finally, they gave up. There's a wonderful New Yorker article, which I should've put on the syllabus, called Leasing The Rain, which I will, I'll post in case anybody is interested, that goes through this story. But so, here you have people thinking that even though this was an efficiency enhancing privatization, the costs of it were not worth having. And indeed, some people thought it was something obnoxious about the very fact of privatizing control of water. Reminds, you might say, it reminds one of John Locke's definition of domination in the first treaties on government, where he defines domination as being in a position to take advantage of another's necessity. Water is a necessity. And so, clearly, people like that felt this way. And interestingly, it's not just Bolivia. The week before last, the former Tory MP, Rory Stewart, kicked out of the party by Boris Johnson as one of the 21. Very interesting man, by the way. He once walked across Afghanistan. Wrote a book about it. But in any event, he was here giving a lecture. And he talked about the privatization of water supply in the UK. And he said exactly what I just said to you here. He said that actually, it's been efficiently enhancing. In this case, it's even, it's not more expensive than it was before, and the delivery is better, but there's still huge opposition to it among many British voters on the grounds, that they don't believe that water should be some, supplying people water to drink should be something out of which money can be made. So we'll come back to public opinion about privatization when we've considered a few more examples. But let's move on to talking about eminent domain. What is eminent domain? Somebody, what is eminent domain? As opposed to non-eminent domain. (student mumbling) Yep. (student mumbling) Exactly, so the right of the government to take private property for a public good. You said a public good. We're gonna have to spend a little time digging into what counts as a public good. In keeping with my idea to avoid theorizing, before we've got some empirical things out on the table, let's start in India now. And we will talk about privatizing eminent domain. This use of government power to actually do things that are a little more in the direction of what Donald Trump was being accused of by Jeb Bush in that opening debate. So in India, in 1984, they had something, they had a almost century old Land Acquisition Act, which took this term public purpose and broadened it to include promoting exports, investments, and public infrastructure. So this starts to broaden the idea. And then in 2007, they widened it further by saying any purpose useful to the general public will be counted as allowing the use of eminent domain. And then they set up a rule that worked the following way because the difficulty with India from the point of view of development is that there are millions of very tiny plots. Land ownership is highly, highly divided up into very small pieces of ground. And obviously, that means if you're gonna use the power of eminent domain, you're gonna have to deal with a lot of land owners. A lot of homeowners. Millions maybe, even. For some of the more ambitious ones. And part of what was going on in the early post Cold War period was that governments were trying to create special economic zones. And we're gonna talk more about them in other parts of the world later in the course, including in Africa. But the idea of a special economic zone would be an area in which, it was essentially to be a magnet for capital. So they would say, well, if you come and invest in our special economic zone, we'll give you regulatory breaks, we'll give you tax breaks, we will protect you from strikes by unions. They are often, therefore, criticized on the left. But just to give you a sense. Since the late '80s, the special economic zones have just taken off worldwide, as attempts to jumpstart economies by creating these, if you like, islands that are supposed to magnetically attract capital. Very popular with governments. They're sometimes called development zones in the US. Very popular often with governments. Not so much with economists. Most economists think that special economic zones don't really work, for a variety of reasons that needn't concern us. Perhaps the biggest exception is China, where the special economic zones are enormous. And they are essentially sort of integrated mini economies, and as I mentioned in my lecture about China, the original idea behind the Chinese special economic zones was actually islands of experimentation with capitalism. They wanted to try out capitalist economies without transforming the whole Chinese economy. This was part of their experimentalism. And of course, in China, there was not much of a problem about moving people around, if they were disinclined to move. So if you look at the difference between, in the red box here on this slide, the Chinese and the Indian special economic zones, you can see that the Chinese ones are enormous, 300 square kilometers, whereas the typical Indian one was 1.4 square kilometers. So, India has this basic problem that if they want to create special economic zones and they want to use eminent domain to do that, and it's a democracy where there's strong tradition of citizen opposition to things that they don't like, how are you gonna do it? So they came up with this 70% rule. And the 70% rule basically said the following. It said if a private developer can buy up 70% of the land for a perspective development, whether it's for a single company, plant, or for a special economic zone, then the local government will come in and compulsorily purchase the other 30%. And then sell that 30%, that 30% would be sold to the company. And so, you can see what the logic would be here. That in order to buy the first 70%, the company would have had to have done a lot of work on the ground with the local people and gotten enough very significant showing of support for what they were going to be doing. But then there's the obvious problem of holdouts that the more you become the veto player or the possible veto player on the decision of this project to move forward, the more you can ransom the whole project by demanding a very high price for a market transaction. And so, the thought was, didn't seem crazy at the time, the thought was the 70% rule will do it. It sort of splits the difference. It forces the companies to do a lot of work with the citizenry, but it also solves the problem of holdouts. So, this is what came to a head in the province of Singur in West Bengal early on in this century. And so, the Tata Corporation wanted to build something called the Nano. Anyone know what the Tata Nano was? Yeah. (student mumbling) The world's cheapest car. And this was their marketing ploy. They wanted to build the world's cheapest car. It was gonna cost about $2500. And they thought this would be, I guess it was sort of the VW Beetle for the 21st century. Though when we see what happened to the VW Beetle in the run up to World War II, that might have been a cause for worry, at least for Tata. But in any event, they decided they were going to build this world's cheapest car for the Indian mass market. And then the next question was where to build it. And so, they decided they would build it in West Bengal. West Bengal was interesting for a number of reasons. And they'd had bids from other parts of the country as well. But the area of West Bengal in which they were planning to build it, very consistent with the themes of this course, was governed by a communist party that had got the zeal of a convert about capitalism. And so, even though it was governed by a communist party, they wanted to get this plant. They wanted to get Tata to come and build this plant. And so, they worked with the company and the company began buying up land. And when they passed the 70% threshold, the single government of West Bengal compulsory purchased the rest, and the project began. So, just to give you a sense of how it played out. - [Narrator] In the countryside outside Calcutta, the electoral battle has turned decidedly nasty. Highlighting some of the problems of India's industrial expansion. Five years ago, the communist government of West Bengal moved thousands of farmers off their land to make way for a giant car factory, owned by the Indian conglomerate Tata. Some of the farmers, like Mahadad Dass, weren't even compensated. (speaking in foreign language) Sensing a political coup, the opposition party backed the farmers. - So there, you see democracy at work because it turned out that this plant was in a very marginal constituency. It had been won by literally a handful of votes by the communist party in the previous election. And the opposition saw this as an opportunity to unhorse the government. And so, they got behind, and just as with in the Bechtel case, there was a very charismatic opposition politician who got behind this cause, and the farmers began not only complaining, but actually attacking the plant. (dramatic music) (yelling in foreign language) - [Narrator] It's a word from the Indian Tata Motors that caused this violence. While Ravi Kant was in the plant assessing its progress, outside the plant, 200 villagers strike and storm in. Police and the protesters engaged in flashes for over an hour, as tear gas shells were fired, which injured two policemen besides some farmers. Even as the situation was brought under control, the protestors made it difficult for them to leave the area, setting up blockades along the highway. (yelling in foreign language) - According to the villagers, this kind of sporadic violence will continue in and around these borders. Many farmers are adjudicated because they didn't get jobs as they were promised. - [Narrator] The tension in Singur is unlikely to end soon. (speaks in foreign language) Times Now. (yelling in foreign language) - And so, it played out very similarly to the efforts of Bolivia. It turned out that even though the government became more and more repressive, there came a point where the cost was too great. There was something of a standoff. And finally, Tata said, "We've had enough". And they announced that they were leaving West Bengal, and as you can see, it's not one of the poorest parts of India, but they moved to right across the country to Gujarat, which is significantly more developed. And they built the Tata Nano there. The whole issue that became the flashpoint was this idea of just compensation. And you might wonder about how it was conceived of and why it became such a flashpoint. Because of 70% of the farmers had actually already sold their property, and it was only the remaining 30 who were being forced to sell, you would think that the politics of this could have been managed. But here are some inconvenient facts that got in the way of this transaction being successful. And one is that, India's a relatively weak state, and one of the ways in which it's weak is that people avoid taxes by when they transfer land, they have an official price and an unofficial price. And the unofficial price is higher than the official price because the official price is used to calculate the taxes. But of course when, when the government compensated people for the land they were taking, they would use the official price. And so, people felt like they were not getting the value of their land. To make matters worse, once Tata announced the project, that the project was going forward, the price of land immediately started to skyrocket. And so, the people who had sold their land earlier felt like they were now getting screwed because they had sold it for a much lower price than was available on the open market once the project had been announced. And to exacerbate that problem, some people who figured out what was coming down, sort of became middle man and ran around buying up lots of land, so that they could bid up the price and sell it later. This really meant that the moral, the 70% rule created a kind of moral hazard because it meant that people could hold out for a higher price, even though it was intended to defeat the holdout problem. And then finally, and perhaps as you saw from the interview with that farmer, many of these farmers who were being thrown off the land were not gonna get jobs working in the Tata plant. That was semiskilled work, these were farmers. And so, what were they actually going to do? Even if they had been compensated for their land. Rajat Tata, that's him there, CEO of Tata, actually a very well respected and well liked company in India. So this was quite an unusual catastrophe for them. Even though he left 350 million dollars on the table in West Bengal, he decided he'd cut his losses. And here, you can see the plant opening in Gujarat. A footnote to this story is that this is the first Nano being produced in Gujarat. But it turned out that it was not a successful car, by the way. The reason being that people don't want to be driving the world's cheapest car. (laughing) They think, well, this is a status economy as well as an economic economy. So, they've done many back flips to try and rescue the Nano. But it hasn't been an economic success. Unfortunately though, this is what's left of the defunct plant in West Bengal. The land continued to belong to the Tata Corporation, and most of it has not gone back to farming. So here again you see a case where the losers from a privatization effort, using the concept of eminent domain as a vehicle for promoting economic development, backfires because the people who are gonna be harmed by it or who believe they're gonna be harmed by it get mobilized, and incidentally, the opposition party did knock off the communist government in the next election. So, it worked to that extent. So these are some cases, the Bolivia and the Indian case, to remind us that when we talk about neoliberalism and the Washington Consensus, we should always remember that it's important to think about how these things look from other parts of the world, and not just as they articulated by policy wonks in Washington or in the World Bank or among Western academics. So, let's talk about the US. Privatizing eminent domain in the US. Now we have something called the Takings Clause in the Fifth Amendment, and the Due Process Clause of the Fifth Amendment, for the lawyers here will know, the Due Process Clause of the 14th Amendment applies the Fifth Amendment to the States, and it says that people shall, the last clause of it, "nor shall private property be taken for public use "without just compensation". So we've seen that in India, they had this very capacious definition of public use to enable economic development in West Bengal that produced a certain amount of backlash. But how should we think about public use? What is a public use? Many people would say, well, Donald Trump's casinos aren't public use either. So what is a public use? How do we know a public use when we trip over it? - [Student] Benefits the public. - Pardon? - [Student] It benefits the public. - Benefits the public, okay. Well. That could be part of it, but presumably, economic growth benefits the public. And this was the argument for what was being done in West Bengal. What more can we say about public-- - [Student] Clear public entity. Clearly a public entity. - A public entity. Like? - [Student] School. A school. - A school. Well, there are a lot of private schools but yeah. Okay. So one place one might start, we shouldn't always start with the economists, but the economists have a definition of a public good. Can anyone tell me what would an economist tell us is a public good? Yeah. - [Student] Non-competitive, non-excludable. - Okay, so you said non-excludable. Was your second criteria. And what does that mean? - [Student] One person having it doesn't exclude another person from having it. - If we create clean air for you, we can't exclude the person next to you from having the clean air as well, right? So that's one feature of the definition of a public good. If you provide it, everybody's gonna get the benefit of it. So that obviously creates incentive to free ride because if I'm gonna get the benefit anyway, why would I pay for it? If somebody else is gonna pay to provide it. So that's one element of a public good. And then you said, the other element? - [Student] Non-competitive. - Right, sometimes non-rivalrous. And what does that mean? What is non-competitive or non-rivalrous mean, somebody? Some economist. Have I frightened every economist out of here already? (student mumbling) Yeah, it's basically the ideas that my having it doesn't stop you enjoying it. So, sunshine. My consuming sunshine doesn't stop you consuming sunshine. We don't need, we're not rivals for the sunshine. So, that's the idea of a public good. But notice that even if you start with such this narrow technical definition of a public good, of things that are non-rivalrous and non-excludable, they're still gonna be politically charged. Why are they gonna be politically charged? One is because there are alternatives. So if we think about building a road. If you build a road, anybody can use the road. And if you build the road, my using it doesn't, it doesn't use up the road. You can also use the road. We can't exclude you and it's non-competitive. But of course, rather than building the road here, we could've built the road someplace else. So, there's still, there are always gonna be alternative courses of action that might have been taken. There is now a plan to build a high speed rail, which I'm sure will never be built, (laughing) from Washington to Boston, and the current plan goes to Hartford, bypassing New Haven. 60 billion dollar enterprise. And of course, the negative externalities of that for Yale would be huge. It would be much better if the high speed rail went through New Haven. So even when you're providing what is technically understood as a public good, there's always gonna be more than one way to provide it. And there'll be winners and losers, depending on which way is chosen. Then secondly, there are gonna be externalities. What economists call externalities. There are gonna be costs, wherever you put it, that some people will have to bear. So even if it had been the case that the Tata plant in West Bengal would have jumpstarted the economy, brought lots of employment, would have brought other industries to the region, would've been a net benefit by just about any definition, these farmers would still have paid a substantial price because they were getting kicked off their land, and they were not likely to be employable in the new industries. So even if our net, we say that providing a public good is desirable, it's still gonna have externalities. And you have to worry about the externalities. And then finally, you're gonna have the problems of valuation that we talked about before. When you do eminent domain, it's compulsory purchase, and so, you have to decide, what is the right value? And of course, there are various mechanisms for doing that. Looking at the value of houses around the house that you're gonna compulsory purchase. But there may be problems with the market, or at least, the people living in the house may perceive problems with the market, or it may be that I was born in this house, my grandmother was born in this house, I value this house differently than I could sell it for on the market. And so, you're gonna get those sorts of problems, and the sorts of, if you try and get around the holdout problem with something like India's 70% rule, you're not going to persuade the holdouts. In fact, what you've just, what you basically do is you shift the holdout problem near the 70% threshold rather than near the 100% threshold. So there's gonna be, even if you start with a very narrow definition of public purpose, it's always gonna be politically charged. And in reality, there's almost no government in the world that sticks to such a narrow definition of a public purpose as a technically what counts as a public good. So, this all came to a head in Connecticut, here in Connecticut 14 years ago, in a case called Kelo versus the City of New London. And it was a very interesting and controversial case because it was exactly the use of eminent domain to facilitate a purely private transaction. Basically, the City of Kelo, which was not a blighted area, there had been some previous Supreme Court decisions which had said if an area is truly blighted, then compulsory purchasing can be done, if there's some plan to revitalize it. This was not an area of blight, and by the way, anyone who knows about the history of debates about urban renewal will know that even compulsory purchase of blighted areas, go take a look at their remains of the Oak Street Connector in downtown New Haven. Or read Douglas Rae's brilliant book called City: Urbanism and Its End if you want to know about urban blight as not being eradicated successfully through eminent domain purchases. In any event, this went way beyond blight, and the City of New London had an offer to build a shopping mall. And they used the eminent domain power to force that transaction through. And they compulsory purchased some homes, forced people out of them, and compensated them, and justified, of course, they were sued. It worked its way up through the courts in Connecticut. But the defense of it was that it would bring economic growth, employment, and higher tax revenues for the City of New London. And what was remarkable about this was it went all the way to the Supreme Court and the City of New London won. They won in a five to four decision, and the decision, interestingly, was written by what we think of as the less Conservative wing of the court in those days. So, it was Justice Stevens with Kennedy, Souter, Ginsburg and Breyer. They all said, well, if careful studies have been done and it's clearly going to bring economic growth and employment and higher tax revenues to the City of New London, it can go forward. And so, it actually survived. And it was the Conservatives who we tend to think like capitalism and like economic development. Sandra Day O'Connor, then Chief Justice Rehnquist, Scalia, and Thomas dissented with a strong defense of property rights. As O'Connor put it in her dissent, she said, "To reason, as the Court does, "that the incidental public benefits resulting from "the subsequent ordinary use of private property "render economic development takings for public use "is to wash out any distinction between "private and public use of property, "and thereby effectively to delete the word "for public use from the Takings Clause "of the Fifth Amendment." And she basically is saying once you can do it for this, you can do it for anything. And so, the Takings Clause really doesn't offer much in the way of real protection. So, that was a pretty dramatic occurrence. It's more dramatic than the 70% rule in India. It's actually saying if a local city government can convince itself and perhaps some fact finding body that this is gonna be good for the city, they can compulsory purchase land. And though we didn't have riots the way that we saw riots in Bolivia and we saw riots in West Bengal, there was huge backlash to this decision. And interestingly, it was a coalition of strange bedfellows that were opposed to this. It was groups like the AARP, the NAACP, the Libertarian Party, and Conservative think thanks all came out against it. And president then, George W. Bush, announced an executive order that federal funds could not be used to facilitate private use of eminent domain in this way. And before the Kelo decisions, only eight states had had limitations of this sort on the use of eminent domain power to facilitate private transactions. As of April of this year, 45 states had enacted curbs upon it. So, as you can see, the forward use of the eminent domain power actually triggered backlash. And as with the Tata Nano plant in Singur, West Bengal, it turns out the development never occurred. And you can, there's now a movie about this whole episode, which you can watch. So the upshot of this discussion so far is that neither efficiency, as in the case of water privatizations, nor just compensation, even if it's effectuated in terms that an economist would approve of, might be sufficient to grease the wheels of privatization. In fact, you might run into problems. Loss aversion might be more important to people than just compensation. They may not want to give up their home, or they may not want to retrain to learn how to work in a car factory. Unanticipated externalities can trigger very effective opposition, particularly when there's a political entrepreneur to mobilize that opposition. And the most important takeaway is that there's no technical answer to this question. It's not having the right theoretical model. These are just political questions where there are distributive benefits and burdens that are gonna be born, no matter what's done or what's not done. And so, it's really, at the end of the day, political. And in that sense, you can't ring the politics out of politics by trying to have a fancy model to tell you the answer. The answer's gonna depend upon figuring out who's going to benefit and who's going to be harmed, and then making sure that the people who are going to be harmed are not mobilized to stop the project, possibly by compensating them or by finding out what else you can do to help them. Let's talk about privatizing local government services. There's a link to the previous class in this that we talked about how the agenda of the anti-tax movement was to starve the beast, to reduce the amount of money available to government on the theory that governments either waste it or further their own nest with it. And if you look, if you go back to, I started with Proposition 13 last time, you can see, indeed, after Proposition 13 was enacted, local government revenues in California just plummeted. And cities and counties where the same kinds of propositions have been adopted have found themself increasingly starved for revenue. So you can see in the second slide, it's the right hand slide there, to extent, they can raise revenue. They've had to depend on things other than real estate taxes because of the bars on raising money in the ways that they traditionally had done. And if you look at California, the growth, this would make the Proposition 13 crowd happy. The California has, local government revenues have indeed grown more slowly than elsewhere and contributed to many of its fiscal woes. So, one of the ways in which local governments have responded to this reality is by moving towards privatizing local government. And what do I mean by that? It's basically they have moved towards these things that Evan McKenzie writes about called common interest developments. What is a common interest development? Yeah. (student mumbling) Right, they're sort of giant condo associations. Where you own your unit that you live in. But you pay a fee to the association, and the association then delivers a lot of services. They arrange for a snow removal if it's in New England, not in California. Most of California, not Southern California. But they arrange for garbage pickup, sewage, some of them local policing, you can get your hot water through them. So you get a lot of services that would traditionally have been provided by local governments from the common interest development. And this is obviously attractive to cash strapped local governments because they find the provision of services expensive. And now it's much harder, as we've seen, for them to raise revenue in the ways that they typically have done. And so, in a lot of these kinds of states, you find cities that actually, the only form of new housing building they permit is common interest developments. And so, to give you some sense of the numbers here, they're really quite dramatic. Go back to 1964. There were fewer than 500 of them in the country. And you can see, after the anti-tax movement gets going, they just start to skyrocket. And this is now 10 years ago. I couldn't find aggregate statistics more recently than that. But by 2009, we had 24 and a half million housing units, 60.1 million people, and almost a fifth of the American population is now living in these common interest developments. And I'm sure that number is significantly higher today than it was a decade ago. That's more than the population of many, many countries. 60 million people are living in these developments. And as I said, a number of cities, particularly in California, in Arizona, in Florida, will not allow any other kinds of development. So it's a big change. It's a big change over the last several years, in large part, driven by the cost savings for local governments. They get to double dip, if you like, because they get to, you still have to pay property taxes, but they don't have to provide you with anything like the number of services that they have to provide to other homeowners. So what should we think about that? What are the consequences of that for democratic politics? (student mumbling) Take the mic for him, yeah. (student mumbling) So they're private, essentially, private governments in these developments. They're usually boards. Sometimes, they're elected boards. But typically, the way it works is the developer creates a board when they're building it. And that board will decide what type of housing they're gonna build, is it gonna be a development that permits animals, is it gonna be a retirement development, no children below a certain age. What the basic structure and rules are gonna be is gonna be set by this board, and then people will have to come in on that basis. And he'll have essentially a charter, constitution, which will likely be very difficult to change because the initial homeowners will not want the rules to be changed on them in midstream. And so, you're gonna essentially have private government of a lot of local activities. Now some might say, well, so what's wrong with that? People can pick and choose. I mean, you go to Fort Lauderdale. And you can see, by the water, these developments go for millions of dollars. Then you go quarter of a mile back and they're cheaper, quarter of a mile back, they're cheaper, and people could sort into just the kinds of housing that they want. The market is hunky-dory here. Everybody can, if you don't like animals, go live in one that doesn't like animals. If you, doesn't allow animals. If you want access to a swimming pool, you go to one that has swimming pool and you pay a little bit more. So, you could say, one might say, there's no problem with this. Yep. - [Student] One of the bigger issues is the people who are elected to serve on these elected boards might not have the expertise needed to properly run the different utilities or any of the different elements of the government. Of the internal government. - So there can be accountability problems because the boards tend to be run by amateurs, as you say, and they might be quite incompetent. They turn out not to be able to manage the accounts and so on. If we had a little bit more time, we could work our way through a number of those kinds of issues that pop up. And which have caused a demand for regulating these boards and so on. It still could be a problem that your neighbor stopped paying their condo fees and the thing goes bankrupt and it's very difficult for you to have redress in these situations. So there are a lot of those issues. Any other issues we should be thinking about? Yeah, at the back. You have to yell or wait for the microphone. (student mumbling) The microphone's gotten to you now. (laughing) - [Student] I meant that no judges can overturn the decisions of the board, the courts, then not to be involved with these CIDs. - So, you have to explain what you mean. - [Student] So that in the readings, there was a lot of people who wanted to sue the boards and the judges that they didn't want to become involved with these associations. - Yeah, so there's potential for incompetence and corruption and related things. Those things are all true. However, there's the same potential for all of those things in local government. So I think that the things to think about that are gonna be significant going forward is one, is the boards can be quite undemocratic because they usually, as I said, the dice is loaded early on. But also, we talked earlier about Hirschman's Exit, Voice, and Loyalty. There are problems here about entry because if all the housing in a part of the country is built in these developments that pick the markets that they want to serve, what about homeless people? Where are homeless people gonna wind up? They're gonna wind up on the streets of San Francisco or somewhere like that. That because if you want to buy into one of these places, of course you've gotta, they don't want you unless they're sure that you can pay, so you're gonna go through financial screening, you're gonna have to prove you can afford to live in the place. And people who can't are going to be wind up not being served. There's gonna be a big mark. If you try and do housing through this kind of market, there's gonna be a market failure that is gonna be probably quite costly for governments to attend to. A third difficulty. Two related things. One, I mentioned Douglas Rae earlier. But his book about urban politics. But he has an essay called Tyrannies of Place where he points to the fact quite prescient. He wrote this more than 25 years ago. That we're increasingly becoming a segmented democracy. That is, people tend to spend time around people like themselves. And this, of course, greatly facilitates that because people will sort by income, if you go to the ones in Florida, often by ethnic groups, into these relatively homogeneous, certainly financially homogeneous, groups. And we know from Khaneman, who I've mentioned to you before, and his research with Cass Sunstein, lately of the federal government and in the Obama Administration, now back at the Harvard Law School, but like-minded people, if they just talk to one another, tend to become more extreme. And so, if we get an increasingly segmented democracy of people only hanging around people who look and talk like themselves, this will reinforce a lot of the divisions that are contributing to polarization of the electorate. And when we talk next week about the segmentation of media markets, we'll see that this reinforces that trend. That sort of out of sight, out of mind of people not like yourself. So, a final point I would make is that these organizations are here to stay. If you go back and look at McKenzie's first book, which he published in 1996, Privatopia, the tone of that book is, this is terrible. The sky is falling. These things should be stopped. For all the reasons I've mentioned and other reasons. But the chapter I gave you from his book from 2011, the tone is completely different. There are 25 million of these things here that are here to stay, and figuring out how to manage them and manage with them is really the order of business going forward. Okay, on Thursday we will talk about prisons and the military. (soft music) |
Power_and_Politics_in_Todays_World | Lecture_1_Introduction_to_Power_and_Politics_in_Todays_World.txt | - Hello everybody and welcome. How is everybody today? Great. Well, I'm delighted to have the opportunity to be giving the DeVane Lectures. And the DeVane Lectures, as you can tell, from looking around you double as being a regular Yale course for credit that students can take for credit and lectures that are open to the general public. These lectures are going to deal with power and politics in today's world and by today's world, I'm gonna mean the 30 years since 1989 and the 30 years since 1989 are and have been an incredibly tumultuous period of very great change. And that's for xxx unusual. For instance, if you compare it to the previous 40 years in most of the advanced capitalist democracies, they were a period of relative stability after World War II. In most countries, it was an era of great prosperity even countries recovering from World War II like the countries of Europe were being rebuilt with Marshal Plan aid and it was a period, partly for demographic reasons, of very great political stability for people who grew up in that period. Internationally, as well, it was a period of very great stability because partly, because of the Cold War. It's true we had episodes like the Cuban Missile Crisis and the Vietnam War but as the Vietnam War indicates, most of the conflicts within the Cold War were played out as you like, as proxy wars in other parts of the world. From the point of view of the citizens of the Western democracies, except for those who are actually fighting in Vietnam, it was a far-off war that didn't have a great impact on the stability of people's lives and that is very different from what has been experienced since 1989; time, if you like, has speeded up a great deal. We've seen incredible change in three decades. And those are the three decades that I am going to be exploring. One pedagogical challenge that presents is that for some of us in this room, the last three decades are etched into our minds as like it was yesterday. We experienced them in real time but there are many people in this room who were never born until long after that. For them, whether it's the last 40 years or the last 60 years, 40 years or the 60 years after World War II, it's all history. One of the first things I need to do and I'm gonna try and make this a regular practice during the course is I need to take everybody back and to make people understand who weren't there what it was like and then to remind people who've lived through it of things that they might have forgotten. So let's just go back to 1989 in Berlin. - The Berlin Wall, once it divided East from West, now on its way to becoming an artifact of history. (upbeat music) This the CBS Evening News. Dan Rather, reporting tonight from in front of the Brandenburg Gate in Berlin, Germany. Good evening. These are the sights and sounds of the continuing celebration of Germans about the symbolic, not the literal, at least, not yet but the symbolic tearing down of the Berlin Wall. It's impossible to completely describe how deeply Germans feel about what's happened here. East German border guards tonight were literally tearing down portions of the wall itself, not the whole wall but portions of the wall to make it easier for East Germans to come into West Berlin and as the joyous hordes of Berliners were still streaming through the wall, the East German communist government said they can come and go permanently, they can come into West Berlin, have a look and then come back home again with no special documents required. Today, what goes through your mind and heart? - Well, of course, I look back to all those years of hardship for the families even more than for the country as a whole and it's moving to see families getting together again. My feeling is that we are very close to an end of the artificial division of Berlin and I also believe we are close to the point where the parts of Germany will become much closer together. This, of course, only within the reasonable European framework. Anyone know who that was? Xxx Billy Braun, the former Chancellor of Germany who actually would die a couple of years, three years after that, so he was one of the celebratory crowd at that time. So that might give you a little sense of the shock and the enthusiasm that people experienced in the latter days of 1989. This had been a period in which the Soviet Union was clearly losing its grip on Eastern Europe. It wasn't just Germany. Right across Eastern Europe, all through 1989, there had been these massive resistance movements developing and the Soviet Union was losing, it was clearly not in a position to intervene in these countries. And they were shedding totalitarian control for the first time in decades. Of course, the great exception which we will be talking about next week was in China where demonstrations in Tiananmen Square that June had come out, meant a very different result but in most of the world, after 1989, it seemed like democracy was on the March. We saw the democratization of all of the former Soviet Union bloc countries in Eastern Europe even places like South Africa which had been mired in a bitter conflict for decades transitioned in the early 1990s to a multi-racial democracy. The problems in Northern Ireland finally settled with the Good Friday Accords in 1997. Even Israel-Palestine which had been one of the most intractable conflicts for decades and decades seemed to be moving towards a resolution in the early 1990s. There were the Oslo Accords, the PLO and the Israeli government were negotiating a settlement. Anyone who was in Israel or the West Bank at that time anticipated that there was going to be a settlement until the assassination of one of the key protagonists, Yitzhak Rabin in November of 1995. And that would lead to an unraveling of that potential settlement but in the early 1990s, it's really I think difficult to overstate the enthusiasm for change. This is the period when Francis Fukuyama was talking about the end of history by which he meant that liberal democracy was sweeping the world. In fact, at the turn of the 21st century, we finally went from a world in which most countries in the world were not democracies by most standard measures to a world in which most countries were democracies by most standard measures. So it's not surprising that one would have had the kind of enthusiasm that Fukuyama had at that time and that was almost, what's the word? Just snowballing across so much of the developed world. Enormous confidence in democratic capitalism and enormous confidence in the idea that many people were gonna be lifted out of poverty and that the world was heading for a kind of benign equilibrium, as an economist would say, Fukuyama's end of history idea. now let's fast-forward three decades and let's stay in Germany and here's a very different thing to look at. (speaking in foreign language) (audience cheering) (speaking in foreign language) So he's a leader of the Alternative for Deutschland, far-right anti-immigrant, anti-system political party and what he is celebrating is that they have crossed the 5% threshold. This is the AFP here on this graph. Germany has a 5% threshold to get seated in the parliament so if you don't get 5%, you get no seats, they hadn't gotten 5% before and now they had won 5% and so they saw themselves as getting a foothold in German electoral politics for the first time. Germany, in 2017 was coming out of a situation in which there had been a grand coalition between the SPD which is the left of said Social Democratic Party and the right of center CDU, Christian Democratic Party led by Angela Merkel. And the SPD were very unhappy, they had been in this grand coalition for a long time and they found that they were paying a huge price with their supporters. They were getting less and less of the vote for reasons we'll talk about later in the course and they announced that they were not gonna participate in a grand coalition again and they were gonna go into opposition and rebuild themselves. Angela Merkel then spent the next six months, Alternative for Deutschland were and still are seen as beyond the pale, nobody will form a government with them. So she spent the next six months trying to put together a coalition with the Green Environmentalist Party and the so-called Free Democrats. The Free Democrats would be what we would think of as, in this country, as Rand Paul libertarians. They are for small government, low regulation across the board. It's not surprising that she couldn't do it because the greens want green regulation. They want environmental regulation. That's their raison d'etre and the Free Democrats want no regulation or certainly, less regulation. So they stumbled along in and out of negotiations but they weren't able to form a coalition. However, over the course of that six months, all the opinion polls showed that the AFD, the Alternative for Deutschland was actually rising in popularity. And so the German president was very against having another election in the face of the stalemate with the Social Democrats refusing to join in a grand coalition on the one hand and Merkel's inability to construct a different coalition on the other. They all knew that if they went for another election, the AFD would do even better. So finally, after much hand-wringing, the SPD was persuaded after extracting a very big set of concessions, including six ministries and the Finance Ministry. They were persuaded to go back into a grand coalition even though a lot of their membership didn't want it. So terrified were they of the prospect of another election in which the far right would do even better. So we thought German politics was kind of settling down at this point but the following year, this is what you see happening. - [Presnter] German Chancellor, Angela Merkel, who's led Germany for 13 years has offered to step down as her party's leader and said she won't run for office again after her term ends in 2021. Her announcement came a day after her party, the Christian Democratic Union, saw a disappointing performance in a key regional election in the state of Hesse. The far-right, Alternative for Germany party claimed more than a dozen seats in Hesse's Parliament for the first time. The entire migrant party now controls seats in all 16 of Germany's state Parliament's plus the National Bundestag and European Parliament. - So there it is. They had also done very poorly in the Bavarian regional elections. They had hemorrhaged about 10% of their vote to the AFD in these regional elections and by the way, the SPD wasn't very happy either because they were hemorrhaging support to the greens. And this is something we'll see playing out in many European countries. The establishment parties are shrinking and becoming weaker and the parties on the fringes are growing and becoming stronger. And it was impossible not to start thinking about the past and thinking particularly about the 1920s and 30s. (speaking in foreign language) - [Presenter] As a result of the elections of July, 1932, the Nazis became the biggest party in Germany with 37% of the vote. - So if you wanna go back to the 1930s, you can see that that video was about the July 1932 election and you can see what happened in the subsequent elections that it was a very unstable system. They were having multiple elections and of course, 18 months later, Hitler came to power. And if you go to Germany, you hear very anxious people talking about is this back to the future? Is 2017 and 2018 some kind of replay of the empowerment of extremist parties and of course, it wasn't just Germany. In 2016 we had massive shocks delivered to establishment parties with the Brexit result in the UK and Donald Trump's populist stampede to the presidency in the US, both widely unanticipated outcomes by most of the establishment parties, pundits and politicians. And you could go around the world. In the Austrian elections of 2016, people are very relieved that in the runoff, the Green candidate actually defeated the far-right quite handsomely by 54 to 46% but if you look at the legislature, again, you see the far-right gaining ground, the establishment parties coming in fourth and fifth in the 2017 legislative elections. These are the parties that would normally have come in first and second while the far-right party increased its vote, putting them a close third with 51 seats, while the Greens fell below the threshold and won nothing or if you look at Belgium, you see a center-right party retains its majority but if you drill down a little bit, you can see that there was an increase of support for the far-right Flemish Vlaams Belang which received almost 12% of the vote gaining seats. If you look at Italy, you can see the center-left party ceding power to the center-right but many of the votes for the center-right party are coming from the league, so-called, again, a far right-wing populist party which ends up with 125 seats and 17% of the popular vote, of 109 seats and I could put up another seven or eight or nine slides of different countries that basically tell the same stories in country after country across Europe both Eastern Europe and Western Europe, including countries that we thought of as bastions of civil social democratic stability like Sweden. You see these far-right parties doing well, Turkey, Latin America, elsewhere, where anti-establishment parties that sometimes also verge on being anti-system parties are gaining ground in many legislatures. So if you think about the contrast between the videos of 1989 and where the world has been since 2016, it couldn't be more dramatic and in some ways, it's a big downer but my first thing I wanna say is don't get too depressed. It's not all of course, for depression. The central questions of this course are three. How did we get from there to here? What are the challenges and prospects going forward? And most importantly, in the last part of the course, how could we get to a better place in many of the countries that we're talking about? We'll spend a lot of time on the US but not exclusively in the US. I should say a little bit about the distinctive approach that I'm gonna be taking in these lectures, not to say it's the best approach; there are many ways to look at this kind of material but it is the approach that I'm going to be using. And the first thing I would say about this approach is that I'm going to be studying history with the tools of political science and political theory on the one hand but also using history to keep political science and political theory honest. So what do I mean by that? Well, one thing that is remarkable about the events of 1989 is that they supply us with a terrific natural experiment from the point of view of social scientists. If you look at the literature, for example, of people who studied European democracies, before 1989, they were essentially cycling endless numbers of theories through the same old data set that everybody had had for four decades. And they didn't have any a big exogenous change. 1989 is a big shock to the system and suddenly, we have, for instance, in Western Europe, in Eastern Europe, the addition of a whole lot of new parliamentary democracies. That creates possibilities of thinking about long-established conventional wisdom, testing theories against new data which is the gold standard for social science rather than testing theories on data out of which the theories have been developed, which tend to result in what statisticians call just-so stories, fitting the curve to the data. Suddenly we have all this new data, coming along in real time. If you think about and we're gonna talk some about some of the standard theories in political science like modernization theory, which is a theory about how as economies modify, certain kinds of political changes become more likely. It was long held that modernization produces democracy. There were many variants of modernization theory. And we will talk about some of the differences among them but now we have new data and whether modernizing economies will produce democracy. It is long been conventional wisdom that democratic systems are incompatible with state-run economies. If we look at what's happened since 1989, we've gone to market economies in some of the post-communist systems but others like China and Vietnam have become state capitalist systems of a certain kind while retaining non democratic politics. So we'll have cause to think about theories of that general sort. There's been a lot of conventional wisdom about the conditions for stable democracy that suddenly can be put to the test on a whole slew of new democracies. Is it all the economy stupid or do beliefs of citizens matter? And what kinds of beliefs matter and what about the beliefs of elites? All of these things, we can look at again, in new contexts. So we have lots of new data to think about them, provided by this dramatic break of 1989. There's a lot of conventional wisdom about the relations between business, government and labor that has built up among political economists and political scientists over the last several decades before 1989. We now have big power shifts partly because communism as an economic system has been taken off the table. Well if communism, as an economic system, is taken off the table, how does that restructure relations between business government and labor? It turns out it really has a big impact on those relationships. So that's another of the topics we will be looking at. How do electoral systems affect things like inequality, provision of environmental legislation and public goods? A lot of conventional wisdom there about which types of democracies are more likely to do that and which are less likely to do that start to look before and after 1989 and we're gonna discover that some of the conventional wisdom needs rethinking. So on the one hand we're bringing the tools of political science and the theories of political science to bear on the data that's thrown up by this last 30 years of history. On the other hand, we're using that data to keep the political scientists honest, precisely because we have this whole smorgasbord of new results as a social scientist would put it but then I also said I'm gonna use the tools of political theory. So political theory, I should confess, truth in advertising by my first profession, if extended is a profession is I'm a political theorist. I cut my teeth in the world of political philosophy. I'm in normative things about what should happen, how the world should be organized, rather than empirical work on how it is organized. We are definitely gonna be thinking about normative questions here as well as we go along, what should happen, what might have happened but the confession I need to make about my home discipline is that it reminds me of the story about the fella who goes up to a farmer in Donegal and says how do I get to Dublin and the answer that comes back is well, I wouldn't start from here, sonny. That is to say much of political philosophy develops theories that take no account of where we actually are and how the theories that people argue about in the journals and in the literature actually could be implemented in the world, if at all. And this spills over into normative arguments made by other scholars. Thomas Piketty in his book, Capital in the Twenty-First Century, argues for a 4% global wealth tax. Well, good luck with that. Who's gonna implement a 4% global wealth tax. So when I think about normative questions, it's gonna be from the perspective of how might the goals that we think are desirable actually be achieved. I am going to spend, as we go along, through the different topics that we're gonna be discussing. I am gonna be focusing on paths not taken, things that might have been done differently and here the sorts of things I'm gonna be talking about are things like NATO expansion, after the collapse of the Soviet Union, the NATO alliance, we'll talk about this next week which was expanded over the next several decades to include most, eventually, all of the former Soviet bloc countries and there were overtures, also to including former Soviet states like Georgia and several others. I'll, talk about this Ukraine, Georgia and Ukraine were both talking about, joining NATO that part of the reason the Soviet Union invaded Ukraine. So we will talk about whether there was a path not taken at that time French president Francois Mitterrand thought that NATO should be shut down at the end of the Cold War it was a defensive alliance that had lost its purpose. How realistic was that? how might the world have been different had that happened we're going to be talking about the Global War on Terror I should in after 9/11 when, we invaded first Afghanistan and then Iraq. Was there another path that would have been politically viable that would have led the world in a different direction. We will talk about the response to the financial crisis of 2008 in 2009, again what were their possibilities that, were, ignored or overlooked that might have been both politically feasible and more polished effective in the sense of policy and so that brings me to underscore the third feature of the approach that I am going to take in this course which is when you look at, people who talk about public policy, it tends either to be policy wonks who just go on about what would be, the best policy or political scientists who talk about why some policies get adopted in some systems and not in other systems but there's very little discussion of what is the effective political way of achieving a good policy or the effective political way of blocking a bad policy and so when I talk about innovations in the last part of the course, it's very much going to be in a way that marries considerations of politics to considerations of good public policy so, that that is the flavor of what we're going to be doing here and as I said it's not the only way one might study these materials ,but it's the way that we're going to be doing it in this course. Let me talk a little bit about the shape of the course. It falls into five, sections. The the first one is going to starting on Monday deal with the collapse of communism and its aftermath. There we going to look at collapse of communism in Eastern Europe in Russia, in China and in Vietnam. Although I will talk about some other countries along the way. We're going to talk about the rise of a unipolar world, until 1989, we had a bipolar world in which, as I said a lot of actual stability, that came about as a byproduct of the Cold War at least among the the major powers even if they maintained this stability, partly by, acting out their conflicts in proxy wars in, Asia, Africa and at America but it by and large the great powers were the nucleus standoff worked and we didn't have a major conflict between them. Now we have gone to a unipolar world dominated, by a single power ,after 1989. How does that restructure politics and the possibilities for politics that will be also part of our concern. And then I'm going to talk about the, politics of the economics. What I'm going to call the rise of neoliberalism at home and the Washington Consensus abroad and this is put on this is basically it comes to, exist because of the collapse of an alternative to capitalism, the disappearance of communism as a viable, political system as a viable economic system and so you have this idea that's often called neoliberalism that basically has three elements. Its trade deregulation, trade deals getting rid of restraints on trade, getting rid of internal regulation within countries, and massive privatization of formerly state assets that was called, the neoliberal approach to political public political economy. And when it's translated into a set of recommendations or requirements for countries in the developing world, it was called the Washington Consensus, was essentially adopted by the World Bank by the IMF as a condition for giving loans to developing countries and it, was essentially taking neoliberalism global. And so the, post communist era is marked by, this massive confidence in, the the capacity of unregulated or mildly regulated free-market capitalism, to deliver the best results, for every country in the world then we're going to talk about the new global order, that is ushered in by this. We're going to look at whether democracy really was on the march the people thought about a potential we had talked again the conventional wisdom in political science was that the democratization had preceded in three waves. The first wave being the gradual expansion of the franchise in what we think of today as the older democracies, the second wave being decolonization in Africa and Asia and Latin America after World War II. The third wave coming in 1989 to 1991 and people started to wonder whether there's a fourth wave now with democratization of South Africa with settlements in places like Northern Ireland, initially with the Arab Spring people wondered whether we were going to start to see more democracy as part of this new global order. Then we're going to look at the international institutions that developed in this global order, we'll talk about things like, the International Criminal Court which for the first time would hold dictators to account for their activities of repression. There was the creation and of something called a doctrine at the United Nations called the responsibility to protect. Responsibility to protect, says that this was in the wake of, things like the Rwanda genocide of 1994 and what happened in Kosovo, in 1999 eventually we got responsibility to protect where the UN was said all governments are going to be held accountable for severe human rights violations within their own territories and if they don't respect them, the UN is going to intervene, militarily this is you know a big change we're saying we're not going to respect the sovereignty of Nations so the new international order seems to be, affecting not only relations among countries but relations within them. The third part of the course I'm calling the end of the end of history and this is really, has its ultimate roots with 9/11 and the emergence of the global war on terror. The invasion of Afghanistan, Iraq, Libya and eventually the collapsing Syria we're going to see that the idea that history was heading in some benign direction, was getting harder and harder for people to hold on to. we're going to look at the resurgence of state capitalism. we're going to look at what China was, has been and is doing in Africa. We are going to be looking at Russia's resurgence as a global power, they had been back on their heels, for much of the 1990s and early 2000 but now they became actively involved, in, the Middle East and elsewhere, we're going to, look at, the new role of business in many political conflicts around the world sometimes with benign purposes in mind, but often not and so we're going to start to see a much messier politics. The fourth part of the course, is about the new politics of insecurity. This is really ushered in not so much by the financial crisis itself, but by the way in which governments responded to the financial crisis and this will be one of the areas where we will be looking at paths not taken but, we will see that there was, growing insecurity in the workforces of many of the advanced democracies that had been building up for decades, and the financial crisis, threw into sharp relief the fact that most governments were not doing much about it if anything at all and in fact that they ended up bailing out the elites, while doing nothing much, for the people who have been most harmed in the crisis and so we will look at the response to, the financial crisis in a number of countries particularly in the US and then we will all in the last part of the course, what is to be done that great Leninists slogan, we will be looking at, two things really. First how the, voter sentiment in these countries was, so poorly misdiagnosed by so many, political elites and people in control of political parties, misdiagnosed to the point where not only did they fail to respond, to the growing economic insecurity, that was experienced by many, many millions of workers within their own countries. They were actually implementing political reforms that were likely to make things worse. And so the last part of the course will be looking at those two interacting issues. What sorts of economic policies were pursued or failed to be pursued. What kinds of policies might have been pursued and how did the reform of political systems over the last 15 or 20 years, make the problem worse and that will lead me, to make some arguments, about how we should, think about, politics going forward. So, that is what we're going, to be doing. It's a big menu, we have, you know we have 26 now soon to be 25 lectures in which to explore it. I'm going to, before breaking today, I'm going to talk about a few logistical matters. One is that I have been asked to announce but you can see for yourself, that the Office of Public Affairs is, taking pictures and these lectures are going to be videotaped. And this will include, we will have microphones so we can have questions from the floor and interaction but you should know that, this is all being tape recorded and so anything you say can be taken down and used in evidence against you. So, a couple of other logistical things. One is I want to introduce Christina Seyfreid. Christina, she's the head teaching fellow for the course so for she will be overseeing the grading of students who are taking the course, for, credit and she will also be working with me, in, running office hours for people from the community, which we're both going to be doing. I'll say a little bit more about that in a minute but, you will see more of Christina as we go along. Office hours I'm gonna be having office hours in Rosenkranz room 201 from 10:00 to 12:00 on Wednesdays and Fridays. The presumption there is that from 10:00 to 12:00, Yale students have priority. These are walk-in office hours not by appointment and on 10:00 to 12:00 on Fridays, people from the community have priority but who knows if anyone will show up but that's the way it'll work 201 Rosenkranz Hall which is right opposite, the new colleges up the street. Christina is going to have office hours for people from the community on Thursdays, from 5:00 to 7:00 at a place yet to be announced and one thing we're going to do, the office hours are not just going to vanish into the ether we're going to, Christina and I are going to film five sessions over the course of the semester, where she and I will discuss what's come up in the office hours and we will post that, on the course website. I do this with my Coursera course and it proves to be quite helpful. So, we will get questions that have come up in the office hours both in our office hours and in office hours, between the students and the Teaching Fellows, that they will be participating as well and we will publish those videos of the office hours on the course website. Access to reading. Anyone who has a Yale ID can get almost all of the readings on Canvas, there's a few, I think three of the book, we're using too much for copyright reasons to put them on Canvas but they're all in the library for anyone who doesn't want to buy any. The, people from the community we have, made available in the New Haven Public Library if you don't have access to Canvas us a couple of sets, of the books that we're being using in the course or you can get them that way. Those who are taking, the students who are taking the course for credit the official exam is the last afternoon of the last day of exams, which I'm sure some of you have already bought plane tickets that are, inconsistent with that. So actually we're going to have to alternate exams. One on December 11th and this is primarily to accommodate SOM students because, of the timing of the SOM elective, exam period and that exam will, if we can arrange it be held at SOM, and then there'll be, but others can take it then as well if they want. And then there'll be an another alternate exam on Friday the 13th so there'll be different times at which people can take the exam. Policy here will be no laptops, no screens, no iPhones. I came to this policy a couple of years ago. It definitely works better. It's just too tempting to be doing that Amazon shopping for your grandmother's birthday present that you forgot about and it's very distracting not only for the people around you but it's also distracting for me, to teach. We will post the slides on Canvas and on the course web page off the lectures so people will have access, to the slides. What you'll find is if you go and you look at the slides on Canvas, they'll look like they'd been turned into a PDF and the video will not play. If you want to play the video, you have to download the slides and then it will read, it will regurgitate as playable videos, so it is possible to play the video, yes so no laptops, no screens, no phones and everything, everything I put up here will be available to you, so you don't have to take copious notes of what appears on slides. Comments, questions. Yes, sir, you're gonna have to yell. We will have microphones here starting next time. (student speaking faintly) You can't. So that's why we put a set of the reasonings in a New Haven library and the reason for it, it's not just Yale being mean. It turns out there're two sets of copyright rules so if you are photocopying chapters from a book, you are allowed to take up to 10% of the book, otherwise you have to buy the book if you're using it but universities, for their students, have a more permissive rule. So we can take more than 10% from a book. I've forgotten what the exact number is but it's much more than 10%. We can take more than 10% from a book and put it up on Canvas if it's only available to Yale students but if we made them available to the general public, Yale would be violating copyright laws. So that's why we have bought two complete sets of the books and put them in a New Haven Public Library, from which we have been xeroxing most of this stuff. So you should be able to get the vast majority of the readings if you wanna go down there and get them. Other questions, comments, observations? If there's a question in your mind, it's probably in the mind of 30 other people so they'll be grateful if you ask it. Yeah, you just have to yell. (student speaking faintly) Are we gonna look at the European Union as well as what? (student speaking faintly) We're certainly gonna look at the growth of the European Union and the expansion of the European Union since, in fact, we're gonna talk about it next class. We're gonna look at the expansion of the European Union since the Cold War. I'll talk very briefly about the early, this course is really post 1989. We've just hired David Engerman and Arne Westad, who are both cold war historians who are gonna be teaching courses on the Cold War. So if you really wanna deal with the post-war Europe from into war Europe II to the end of the Cold War, these are the guys who are gonna be teaching that stuff going forward in Yale College. So I'm very consciously not doing that. So I will talk about the way in which the European Union was formed which has had a lot to do with its current troubles but mostly, I'm gonna focus on since 1989. Yeah, pardon? (student speaking faintly) There are no sections in this course. This is why we have an hour and a quarter lecture rather than a 50-minute lecture. Yeah. (student speaking faintly) The topics for the papers will be posted two weeks before the papers are due and we will give, I should say a couple of things about that. This is not a research course so you can write a first-rate paper based on reading the material on this syllabus. You are not expected to go beyond the syllabus. And we will post, two weeks before the papers are due, a list of topics of and it will be, there will be very significant amount of choice on it so we'll probably give you four or five possible topics to write about. Yes, sir. What is the expectation for graduate students? The same course requirements, two papers and a final exam. Yeah. Yes, sir. Pardon. - [Student] Is there a limit for (mumbles) students to join the class? - No, there's no limit. This is an uncapped course. I'm hoping we can stay in this room, it looks like it. Numbers tend to go down rather than up during shopping period. It's not a capped course. Yeah. (student speaking faintly) People have taken courses with me before know that I have a pretty interactive lecturing style. So yes, there will be microphones and there will be questions. I'm not planning to talk for an hour and a quarter every Tuesday and Thursday. So there will be significant amount of interaction. Yes sir. (student speaking faintly) Yeah, we currently, I believe, have six in addition to Christina's Teaching Fellows, all of whom will be holding office hours. They'll be available and you will be able to go and talk to them about what's come up in the class, about your papers and so on. And the conversations you have with those teaching fellows, we will also address then in the videotaped office hours that we post. Okay. So any other questions, comments, reactions? So on Tuesday, we're gonna talk about the collapse of the Soviet Union from Soviet Communism to Russian gangster capitalism. I will see you then. (music) |
Power_and_Politics_in_Todays_World | Lecture_24_Unemployment_Reemployment_Income_Security.txt | - Today we're gonna talk about Unemployment, Reemployment, and Income Security, or income insecurity as the case might be. So, let's start with some something to get our attention focused on that question. - [Narrator] We're hauling 20,000 pounds of freight down the Florida Turnpike in a self-driving, robotic truck. It's been retrofitted with a self-driving kit made by Starsky Robotics. Kartik Tiwari and Stefan Seltz-Axmacher founded the company in 2016. - We think that sometime towards the end of the year, we could be doing this run without a person behind the wheel. - This year? - Yeah. - Self-driving trucks this year? - Yep. - [Narrator] And if it's not his company, it might be OTTO, whose truck made headlines last October by driving itself across Colorado to deliver a shipment of beer. OTTO is owned by Uber, which has also been testing self-driving taxis in Pennsylvania and Arizona. But here's the thing. Once our trucks and taxis drive themselves, what will happen to the people who used to do those jobs? In the US, that's 180,000 taxi drivers, 600,000 Uber drivers, and three and a half million truck drivers. - We really need to start to think very seriously about this. - [Narrator] Martin Ford is the author of "Rise of the Robots". - [Martin] This is it. - [Narrator] He says driverless cars and trucks are just the beginning of a wave of automation that will threaten millions of jobs in every industry at once. Like America's nearly five million store workers. - The cashiers are totally gone. You're gonna end up with the equivalent of a Walmart with a handful of employees. You scale that out, and that's just extraordinarily disruptive. - [Narrator] You name an occupation and there's somebody considering a robot to take it over. - Look how delicate. - Perfect every time. - [Narrator] At Zume Pizza in Silicon Valley, four specialized robots help make the pizza. Eventually, the company plans to replace the remaining humans on the line too. The common wisdom is that robots primarily threaten repetitive, blue-collar jobs. "Not so," says Martin Ford. - We're seeing dramatic advances in the area of computers analyzing tumors, recognizing medical scans, mammograms, and being able to find disease. We're seeing algorithms moving to areas like journalism, for example. - Wait, wait wait. Certainly not journalism. - Oh yeah, absolutely journalism. By one account, every 30 seconds there's a new story published on the web or maybe in a newspaper that's machine-generated. - [Narrator] Algorithms are even threatening the masters of the universe. Earlier this year, BlackRock, the world's largest money manager, announced that it's laying off dozens of human stock pickers and replacing them with robots. By 2025, across the financial industry, artificial intelligence is expected to replace 230,000 human workers. - Bring on the disruption that is automation. - [Narrator] Elisha Wiesel is the chief information officer at Goldman Sachs. The company now hires nearly as many computer engineers as financial workers. A quarter of these people aren't traders. They're coders, writing software to automate the routine grunt work of employees all across the company. According to one recent study, 47% of American jobs could be lost to automation in the next 20 years. - I'm sure 20 years from now, almost no one will be driving a vehicle. Young people are forward looking and they say, "Well, I guess I'm not gonna have a driving career, "so I'm not gonna go there." - Well, except that these young people might think, "Well, maybe I'll go into retail," but that's also going away. "Well, maybe I'm be a chef," but that's also going away. "Well, maybe I'll be a paralegal," but that's also going away. - So, let's do the following thought exercise. It's the year 1900, and 40% of all employment is in agriculture, right? And so some twerpy economist from MIT teleports back in time to Farmer Pogue here and says, "A hundred years from now, "only 2% of people will be working in agriculture. "What do you think the other 38% of people are gonna do?" - Well, I wouldn't know. - Would you say, "Oh, search engine optimization, "health and wellness, software and mobile devices,"? Most of what we do barely existed 100 years ago. - [Narrator] In other words, just because we can't predict what we'll be doing doesn't mean we'll be doing nothing. - So, there you see the great debate that is now being fostered among economists and others about the likely effects of automation. On the one hand, there are people who are saying this is a fundamentally new development, that these, that essentially all employment is going away and there's gonna be nothing for people to do. On the other hand, you see people say, "Well," as was said at the end in the last part of that clip, "In 1900, if people had had to speculate "where everybody who was gonna be put out of work "in the agricultural economy was gonna wind up, "they wouldn't have had any idea as to where they would go." And so just as all kinds of industries have emerged over the course of the last century that nobody could have dreamed of at the beginning of it, so the same thing will happen in the next century, and there will be all kinds of new employment. Life has more imagination than us. People will innovate in new ways, with techniques we can't even consider. And that is certainly gonna be an open question, not a debate we would settle in this course, let alone whether anybody else can settle it. What is clear, though, as I've emphasized in previous lectures is that the idea of long-term employment security has gone away. And so our agenda today, whether people are gonna ever get new employment or they're gonna work through a series of up to 13 of 15 or 20 jobs over the course of a lifetime, the idea of a continuous stream of income from one job that people get upon finishing their formal K through 12 or perhaps K through college education has gone away. And so our agenda to day is how to think about the best responses to endemic and permanent wage insecurity. But I wanna do it with one eye on last Tuesday's lecture. We wanna think not only about the economics and the ethics of what we want to consider as desirable policies but also the political feasibility. Where is it likely that we can build coalitions to actually implement the sorts of policies that might be desirable and then find the resources for them, the proximate goals to make sure they move steadily forward, the entrenchment to stop them from being undone, and the leadership to put together and hold together the coalitions to do that, along with a moral narrative that will be compelling to people and can head off potential blocking coalitions. So, we wanna keep all of that in mind. And I'm gonna talk about four different topics in connection with this. The first is an idea that has actually been around since the 1970s. Then this is the idea of UBI. That used to stand for unconditional basic income. Now refers to the idea of universal basic income. But this is the idea that people should be given some kind of financial grant unrelated to work, unrelated to employment. Second, we'll focus on minimum wages and the political pressure to put up minimum wages in response to employment wage stagnation and the relative decline of wages that we have talked about at some length. Then we will look at something called the EITC, the Earned Income Tax Credit, which is a kind of wage subsidy that has also emerged in the 1970s and gets a lot less press, but we, in my forthcoming book with Michael Grant, think is much more interesting and promising. And then finally we will talk about the history and prospects for transitional assistance. There have been programs since the Kennedy administration in this country to provide people with transitional assistance in moving from one job to another that have not fared very well, and we'll talk about why and what the prospects for them might be. So, let's start by going first to the idea of an unconditional or universal basic income. (gentle music) - Well, there is now, in Europe and beyond, more interest in the idea of an unconditional basic income than there has ever been in the history of mankind. Why? Some people say it's because we predict more automation, robotization. There is some truth in that, but this is combined with a greater uncertainty, greater doubts about the feasibility, the desirability of what has always been offered as the central answer to the problem of unemployment by both the right and the left, namely growth. And so if we think that growth is either, continued growth is either not possible or not desirable and that anyway it hasn't lost the problem of unemployment despite all the growth we've had, then it's high time to think about something else, think seriously about an unconditional basic income. A floor instead of a safety net. Something that would be a central element in any project for a free society and a sane economy. (gentle music) - So, that is Philippe Van Parijs. He is a Belgian political philosopher who has been championing this idea for some time. He runs something called the Basic Income European Network, since renamed the Basic Income Earth Network. Was originally formed in 1986. The book in which he developed this idea was called "Real Freedom for All", it's the subtitle, "What (if anything) can justify capitalism?", which he published over two decades ago. And he's a Rawlsian political philosopher. You might call him a sort of left Rawlsian, and he took off from the core insight in John Rawls's work, in John Rawls's work "A Theory of Justice". And this is a, in the first instance is a philosophical argument for the idea that people, every human being in a society should be entitled to an unconditional basic income. And so what Rawls did, and what was considered to be revolutionary and is at the analytical core of his "Theory of Justice", was he said, "Well, think about the differences "in capabilities between us." Some people say, differences in IQ for example. Some people say it's all about genetics. Some people say it's all about environment. And there, usually, there are these big, ideologically charged fights. Is it genetics, is it not? Are there IQ differences between different ethnic groups, provoked Charles Murray and others, got into these hugely charged, ideological fights over this question in recent decades. Rawls says, "If the differences between us are due to nature "well, then it's luck in the genetic lottery. "If the differences between us are due to nurture, "then it's luck in where you happened to be born." Either way, it's morally arbitrary. After all, you didn't choose to have the genes that you've got, and you didn't choose to grow up in the country, never mind the family that you found yourself in. You were either born into or perhaps adopted into, it's nothing to do with you. And so the debate about nature versus nurture as to what people are entitled to is really beside the point. And this is sometimes called luck egalitarianism because Rawls made an argument on that basis for the proposition that all the differences between us are morally arbitrary and we should think about principles of justice that take account of that moral arbitrariness. And he developed his own theory, which need not concern us here, the bottom line of which was that inequalities can only be justified if they benefit the least advantaged people in the society. And he thought that that would trickle up, a kind of Keynesian story if you like, and that would benefit everyone. So, Van Parijs had a slightly different take on this Rawlsian insight. He said, "What it means is that everybody should be entitled "to an unconditional basic income "at the highest sustainable level." Be up to political economists to figure out what that is, but that's what we should all be entitled to. It should have nothing to do with our capacities to work at different levels of productivity. Should be independent of work. And so this was the egalitarian case for as high as sustainable an unconditional basic income. This generated a huge debate among philosophers. Nobody paid a great deal of attention to it outside of the academy, and, indeed, a fatal objection was launched, eventually, by a philosopher called Susan Hurley in a critique of Rawls in which she said, "Well, you know,"... This is, I put the quotation up on the board, but the gist of it is the following. She said, "Well, it may be true "that the differences among us are morally arbitrary, "whether it's genetic luck "or luck in the lottery of where we happen to be born "and so no one is, "in some natural sense, entitled to what they produce, "but why should that create any presumption "in favor of an equal distribution? "It could just as easily create some different presumption "in some people's minds." And so the gist of Hurley's point was, actually, it doesn't justify any distributive presumption, including, perhaps, that we're entitled to the fruits of our individual labor. I'll come back to that. So, it no more justifies a presumption in favor of equality than a presumption in favor of any other distribution. A case will have to be made, and so it doesn't follow sort of in any pristine, analytical sense that we should be entitled to have a distribution that works to the benefit of the least advantaged, in Rawls's case, or in Van Parijs's case, it doesn't follow in any analytically tight sense that everybody should be entitled to the highest possible sustainable income. And so the normative case for this strong egalitarian presumption sort of goes away. Now, you might say, "Well, that's all true," but particularly in light of what we've been discussing in this course, there might still be a powerful case for a universal basic income. Not, perhaps not the highest sustainable universal basic income, but, nonetheless, it's something that would give everybody a cushion. It would enable people to deal with the vicissitudes of employment markets and the vagaries of secure jobs that we have been attending to. And, indeed, one of the presidential candidates, who has just made it into the debates last night, is a strong proponent of a UBI. So, let's give him a little more air time than he got last night. (students laughing) - And you're running for president based on this issue of UBI, universal basic income. What is a universal basic income? What is your plan? - Universal basic income is a policy where every member of a society, let's say every American adult, receives a certain amount of money to meet his or her basic needs every month. And my plan, the Freedom Dividend, is that every American adult would receive $1,000 per month, free and clear. - [Interviewer] $1,000 per month, every American over the age of... - 18. - 18. This, to a lot of people, is gonna sound like a completely nutso idea. So, just to be clear, has the US ever, has any government body in the US, existing or past, ever done anything like this? - Well, a law essentially identical to this passed the House of Representatives in 1971 under Richard Nixon. Martin Luther King was for it, Milton Friedman was for it, and a thousand economists signed a letter saying this would be great for the economy and society. And then a number of years later, Alaska passed something identical to this in the form of the Petroleum Dividend, where everyone in Alaska today receives between one and two thousand dollars a year, no questions asked, from the petroleum fund. And if you think about this for a moment, Alaska's a deep red, conservative state, and the Republican governor made this case to the Alaskan people. He said, "Who would you rather get the oil money, "the government, who's just gonna screw it up, "or you, the Alaskan people?" And the Alaskans said, "Us." And then now the Petroleum Dividend is wildly popular. It has increased children's, it's improved children's health and nutrition, it has created thousands of jobs, and no one can touch it. And that's been in effect for 36 years. So, if you think about it, what is the oil of the 21st century? The oil of the 21st century is data, AI, autonomous vehicles, and advanced technologies, and that's how we're gonna pay for our Freedom Dividend for all Americans. I want everyone here to reflect. Why is Donald Trump our president today? And the reason why Donald Trump is our president today is this. We automated away four million manufacturing jobs in Michigan, Ohio, Pennsylvania, Wisconsin, Missouri, Iowa. All of the swing states that he needed to win. And now we're about to do the same thing to millions of retail jobs, call center jobs, truck driving jobs, fast food jobs, which are unfortunately the four most common jobs in the economy. I have many friends in Silicon Valley who are working on this, who know full well that this trend is about to accelerate. So, if we don't wake up and start implementing meaningful solutions to the fact that we are going through the greatest economic and technological transition in the history of the world, then we are doomed to much, much worse than Donald Trump. (students murmuring nervously) - So, that is Andrew Yang's platform. And he makes some very, very convincing points. And if you think about it, if you have a universal basic income, it can also have a desirable effect on relations within firms. And this goes back to our old friend Albert Hirschman. If you think about... If you think about a world, so, where there's a continuum from what we might call a Dickensian nightmare, where there's no social support of any kind for people who lose their jobs, and at at the other hand a surfer's paradise. I use the term surfer's paradise because one of Van Parijs's famous lines was that even surfer's should be paid. So, surfer's paradise, it's Van Parijs's utopia. So, you should, you could then think that the power relations within the workplace are greatly affected by where you are on this continuum. Because if you are living near a Dickensian nightmare, where the costs of losing your job are extremely high, and your employer says to you, "I wouldn't mind a cup of coffee," you're gonna scurry off and get that cup of coffee and come scurrying back quickly, whereas if you are closer to the surfer's paradise and you could walk away from your job, you might say, "Well, go get your own coffee," right? And so there was a thought, there was big literature in the '80s and '90s which basically said that if you, rather than trying to have the government produce more democracy within firms to protect workers from exploitative practices, it would be better to just decrease the exit costs and then they'll have to treat you better. Just as, for example, at Yale, in 1992, the Supreme Court ruled that compulsory retirement was illegal, age discrimination. And so the interaction of tenure and the abolition of compulsory retirement created a huge problem for universities. And until that time, universities had treated emeritus faculty with contempt. They had pretty much ignored them. Once you were out of here, you were out of sight, out of mind. In order to get them to leave, the superannuated people like me, they had to start coming up with much better retirement packages to reduce the exit costs from employment. They had to start treating emeritus faculty better. They built a building where they can have offices, they gave them email accounts, and so on and so forth. So, if your exit costs are very high, you're much more vulnerable within a firm than if your exit costs are low. And so some people even argued, and I was one of them, that you might be able to get business support for a much more robust social wage, giving people benefits that are not tied to their work, if you, on the other hand, also promise business less regulation. So, you'd say, "Well, if,"... There were debates about this, for example, in the European Union that in countries that didn't supply a lot of worker protection, maybe the social chapter should provide stronger, stronger appeals processes when workers are abused and so on. And the idea was that since what business wants is flexibility at the plant level, you could try and get business support for a coalition to increase social wages, the quid for the quo being that they would get less regulation in the workplace. It would be easier for them to lay-off people because it would be less catastrophic for the people and so on. And so I and others pushed these kinds of arguments. The difficulty with them was political, for the most part. That because of the declining power of labor anyway, the notion that business had any incentive to get behind these proposals was pie in the sky. And so they were stillborn, and it's very difficult to make the case today. It might, nonetheless, be true that we should think that UBIs now is an idea has time has come. But why does it get so little support? Now, you can get polls. You can run polls where you can get 48% of the American public to say they're in favor of it, but they're standalone polls that don't say anything about the costs or how it would be funded. And if you look at efforts to actually produce it, you find that people are very resistant to this idea of writing checks to anyone that are completely unrelated to work. Why might that be? Why might that be? - [Student] American work ethic. - Work, there's a work ethic. So, what is a work ethic? (multiple students responding) We must bootstrap ourselves up. This idea that people get something for nothing or that they haven't done, they haven't made the requisite effort is extremely strong. It's something that political theorists sometimes call the workmanship ideal, this idea that when you... Working for something creates, it creates an entitlement over it, and because you know that, it gives you the incentive to work hard. And, indeed, if you go back to the John Rawls's argument that I mentioned at the outset, he realized, not surprisingly, he was a very intelligent guy, he realized that his moral arbitrariness argument was sort of at war with the work ethic, right? Because if you say nature, nurture doesn't matter, however you got your skills and capacities, it's all morally arbitrary. You don't deserve, none of you Yale students deserve to be here, right? There's no reason. You just got lucky. And you wanna say, "No, I worked hard to get to Yale." And so Rawls had a kind of cop out, that he said, "Well, yeah, the distribution "of our capacities and skills is morally arbitrary, "but the, what we choose to do with them is not." So, if one person decides to work hard every day and has got lots of Protestant work ethic, and the other one sits on the couch watching ESPN all day, then the one who worked choose, even if they have the same capacities, the one that chooses to work harder is deserving of their differential benefit. But, of course, that and two dollars and 75 cents might get you a seat on the New York City subway. Because the capacity to use your capacities is also morally arbitrary, arbitrarily distributed. If one person has parents who pump the work ethic into them a mile a minute and gets, teaches them, sits and does homework with them every night for 15 years and so on, that's very different from a parent who's maybe in an alcoholic daze, ignoring their children. And so the one child grows up with the capacity to use their capacities much better than the next person. And, indeed, maybe the capacity to stick at things even has a genetic component. Some people may be genetically more able to focus, all the research on attention deficit disorder and all of that. So, Rawls's answer is just a way of avoiding the problem that he's identified. And so here's the funny thing about Rawls, and when I teach this in political philosophy classes we spend a lot more time on it, but what you find is almost nobody can see anything wrong with Rawls's moral arbitrariness argument except for his cop out, which any smart student can shred in a matter of minutes. But nobody really wants to go all the way with it because it threatens the very idea of individual responsibility, right? Because how can you hold anyone responsible for anything if even our capacity to behave responsibly is distributed in morally arbitrary ways. And so even though it is true that our capacities to do things are distributed in morally arbitrary ways, people don't like to live with that implication. Even if it's true that some people got a lot of help from whoever it is, Sunday School teachers, that other people didn't get and more diligent parents and so on, people do not want to live with the implications of that. And if you want a political illustration of that, think back to the firestorm in the 2012 election when President Obama took the Rawlsian position without the Rawlsian cop out. - If you were successful, somebody along the line gave you some help. There was a great teacher somewhere in your life. (crowd cheering) Somebody helped to create this unbelievable American system that we have that allowed you to thrive. Somebody invested in roads and bridges. If you got a business, that, you didn't build that. Somebody else made that happen. - So, all hell broke loose, as you might recall. Here's just one of a million ads you could've found in response. - To say that Steve Jobs didn't build Apple, that Henry Ford didn't build Ford Motor, to say something like that is not just foolishness, it's insulting to every entrepreneur, every innovator in America, and it's wrong. - My father's hands didn't build this company? My hands didn't build this company? My son's hands aren't building this company? Did somebody else take out the loan on my father's house to finance the equipment? Does somebody else make payroll every week and figure out where it's coming from? President Obama, you're killing us out here. Through hard work and a little bit of luck, we built this business. Why are you demonizing us for it? (energetic music) We are the solution, not the problem. It's time we had somebody who believes in us. Someone who believes that achievement should be rewarded, not punished. We need somebody who believes in America. - So, it didn't turn out to be Mitt Romney, this, but it is a portent of things to come because this ad was used to reinforce the perception of sort of "Let them eat cake" elites, that he was sitting there in the comfort of the Oval Office, saying, "You didn't build that," and denying the efficacy of the work effic, of the work ethic. And so you might say it's a difficult case to make. Nonetheless, you might have thought, "I missed something here," because, as Andrew Yang says, not only is he pushing this idea now but he noted that people like Charles Murray are supporting it and, indeed, Milton Friedman has supported a universal basic income since the 1970s. But they have a slightly different thing in mind. So, here's Murray's take on it. - The basics are this. At the age of 21, every American starts to receive a monthly deposit electronically to a known bank account of 13,000 divided by 12. 13,000 is the total basic income, of which 3,000 must be devoted to healthcare. So, you have 10,000 of disposable. I think if we simply add that on to the current benefit system, it's not affordable, but what if it replaces everything else? There's no more Medicare, no more health, Social Security, there's no more welfare, there's no more corporate welfare, there's no more agricultural subsidies. I can go through the whole list of transfers. There's none of that stuff and there's also none of the bureaucracies that administer it. You have $10,000. And you say, "Well, you can't live on $10,000." Well, if you really wanna live all by yourself, no. You'll probably, you'll live a pretty crubby existence on $10,000. You get a minimum wage job, just a minimum wage job, and you work 2,000 hours a year at, let's say, 7.50, that's $15,000 plus 10, that's $25,000. Yeah, that's well above, way above the poverty line. Well, suppose for some reason you can't work or let's suppose you don't wanna work. You can still have a decent existence if you're cooperative. If you're cooperative enough to convince a friend, a girlfriend or boyfriend, a relative, or something else to pool your resources, if you just do it with two people, you got 20 with no working whatsoever. You got 30 if you got three people. You add a little bit of income and you're in the middle class. That $10,000 given to everyone has a huge effect on their ability to live a decent existence by doing reasonable things. Right now, we claw back benefits at a very low rate, so that if you're on food stamps and Medicare and you take a job, you very soon face a very punitive loss of benefits. I say, look, you keep your entire 10,000. We don't claw back a cent of it until you have 30,000 of earned income. - So, there you have it. The conservative case for universal basic income is something that would replace the entire welfare state. And if you look at the way this has been costed out, the Center on Budget and Policy Priorities in Washington, DC, which is a think tank, estimates that giving everybody a 12,000, he says 13, taking three out for medical care, but at the 12,000 minimum wage that somebody like Andrew Yang is pushing would cost three trillion dollars to fund per year. And even if you got rid of Social Security and Medicare, that would not be enough to fund it. And so if you start to think about what would the political coalition for this actually be, it dissolves before your eyes because the AARP would quickly become a leader in the opposition to this. You could think about universe, hospitals, hospital interest groups and administrators. All kinds of people, public interest groups protecting welfare benefits, would rapidly start to oppose it. And it's also worth saying that if you look at where there have been efforts to institute universal unconditional basic incomes elsewhere, they haven't fared much better. So, setting aside something like the Alaska sort of windfall benefit or the Norwegian, the Norwegian Sovereign Wealth Fund that's been funded by their oil, which is a different idea. It's a sort of... It's a sudden... A sudden benefit occurs and there's a notion to take, to spend it on the whole community. But if you think in general, in Switzerland they had a referendum on a UBI and it went down massively and, by the way, was strongly opposed by labor unions, as you might suspect if you start to think about it because it would erode labor unions' bargaining power. In Finland they had a modest experiment, a selected experiment with UBIs to see what impact it would have and abandoned it after a few years. So, the notion that even though it might appear that you could get a great deal of support for a universal basic income and it seems like it might not be such a bad idea from a variety of perspectives, I think that it's a tall order to think your way through a political coalition that could effectively achieve the kind of universal basic income that most people would want and be, certainly would be willing to accept. Between its sort of frontal assault on the work ethic that many people regard as necessary to get people to be productive and the fact that it would replace some of the policies that Americans are so strongly attached to, I think, Andrew Yang, there's a good reason why even though he made it just into the debates, he's not really getting much attention for this view. So, I wouldn't buy a lot of stock in it right now. Let's talk about minimum wages. Of course, this doesn't help with unemployment, but it certainly would help people who have been victims of long-term wage stagnation. - The fight for a $15 an hour minimum wage just netted its biggest victories yet this week. Yesterday, lawmakers in California voted for a bill that will raise the state's minimum wage to $15 by 2022. On Monday, Governor Jerry Brown is expected to sign it into law. Last night, New York governor Andrew Cuomo said he reached a budget deal, hiking the minimum to $15 in New York City by the end of 2018. Suburbs and other parts of the state would reach that level in 2021. The combined moves could affect as many as 10 million workers. - You know by today, many of the CEOs of the companies that we're fighting to get the minimum wage increased on, they have earned as much in the first couple of weeks in January as they will pay these employees for the whole year. That is not justice. - That ain't right. - That ain't fair, that isn't right. And we don't begrudge anyone their success and their income, but we do begrudge exploitation of the workers to that extent. - 90% of Americans have not had a raise for 40 years. - Pay your workers a living wage, 15 bucks an hour. - It is a fight for fair wages. - And so we could find clips of every one of the Democratic candidates now running calling for increases in the minimum wage, at least, at least $15 an hour being phased in or not. So, and this has been a constant theme of American politics that the Democratic party tends to fight for increased minimum wages, and there are efforts, constant efforts to increase minimum wages in urban, within city and state legislatures as well. But there are a lot of difficulties with minimum wages as policy, when we think. It's very easy politics, as you can see, to make that case, at least from the perspective of the base of the Democratic party. One question is, well, who actually pays for increases in the minimum wage? It might well be the case that the cost of the increased wages is simply passed on in increased prices with the, by the companies that have to pay higher wages. And, indeed, there was a interesting effort to raise minimum wages in Los Angeles a couple of years ago, and we did some interviewing of employers because the local Employers Association was supporting it, and they reasoned two things. One was the money would be spent in their stores in the city so that it would come back, but also that it would be possible to increase prices. So, it's not clear who benefits in that sense. But more indirectly, if you're thinking about what are the effects gonna be on employment? And here we have to think about whether you'll get higher minimum wages for some at the expense of lower employment in the economy overall. And in some counties, this takes on very extreme forms. In contemporary South Africa, as I think I have mentioned to you, where they have very strong protection of wages in the formal sector, they have extremely high levels of unemployment in the informal and very different, much lower wages in the informal sector. And, furthermore, they have hemorrhaged jobs to places like Lesotho and textile jobs even to China. And so there's a real cost, if you like, of increasing minimum wages from the standpoint of employment. And particularly now, when we're thinking about the temptations to go to automation, raising minimum wages might have even more deleterious effects on employment. I think I described to you the Nike factory I visited outside Ho Chi Min City, and one of the notable features of that Nike factory was it was extremely labor intensive. They were using technology that was maybe 30 or 40 years old. But the reason was that wages were low. If you start to ramp up wages, the incentive to send more jobs to robots will become stronger. So, the impulse to improve minimum wages is completely understandable, but it has these costs. Moreover, in a federal system like we have here, if you increase minimum wages in one state, it creates an incentive for employers just to go to other states. Just as if you have strong union contracts in one state, such as the auto workers unions in Michigan, the employers can simply move to Tennessee. So, that's, there's a kind of internal race to the bottom problem. And if you have a federal minimum wage, you can avoid that, but only at the price of contributing to outsourcing, the flying geese problem and so on. In an era when capital is very mobile and can produce things in anywhere in the world, propping up minimum wages in one place is water flows around a rock and it's just gonna flow somewhere else. So, it feels like a stop-gap solution. It's also worth pointing out, on the politics front, how difficult increasing the minimum wage is. For example, we talked earlier about the estate tax in the US. When the estate tax was made permanent in the 2013 legislation, the threshold at which you pay it was indexed for inflation, so now it keeps going up. Nobody has ever managed to index the minimum wage, and it has, that's why it hasn't been put up for years. There's very powerful resistance, and I think that reflects the relative weakness of organized labor in getting those kinds of things entrenched, if you like. So, even if we get, we have a five year campaign to get the minimum wage to $15 an hour, it probably will have been eroded by inflation by then and a new campaign will have to be, have to be instituted to get it raised again. The idea of in 1988, Bill Bradley, when running for the Democratic nomination, ran on the platform of indexing the minimum wage to the median income, got nowhere politically. A very difficult sell. Now, you could say, "Well, okay, "so if there's a race to the bottom, "what we need to do is raise the bottom," right? And so can we think about global minimum wage? And there have been people who have, and even movements, that have committed themselves to this. A man called Sullivan, who was famous for the Sullivan Principles applied to American, he was a pastor and the Sullivan Principles were applied to American corporations doing business in apartheid South Africa. And the Sullivan Principles were, they tried to force all the multinationals operating in South Africa to say they wouldn't do business in South Africa except on these terms, which defied some of the worst features of apartheid. They were quite successful. He subsequently got behind various other campaigns to drive up wages in very poor countries. And, of course, there's some precedent for this. If you look at the ILO, the International Labor Organization, they have created international labor standards, and, for example, in the 1984 Trade Act, the federal government said they will not trade with countries that don't meet certain labor standards. In the... When the Bush admin, the Bush Senior administration negotiated the NAFTA agreement and then Clinton came in, it's true that he signed the legislation, but not before so-called labor side agreements had been added to the, to NAFTA which said that countries had to meet certain minimum wage and labor standards if they were going to trade, sell their products in the US. And so there are these efforts, and they often involve sort of publicity campaigns. There was huge publicity attacks on Nike for their labor practices in sweatshops in Malaysia and elsewhere that actually had a pretty good effect on Nike jacking up their wages, and Nike became an advocate of pushing for increases in global minimum wages. And I, in an earlier book, spent some time exploring the prospects for this, but they're very, very difficult to achieve politically. And even in poor countries, they can have deleterious employment effects if the country's sufficiently poor. So, it's not that it's a, it's not that it's a terrible idea, but if you start to think about building and sustaining coalitions that gonna be able to do this in an era of highly mobile capital, again, it's not an idea that is likely to have much effect. And certainly, if you look at the amount by which minimum wages have been increased as a result of these campaigns in countries in East Asia and Southeast Asia, they're minuscule from a perspective of having any effect on American wages. It's very unlikely that they'll be able to do that. So, this is the difficulty with minimum wages. It's an idea that is obviously appealing. It's very, it's an easy soundbite. You see all these Democratic candidates running on it. But in terms of having an actual impact, it seems like it's not gonna do very much, and it's certainly not gonna affect the stability of employment. So, let's talk about the Earned Income Tax Credit. This is a much less well known about phenomenon and although ideas of this kind have been around for a long time. Indeed, in Britain in the 19th century, there was a poor law system called the Speenhamland system, which was essentially a subsidy of very low wages to keep people out of poverty. So, it's a subsidy to business, but it's paid to the worker, okay? That's how the Earned Income Tax Credit works. And, interestingly, it was signed into law by President Ford in 1975, although it had begun its life in the Nixon administration and he had been overtaken by events, but... (students laughing) The program survived, nonetheless, and it started out very small. So, the first Earned Income Tax Credit worked like this. You got a 10% credit for the first $4,000 of your earned income, and it then was phased out after $8,000. So, the idea was that if your credit exceeded the amount of tax you owed, the government would write you a check for the difference, right? So, this is the idea of the Earned Income Tax Credit. That's, economists say it is refundable. That's what a refundable credit means, that if it falls below zero, you don't stop getting it. The government actually writes you a check for the difference. And so it took a lot of people off the tax rolls entirely, and then people would start to get a check. Now, you might say, "Well, the numbers were so small." Furthermore, originally it was for single mothers with children only, but here we should call back the language approximate goals that I talked about last time. A very important book called "The Hidden Welfare State" by a man called Christopher Howard was published in 1985, in which he talked about the world of tax expenditures. This is a classic expenditure. It's not seen as a transfer payment in the budget. It's called a tax expenditure. It's an expenditure through the tax system. Even though the great majority of it is the government actually writing checks to people, it's not tax refunds because of the refundable character of it. So, "The Hidden Welfare State" is the world of tax expenditures, and his book gives chapter and verse of the little-known truth that the hidden welfare state is actually as big as the welfare state, as the visible welfare state. Because what tends to happen with tax expenditures is that they start out small and then they grow over time, and so it's a kind of get, don't worry that it's small to begin with. Get your foot in the door in the dark, as it were, and then push to open the door further later. And an interesting feature is that the, the EITC has had strong bipartisan support from Republican and Democratic administrations and has been expanded numerous times. By 2017, it was costing the federal government more than 70 billion in wage subsidies to more than 25 million families, and it doesn't now phase out completely until $50,000, which is above the median income. So, it's 25 million families. You're talking about 100 million people probably, close to it, are being supported, in some measure, by the Earned Income Tax Credit. They're either getting a tax deduction or they're actually getting a check from the federal government. Now, this is... This is not chump change, right? Another feature of it that is very appealing is that it's very cheap to administer. You don't need a new, you don't need a new bureaucracy to administer it. It's administered, basically, by the IRS, right? You don't need a whole department of the Earned Income Tax Credit. So, that's an additional feature of it. And also, it has a very interesting set of incentives attached to it, which is, unlike a minimum wage which creates the incentive of race to the bottom, this actually creates an incentive of the race to the top. Because if you have state EITCs, as many states do, what's gonna happen is if you, if a particular state offers a better, a more generous EITC than a different state, all else equal, the company will move to the state that has the better EITC benefit because it's a wage subsidy. It's a wage subsidy but it's, so it's making things cheaper for employers and so it can be a magnet for firms rather than a disincentive for firms to move into your state. So, this is a hugely appealing feature of it from the standpoint of avoiding the race to the bottom and, on the contrary, creating an opposite incentive. Another desirable feature of it is that while it is a subsidy to the firm, it's paid directly to the worker. And if you think about that rather than, say if you look at the infrastructure bill that the Congress adopted last year with tax breaks for firms that invest in infrastructure, one of the stock critiques is it's very difficult to make sure they actually make those investments in the infrastructure. They do the planning in R&D and the planning, and there's money paid to consultants, and maybe they do it or maybe they don't. Or if you think about place-based incentives to firms to move to different areas, there's lots of criticism of them by economists that they, that they lead to a lot of clientelism and corruption and so on. When your payment is going straight to the worker, there's no, there's a much smaller possibility of that. Now, economists will still object that this is distorting labor markets, and some people think it's morally obnoxious because it's giving firms incentives to pay lower wages, and it does give firms incentives to pay lower wages. Nonetheless, you can see the potential here for a coalition. And, indeed, business has tended to like the Earned Income Tax Credit. I think that it's one reason why Republicans have tended to get behind it. And certainly in the last decade, the increases in the EITC are by far the biggest increase in downward redistribution that we have seen at the hands of the federal government. This is, I think, genuinely is an idea that has time has come, and you should think about the likely expansion of the EICT, EITC as a way of generating employment and increasing wages that is going to have more rather than less of a future. Of course, it is, the other reason Republicans are willing to get onboard with it is that it's linked to work. There's no, this is unlike the, unlike unconditional basic income. You only get it if you actually work. Nonetheless, employers can keep wages low, knowing that they're gonna get the subsidy for providing the employment. So, I wanted to spend our final few minutes talking about unemployment and adjustment assistance. Again, I think an extremely important set of policies in view of the, the insecurity of insistence. Now, unemployment insurance in America is a very, has a very depressing history. If you go back to the Social Security Act of 1935, there was a lot of talk at that time of creating a national unemployment insurance system based on a tax you pay when employed, but there was a strong body of opinion that said the Supreme Court would strike it down as unconstitutional for technical reasons that needn't concern us. So, instead what we got was unemployment legis... The unemployment system was created as part of the Social Security Act, but essentially it's a kind of unfunded mandate on the states, and all the federal government supplied was some guidelines and some administrative support, and in particularly extreme circumstances such as the Depression or after the financial crisis, the government, the federal government will make lump sum contributions to sort of prop up unemployment insurance in the states. But basically, it's a state-based system. And that immediately has one very bad feature, which is state legislators do not want to have very generous unemployment insurance because they worry about the race to them. They don't wanna be paying it and so there's a lot of resistance at the state level. The benefits that they create, state unemploy, they're often very short-term, they replace about 40% of prior earnings, and they mostly don't keep people out of poverty. There's particularly low coverage for low income workers, so it's not a good country in which to be unemployed. And, indeed, the short-term character of the support, the hurdles to getting it, and the prevalence of longer-term unemployment I was telling about in an earlier lecture may mean that in 2016, only about a quarter of the unemployed workers in America were receiving benefits, down from over a third in 2007, so, you see there the effects of the long-term unemployment and the difficulties of getting it. And also another thing that's happening is that more and more long-term unemployment, employed workers are just choosing to go onto Social Security disability benefits. Applications for disability insurance have skyrocketed, so they're leaving the workforce permanently. Again, it's a very costly solution to the problem. And because the wages that are taxed to fund unemployment benefits, there's a cap, it's also a very regressively financed system. So, the unemployment insurance in America is never been robust, it's difficult to get, and it's failing badly I think is about the best that you can say of it. But there's a more interesting set of programs that I wanna end by spending a little bit of, a few minutes talking about, and this, this is the idea that was known as trade adjustment insurance, trade adjustment assistance, and it was a missed opportunity politically for the, because of the way in which it came to be enacted and then implemented. So, it was adopted in the Kennedy Administration in 1962. This was an era of trade deregulation, not unlike the era in the post-Cold War, and the Kennedy administration really wanted to get trade agreements with other countries to improve trade. And the unions were a lot stronger then than they are now, and the way the administration got the unions to support it was to get behind this idea of trade adjustment assistance. And it was gonna be... It was going to be robust unemployment insurance coupled with funds to support relocation and free retraining. And so the AF-of-L-CIO supported it. They supported the Free Trade Act on the grounds that this would be the quid for the quo, if you like, to use the language of the week. That we would start to get this adjustment assistance paid to workers. They would get better unemployment, longer-term unemployment support, and they would get support to relocate, and they would get free retraining to re-enter the workforce in, in another occupation. It was a huge administrative catastrophe. President Kennedy was assassinated, and when the Johnson administration came in, they had different priorities, and they were doing civil rights, they were doing voting rights, they were preoccupied with the Vietnam War, so this was not a high priority of the administration, and they didn't set up an effective implementing strategy for this program. And so throughout the 1960s, actually, not a single claim for trade adjustment assistance was actually funded. And business had been very skeptical of this program and had not really supported it, but as the 1960s wore on and the desire for free trade agreements in many sectors got stronger, business decided that they liked this idea. But by then, the AF-of-L-CIO had concluded they had been so badly burned that George Meany, who was it's head and who had actually championed the idea, he rechristened it burial insurance, and the unions have since opposed all free trade measures. And so it's really a lost opportunity because it was the one time in American history that American unions had actually been in favor of free trade. A real lost opportunity. And this is an idea that has really taken hold in some other countries. The gold standard here is Denmark, which supplies 60% of your lost wages for three years plus huge amounts of retraining for workers to re-enter the workforce. Very tailored to individuals. They spend a lot of resources on it. This to is an area where we argue in "The Wolf at the Door" is an idea whose time has come, but it has to be rethought in the following, I'll just finish with this, has to be rethought in the following way. One of the problems with it was that the unions had a lot of trouble supporting it, not only because they had thought they had been burned but they said, "Why should we only help people "who've lost their jobs to trade?" Indeed, one of the reasons it was so hard to get it was you had to prove you'd lost your job due to trade rather than for some other reason. They said, "That's unfair "because what about industries where people lose jobs "that are not due to trade?" And, indeed, in the Reagan administration when they decided to essentially underfund the program so much as to effectively kill it, Reagan made exactly the same argument. He said, "It's arbitrary just to help people "who've lost jobs to trade." Well, in our era, where jobs are probably increasingly going to technology, as we have heard about, it should be re-conceived as universal adjustment assistance, and I think it would be much easier to build and sustain the coalition to protect it. So, what business should fear if they don't get behind these kind of proposals is that Andrew Yang will turn out to have been right. (students laughing) And just, we'll let... I'll leave you with this epitaph on the 2016 election, and I won't tell you who said it until after I've read it to you. Trump's election wasn't about Trump. It was a throbbing middle finger in the face of America's ruling class. It was a gesture of contempt, a howl of rage, the end result of decades of selfish and unwise decisions made by selfish and unwise leaders. Happy countries don't elect Donald Trump president. Desperate ones do. In retrospect, the lesson seems obvious. Ignore voters for long enough and you'll get Donald Trump. Yet the people at whom the message was aims never received it. Instead of pausing, listening, thinking, and changing, America's ruling class withdrew into a defensive crouch, beginning on election night. Well, that's enough of it. We're out of time but it's Tucker Carlson. You might be surprised to know. All right, we will see you after next week. (bright music) |
Power_and_Politics_in_Todays_World | Lecture_26_Agendas_for_Democratic_Reform.txt | - Hello and good morning everybody. - [Students] Good morning. - Today our agenda is going to be threefold. I'm going to review some of the central themes of the course to begin with, then I'm gonna talk about four paths not taken as a way of, rather than giving a mechanical summary of the previous 25 lectures, to pick out the four central areas where, as I said early on in the course, we're gonna be paying attention to paths that might've been taken that weren't taken when they were a scope for choice, and then finally we're gonna end with agendas for democratic reform, the sorts of political reforms that would make it more likely that the kind of policies I have been discussing in the last two lectures could actually be successfully pursued, but let's go all the way back first of all to September. There were really two central motivating questions of the course. The first is, how did we get from the euphoria and optimism that we say at the end of the Cold War to the politics of fear and resentment that has swept so much of the democratic world since 2016? And then, secondly, what are the prospects for a better politics going forward? So if we reflect on those two questions from the standpoint of what we have been doing over the past three and a half months, I think there are two related answers to the first question. The first one is, especially in the period of what I was calling the early post-Cold War era that is up to the financial crisis, geopolitical arrogance and uncritical embrace of neoliberalism at home and the Washington consensus abroad. If you wanted the bumper sticker for the geopolitical arrogance, there are plenty of candidates, but I think this anonymous, what turned out to be Karl Rove quotation that I showed you earlier is probably the winning candidate when he said, "We're an empire now, and when we act, "we create our own reality. "And while you're studying that reality, "judiciously, as you will, "we'll act again, creating other new realities, "which you can study too, "and that's how things will sort out. "We are history's actors and you, all of you, "will be left just to study what we do," and this was in October, 2004, in the run-up to the election in which George W. Bush would be reelected and unusually with increased majorities on Capitol Hill. Usually the president's party loses seats in his reelection bid, but he actually increased his majority in 2002 as well. So there was some reason for that arrogance you might think and it certainly informed their policies in the international arena. If you wanted the winning bumper sticker candidate for the uncritical acceptance of neoliberalism at home and the Washington consensus abroad, it has to be this statement by Alan Greenspan. - Have looked to the self interest of lending institutions to protect shareholders' equity, myself especially, are in a state of shocked disbelief. - As you will recall, we played the longer excerpt of that interview in which he admitted that his models had broken down, and I think that that again captures the uncritical acceptance of the neoliberalism at home, Washington consensus abroad. If I talk about the second big answer to the first question, it has to be the failure to take account of the long-term changes in the national and global political economies. Just to remind you again here of some of the key points about this, this is a slide I showed you of the mirror image of the graphs depicting the decline of industrial jobs over, this is 26 OEDC countries between 1960 and 2018 and the fragmentation of political parties, the weakening of political parties, and I told you a story about that. As left-of-center parties became less able to defend unionized workers, they defended them less well, and they defended fewer and fewer of them and so workers hemorrhaged to other parties as in, say in Germany to Die Linke, the far-left Marxist party and The Greens, and eventually to populist parties such as the Alternative fur Deutschland. This in turn led to fragmentation of parties on the right as they saw opportunistic possibilities of picking up seats and so we get this very substantial fragmentation of political parties, which is gonna have an impact on their capacity to protect the interests of workers over time. This reflects a much broader trend. This is the picture from the U.S. Of the decline of union membership. As industrial jobs have gone away, unions have gotten smaller and smaller, particularly private-sector unions, and we saw how that coincided with the growth of inequality, the hollowing out and stagnation of middle-class incomes, and the declining fortunes of the middle-class, both in absolute terms as they faced harder and harder difficulties in achieving the same real income, often having to go from one to two earners, being decreasingly able to afford college for their children, and often having to finance their children well into adulthood. So this is a broader context where we've seen this impact of globalization, technology, and slow growth. We have aging populations in many of these developed countries, which has put increasing stress on their fiscal states as ever shrinking employed, what we call the non-dependent part of the population is supporting an ever increasing dependent population. The collapse of the corporatist consensus between business and labor that underpin many of the social democratic welfare states in the post-World War II era which really depended on relatively large, stable, labor-oriented parties making deals with relatively large stable business-oriented parties that were then blessed through the political system and maintain the social democratic order. All of that has been weakening shrinking and so we have had this fragmentation of parties. All of this has made it much harder for governments to govern and has led voters to make demands for reform, some of which we have talked about in a previous lecture and some more of which we are gonna talk about today, and it's led to the rise of anti-system parties. One of the if you like tragedies of the politics of the last several decades is the extent to which political reforms have been fixing things that were not broken. We saw this I think particularly dramatically in the United Kingdom in the 1970s. Aside from the war decades, the 1970s were surely the most challenging economic decade that Britain experienced, although they certainly were not alone. The '70s were very difficult here because of the oil shocks from the Yom Kippur War in 1973. The advent of stagflation is hitherto an unexperienced combination of high inflation and low growth, job stagnation, high inflation, high unemployment, very difficult to manage, and as Anthony King's excellent book "The British Constitution" argues in was this period which led people to say, the political system is failing us and we need to start reforming it. Very comparable developments occurred in many of the other capitalist democracies. So when the economy is going badly and if the sources of the pain are not addressed by the political system, people all want to attack the political system and say it's failing us, even if the problems, as was the case with much of the stagflation of the 1970s, were actually exogenous. So I think these are the two big answers if you like to course: the arrogance and uncritical acceptance of the dominance of the American's model of democratic capitalism and the American approach to international geopolitics on the one hand, and the failure to appreciate and think about the implications of these long-term changes in the political economy, which had indeed been going on since well before the end of the Cold War and would play themselves out in the subsequent decades. So I think those are the two central answers that I would give us to the question that I posed at the start of the course. What I wanna do now, though, is think about what might have happened? What could've been done differently? And I wanna pick out as a way of summarizing and reviewing much of the material that we have covered four particular types of path, taken or not taken, where there was relatively unusual scope for choice, for agency, as I'm gonna call it. You might recall that earlier in the semester I said that we need to think about political outcomes by reference to three of my I words. So I have a penchant for I words. I said one was institutions, one is interest, and that one is ideals, or we might say ideas or ideologies, and the constellations of interest, ideals, and institutions shape what can be done. You might think that the capacity for agency as I'm calling it here is sometimes more constrained and sometimes less constrained by the interests, ideas, and institutions that structure a given situation. So if we think about the Cold War, this was a period in which there was very little capacity for agency because the interests on the two sides were pretty much baked in. The coalitions, at least until Nixon's opening to China started to split the Communist coalition in the Cold War, but at least until that point, the interest were really, we had these two snarling coalitions of forces, institutions were largely either baked into the conflict as with the Warsaw Pact and NATO, or irrelevant, as with the United Nations, which was largely beside the point in such a realist-dominated conflict, and then ideas and ideologies were also largely baked into the conflict. George Kennan had sort of summed up the American attitude that it wasn't worth trying to argue the merits with the Soviets because they were locked into their own ideological prism and they certainly saw the West in the same way, that it was all bourgeois can't coming out of radio through Europe or wherever it might be. So the Cold War was an era in which the capacity for agency, to actually change things, was very constrained that, if you like, that triangle of constraints was very small, but there are times and I mentioned this to you right at the beginning of the course when those constraints can loosen up, when suddenly, we know now from the divide-a-dollar story, for example, coalitions can start to break apart and unlikely coalitions might begin to form. We saw this for example in the South Africa case where the traditional coalitions completely fell apart and the ANC and the Inkatha Freedom and the national party government formed a kind of coalition against Inkatha to get the new rules adopted. We also saw in the post-Cold War era and also in the South Africa case that ideologies and ideas can suddenly become freed up. So the hard right wing Afrikaners ended up making an alliance with Buthelezi's Inkatha Freedom Party and would be photographed standing together in ways that would've been unthinkable in terms of the Afrikaner ideologies of even a few years prior, and institutions can change. We will see how the institutions did indeed change once the Cold War was over. We saw that there was a scope both for the reorientation of existing institutions, most notably, NATO, and also the creation of new institutions such as the system of rules enacted for the responsibility to protect in a post-Cold War era. So there are circumstances in which this triangle becomes larger, and suddenly it's possible to do things differently and there are real choices for the players at the time, and of course that window, something of a mixed metaphor here of triangles and windows, but that window of opportunity doesn't stay open indefinitely. It will close again as things solidify and it becomes harder to make changes that might've been easier at a different time. We've seen this somewhat, but classic example is if there's a lot of dissatisfaction and a country changes its electoral system that then new politicians get elected under the new rules, so they are gonna be very disinclined to change it again in the future. So as I said the 1970s create a lot of dissatisfaction with existing electoral institutions in many parts of the world, changes were made, but then it becomes harder to change them again. So let's start with the first of these paths not taken and I think that the most important one here, certainly the first one I talked about during the Cold War was the decision to expand NATO after the Cold War to consume what would eventually be all of the Warsaw Pact countries and even make overtures to Georgia and Ukraine. It's worth going back to thinking about this. How this started, this first came onto the agenda when the Berlin wall came down and it was obvious that Germany was gonna be reunited and of course Gorbachev was very worried that this would mean East Germany would become a member of NATO, and East Germany had been a member of the Warsaw Pact. Very substantial assurances, there was gossip about it at the time, but we now can see from declassified documents that it wasn't just Kohl and Baker who promised Gorbachev that there would be no expansion of NATO into the former Soviet empire, but many others, you can see here on the bottom of the slide, all gave assurances to that effect in early 1990, and this was a promise that was gone back on first in the Clinton administration. In 1998, the Clinton administration took the decision to expand NATO to include Hungary, Poland, and the Czech Republic. The decision was made in about 1996, but it actually came on board in 1998. In May of that year, the journalist, Tom Friedman, called up George Kennan, the architect of containment about whom you heard in earlier lectures, and he asked him for his opinion about this and it's worth paying some attention to what Kennan said. He said, "I think it's the beginning of a new Cold War. "I think the Russians will gradually react quite adversely "and it will affect their policies. "I think it's a tragic mistake. "There was no reason for this whatsoever. "No one was threatening anyone else. "This expansion would make the Founding Fathers," of course he's alluding to the fact here that the Founding Fathers opposed all permanent alliances, "would make the Founding Fathers "of this country turn over in their graves. "We have signed up to protect a whole series of countries, "even though we have neither the resources "nor the intention to do so in any serious way. "NATO expansion was simply a lighthearted action by a Senate "that has no real interest in foreign affairs." So this was a prescient observation that this was not going to go over well with the Russians as by this time it was of course Yeltsin in charge of Russia and he was deeply troubled by it. Indeed there's more to the prehistory of this than I told you during the original lecture in that during the late 1980s, Gorbachev had made several overtures to the West. He, three times brought up the prospect of, he thought at that time the Soviet Union was not gonna fall apart, it wasn't being thought about, but that they might themselves join NATO. There was some discussion of this with Western leaders and it was sort of deferred into the future and Yeltsin too had, in print, said that it was a long-term plan or policy of the Soviet Union eventually to join NATO. Indeed even in 2000, when Yeltsin came to power in Russia, he approached the Clinton administration about the possibility of Russia, not only of Russia not only joining NATO, but also the European Union, and they had been brushed off by this. If you're interested in more granular depth on these proposals and rejections and proposals and rejection, if you Google up the journalist, Vladimir Pozner, a very interesting, you would never believe he's actually French by birth, French journalist, but who's worked for many years in the U.S. and now works in Russia, he has an excellent video. It was actually a lecture he gave at Yale. I think he was probably a Poynter Fellow or something. If you Google up Vladimir Pozner at Yale University, you'll find this lecture on YouTube, really a riveting account of the missed opportunities to incorporate Russia into the New World order in a way that would have made them much more likely to be part of the solution than part of the problem, and of course this isn't just retrospective. In 1991, when the Soviet Union finally fell apart, Mitterrand, who was at that time the president of France, said, "We should just disband NATO." The French, as we saw playing out yesterday, the French have a long history of skepticism toward NATO. de Gaulle didn't really like it and actually he kicked NATO troops out of France on several occasions and took France out of the command-and-control systems. So it's never been a lot of love lost between the French and NATO and indeed they insisted on having their own independent nuclear deterrent, which NATO was designed to prevent the incentive to do that sort of thing. So when the current French president said NATO is brain-dead, he's walking into a well-worn tradition of France's skepticism towards NATO, but there it is, there was a path not taken. Mitterrand could have prevailed, and it does seem like there was surprisingly little discussion of the possibility of creating a different world, that instead that decision that was taken in 1996 and implemented in 1998 as I detailed to you was followed up by the gradual inclusion of all of the NATO countries and overtures to Georgia and Ukraine, which had a lot to do with the Soviet policy toward Georgia in 2008 when they went in there and their attitude toward Ukraine once the election in 2014 came around and it was clear that their base at Sevastopol was going to be jeopardized and that was a real concern if Ukraine had joined NATO and/or the European Union. So you could ask yourself sort of in the domain of counterfactual history, what kind of geopolitics might we have today if that was not the choice that had been made and instead either the incorporation of the Soviet Union, and then Russia into the new geopolitical order as a full partner and participant, or the disbanding of NATO which after all was a defensive alliance against a foe that had no longer existed and was gonna turn into an alliance in search of a mission which we saw playing out rather depressingly in the Libyan conflict about which I'll have more to say in a few minutes. So that's a clear case of a path not taken. Of course we don't know what the path would have yielded if we had gone down that, but it would've made the New World order, once America's capacity for unilateral action was revealed to be chimerical after the debacles in Afghanistan and then Iraq, it would've made it much less likely that geopolitical relations would've settled back into something not that different from what they had looked like during the Cold War. So I think Kennan, who had been prescient about so much in his life when he had advised against Vietnam War, when he had advised against militarizing NATO in the early years of the Cold War, was extremely prescient in 1998 in pointing out the folly of the Clinton administration in expanding NATO onto Russia's doorstep or beginning the expansion of NATO which would eventually extend not only to Russia's doorstep, but to at least two of the former republics of the old Soviet Union. So a second path not taken also in the realm of geopolitics I put under the heading of the management of geopolitical crises and humanitarian intervention, and here George Herbert Walker Bush really pointed the way to a new containment regime in the first crisis after the Cold War when Saddam Hussein's Iraq invaded Kuwait. The question was: how would the international community handle this? And the answer was essentially a global containment regime that the bumper sticker for containment as I have said to you in earlier lectures is stop the bully without becoming a bully yourself. So what did Bush do? First of all, he got authorization from the international institutions as they then existed, namely the Security Council. He put together a massive coalition that included very diverse players including every Arab country except Jordan. He then expelled, with that coalition, expelled Iraq from Kuwait, and against the advice of many hawks in the U.S., and particularly the neoconservatives who we talked about in an earlier lecture, he chose not to go for regime change in Iraq, but instead essentially create no-fly zones and buttoned-down the Iraqi regime and make it impossible for them to threaten their neighbors. So if that had become the model for dealing with geopolitical crises in the wake of the end of the Cold War, we would again I think be looking at a very different kind of geopolitical reality. Instead though George W. Bush abandoned this with his unilateralism after 9/11 where he talked about coalitions of the willing, forced countries to choose, if you're not with us, you're with the terrorist. By the way the big impetus for the expansion of NATO as I mentioned came about because he wanted to get many East European countries to commit to the war effort in Iraq and the quid pro quo for that was that they would be able to join NATO, which many of them did the following year. So the paths not taken here in Afghanistan: the possibility which was advocated by some in the military at the time was to ignore the regime in Kabul when they refused to turn over Al Qaeda and to close the camps and instead to have just gone in ourselves and finish the war with Al Qaeda. Instead because we were committed first of all to fighting that war on the cheap, we wanted to fight it with the Northern Alliance, and their agenda was regime change, and so by mission creed it became America's agenda, saddling us with a puppet regime that wouldn't be able to sustain itself into the future. The commanders on the ground were telling the White House before the end of 2002 that a deal would have to be made with the Taliban as the other forces were not gonna be sufficiently powerful to govern the country. Again that was ignored and instead we created a puppet government. "If you break it, you own it," as Colin Powell once said. We then became responsible for the administration of this country which at some level continues to this day. We also failed to defeat Al Qaeda which escaped through the mountains into Pakistan and indeed became the staging area not only for future reconstitution of Al Qaeda, but also for the exiled Taliban government that would eventually work its way back into Afghanistan. In Iraq, we failed to see that knocking off the regime in accordance with the desires of the neoconservatives would first of all not produce a government that was likely to be democratic, and even if democratic, would not necessarily produce a government that would be friendly to the United States. We talked at some length about the naivety of the neoconservative view that if you just get rid of authoritarianism that not only will democracy flourish, but it will be pro-American democracy. Neither of these things turns out necessarily to be true and so they had their eyes firmly fixed on the Eastern European countries, which would all have pre-World War II histories of democracy. They would've done a lot better to look at the former Asiatic Republics of the Soviet Union for what was likely to happen in a country like Iraq with its per capita income, its lack of a diversified economy or a middle class that was independent of the government's control of the gateway to wealth and prosperity. So we became embroiled instead in this ongoing conflict in Iraq. The other thing that they failed to see was that the capacity to defeat an army, we defeated the Iraqi army in four days, does not translate in the capacity to govern a country. This is the fatal error of thinking just because we are the most powerful military in the world, we can knock off governments and putting governments who will then be able to govern. There are many economies of smallness in governing. You need street-level buy-in, you need street-level commitment from various broker groups and you need to empower local institutions, which we had obliterated. We largely obliterated the army, the civil service and the police and it was not surprising that that country would be headed for decades, more than a decade of chaos. So then we can think not only of that tragedy, but the tragedy of responsibility to protect from Rwanda through Libya. Rwanda was the never again. We then had Kosovo and it seemed like a unilateral intervention. It was by NATO, again without the permission of the UN Security Council because the Russians and Chinese were supporting Serbian sides so there was no question of international authorization. There was an intervention that appeared to be successful, even though it provoked virulent attacks in the global south, from South Africa, from India, and from others as expansions of Western imperialism. It provoked the African Union to declare that it had unilateral authority to intervene in member countries on humanitarian grounds, even though the UN Charter extensively reserves that authority to the Security Council. So what we were doing there was somewhat eroding the authority of the UN and the Security Council. We saw how the UN tried to sort of reign in and domesticate this idea of responsibility to protect by authorizing the Security Council to authorize it, and then we saw the tragic way it played out in Libya when the Obama administration, which was quite divided and against the advice of his generals but in the thrall of the ideology of responsibility to protect, being pressed by some in the administration, and eventually Hillary Clinton at the behest of not very credible interests of people who were looking to ensure oil concessions from what they thought was gonna be the new government, this whole doctrine became a smokescreen for knocking off the Gadhafi regime, which then, because of what IR theorists call the moral hazard of intervention, became an accelerant of the Syrian conflict as the Syrian demonstrators against the Assad regime reasoned that if only they rose up, NATO would come and help them too, which of course they didn't, and then the Assad regime came down on them like a ton of bricks, not to mention the collapse of the Libyan state created an arms bizarre across the whole of North Africa as many people made lots of money, and it also created a foothold by the way for Al Qaeda in Libya. So terrible tragedy, and this doctrine of the responsibility to protect was essentially hijacked by American and its allies' unilateralism as expressed through this now NATO organization that had lost its raison d'etre and was becoming the self-appointed policeman of the world, which is unlikely to win it much legitimacy, except among people whose acceptance of it was not needed. So you might ask yourself again, as with the expansion of NATO, what world might we be living in had the responsibility to protect doctrine not morphed into one more piece of camouflage for the unilateralism that had become so fashionable after the collapse of the Soviet Union, and particularly after 9/11? So that's a second area where there were real choices that could've been made differently, but were not and have left the world in a much different place and the idea that responsibility to protect could be rehabilitated now is extremely difficult because the way it played out has been baked into the conflict in the Middle East as it since evolved with Russia filling the gap in the Syrian Civil War and if you like my triangular window of opportunity, to experiment with that doctrine having been dramatically shrunk. Paths taken and not taken by business. So here it's just worth pausing to ask the question: why is business so important? Why focus on business? And the answer to that question is that once communism as an economic system has been taken off the table, capitalism essentially is the only game in town. There is no serious competitor, maybe not even a frivolous competitor to capitalism. You think about Islamic fundamentalism. One thing Islamic states lack is any kind of political economy. If they have an economic theory, it comes from the 14th century. They have no conception of an alternative way of running political economy than capitalism. If you look at Islamic fundamentalist countries such as Libya or Saudi Arabia, they just do it badly. There's no alternative. Well, if there's no alternative out there, that means that capital is gonna be in a uniquely powerful position, particularly in this world in which labor is weaker and weaker and organized labor has shrunk. This is basically a Hirschman story that business can go anywhere at the click of a mouse, capital can move anywhere at the click of a mouse and labor has very high exit cost because you have to pack up and move and find another job, and move your family and so on. So the differential capacity for exit between business and labor means that business is more and more important and the idea that informed some business elites during the new deal that we need to worry that America's workers might conclude they have nothing to lose but their change ceases to be important because there isn't an alternative out there competing for the hearts and minds of American workers. So businesses is now in this unique position that it's hugely consequential. It's not really counteracted any other interest. Maybe there are conflicts of interest within business elites, but it's not really counteracted by any other interest, and that means that unless business is part of the solution, it's gonna be part of the problem. You could imagine a continuum and ask yourself, a continuum. One is where you have predatory capitalist class, as in Russia, and another would be if you have a very progressive capitalist class, as in some Nordic countries and has seem to be possibly happening in South Africa, at least during the 1980s. And then the questions become, so what makes it more likely that business will move toward the progressive end of the continuum rather than down toward the predatory end of the continuum? So that's why business matters. Indeed, as we argue in my book with Michael Graetz, "The Wolf at the Door" thinking about changes in social policy that are not gonna command support from substantial amounts of business is probably a waste of time. So if we think about business in sort of under two headings, one we talked about some with the role of business in regime change, which we talked about in relation to the Soviet Union and we talked also about it in relation to South Africa, you can say two cheers for the South African story, two, but not three. So the two cheers are devoted to the fact that in the 1980s, the South African business elites did discern the possibility of both facilitating regime change and influencing the ANC to think about how the world economy was then operating and was gonna be operating into the future. They became the kind of handmaidens of the transition in that as I said business did encourage the South African transition. The old regime was falling apart. There was no question that apartheid was unsustainable, but that didn't mean it was gonna have a transition to a democracy or that the transition was going to be peaceful. It might well have been replaced in a military coup, which people at the time, some feared it might have just become increasingly repressive and cease to be even sort of quasi-democratic as it had been. It might have descended into a civil war, which by the 1990s had already erupted in the eastern part of the country. So what business did was a huge amount of very creative and intelligent work that we talked about in that case to broker the deals and get the transition, keep it on track when it seemed to be falling apart. And when it did finally fall apart, do the back channel negotiations to produce an agreement and eventually the election in 1994. The reason I say two cheers rather than three cheers is that thereafter South African business did very little. It retreated into the sort of normal stereotypical view of business that the business of business is business and leave politics to the politicians. A few lone voices like Christo Nel who said, "No, we should be actively transforming our companies now," before we are hit over the head with what would eventually become black economic empowerment and talk of seizing land and so on. Business should continue to be ahead of that curve, so if you like shape that destiny rather than ending up being shaped by it. So we see this has played out in the social indicators I showed you of South Africa today. The morals of Israel and Palestine story. Again there've been business groups that have tried to participate, but it's very difficult if they don't have a K group. It's very difficult if they don't agree. It's very difficult if they don't show themselves to be willing to actually internalize some cost to achieve change. So the impasse group never really had an effect, but you can still see and we showed this with (mumbles) and SodaStream and some of the other business interventions that I have talked about that business, individual business leaders, if there's no capacity to do things on a regime-level scale can nonetheless be thinking about ways of changing things at the margin so that even if this is parried as legitimating the status quo, it can be legitimating a status quo that is itself evolving over time. This was Douglas Ray's idea of utility drift that I talked to you about and could put business in a position to be a constructive force when the other things line up more propitiously. We have ongoing projects looking at this in other contexts. We have a project looking at this in Columbia where various Colombian business groups, I didn't have time to share this with you during the lecture, but I have for decades actually been involved behind the scenes in negotiations between various Colombian governments and the FARC and were instrumental in the agreement that was eventually settled. Unlike South African business after the transition, there are a number of interesting Colombian business enterprises aimed at trying to incorporate the FARC into the new legitimate economy and revivify it in ways that will make backsliding less likely. By no means sure that they'll succeed, but they perceive the possibility for agency at the margin and/or trying to do that. One might think of Syria where the Civil War is going to be over in the next year. There's a huge amount of expatriate Syrian capital parked in various countries and convening groups may well have the opportunity of thinking about how it might be constructively deployed in rebuilding what's probably the most badly devastated country by Civil War, at least since Rwanda. Business and domestic politics. Here I would say that the missed opportunity with transitional adjustment assistance is symptomatic of a bigger sense that business has not seen it as their problem to worry about the fortunes of workers. As I said to you in that lecture, in 1962, the Kennedy administration, it actually lined up the AFL-CIO to support tariff reduction as a quid for the quo being transitional adjustment assistance. By the time business finally decided that was a good idea, the unions had abandoned it and were no longer supportive of it, and that was because the implementation by the Johnson administration had been so desultory, but if business had gotten behind that at the beginning we might have seen something different. Full-bore support for the anti-tax movement by American business from the beginning even if this was gonna mean destruction of many programs that are important to the most vulnerable populations. We saw this with Proposition 13 in California, and its effect on the public school system, its effects on local government services and so on. A new model of the corporation became triumphant in the late 1980s where shareholder value triumphs over everything else, trumps everything else, and the notion that business corporations should worry about anything else. If you look at the changes in business school curriculums, for example, largely went out of the window. That all adds up to the fact that there was a failure to see the costs of ignoring the long-term employment insecurity of labor. Andrew Wang's warning at the end of the video that I showed you that if these things are not addressed, worse than Trump is likely to come along. A final strategy I would just mention or set of a path not taken I would mention is the strategies of left political parties at home. - Here's the truth. No politician can reopen this factory or bring back the shipyard jobs or make your union strong again. No politician can make it be the way it used to be because we're living in a new world now, a world without economic borders. A guy can push a button in New York and move a billion dollars to Tokyo in a blink of an eye and in that world muscle jobs go where muscle labor is cheap and that is not here. So if you wanna compete, you're gonna have to exercise a different set of muscles, the one between your ears. - So that was from the movie, "Primary Colors," and I showed you a more extended version of that clip, but there it is in 1992, Bill Clinton nailing it, exactly what the problem was going to be, that the long-term employment security that had been taken for granted by generations of American workers had gone away and was not coming back, but what did the Clinton administration do about that? Nothing. Instead, what did they do? They triangulated to neoliberal policies. They engaged in the expansion of NATO. They got behind things like welfare reform essentially to make it more punitive and they cozy-ed up to Wall Street by getting behind the kind of financial deregulation that Wall Street was demanding. So the fact that even though the candidate Clinton had identified the problem, he didn't really take seriously the effort to do anything about it. Indeed in the class on the subprime mortgage crisis he became one of the champions of getting mortgages to people who are gonna be unlikely to be able to service them as soon as the recession came. So very much he identified the problem and the different do anything of a serious time to address it and he squandered whatever political capital he had in the first years of his administration when he controlled all the houses and so on. And by the way, it wasn't just here. We saw Tony Blair's government in the UK did many comparable things. Even in Germany, I believe I mentioned to you, if you look at the Hartz reforms which were neoliberal reforms, not that different than those pursued by the Blair government, or by the Clinton administration, or indeed in countries even like France where Mitterrand did an about turn, who were the Hartz reforms implemented by in the early years of this century? It was a coalition of the Social Democrats and the Greens. It was not a right-wing imposition. So all of these parties were essentially engaging in triangulation instead of addressing the long-term employment insecurity that Bill Clinton did indeed identify in the 1992 New Hampshire primary. The difficulty with triangulation as I've said is that it's good tactics, but bad strategy. It's good tactics because if you just think about it as a one-shot game as game theorists say, you can move to the middle, peel off some support from your opponent and the people on your flank have no place to go. But if you think about it as a dynamic game over time, once the other side figures out that that's what you're doing, their incentive is to just keep moving the goal post and dragging you along as they go. This was the story of the anti-tax movement that culminated in the 2001 bill that I described at some length in our class about the estate tax. Indeed the sort of tragic denouement of that story was when in 2012 the Obama administration made virtually all of the 2001 tax cuts, biggest regressive change in the tax system in American history permanent. So this was I think a big missed opportunity reinforced by this sort of "living right and thinking left," to quote the phrase from the book that I mentioned to you when I was giving that lecture on things like affirmative action and managing the financial crisis. Affirmative action is about promotions in the police department and the fire department. That's where the rubber meets the road. Doesn't really affect people living in Scarsdale. In the financial crisis, unlike what FDR had done often the depression, I showed you how the Obama administration essentially got on board with the Bush administration's policies and allowed the financial sector to write the vast majority of the bill and not to pay any cost for the crisis that they had created thereby fueling what would eventually erupt as the kind of enraged politics of anti-elitism in 2016. Given that they were converging on questions of economic management, instead what left-wing parties tended to do or left-of-center parties tended to do was to begin investing in culture and identity issues as ways of differentiating themselves from the Republicans. This goes back to the McGovern-Fraser reforms which start to emphasize the importance of descriptive representation in politics which I'll have more to say shortly. So what is to be done? So the bottom line is, and this is picking up on the lecture about parties, that the most important thing that needs to happen is to strengthen political parties so that they can operate more like teams and thereby get themselves behind the sorts of policies that are gonna lead large numbers of insecure former or post industrial workers and middle-class people to support them. In all systems, all political systems, that basically means a number of things. One is disempower party members and outside groups in the selection of candidates, leaders, and platform. Members sound like shareholders, but as we saw, they tend to be unrepresentative of the voters for a party, particularly when you have a large party, never mind the voters in the electorate. Get away from referendums and other forms of unbundling. Parties are there to bundle. We had a long lecture on bundling, why bundling is important, because it forces us to discount what we want by everything else that we want and come up with policies that can appeal to a broad swath of voters. Empower backbenchers to pick frontbenchers so the parties can operate as teams in the parliament and empower frontbenchers to pick backbenchers for the same reason so that the people who get picked can both win electorally, but will also support the program that the leadership desires. So you give a lot of delegation to the leadership, but they also have enough rope to hang themselves. If they can't deliver successes, they won't be there for long. In multiparty systems, this means increased thresholds to try and get parties to combine so you combat the fragmentation of parties that creates huge opportunities for rent extraction when people are needed to join coalitions. Push for reelection coalitions so at least voters have a better idea of what governments are likely to form after the election. Prefer closed PR to open list PR because open list PR encourages celebrities to run. It encourages people, intraparty competition to make private promises, and that's exactly inimical to the idea of parties operating as teams. Turning to two-party systems, in Britain, change the leadership selection rules back to something more like they used to have when the backbenchers picked the frontbenchers and got rid of them if the frontbenchers were not delivering. Get rid of the fixed parliaments act, which again is a way of disempowering the authority of leaders. Bigger constituencies. They a couple years ago had a commission recommending that they reduce the House of Commons from 650 to 600, a very modest reduction. They've even rejected that, but one of the problems in Britain is that all the wealth is in London and many parts of the north of England have per capita income below Mississippi. You really want a sliver of London in every constituency in Britain if you want a healthier system and certainly the constituencies which are a tenth of the size of ours and a third of the size of Germany are too small. Stop having referendums. In the United States, first of all don't think abolishing the electoral college is a solution to anything. Everybody says, "It's so unfair. "Hillary won three million more votes." Abolishing the electoral college, it will increase the independent authority of the president. It will make us function more like a Latin American country. That's exactly what we shouldn't be doing. We should strengthening the power of the parties in Congress at the expense of the presidency, going back to something more like what prevailed until 1824 when the congressional parties pick the presidential candidates and made the system operate more like a parliamentary system. Deemphasize the importance of primaries, both Congressional and Presidential. Presidential primaries have really only been important since the McGovern-Fraser reforms of the 1971. What people don't know about primaries is the very low turnout. So Donald Trump, for example, was selected as the Republican candidate by less than 5% of the American electorate, but then he's the candidate. If there's a weak candidate on the other side, he might end up, as he did, being the president. So one proposal would be, since most American voters don't realize that primaries are so unrepresentative and would be a proposal whereby if the turnout in a primary falls below some threshold, such as 75% of the general election turnout in the previous election, then you discount the primary and let the party choose the candidate. Similarly, if you think about legislative elections, primaries have been around since progressives. This is an example of a reform dreamed up on a starry-eyed left that has since been hijacked by the steely-eyed well-endowed forces in other parts of the electorate as well. You could have a similar kind of rule. What we have is a world in which, I should say the primaries have been around for 120 years, but what's changed is the number of safe seats. We have many more safe seats than we used to have which in unsafe seats the only election that matters is the primary election. Again so you can have AOC elected in the 14th district in New York on a 11% turnout in a primary or Jim Jordan elected in Ohio's 4th district on a 15% turnout in the primary. Again you could have a rule whereby if the turnout in the primary fell below some number or some threshold that the party leadership would have a say. No jungle primaries. This is a new flavor du jour they put in California where everybody runs in the primary. The thought was it would produce more moderate candidates. In fact it doesn't because the, even in jungle primaries, it tends to be the activists who turn out in the primaries. For the Senate, get behind plans to admit DC and Puerto Rico and proposals to split up states like California and Texas so that the massive malapportionment in the center could be redressed somewhat at the margin. I'm talking here about things that could be done without constitutional amendments or without a new constitutional convention. House, in the House, we should be redistricting for interparty competitiveness by reducing the number of safe seats and the way to do this is to change the redistricting. The redistricting historically has been done by state legislatures, and of course whoever controls the state legislature does it to maximize the benefit for their party. Earlier this year, the Supreme Court has decided it will never get involved in partisan gerrymandering. It's too much of a political issue and they don't know what to do about it. So what happens is the parties draw the districts to benefit themselves. So I think there now 21 states have gone to so-called independent commissions. This could be a good idea. In Britain, for example, they have an independent commission after the census that draws up boundaries, boundary commission they call it, and then parliament can vote it up or down, but they can't change it. So that's one way of doing it. The problem is, so we have now, I think we have 14 in operation since the 2010 census and another seven which will be operating after the 2020 census, but if you look at them, they're either so-called bipartisan or nonpartisan. The bipartisan ones are hopeless in that the parties carve out the states. They essentially make deals. The nonpartisan ones, the difficulty with them is that they're not given the mandate to make districts competitive across the parties. Indeed they are told to do things like not split up cities, keep neighborhoods together and so on. Really what you need and what would make for the healthiest interparty competition would be every district in America having some urban voters, some suburban voters and some rural voters so that the politician who is running for office in those districts would discount the preferences of all three groups against one another in coming up with their platform. So for flower-petal districts, you might think to achieve that goal, and then we should think about better approaches to descriptive representation because the trouble with majority-minority districts is you achieve, they did indeed increase the number of African-Americans sent to Congress, particularly from the South where none had been sent since Jim Crow in many cases, but it comes at the price of exactly what we don't want. We do not want blue cities in red states. We do not want all the Democrats, African-Americans, to be herded into one district, and then have Republican nonminority districts electing separate representatives. That is exactly what we don't want. There are better ways to get descriptive representation. In some Nordic countries, for example, with a PR system they just say, well, every third person on the list must be a woman, or in India in 1932 there was the Poona Pact which essentially it said in a number of constituencies, believe it was 143 constituencies, the only person who could run what was in those days called an untouchable, now is referred to as a member of an unscheduled caste. So you can get descriptive representation without screwing up the boundaries of the constituencies. So these would be better ways to get descriptive representation. Majority-minority districts, which now account for a fifth of Congress' districts, are not. So the upshot is that stronger and more disciplined parties will have better incentives and will be better able to address the long term employment insecurity that we've been talking about and head off the dangerously destructive politics that we've seen since 2016, or to put it differently, what we should be trying to do is to marry the arguments of these two books. So I thank you for indulging me over these previous 26 lectures and I hope I'll see some of you in the future. (students clapping) (instrumental music) |
Power_and_Politics_in_Todays_World | Lecture_18_Political_Limits_of_Business_The_IsraelPalestine_Case.txt | - Okay so today we're gonna talk about the Israel-Palestine case that we made at the School of Management. And this is in many ways complementary to the South African study that we talked about earlier in the semester. - Distinguished participants, I would like to welcome very cordially President Shimon Peres, President Mahmoud Abbas and Secretary John Kerry. (audience applauds) We want to share first with you, with all of you in this hall and beyond a call for action to break the impasse. This call for action from all 300 key business and societal leaders, from both Israel and Palestine representing large parts of the economy is a powerful, and represents a powerful mainstream constituency to accelerate and support the political process. - When I look back, I see that most of my life, 45 years was spent witnessing the conflict between us and our neighbors. It is too much, too long, and too painful. There is almost not a single Israeli or Palestinian family that didn't pay a heavy and bitter price. Too many mothers in both sides, shed too many tears. Enough! During the entire last year, a group of civic society Palestinian and Israelis worked closely together trying to find a way in how they can help accelerate the talks towards creating two states for two people living side-by-side in dignity, peace, security, justice, and prosperity. In the beginning, we were few, but very quickly many other joined. The group we have brought together today in front of you consists of some of the leading business people, top executives, academicians, security and negotiation experts, and others. Partners who have come from multitude of opinions are from the left and the right, secular and religious, with different views and persuasions, however, we are all united by our strong desire to see an end to the conflict and move towards a two-state solution. This will be a long road. We have to build trust among ourselves. We have to recognize that each side has its own narratives as to the reasons and the backgrounds of the conflict. We knew that we are unable to breach these narratives. So we decided to concentrate on focusing on moving to a better future. We believe that this can be done. We are not here to submit a peace plan, this is the role of the leaders, and only they can do the painful compromises which will be needed in order to move the process. But we are here to encourage them. We think that the civic society has a role to play. Our voice should be heard, loud and clear. - Members of the business community of Palestine, we are issuing an urgent call to action, to all leaders to break the impasse and put as their top priority the serious and expedited pursuit of peace process. This process must end with the establishment in the short term, of an independent Palestinian state standing tall alongside the state of Israel. As business leaders, responsible for over a substantial amount of the GDP of both Palestine and Israel, we have decided to take collective stand and upset the status quo. Enough is enough. Break the impasse. We shall support you and contribute in any way we can. We will not entertain joint economic or commercial schemes that would legitimize the status quo, or illegal occupation. The time for choosing which side of history we wish to be on is now. - I want to thank those who took part, and are taking part, and will continue to take part, in the Breaking the Impasse. My good friends Munib Masri, whom I have known and worked with, and been to some of those private and quiet meetings with him in various places, and Yossi Vardi, thank you both of you for stepping up and being courageous. And even as they found plenty to disagree on, I understand they did in the course of their discussions, even as they fully understand the difficult history that is embedded in this conflict, they refuse to underestimate the potential for the future. And that's because Breaking the Impasse's guiding principle is to respect the freedom and the dignity of all peoples. - So that was the official public launching of Breaking the Impasse, this group of some 300, eventually grew to closer to 500 Israeli and Palestinian businesses that were committed to trying to move the peace process forward. And in some respects it was, and certainly in the minds of some its founders it was loosely modeled on the Consultative Business Movement that had been so consequential in facilitating the South African transition starting in 1988, that we talked about earlier. The first person you heard there was Klaus Schwab, the founder and President of the World Economic Forum. The second was Yossi Vardi, Israeli's wealthiest tech billionaire. The third was Munib al-Masri, the wealthiest Palestinian, business leader, also a billionaire. And then finally of course John Kerry. Breaking the Impasse was wheeled out with great fanfare and it was also a spectacular failure, for reasons that will concern us. So lemme just spell out what today's agenda is going to be. First I wanna introduce Nick Strong over here. Nick is a graduate of Yale College, major in Ethics Policy Economics, graduate of Yale College class of 2018. And he and I teamed up on this case. He took a course from me as a freshman and we got to know each other, one another early. And he, the truth be told, is the real creative energy behind the case that we wrote with the SOM case writers that we've made available for you today. And it explores a number of projects that have been pursued by business in the wake of the failure of BTI. And we're gonna focus on three of them. I would urge you to look at the ones we're not gonna focus on just for want of time. One is the Palestinian internship program, created by Yadin Kaufmann, through his company Sadara. And the other is a man Shlomi Fogel, who was the energy behind the Jordan Gateway Project, an SEZ of the sort Gustino was talking about but that goes across the Jordan River, and is one of the few Israeli businessmen doing business, collaborative business with Palestinians in Gaza today. All five of the projects, those two and the three we are gonna talk about, have the feature that they get attacked, both by the BDS movement, BDS standing for Boycott, Divestment, and Sanctions. This is largely, initially Palestinian-founded movement, officially launched in 2007 at a conference in Ramallah, but subsequently, I'm sorry, has antecedents going really back to about 2001, they all get attacked by BDS, and they also get attacked by the Israeli right, on the same grounds. They get attacked for normalizing the status quo. And neither the BDS people, nor the Israeli right wanna do that. We'll talk about that, and indeed we'll also talk about what normalizing might actually mean. But we'll do after we have worked our way through the cases that Nick is gonna end up describing to you shortly, Nick welcome. (audience applauds) So the agenda, the rest of the agenda. We're gonna be talking about BTI and its failure. Then the subsequent business efforts and that's when Nick will take over. And then what are the lessons learned. And what should we think, both about the possibilities and limits for business involved in high-conflict political possible transitions and other more general lessons that might be drawn. So what happened to BTI? There was as you saw great fanfare there on the Dead Sea, when it was launched in the summer of 2013. It had actually been conceived the year before in Davos. And then launched the following year. It was quickly walloped the following summer, because that was the summer that Israel went into Gaza in 2014, in one of the most highly conflictual military confrontations between Israel and the Palestinians in recent times. And that was then rapidly followed by the March 2015 election in which Netanyahu did much better than anticipated, actually against a combined coalition of the left of center parties, which didn't manage to stop him. And he formed a very right-wing government, which was a harbinger of things to come. And so most of the people who had been involved in the BTI movement got pretty demoralized and it didn't cease to exist, indeed it exists to this day. I went to one of their meetings earlier this year in Jordan. But for all intents and purposes it became moribund. Then we're gonna talk about the subsequent evolution of business efforts in the wake of BTI's failure. Which is what Nick and I worked on and what he is gonna discuss. But first let me just draw out some contrasts between BTI and our earlier discussion of the Consultative Business Movement in South Africa. And I should say right off the bat that BTI's emergence surprised me. And the reason it surprised was the following, in the wake of learning how big a role South African business had played in the 1980s, in facilitating conversations between the government and the ANC, I had gotten interested in what business might be doing in other high-conflicts settings, in Northern Ireland, in the Middle East, and elsewhere. And so I actually went in 2006 and 2007, I went to Israel and the West Bank, and I interviewed a lot of Israeli business leaders. And talked to them about possible involvement in the political process. And I interviewed some of the leaders of Israel's biggest companies and there wasn't a whiff of interest. They said what you would expect them to say, they said, our business is business not politics. The politicians wouldn't listen to us anyway. And indeed when we lobby the government it's usually about doing cross-border deals in Jordan and Egypt, you know, we have nothing to do with the peace process. And I would say things like, well imagine if there was peace? You would have cheap Palestinian labor. You would have, you know Gaza is on the Mediterranean. Think about the possibilities if there were a real settlement. And to a person, no interest. We have cheap Russian labor, was one answer I got from a couple of people. So I concluded that for whatever reason there was no Israeli business interest involving themselves in the political process. So I was very surprised when in 2012 and 2013 I started learning about BTI. So business had not been interested. As you could see from that video, BTI, one difference between it and the Consultative Business Movement, it was very much dreamed up outside of the conflict by people in Davos and it was very much a top down movement. Whereas the Consultative Business Movement, as I said to you when I gave you my lecture about it, really bubbled up and indeed was only, the CBM was only created when ANC leaders got fed up with having informal conversations in people's living rooms and said we want an organization to talk to. So there was a much, even though they were leaders of large businesses, it was much more bottom up and generated from inside the conflict than the BTI movement ever was. Another important difference is, as you heard from Yossi Vardi, and Munib al-Masri, to a lesser extent, they both said we have all political persuasions here from the left to the right, and we are not, we're not taking a position, we're gonna be politically neutral, it's up to the politicians to hammer it out. Well so one thing to say is the position of BTI was never neutral, because the were committed to a two-state solution. There are people on the left and on the right, who have one-state solutions in mind, very different one-state solutions in mind, but nonetheless, one-state solutions in mind. So while talking about a two-state solution as the politically-neutral answer, in fact they were not politically neutral and were committed to a two-state solution. And one thing to say is, if you spend, and Nick will have more to say about this, but if you spend any time in the West Bank, it doesn't take you long to realize that the probability that Israel is going to leave the West Bank is somewhere in the range of the probability that the federal government is gonna agree to give North Dakota back to the Sioux Indians. It is so far beyond the realm of what is feasible or likely in the next several decades, that for everybody to keep saying, repeating this mantra, that we're all committed to this two-state solution, tells you that a lot of people are pretty disconnected from reality. Another difference with the Consultative Business Movement was the resolute refusal of anyone involved in BTI to talk to Hamas. This would have been the equivalent of the Consultative Business Movement saying we're not talking to the ANC, we're only gonna talk to Inkatha. And indeed in December of 2013, I went to a meeting of the CBM, of the BTI in Washington, D.C. and I interviewed both Vardi and Munib al-Masri. And we then had this big forum, there were 200 people there, and they made speeches, similar to the speeches you saw there. And then there were questions. And I asked a question, I said, you know, it was sort of about the elephant in the room, I said, nobody is talking, nobody has mentioned Hamas. What do you think about Hamas? And they politely, but definitely, unambiguously, completely ignored my question. They didn't answer it at all. They didn't even go through the motions of answering it. It was as though I hadn't spoken. And so again, one of the key potential veto players, on any possible settlement, was not at the table, even implicitly. Then they also lacked a K-Group. They indeed trumpeted the fact that they were 300, and eventually closer to 500 businesses involved, but even in the smaller meetings of BTI, where there were 15 to 20 businesses represented, people you interviewed there would tell you that there was never any agreement on what it was that they should actually be trying to do. And there was not any equivalent of Mike Spicer's declaration to me that in South Africa in the 1980s, you could get nine people around a table and it was the whole economy. So there was no K-Group, and therefore no capacity to turn this movement into something that could pursue a goal in a disciplined way. Again the refusal to get involved in the negotiation themselves, a big difference from the Consultative Business Movement, where the leaders of the movement rolled up their sleeves, started themselves talking to members of the liberation movement, committed themselves to majority-rule democracy, long before any politician was doing that, and so showed themselves, as the way I put it in the lecture, showed themselves as willing to internalize some of the costs of achieving a transition. And none of these business leaders was really doing that. They were simply running ads and making speeches, telling the politicians to make settlements. Almost no effort to build grassroots support, was a corollary of that. And one symptom of it is, if you look at the BTI website it's in two languages, what are they, they're in Hebrew and in English. There's no Arabic on the website. Again, just not so subtle signal that there was not much tactical, never mind strategic thinking about building grassroots support for the new order. And indeed, as Yossi Vardi said in the clip that I showed you, they had actually they thought made a smart decision. They said we have these conflicting narratives about the past, and we don't agree on what they are, and we've decided to stop talking about our conflicting narratives and simply move forward. But of course, if you wanna move forward without confronting the conflicting narratives at all, that puts a different question on the table, namely how to build support for the new order that you're trying to construct, among populations that are deeply wedded to these conflicting narratives. So the short answer to this is that the failure of BTI was heavily overdetermined. I think it was not well conceived and not well executed. But in the wake of the failure, first I, and then Nick and I, got interested in what some of the principles behind this were doing, instead. And that's what he's spent much of his undergraduate career doing, and then as I said, he was the main energy behind the case that we wrote, so over to Nick. (audience applauds) - Hi everyone. Thanks so much for having me on this rainy Thursday, and also for making your way over to SSF, which I remember fondly as an undergrad. So I think the place to begin is a little post-mortem on BTI. So we began by hearing the actual start of the movement, of the organization at the WEF in 2013. And now, we can look at some footage that we shot, during the summer of 2018, of those same principles of Vardi, of al-Masri, discussing what happened and what went wrong in BTI. I want you all to focus on this too, what different narratives come out of this? The Israeli narrative, Vardi talks about, the Palestinian narrative that al-Masri talks about, and the differences in how that informs perhaps some of the troubles that have plagued the organization from the beginning. - I think first of all, that the voice that we should have a two-state solution, and the voice that the business community will support, or at least parts, or big parts of the business community will support the government in doing it. It's a very important input in the future equation. We should continue to meet. - I believe in having the political solution first. I would never cooperate in any way after my experience with the Israelis. After experience with the 500 business people, the Israelis who said yes they wanted two-state solution, they wanted to do this, and when it came, when they WEF, the World Economic Forum, tried to do something they vanished. - We felt that the Israeli business community was not what we were thinking that they would be, in terms of being aggressive and push the political leadership to come in line with what they were privately saying. - These days it's not very popular to be a peacemaker. Not in the Israeli side, not in the Palestinian side. It's not very popular nowadays. There's no negotiation today and it's not very popular because the society doesn't support people that believe in negotiation and peace. The same happens in the Palestinian side. Maybe even more. - And to be frank, the failure came mostly after the war on Gaza in 2014. During the war, when you have suffering of people for 50, 55 days, and you don't hear a voice coming from Israel, trying to, or asking to stop that war, or to stop that part of grievance that we have, we thought that it seems all that we have done does not have a good sound on the other side. They could not manage to say something publicly against. And we felt alone. - In the minds of all Palestinians it's dead. In the minds of the Israelis they want to keep it this way for their own benefit, to create, to say that we are working. We're doing something with the Palestinians. But to me it's completely dead. - If you go down the street here, you ask any Palestinian what do you think, one-state, two-state, they'll tell you look, we take one state, we have one condition only, equal rights. - So very different in 2018 from the initial excitement about the organization in 2013. I think there's a lotta things that jump out. But to me, the most important part is you have the Palestinian leadership saying that this organization no longer exists and then you're also interviewing a CEO of the same organization, which is moving forward and still holding events. I think that dovetails into just a very brief history of the conflict. And what I want everyone to think about in terms of narratives, as we think about this conflict and the business that's operating within it, so much ink has been spilled on the history of the conflict. There's plenty of wonderful courses that you can take here at Yale on it. So our mission today isn't to actually understand the conflict, to determine what events took place to shape it, but more to be comfortable in the fact, that there are multiple narratives within it. If you read the case, I begin the history of the conflict with the first massive wave of European Jewish immigration to what is now Israel in the late 1900's, rather 19th century. If you talk to some, they would define the start of this conflict of this story, particularly religious, or far-right Jews, as going back to the times of the Old Testament. For Palestinians it could be during the 17th century, where there were efforts to gain autonomy from the Ottoman Empire. So our goal here is not to establish the history, but to recognize that there are multiple narratives, and that multiple narratives can be true. I've two images up here to show this conflict, or this contrast. On one hand, you see the separation wall. It is two different people. A separation wall, a security border, and an apartheid wall. Also on the left, you see not only is there a conflict of separation that you see in the barrier, but there's also one of co-existence and of proximity. This is in the east of Jerusalem. You see an ultra-Orthodox family, passing Palestinian Arabs. So I want everyone to get comfortable with the fact that when we think about the principles in the case we'll be discussing and how they operate that there isn't a single narrative that defines the conflict itself. I think everyone subscribes to different narratives. And one of the challenges that they face, and anyone I think in a conflict zone like this faces is actually understanding that there's multiple narratives. That what you're operating and the assumptions that you're taking are not necessarily universal. So jumping back, Professor Shapiro discussed the two-state solutions and the facts on the ground that now make it quite difficult to imagine that it will take hold any time in the near future, but if you've been following contemporary Israeli politics at all, you know that this has perhaps become more and more moribund in the last few months. So the left in Israel, the Labor Party, which had been the traditional left, has effectively dried up as a major force. It's now Likud on the right, and then a more centrist party Blue and White, under Benny Gantz. And as you saw in the spring, when there was this first period of campaigning and now they're going through renewed cycles of trying to form a government, Likud especially under Netanyahu was pushing facts on the ground increasingly away from a two-state solution, openly discussing annexing the West Bank. And further, there wasn't a left left, or even a center party that was providing a viable alternative. So one of the things that I think we need to discuss and think about BTI, is where do you move if you're operating in this conflict, if what's been the dogma for it, since really the beginning of two-state solution, and that's both in Israel, among Fatah and the Palestinians and in the U.S. too. Congress right now is discussing passing a resolution affirming its support for the two-state solution. If you're moving away from that, if the status quo is changing, then how do you actually operate? Where are you moving towards? The economic realities of today are really one of separation. Israel, if you've kept up with their economy, is truly a technology powerhouse. Start-up nation is used colloquially often to describe it. But it's one of the leading technologically-advanced countries in the world. In terms of startups per capita, it is the world's leader. The Palestinian territories, so speaking about the West Bank and Gaza, are diverging, but also far behind. Though after the Oslo Accords, there's a bit of an uptick in the economic fortunes of the West Bank. Increasingly, this has become more and more difficult, as intifadas have created conditions, and you're seeing above a shot on a street in Hebron, a wall separating a Jewish portion from a Palestinian portion. Destruction of capital, the difficulty of moving goods, and this is particularly true in Gaza. The economic situation in the West Bank, though nowhere approximating that of Israel, is still far more advanced than Gaza, which has been under blockade for some years now. But I think the second point to make, is there's always going to be a sense of intractability in the conflict too, even as you have this divergence economically. Israel's a very different place, if you think about its economy from Palestine, because of the geography of the conflict, because of contours of demographics, of shared history, there's always going to be this element of shared fate. And I put this photo up, which at first glance looks like a standard Israeli beach, overlooking the Mediterranean, to make that very fact. So if you went through the case, you'll know that what we're looking at here is a beach just south of the Israeli city of Ashkelon, looking south to Gaza City. It's quite interesting to go to this beach because while it's well within range of rocket fire, during periods of conflict, it's quite popular among local Israelis to go to the beach. But perhaps more importantly to us, when we think about the interconnectedness of geography, is just north of this beach is the Ashkelon desalination plant. So Israel in a part of the region that is quite water scarce, has turned towards desalination as a way to get much of its water capacity. And it was found out that there's a period in recent years, where the Ashkelon desalination plant had to be closed because of raw sewage being let out in Gaza due to issues processing it with lack of electricity, capital infrastructure et cetera, had been dumping raw sewage into the Mediterranean, flowing up to Israel and shutting down the desalination plant. And all this is to say, that even as you have this divergence, even as you try to establish a status quo of separation, even as you try to kick the can down the road and pass the buck on what a solution is, there's always going to be this intractability. And I think that's something we're going to continue seeing as we discuss the different case examples. Now let's talk very briefly about geography. Who here has been to Israel or the West Bank? Any impressions on its size when traveling through it? You can just shout it out. - [Man] Small. - Small, it's about the size of New Jersey. At its narrowest point it's about 15k across. And you can see, that part of that means, that even if you don't have the two-state solution, if you have a separation of the Palestinians, of the Israelis, that there is shared fortune just by nature of the geography. And the geography is quite contested. You can have a situation like we see in Gaza, where you have a blockade. Where you have Israel totally removing itself from connections to Gaza and yet because of the environment, because of the closeness of the geography there's still has to necessarily be connections. And again, geography is so important to understanding this conflict. You think about the West Bank there. Right now we're using a United Nations map, but the borders that are drawn here are very much ones that are understand contest. Similarly, if we think about what to even call the West Bank. There's huge contest about what that would be. There would be those that would call it the West Bank, those that would call it the state of Palestine and those who would call it Judea and Samaria. So again, when we think about contested geography, when we think about different narratives, there's the fact that all of these folks, all these principals, are operating with very different assumptions in a very confined space. Now, I think the central question we're all going to consider together today, is what can business do, so what can BTI and similar folks do, when the political process fails? And to begin, let's go quickly through the major peace talks for the Israel and Palestine conflict in the last 25 or so years. So beginning in 1991, the lead up to Oslo, you had the Madrid conference, you had Oslo 1, Oslo 2, the Hebron Protocol, the Wye River Memorandum, the Sharm el-Sheik Memorandum, the Camp David Summit, the Taba Summit, the Annapolis Conference, you then had Senator George Mitchell leading talks, the John Kerry-led talks, which dovetails into BTI, and what you saw at the start of class today, and then 2019, or coming sometime soon, the Kushner Peace talks. And think I've had this slide for a couple years now and I was telling Professor Shapiro, every time I do this talk I have to update that date. But hopefully sometime in the near future we'll hear about the Kushner Peace Plan. But I think the takeaway here is what's the actual progress towards any solution, and it's nearly none. So if you're a business looking at this, despite all the efforts, all the talks and expend by the Arab League, the belligerents themselves, the U.S., Russia has been involved in the quartet, where do you actually go from here, when the political process fails and is stagnant? So I think there's three options for the private sector. And for the next portion of the lecture, I want everyone to imagine themselves as a business person in Israel, in Palestine, facing this stagnant political process, and thinking about what options you have available if you wanna move things along? And the three I'm gonna layout is, one, and this I think fits into BTI's approach is, you can revive, or influence the former political process. So you can, like BTI, say we are gonna take our collective influence in the business community, we have 30% of the GDP, and we're gonna pressure politicians on both sides to re-engage in the political process. We're not going to actually shape it, we're not gonna say what a solution's gonna look like, but we're gonna be a pressure group that makes our politicians move forward with it. The second, and we'll see some examples of this, though I think it becomes quite fraught, is you can act completely independently of the political process. So you can say that the status quo is not moving anywhere from the political process. Talks aren't doing anything, there's no discussion. So what I'm gonna do, is I'm gonna change facts on the ground myself. I'm gonna remove myself from the political process, and I'm gonna act independently of it, and I'm gonna change the actual facts by using my influence as the private sector. And then the third, and I think the one that you'll see most of the cases we discussed aligned to because of the nature of the conflict, is you act both within and without the political process. Fundamentally what I mean by that is you cant's wring the politics outta politics, particularly in a high-conflict zone like Israel and Palestine, even when you are trying to act independently of the political process. And we'll see this more as we go through the case. You are necessarily being political in doing so. Even the act of operating outside the political process is political, and quickly becomes political. So there's three main principles I wanna talk about today. And you likely saw them in the case. They are SodaStream, Rawabi City, and then also EcoPeace Middle East, all of which are very different, but also I want everyone to consider, and we'll discuss at the end, what are the common threads, we have one of them, which is the challenge both from the left and the right to their progress. But consider, what are the similar challenges, the similar narratives, perspectives, shared by all these? So we'll begin with SodaStream and a quick video. - [Reporter] Here's one commercial millions saw during the Super Bowl. - Mix in the perfect flavor. Look, soda that's better for you and all of us. - [Reporter] Johansson endorses SodaStream and for years served as a spokesperson for Oxfam, a group fighting global poverty. But Oxfam recently told Johansson, her work with SodaStream was incompatible, because it operates a factory in what they call, the occupied West Bank. - This is the plant at the heart of the controversy. It's about a 20-minute drive outside of Jerusalem and it's become the target of the BDS movement. BDS means boycott, divestment, and sanctions. The factory employs 13 hundred workers, 500 of them are Palestinian. Some in the BDS movement and groups like Oxfam would like to shut the factory down. SodaStream's CEO says they should come to his factory. - I say they should come here and look my workers in the eye and then tell me that they wanna close the factory. 'Cause you know, it's easy to wanna close the factory when you're sitting behind a Oxfam desk in New York, or wherever they sit, and claim that we are the ones who are hurting human rights. - [Reporter] Oxfam claims, businesses such as SodaStream, further the on-going poverty of the Palestinian communities that we work to support. Yet SodaStream workers, who make three to four times what other Palestinians make in the West Bank, tell a different story. - We love SodaStream. We need SodaStream. Where do we go after this job? - Anyone have a SodaStream at home? So SodaStream is a unique company, in that today if you go to the Negev in southern Israel and walk into their factory, you will see a scene similar to this, which was taken the summer of 2018. Which is Israelis and Palestinians working together on a factory line. And what I'll add in there, it's not just Israeli workers supervising Palestinian labor. There are Palestinian foremen and managers on the SodaStream factory floor, which is quite unique within the conflict and within the region. Now how did we get to this place of SodaStream, which Daniel Birnbaum who you saw at the end of the clip, and factors heavily into the case, refers to as an island of peace within the conflict. An example of cooperation between Israelis and Palestinians coming together in business to change the facts on the ground and change narratives around cooperation. It began with I think, much lower sights. So SodaStream, and you saw a clip from 2013, of when that factory was originally in the West Bank, began hiring Palestinians because of a lack of labor. It wasn't a principled move. Daniel Birnbaum, when he became CEO, needed additional labor to deal with growing demand, and hired Palestinians. Very quickly though, I think he realized that there was something there. That Israelis and Palestinians could work together. There were of course challenges. Some Israeli employees initially quit the West Bank factory when Palestinians were introduced. And you had challenges, coming both from the left and right, to this experiment. So SodaStream moved from something that was originally using Palestinians as additional labor, into them becoming an important part of the company, serving as foremen, as floor managers, getting full health benefits, commensurate of that provided by the Israeli state system. But as soon as that began, and soon after, as you saw in that clip, there were challenges from both the left and the right. On the left as you saw was BDS, which looked at something like SodaStream as normalizing the status quo. As employing Palestinians, but not actually making any fundamental changes to the political situation on the ground. And what it was doing, by operating in the West Bank, was just normalizing an occupation. And this was something that was heavy enough in pressure, that depending on who you believe in the decision making, but something that I lean towards, was a significant factor in SodaStream leaving the West Bank altogether. The second was on the Israeli right, which saw Israelis and Palestinians working together, as fundamentally challenging the narrative of the conflict. So very interestingly it was even once SodaStream left the West Bank, for a new factory in Israel, they brought their Palestinian workers with them. They worked with the Israeli government to get work permits to bring the Palestinian workers along with with them to the Negev. So I think there's two reasons for this. One is they were fundamental to the production process. They were skilled labor. They were in management positions. And two because Daniel Birnbaum had begun to see this cooperation as something that could actually be influencing consumers. If you have a SodaStream now, and you look on the back of the box it comes in, you'll see that it says, produced by Arabs, and Jews, and Israelis, working together in peace and harmony. So again, it was a way to actually change the narrative around the conflict. But, when I first started getting involved in SodaStream, and this is in 2016, they had just had their last employees in the Israeli factory, the last Palestinian employees, lose their work permits. And there's this question, and Birnbaum became a very outspoken critic of Prime Minister Benjamin Netanyahu over this, about these permits being canceled. If you ask Birnbaum, the reason they were canceled, was because for the right, for those who were not looking to change the political status quo, not looking for cooperation, it represented a threat to that narrative. So the Palestinian workers were reinstated, in waves after a heavy pressure campaign by SodaStream, but it showed that it was a capricious solution at best, both on the left, BDS, who still criticizes SodaStream. And there are several Nordic countries, before this factory was moved, that wouldn't accept SodaStream products made in the West Bank. They had to outsource it to China. But also on the right, if you look at Prime Minister Netanyahu and the idea that, well why would we bring Palestinian labor into Israel when we have the unemployed in Israel continues to be a challenge to it. But I think what you see from the example of SodaStream is the ability to influence the status quo on a grassroots level. You're able to build business and people-to-people connections that actually change the outlook. And it's quite interesting, and there's some quotes in the case, about if you speak to folks, both Israeli and Palestinian who work at SodaStream, to hear how their outlook on the conflict, and on the other side the other, changes by actually working together. Having Palestinians manage Israelis, Israelis manage Palestinians, and as you see, sit together on the factory line for hours every day. So this, I think if you look at it, Birnbaum would say, he's trying to operate outside of the political process. This is about business. This is about what he thinks is right. But if we think about the realities, it is inherently political, because you are challenging the facts on the ground. And I think both the reaction of the left, with BDS, and the right, with work permits being canceled and this being challenges bringing Palestinians into Israel, show exactly that. SodaStream today, continues to operate in this way, but has to inherently be political to renew work permits. It needs to lobby the Israeli government and it further needs to lobby those aligned with BDS to show that what they are doing is not actually normalizing an occupation. And it shows, even examples like this, which you would think is straightforward, you have Israelis and Palestinians working together, it can become very fraught, and very political, very quickly. And adding to the complication recently, I believe the summer of 2018, SodaStream was sold to Pepsi. So they are now a subsidiary of Pepsi and it leads the question, at what point do the needs of a business trump whatever social good it's trying to promote, or social mission it has. So, so far, SodaStream's kept up with the model of having Israelis and Palestinians together. But the question remains of, will Pepsi have the stomach to continue that in the long run? The next I wanna jump into is Rawabi City. And Rawabi City, and we'll hear from Bashar al-Masri, who's actually the nephew of Munib al-Masri, shortly. But Rawabi City is the first planned city in Palestine, and in the West Bank specifically. The most significant construction project in terms of scale, that's really ever been undertaken in the West Bank. - I feel more of a politician sometimes than a business person. Politics is all over us, you cannot take the politics out of the Palestinian lives, not just business people. With Rawabi it was difficult for the Israeli, that Israeli propaganda, the negative propaganda, to paint Rawabi they way they paint the whole Palestinian situation. One of the most important issues of Rawabi is the Palestinians' feeling of pride, that yes we can take matters with our own hand, despite the fact that we're under occupation, we don't have to give in to the occupation and leave. We can defy the occupation and build on our land. Of course it's also a great message to the international community that we are sophisticated people. We're not just as we are perceived the terrorists and the victims. We are the victims that's for sure, but we're not simply terrorists, we're just people who are fighting for our rights. And now we are working hard to build our nation. 140,000 Palestinian workers every day cross into Israel, to work in Israel. Are they normalizers? Of course they are. Do they have a choice? They have a choice of being unemployed definitely. And what I'm doing with normalizing is I'm creating credible jobs for these people, so they're not humiliated. They're getting salaries, higher salaries. They're getting their rights. And IT professional does not need to work as a gardener, or as a laborer in Israel, when they can work here as a programmer, which is what they studied to be. The ultimate evil is the settlers of course. And then the occupation, which is the cover for the settlers. We don't have issues with the ordinary Israeli people. The settlement movement is an ideological movement. In the core of it, we should not exist here and they should be on this land. Ironically, we found it was Israelis and pro-Israelis that lobbied our case. The water crisis was resolved by pro-Israeli groups and Israeli groups, and some even Israeli officials, including the late President of Israel, who was calling his Prime Minister and his Defense Minister and say, why aren't you giving water to this city? So perhaps you could see this as an alternative way of struggle and to reaching your rights. - So if SodaStream was meant to be incremental in changing the facts on the ground, by bringing Israelis and Palestinians together, and promoting understanding. Rawabi was meant to shift the status quo altogether. As you can see, this is a picture of Rawabi, taken during construction. It's a massive endeavor. It's meant to be a planned community for 40,000 residents. Schools, hospitals, and also employment opportunities all within. And if we think about the three options that we discussed at the beginning available to the private sector, Bashar al-Masri basically looked at them and said, well what I'm going to do is totally remove myself from a flawed political process. I'm going to build on land that is in area A. So if we think about Israeli territory in the West Bank, there's area C, which is administered by Israel. There's area B, which has security oversight by Israel. And then area A, which is completely under Palestinian authority. And he said, to remove myself from the political process, I'm gonna build only in area A. And what I'm going to do is create a community that changes the status quo on a couple levels. One is I'm going to change the economics of the West Bank. Right now we have 3,000 university graduates per year in the technology fields and we're unable to sufficiently get jobs for all of them. So I'm gonna promote a tech hub. What I'm also going to do is change the narrative around Palestinians. No longer are we going to have to go and cross into Israel for employment. We're going to have employment opportunities here in the West Bank. Now, getting back into the political side of the conflict, very quickly, something that was meant to be fully removed from the political process, Masri went as far as to build only on area A to ensure this, became highly political. There was of course the detractors, and he spoke a little bit about this, on the Israeli right. Settlers who saw this as a challenge to Israeli sovereignty over the West Bank. And they provided more of a nuisance than an actual challenge, when it came to it, but were certainly an opposing force. The second being BDS, which looked at this and said, well what you're doing is again normalizing the occupation. Though you're improving the material conditions of those who live in Rawabi, you are normalizing the same conditions. You're making life under the same occupation more bearable. And then there was the third one, that in a contested geography like this, even when you build on Palestinian territory in area A, you still are deeply connected to an Israeli political system. So you heard Masri talk about water. One of the major challenges that faced Rawabi was actually getting water access through area C. And for a long time, construction of the project, and having residents moved in, was delayed because the Israeli government COGAT, which administers the territories, did not approve water to be sent over. So it actually took pressure, from both the Palestinian side, and as you're heard the Israeli left, including Shimon Peres, to get the water approved. But even now, with Rawabi welcoming its first residents, with water flowing to the city, with the tech hub up and running, it remains a highly-political process to ensure its success. So one of the things that you'll notice on this photo of the city, is how narrow the road is leading into Rawabi. For a city of 40,000 people it has a two-lane road, so car on either side, coming in and out. And this can create massive traffic buildups, when you have folks coming into Rawabi for a concert or an event. And again it will depend on the Israeli administration to approve expanding this road. And even having residents cross from a major Palestinian city like Ramallah, through area C, back into area A to go to Rawabi. So again, even when you change the conditions on the ground, when you try to insulate yourself from the political process within this conflict, it's quite hard to do so. At the same time, what Masri is hoping to do is change the status quo. If you're able to create this and have it be successful then you influence the political process in turn. You see what Rawabi could be, multiplied 10 over, 20 over, which is a Palestinian state. You start on a small basis, as an example, and then you move forward with the opportunity to show what Rawabi could be, and ultimately what a Palestinian state could be. Now we're slightly tight on time, so I'm going to go through the video we had for EcoPeace. But this gets back to that original photo that I showed when we started thinking about the geography of Israel and of Palestine. This is a photo in Gaza City of the Mediterranean. So again, after conflict, limited electricity, the ability of Gaza to process garbage, sewage, et cetera, is quite limited. And what this does, is not only impacts their coastline, but goes up the coastline towards Israel. And moreover, if you're thinking about ground water, which is another key water source within the region, it can seep across the border from Gaza, into aquifers in Israel. And what this points out, and what EcoPeace, which formed in the wake of the Oslo Accords, to actually bring Palestinians, Jordanians, and Israelis together, the first NGO, and really the only lasting one to bring those three groups together, is to recognize that even if the conflict creates disengagement, even if Gaza is separated from Israel, the demarcations of the state, don't define the actual geography of it, the environment. And what they've basically come and looked at and seen that even you have a political process that is flawed, that isn't moving forward, you can't disentangle the environmental implications of this. So one of their main focus projects has been looking at the connection of water between Israel and Palestine. Finding ways for Israelis and Palestinians to improve and work together to improve the environment, particularly water. So Gaza is a great example of this because Israel, despite being fully disengaged from it, has had to recognize the environmental implications of what's occurring in Gaza, the destruction of capital infrastructure, of the lack of electricity, et cetera. But they've taken it a step further too. We're having this idea of you have to leverage the environment, or pay attention to the environment because of the connection. They've taken it a step further and said, well if we have a fundamental connection through the environment, is there a way that we can introduce a business, regardless of the political process, to actually take that reality of environmental interconnectedness, and use it as a way to create cooperation. And there's this idea, and I encourage you all to go into the case and read more about it, but an idea by Gidon Bromberg, who is the founder and Israel head of EcoPeace Middle East, of a water/energy nexus. Looking at the differences between the environmental resources of Israel, Jordan, and the Palestinian territories in Gaza, and finding a way to create an actual and business incentive for them all to collaborate. So Gidon looks at the situation and says, Jordan is limited in its water, but it has huge amounts of low-cost electricity because of its deserts and solar energy. What if there is a way to have Jordan send its low-cost electricity across Israel into a place like Gaza, which has access to ocean water which can be desalinated. Use that electricity to produce, at low-cost desalinated water, which is then sent back to Jordan via Israel? Is there a way to align business incentives within an environmental context, you have a problem with water scarcity, and electricity being scarce, is there a way to combine these incentives, to find a solution that brings all the principals together, from an environmental perspective? So that's something that, when we first talked to him in the summer of 2018, was in its infant stages, but he continues to push on, as a way to take the realities of the environment of scarce resources and create cooperation from that. So very quickly, this I think is scratching the surface of a lotta what's in the case. And I do hope you'll all take the time to go through it. But to summarize and go through the lessons learned quickly, I think the things that stick out to me, and the things to consider is how hard it is for business to have a political impact on the conflict because one, the arguments about normalizing the status quo actually ignore the way that it changes over time. Oftentimes, on both criticisms from the left and the right, you'll hear about normalizing a status quo. But how do we actually define that if it's changing? And a few examples is the economic divergence between the West Bank and Gaza, and then also from Israel and the West Bank and Gaza. A water crisis and general humanitarian crisis in Gaza growing worse and worse, and its environmental implications. And then also the changing regional geopolitics. The fact that the principals, who would have supported the Palestinian cause, perhaps more stringently, thinking about the Gulf countries especially, are no longer as involved or as focused. The second is, there's neither sticks or carrots present. If you're thinking about Israel, and you're thinking about being in Tel Aviv, it feels quite removed from the conflict nowaday. If Israeli's economy is thriving and most business leaders are not seeing necessarily the impetus for a change, particularly on the Israeli side, then what is the real incentive there? It has been dangled, you know since the Oslo Accords, of accessing Gulf markets, but if you're having other factors in the region, such as alliances against Iran, bringing countries like Israel and Saudi Arabia closer and closer together, then is there a need to actually resolve the conflict? And additionally, as Professor Shapiro talked about in the context of BTI, there isn't a K-Group. When you have such a diverse economy, even an organization like BTI only represents a fraction of the interests involved. The last one is the fact that political neutrality in a conflict like this is not an option. And again, getting back to the point about why, even when you're within the political process, and without the political process, you can't wring the politics out of politics. Even something like Rawabi, which is meant to be outside of the status quo, outside of the political process, is inherently political. So I think what to step back and look at, from both of these three examples, and the additional examples in the case, is the best option forward is to actually look at scalable demonstration projects, projects that operate on a smaller scale, and use them as an example that can be built on to have a positive impact. For business leaders to start small and use that to influence the political process as it moves forward. And I think these three, though all very different, are an example of that. So what I'll do now is hand it back over to Professor Shapiro, to talk about some of the more conceptual components that we have learned about. - Thank you very much Nick. (audience applauds) Thank you. So as he said several times, he's really scratched the surface of this incredibly rich case, that he has done most of the heavy lifting to put together. He spent, I don't know how he graduated from Yale College, 'cause he must have spent half of his time, when he was supposed to be in classes here, wandering around the Middle East. But nonetheless, I think that the place to start is where he ended. When you look at these conflicts there are these endlessly repeated phrases and they're repeated so often that you have to wonder whether they have any real referent. One is that the two-state solution is, everybody repeats it as the only option, whereas it's clearly the case that it's not an option. And it hasn't been an option for a long time. And so, that's something to think about. But the other is this idea of normalizing the status quo. And the thing about the phrase normalizing the status quo, it sounds like what you're normalizing is the absence of change. But in fact, there was a very famous paper published in 1975 by a Yale faculty member now retired called The limits of consensual decision, in which he talked about the concept of utility drift. And he had in mind the fact that making things hard to change, also makes it hard to change things if the facts on the ground change. So the American founders made the Constitution very difficult to change, but that doesn't mean that we are locked into the world as it existed in 1787. On the contrary, the facts on the ground change and then that can bubble up and influence the way in which people think about the status quo. And so, a few things just to highlight that have come out of Nick's presentation here, and earlier discussion the failure of BTI. One is I think you cannot overstate, and this is not just Israel or Palestine, but the issue of water in the upcoming years and decades, across the Middle East and North Africa, that are either gonna create catastrophic crises, or force greater integration. The fact is, what EcoPeace is doing, and by the way EcoPeace has been around 30 years. It has directors from Jordan, Palestine, and Israel. It's a genuinely tripartite organization. But they are raising capital for the project that Nick talked about, that is essentially going to produce cheap electricity in Jordan, to supply water for Gaza to manage its sewage, so that Israeli desalination plants can actually operate. And it's an example of the crisis forcing forms of cooperation that hadn't even been thought about five years ago. Another, you get a brief sense from the Rawabi story, is that most outsider's perception of Palestinian workers are laborers who will be going and working on building sites in Tel Aviv. But in fact, there are a number of universities now in the West Bank, graduating 25 hundred software engineers a year. These are people who are well educated and are employable in many walks of life. Yadin Kaufmann in one of the other things we mentioned in the case has created something called the Palestinian Internship Program, where people in, by the way Bashar al-Masri was one of the people who did it, you spend a, Palestinian goes and spends a year in an Israeli tech company, and then has to go back. And several of the startups in Rawabi are by graduates of, their tech startups, by graduates of the PIP, the Palestinian Internship Program. He has also started a venture capital fund, with no Israeli money, so it can't be seen as neo-colonialism to fund Palestinian startups in the West Bank, particularly in Rawabi. So again this is, it's producing a new population, a new generation of Palestinian youth with very different education, background, and credentials, and aspirations. And these people are growing, they're multiplying. By the way, it produces some over things that BDS doesn't like. Like growing inequality within the West Bank, because you're seeing a much more differentiated economy. And by the way, considerable divergence between the economies of the West Bank and Gaza. Furthermore, the far right wing Israelis, their agenda for a long time, as in the video, has basically been to make life so unpleasant for the Palestinians, that eventually they will go to Jordan. But the truth is that because of the economic dynamism in parts of the West Bank, and the sorts of things like Rawabi you're seeing, is that the per capita income in the West Bank, not only is diverging from Gaza, but is actually soon going to equal and exceed the per capita in Jordan. So you may well start to see the settlers' worst nightmare, you may well actually start to see Palestinians from Jordan coming into the West Bank. And so this is another example where, the word status quo suggests no change, but in fact the change on the ground is likely to start forcing political change. And as you saw at the end of the post mortem video, one of the Palestinian businessmen saying, actually we have no problem with one state, but what we want is human rights. And I think that this is ultimately where this conflict is now headed. That people instead of continuing to repeat the mantra of a two-state solution, are gonna start demanding the vote. And of course that is gonna put a very different kinds of pressures on Israeli politics. But I think it is clearly a case where, unlike South Africa, where the business community could take a role in finding a settlement that was potentially there, but needed assistance along the way, this conflict isn't at that point, but it certainly is heading to the point at which the sorts of things that these businesses have been doing will start to make new possibilities come onto the horizon, and perhaps, which really has to always happen for these kinds of conflicts to be solved, is for the unthinkable to become thinkable. Okay we will talk about crisis, crash, and response, next Tuesday. (relaxing music) |
Power_and_Politics_in_Todays_World | Lecture_17_Filling_the_Void_China_in_Africa.txt | - Good morning everybody. So most of you know that Christina Seyfried is the head teaching fellow in the course. But what you don't know about her is that she's writing a dissertation, she's an advanced graduate student in political science writing a dissertation about China in Africa. And has spent much of the last two years in Nigeria, Ethiopia, South Africa and elsewhere studying just what the Chinese are doing there. She is also the brains and energy behind Yale's China-Africa initiative that has been going on for the last several years. So it seemed absurd for me to be giving this lecture when there's somebody who knows in such more intricate detail about what's going on then Christina so, Christina, welcome. The floor is yours. (audience clapping) - Great, thank you so much, professor. Welcome, everyone. I just wanna say I feel extremely honored to be giving this lecture today on China in Africa, especially because my journey really with this topic started three years ago. In my first semester here at Yale when I took a class with Professor Shapiro at the business school, and one of the lectures was on China in Africa. And I found the topic extremely interesting. And I started to read more about it. But I was a little bit not completely satisfied with the rather China centric angle that literature was taking on the topic. So I decided to write my dissertation on how African governments strategize around Chinese investments in their countries. And I spent 10 months of doing field research now in the African region. And as I said, I feel extremely honored now to be back and to be standing on the other side now of the classroom, and to teach a lecture on China-Africa myself and to share some of the lessons from my research with you today. So to start us off, I brought you a video from the forum on China-Africa cooperation, that just happened in September of last year. And China started these forums in 2000. And they have been happening every three years. And China uses them as a platform in order to announce a strategy for Africa. So in terms of loans, trade and investment for the next three year period. - [Woman] Building a community with a shared future and seeking inclusive growth, the opening ceremony of this year's summit for the forum on China-Africa cooperation, saw President Xi Jinping outline his vision for China's ties with the African continent in the new era. (speaking foreign language) - [Translator] China is the largest developing country in the world, and Africa is home to the largest number of developing countries. China and Africa have already formed a community with a shared future. We stand united with Africa's people thinking with one mind and working with one heart. - [Woman] President Xi Jinping has announced eight major initiatives in the coming three years. (speaking foreign language) - [Translator] The framework of the 10 major cooperation plans, China will work closely with African countries and implement eight major initiatives in the next three years and beyond. It feels like industrial acceleration, infrastructure connectivity, trade facilitation, green development, capacity building, healthcare, people to people exchanges, and peace and security. (speaking foreign language) In order to promote the successful implementation of the eight major initiatives, China will provide $60 billion in support for Africa. For government assistance, financial institutions, and financing from Chinese enterprises. It will include $15 billion in free aid, interest free loans, and concessional loans. $20 billion in credit funds, $10 billion in China-Africa developmental financing, $5 billion in special funds for import trade from Africa, no less than $10 billion in investment from Chinese companies over the next three years. - [Woman] Already the summit has passed two outcome documents, including the Beijing declaration and the FOCAC action plan 2019 to 2021. Participants say that in a current era of globalization, all nations have a common destiny. And to put it in the words of President Xi, everyone has in himself, a little bit of others. - So as you can see, a lot of big announcements were made at this forum. And this last image here has actually become very famous because for the first time in the history of these forums, 53 out of 54 African countries were represented. And I remember an African newspaper called Africa News wrote at the time that this was the week where the African continent was bereft of its leaders because everyone was hanging out in Beijing. So this is just really the most recent peak of China-Africa relations that have gained more and more international attention from practitioners and academics over really the past 10 years or so. But why is it that this has become such a hot topic from a developmental point of view I mean? And in order to answer this question, I would like us to take a step back and think about what are the African regions persistent growth obstacles? And I brought you a comparison here from 1982 to 2016. And you see in the 1980s, China really was poor agricultural, and had just emerged out of a series of intense conflict, namely the Cultural Revolution. And you see that China was in fact poorer than most African countries, except for Guinea Bissau and Uganda. But fast forward to 2016, China is now richer than most African countries except for Equatorial Guinea and the Seychelles. And I know in lecture four we studied very much in detail China's growth miracle. But what was meanwhile happening in the African region. So after decolonization, several African leaders starting in Nigeria, Zambia and Tanzania, tried to import substitution industrialization. And that means you basically keep your economy closed to international trade and you try to grow and develop local industries first that compete within the domestic market. But that is an economic growth recipe really failed. So after several debts and oil crisis, the IMF and the World Bank stepped into Africa and implemented so called structural adjustment programs. And the idea was that everything needed to solve Africa's poverty was real neoliberalism through the Washington Consensus. So deregulation, privatization, and liberalization. And after that, everything would fall into place. But that logic was flawed. And even after these countries went through structural adjustment programs, foreign investors weren't really willing to come to Africa still. So at the turn of the millennial, Jeffrey Sachs who was a very famous economist at Columbia. He also worked on the Millennium Development Goals with the UN. He wrote this famous piece on Africa being caught in a poverty trap. And what he's saying is that because of diseases, low stocks of human capital, technology, high transportation costs and an adverse geopolitical location, Africa is really stuck with both weak growth and weak institutions. And we have studied the modernization theory that says that growth supposedly leads to strong institutions. But the question here is what leads to growth in the first place. And he's saying that foreign investors don't want to come to the African region, because they aren't strong institutions that could enforce property rights, the rule of law, et cetera. And he's really advocating for aid as the solution to Africa. And there are economists such as William Easterly or Dambisa Moyo who highly criticize this call for aid as a solution for poverty. But it remained the fact that the African region wasn't able to attract much investment. And then, with the wake of the financial crisis in 2008, it looked even more dire because you have the collapse of the Washington Consensus. And if this hit to this overconfidence and neoliberalism as a concept and then the solution to end poverty, and you really see a retraction of Western investment in Africa, also because the West is running capital shortages, as they have to bail out banks at home. So now the question is, where can Africa attract investment from now? But as Deborah Brautigam in her book that you were also supposed to read for this lecture today, predicted, China meanwhile is rising and is running huge capital surpluses. So where others are exiting as the path of least resistance going back to Hirschman, who we have studied as well. China now fills the void that the West is leaving and increases its presence in Africa. Hence we know that China has lifted millions of people out of poverty. And I would really recommend this book by Ang "How China Escaped the Poverty Trap". So now the question is, is China just a new colonialist? Or is its presence going to add something beneficial to the African continent? And as I walk you through the slides, I would like you to think about this question, whether it is even the right question to ask and I'll also offer you my opinion, as we move along. So for today's agenda, I will provide you with a quick historic overview on China-Africa relations, then I'll give you a big picture of what China is doing in Africa today. And then we'll dive into some key questions that I think are important to ask about this relationship. And I will use two case studies, Nigeria and Ethiopia, which are among the top receivers of Chinese assistance in Africa. And then I'll talk about some lessons and I wanna bring up this idea of African bargaining power here and how much African agency matters in this relationship. And then finally I'll give you my two cents on where I think this relationship with China is going in the future. So a lot of people think that China really started to become present in Africa in the early 2000s. When in fact, we have to go back to the end of World War II. And African countries are slowly becoming independent, while China is meanwhile fighting its own domestic conflict, where the communists had just driven out the nationalists to Taiwan. And Beijing, China is now looking for allies in this Beijing-Taiwan conflict. And after independence, 14 African countries actually decided to side with Beijing here. And we have very skillful leaders like Sekou Toure in Guinea who played the United States and the Soviet Union against each other in their aid offers. And then he secretly accepted $25 million from China on the backside. China at that time is also rewarding socialists and communist movements as in Guinea but also in Ghana and Mali but after the Sino-Soviet split, which was basically about personal differences between Mao Zedong and Nikita Khrushchev, but also different ideological interpretation of Marx Leninism. Sorry, Marxism-Leninism. China has an interest in keeping up this non aligned world that neither sides with the US or the Soviet Union here. And what's really interesting is that in 1971, Beijing, China is actually able to gain back a seat at the UN that had been previously occupied by Taiwan. And Tanzania was very, very vocal here in these negotiations. So we studied how in 1978, China opened up its borders to the world. And it officially starts its bringing in period where it is trying to attract as much foreign capital as possible. And what's really striking here is that China is the only developing country in the world that is also offering aid to other developing countries. And the strategy here is very much based on its own experience as an aid recipient for example from Japan. That during that time knocked on China stores and said, we'll offer you low interest, low interest loans for technology and material and in exchange for crude oil. And we see very similar deals still being made today between China and Africa. And we studied how China was able to rise very fast. So starting in the early 1990s, China's running huge capital surpluses and is officially entering this going out period, that we're still in today. And China is trying to put its capital abroad, to invest it abroad, and it really uses its aid as a strategic tool to promote Chinese exports, and to help Chinese infant companies really grow and learn in other countries. And it also positions itself sort of as an alternative to Western aid. So in 2005 in Tanzania President Mkapa, he really when he stepped down, wanted to give his population kind of as a departure gift, a football stadium. But the problem was that Tanzania was on the list of highly indebted poor nations. And they were supposed to actually run austerity measures, and couldn't spend all that money. But then he went to China and China said, okay, we're willing to build you that football stadium. It's gonna be smaller, but we're still gonna build you one. So he was happy. But what's very interesting is that in the early 2000s, China seems to have accepted the rules of neoliberalism. And the US is an economic headroom at that time, because it joins the WTO at a lot of pressure from the US and it seems like it wants to play by the capitalist international rules that were set up by the US. But then as we studied the financial crisis hits in 2008, and this is also a hit to this overconfidence that neoliberalism is sort of the way to go. And it gets a lot of affirmation to the Beijing, China model which is the state capitalism and that it sometimes it's better if the government owns companies in a certain sector that is deemed strategic. And if you have large infrastructure investments and that sort of an interventionist approach by the government is more effective than leaving everything to the private market. And an embodiment of this new confidence that China has found is the Asian Infrastructure Investment Bank, and the Belt and Road Initiative that I'll talk about in a little bit. But I would really recommend this book by Kurlantzick here, where he also talks about the effect of state capitalism and the spread of it in today's world on democracy if you're interested. And as I said, the most recent peak of China-Africa relations has really been this FOCAC meeting of 2018 where you saw the clip, where really China is trying to rebrand this relationship with Africa and it has signal that it's been listening to the concerns about debt traps, et cetera. So it seems like the relationship has matured or that's at least the narrative about it. So what is China doing an Africa today and I'll briefly walk you through loans and trade, which are oftentimes in the news. But then I also wanna talk about investments, which is what I'm focusing on and it's a smaller sector of Chinese involvement, but growing. And then I'll also talk very briefly about China's military presence in Africa. So China's loan strategy in Africa is very much based on this idea of state capitalism, that you need to take a lot of money into your hands first, and invest in infrastructure and sort of get in the early stages of development, kind of the state to make growth happen. And as we know, African countries sometimes have shortages of money. So these loans serve that they lend to these countries now so that they can build infrastructure which is pre-stage of industrialization, in order to make growth and development happen. And just to give you an idea of how big these loans are today, we see that really, the loans from the China Exim Bank are already competing with World Bank loans. So the same magnitude that we're talking about today. But there's a big difference between World Bank and Chinese loans, namely that World Bank loans come with strings attached. So this is based on the structural adjustment programs today so we have austerity measures, and if you're expected to have fiscal prudence, where Chinese loans really have no strings attached. And a lot of people are concerned about this because if this money gets into the hands of politicians, they could misuse this, could also be a way of China sort of making African countries more dependent on itself. Or it could also be kind of China thinking long term, it gives loans to countries that can't pay back these loans. And then it can take over, let's say ports as it happened in Sri Lanka in 2017. And we'll talk about this a little bit later whether these concerns are justified. So China also gives us these loans as part of the Belt and Road initiative. And this initiative as I said, was launched in 2013. And it is part of China's going out and going global strategy. And China right now is building trade infrastructure in Eurasia. And it's really meant to build trade routes in Eurasia, that is going to compete with the Transatlantic trade route that is currently being dominated by the United States. So here you again see this confidence that China has in its own development approach. And what's interesting, China's already Africa's most important trading partner. And you can look at these graphs, I'm gonna post them up. But really, the essence of this is that China is exporting more to Africa than Africa is exporting to China. And that was also one of the big topics that the FOCAC of 2018 that China was asked by African countries to also import more from Africa. So as I said, loans and trade kind of dominating the headlines, but Chinese investments in Africa, also a growing sector. So you see that the United States and the former colonial powers is still really leading here. But if you look at a change of five years, they're pretty stable. But when you look at China, this is really rapidly increasing. And I really like this graph here, which is the comparison between US and Chinese investment flows to Africa. And when you look at the blue line here, I think it shows very nicely that you have an upward sloping trend line until the financial crisis and then under the Obama area, those flows really drop. And now if we could extend this to the Trump area, Trump he focuses more, he has more military concerns than kind of foreign direct investment priorities. But when we look at China here, you see an upward sloping trend line and there's one outlier. That's when China bought 20% of the shares of the Standard Bank in South Africa. But I think the essence of this graph is really external shocks really don't matter as much to investment flows of China, and also changes in political leadership, et cetera. So I think this shows that China intends to be in Africa for a longer, longer future. And we can't say really the same about the US, we don't know what's gonna happen. And here just to give you an idea, investments, Chinese investments in Africa, very diverse, and the biggest sectors are construction, transportation, manufacturing. And I think this really nicely fits into Sun's story of the Flying Geese Theory and the geese flying to Africa now. Where we have talked about the Flying Geese Theory a little bit when we talked about China. Capital basically moves to wherever labor is the cheapest, and we saw Japan industrialize in the '50s. And as labor prices increase there they moved to other areas in East Asia. And now because labor is increasing, labor prices are increasing so much in China. We see a lot of Chinese manufacturers moving to Africa and actually, I conducted a lot of interviews with Chinese manufacturers in Africa, in Nigeria, Tanzania, Kenya, other places that are really said, I'm here because this looks like China in the '70s. And it's really amazing labor is so cheap, land is so cheap. So officially, China's saying that it is interested in an economic and sort of cultural exchange with Africa. But people are getting more and more concerned also about China's military presence, because right now China is already the Security Council member that deploys the largest amount of troops under UN command in Africa. And what you need to know here is that African countries don't really like if soldiers from neighboring countries are deployed inside their countries because ethnically which is oftentimes cross national borders. So now the AU, the African Union and the UN might like to employ and have Chinese military stationed in these countries to avoid this problem. And then China also set up its first military base in Djibouti that also raised a lot of concerns and I have a little clip. - [Woman] A fleet of Chinese ships en route to the Horn of Africa on a mission to set up China's first overseas military base in Djibouti. China's is the base will be used to support peacekeeping and relief missions. (speaking foreign language) - [Translator] The establishment of the base will help Chinese vessels escort missions in the Gulf of Eden and Somalia and international obligations such as humanitarian rescue missions. It helps China to continue to make even greater contributions to the peace and stability of Africa and the rest of the world. - [Woman] Construction began on the base last year. Although small, Djibouti is strategically placed, sitting on a busy shipping route and a stable country in an otherwise volatile region. China now joins the ranks of several other nations with military bases in Djibouti, including the US, France and Japan. But China's new base and rapid modernization of its military is raising alarm bells for neighboring countries, particularly India. With concern this base could become another one of China's so called string of pearls, building up strategic military alliances and sea lanes in the region. - So I think we can all agree that whatever China's doing in Africa, it's a very wholistic strategy. And China has really increased its presence in Africa in the early 2000s. And it seems like it wants to stay for a near and medium future. So I think there are a couple of key questions that we should be asking on this relationship that really try to get at what are China's intentions but also what is in it for the African side. So first of all, going back to our central question from the beginning, is China a new colonialist? Or is this presence adding to local development on the ground in Africa? And are the final colonialists here by a foreign power that tries to take partial or complete political control over another nation, as well as engages in economic exploitation? Then the second central question is how does China's presence add to institutional development? So sort of more on the political side. Does it have good or bad governance? Thirdly, are debt trap claims justified? And these are all questions that really media and scholarship has been trying to analyze and answer. And I would like to throw in a fourth question, which is how much bargaining power do African governments have in their relationship with China? And how do they use it? Is there leverage on the African side to influence what this China-Africa story might look like? And so on this key question of China, is China new colonialist, China's officially running a five nos policy in Africa. So it's this five nos policy is very much based on its general policy of non intervention. So that's also why China is not supporting the responsibility to protect or to keep because it's basically saying every nation is responsible for its own population. And we're not going to intervene with anyone else's affairs, as we also don't want anyone to step into our domestic politics. That's also why China oftentimes vetoes sort of intervention. So resolutions at the UN, also with Syria for example. So here you see Xi Jinping talk about his five nos policy in Africa. (speaks foreign language) - [Translator] China values sincerity, friendship and equality in pursuing cooperation. The over 1.3 billion times people have been with the over 1.2 billion African people in pursuing a shared future. We respect Africa, we love Africa and support Africa. We follow a five no approach in our relations with Africa. No interference in African countries pursuit of development path fitting their national conditions. No interference in African countries internal affairs. No imposition of our will on African countries. No attachment of political strings to systems to Africa. And not seeking of selfish political gains in investment and financing cooperation with Africa. - So you see here China's official stance on this is, of course, we're not a new colonialist. We're not interested in kind of political control. We're just really as I said before, interested in economic and cultural exchange. But of course they're also more critical voices on this China-Africa relationship. And I brought you here a comic by a famous South African artist. And you see basically China coming in with those big machines, and it says China's industrial colonialism. And these machines are going against the local population. And here you also see, Africa's environment is also suffering from so the expansionary vision of China and Africa. And when you look at Western newspaper reports, there are also a lot of articles like on Chinese debt trap diplomacy again with those loans, economical cold war. Is it a new colonialism? Chinese investments bring racism and discrimination. And I also talked to locals that have told me that the Chinese they just take all the jobs. So it really, everyone is split about this topic. So I would like to have a poll now on sort of from your background knowledge on China and Africa, what do you have heard from the news, et cetera, who thinks that China is a new colonialist in Africa? Okay, that's a lot of people. I didn't expect that. So who thinks China's influence in Africa is really kind of they're actually better than what the former colonialists have done? Okay, yeah. And who was split on this topic? Yeah. So I think the vast majority thinks that China is the new colonialists. So I'll just give you my opinion on this issue. So I think if China is a new colonialist, which I actually doubt is, first of all, it is a very different kind of colonialism. Because as you've seen in that famous image, 53 out of 54 African countries actually deliberately selected into it. So it wasn't like China, just coming to Africa and kind of taking over political control. but it is they are offering deals and the Chinese side is can either accept them or not. And there are a lot of speeches from African leaders from this forum that talk very favorably about China and how their relationship is different with China than with the West because of this non intervention. We also talked about how altropy is is sort of Europe centric. And it could also lend itself to again, like colonialist aspirations, et cetera. So this is sort of an agreement here. But I think obviously, the Chinese government, the West and African leaders, they're self interested actors, and they're very irrational in sort of what narrative they put out there. So I think instead of listening to what different leaders are saying here, we should focus more on impact and the variables that we can actually measure of what China is doing in Africa. And I think there are three questions that are probably the most productive to focus on if we want to have a real assessment here of whether China is a new colonialist. The first one is, is China helping to grow and diversify African economies? And this one is really important from both an economic development perspective but also from an institutional one and also in terms of democratic one. First of all, economic one because we want to see with China's presence in Africa, are African countries that are mostly agricultural now, are they moving more towards industry? And from an institutional perspective, I know we started to risk it all in detail where they're all saying we can't really predict when a democracy, like when the transformation to a democracy becomes more likely. But what can we say about its survival? So we know that democracies are more likely to survive if they have a higher per capita income than $14,000 as well as they have a diversified economy. And then I want to ask two more questions, which are is China adding local human capital and technology? And how does the competition with local businesses look like? And these two questions are really inspired by China's own growth miracle, because when China opened up in 1978, it invited a lot of foreign investors and provided incentives, all of that, to come into specific industries. But what's very interesting about China is that it actually kept a pretty strict foreign investment catalog, where it laid out certain sectors that were completely restricted to foreign investment and others where foreign investment was allowed and encouraged but had caps on foreign equity that was allowed for a joint venture requirements and also requirements and how many foreigners were allowed on the boards of companies. So they're much I summarize, this is local participation requirements. And China was using them really as a magnifier effect because China was able to import the technology from abroad and get the money from abroad, but it built a second of local managers that could then run the show. And this is also one of the explanations as to why China was able to grow and develop so fast, because locals learned very much and very fast from foreigners in their countries. And before I dive into my case studies, Nigeria and Ethiopia, I just wanna point out that, this idea that there's One China Inc. Africa is also flawed. So I know we talk about China and Africa, now also use the term just for simplicity reasons. But through my research, I found that there actually at least three different types of Chinese companies in Africa that have very different incentives that are worth studying. So first of all, you have state owned enterprises, and they're very much linked to China's sort of state capitalism model. They are very dependent on the Chinese government's control. And oftentimes, they're both driven by an economic incentive as well as sort of a different diplomatic kind of relationship, let's say between China and Nigeria where these companies are allowed to operate at razor thin profit margins. If this is important enough for the strategic relationship between an African country and China. Then you also have big private firms like Huawei who come to Africa to really they're very profit driven. I've talked to CEOs from Huawei who really want to compete with Western companies in Africa. But they have such a high profile that they can't really hide from the Chinese government. And so they're still dependent sort of, on China's, the Chinese government's strategy in Africa, and they can just do everything they want, because China's actually very concerned about its image in Africa. So they have still have to obey by certain rules. And then there's a third category which I particularly find very interesting to study which is small, private firms. So how they come to Africa is in two ways. So sometimes you have Chinese state owned enterprises being sent to Africa. And then employees in the state owned enterprises are realizing, okay, we're not really making any money. So individual entrepreneurs, they venture off and they open a small company in Africa to really make more money than that sit on the enterprise. Or we have some small Chinese private companies already in Africa that do have friends back home in China. And they will tell them all like land here is really great. Like here, the taxes really low, et cetera. So you should come from China to here. And they keep a very low profile, and they hide a lot from the Chinese government as well. But what I would also like to add is the Chinese government doesn't really care if they hide or not as long as you're not big enough to really matter. And if you make an impression, if like your company screws up and it's gonna be in the news and stuff that could be could influence this image that China has in Africa. So I also talked a lot of two or three types. And as I said, they all have very different incentives. And that matters for the stories that I'm going to tell you now. So let's first focus on Nigeria. So Nigeria as a country in West Africa. And just to give you a little bit of background on Nigeria, it has the largest population in Africa, and also one of the largest youth populations in the world behind China and India. And it's a former British colony, and it went through a military period, military regime period. And it's been formally democratic since 2015, when Jonathan Goodluck deliberately stepped down to Muhammadu Buhari and democratic elections. And Muhammadu Buhari he just ran for a second term in February and I was actually in the country during that time. So it was very interesting to observe. So what's really interesting about Nigeria is that in the '70s and early '80s, it actually had the largest textile industry in Africa. So we had some diversification of the economy happening. But when we fast forward to today, 80% of all finished consumer goods are imported. And really, the economy is not diversified at all. So what happened there? So in 1958, Nigeria discovered oil, and by the 70s, this was really the largest sector of the economy. And I think that led to really three outcomes. So first of all, as we talked with the Soviet Union, if there's one dominant sector in a country, the political elites really have an interest in keeping that one sector dominant because they can grant access to the sector and they can make a lot of money versus if you try to diversify the economy and try to provide alternatives to the population, they would not pay you anymore to get access to the sector. Secondly, there's also the Dutch disease and basically as your revenues increase in one particular sector, your currency appreciates and that makes exports from any other sector really expensive. So Norway is also an oil rich country, has tried to circumvent this issue by when it gets the revenues for its oil, it doesn't even convert them into its own currency but puts them into funds again abroad. So because Nigeria has all this oil, its GDP per capita is pretty high, but we have very high inequality. And it is one of the top receivers of Chinese investments. And those are not just investments in natural resources. But I also talked to a lot of manufacturers there, people in the service sector so it's a very attractive market because you have a huge population, very cheap land and very cheap labor. And what's important to note about Nigeria as well finally is that it really adapted the structural adjustment programs and went with it. So in 1995 it allowed 100% foreign ownership in all sectors. And really, if you're a foreign investor in Nigeria, you can basically do whatever you want. So you can compete with locals, you can make money there and you can export it abroad. So you can really have a lot of leeway there. So let's look at growth and diversification, as I said like that is one of our key questions, is China's presence adding to growth and diversification? And here you see that Nigeria actually just emerged out of a recession. And that was because of a drop in global oil prices and output chalking, Nigerian and advocate policy to really respond to this. And here you see how much Nigeria is, is really dependent on the oil industry, because over the past 20 years, the value added from industry actually declines so we don't see any diversification happening over the past 20 years unfortunately. So let's hone in a little bit on how China's presence is really adding maybe something to diversification and local participation in Nigeria. And what's interesting to note here is that in 2006, at another form of China-Africa cooperation, China announced that it wants to build eight special economic zones in Africa. And that was really based on this idea that China used special economic zones in the beginning of its opening up to really experiment with capitalism. And so they suggested kind of as a policy strategy, you should also build these zones and have tax incentives, tax holidays, and let's see what happens but the difference is that these zones are much smaller than the ones in China. So that's why they might not be working effectively. But Nigeria got two zones during this bidding process, the Lekki Free Trade Zone and the Ogun Guangdong Zone, and both are very much top down joint venture, joint venture project between state governments in Nigeria, as well as Chinese state owned enterprises. And I will show you a video now from the Lekki Free Trade Zone. - For many African countries, attracting foreign investments has always been a challenge. And now it's become even more difficult with an economic downturn in Europe and the United States. So here's how some African countries have responded. They set up what are known as free trade zones. These are places where multinationals can come in, set up shop and in some cases, a little or sometimes no tax at all. It really does beg the question, what's in it for the ordinary African. Mark Edwards been checking up on that in Nigeria, Africa's most populous nation. - [Mark] In Lagos they hope this will be the road to economic success. The Lekki Free Trade Zone is being carved out of thousands of hectares of untamed land near the coast of Nigeria's commercial capital. - The first part is 3,000 hectares. - [Mark] They call Sola Oworu the Empress of the zone. This former city banker is in charge of creating what she believes will become the Dubai of West Africa. - The idea is actually to create a new model city, not just industrial, also a residential, commercial area, so that you have a self sustaining city. - [Mark] While Lagos State providing the land, a Chinese consortium is responsible for the actual construction. There's a lot to build. The zone will have its own international airport, deep sea port, and its own water, power and sewage systems. - The idea is to have the zone as a catalyst for economic development in Lagos State, diversify the economy, we would want to attract foreign direct investments into Lagos State. - There are 23 other Free Trade Zones in Nigeria, all vying for the attentions of international investors. But the Lekki Free Trade Zone here in Lagos has one big advantage, size. At 16,500 hectares it could become the biggest in Africa, four times the size of Manhattan in New York. - So this is a video from 2013. And what's striking here is that you can't really see the zone itself, right? So I tried to find videos where you have some video footage, what it actually looks like, but I couldn't so I would have liked to show you. So I was wondering about that myself. So I went to the zone and I drove there and when you arrive there's a really nice building in front with the administration and like they welcomed me and I could talk to the MD, the managing director on the Nigerian side as well as on the Chinese side and everyone said, yeah, the zone is gonna be amazing and like all of that potential and it really is going super well. And then a team of engineers took me around the zone. So we came behind the administration building, and what you could really see was not that much, because there wasn't. So there was some basic infrastructure built. But I was driven around and I talked to a few companies that were really only 14 fully operating Chinese companies in the zone. And that was after 13 years of being built, right? So I think it lags a little bit behind expectations. And one of the reasons I see is really, because it's more of a top down joint venture between two governments. So as I said before, you have Chinese state owned enterprise involved. And it is kind of bound to this diplomatic relationship between Nigeria and China and it is a prestige project. So that's why they build a lot of infrastructure there. But how to make the zone actually work is a different story. It is a very profit driven story. And when I saw kind of the management building and then like the office is there, they're very comfortable in that position, I don't really see that big of an incentive to go out of your way to really attract all of these companies. So while there's tax holidays, so no taxes inside the zone, there's a very high management fee actually. And I talked to companies that had considered moving into Lekki Free Trade Zone, but they didn't find it profitable because they were too small for the management fee to make sense. And finally, they were also locals that have been protesting because of resettlements. 'Cause that used to be their land who have been protesting a lot. And in 2012 actually, one of the managers of the zone got killed. So they'll have problems to make the zone really, really work and profitable. And we have a similar story for Ogun Guangdong. So now though I would say it is performing slightly better. So it was started with a Chinese state owned enterprise. And then they were caught up in some smuggling problems. And then a Chinese company actually took over the zone, the management of the zones. And they were more interested in kind of making the zone work because they were more profit driven. But then the Chinese government kicked them out, I believe two years ago and put in again, a Chinese Special Economic, SOE, State Owned Enterprise, because they felt like they had more control over what that company could be doing with this prestige project, which meant a lot for the diplomatic relationship between Nigeria and China. And there's a very interesting story that I would like to share with you. Professor Shapiro and I, we both separately, interviewed the private management at that time of the zone. And they told us they had one big problem, which was someone was blowing up the gas pipelines that were coming to the zone. So who do you think was blowing up these pipelines? Yes. No, it was not terrorists. Yeah. So what's really interesting is that you actually had locals blowing up these gas pipelines because they started sort of as a local competition, or like as a foreign project that they weren't really being able to have any advantage of. So again, have locals seeing the zone and like seeing the infrastructure, but not really being able to participate. And as I said like in Nigeria, you can kind of just come in as a foreign company and you don't have to employ locals or train locals or have them on or kind of give them ownership. So I've heard a lot of stories of competition between Chinese companies and Nigerian ones. And obviously competition is good for kind of growth. But there's also the flip side of local inclusion. So I guess that is the trade off here. And I know these stories are just anecdotal. So I have like, for example, in the bike ride hailing services, and there's a really aggressive Chinese company right now or in the FinTech industry, whether a lot of Chinese companies Offering loans to locals, just competing on price alone. And I'm going to conduct a survey experiment with Chinese companies next semester in Nigeria to kind of get more at the trend here. So I don't only wanna talk about negative things about Nigeria, so they also some positive developments here. So for example, Aliko Dangote, he's the richest African right now. And what you need to know about Nigeria is that it exports all of its crude oil and imports the refined oil, because it's not doing the refining itself domestically. So Aliko Dangote is now currently building an oil gas refinery, in Lekki Free Trade Zone to solve this problem. And I talked to locals and they generally actually feel positive about it because they're excited that this is now finally that there's some domestic refinery here. I would be a little bit cautious because 20 years ago, Nigeria had this law that not one private businessman could own that much oil in that industry just because of national security concerns, right? And I think Aliko Dangote is getting a lot of political power because of this now, because he controls, he can just stop the oil immediately. He is also, I heard he wants to also sell it at kind of international prices. And if the government doesn't really comply with what he wants, he can just easily export the oil as well, because he's in the Lekki Free Trade Zone where they have an airport and a seaport. But what I'm more excited about are these small Chinese industrial parks. And I know you also had to read an article on this. Where through an interview, a local Nigerian broker who work with Chinese companies, he invited me to his village and he said, I live in this village where the local chief he set up an industrial park with Chinese and I should come visit. So I drove there for a weekend and I interviewed all the Chinese companies there and I also talked to the local chief. He invited me for tea and we just sat and chatted. And this looks like much more like a bottom up sort of win-win situation where in the '90s, a Chinese businessman, a private one came to that village and he wanted to do some fishing there. So his people they built two houses for the village chief, for his palace. And kind of as an exchange, he was allowed to fish there and then he brought more and more people, small, private Chinese firms and they build this industrial park. And they prefer being in that small industrial park to being in the big ones because it's located on a very strategic street that goes from the south to the north and they produce mattresses and slippers and all of that and they actually also get tax benefits from the state government because the village chief talks to the state government and they also employ locals and have them not as office managers, but at least as floor managers. So there seems to be more promise here. So I wanna contrast the Nigerian case with the Ethiopian one and Ethiopia is a country in East Africa. And I'm comparing it because it is the second most populous country in Africa. So similar in size, it has never been colonized. Except for a short period of where the Italians occupied Ethiopia during the Second World War. It went through a communist era. But then what's really interesting is starting in the 1990s, where you have an authoritarian regime, arising the Ethiopian People's Revolutionary Democratic Front. And Ethiopia also went through structural adjustment programs, but they didn't apply them one to one. They sort of like China that also went through structural adjustment, took the pieces that they felt like that would work in their country and left everything else out. So in 2002, they adopted an industrial policy strategy. And that transformed into two growth and transformations plans that are very much based on the Chinese and Korean developmental models. And when we look at Western investment flows to Ethiopia, after 2003, they have like really declined because there was sort of, the Western companies weren't able to just operate in this market anymore, because like in China it was more strategic, like some sectors were closed, some are open to foreign investment. But Chinese investments stocks meanwhile, rose a lot. And it's already ranked 11th for amount of Chinese investment stocks in Africa. And I expect it's gonna join the top 10 very soon. So when we think about growth and diversification in Ethiopia, this is a much more positive story. Although Ethiopia is also starting from a much lower level, of course, and it's much poorer than Nigeria, because it doesn't have this oil wealth. So now Ethiopia is actually the fastest growing economy in Africa right now. And it's in the news for that a lot. When we look at industrialization, this graph looks very positive, but actually when you compare over the past 20 years, the shelf industry has only grown by I believe it's 11%. And most of it is coming from construction, which is kind of the pre-stage for industrialization. So it looks promising, but we aren't there yet. And actually, we see a lot of movement from the agricultural sector towards services. So Ethiopia looks more like India now, than Korea or China, although it has this industrialization policy so. What are some positive developments here by the Ethiopian government that were implemented? So first of all, Ethiopia has really tried to ramp up primary education enrollment over the past 10 years, and it isn't that 100% now so obviously, this isn't the perfect measure of quality education, but it's one measure to look at sort of human capital stocks and then how there's human capital growing in the economy. And as I said Ethiopia is also very strategic and has a local content initiative where some sectors are completely closed off to foreign investors. In some sectors, they are open but there are a lot of incentives for joint ventures. And even informal sort of deals. Investors are always expected to come up with a strategy and a proposal of how locals are going to benefit from this investment that is being made. However, there's the study by Blattman and Dercon, which I found very interesting. They tracked industrial workers in five different factories in Ethiopia, and actually 70% of them quit after one year. So again like there's potential in Ethiopia for industrialization, but it seems like there are more options still in the agricultural and informal sectors and people aren't really picking up these industrial jobs yet on or running with them. So I do think industrialization is happening in Ethiopia right now. But locals should be incentivized to pick up these jobs and stay for longer and they were also health concerns. So there were a bunch several blocks to move out of the way. And here's the Eastern Industry Park. And this is also one of these eight Chinese special economic zones that were placed in Africa, after China announced they want to build them in 2006. But this is a very, very different approach to what we saw in Nigeria, because this is actually a bottom up, very private initiative. So it's a private Park. And there was one Chinese businessman that came to Ethiopia in 2006. And he wanted to build a pipe factory. But then he heard about these biddings for these special economic zones. And he applied for one and he brought one of them to Ethiopia. And in the beginning, the Ethiopian government was very much against it, it was like very skeptical of what was happening there. But this private Chinese businessman, he brought government officials from Ethiopia to see the zones in China. And over a while the Ethiopian government really got into this idea of special economic zones. And in 2015, they even implemented industrial zone framework and they have four more zones planned now here and there are 83 Chinese companies operating in the zone already, although it's much smaller than the ones in Nigeria. So you have some footage from there. - [Girum] If you look for some of the best China based companies operating in Africa, chances are you'll find most in this humongous industry zone. Located about 30 kilometers outside of Ethiopia's capital, Addis Ababa, the Eastern Industry Zone is a place for the multi faceted industries from China. - This industry is first industry in Ethiopia. And this industry is a wide stage for the foreign investor. And this one is we get support from both of Ethiopia and China government, because they know in China government, we have the experience, how to build these industries, and how important to attract the foreign investor. Also now Ethiopia, they want to develop the industry. And now they know. - [Girum] Since its inception seven years ago, the Eastern Industry Zone has been showcasing Chinese industrial developments positive impact here. The zone now houses best Chinese manufacturers such as Huazhan Shoes and Lifan Motors. Here there is already a conducive environment for any investor to operate. And that according to the management of the Eastern Industry Zone is key. - Now Ethiopian they give good condition. They support us like one stop surveys, like the customers credit lines, inside industries, they give good service very easy. And the here is peaceful. - Now the future equally matters for Ethiopia's manufacturing sector. There is already a big ambition to make the nation Africa's manufacturing hub. For that Ethiopia eyes the new Belt and Road Initiative to make all the dreams a reality. - So actually there have been a few academic studies done on the zone. And what they found was that actually companies are training locals. Because as I said, the Ethiopian government sort of informally requires that you come up with a proposal on how locals benefit. It isn't like for the entire zone, that there's one sort of strategy so I think that could be one improvement here, but at least you see some development. That's the locals are being trained and that local human stocks of capital are going up. So I wanna make a final point here on sort of what's in it for foreign investors on obeying with local participation requirements. And I wanna talk about the state of emergency in Ethiopia. I was actually there during that time. And basically the state of emergency broke out after the Ethiopian government wanted to expand the borders of Addis Ababa, the capital to the south to have like an industrial kind of area there and protests broke out in the Oromo and Amhara regions, and it cost them ethnic tensions, and a lot of protests and arrests, et cetera. And during these protests, locals really went to foreign companies and destroyed them because they felt like they weren't part of this growth story that Ethiopia was telling them and also you could hear in Western media. So I had a very interesting story from an interviewee there that there was Aliko Dangote, which is the richest man of Africa had one factory there. And right next to it was a Dutch South African factory. And what's really interesting is the protesters walked around the Dutch South African factory and destroyed Aliko Dangote's factory. Why? Because the Dutch South African factory, they employed a lot of locals, also gave them shares from the company. It was perceived as a local project. So in some sense, if you incorporate kind of local partners, it also mediates political risks. So let's now turn over to the lessons that we can draw from these two case studies. So first of all I think the central lesson is that deal conditions matter for development outcomes. And going back to this first question that we had in beginning, like is China new colonialist? Or can it sort of put its model in Africa and then now Africa is gonna rise, I really think that this is more about African side, strategizing rather than the Chinese dictating any model, and you see that in different sort of policy environments. The Chinese are gonna act very differently as international actors are gonna act differently as well. So I think it's more sort of African agency that is going to determine whether this is a positive or a negative China-Africa story. I also doubt that there's one China-Africa story, because we have 54 different African countries, but very different policies. And the conventional wisdom here is actually on these two conditions that African countries can't set them because they have to be deal takers in order to attract any kind of foreign capital. And this is very much this Washington Consensus Narrative, the Poverty Trap Narrative, that the region historically has not been able to attract significant amounts of investment. So you have to make it as easy for foreign investors as possible to kind of attract them. This is because Africa's population is young and under educated, low stocks of human capital poor infrastructure and political risk, sort of what Jeffrey Sachs laid out in his piece. And even with China in the picture, the African countries still have no exit options to the Chinese offer, right? So it's kind of this idea of whatever the Chinese offering, you have to accept, because there's no one else who will step in and build this for you. So as I said, I conducted 10 months of field research in Africa, kind of to try to find evidence for this. And now based on the evidence that I found, I highly doubt this entire narrative. And I say that paradoxically, I believe African countries actually have significant bargaining power with China. And I think this is fundamentally based on the idea that Africa doesn't just have all these weaknesses, that the poverty trap list here, but in fact, Africa is the region in the world with the most amount of natural resources. The cheapest labor in the world and the cheapest land. And in contrast to the West, China has both willingness and capacity right now at its developmental stage, to put its money into Africa and to really get returns from these factors that look very attractive and lucrative to China here. So completely, China has massive capital surpluses as I said and it needs raw material to fuel its growth. Africa is the last region to industrialize. So as I said, I interviewed a lot of Chinese manufacturers that are really excited about the African markets, because it's so much cheaper there to run a factory now. Africa also has export deals with the UN, the US. So actually, when they want to export goods, they don't have to pay any tariffs. So as a Chinese if you produce in Africa, it is cheaper for you and in a lot of cases. What's also interesting, I want to focus data in and I interviewed someone who is still in Nigeria and the President said that in 2016, China increased its environmental regulations in symmetric market. And in that year, you saw a lot of Chinese companies that did not want to obey with these environmental regulations to Africa because, for example Nigeria doesn't have these strict government regulations. And finally, Chinese companies also seem to be less risk averse to political instability. So you also have a lot of Chinese investments in South Sudan, for example, which is not politically very stable. So I think that Western investors are much more sensitive, they also stay to these kind of political and safety concerns. They stay a lot of urban areas, where you see a lot of Chinese in rural areas. And I think that what's the most important is that certain concerns with Western investment that limit African bargaining power, are less alien with Chinese one. So first of all, you could argue that because China is offering all these loans to Africa, it makes these African countries very dependent on China. But actually Deborah Brautigam, who is one of the leading scholars here, she wrote a lot on this and saying, the magnitude of this is really overstated. And then you see countries like Ethiopia who have taken out much more debt with China actually being more aggressive with China than others which are smaller debtors like Nigeria. Secondly, China is also really trying to compete with the West. And the number of allies matters. Also at the UN we see that UN General Assembly, these African countries on certain things they vote with China. And if you're Ghana or if you're Nigeria, you have the same vote, it's one country, one vote. So that's why China's interested in having a relationship with all these countries and economic relationship and there is the loan, trade and investment. And I think this is very important for the central point here that these African countries, there's less of a need for intercontinental competition for Chinese investments because it seems less rivalrous than what's coming from the West. And even if China's the only sort of investor here, China oftentimes prefers to be in your economy, than not be in your economy. Even if they have to adapt slightly to policies that countries imposing and you see this year, you see investment stocks from the US where like you have maybe five countries where the top receivers and the rest is getting very little. And there's a lot of competition here for US investment. But when you look at Chinese investment stocks in Africa, the first ones surely to get more, but then you have a much more even distribution of those investment pies, and you have 51 African countries that actually receive an investment pie. And yes, that is dependent on sort of China's political and economic considerations. But I run some regressions and it is very independent of whether African countries are imposing local participation requirements or not. And let's take one of these investment pies. Let's say Zimbabwe. And let's see what a policy shock does to these investments. So in 2008, Zimbabwe implemented this Indigenisation and Economic Empowerment Act, where all companies inside Zimbabwe had to have 51% of local ownership. And when we look at US investment flows here, investment stocks in Zimbabwe, you see that after 2006, they were not even counted anymore because they really decreased a lot. And I found a lot of qualitative evidence from Western companies here, complaining about 51% local ownership all of that and it just made Zimbabwe very, very uncompetitive compared to African countries. And of course, you also have the financial crisis hit here in 2008. That also is part of the story. But when you look at Chinese investment stocks here, they really don't seem to be bothered by the 51% local ownership here where you see starting in 2010, these stocks really skyrocket. So this is puzzling and then when we look at Algeria, we have the exact same policy change also in 2008, with 50% of local ownership required with companies. And here you see, starting in 2007, when these were announced it really like the curve really gets much more flat versus Chinese investment stocks. They just you have that upward sloping trend line regardless. And I also conducted a lot of interviews with Chinese businesses and Western businesses, try and get at these sensitivities, to local participation requirements. And really, across the board, when I asked the Chinese they were just like, honestly, like as I said, this is China in the '70s and '80s. And it's great here, I would stay under any kind of policy regulations. Versus when you ask a Western investor, they were like no, like if I had to employ locals, I would think about it twice. And I found this quote from Algeria as well. So the question is, why do some countries like Nigeria, Kenya, negotiate weaker deals in the sense of like, diversification issues and local participation, weaker deals with China than others like Ethiopia and Tanzania? And I think common explanations here are, for example corruption that countries know that they could be asking for more from the Chinese investors, but they aren't because politicians like to just pocket some money and let them do whatever they want. Or another explanation is ideology so that they know they could be imposing these restrictions and the Chinese would be doing it but let's just keep the economy open because we really support capitalism and let market forces do their job. Or capacity where the government knows they could be asking for more but they don't have trained or skilled negotiators that can face sort of the Chinese negotiators that have more experience. But I would like to bring up a new explanation here, which is the government's perceptions on their own bargaining power matter. And why could this be one of the explanations? I mean, it isn't exactly obvious that these economies do have bargaining power, right? Like we just talked about all the weaknesses and how the Washington Consensus and the Poverty Trap Narrative really suggests that these countries are negotiating from a position of weakness. So I conducted interviews, especially very, very intensively in Nigeria, to understand how the Nigerian government thinks about Chinese companies and how Chinese companies think about the Nigerian government in terms of kind of policy sensitivity and exit thresholds and of what would be the policies that would drive the Chinese out. And the Nigerian government I interviewed national ministers who told me honestly, we have no bargaining power here. The Chinese and the West, they're bringing all the money and we have nothing to offer. So we kind of just have let them do whatever they want to, versus with the Chinese companies, as I said, I got a lot of them just said, you know, I would do everything to be able to stay here. And in other countries like Tanzania, there was much more of this awareness, this China really needs us. And then Ethiopia I've heard that a lot like we are really pulling the Chinese in here right now because we have cheap labor and we have cheap land. And this is exactly what China is looking for right now. So what are the factors that influence African governments' perceptions on bargaining power? And I tried to understand the political economy behind this. So first of all, I think the historic relationship to the West versus the East matters to where I said, Nigeria was a British colony for a very long time. And we went through the Washington Consensus and really adapted what Western investors were telling Nigeria at that time, versus Ethiopia had never been colonized. Tanzania also has a very long historical relationship with China. They've looked very much more towards the East also during the Cold War. There is also federalism and ministry constellation where in Nigeria, a lot of different states, they actually compete for investments. So they're not likely to share information, how much Chinese investment is actually coming in. And a lot of them really underestimated Nigeria's position and attractiveness to China. And it's the same with ministry constellation, where in Nigeria you have a lot of ministries sort of competing for these deals, and there's not one collaborative effort versus when you have one sort of very centralized government. That government has all the information on sort of what investment comes in. And it's also overtime learning if you have one party in power for a long time, they might have tested out kind of what can we asked for also more informal deals. And then finally individual agency also matters, where I've just met people, for example in Tanzania, someone in a very senior position in the ministry. He was educated in Japan, and just really studied industrial policy through and through. And he just knew, he was very confident that China was going to obey by these rules if Tanzania asked for them. And Tanzania also has similar policies to Ethiopia. So what are some policy implications here? I just think incentives are more effective than requirements. Mostly because when you look at these different graphs from the West and China, you're trying to attract two different kinds and types of investors. So if you offer like a menu like Ethiopia has here, different types of investors can pick whatever they like the most. So if I care more about low taxes, maybe I'm more willing to have a joint venture and so forth. So you can kind of mix up different policies and offer a different menu to different types of investors that can pick whatever they wanna do. I think because these countries also have weak formal policies, it leaves room for a case-by-case negotiations as we see, for example, with these industrial parks that I talked about. The hidden industrial parks of the villages that kind of set up a deal that work for all sides. And you should also account for enter and exit thresholds where I think the decision of companies to enter your country, are probably, they're more sensitive to kind of your policy environment rather than if you're already in the country. And then certainly, you have to pay higher taxes. So your decision to exit because you already set up kind of a factory, et cetera. It's gonna be less sensitive than if you still need to put in all the capital in the beginning. So what are some expectations for the future? So I think there's certain things to consider on the Chinese side. Which is can China maintain its growth but I also heard that because China's growth is slowing now all these Chinese manufacturers are moving to Africa in search for new opportunities, and they feel like the African region is the next one to industrialize. So it's kind of, literature here is not clear whether China slowing its growth, is actually good or bad for the African region. How its rivalry will play out with the US. And what about automation and the flying geese. So I think that's something we should be concerned about that if labor is being replaced with machines, maybe the cheap labor in Africa is gonna be less needed. So factors on the African side, I think the biggest question is, can countries start to use their bargaining power and negotiate stronger deals? And just coming up with a strategy, I think here is essential, rather than just leaving the entire economy open and letting investors do whatever they want. Sort of what China did when it was poor and agricultural. It did come up with a strategy that propelled up the economy. And I think this is interesting on debts and trade, because as I said, at the former China-Africa cooperation in 2008, African countries already complained that there was an in trade imbalance between China and Africa and that China should also import more from the African region. So it seems like around debts and trade, there is more of an awareness that this relationship should be more equal. And I hope that investments also are going into the right direction. So just to close, I brought you this comparison here again, which shows very nicely, China's economic growth miracle. And what I find really interesting here is that in the '80s, it was really the West investing in China and doing a lot of trade with China, but we still don't call it the West in China, how we call China in Africa today. And this is because China took some lessons from other countries in the past, like some of the structural adjustment, it's liked and implemented others, it came up with the zones model, which was the state capitalism. And I think similarly, the African region also can just copy and paste one model from China now or from the West, but it should strategize and see sort of what fits to local realities and the common narrative here has always been that the African region can't afford to strategize because otherwise you can't really attract foreign investment to grow. But I think because Africa is the region that has the highest stocks of raw material in the world, the cheapest labor right now and the cheapest land, I do think that there's some leverage here for the African region. And hopefully in the future, we can recolor this map and the African region can catch up with China. Thank you. (audience clapping) (soft music) |
Power_and_Politics_in_Todays_World | Lecture_4_Fusing_Capitalist_Economics_with_Communist_Politics_China_and_Vietnam.txt | - So today's agenda, we're gonna start with why am I talking China and Vietnam. We're then gonna talk about reform in China in the period leading up to the 4th of June, 1989 which is the Tiananmen Square massacre. We'll then talk about Tiananmen in the 1990s that will lead us into a discussion of what I'm calling the sequencing debate of political and economic reform. Once it was clear that communism was gonna be replaced by capitalism, there was a huge debate about, well, is it better to have political reform first or economic reform first. We'll try to do them together and we'll talk about that sequencing debate which will lead us into a larger discussion of what since the 1950s has been known as modernization theory. The thesis that economic modernization will eventually produce demand for and the establishment of democracy and that will then leave us to think about the future. So, China and Vietnam Today. - [Narrator] China so far has built the equivalent of Europe's entire housing stock in just 15 years. In November 2015, Beijing replaced the substantially larger 1300 ton Sanyuan Bridge in just 43 hours. Between 1996 and 2016, China has built 2.6 million miles of roads including 70,000 miles of highways connecting 95% of the country's villages and overtaking the US as the country with the most extensive highway system by almost 50%. Over the past decade, China has constructed the world's longest high speed rail network. 12,000 miles of rail lines that carry passengers between cities, at speeds up to 180 miles per hour. China now has more high speed rail tracks than the rest of the world combined. - So that's one of any number of video clips one could pick to just give a snapshot of the incredible transformation of the Chinese economy over the last couple of decades and indeed, over the last decade, I went to Beijing last year for the first time in about 12 years and 12 years ago, there were lots of potholes in the streets, there was still rickshaws on the highways. You go to Beijing today, it's just miles and miles of skyscrapers and every conceivable multinational is there. It looks like just West Los Angeles times a hundred. So you can't overstate and I will put some statistics up in a little while, but the economic explosion in China since the 1970s has been phenomenal. And then here's another snapshot versus Vietnam. - If you're talking about the 80s what Vietnam look like. You go to cities, you didn't see the light, electricity cut, no car in the street. Vietnam was so isolated from the world. - [Narrator] With the Doi Moi economic reforms of 1986, the government hoped to move toward what is now called a socialist-oriented market economy. When the Vietnamese government asked for hands-on advice, the IMF did not refuse the call. (speaking in foreign language) - If you use the analogy of Vietnam's economy as a eagle flying up, there are two wings. One is internal institutional reform, and the other is opening to trade. - [Narrator] Free to start their own companies, the Vietnamese people created more than 30,000 private businesses in the first 10 years, spurring growth that averaged 7% throughout the 1990s. - If you look back in the late 80s and early 90s we had the first FDI law ever which opened up the opportunity for newly established private firms to engage international trade. - And that was complemented with a major push to attract foreign direct investment into Vietnam. - [Narrator] The economy has grown at about 7% per year since 1990, second globally only to China. - So these are the two most successful examples of post communist systems that have remained authoritarian but have been transforming their economies into capitalist arrangements. And that differentiates them from the other post-communist systems that have retained authoritarian rule if we think about North Korea, or we think about Cuba, they have not transformed their economies. So this is one thing that differentiates them from the other post-communist authoritarian systems, post-communist economies retaining communist authoritarian political systems. And they account for a large part of the economic success in the world in the last decade just to take one or two decades. Just to take one example, one of the encouraging stories of our time is if you if you take the long view of things, number of people and even the percentage of the population both living in extreme poverty has really gone down over the past two centuries and much of that decline has come in the last few decades. Of course, this is a worldwide indicator. But if you look at what is accounting for most of this decline, you can see that the main action has been in East Asia. So the this is the blue line up here where poverty again, this is number of people living on a less than $1.25 a day, which is a World Bank's definition of extreme poverty. So, most of the changes of course, have come in the developing world. But if you compare East Asia, for example, to Africa, East Asia is by far the lion's share of the reduction in extreme poverty. Now of course, not all of that is the is China and Vietnam. There're countries like South Korea that have massively improved. There are other countries in that part of the world not to mention Singapore, where poverty rates have decreased massively. Nonetheless, China and Vietnam are relatively impressive stories on this front as well. This is a picture of reductions in poverty in Vietnam and if I'd had gone back before 2010, the numbers would have been higher up there as well. But you can see, even in this last difficult decade, and despite the fact that ethnic minorities within Vietnam are not doing as well as as others, we're seeing very low rates of people living in poverty in Vietnam. China, you can see the head count living below the official poverty line is the red line. The head count below 1.9 US dollars a day is the green line. Again, you can see massive reductions in poverty by many estimates they vary, but many estimates say that since the reforms in China ushered in in the late 1970s, somewhere between 400 million and half a billion people have moved out of poverty. Now, there's another line on this chart, which I don't believe, which is the Gini coefficient which is a measure of inequality and I believe that the inequality in China has gone up a good deal more. If you look at here, this is comparing the top and bottom quintile. But again, this is data provided by the Chinese government, you can see that in terms of per capita income, they're showing an increase inequality from three times to about eight times and land and consumption, less of a change. So I would just say two things about these statistics, these are based on Chinese government statistics. My guess is that the inequality in Chinese is a lot higher, and particularly this is comparing quintiles that's the top 20%. But just as in the US, and just as in many other capitalist countries that we'll talk about later in the course, if you look at the top 5%, the top 1%, or the top .1%, you're gonna see massive increases in inequality and that puts on the table a dilemma that's gonna be worth thinking about as we progress through the course and talk about some other countries. Which is that while inequality within countries just about anywhere in the world has gone up over the last four decades, worldwide, global inequality has actually come down, and the number of people living in poverty has come down dramatically as we've seen. But if you read say the Millennium Goals or you go to the World Economic Forum meetings, which I have done from time to time, you'll see politician after politician and policy wonk after policy wonk standing there saying, "We need policies to reduce poverty and inequality." It's like almost one word, poverty and inequality is one word, but what we're seeing here is something that' gonna come back to us in multiple ways during the course. It's not obvious that all good things go together. And it may well be the case that the price of reducing inequality, the price of reducing poverty is tolerating significant increases in inequality, we'll have to think about those trade offs in various settings. In any event in both China and in Vietnam, you've seen massive decreases in poverty, big increases inequality in not as high by most measures in Vietnam again, now it's hard to know what statistics to believe on these matters but most scholars seem to think there's less inequality in Vietnam than in China. So the first answer why are we talking about China and Vietnam today is that by virtually any measure they have been economically successful. But then the second and in some ways more interesting reason we're gonna look at them is they challenge a lot of accepted wisdom about capitalism, democracy and the relations between them because here we have two very repressive authoritarian systems, transforming their economies into successful capitalist ones. So that's the puzzle we're gonna be wrestling with for the rest of today's class in one form or another. So let's think about what was happening in China before Tiananmen. After Mao's death in 1978, things were gonna change very rapidly, just to give you a sense, up until the end of the Cultural Revolution in China, this is a picture comparing Chinese and Japanese per capita GDP. And as you can see, this is a period of explosive economic growth in Japan after World War Two, of course, from a very low base because the whole economy had been destroyed. When people talk about the Japanese miracle, it's a little bit misleading. If we took this back to 1925, this line would have come down because of the collapse produced during the war. It's a recreation to some degree of what it existed before the war, but nothing like the levels that were seen in years afterward. But as you can see, per capita GDP in China was basically flat. There was no economic development to speak of. 1978 Mao is now dead. The third plenum of the Communist Party Central Committee, led vice premier soon to be premier, Deng Xiaoping calls for officials to start liberating their thinking, to start seeking truth from facts. An ironic statement perhaps in light of our own discussions of fake news and manipulation of the truth that we're coping with today, but to innovate and experiment in the economy, and officials were instructed, this is when they started, China had been very closed and inward looking and they began going in large numbers abroad, they started applying for courses at places like MIT business school, they started trying to learn from other development models and this is gonna be the beginning of experimentation with new forms of capitalism. This is also when the big special economic zones in China were created, these massive islands of experimenting with capitalism, if you like, most of the world's special economic zones and we'll talk a bit more about them in other countries later are sort of industrial parks. They are 40 or 50 firms, they get some preferential tax treatment and the idea is to attract foreign capital. The ones being set up in China were enormous, they were sort of giant cities with schools and shops and people lived in them in their entire lives. So the idea was to experiment with capitalist economics in self contained ways. This is also when they start opening up to foreign direct investment. This is when companies like Cummins truck engines come to China, one of the biggest multinationals in China today started in the in the late 1970s, early 1980s. And so this seemed to be a little bit like the Prague Spring. And indeed, in the late 1970s, people started talking about the Beijing spring and there was much more criticism of the government tolerated, people started publishing things that you would have been thrown in prison for five years or even three years earlier, and there was something of a thaw, there was something called the Democracy Wall in the Shidong district of Beijing was created, a kind of graffiti wall if you like on which people published criticisms of the government, but the vast majority of the political criticism was backward looking. It was criticism of the Cultural Revolution, criticism of Mao, criticism of what had been done in the past, mistakes that had been made in the past. And of course, that wasn't gonna stay that way and so before long, the Democracy Wall protesters started shifting their attention from attacking the past to attacking the current leadership in China. Again, a lot of demands now for more democracy. Wei Jinsheng was a protester who came out and made very strong criticisms of the government and this went on until late 1980s, early 1981 and then there was a lot of pushback from the regime. And so much of the 1980s were this era of testing, political testing, by student movements, there were reorganization of student groups on Chinese campuses, there was a lot of criticism of the government interspersed with pushback. So Deng, came out and made very clear that when he was talking about experimentation with reform, he was talking about economic reform, he was not talking about reforming the political system and he reaffirmed what he called party rule and the socialist path. He focused on technical questions. His four modernizations had to do with agriculture, the industrial economy, the military, and technology, technological innovation. They started attacking what they called Spiritual Pollution. There was an anti-Spiritual Pollution campaign in 1983, and there was an anti-Bourgeois Liberalization campaign in 1987. And Hu-Yaobang who had been a member of the Central Committee, who had been one of the most energetic reformers was actually fired from the Central Committee of the Communist Party. So it was this period of back and forth, but a lot of growing protests nonetheless, in the 1980s. And it was also a time when the economy wasn't doing that well, particularly for consumers in urban areas, we saw a lot of inflation and that led to austerity programs, cutbacks on spending by the government, which produced further escalation of protests although the economic reforms continued. So if you look at foreign direct investment in China, you can see it's very low there, but it's going to accelerate dramatically over the 1990s. So then, in 1989, in the summer of 1989, if somebody had said, "Where is there gonna be political reform?" It would have been a pretty good bet especially given what was going on Eastern Europe, it would have been a pretty good bet that there was gonna be significant reform In China, the student groups were active just as in many European countries. And this erupts in June 1989 - Hundreds of thousands of people here in Tiananmen Square perhaps as many as a half a million, even more. In the history of Communist China there has never been anything like this. For the first time in huge numbers the ordinary men and women of Beijing, the old and the young, professors and taxi drivers have joined the student protests, lending their support to what has now taken all the appearances of a peaceful popular uprising against the oppressiveness of communist rule. As the afternoon wears on here, the crowd gets bigger and bigger hoping that this protest will produce results, that the government will respond to the students demand for meaningful dialogue and begin to implement the kind of genuine reforms as opposed to cosmetic reforms that the students want. - So these protests were going on and it soon became clear that the regime was not going tolerate them and tanks started appearing in Tiananmen Square. (crowd chattering) (crowd yelling) So that's not that different from the video we were looking at outside the White House in Russia when Yeltsin climbed the tank, and there was the real expectation among many of the protesters that the Chinese army would not attack its own citizens. And they certainly were appeals going out to that effect, but they turned out that they were wrong. (gunshots blazing) - The noise of gunfire rose from all over the center of Peking. It was unremitting. On the streets leading down to the main road to Tiananmen Square, furious people stared in disbelief at the glow in the sky, listening to the sound of shots. Heading down the road was a hazardous business but hundreds of people cheered as buses were set alight and army trucks caught fire too. (crowd yelling) They yelled and shouted. And then as troop lorries were seen moving down the road, there was gunfire from those lorries. (gunshots blazing) The troops have been firing indiscriminately, but still, there are thousands of people on the streets who will not wield back. The bicycle rickshaws scooped up the injured, others were shunted onto bikes and pedaled to hospital, many were carried away by frantic local residents. There was confusion and despair among those who could hardly credit that their own army was firing wildly at them. Young people was singing the international to a background of gunfire. (crowd singing) After hours of shooting and facing a line of troops, the crowd is still here. They're shouting, "Stop the killing "and down with the government." (gunshots blazing) A huge volley of shots just as I left the front lined cause panic. The young man in front of me fell dead, I fell over him. Two others were killed yards away. Two more people lay wounded on the ground near me. In the streets, many came up to us shaking with anger and disbelief and fear. Many were terrified, saying there would be retribution. There was not one voice on the streets which did not express despair and rage. (crowd chattering) "Tell the world." They said to us. - So that was the beginning of a massive crackdown, huge numbers of people. We never really got statistics. We believe just a number of people were killed in Tiananmen but it was very large numbers of people were killed. Huge numbers of people were imprisoned after that massive political crackdown. And in sharp contrast, again, this is June of 1989, when the the revolution is all over Eastern Europe, the thaw is going on, the Soviet Union is not intervening in the liberalization movements in Eastern Europe and of course, five months later, the wall came down in Germany, but China went a very different path. That the massive political repression of this descent but it did not stop the economic reforms and the economic transformation. And as you can see here, this is just foreign direct investment in China, it just throughout the 1990s it increases. The fact that it's decreasing as a share of Chinese GDP tells you how much the domestic economy is developing as well. So that even though as in absolute numbers foreign direct investment is going up by leaps and bounds, it's a declining share of the growing Chinese economy which is becoming a powerhouse. And here you can seethe astonishing growth rates that have been recorded in China since since 1989. You can see by 1991, they were recording already double digit growth rates and they have been the highest growth rates in the world, virtually every year since. If you look at manufacturing in China, and we'll talk more about manufacturing, this is between 2006 and 2016 as modest increases in the US and Japan, just threefold increase in manufacturing in China, just absolutely dramatic industrial take off. Comparable story on a smaller scale about Vietnam, transforms itself from an extremely poor country into a very successful middle income economy. I think it's the fifth largest economy in East Asia, but it is second only to China in terms of economic growth. So these huge success stories, and they have been studied to death. There have been armies of scholars trying to understand the reasons for this success. And in the readings I gave you, you get some sense of the variety of explanations that are out there that are trying to account for the dramatic success in China and Vietnam. And most of these explanations suffer from what economists and political scientists refer to as the pathology of selecting on the dependent variable. That is, everybody focuses all of their attention on China and Vietnam in trying to look for characteristics that explain the outcome, but of course, that's not how you do social science, we have to think about the independent variable because if you select only on the dependent variable namely in this case, economic success, you may be missing what's really causing it by not seeing other countries that have done the same things that have not succeeded. So, for example, Alexander Gerschenkron was famous for the idea that late developers have an advantage, they have an advantage because they don't have to necessarily go through all the stages. This is sort of the opposite of a first mover advantage, which we talked about in tariffs in the EU. So the idea is that late developers can skip some of the stages. You look In Africa, you can create a banking system without banks today because people can do banking on their cell phones, right? They don't have to go through all the stages. So these were relatively late developers. Well, they were relatively late developers, but they're plenty of relatively late developers that haven't had this kind of success, all over Africa, or in much of Latin America, elsewhere in East Asia. So the fact of being a late developer alone is obviously not the advantage. They were predominantly rural at the start. So the argument this is in Moleskine London paper, the fact that they didn't have inefficient urban economies to dismantle, they didn't have a big, inefficient economies to dismantle, made it easier that as they urbanized, they urbanized in relatively efficient ways. Again, there are many predominantly agrarian countries that have not been as successful as them. People point to Confucian values. Confucianism, above everything else values order. So they say, "Oh, well, it's Confucian values." But of course, there are countries with Confucian values that have not been particularly successful economically. The most obvious and dramatic illustration being North Korea. Confucian values haven't done anything for the North Korean economy. Education, both China and Vietnam have made huge investments in education. They're making bigger investments for example, than we are making in education. A K through 12 education but also technical education, science education. The few things to say about that most of the investments in education have come later after the economic take off, but also there're country. Russia is an extremely well educated population, Ethiopia has now got every child in school. They're one of the first African countries to have such high levels of K through 12 education but they have not been replicating this success. There was a less entrenched command communist system, the Moleskine London piece put a lot of credence in this idea that the Soviet economy, for example, was this massively inefficient, but deeply entrenched communist economic system that once they started to dismantle it, the whole economy fell apart. China and Vietnam partly because they were mainly rural economies in the 1970s, didn't have these a big entrenched communist economic command economies that had to be dismantled. They involved gradual reform, they did not, and we'll come back to this in more detail in a few minutes, but they did not simply replace the communist system with the capitalist system. In the Soviet Union, there was no choice because the communist system itself just imploded, it just collapsed. It became completely dysfunctional and so they were starting from ground zero. In both China and Vietnam, there was a gradual experimental approach, as I've already mentioned in China in the special economic zones, opening their economies to foreign direct investment which enabled the creation of joint ventures with foreign firms and allowed the economy to grow in particular sectors, not trying to reinvent wholesale a new system. Experimentalism and decentralization. They devolved a lot of power to make decisions about economic performance to regions. Provided that they delivered what was required by the central government, there was much less micromanaging from the center of how the different economic enterprises were run and how local officials met their goals, again, putting less demand on the central leadership to know what to do in every circumstance when very often they didn't know what to do. And they knew that they didn't know what to do. Competition within the state sector and between the state sector and the private sector, so they created big state owned enterprises, particularly in China, but they introduce competition among them and they required them to compete in the private sector and they shut down inefficient state companies. It should be said that it's very difficult to get a grip on the differences between the state sector and the private sector that means very much in China, partly because the many of the private sector companies get funding from state banks, but also because many of the top leadership of private sector firms are members of the Communist Party, and have to be members of the Communist Party. And so that raises the question while they might be nominally private, but what does it mean in terms of their influence ability by the party. On the other hand, the flip side of that is that it means that the leadership of the Communist Party is hearing a lot about the needs of business all the time, because there are many business people, business leaders are in the Communist Party and making the case for what they need. So the system is relatively competitive for a state sector dominated economy. Export oriented growth model. Again, this is pointed to by many scholars that this export basic producing for foreign markets, manufacturing goods for foreign markets has been extremely successful. There are many other countries in the world, again, that have produced for foreign markets that have not been as successful. Relatively good accountability and the rule of law. So, the system of accountability it's interesting. In China while you can't attack people for political reasons, you can attack officials for being inefficient and corrupt. And that has usually been the path by which people have pushed for accountability and reform. Yuhua Wang in the chapter I put in with the readings from his book, "Tying the Autocrat's Hands" points out that there's been a pretty effective increase in the demand for the rule of law with respect to commercial matters, not political matters, commercial matters, in the sectors of the economy that are dominated by foreign investment. And the reason he gives and I think it's plausible. The reason he gives is that domestic firms have access to elites to solve their problems through informal networks, the so called guangxi relationships that they build. Distinguishing guangxi from corruption is often easier said than done, but they have access to informal dispute resolution mechanisms, whereas the foreign firms do not. And so the parts of the economy in which the foreign firms are big have been where there's been a lot of demand for the creation of courts to resolve disputes. And Wang who's a very impressive political scientist at Harvard makes the case that it's this demand for the rule of law as a condition for attracting foreign investment that has fed that development. Of course, the subtitle of his book is Tying the Autocrat's Hands something and The Rise of The Rule of Law in China. I was recently at a panel with him where he said, "Maybe we should now be talking about the rise and decline "of the rule of law in China." Because, as I said earlier, slide showed you foreign investment is becoming a less important part of the Chinese economy to the extent that the demand for the rule of law is coming from the foreign investment sector, it might be diminishing. Leadership. So, leadership. Several people have noticed that if you look at authoritarian systems, that one way to classify them is that those that have managed to renew themselves will find mechanisms of renewal versus those that have not. So, if you think about the Arab monarchies in the Middle East, they stagger on until the king dies. Pretty much that's what happens in monarchies or dictatorships in Africa and Latin America, they stagger on until the dictator dies. And even in the Middle East, Assad is followed by his son, you get another Assad. And this is true in some of the communist authoritarian systems like North Korea, Kim Jong-il, Kim Jong-un, again, you have you have aging ossifying leadership's that don't manage to renew themselves and so kind of drag the country into the abyss as they become geriatric, whereas some authoritarian systems seem to have come up with mechanisms to renew themselves. If you said to most people, maybe most people in this room, who are the leaders of Vietnam? Most people would have no idea who they are. It's a triumvirate, they do get replaced over time, or you think about Burma. Again, there's no cult or personality. These are countries in which they have renewed themselves. And in China, after Mao, it seemed like they had really figured out a way to avoid the trap of charismatic leaders. Indeed, Bo Xilai had been a member of the Communist Party who was quite charismatic and some years ago, was kicked out and thrown in prison and his wife had allegedly been involved in a murder. Many Western scholars interpreted that as he was too charismatic and the Chinese didn't want, they tried charisma, they saw where it went with Mao and they were into this model where people would not be able to be at the helm for more than a decade. And so it seemed like to extent that you need young, vigorous leadership, countries like China and Vietnam had solved this problem which is hard for authoritarian systems to do, because there's no automatic mechanism of renewal and if a charismatic figure comes along, it's very difficult to get them out. So that was another reason people pointed to. So this is, I think, 11 reasons you could pull out of those review essays and no doubt people could come up with more. And the truth is, nobody knows, there's no silver bullet here. There is some combinations of these things, there things like luck but these are out there as the main explanations for why these countries have been so successful, right? It then puts on the table the question, well, how sustainable is this success? Should we be expecting it to continue going forward? And can we have long-term successful capitalist authoritarianism? This, if you like, is the question put on the table by the success of these countries now over a period of two decades beyond the wildest dreams economically of any observer, and maybe anybody even in those countries along with very repressive political regimes. And, so how do think about that. Well, there's some questions to wonder about with respect just to the economies. One is that since 2008, we have seen a resurgence in the growth of the state sectors of these economies. This is partly, as I intimated, at right at the beginning of today's lecture, that now that the Washington Consensus is in bad odor, people are now talking about the Beijing Consensus. Until 2008, China made no attempt to export its development model elsewhere, they would kind of back on their heels about the Washington Consensus. That is no longer true. Not only is the state sector in the Chinese economy expanded, they are now training people from many developing countries in economic management pushing a very different model of state-led growth from the model that was being purveyed in the Washington Consensus. And to the extent that the state sector is gonna be less efficient, that raises questions about how sustainable these growth rates are. Because even though relatively efficient, compared to other command economies, they're less efficient than the private sector in those countries and elsewhere. Then there's what Steinfeld in that "Playing Our Game" book that I put on there calls, the smile curve of profitability, and so, this is the idea which he actually takes from somebody calls Stan Shi. This is the idea that if you look at where the biggest profitability in manufacturing comes from, most of it is either at the R&D phase, research and development upstream, or it's in the manufacturing and marketing downstream. If you think about where's the money made out of an iPhone? It's people who come up with the basic technologies, make a lot of money, and then the people who put it together, do the marketing, make a lot of money. So the solid curve here, what it's telling you is that there's not that much money to be made in the manufacturing, this is just low wage labor and these are these people working on assembly lines. That's not the source of great profitability but that has been where most of the economic action has been in these economies. The dotted line is that realizing that the Chinese have started to try and move further up the smiley curve in both directions, so they tried to get more involved in research and development and more involved in marketing and sales. But of course, they do that largely by imitating, sort of creating if you like, generic versions that are copying technologies and then copying marketing methods. And if you do that, that tends to move, it means everybody else can do it too. And so if you like are sawing off the branch you're sitting on, and so it's not clear that they're gonna be able to continue to be profitable. Picking winners. Economists always criticize central planning because governments pick winners and governments don't know how to pick winners. They don't know where are gonna be the successful parts of the economy. That is less of a problem in the early stages of industrialization. You know what you have to do, you have to build cities, you have to build roads, you have to build housing, you have to put in electricity, you have to put in water. Once you get into complex economies, it's very much more difficult to know where the successful innovations are gonna be and where they're not gonna be. And so the government's might start to make a lot of bets as governments often do when they start to pick winners. So as growth becomes more complex, it becomes less likely that you're gonna have even relatively efficient political leadership's successfully keep picking winners in the economy. Changing leadership and corruption. The story I told about China's leadership in the last five years seems to have gone out the window. So now as Donald Trump complains, Xi Jinping might be there for life and Trump is clearly somewhat envious about that idea. (students laughing) But the 10 year term limit has now been abandoned at his request and he is a very charismatic figure, certainly the most powerful Chinese leader since Mao. And so it may well be that the mechanism for ensuring leadership renewal is breaking down, or has perhaps even broken down. In Vietnam, they had almost a separation of powers system with different leaders, responsible for different parts of the political and legal system and that also seems to be breaking down or at least under strain from reports that it's possible to get. I already mentioned Yuhua Wang's point that to the extent that the rule of law in the economy was dependent on demand from the foreign sector, that might be going away as foreign investment as a proportion of the Chinese economy is actually shrinking. And given the trade wars that we are now in, they might be shrinking a lot further than we talked about before. And so that plus the leadership changes might mean that we're gonna see increases in corruption again, inefficient management, different forms of corruption. Then the next thing and this is gonna now we're gonna start transitioning to talking about politics is that when you create big industrial workforces working on assembly lines, eventually they start to make demands, they start to want higher wages. And this is so called flying geese theory of economic development. This is the idea that capital is gonna go to wherever labor is cheapest. And so, if you look here, these are wages in China, Mexico and Vietnam, what you can see is that since the beginning of this century, wages in China have gone up quite dramatically and they're now substantially above wages in Mexico and wages in Vietnam, which while they've gone up, have gone up nowhere near as much. And so, if you look at this is called the labor cost index, which basically measures the cost of producing one unit of output. You can see that they are two. The cost of labor has gone up in China quite dramatically during this century and that has meant that the geese are flying out of China, they're flying to places where labor is cheaper. This is the Nike plant in Ho Chi Minh City, which I visited in March of 2001. It was run by a Vietnamese management company, with about 20,000 Vietnamese women working in it there. The whole thing consisted of, it looked like four giant football stadiums. This is about an hour's drive outside of Ho Chi Minh City, which used to be Saigon. It looks like four giant football fields but all of them built out of aluminum. You can see they could be packed up and put into a bunch of planes and flown off to somewhere else at relatively low cost. And here are some of the people working, what strikes you about these pictures? What's notable about these pictures? Yeah. - [Student] They're all women. - They're all women and they are all Vietnamese, the management of old Taiwanese and all the men, by the way, almost all men. Yeah, what else is striking? I don't know if the picture is quite captured. It's incredibly labor-intensive. These machines looked like they're 20 or 30 years old. Making a Nike running shoe is split into about 35 different tasks, one person is cutting the little pieces with a cutting machine, one person is putting glue, it's an unbelievably labor-intensive because labor is so cheap. That after all is why they went there and indeed, I spent the day there and walking around with the mid-level manager, and I asked him his personal history and he said, "Well, I used to manage a plant like this in China." So I said, well, what's the biggest difference between managing a plan like this in China and manage a plan like this in Vietnam? And the answer was, "The Vietnamese argue less at work harder." (students laughing) And I'm sure that was exactly true because the Chinese workers had been doing this kind of work for longer and they had started developing aspirations and so Nike had moved the plant to Vietnam. And so the idea would be that of the flying geese theory, that eventually demands for higher wages would cause the geese to fly. And, Vietnam will get this benefit for a while, but then the geese will fly to Cambodia, or wherever the next place is or maybe Zimbabwe. Now, this is also exacerbated the coming economy by the probability that a lot of these jobs are gonna eventually be going to technology that is, the technology is gonna become so cheap, that it will no longer be worth it to pay even starvation wages and so the geese might stop flying all together. But it's again gonna put stress on economies that are dependent a lot on manufacturing for feeding their populations. Slowing growth and regime legitimacy. A lot of people have argued that the legitimacy of these regimes depends upon their delivery growth, yes, they don't have elections, so they don't have political accountability but people have been satisfied because the standard of living has been going up, they've been doing a lot better economically, they're employed, and they have had prospects for upward mobility. As as growth slows in these economies, which almost every economic projection predicts that it will, that's gonna become harder and harder to sustain and so you're gonna get pressure, and maybe resurgence of political dissent and they're gonna be strains on those economies. So despite the many reasons that people cite for the success, there are reasons also to wonder how sustainable these models might be. And this leads us to thinking about the relationships between political and economic reform. During the transition in Eastern Europe in 1989, people started worrying a lot about this and these were the debates about shock therapy, should we have a dramatic reform right away, get the paint over with and then do political reform? Or should we do political reform first, or do them in tandem. And I've put three people up here because they ideologically from very different places. Adam Przeworski is a former Marxist, he wrote a book called "Democracy and the Market" in 1991. Janos Kornai "The Road to a Free Economy" is a riff off Hayex road to Saf deb So he was a kind of libertarian, and Jeff Sachs was a kind of middle of the road policy wonkish economy but they all made the same argument, which was essentially this, If you think about a collapsing communist economy, as the old system is dismantled, eventually people are gonna be thrown into distress and it's gonna get so bad per capita income is gonna fall below the poverty line. Eventually, things will get better, yeah. But what if you democratize here, then all of these people are gonna be angry and they're gonna use their democratic rights to stop the reforms and so instead of turning the corner, you're gonna just keep crashing down. And so the shock therapy people said, No, let's not democratize until this point, once the worst economic pain has been experienced. So these technocrats had the same kind of hubris as we were talking about last time, they really thought that they understood everything about market economies but they got a lot of things wrong. For example, both paths have turned out to be successful. If you compare Poland that had shock therapy with the Czech Republic that went with gradualism, they both been among the two most successful former communist economies in Eastern Europe. Of course, Poland's four times the size and so on, but they both have had very successful development pads. So it's not clear what made that much difference. The other thing that it ignores and this is really important, it's not only the losers from reform who mobilized, it's the winners who mobilized as well. The people who get things out of the reforms don't want things to go away. And so it's far from obvious that democratizing while you're doing economic development is only gonna empower the people who wanna stop it but on the contrary, it might actually help economic reform go along. The economic advantages of gradualism get lost. So one of the big secrets of success in China and Vietnam was this to track pricing system where they created a second economy next to the state sector, not instead of the state sector. So basically people were told, if you can produce at the government set price, what's your quotas required, on top of that you can produce as much as you want and sell it at market prices. And so that was one of the ways in which a market economy was created alongside the socialist sector and didn't depend upon its destruction. Those are the kind of advantages you forgo when you go for shock therapy. And this leads to a broader set of questions about modernization that I just want us to end with, which is that the old modernization theories said that economic development would inevitably produce demands for democracy. And I'm giving here one, you can find a million statements like this but here the person is saying, "The effects of socio and economic change, "rising literacy and common urbanization rights, "along with the improvements "of communications technologies "reduce the cost of collective action, "delegitimize autocratic rule, "and foster demands for greater democracy. "As a result, authoritarian regimes "which have a relatively easy time "ruling poor and agrarian societies "find it increasingly difficult and ultimately impossible "to maintain their rule once socioeconomic development "reaches a certain level." That was a scholar talking about China five years ago. The trouble with modernization theory is that that logic is very unclear. Why exactly is modernization gonna produce a demand for democracy? And if you sort of trawl through it, you find at least three arguments there. So the most famous one, first made famous by Barrington Moore was that it's all about the bourgeoisie. No bourgeoisie, no democracy. And I've mentioned some other modernizations theories there. And so there the thinking seemed to be that as you got a middle class, a middle class is gonna want bourgeois freedoms, they're gonna want to be able to travel, they're gonna want their kids to have opportunities and so they're gonna put pressure on the regime to change. Nonetheless, if the bourgeoisie are getting everything they want, it's not obvious that they really gonna demand changes, particularly if their lifestyles are improving. Certainly no inexorable logic why that might happen. So then a second variant, this really goes back to Marx and Engels, but I put up some more recent scholars of it. It's not about the bourgeoisie, it's about the working class. It's the working class, no working class, no democracy. And you've gotta have an industrial working class that is gonna get angry as its relative and perhaps absolute position eventually deteriorate and they're gonna demand. So it's gonna be pressure from below. And so this is what these scholars argue in different ways. Then a third group of scholars say actually it's all about war, that when governments fight wars, what they need is people to pay for the wars and people that fight the wars. And so they're gonna have to make concessions when they tax people with income and they're gonna have to make concessions when they ask people to go and fight. I should have put Michael Klarman's book up here, "Unfinished Business" he points out that every single increase in political conditions for African Americans in the US has been associated with a war. That getting them to fight has involved making political concessions. There again you have to wonder, Well, now when you no longer need large armies to fight wars, you can be doing it with drones and so forth, maybe that pressure won't be there. And we've been fighting wars for the last two decades on debt. So we don't have to make concessions to the unborn who are gonna be paying these debts. So there are these different logics but the upshot for the future in thinking about the future is that democracy is not inevitable; it depends upon contingent historical struggles. There's no reason to think it will spring from processes of economic development and modernization. They're conditions unto which it will, becomes more likely and we'll talk about them later but it's not obvious. Even if these economies do falter and wages stagnate and they start to lose legitimacy, they might just become more repressive. There's no obvious reason to think that they will democratize, perhaps they will, but perhaps they will double down. And in a post-Foucauldian world, this may take the form of what he referred to as the gentle way in punishment rather than throwing large numbers of people into prison. (upbeat music) - [Narrator] So picture your life in a place where everything you do, what you buy, how you behave, is tracked. The government gives you a score. And the score is a measure of how trustworthy you are as a citizen, and determines what you're allowed to do, like ever. boarding a train, getting a mortgage, all goes back to this score. It's called social credit. (ungainly music) The system's eyes are in big data, artificial intelligence, and roughly 200 million surveillance cameras. The scores go from 350 to 950 and are based on habits and behavior. Buy clothes or diapers, it's good. A lot of alcohol, too many video games. Not so good. And it's mandatory. When it goes national, social credit scores will be assigned to every one of China's citizens for life. In 2014, the Communist Party called for a system to allow the trustworthy to probe everywhere under heaven, while making it hard for the discredited to take a single step. The model is a place called Rong Chang it's in Chandon province. It's orderly, streets are spotless, cars slow down, which is unheard of in China. (dogs barking) Words like honesty and credibility appear on propaganda posters. Display cases show pictures of Rong Chang most honored citizens. You start with 1000 points, and you can print out your report anytime. You lose points for things like jaywalking, littering, tossing cigarette butts, or spreading rumors. Cameras do a lot of the surveillance work, but it's pretty analog too like posters that list the ways you can gain or lose points. (speaking in foreign language) And each night, local TV shows the surveillance highlights of the day. (speaks in foreign language) And then there's Juan Yi, she's an information collector, a paid enforcer who walks around and writes down deeds about her neighbors. (speaks in foreign language) Like the man who carried a drunk person home, things like this are good deeds she said. But the farmer overheard swearing and being rude, yeah, bad deeds. (speaking in foreign language) Her quota is 10 a month. She likes the work, thinks the city's better for it. - Will see you on Thursday. (students laughing) (calm music) |
Microeconomics | Lecture19_Producer_Surplus.txt | Now, let us move to the Producer Surplus. Producer surplus of a seller in a transaction is the difference between the price he gets in the market and the minimum price he is willing to receive for that product. So, producer surplus of a seller in a transaction is the difference between the price he gets in the market and the minimum price he is willing to receive for that product. What is, from where do we get this minimum price? Student: Marginal cost It is his marginal cost, it is his marginal cost. Seller would not like to receive a price that is less than his marginal. Student: Marginal. Cost and price, what is the price he receives in the market? That is the market price. So, little bit you know little bit about cost, later on when we start talking about producers theory, we will talk more about the cost, but just basic little bit about the cost. What we have is the total cost? Total cost is a cost that you incurred to produce let us say Q amount of the good time that is the total cost. We can divide it into two part; fixed cost and variable cost. What is fixed cost? The component of the total cost that does not vary with the quantity produced by the farmer produced by the producer, why do we get the fixed cost. Let us think about someone think about a company producing chewing gum, that company will have a CEO does not matter whether the company produces 5 chewing gum or fifty thousand chewing gums, the company will have probably one CFO also. So, these would be or let us say it will have a building also where the chewing gum is produced whether you produce 5 chewing gum or 50 chewing gum or 500 chewing gum does not matter. You need a place to produce it and to the certain large extent that you do not need to increase the size of the place as long as you do not need to change, by change as long as your cost does not change the component of course, that does not change if you increase or decrease the quantity that component is called fixed cost it is independent of Q. Student: Yes sir. And variable cost is a function of. Student: Q. Q if you produce more, you will have higher variable cost. So, this is we can say variable cost is a function of Q and this is fixed cost small digression and now marginal cost as I said earlier let me remind you again, it is quite important concept - MC is the cost to produce one more unit at least that is how we are starting. So, what we can say that MC is total cost, if this is represented by total cost marginal cost is nothing, but total cost of producing Q plus 1 unit minus total cost of producing Q unit. We can write it like this also. And more as I said earlier in words I am telling you today mathematically that rather than thinking about change in terms of one unit we can also think about terms in a. Student: Q Small smaller units and that smaller unit is let us say delta Q. So, what we can say the better definition would be marginal cost is nothing, but Q plus delta Q where delta Q is the very small entity and this is Q plus delta Q divided by Q. And this we can write as delta C or delta TC if you want and this is delta Q and if you know calculus we will use the concept of limit and if you do not know calculus do not worry about it you can ignore the part what we can do is that we can take delta Q tending to 0. So, what do we get then MC is equal to limit delta Q tending to 0 delta TC delta Q and this is nothing, but the. Student: Partial. Partial derivative of total cost with respect to Q fine. So, that is why these two definitions indicate the same concept. This one is more precise, but this is used when we are describing things in words because this is more convenient to think, the changes in terms of one unit rather than thinking in terms of change in infinitesimal unit, very very small unit and that also change is going to 0, at least tending to 0. So, that is why in words will keep on using this concept. And also you should notice that we are talking about marginal cost. Let us write this total cost in terms of fixed cost as well as variable cost. Student: Variable cost. So, what we get here the total cost we can write it as variable cost for Q plus 1 unit plus fixed cost. Remember fixed cost does not depend on number of units. So, we will just write as fixed cost and what we will get here total cost for Q unit and that again will be divided. Student: (Refer Time: 07:47). Into two component variable cost for Q unit and fixed cost for Q units that is again going to be the same as earlier divided by Q plus 1 minus Q. So, this will get canceled and what you have basically is VC into Q plus 1 minus VC Q. So, what we can say that marginal cost is nothing, but equal to variable cost producing one more unit. So, now the concepts that the graphs that we used what we can, let us say that the marginal cost is gradually increasing. To produce first unit you need to spend 1 rupee, to produce second unit from first to second unit you need 2 rupees. Let me write here the quantity, total cost and variable cost. So, from 0 to 1 total cost is 2 and here we have fixed cost. For quantity 0 total cost is let say 2 rupee variable cost is 1 rupees, fixed cost is 1 rupee sorry; variable cost is 0 and the fixed cost is 2 because we are not producing anything is going to be equal to 0. Student: 0. While you are still because even though you are not producing anything, but you have that building you have to pay the rent for the building. You will get into more details about it little later when we start talking about producer theory. Now, from 0 to 1 what you have is let say fixed cost would remain same, it will be always be equal 2 as it is fixed cost and variable cost is let us say now you have 1, here you have 3, here you have 6, here you have 10, again made up number. What will be your total cost? 3, 5, 8 and 12 and here we can write the marginal cost. Here there will be no marginal cost what is the marginal cost here to produce the 1 unit. Student: 1. How can we calculate the marginal cost either we substract this 2 from 3 or we substract this 0 from 1. In both cases we will get one similarly how we can get from here we subtract 3 from 5 or 1 from 3 to get 2 and similarly here we get 2, here we get 4 fine. Now, we can draw marginal cost versus quantity how would it look like. Student: Increasing. 1, 2 something like this it will look like, it is increasing. Now, let us say market price is 4, 4 rupees, how many units will be transacted in the market; up to 4 units because marginal cost is equal to 4 for the fourth unit. So, up to 4 unit will be transacted in the market. What will be gain from that selling the first unit? If you are selling the first unit, what how much is going to be the gain? Student: 2. Student: 2. 2, how? Student: 1. When you sell it in the market, market price is 4 rupees. How much is your cost? 1 rupee, how much is going to be the gain? Student: 3. 3, here it is going to be equal to 3. When you sell the second unit how much did it cost you to fabricate this extra unit Student: 2. 2, how much are you going to get paid in the market? 4. Student: 4. So, your gain is. Student: 2. 2, and this is 2, and then 1. Student: 1. And then 0. So, how much will be the total surplus for you 3 plus 2 plus 1. Student: 6. 6. So, one thing that I did not mention earlier, let us go back to the demand curve. I have been drawing, notice that I have been drawing whenever I am saying demand curve what I am drawing, I am drawing the marginal value with respect to quantity. In the beginning I had already shown that marginal value with respect to quantity is basically inverse of demand curve, but when we draw, the two curves are the same, that is why I have been using these things interchangeably. Similarly in case of the supply curve what I have drawn the marginal cost versus quantity, but this also gives, the graph also gives the supply curve you understand. So, do not get confused between these two, that we have already established in the last class. So, in other word when we have a continuous supply curve how would it look like, some upward sloping curve and how can we give the producer surplus, we will take the market equilibrium price and this area gives you producer surplus. |
Microeconomics | Lecture55_Utility_Maximization_tangency_criterion.txt | Now, let us interpret what does it mean, we did it mathematically, what does it mean? This is the point. Of course, this is bad drawing, but this is the point. What is the special property of this point? It is tangent. Point of tangency; in other word if we get the slope of indifference curve at this point. It is equal to the slope of budget line at that point. Now, what how do we get the slope of just to interpret, how do we get the slope of an indifference curve? Very good, but let us see how do we reach there. What do we, how can we denote the indifference curve mathematically? It is basically uxy of course, I am talking about 2 good world, and is equal to some constant. Let us say k 0, or we can say u 0, does not matter, because it is constant. And we will differentiate it totally, total differentiation. What we will get? Equal to 0, why? Because this is a fixed number, if we differentiate with respect to any variable what do we get? 0. So, what do I mean by total differentiation? Let us look at it mathematically. We are going back and back in the description, but it is important to understand this. Slope of indifference curve, we want to get at a point. Let us say we want to get at this point. What we can do? That we start at this point and we pick any other point. Let us say 1 and 2. How can we move from 1 to 2? 1 and 2 what is happening? X is going up and. y. y is coming down. So, so x is going up, x is going up if, x is going up, how much increase in the utility you will get? What is del u by del x? Del u by del x is nothing but rate of change in u, with respect to x, fine? So, du by dx is rate of change in u with respect to x. And of course, we are keeping y constant. So, x is let say x is going up by delta x. So, total change in utility will be du by dx, del u by del x, that is the rate of change in u with respect to x multiplied by delta x. And similarly, how much we will get here? Del u by del y multiplied by delta y. Not minus, it will take care of itself of it on it is own we do not have to put. So, total change is going to be sum of these 2. And this is has to be equal to 0. If I want him to if I want this person to bring back to the same utility level. So, when I move from here to here by increasing delta x utility goes up. And to bring him back, what we have to do is; We have to decrease y by delta y amount. And so, these are the changes and this is equal to 0. So, what would be the slope? Let us say this is the slope at this point, this is the slope. Delta y by delta x is the slope. So, this is equal to from this formulation, this is equal to del y as u equal to u naught, fine? And of course, what we are doing? This is we are picking this point that is little further from the original point. What we can do? We can keep on decreasing the change in x; that is, delta x, we keep on decreasing the change in x. Accordingly change in y will also decrease. Because see ultimate aim is to bring this person back on the same utility level. So, ultimately this is what we get. This is what we will get, and same thing you can get from here. This is mathematical route, this is a mix route that graph as well as mathematics. Nothing fancy from here, what we are doing we are taking here limit delta x goes to 0, fine? And we are obtaining the slope, fine? And what we are getting? Delta y by delta x is equal to, and that the sort of this is u of x divided by u of y, u of x is nothing but del u by del x. And what is this? How did we define it? Earlier MRS this was earlier defined as MRS. Of course, we talked about a special case of MRS where change in x was given by one unit. But this is more general definition of MRS. So, this is a MRS marginal rate of substitution of good one with respect to good 2. So, in other word MRS is nothing but the slope of the indifference curve fine. So now, what we are talking about that slope of indifference curve should be equal to the slope of budget line. That is the tangency criteria, because if it is not true then, indifference curve would intersect the budget line at 2 point. They have to be equal, then only you will get that indifference curve is just touching the budget line, fine? So, what we have talked about? That MRS has to equal to slope of budget line at the point of contact. Fine, is it clear? We will do just, we will do this thing again using just argument so that we not only learn the mathematics, but also economic intuition behind it why we are doing it something like this? But one thing at a time. And what is the slope of budget line? Sir Minus P minus Px by Py, because remember the line is Px x plus Py y is equal to I. Remember, I started with less than or equal to 1, but we figured out because of monotonicity what we need? The constraint is no longer this, but constraint becomes un equality. And how much is; this is what we have figured out. So, what we have figured out is, dy by dx is equal to minus ux by uy, and this should be equal to that is the tangency criteria? That this is equal to minus Px by Py. Or if we leave this part, what we get is; ux by uy should be equal to Px by Py. Or in other word what we are saying is, ux by Px should be equal to uy by Py. |
Microeconomics | Lecture58_Example_Revisited.txt | So now let us begin with the problem, where we have utility function as x plus 2y ok. This is the utility function, what i am saying that this problem is from a 2-good world ok. So, on x axis the quantity of first good is given in x and quantity of second good in y ok. The idea is to maximise this utility, maximise this utility such that x plus y is equal to 10. So, this is the problem in in the mathematical way let us go in the opposite direction and let us get the economic problem out of it ok One way is to typically, what would happen that in life you get an economic problem and you will translate it into a mathematical problem, but here we are going in the opposite in the opposite directions. So, you would understand how it was done. And also, we should add that x is greater than or equal to 0 and y is greater than or equal to 0. We are going to solve this problem using two three techniques. But let us say what does it mean; that a person lives in a 2-good world and good 1 can be let us say for example, good 1 can be tea does not matter what you take, and good 2 is cola. Amount of tea is given in given as x and amount of cola is given as y ok. And this person what he cares about is his consumption in in this particular fashion; that choice of cola plus tea. You can simply say that he is willing to exchange 1 unit of cola for 2 unit of tea or 2 units of tea for 1 unit of cola. So, individual interest, individual valuation is same at all the point. And what is this value? That 1 unit of good 1 is equivalent to half unit of good 2, or you can say 2 units of good 1 that is of course, good 1 is tea and it is equivalent to 1 unit of cola. What it is saying that when you go to canteen or somewhere or you know one of your when you visit or what you want is either 2 cups of tea or 1 cola ok. It can be 2 cups or 3 does not matter we are not talking in terms of cups, but i am just saying to understand; that you are indifferent between 2 cups of tea and 1 glass of cola. If we measure tea in cups and cola in glass ok; that is what we have here fine and this is your own exchange rate. Now, let us say that what we have from the market, market we have x plus y is equal to 10. So, what we are saying basically, typical our budget curve is Pxx plus Pyy is equal to I. So, in this sense 10 is nothing but income P y is equal to 1 and P x is equal to 1. So, in market basically you can exchange 1 cup of tea for 1 glass of cola, that is what you can do in the market fine. But now let us look at your your own this is the markets valuation, and this is your individual valuation. Now in market if you give up 1 cup of tea you will get 1 glass of cola in lieu of that cup of tea. Now what is happening, what you have in your mind that 1 unit of cola is equivalent to 2 units of tea fine. So, 1 cup if we compare it with market in your mind valuation of cola is higher than the valuation of cola in the market. So, what you will do? You will consume more of cola rather than more of tea. And you will continue doing that, but here let us say earlier remember we talked about diminishing marginal rate of substitution, that is not true here it is not decreasing; again, i am going to come back and talk about marginal rate of substitution in little bit more detail, but this is not true here. The ratio for yourself is same at all the point and exchange rate for tea and cola is also the same in the market it is not changing. And your valuation in your own valuation you value more than cola in comparison to what market the market valuation of cola. So, what is the best idea for you? If if you have a bundle where you have tea and cola both you should give up tea and get more of cola, keep getting more of cola and till when you can do that till you have no tea left to exchange it for cola. So, in this case you will consume only cola no tea. Now, let us look at it this is logically speaking, let us look at it using a table ok. What we have here? Let us say here we have tea, here we have cola and here we have total utility. When we have 1 unit of tea how many cola we can have? 9; why, because 1 unit of tea is costing 1 unit 1 rupee we can say we can say everything is in rupee. So, it is costing 1 rupee. So, we have 9 rupees left and the price of cola 1 glass of cola is 1 rupee of course, this is all made up an imaginary, it does not match it does not match with the market, but the idea here is to learn the concept. So, 9 rupee you can afford 9 glasses of cola. So, here you have 9 and what will be your utility? It 9 multiplied by 2 because what we have utility is equal to x plus 2 y. So, 9 multiplied by 2, 18 and 1 19. Let us look at here if we have 0 units of tea; it means we have 10 rupees that we can spend on cola. So, how many cola we can get? 10 colas and total utility is 20, why? Because 0 plus 2 multiplied by amount of cola and that is 20. And similarly, we can do that let us say we increase the amount of tea what would happen? Of course, we will have to decrease the amount of cola. And it is if we are increasing it by 1-unit cola is decreasing by 1 unit. If we increasing tea again by 1-unit cola is again decreasing by 1 unit why? Because market exchange rate is rate for tea and cola is fixed it is 1 is to 1. So, it will happen in this fashion. And what is going to be the utility here 18. And let us do it for one more from 2 to 3 from 8 we will have to decrease it by again 1 unit, and it is going to be 17. And if we keep on doing this let us say now, we have 10 units of tea then of course, we will have 0 units of cola and what we will get 10 units of utility. And we have we want here to maximise our utility ok. So, which bundle should be pick? 0 comma and 10; this is the bundle that we should be selecting, why because 0 comma 10 bundle leads to maximum possible utility. |
Microeconomics | Lecture91_Returns_to_Scale_through_Graphs.txt | Now, as we are talking about graphs, let us use these graphs to also understand constant returns to scale, increasing return to scale and decreasing returns to scale. Here we have again we are drawing an isoquant on y-axis, we have K and on x-axis we have L and here we have one isoquant fine. And let us say we take this particular or let me say let me take another we are taking this particular combination of L naught labour and capital. Now if we double the amount of labour means let us say roughly, we are here, and we double the amount of capital we are here. So, what we will get? We will have another isoquant passing through this point and if it is constant returns to scale what we are going to get this is Q naught, then we are going to get here as. 2 Q naught. Two Q naught and if it is increasing returns to scale, then what we will get? Greater than Q naught. Greater than 2 Q naught and if it is decreasing returns to scale then it will be less than 2 Q naught. So, in other words if we can say, that when we have constant returns to scale and we are drawing isoquant with the equal you know the equal difference is like 10, 20, 30, 40. If this is the case again the pictures are not that nice, but if this is the case for let us say constant return to scale, what we will get in the increasing return to scale that, isoquant will become nearer for the same value 10, 20, 30, 40 they will become nearer and nearer why? Because if we double all the inputs then. Greater. Output will more than. Double Double and what we are saying that here we are saying that we are taking isoquant for quantity 10, then 20, then 30, then 40 again we are taking for 10, 20, 30 and 40. So, what is happening here in constant return to scale from this point you know if we are moving here we need to double, we can find a corresponding point here let us say here we have again this is not to the scale, we need to double when we reach here. But here if we double constant increasing return to scale what will we get? We will get more than 20, but if we want to draw the isoquant for 20 what do we need to do? We need to increase the capital and labour by less than the double amount. So, in that case it would be closer. Closer. To the isoquant 10 in comparison to constant return to scale isoquant 10 and 20; so, they will become nearer and nearer, and you can use the same logic if you have decreasing return to scale the isoquant will be spaced further and further for the same values in comparison to constant return to scale. Now we can talk about let us start a new topic again in the producer theory and that is cost. Is it clear? The real thing that we talked about, and it is quite important to understand not just mathematics, but also graphs it helps you understand the intuition behind these things. |
Microeconomics | Lecture53_Summary.txt | So, we will come back to these the uses of these assumptions and these axioms, but before that I just want to summarize that what we have learned through these axioms through these properties. So, the first we learned rationality, rationality axioms - completeness, reflexivity and transitivity. These three we have learned. And when our when someone’s preference exhibit these three properties, what does it mean that he is able to completely rank all his choices all his potential choices, with also possibility, with possibility that there are more than one bundle at some ranks this is a possibility. One thing also I should add that he will be able to rank only if he has finite consumption set and also what we have learned that this will translate into a utility function. So, if these 3 axioms are satisfied you will get one utility function. In case, of finite consumption good, fine. Next we add I am not saying just continuity implies this, but continuity in addition to earlier three axioms that we have learned, continuity. And what do we get? What we get is continuous utility function and now we do not have to impose the restriction that our consumption set is the consumption set of this individual is finite. What else, what does it mean continuity? No, break indifference curve, can I say that, no break in the indifference curve would be continuous. What it means is let us see that we cannot have we cannot this kind of indifference curve and also using all 4 above what we have learned earlier in the class that indifference curve of an individual cannot intersect two indifference curve of a particular individual would not intersect. That is also we will get got from continuity, fine. Now, let us add non-satiation, but let us take example of your monotonicity, monotonicity, and what would it imply? In other words, let us say that is clear that, that is through definition, but it if we translate it into indifference map what we will get that the indifference curve has to be a downward-sloping curve. Why? Because when we take here say a point we are talking, all the bundles here are, all the bundles here are let us say here is the bundle that we are talking about x comma y all the bundles in this blue zone are preferred over x y and here they would be least preferred. So, if we combine we bring continuity there will be a bundle in its nearby zone which a person would be indifferent between that bundle and x y bundle. So, it cannot in this zone, it cannot be in this zone, where will you find that bundle? Either in this zone or in this zone. So, if you keep on continuing how you will get the indifference curve? So, indifference curve has to be downwards sloping, fine. So, from here we learned that indifference curve has to be downward sloping. What else? Anything else the monotonicity would tell us? Of course, so that is clear from here the downward-sloping indifference curve because whenever you have downward-sloping indifference curve, MRS will be negative that is clear, that is clear. The other thing that we can get from monotonicity is that indifference curve cannot be a thick curve. For example, we cannot have let me draw really the thickest possible it cannot be like this. Why? If I do this let me take this zone and look at it in, at the bigger scale what will happen, what we are basically seeing is that indifference curve is like this. So, monotonicity will be violated because if you take a bundle here then here you have a bundle which has more of both the goods, but still you are indifferent between the original bundle and this bundle that can that is not possible. In fact, thickness what we talked about thickness, see downward sloping curve you will not get from local non-satiation, but local non-satiation will very clearly give you thin indifference curve because the same you can take the same logic here. Let us say here and you have really thick indifference curve like this and here is a bundle that we have taken and we take small neighborhoods around this ok. So, you are not able to find a bundle which is preferred over the original bundle. Why? Because the indifference curve is a thick one. So, even local non-satiation would say that indifference curve has to be very thin, fine, it is clear. And now we have convexity plus convexity. Let me say that forget about fulfilling the requirement imposed by convexity axiom. Let us say a person’s preference satisfies all the other axioms, and I draw let us say this is his indifference curve is there anything wrong with this indifference curve what is wrong. I am not talking about convexity I am not; nothing wrong everything is satisfied. Of course, what I mean here is of course I am not drawing the family of indifference curve, but what I mean this person’s indifference curve is like this, fine. So, everything is satisfied, but convexity is not satisfied, where? In this zone, in this zone. So, whenever we have convexity the indifference curve will look like this. Can I say it is bowed out? As opposed to this what you here, if convexity is not satisfied you get something like this. We covered all these axioms let us look at one more case because we discussed. When we have satiation or in other words when we have bliss points. How would the indifference curve look like? Concentric circle around like this why I am talking about it good 2 and good 1. Now tell me is it convex in this case preference of course, this is not the zone it is just poor drawing is it convex or it is not convex? Ajit, is it convex or it is not convex? Devesh? Student: It is a partial. Partial, I did not define anything partial in the class. It is convex, let us look at it, let us look at it. Let us take one indifference curve. Let us take one indifference curve, let us pick any two bundle on this indifference curve. All the bundles here are preferred over the original two bundles, why because here at this point utility is increasing in this direction and at this point, utility increasing in this direction. So, it is convex. Convexity is satisfied, continuity is satisfied, rationality axioms are satisfied, what is not satisfied here is monotonicity or local non-satiation is satisfied everywhere except at this bliss point, is it clear? So, now, what we will do we will start talking about mathematical problems, optimization, and all. |
Microeconomics | Lecture67_Giffen_Good.txt | So let me say here just the repetition, normal, inferior. Let us say we are talking about price of good 1 has gone up substitution, income these are the effects and overall. When P 1 goes up subs because of substitution effect x 1 will. Come down. Come down. And for income also because of income effect it will come down. So, overall it comes down. Come down. While inferior goods substitution effect quantity demanded would come down while income effect is. Up. Up. So, overall there are 2 possibilities either. The price of good 1 comes down and overall the total quantity demanded of good 1 comes down or. It goes up. It goes up theoretically it is possible. What does it mean that what when we can this scenario that price is going up and the quantity demanded of good 1 is going up. When do we get the scenario? When 2 things are important; first that this good has to be an inferior good, then only the substitution effect and income effect would work in the Opposite. Opposite direction so that is definitely required. What else not only this, but income effects should be greater than substitution effect in the so first requirement is inferior good, second income effect is larger in magnitude, larger than substitution effect in. So, theoretically speaking what is happening? The price of good 1 is increasing and if these two are true then the quantity demanded of good 1 is increasing such good are called. Giffen good. Giffen goods. Giffen good. What it means is P 1 goes up x 1 also Goes up. Goes up. Can you give me an example its very very difficult to find Giffen good in real life why? Why cannot we find Giffen good in the real life the problem is that the second point typically income it is very-vey difficult to have income effect larger than substitution effect. When can we have income effect larger than substitution effect when majority of your income is spent on this particular good then only it is possible? Sir just one question. Sir can luxury goods can be termed as Giffen good because. No. Luxury goods are not Giffen goods. Sir, like if Gucci is a brand and it is. Prices comes down by significant amount. No one could buy a Gucci, it is it is at that level only because of it is prices. I am not certain about this statement then when price of these goods would come down then people would not buy what you are saying is that let us say that the price of a purse you know these expensive purses; let us say 10000 rupees and if it comes down to 9000 rupees, then less number of people would buy these purses that is not possible. No, no if some purse is very expensive like 10 lakh rupees and it symbolizes your status. And other day it comes down to 1 lakh rupees. Then there is no point that the purse the set of people who were consuming that purse that good for 10 lakh rupees to show their status would buy the same purse at 1 lakh rupees because it will become more common, then, it would not symbolize their status. To see what again there are a lot of assumptions that you are making in giving this statement because see at 1 lakh rupees also it will be a status symbol for different group of people and given the income distribution in this world probably more number of people would be comfortable buying it at 1 lakh rupees price. So, probably the not probably the consumption of these goods will go up because price has come down, maybe not same set of people would buy, but when we are talking about from for the market. Sir the persons who have income who are like who can afford it. And also see this is the another requirement here that income effect has to be larger than substitution effect it means significant portion of your income is gone in buying these items so that is not true. That is not true. So, these luxury goods are not Giffen goods fine, it is clear? Let me give you just an example typically it is very very difficult to find Giffen goods. You know if you take an example of potato that I gave you earlier and it was it was talked about for a very long time in economic discipline that during Irish famine the potato was Giffen good because people were using most of their income to buy potato because that is what they would consume. So, when the price of potato went up they could not even spare little bit of money for meat and other food products. So, to get enough calorie; they had to increase the consumption of potato. So, in that case if this story is true then potato is Giffen good ok, but the problem is with this the second requirement that income effect is larger than the substitution effect. Typically this is not fulfilled that is why we have very-very hard time getting finding out, figuring out Giffen good in an economy. |
Microeconomics | Lecture92_Cost_in_Long_Run.txt | So, cost of course, we can talk about cost in two different settings. We have already learned about short run and long run what do we mean by short run? changing at least one input. At least one input cannot be varied. Varied. Cannot be changed and what do we mean by long run? All the input. All the inputs can be; Varied. Varied; so, let us talk about long run first ok. And we take the same production function that Q can be produced using capital and Labour. Labour we are talking about cost why are we talking about cost what is the aim? To determine on which cost we could supply. Fine, but let us think about a firm. What does the firm do of course, it produces certain output, but why does it do it? The first thing that you can say that it. Profit. Wants to Generate profit. Wants to not just generate Maximize. Maximize profit or it wants to get maximum profit from the venture and for that, it needs to decide two things. One how much to produce, how many units you if you want to say, how many units of output to produce. How to And second. How to produce. How to produce in other words. What. What combination because we know that same label of output can be produced by using different combinations of inputs. Inputs. So, which particular combination of inputs it should select to produce the amount of output decided in the first question? So, here of course, rather than talking about maximizing profit and rather than talking about how many units to produce, we are going in the backward direction and there is a reason to it, it will become clear to you later. That first, we are talking about that which particular combination of inputs it should select to produce a party the pre-decided amount of output. So, what we are saying is basically how would the company decide let us say this is the isoquant here we have labour here we have capital, which particular combination it should check it should take. Which cost minimum? Which costs minimum is because the idea is to produce Q naught amount of output, but with different combinations of inputs Q naught can be produced, but different combinations would cost different. So, the idea is to select a combination of inputs, which is least costly. So, what we are basically talking about it is, let us say if we have only two factors of production that is K and L this is our production function. Q is equal to F of K comma L fine what will be the cost of production let us say we do not have any other factors of course, we are assuming it is far from reality, but this is an assumption that we are making that we are using only capital and labour to produce. So, how much will it cost? Cost of wages to labour. Cost of capital and wages to labour. Labour. So, let us say cost to capital is per unit per unit capital for per unit time. Remember one thing because if we change the duration the cost will change. So, always remember that cost is here a flow variable cost is a flow variable, but we are talking about, and we will not mention it again and again that the cost of capital when I say. R what we are saying that r probably INR r per unit of capital. Per unit of time per unit of time. Per unit of time that we will not mention again and again, but we will just say that r is the cost to one unit of capital. So, how much would be the cost of capital? R. R k. K. R multiplied by k and let us say L w is the wage paid to the worker again w we need to specify that w is per u it of worker, per unit of time probably per day per month depending on how we are we have defined the problem. So, the cost is going to be w L. L. And thus, the total cost is going to be r k plus. w L. W L. Now my question is what if a person owns some amount of capital? He does not need because one way to look at it r is the rental price of capital why you are paying to the workers? Because you are hiring workers from the market. So, in lieu of their work, you need to pay them, but what about the possibility that capital you already own. Sir so. Then do we need to put here r k or we need to ignore r k. They were rented. They were no I am saying if they are rented then you need to put it definitely r k, but what if they are not rented? It is the opportunity cost of the capital. It is what you could do that you can let us say you can use your machinery capital represents some sort of machinery, you can rent out those machineries in the market and you would earn some money the rent for that capital you are using it for your own purpose. So, remember in economics we always talk about the opportunity cost, and what is the second best you can do of course, you can say I will otherwise I would keep idle ok. If that is your if you are not going to put it in the market, then it is fine you can put it at 0 because that is the second best alternative available to you, but the businesses they are very concerned about cost, and you know how they produce. So, they frequently rent out capitals and rent take you know higher capitals from outside. So, it is very important that they consider this cost and this is the second best they can do if they do not use it for their own production they will rent it out and then they would earn r per capital per unit of time. So, that is why it is important for us to mention it here ok. So, the; what we have now the cost is r k plus w L. |
Microeconomics | Lecture03_Allocation_Command_Market_and_Mixed_Economy.txt | When you do not have unlimited amount of everything, unlimited amount of resource to produce unlimited amount of all the goods that you desire, you will have to make a choice, what to produce? In what quantity you should produce a particular good. So, that leads to our next term and that is allocation. So, what do we mean by allocation? Allocation is nothing but assignment, allotment, share, but in economics we are more technical about this particular term allocation. So, allocation here means solving these three fundamental or basic questions of economics, and the first question is what to produce? The second is how to produce? And third is for whom to produce? So, let us talk about this question, first what to produce? Let us look at to answer these three questions one by one starting with what to produce. Let us imagine a very simple economy. We will time and again in this course, to make our life simpler, we will assume a very simple version of it. So, let us talk about island economy, this island economy is populated by only two individuals. One is Robinson Crusoe, and second is Xena, the warrior princess. Names are immaterial just to make story interesting. So, do not pay attention to names, and there are only two activities on this island. The first activity is catching fish, and second activity is gather coconut. So now, of course, here no literal production is taking place, but again we will use it in a very loose sense. The production here involves catching fish or two production activities, catching fish or gathering coconut. So now, the first question is how they should do? Should they devote all their time to catch fish or should they devote all their time for gathering coconuts, because see time here is scarce, scarce resource available to them. So, if they spend all their time catching fish, they would not be able to gather any coconut, or otherwise also if they spend all their time gathering coconut, they cannot catch any fish. So, they have several options there; devote part of the time to catch fish or devote all the time that they have to catch fish depending on their need. So, how this allocation involves making this decision what you produce. The next is how to produce. Now let us bring, just to understand let us bring the third activity, that they can take up is that, they can fabricate net that would help them catch fish. So, how to produce involve; are they going to use their bare hand to catch fish, or they are going to make net to catch fish? So of course, this decision is about technology and what kind of process they would use to produce these two items. And third is about once they have produced, how they are going to allocate? How they are going to share? How they are going to give these collected coconuts and fish among themselves, Robinson and Xena. So, these are the three fundamental questions. And depending on who makes these decisions we have different form of economy, who makes these decisions. So, let us look at something, what we call command economy. In command economy essential authority typically, government some sort of government determines, how to produce? What to produce? And for whom to produce. So, government or a central authority is the decision-making body. So, here individuals do not have to make many choices. Government decides for the individual. The second we can think of is market economy. Here, as opposed to command economy, decisions are made by individuals. For example, let us say a farmer in west bengal produces some rice, nobody tells him how he should produce, what he should, how he should, where he should send his produce. So, what is happening there the individual is deciding on his own about these allocation questions. Third is mixed economy, here it is somewhat a mixture of command and market-based economy. So, some decisions are made by central authority while some other decisions are left to individuals. Now think of it how would you categorize Indian economic system? Do we have command economy? Or we have market-based economy? Or do we have mixed economy? Also, it is good idea to think about what we had before 1991, and what we have after 1991. Just to let you know earlier we had mixed economy, but heavily tilted towards command economy central planning, but after 1991 after economic liberalization we are slowly moving towards market based economy, but we are not yet there so, still I would say that we have mixed economy, but realistically speaking these are 2 extreme command economy, and market economy. All the economies of this world we can say we can characterize them near to command economy or market based economy, but truly speaking none of the economies are purely command economy or purely market based economy. One concept that I would like to emphasize here is laissez faire, this is a French word it means leaving it alone. So, what do we mean by leaving it alone? That here in laissez faire kind of economy of course, it is an extreme version of market economy where private parties are absolutely free from government interaction, intervention such as tariffs, tax, regulations. So, individuals transact without any interference of the government. So, this is a concept, an extreme concept that we typically do not find, but you would see in newspapers talk about moving towards laissez faire or moving away from laissez faire. |
Microeconomics | Lecture51_Marginal_Rate_of_Substitution_MRS.txt | The second way to look at it let us say, let us take again little different example then tea and cola, what I have again if I go back to the Island and now let us say that we have mango on that Island; mango and coconut, and there is no other good. Now let us say its Robinson fine, and let us take a bundle here when he has 4 units of coconut and 1 unit of mango. Now let us say that suddenly Robinson finds someone selling mango on this Island. And since on this island there is no other good available, the only thing that you can do is barter, you can give some mango take coconut back for those mangoes or give some coconut and take mango back fine. So, now let us say the Robinson wants to increase his amount of mango from 1 to 2. Here we have 1, here we have 4, let us say here is 2, earlier he is here. So, he wants 1 unit of mango. Now let us say the person who is selling mango says that I need c amount of coconut to give you 1 mango. Let us say, let us not worried about whether he would become better off or worse off, let us say if he participates in this transaction then what happens. Then he will get, he will have ultimately 4 minus c amount of coconut and 2 units of mangoes. Now there are three possibilities. Oh sorry, thank you 2 comma 4 minus c fine. Now there are three possibilities, either this person is better off, worse off or indifferent. Of course, what we are doing here, we are changing the amount of mango by 1 unit. Let us say that we have figured out a particular c star for which this person is indifferent. There will be a c star, because we are talking about continuity assumption. So, there has to be a c star that would make this person indifferent between 1 comma 4 and the new bundle fine. Now, let us say, if he is moving from 2 to 3, now he wants one more. So, now, let us say here we have started from 1 comma 4 to 2 comma 4 minus c star and from there we are moving to 3, and the idea is to keep this person Robinson indifferent between these bundles, and of course, he has to pay further to get this mango. Now what would happen; of course, it would be 4 minus c star, let us now put here 1. This is for the first unit of mango, and now he is going to get, he is going to give up c2 star amount of coconut, but the thing is that he is in different between these two, and he is indifferent between these two. So, using transitivity, this person is indifferent among all three bundles, but now my question is, when he had one mango, he paid c1 star coconut for 1 unit of mango. Now he has 2 mangoes. Do you think he would pay. Now he is paying, he pays c2 star for 1 unit of mango. When you talk about a relationship between c star 1 and c star 2. So, there are three possibility that c1 star is equal to c 2 star, second is c 1 star is more than c 2 star, 3 is c 1 star is less than c2 star, which one would be typically true. Of course, different person would have different kind of preference, but we are talking about a typical individual, which one do you think would be true, 3. So, what you are saying that as he has more mangoes, his willingness to pay for one more mango would go down fine. Now let us try to draw it, how would it look like, how would it look like, what you are saying, if what you are saying is true how would it look like, starting with 1 comma 4 and we increase here to 2, this is 1, this is 2 and here is the fall this is c 1 star. The fall is going to be more or less you said 2, fall is going to be less now, c 1 star that is what we are. Discussing. So, less. So, for this we will be here, and if we continue with this then fall will be lesser, lesser and lesser. And of course, we are talking about change in mango by 1 unit, but if we, let us say mango is divisible and we are talking about change by small amount and then we will get more points and we will be able to draw unindifference curve, because what we did? We did these exchange in a way that this person remained indifferent. So, how would it look like. And if the case number 3 is true, if case number 3 is true then what would happen, it will keep on increasing and it will be like this; good 1 and good 2. Typically what has been observed that people exhibit this rule number 2 and that is the reason, that is another explanation that I talked about the balance that why do we get, why does preference exhibit convexity, because balance bundle is preferred over extreme bundle, this is another interpretation. Why do we get convex preferences, because our willingness to pay for a good decreases in terms of other good as we have more of the first good, and this has a technical term in economics. This is basically diminishing marginal rate of substitution, but before that let us talk about what is MRS then it would be clear, marginal rate of substitution, fine. So, what is marginal rate of substitution, what do we mean by marginal rate of substitution. Again how you, what you are talking about is important. So, let me say again marginal rate, just saying marginal rate of substitution is not good enough, we have to specify that marginal rate of substitution for good 1 with respect to good 2. Now you can say that marginal rate of substitution for good 1 with respect to good 2. It is basically as the name says it is a rate, at which consumer is willing to give up good 2 to good get 1 more unit of good 1. So, marginal rate of substitution for good 1 with respect to good 2, is the rate at which the consumer is willing to give up good 2 to get one more unit of good 1. Since it is a rate let me reverse this, that marginal rate of substitution for good 1 with respect to good 2 is the rate at which the consumer is willing to exchange good 2 to get one more unit of good 1, whether the consumer is giving up or taking good 2 to reduce the amount of good 1 by 1 unit. So, it is rate of exchange, not the market exchange rate, but the individual exchange rate and this exchange rate is such that, if you participate in this exchange your utility or your level of satisfaction would not change, you will remain on the same indifference curve, your utility level is not changing, what does an indifference curve represent, a particular level of utility fine, you had a question See opportunity cost is typically we talk about in the market sense. Here we are talking about what an individual would feel. So, it is possible. Again there are two parts, why do you buy a good in the market, because the value, the price of that good is less than the value that you assign to that good. So, we are talking about, in the individual sense, not in the market sense, the opportunity cost is related to the market, what is the price of that good in the market ok. So, its bit different, but the concept is very much similar. So, we will talk about the opportunity cost, when we talk about the budget constraint, because there the budget constraint relates to the, how market is allowing the, what is the market rate of exchange. This is your individual rate of exchange, you understand that is why its bit different. So, marginal rate of substitution is very clear to you. Now what we are saying is, whenever we have diminishing marginal rate of substitution, then the preference would exhibit convexity. So, convexity implies DMRS, D sort for diminishing decreasing, but this is not always true, this is not always true. What I am not saying that convex, its convexity, always implies DMRS and DMRS always implies convexity, but I am saying that its not necessary that an individual preference would exhibit convexity, fine. |
Microeconomics | Lecture24_Elastic_Inelastic_and_Unit_elastic_Demand.txt | Now I gave you the definition of price elasticity of demand, when do you think the demand is elastic enough? And when do you think the demand is not elastic enough or in other words, inelastic that is the opposite of elastic? sir when we have many substitutes. No, I am not talking about the factors affecting the price elasticity of demand. I am talking about mathematically. So, what I am saying, let us say, I do not have those many products, but here I put a little bit of force and I am able to stretch, let us say if I am putting a lot of force and I am not able to stretch it or if I am able to stretch it a little bit, when can I, what should be the cut-off, that then, in this case, it is elastic and in the other case it is not elastic. vertical then it is not elastic and horizontal then it is elastic. You are ahead of you know where we are in the class right now, but how about just to continue with your thought, how about when it has some slope? mathematical number. So, mathematical number. So, how do we get the number? if it is less than 1. So, one how do you know my point is not that I am going to describe in a minute that why we take one? As that number why not 2? Why not 3? sir because if we apply a force like you gave an example if it stretches the same like rubber; that means, it is responding 100 percent with the means 100 percent in proportional with the price. So, we take one as. Good. So, what you are saying is that if there is a 25 percent change in price and if we get at least 25 percent change in quantity then we call it elastic, it means what we are saying is the price elasticity of demand, is at least equal to one and mind we are using the minus sign in the definition and when we put 25 percent change in the price and we observe less than 25 percent change in the quantity demanded then we have inelastic demand so one is that cut off point one comes out as a natural number, but also there is a reason. . Let’s look at the total revenue and what is total revenue price multiplied by quantity. It is the revenue earned by a seller if he sells a particular quantity at particular price. So, total revenue is PQ and now I am going to use calculus a little bit to describe. Because we already talked about in words why we are doing it, but here using calculus we will do it more precisely. Now, let us say, that the cut-off would come from here if P goes up if seller has increased its price and PQ also goes up then do you call this product elastic or inelastic? inelastic Inelastic and when PQ has come down why? It involves a certain bit of logic. Let us look at it P has gone up and what is the effect of P on Q? Q decreases P goes up Q decreases. So, if P goes up by a little bit, and Q does not decrease as much as the P has decreased. Then what is the impact on total revenue? Total revenue would go up, but that is case, when p goes up by 25 percent and Q goes up by less than 25 percent that is the case of inelasticity. So, this is what we get here inelasticity and when P goes up Q comes down, but overall impact on revenue is that total revenue comes down it means, it is elastic why it is elastic? Again, P has gone up if I continue with the same example, P has gone up by 25 percent and Q has gone up by more than 25 percent what would be the end result? PQ will decrease. PQ will decrease, that is why it is elastic not because PQ has decreased, but because. Q has decreased. Q has decreased more with respect with proportional increase in P that is why it is elastic. Now, let us look at it using calculus let us denote total revenue by TR if you could not understand calculus you can skip. Total revenue is P multiplied by Q. Now let us look at it what happens to the total revenue if we change P, and how can we get that change using differentiation, partial differentiation of this expression both sides of this expression with respect to P. So, what we get? And this will be equal to Q plus P DQ by del Q by del P. First, we are keeping Q fixed and differentiating with respect to P. So, when you differentiate P with respect to P you get one. So, one multiplied by Q, Q now we are keeping P fixed and differentiating Q with respect to P what we get del Q by del P. So, if you take, Q common 1 plus P divided by Q and what is the definition we put? Minus sign. So, what is this? This is equal to 1 minus epsilon. If by the way, I forgot to mention although I have been using it. That elasticity is denoted by epsilon. So now let us look at it, if epsilon is greater than one, what happens? P goes up, total revenue goes up no, total revenue goes down So, in this case, we have already discussed this is elastic. And when epsilon is less than one, P goes up, total revenue goes up. Inelastic and we should also discuss the case when epsilon is equal to one, this is called unit elastic P goes up TR remains the same unit elastic. Now, let us look at it, the demand function that we discussed Q is equal to 10 minus P, example we are talking about Q is equal to 10 minus p can you tell me? That this demand function is elastic or inelastic? it is unit elastic someone is saying it is unit elastic. Why do not you draw? So, 2 3 ways now we have to look at it. One way to look at it, if we use the definition what is the definition of elasticity? P by Q into delta Q by delta P here we were using x in place of Q it does not matter this is just a variable and delta Q by delta P what is this? I will give you through this example; I will give you one more method. What is this delta Q by delta P? For this not just minus 1 for this it is minus 1, but in general, what is this? Delta Q by delta P, it is inverse of the slope. So, also it can be written here that elasticity can be given as, one divided by slope. So now, how much is the slope? It is minus 1 and P divided by how much is Q? 10 minus P or in other word, what we can write? 10 by P minus 1 is it true? So, when P is equal to 0 what do we get? P is equal to 0 when put here P is equal to 0 what you get in the denominator when P is equal to let us write it here in detail P is equal to 0 it implies 10 by P is infinity. It implies 10 by p minus 1 is tending to infinity. So, what is happening here? This is tending to 0. So, here, at this point what is it is inelastic. Fine how about here? At this point? elastic It is at this point, what do we have Q is equal to 0. So, P by Q is tending to infinity. So, P by Q multiplied by 1 by slope, is also tending to infinity. So, here it is elastic. So, similarly if you continue with this thought, at some place in the middle you will get where elasticity is equal to one because, it starts with inelastic and becomes elastic and somewhere you will get unit elastic. unit elastic. What is that point can you tell me where? now we know we are trying to find a point where elasticity is unit elastic; it means, that epsilon is equal to one and this should be equal to minus P by 10 minus P multiplied by minus 1, or in other words P should be equal to 10 minus P. So, P should be equal to 5. This is the one way we have used to calculate. We have also another way to calculate it. How can we do it? Here we have the same line. Here we have P, here we have Q, on the same scale here we can draw Q on the x-axis and total revenue on the y-axis. When Q is equal to 0 how much is the total revenue 0 and when at this point when P is equal to 0 how much is the total revenue? When P is equal to 0 total revenue is 0, what is the total revenue P multiplied by Q. So, one of these 2 is 0. is 0. Then the whole thing would be equal to 0. So, the total revenue is equal to 0. So, at this point, it is 0, at this point this is equal to 0, how about in the middle? PQ or in the other word what we are saying because we are writing with respect to Q. So, P is equal to 10 minus Q and this is if you use a little bit of algebra, if you do not know the algebra then create a table and try to draw it, but if you know the algebra, this is an equation of, it is an equation of quadratic type. So, what will you get? It is also an equation of, if you use coordinate geometry, this is an equation of a parabola. You do not need this much of information to draw. I am just saying. In case you know you can also identify. The simple way to draw it is that you start, you make a table Q is equal to 0, Q is equal to 1 2 3 and draw more points you take better your drawing will be, but another way to draw is to understand that this is a quadratic equation and this is also the equation of parabola and then it would be much simpler it is going to look like this. Here you have 5, here you have 5. So, what is happening in this zone? In this zone from 0 to 5 quantity, you increase the quantity. Let us let me erase this, simple thing that you should look at it, that 0 to 5 range for quantity. Q is going up it means P is coming down any way look at it here. In this zone, if you move in this direction p is moving in the downward direction. So, Q is going up P is coming down and what is happening to TR? TR is going up TR is going up. So, if you look at this, that price is going down or TR is going up. Similarly, just to understand you move from 5 to 0 direction; what is happening? Q is decreasing, P is increasing, and TR is decreasing. So, simply P goes up TR is coming down, it simply means it is elastic zone. So, this zone is elastic zone. How about at 5, little bit of change because, here it is flat or in other words, this total revenue is maximised at this point. So, little bit of change in price will not bring any change in the total revenue. So, here Q goes up or down while; P remains the same, TR remains the same. So, unit elastic. So, at this point, it is unit elastic. Fine now in this zone what is happening? Just reverse of this, and here you get simply what is happening p is going up and TR is also going up. When TR is coming down TR is also coming down. So, here it is clearly inelastic. So, in the zone where price and total revenue move in the same direction then what do we get? Inelastic zone. When price and total revenue move in the opposite direction, we get elastic zone. |
Microeconomics | Lecture22_Implications_of_Market_Equilibrium.txt | One property I can say of this market equilibrium it maximizes the gain to the society. Mind you one thing you should always keep in mind that this scenario is valid only when you have large number of buyers and large number of sellers. Later on, we will learn this thing in a more technical way in more concrete manner; we will learn that this is a property of perfectly competitive equilibrium. So, keep this remember this term yes, we will come back to this term what is a perfectly competitive market and what is a perfectly competitive equilibrium. But one thing that we could see that equilibrium maximizes the gain to the society. So, we can say this is the one role that the equilibrium price, the equilibrium is playing in the market. Now, let us look at some other roles that this equilibrium price plays in the market, and let me name them what we have is something called rationing function. Let me write it here, role of market equilibrium rationing function. What is rationing function? See what is happening in this case now let us forget about that we do not have any intervention from government or anybody outside the market. So, what is happening with this demand function and with this supply function, what we are getting is that P star is equal to 4; that means, goods will be bought, this particular kind of good will be bought and sold at 4 unit price. So, it means people, whoever values this product at level lower than 4. Let us say for an example for one unit of this good will give me marginal value of 3 unit will I buy this product in the market? The price is 4 units, the gain to me is if marginal value is 3 unit. So, if I get this product my gain would be 3 units. So, price is 4 units gain is 3 units if in case, I go for this unit, then will I buy this product? No because by buying I will be incurring a loss of. 1. One unit that is why I will not buy. So, see the price is automatically rationing, it is excluding the people who value this product at a price lower than the market price. People who value this product their valuation is lower than the market price, they would not be able to buy. So, it’s a rationing what is rationing? Basically, you control the buying behaviour of the people. So, it’s controlling in a way that you do not value the product good high enough you should not buy this product. So, that is the rationing role that it is playing fine any question about it ok. Now, the second that I want to talk about is efficient allocation; and in efficient allocation I will talk about efficiency, first, something called productive efficiency and then allocative efficiency. So, because you know we are talking about large number of buyers and large number of sellers. So, sellers are competing against each other to sell their own product. So, now, consider two scenarios, where seller A has marginal cost of 4 rupees and seller B has marginal cost of 3 rupees. So, if just consider this scenario; if seller B says I will sell my product at 3 rupees 99 paisa what will happen? He will be able to sell the product and make a profit of 99 paisa, but what about the seller A? Seller A cannot bring down his price to 3 rupees 99 paisa. The reason being is that he will incur a loss. So, what it is doing, it is in a way its forcing the producer to use the least costly technology and the combination of input that reduces the price of the output. So, no one is telling him to do this, but he finds it, to do it in his own benefit. So, this is basically productive efficiency. And we will come to all these when we start talking about welfare economics towards the end in much more technical manner, in much more detailed manner. But right now I just want you to understand the basic these basic concepts that how the market equilibrium is leading to the productive efficiency. Fine, is it clear how it is leading to productive efficiency ok? Now, allocative efficiency; now imagine a scenario when you have only two kinds of input and combining these two inputs in various different ways, you can get various different kinds of output fine. So, now, you can combine it in different ways, and you get different kinds of outputs using different technology or whatever you have. For example, let us say one input is time, and let us say you are living on an island, and you can participate into two activities either you can gather coconut, or you can catch fish. So, input is just here all even I am using only one kind of input. So, what I said even if you have two kinds of inputs, you can combine it differently or use different technology and you get different output. And then what happens in society people consume different types of outputs, but society needs a particular mix of goods and services. So, in a way, I will tell you how, but in a way this market equilibrium is forcing producer to produce and the optimal mix of those goods and services why? Because they want to make a maximum amount of money. So, they would let us say if the demand goes up too high what will happen to the market equilibrium price? demand goes up, it will rise. It will rise and what happens if let us say if you are supplying too many of a particular kind of a good? come down. It will come down. So, automatically no one is telling you to adjust your production, but you will do it in order to maximise your profit. So, society would end up producing that optimal mix of goods and services. So, it will also lead to allocative efficiency, fine by the way I will come back to these two topics later on also. |
Microeconomics | Lecture87_Economic_Terminology_Opportunity_Cost.txt | Now, we are going to talk about some economic terminology before we move to another topic. So, the first is we are going to talk about opportunity cost and also sunk cost. Opportunity cost is the value of best alternative foregone. Foregone. Let us take it is not very clear from the definition but let us take some examples. Currently, you are studying at IIT Kanpur, I do not know the exact fee that you pay, let us assume that you pay per semester 25000 rupees am I right; how much do you pay? 44. 34. 50 approx. Let us keep it at 48 means, 4000 rupees per month you pay. The second is, of course, you have to pay for. Mess. Mess and let us say that you pay 2000 rupees per month fine. And let us say you need books and other items for the other items you are spending 1000 rupees per month. What is your cost of education? 7000. Because, I have deliberately I have moved you in this direction. Do you think it is exactly 7 it is of course; with these assumptions do you think it is 7000 rupees per month? More than 7000 per month. It is more than 7000; why it is more than 7000? Do not say that you did not include some other expenses. You know like you did not include the football that you bought last month, I am not I am assuming that this covers your tuition fee, this covers your living charges, and this covers the remaining item the money that you spend on remaining items. So, 7000 rupees per month would you still say it is more than 7000? Yes. Why? Because instead of stud studying over here, we might have worked somewhere. Ok. Or we might have done some other thing. But no one works. Particularly in our country no one at you know most. Sir, the services are not included What kind of services? Like the facilities you have provided on like the assurance factors like. No, I am talking about, I am not talking about what is provided to you; I am talking about the cost that you incur, because you are studying in IIT Kanpur and that this 1000 includes all the other costs. So, then if we do not work then leisure cost. Leisure never, but the idea is that it is a different thing that we do not you if you do not work that is fine, but what you could do see we come back to this best alternative for gone. So, for different people, it would be different. But it may be let us say for him that if. He is not in the IIT, he would have started his own business and again we do not know for certain, but expectation is that he would earn 50000 rupees per month because he has some experience, he has some family business, his contribution would be 50000 per month for his family. So, now, because he is studying here, he is not earning this 50000 rupees. So, his cost would be total 7000 and this 50000 and this is the opportunity cost that we are not including here in the cost fine. For example, let us say you do not have any business, if you were not here you would have not studied at all. The best alternative for you would have been to sleep in the daytime, play some games and not earn anything. So, your opportunity cost is 0, because you would not have earned anything. So, again see it is not explicit in the sense it is implicit, and it depends on the; what is the second best available to you? You selected this because this was the best available to you. So, now, you have to look at the second best and what is the value of that second best that you have to figure out and that is your opportunity cost. So, it varies, it depends on the person; is it clear? So let us take an example; another example, again made-up example not a perfect example, but made-up example. Let us say he has, you know at present he has two different options available, one that he can work for me. You know I have some research assistants’ work that I can give, and I can pay him, let us say 500 rupees for a day work today fine; that is one example. Second, thing that he can do today although, we do not have any cricket match, but let us say that today we have India versus Pakistan match, and what he can do is to, watch this cricket match. (Refer Slide Time: 06:34) And let me give you two scenarios that in scenario A: that he has selected to work for me. What is his opportunity case in this case and in the second scenario is that he decided to watch the match what is his opportunity cost in. In case. For watching this cricket match. 500 approximately 500. So, in this case, if he watches the match and we are assuming this is the only other option he has, in that case, his opportunity cost is 500 rupees, but how about here? 0. 0. Considering there is no value of leisure. But it is not leisure, it is pleasure. There is a difference between leisure and pleasure. If he works for me, the second best is that he would watch this cricket match and there is some value associated with that cricket match, what we know that it is definitely less than 500 rupees, but how much we do not know, but the opportunity cost is simply the value that he associates. With the cricket match, with watching this cricket match so, you see you have to figure out the second best alternative and that would be your opportunity cost, the value of that alternative. Is it clear, perfectly clear? |
Microeconomics | Lecture78_Average_and_Marginal_Product_of_Labour_APL_MPL.txt | So, now let us take numerical example, what we have here is let say capital as we have fixed at 5. And here we have labour and, here we have output. So, let say let us take some example, labour is 0, it is possible that to produce something, but we will take that here in this example when labour is 0 output is also 0. It is a made-up example, then what we have that capital at the same labour; labour, goes up to1, then let us a total output is 2, I am trying to create a table and then again it is fixed at 5 it is labour is increased from1 to 2 or output goes up from 2 to 6. Now, you got that draft what I am doing I am again keeping the capital fixed at 5 and increasing labour from 2 to 3 and, output goes up from 6 to 12 and. So, on I have already this there is nothing sacrosanct about this example, this is an artificially example that I have created to illustrate certain concepts. So, let me just quickly make it 12 and then 20, if you say 20, 25. I am not see function, if you remember, function is thus just the combination of you know some inputs and outputs, some independent variable and dependent variable, some time we get algebraic expression, some time we do not get algebraic expression. So, let us not worried about let us not worried about algebraic expression that is not the aim right now, 6 it should be 5 6 27. Let say 5 4 28 5 8 28 5 9 27 10 5 10 and 25, now what we can talk about is that average productivity of labour APL, what is average productivity of labour, average productivity of labour is Q by L, of course for fixed capital; capital, has to be fixed. We are talking about and of course, average product, but first let me tell you what average productivity of labour in words is that it is the on average contribution of one labour in production and of course, production depends here not only on labour, but also on capital, but since we are talking about labour, we are we have we are keeping any way we are talking about short run. So, we are keeping capital fixed. So, what is the average contribution here, we cannot define how about here 2 how about here 4, 20 by 4, 5, 25 by 5 5 here 4.5 here 4 here 3.5 3 2.5, now deliberately I have created this fine it is clear. Now, we can also define similarly marginal productivity of labour. And what is marginal productivity of labour, change in output divided by change in labour, or in other word that incremental contribution of additional labour, you know incremental contribution. So, here we can start calculating from here labour went up from 0 to1 what happened to the output it went up by 2 unit. So, by adding 1 more labour we increased output by 2 units. So, marginal contribution marginal productivity of labour is 2. Similarly, from here to here again labour went up by labour is has gone up by 1 unit. So, how much we cannot look at here, we have to look at here, why because earlier output was 2, but by adding 1 more unit of labour, we are able to increase output to 6 units. So, here it is 4 and so on if you calculate. Here we get 8 basically see here the jump is of jump is worth1 unit. So, what we have to get basically is difference between these two, and 20 minus 12 is 8. Similarly, for here we have to get difference from between these two and it is equal to 5 here 2, here1, here it is 0 minus1 here minus 2. So, now I believe you understand the concept of average productivity of labour as well as marginal productivity of labour. If we plot this, here we have total output and here we have labour how would it look like. first increasing Just that nothing more. decreasing slope First it is increasing at the here let say, here we have1, here we have 2, here we have 3, here we have 4. So. straight line Not a straight line in is if you do not know what happening in the middle. So, initially it is increasing at the slower, then it picks up, then it rate of increase decreases and then it starts decreasing overall. So, what we have basically is something like something I let us not put it to the scale roughly I am saying, what we have is basically like let us we put let me put here. But it is 2 4 6 8. So, it is following the graph. Not to 4 I am here is Q total output it is 2 6 12 20 and, it picks up where at 28. So, like we can say this is 7, or here it is like this like this and here also it is this is 8 fine it is clear. To produce one thing more than labour required are working on it. So, they are getting into each other’s way and they is not much work as per the. So, one story that you can make that here, capital represents number of computer and labour represents number of engineers coders, then what is happening that earlier the you have 0, you have no coder you have no engineer. So, all computers are lying ideal. So, you have 0 output, as you start adding engineers, they start occupying computer and start working. So, you have basically 5 computers but till 7 because you know not all of them are working all the time. So, they can share computer, but after certain labour what is happening is that the there is some kind of crowding, you know like there are more people than number of terminals. Of course, I am not saying that happens all the time, but it is possibility that if there are more people they would start chatting and the productivity would go down. So, typically it has been observed that if you increase1 factor beyond the limit you know limit would vary in each of these cases, then the productivity starts decreasing that is what happening here fine is it clear. Now, tell me how can we represent average productivity of labour on this graph, what is average productivity of labour on this graph, slope of line from origin to that particular level of output, like let say if I want to calculate average productivity of labour at this point what we will do, I will go here and see how much is the output, this is the output at this particular level, and we will draw a line from this point to the origin, hum what is this is Q at this level let say it is a Q not and what is this L not. So, this is how we have defined the Q not divided by L not is the average product of labour in, we are keeping the capital fixed fine. So, at any point we want to get the average product of labour. So, what we have to do is we have to draw a line. So, what we can observe here, in this graph that first average product of line, average product of labour is increasing till probably this point and, then it starts decreasing. So, average product if we draw here let say we can say that average product of labour is increasing and then it start decreasing, how about marginal product of labour, how can we get the marginal product of labour at any this point. Slope of tangent at each point. Remember how we have defined? We have defined marginal product of labour as incremental contribution of additional labour. So, marginal product of labour is delta Q divided by delta, what is delta Q change in amount of output and, what is delta L change in amount of labour, but the way when we say again I keep repeating this again and again, but so I will do it 1 more time that when we are talking about incremental contribution of an additional labour, what we are talking about that Q of L plus 1 minus Q of L, and L plus 1 minus L, but this is not very precise. The above definition is precise here, we are the increment the increment is equal to 1 labour, but if we have a continuous function, then we can take even smaller increments and, if we take small enough increment what we will get as marginal product of labour del Q by del L, why del, why this partial derivative because capital is fixed ok. So, we can say capital fixed, and what is this del Q by del L how can we represent it in this graph this is the tangent, this is the slope of the tangent to the total output curve at 1 particular level of labour ok. So, let see if you want to in this graph in this graph, we want to figure out what is marginal product of labour, what we can do,1 way to do it is the decrease labour by1 unit of course, we will get here and see how much how much is the change in output and, this much is the change in output ok. And this is the change in labour that is equal to 1. So, marginal product of labour is nothing, but this amount, but this is basically crude way of getting the marginal product of labour better would be to have just small you know, or we can talk about small increment here, we do small increment and see how much the capital has changed. If the increment is small enough how much output has changed, if they increment is small enough, what do we get we get the slope of, this we get the tangent at that particular point. And the slope of this tangent is marginal product of labour fine. So, we get the marginal product of labour given here as the above ok. So, what we can say here, if you look at this point, if you look at this point which1 is higher MPL is higher why because, here it is steeper. So, steeper means higher slope. So, tell this point if you where it was tangent let me draw again, I have already made that graph. Let us draw simpler one, this would be doted rather, why I am using doted by the way why not the. Sir because we put the assumption that firms would increase their productivity and, if a product is coming down then this. And this closes to have Definitely; So, what we are, but again the definition of the production function we are drawing this, this is basically production function Q as function of L, production function in 1 variable and what was the products definition of production function that, we gave earlier the maximum level that can be produced. So, here if you come if you look at this table using 9 how much maximum you can produce 28 not 27 not 27 free, you what you will do of course, when you are using 9 whole units, you will produce 27 as the table gives, but what did we talk about earlier, that you can freely dispose. You can freely dispose some of the inputs and you will dispose probably 2 unit1 unit or 2 unit and you will come to this, that is why of course, here I am drawing the functions that is why it is going down. But the production function cannot go down beyond this point, it will be fixed you will freely dispose off because; we are talking about efficient level of production as you said fine ok. So, here we have Q and here we have L and let say at this point or rather take this. So, slope of course, this is from origin the slope of this line is the average product of labour and, how much is the marginal product of labour at this point, this is the slope of tangent it is clear. And so, marginal product of labour is more than average product of labour and so, here if we draw. So, till this point it seems ok, we have another here it achieves maximum. So, if we follow this. So, what is happening is basically the average product of labour is increasing, till this point and then it is decreasing, it’s not accurate, it is very rough done. And they if we take the marginal product of labour what is happening at this quantity same that here, it is higher and how about at this point, they are equal because this line becomes tangent. So, basically just what will happen that at this point it will be like this ok. |
Microeconomics | Lecture62_An_Example_with_Quasi_Linear_Preferences.txt | Now, I am going to give you another problem where we will use mathematical, some mathematical concept that we have learned to solve, that one is also difficult. Now, we are talk, going to talk about something called quasi linear preferences, let us say in 2 good world, preference is called quasi linear if utility function representing this preference can be written as a V x plus by. What it means is that utility function is linear in one of the arguments, are not linear in the other argument then we say the utility function the utility function represents quasi linear preference fine, is it clear? So, now let us take a problem where, what we have is maximise a lets instead of vx just for simplicity let us take log, a log x plus by or if we are using x 1 sorry . . V is a function it is not it is not same v as the earlier one, if it is may confusing you take it as w that it is a function of x, what I mean to say an individual is said to be having quasi linear preference if the utility for function representing his preferences can be written in this particular form and what is this particular form that it is linear in y and also this is addictively separable in x and y. That x and y are not appearing as in a multiplicative term, they are addictively separable then the preference is called quasi linear preference fine is it clear. So, let say its x 1 and it is x 2 and here we have maximise this is the last example we are talking about and then we will move to next topic x 1, x 2, alnx1 this is natural log ln is natural log and let say bx2 fine and the budget is p1, p1, x1 p2, x2 and it should be less than or equal to r. How will you proceed, indifference curve draw indifference curve how would it look like of course, you can solve it using indifference curve later on we will see that, but right now I want to give you flavour of mathematical techniques. So, let us solve it using the concept of marginal rate of substitution the concept that we have learned earlier ok. So, what we can see here. Let say here we have amount of x 1 and what we should put here is mu1 divided by p 1 and what is this mu1 divided by p 1 the gain in utility if you spend 1 rupee on x 1. So, to do this what we need we need to get mu1 and what is mu1 a by x and mu1 by p 1 will be a by p1x is not it. So, how will it look like p 1 is of course, given to you, you do not have any control over p 1 it is parameter that does not change in short period you can change it as modular you can play with it by you can study the effect of change in p 1 on this entity, but typically in its given to you ok. So, how would it look like 1 by x graph, something like this fine and how about let us calculate mu2 divide by p2, again this is for good 2 and what do we get how much is mu2 b divided by p2. So, how will it look like it is the same straight line horizontal line. So, what we are doing is here we are drawing first with respect to mu 1, p 1 with respect to x 1 and then m u 2, let me use the different colour m u 2 divided by p 2 and this is this is with respect to x 2 and deliberately I have put it on the same graph. So, what it means is that up to this level lets I am calling in it x 1 star why it will become clear immediately from 0 to x 1 star, m u 1 p 1 is greater than m u 2 p 2, no matter what is the level of the second good it is independent of the second good fine. What does it mean? What does this equation mean what will be your interpretation for this equation. He would buy more of first good. More of first good he will buy only the first good up to this level, unless this level is achieved we will not buy the second good because let say right now he has k comma 0 and k is less than x 1 star. So, it makes sense for him to spend one more rupee on good 1 rather than on good 2. So, he will keep on buying good 1 you can think you can stretch it little further you can say he is starting from 0 comma 0. And of course, first unit he should buy of good 1 why? Because mu one divided by p 1 is greater than m u 2 divided by p 2 and he will keep on buying good 1 until he reaches the level of x 1 star and once it reaches the level of x 1 star and he is still has some money left because there is a possibility; income is again a parameter, income is given to him there is a possibility that given his income is not able to buy x 1 star of good 1. If he is not able to buy x 1 star of good 1 in that case is optimal level of consumption would be the maximum amount of good 1 he can buy and 0 amount of good 2. Is it clear? But he second condition, second situation is that he has let us say he has already bought x 1 star amount of good 1 and he still has some money left, now what is happening that in after beyond this beyond x 1, greater than x 1 star m u 2 divided by p 2 is always greater than m u 1 divided by p 1. So, it means that once he reaches to the level of x 1 star he would no longer buy any more of good 1 and he will start buying good 2. So, what is happening here this is a very nice situation, if you understand it that he has income what he do he starts buying good 1 and he keeps on buying good 1 till he reaches, till he reaches x 1 star and how can we get the x 1 star? Do we have any way to figure out x 1 star. How that at this level its m u 1 divided by p 1 is equal to m u 2 divided by p 2 and we have value for both of these entities and how much is m u 1 divided by p 1, a by p 1 x and m u 2 divided by p 2 is b by p 2. So, it is independent of y. So, this gives us x 1 star and x 1 star is a divided by b p 2 divided by p 1. So, unless he accumulates this amount of good 1 he will not buy good 2 and to by this much of good 1, how much income does he need? P 1 x 1 star and that is a by b p 2. So, what we are saying that if income is less than a by b p 2 and of course, greater than 0 then he will buy in this case x 1 star is going to be equal to I buy p 1 because he will buy only good 1 and x 2 star is going to be equal to 0 fine is it clear. What if he has, what if it is more than a by b p 2 what happens in this case, x 1 star is going to be equal to a by b p 2 and how about x 2 star because this much is going to buy good 1. So, it is going to be I minus divided by p 2. So, can you say anything can you make any comment about this scenario that x 1 is kind of a necessary good, you know this person does not consume any of good 2 unless he has sufficient amount of good one, fine that is the situation described by this particular kind of utility function. Is it clear. |
Microeconomics | Lecture05_Rationality_Self_Interest_and_Optimization.txt | Let us endow our individual with 2 properties. The first is rationality, and the second is self-interest. It is not exactly endowment; it is like if we want to study individual systematically then we say that the individual is rational. Let us say if someone is acting irrationally. Someone is acting very, very irrationally; then of course, we cannot analyze that situation. For an example let us say that someone did something. And then you may contemplating, why that person has acted in this particular way. One simple answer could be this is the way he wanted to act. Everything can be explained on this basis, that this is the particular way this person wanted to act. And that is why he has acted in this fashion. So, here this explanation would explain everything in the world. So, in that way this explanation does not give you any more information that you already had. So, for this reason, this explanation is useless. So, that is why we want to fix one side of the story, that individuals are always acting rationally. Now it has two side. Because this rationality can be about the process, how a person has reached to this decision, or it is about result acting in a particular way, that gives you the maximum, the best possible result. And these two are not the same. These two are different. Sometimes a rational process may lead to, not the best possible result. But still, we talked about it in economics, we emphasize process rather than result. The reason is very simple. Because result, it is very-very difficult to internalize. That it is almost impossible to internalize; how did someone reach to that result. One can say this is the best result, and he should act to achieve that result, but that does not say anything about the process, the thought process that is going on. So, here in economics we are going to concentrate on process. The next term is self-interest. what do we mean by self-interest? So, this is the most fundamental assumption that we make in economics is that all individuals act in self-interested fashion. Now what do we mean by self-interest? It is like acting for your own benefit, for the self-benefit and that is self-interest. Someone may ask off immediately that, why you are making self-interest as the central assumption of this subject. I know many people, you can say I know many people who do not act to fulfill their self-interest. They act for the betterment of the society, but here in economics the common, the basic assumption is that people humans are self-interested. How they have reached to this level, it is not typically discussed. We take a person’s preferences, a person’s tastes as given. So, we do not say why that person has become so self-interested. We do not care about it. We take that person’s nature as given, and then we analyze. But here another way to look at it, this self-interest can be broad enough; that you may care for just one person or you may care for the society. So, when you care for society you become altruistic, but this also includes Here the narrow definition would say, here we can use it to say this is your this is what you want to do. Whatever, if you want to do well for the society, that is what you would like to do. So, here we are completely silent about how you have reached to this assumption, but we take this assumption as given. So now we have seen this rationality and self-interest. Rationality is about method that your methods would be well thought, should be well reasoned. It should not be guided by emotions. It should not be guided by your fear. It should be well thought and reasoned. That is what we are talking about method. And self-interest, it gives you an objective. What you do? You do anything to satisfy your need, your wants. So, it gives you objective. And these two leads to optimization. So, what is optimization? That here in optimization we assume that individuals try to do the best they can. And also, under given constraints. It is not that you can do anything that you want to. You have to consider the constraints. So, why do we have constraints? The typical reasons for having these constraints are scarcity. We have already discussed what does it mean. So, scarcity leads to constraint. Self-interest leads to this and this is related to self-interest. And here we are talking about how they try to do, but how using reason. Let us look at it again, at optimization why do we get into optimization, why do we get to this problem. That individuals try to do the best they can under given constraint. Let us look at the definition of economics once again. What do we say that what is the definition? Economics is the study of allocation resources, now let us look at it. What do we have economics is the study of allocation of scarce resources to satisfy individual wants. So, the problem here is that individual wants are unlimited. There is no end to yours’s or mine’s want. And how about resources? Look at resources are scarce in nature, resources are limited. So, you have limited resources, and from limited resources you want to satisfy unlimited wants. How could you do that? Of course, when you have limited resources, and your wants are unlimited or infinite the only way to do it is to figure out that which wants are more important than others. And then given the availability of resources, you would satisfy the wan which has higher priority than other wants. So, that is why you have optimization. And not only optimization, we have very special kind of optimization, that optimization where you need to consider constraints, because of scarcity. So, from here you get constrained optimization. More on optimization little later. We will talk about optimization in detail. |
Microeconomics | Lecture25_Perfectly_Elastic_and_Perfectly_Inelastic_Demand.txt | We looked at elastic zone of the demand curve, inelastic zone of the demand curve, and in the middle, we get a point where demand is unit elastic that is what we have discussed. Now, let us discuss the 2 extreme cases, the one extreme case is perfectly elastic, what it means is that you change price by tiny bit amount, and you get the maximum possible change in quantity demanded. Now what is that maximal possible change in quantity demanded that your quantity demanded is becomes equal to. Student: 0. 0 if price goes up by even tiny amount. So, let us look at the equation what happens; minus p by q delta q by delta p, and if we get the huge change if we have very, very small change in price. So, what we are saying that if we use this particular way to write it, that this is even if this is very, very small, and this is going to be very, very high. So, in that case, what should be the value of elasticity, when we have perfectly elastic demand; infinite it should be infinite, and how can you draw it, how would the demand curve look like. We have q on x axis and p on y axis, it is horizontal, and we will come back to this is a very special case of the demand function we will come back to it is very important, it is like in physics you never get frictionless word, but you have an idealised motion where friction does not work, similarly we will talk about this particular situation. We in reality it is never observed in any real-life situation, but we will talk about it a great deal just to understand how the market functions, but not today sometime later. And now similarly we have the opposite extreme; perfectly inelastic, and that is no matter how big the proportional change you bring in price, you will not get any response in. Student: change in quantity. Change in quantity. So, there what we have is, this is equal to 0, and it does not matter how much how big the change you have in price. So, in that case this is equal to 0, and if you draw you will get a vertical line, this is perfectly inelastic demand curve. Now, also just think about it what happens to the total revenue, in case of perfectly elastic demand curve. If there is a small increase in the price total revenue becomes equal to 0. Because as soon as you increase the price, the quantity demanded would fall to 0, I am talking about the highly idealized situation of. Student: inelastic. Perfectly inelastic, perfectly elastic. So, in that case it will be 0. And how about perfectly inelastic case. You keep on increasing the price your revenue will keep on increasing, because people would demand same quantity of this particular good at all the prices. Fine even if you increase the price quantity demanded would be the same. So, what is total revenue equal to p multiplied by q, p is going up while q remains the same. So, overall impact would be total increase the increase in total revenue fine. Now, one thing that you should notice that whenever we are talking about, elasticity price elasticity of demand. We are talking about, we are picking a particular demand function, and as it is demand function it is a downward sloping curve. It is always a downward-sloping curve. You know you can have a scenario, when you have perfectly inelastic then it is flat, but it would never be, demand curve would never be an upward-sloping curve. Now, let me talk about a scenario, let us say that this is an equation for banana, and apple and banana are substitute, and price of apple goes up what would happen to the demand curve? It shifts rightward. Now of course, if we bring here the supply function also let us say the supply function is this, earlier the price market price was p star, and the new market price is p dash star, can we talk about price elasticity of demand. Student: no sir. We cannot talk about price elasticity of demand, because here what has happened, the demand curve has changed, we have moved away from this demand curve, we have shifted the demand function now we are on a new demand curve. So, whenever we talk about elasticity when only we move along a particular demand curve while everything else is. Student: constant. Fixed; ceteris paribus and only the price of that particular good is changing, then only we talk about price elasticity of demand fine. Now, let us take another example just to understand. Have you ever heard of farmer saying that this year we had bumper crop? but that is why I am making the loss, the price has come down so heavily that I am making a loss, have you ever heard such a statement? This is quite common whenever we have very good monsoon very, very good monsoon farmers are not very happy, probably they have the same feeling which they express when they have very poor monsoon and drought. So, why do, can you explain this particular phenomenon, do you think farmers are lying or is there any economic reason behind exhibiting such a behaviour. So, tell me what is that economic reason. Student: sir because them supply is supply has gone up, because production of the every farmer has increased. Ok. Student: So, whereas, supply demand remains the same. So, what you are saying is, that demand for crop, demand for food item let us say for rice for wheat for maize is almost inelastic, you demand the same quantity of rice, in almost all situations. Again I am talking in a highly idealized fashion, not in the exact manner fine. So, if I can draw, let me draw it here demand curve is almost vertical, I can draw it like this, and let me say this is the supply curve in the beginning, p q this is supply this is demand, and when you have bumper crop what happens? Student: outward. Outward shift. Let me redraw it, let me say that this is the new one the earlier one is, this is the earlier one and this is the shift, because of better monsoon production will go up, what is happening to the market price. Student: it is falling. Market price is falling heavily, p has fallen and demand is inelastic. So, what happens to the total revenue? P has fallen of course; p has fallen so. Student: q also falls down. Q has gone up though q has gone up. Student: (Refer Time 09:35) To what happens to the total revenue. Student: increases. Now can I say the farmer farmer’s happiness is a function of not the amount of quantity they sell, but a function of total revenue that they earn. In fact, total revenue minus total cost, but right now we are ignoring the cost part just to understand. So, what is happening since demand is almost perfectly inelastic, it is not perfectly inelastic, but I am saying almost, p has come down. So, q has gone up, but how about TR. Student: it is decreased. It has decreased and that is the reason that farmers express unhappiness, even though they have bumper crop now this is one example bumper crop Next example I will not describe it in completely because we need more concept, but just I will give you an idea. In airline let us say you want to travel 6 months from now, and you buy a ticket from Kanpur to Delhi, and let us say you forgot to buy the ticket today, and you want to buy just one day before you travel, what would be the price can you compare the price of these 2 tickets. Student: yes sir. Which one will be more? Student: one day before. One day before and significantly more. Student: yes. The way airline tickets are priced, why do you think that airlines have this particular kind of pricing? Student: because sir at one day before the demand is quite inelastic. So, if the demand is quite high. Just demand is, one thing let me just tell you are in the right track, but let me just point out one thing, even when demand is quite high, but if price is also very high, quantity demanded would be very, very less. Student: yes sir. Demand high means, the demand function whenever we use this term demand, we are talking about demand function, not quantity demanded, these 2 things are different quantity demanded is a point on that demand function. Student: yes sir. Fine. So, when demand may be very high, but price is also very high. So, quantity demanded would be very less. Student: yes sir. So, that is not the reason the reason that he was talking about, that one day in advance the demand is very, very inelastic. You want to go tomorrow, you need to go tomorrow, 6 months from now if you are buying again airline would not be able to, basically also you have to understand from a little bit from their perspective. They do not know how inelastic a particular person’s demand is, but they have one way to distinguish that if a person purchases a ticket well in advance, it means that person is buying for some leisure travel, that he is planning for vacation and all. So that person has fairly elastic demand, if price goes up he would not travel, he would travel to some other place, or he may travel using train. So, that person has very elastic demand, but a person who typically buys ticket on short notice, they can say they may not know that person, it may very well may be that person has poor memory or poor planning skill, but what they think that majority of people who are buying tickets on short notice they are business travellers, and why are they buying tickets because they need to attend certain meetings, where they need to go somewhere for some work. So, their demand is relatively inelastic. So now think about total revenue when your demand is inelastic, and now here airline in this particular way airline finds a way to distinguish people who have elastic demand and who have inelastic demand. If airline couldn’t distinguish these two sorts different kind of people then airline will have to charge the same prices. Fine, but here using this airline can separate, segment the consumer into 2 different categories, people with elastic demand and people with inelastic demand. So, when I know, I can guess that you have inelastic demand, in that case to increase my total revenue what should I do, I should increase the. Student: price. Price and when I know you have inelastic demand what should I do. Student: has to decrease. I should decrease the price so, that is what airlines do. Student: Sir also the concept of time value of service is here. Like you are demanding a ticket now, and I am providing you right tomorrow. So, the time value you are giving me, it has a high time value right now than for 6 months. See again that would be do you think, if you want to let us say one way to include the time value for the logic sake, is that let me put it in more in econ in economic terms, that what you are saying if rather than using 6 month let us you used one year. Just for example, that I am putting I am giving 100 rupees right now, instead of giving one day before. So, this 100 rupees they can use it for their own work. So, one way to look at it that what is the value that the interest that you can earn from the bank, or interest you can earn in share market. So, roughly interest that you can earn by fare means, I can say it is around 10 to 15 percent in India in this particular time, but the price if you look at the ticket of the price tickets, the difference is not 10 to 15 percent it is sometime it is 4 times 5 times. Student: yes sir. Means 500 percent. So, the logic that you are giving can be an auxiliary reason, but not the primary reason because, if that is the reason then the change should be just 10 to 15 percent, but what I suggested that the demand is again you have to whenever you are explaining you have to look at the primary reasons, I am not saying that is not a factor that is definitely a factor. That is the reason that you give that is why when you pay for magazine, where there it would be more applicable when you pay for 3 years right in the beginning you pay significantly less amount, 30 forty percent less. Student: yes sir. One component is coming from the fact, that the money they get immediately that they can use to, generate more revenue for themself. So, that is why they are giving you discount. And the second when you pay immediately you are guarantying that for 3 years you will keep on buying, if you do not register you may buy you may not buy. So, for to compensate for these 2 factors they give you huge discount, even in airline that can be a reason, but that explains very small percentage. Fine, we digress. So, these 2 examples bumper crops and airline ticket we will come back to airline ticket, when we talk about monopoly later. |
Microeconomics | Lecture07_Introduction_to_Demand_Supply.txt | Welcome again to NPTEL lectures on microeconomics. Today we are going to discuss demand, supply and market equilibrium. I will begin with a simple story. A very simple story, a short story and then I will talk about some facts. And we will discuss why do we observe such kind of things in our society in our economy. So, the title of the story is, let there be cheapness. The story is from roman empire. Approximately around 302 A.D, there was an emperor named Diocletian, and what he observed that wherever his army marched, the price of food items all the consumables increased tremendously. So, he was very worried. He was very disturbed. He thought that sellers are trying to make easy money. So, he was very unhappy. So, he said let there be cheapness. In fact, he issued an edict, which reads that unprincipled greed appears wherever our armies following the command of public wheel march. Not only in villages and cities, but also upon all highways. With the result that prices of foodstuffs mount not only fourfold and eightfold but transcend all measures. Our law shall fix a measure and put limit to this greed. So, the idea of this simple story is; that he tried, he tried to fix the prices in the market through his edit. Just imagine can you think of what happened after he issued this edict, we will come back to this question. But I want you to ponder that what happened once he issued this edict. Another thing, another let us focus some at some other facts. Have you ever seen long line waiting to get gas cylinder outside a shop? Yes sir, yeah. Have you ever seen long line of cars waiting to get their tank filled? Yes sir. That is fine, also think about, think about just another scenario. Whenever let us say mangoes are in the season, what happens to the mango prices? Go down. The prices. Go. Go down. Of all the items, all the different kind of mangoes, but think about when there is a season for hotels; like, some are in similar, what happens to the prices of. They rise. Room? It rises, why? If mangoes are in the season, the price of a mango goes down. But when rooms are in the season, a price of a room or here, rent of that room goes up. What is happening? So, all these sort of things and also some other let me also add one more that these days it is in the news; that the food grains are rotting. Why do you think it is happening? Why do you think government procures so much of food grain that it cannot save it, that much it cannot safely keep, what is happening there? So, to understand all these phenomena, we are going to concentrate on the basic workhorse of microeconomics. Demand, supply and market equilibrium. And this is the title of this chapter. Demand, supply and market equilibrium. Let me write; in this chapter we are going to study six broad topics. The first topic is of course demand, what is demand? What do we mean by demand? That how many units of a good buyers are willing to purchase at a particular price. So, we are going to learn about it, the factors that influence the buying decision and so on. Similarly, we are going to learn about supply. How many units of a good. Produced in a. A particular seller is willing to sell at a particular price. And when we take all such price and quantity combinations, we get supply curve. So, after learning about demand and supply, we are going to learn about market equilibrium. And what is market equilibrium? A. A particular price where quantity demanded is equal to. Quantity. Quantity supplied. So, we are going to study all these things in more details. The fourth topic that we are going to cover is consumer. . And producer surpluses. What is consumer surplus? When you buy an item. Let us say when you want to buy an apple. Let us say the price of that apple at present is 5 rupees in the market. You can buy one unit of apple, one apple by paying 5 rupees. But that is not your maximum willingness to pay. It is possible that the price of this apple was 6 rupees, and you still have bought it. Or maybe 7 rupees, you may still have bought it. So, you have something called maximum willingness to buy. The money that you are willing to pay, maximum amount of money that you are willing to pay to get one unit of apple. That is your maximum willingness to buy. So, you are willing to pay something higher, but the price is, if price is lower then only you will buy. So, there is a difference between the maximum willingness to pay, and the price of that apple. So, when you buy this apple you are gaining the difference. For example, let us say your maximum willingness to pay is 10 rupees. And that apple is available for 5 rupees in the market. So, your value for that apple is 10 rupees. That is why if your value was 11 rupees you would have even paid 10 rupees 50 paisa to get that apple. So, maximum willingness to pay indicates your valuation of that good. So, when you buy that apple in 5 rupees, it means you are gaining 5 rupees out of this transaction. Yes sir. And that is your surplus from this transaction. So, similarly we are going to study producer surplus also. And then we are going to talk about social surplus also. The fifth topic that we are going to talk about, the applications of the concepts of demand and supply. And we will pick two particular applications, the first is going to be effect of taxation. Effects of taxation, and second application we are going to study about price control. I already gave you an example of price control through the story of Diocletian, let there be cheapness. It is the Diocletian was trying to control the price. So, we will see the repercussion of an attempt of controlling the price through mechanism other than generated by the market. Fine, and last, we will introduce the concept of elasticity. Because we are not only interested in how much money we pay to buy one unit of apple. We are also interested in how our demand of apple changes with a change in price of that apple. So, we will learn all those things in this particular section. |
Microeconomics | Lecture54_Utility_Maximization.txt | So, far we have been talking about different building blocks of consumer theory or consumer behaviour. We started with consumption set. We started with consumption set then; we talked about feasible set or affordable set, from there we moved to preference to describe taste and liking of an individual. And what we did? We developed axiomatic theory of preferences ok. And then the 4th building block is very simple it is again just psychological building block; which says that most preferred, it is the psychological assumption that the most preferred, affordable bundle is chosen for consumption; that is very simple. So, we have completed all these 4 buildings blocks consumption set, feasible set, preferences and most preferred this psychological assumption. So, the now next step is given the next step is given the constraints selecting the most preferred word bundle ok. Now of course, unless it is stated unless it is clearly stated what we will assume that all the axioms that we have talked about; rationality axioms, then continuity, then monotonicity, and then strict convexity will be satisfied by individual preference; from time to time some it will be explicitly or implicitly asserted stated that one of these axioms are, one of one of these axioms is violated. And then we will have to modify our technique accordingly, but right now I am talking about a scenario where all these axioms are satisfied, all these properties are satisfied. And then we are talking about selecting the most preferred bundle from the set of affordable bundles fine. So now, if you remember your lesson on budget set typically for most of the problem what we figured out that the budget set is unless we have any more constraint let us say 2 good world on x axis we have x, y axis y; x denotes the amount of good 1 y denotes the amount of good 2 then the budget set is like this ok. And of course, here we are taking x is greater than or equal to 0 and y is greater than or equal to 0. You cannot consume negative amount of any of these goods and what we had figured out earlier that our budget set will be given by this triangular region. Now, what we are going to do is, here it is the budget set now let us draw the indifference curve. And this we are getting from feasible set that is our building block number 2. And here we are going to describe preference of a particular individual with help of indifference map ok. That will indicate the choice the the likings of a particular person; and if let us say preference satisfies all the axioms then the typically qualitatively speaking the indifference map is going to look like this. And also, we know that utility would be increasing in this direction. If we superimpose this diagram, this is 1 and this is 2, instead of drawing a drawing it separately if I superimpose it on one another how would it look like of course, it is not to the scale. So, just qualitatively I am putting it here. So, what I have here is like this, this is the budget line and inside is the budget set including the bundles on this line. And here we have something like this this is the; it is increasing in this direction let me it is increasing in this direction ok. This these indifference curves they represent preference of a particular person. Let say if this person is trying to maximise his happiness. Why do you want why do you buy a particular bundle? Remember we talked about the 4th we talked about it in the 4th building block. That a person selects most preferred bundle from the affordable set; so, that is what; so, I can I can translate the, I can translate this consumer problem into, I can say that a consumer wants to reach to highest ranked indifference curve given his budget or other constraint. Although I talk about other constraint, but in graph and in mathematically we deal with only the budget constraint. So now let us look at this diagram. What is the aim? Let me draw few more indifference curve what is the aim of this person? Of course, the value let us me put here random value 10, 20, 30, 40, 50 and 60 and so on ok. You can give them different values because utility is ordinal. So, these are just ranking what it says that 20 is more than 10 nothing more nothing less. We cannot say that this person prefers bundle on indifference curve given by value 20 twice more than a bundle on indifference curve given by value 10 we cannot say that we can only compare. So, here it does not matter whether I say 10, 20, 30, 40 I can say 1, 2, 3, 4, 5, 6 does not matter ok. So, what is the aim now? Let us say how can this person maximise his utility; in other word how can this person get maximum possible satisfaction maximum possible happiness given his budget constraint ok. How is it possible? Which bundle should he select? Tangent Tangent so right now let us not make it fancy with the word tangent. Let us let us rule out certain scenario first. Let us say that this point is here we have y here we have x. Let us take this say this point is 1, this point is 2, this point is 3, and this point is 4, this point is 5 ok. Just I have selected few points not so randomly because I want to illustrate how it is happening. So, do you think that if this person selects bundle 2 he would achieve maximum level of satisfaction 5, no why? 2 why? Because he can still afford We can still afford more. So, the thing is here the basic you should you should pay attention that to this zone, this zone ok, not the zone. Let us look at this part ok, what is happening? That I have drawn axis taking 2 as origin and the axis are parallel to the earlier earlier axis. What is happening? That in this direction if you move, what will happen? This person will get higher satisfaction more utility that is from monotonicity and in this zone that the zone that I have coloured I have shaded with the green colour, what is happening there? That we are taking this indifference curve. Let us say indifference curve 10. What is happening there? That in all the green zone because this indifference curve is dividing the whole of consumption set into 3 parts, 1 is this part where all the bundles are less preferred than the bundles on this indifference curve denoted by number 10 ok. Similarly, but in this zone second is that all the bundles on this indifference curve and third is all the bundles above this indifference curve here this person gets higher utility, but what is happening? Because some of these bundles are still affordable. So, this person can move in this direction and increase his happiness is utility. So, 2 is not the bundle that we are looking for. Second thing that we have figured out using the concept of monotonicity is that the bundle that would give maximum happiness has to lie on the budget line because, if let us say let us look at it let us say that we are picking this is the budget line and we are picking a bundle not on the budget line and it is of the bundle is affordable we will pick here here here somewhere. And there we can of course, move in this direction and that is, north I think it is northwest direction actually northeast direction we can move in northeast direction and utility will increase. So, the optimal bundle has to lie on the budget line ok. Mind you it is true only if, only if preference satisfies monotonicity axiom if preference does not satisfy monotony axiom, then we cannot say that this is true you understand fine this is one thing that we have figured out. How about bundle 4? Bundle 4 is not affordable; may give higher utility, but it simply we cannot buy it how about 1 and 3? Same as 2 Same as 2, the 1 and 3 bundle, the bundle at point 1 and 3, they give you the same satisfaction level as 2 if you look at this map the way it has been drawn ok. If you look at it here 1 2, but why can not we select 1 and 2 while 1 into both are on the budget line what is the problem there? Sir we can get a bundle with which has more utility than 1 and 3 and even that monotonicity we can say that by 2 extreme cases 1 and 3. Yes, using convexity if we take 1 and 2 and we connect them then all the bundles on this line are more preferred than any of these 2 bundles. So, we we would not so where we will get basically, where if if the indifference curve is crossing the budget line at 2 points then we cannot get the our optimal bundle on that indifference curve is it clear. So, there is only one possibility left at our optimal bundle where the indifference curve touches the budget line at only 1 point and that point will denote the bundle that gives the maximum have possible satisfaction to this individual fine. And that bundle is of course, in this graph it is 5th. |
Microeconomics | Lecture26_More_on_Elasticity.txt | Now, what we have done is we have studied price elasticity of demand. Similarly, one can study price elasticity of supply. And what is price elasticity of supply? Again, remember what is elasticity basically? Proportional response It measures the proportional responsiveness of one variable with respect to proportional change in the other variable. So, here now we are talking about price elasticity of supply. So, what we are studying, the proportional change in the supply not just supply function. The proportional change in quantity supplied with respect to the proportional change in price. So, I can write here this is Q if you want to distinguish this with respect to price elasticity of demand you can use the different notation you can say it is eta. Epsilon is for price elasticity of demand and eta is for price elasticity of supply. This is and here also if you want to distinguish you can put s superscript just to say that we are talking about quantity supplied not quantity demanded. I am not going to talk about lot more for this price elasticity of supply because mathematically speaking they are basically the same concept. The 2 differences that you should keep in mind. That one that demand function is a downward-sloping function while supply function is an upward-sloping function. So, you should not put a negative sign there to make it positive, it is already positive. And second is also related the same thing just the restatement that P goes up Q goes up, that is the natural response of supply. Natural response of a buyer was that P goes up Q comes down. So, qualitatively we establish that demand is a downward sloping curve, but we want to also understand the responsiveness, the proportional changes. Here P goes up Q goes up. So, eventually, you will get it a positive. So, do not put it a negative sign and although again we will not about not talk about new cut offs, you should have the same cut-offs in the mind. That eta is equal to 1-unit elastic, eta greater than 1 elastic this is just a repetition and revision, and this is less than 1 inelastic. How about here perfectly, perfectly elastic supply? How would it look like? Again, horizontal, a perfectly inelastic supply vertical. This you will never observe. You know you will never observe this. This is just a theoretical construct that you will never ever observe even in approximate manner you will never observe this, now let us talk about now what we have already said that elasticity, we are measuring the proportional responsiveness of one variable with respect to proportional change in other variables. So, just talking about demand and supply we can bring more such variables in the picture. So, let us do that. What we can do? Talk about is something called cross elasticity. And now I will not talk in terms of supply only in terms of demand concepts are very similar. The cross elasticity of demand. What we are talking about is the price of that good remain the same, but the price of other good has gone up or price of the other good which are substitutes or complement they have gone up. So, let us say cross elasticity this is denoted by Exy, why we are putting Exy? We are talking about change in quantity of good one even by x proportional change of quantity of good 1 with respect to change in price of y good 2. And that we have here fine? Here we will not put minus sign or anything. It can be positive it can be negative. It is possible that price of one good is one good goes up and its response on the other good is that the quantity demanded of other good comes down, when does it happen? Substitutes Let us look at it, very good let us look at it substitute, substitutes and complement. Let us talk about x and y and x and y. Here we have Py x. If they are substitute if Py goes up what would happen to the quantity of good 1, if Py goes up x will go up. So, in this case what will be the cross elasticity? Positive, and if when they are compliment cross elasticity is going to be negative. One more thing that you should notice here whenever we are talking about cross elasticity of demand we are talking about shift in the demand function. Either in this direction or in this direction. Why there is shift in the demand function? Because we already know that if one of the variable which effect the demand function changes, its impact is a change in the or shift in the demand function, if that variable is other than its own price fine. That is why we are talking about shift here. So, here we will not insist on having a movement along the same demand function is it clear? So, do not get confused between these two. Similarly, we can also talk about something called income, income elasticity of demand. Can you tell me what is it is equal to? delta x by x. Delta x by x divided by delta I by I, where delta x denotes the change in quantity demand and delta x by x is proportional change in quantity demanded proportional change in quantity demanded and what is delta I by I? I denotes the income. I denotes the income and delta I is proportional change in the income fine we can also write it by I divided by x. Delta x by delta I, this is income elasticity of demand. So, is it positive or negative can you tell me? Always remember we talked about 2 different kinds of goods normal good different good. We have not even discussed yet normal good and inferior good for normal good what we will observe. positive income. Positive, income goes up you consume more of that particular good again also here when we are talking about income elasticity of demand what we are talking about is shift in the demand curve. Similarly, for inferior good, what do we observe negative income goes up you consume less of that particular item. |
Microeconomics | Lecture31_Incidence_of_Tax_Four_Extremes.txt | Now, we already have learned the concept of elasticity. Now, I am going to talk about certain extreme scenario and I will let you to interpret what happens in the middle cases. Now, I would say in all these cases just it does not matter whether it is imposed on buyer or it is imposed on seller, but just to understand let us say that the tax t, unit tax of t is imposed on seller. Now, let us see what is happening. We will consider 4 cases elastic demand, inelastic demand, elastic supply, inelastic supply. How would the demand curve look like when we have perfectly elastic demand? I should as add here the perfect word here, perfectly in all perfectly elastic demand, perfectly inelastic demand, perfectly elastic supply and perfectly inelastic supply, these 4 extreme cases we are going to talk about. So, when we have perfectly elastic demand how does it look like? Horizontal, so it is horizontal. How about the supply curve? Supply we can pick any supply curve and tax t is imposed on the supply curve what would happen? Shift upward. Upward. That we have just now learned, shift upward, fine. Yes sir. This is the shift, equivalent to t; this was the equilibrium quantity Q star earlier, this is the new equilibrium quantity Q star dash and this is P star, fine. Demand curve is not shifting; supply curve has shifted upward, what is happening in this case? The Q star. Increasing. Has decreased. Q star has decreased, fine. How about P star B? Not changed. It have not changed, no change. How about P star S it has decreased. Decreased. By t. T. Let us look at it. How do we calculate? From here, we moved here. See, now let us look at look at it through logic, just not graph, look at it through logic. What is happening? Demand is perfectly elastic, what it means is that price goes up by infinitesimal amount and. Change in quantity. Buyers would not buy even a single unit of good, fine. Now, supply curve has shifted upward, the usual thing that seller can do that seller would pass the whole of the tax to buyer; when the seller passes on the tax to buyer, what would happen the buyer would not buy single unit, what would happen to revenue? 0. 0. So, seller would decrease the price. Let us say, if they decide to pass on half of the tax to buyer still buyer would not buy any of these good. So, seller does not have any other option, buyers are not willing to buy these goods at price higher than P star. So, in this case seller has to bear. (Refer Time: 04:57). The burden of entire tax and in graph you see, here Q star dash is the new equilibrium, this is the new supply curve to get the price that seller would get, we will have to move to the original supply curve from here and this much P star minus t seller would receive. So, in this case the entire tax burden is borne by. Sellers. The seller, fine; so case 1 we discussed the perfectly elastic demand and we saw that in this case the entire tax burden falls on the seller. now, we are talking about case 2 and that is when we have perfectly inelastic demand. Perfectly inelastic demand means? Demand is vertical, this is demand and supply is any supply curve we take up and any upward sloping supply curve that we are taking and when tax is imposed supply curve shifts upward. So, this was the original P star and the new this is P star dash or P star new, Q star remains the same. So, here are 2 scenarios on top we had perfectly elastic demand, what happened that Q star decreased, P star for the buyer remained the same, P star for seller also decreased. Now, see let us see what is happening here. Q star remains the same, it does not change. We will immediately get into just in a moment we will get into why? Now, what is happening to P star for seller? And what is happening to the P star buyer? P star seller is not changed, yes sir P star buyer is increasing. So, how do we find out P star seller? Now, at the new equilibrium quantity that is Q star it is the same, we take a vertical shift to move to the original supply curve. So, when we take that shift and this shift is equal to t, we go back to the original point. So, P seller is getting paid even in the new scenario it is still P star. So, remains the same, but how much the buyer is now paying, that is P star dash, so it has gone up. Now, let us see what is happening rather than using graph. Let us use, let us look at it using logic. Because of tax the marginal cost of sellers have gone up, if for example, if sellers were willing to supply 5 units of good at 5 rupees per unit and now let us say 1 rupees tax government has imposed on seller then seller will be willing to supply 5 units at 6 rupees cost, 1 rupee increase. Now, what is happening on the buyer side? The buyer side demand is perfectly inelastic, it means that price goes up quantity does not change. I am not saying market works in this particular fashion, but the earlier example that I gave that seller would try to pass on the tax to buyer and if seller is passing on this tax to buyer it means the new equilibrium price is just t amount above the earlier price, buyers would buy the same amount of good because buyers demand is perfectly inelastic, it is a vertical curve, buyers they buy the same amount of good at all the prices. So, here seller would be able to successfully pass on the tax to the buyer. So, who is paying the tax here? Buyer. Buyer, now we have, so far we have looked at 2 cases, in first case what we had was perfectly elastic and in the second case we had perfectly inelastic demand and when we had perfectly elastic demand tax burden fell on the. Seller. Seller. Seller. When we have perfectly inelastic demand then buyer had to pay the whole tax. So, can we get some idea that it has something to do with the. Elasticity. Elasticity, it will become more clear when we look at 2 more examples, fine. Case 3, when we have supply is perfectly elastic; perfectly elastic supply means a horizontal curve and we take any demand function, any downward sloping demand function. If government imposes tax what happens? Downward shift. Downward shift or upward shift, think about it. When supply curve is perfectly elastic or it is horizontal what does it mean that marginal cost is same at all point and let say that marginal cost is in this case P star, when government imposes a tax of t unit a specific tax of t unit then what happens marginal cost goes up by. T. T. So, supply curve will shift. Upward. Upward, exactly t amount, it shifts upward exactly t amount and how much is the shift? This shift is t, is it clear now? Why it will shift upward? No, I am saying that here I said right in the beginning sorry if I said it is a demand curve, then I was wrong because we are talking about scenarios where tax is always imposed on seller those 4 cases we are discussing. You can do it same, when tax is imposed on buyers, but here right now we are talking about scenarios where tax is imposed on seller. Now, what happens? Let see, that earlier equilibrium quantity was Q star, the new equilibrium quantity is Q star dash, this is P star plus t. So, Q star decreases what happens to P star S, let us look at the graph at this new quantity Q star dash. If what is the price that seller is getting? P star; so remains the same and how about P star b it is rises by t. So, again who is paying the tax? Buyer is paying. Buyer is paying the tax or can I say this, the side which is relatively less in less elastic is paying the higher proportion of tax. Here, it is the perfect it is perfectly extreme case where supply curve is perfectly elastic, so supplier does not have to pay any tax, the perfectly elastic side does not pay any tax, is it clear? So, here supply perfectly elastic supply and the 4th case, the logic that I gave you in words that you can apply here too in this case also, but let us move on to the 4th case, we have downward sloping curve and perfectly elastic. Inelastic sir. Perfectly inelastic supply function, what does it mean? That supply function is the vertical. Vertical. This is supply, this is demand. Now, a tax is imposed on seller what would happen here? The supplied, the new supply curve is the same as the earlier supply curve. Yes sir. Why? because this curve has to shift upward, by how much? T. Yes sir. It is the same thing when you shift a vertical curve upward you get the same curve. What it is saying? That at all price this person would supply the same amount, fine. Is it clear? Now, what is happening to the equilibrium quantity and equilibrium price, Q star remains the same no change, how about P star b? Remains the same, does not change? How about P star s? Decreases by t. Decreases by t and it is difficult to see in this graph, but what is happening that because there is another graph super imposed on this and this graph is P plus t. So, earlier that the new equilibrium point does not give P star s, but it is giving P star S plus t, which is same as the earlier P star, so what is happening to the equilibrium price that seller is getting, it is t unit less than the earlier equilibrium price. So here, the whole tax the complete burden is borne by. Seller sir. By the seller and seller is inelastic, perfectly inelastic supply, fine. Is it clear? Yes sir. So, here we have taken the extreme cases. Now, let us talk about you know in between cases, when we do not have perfectly elastic demand and perfectly elastic or inelastic supply what happens? Can I say again I am saying heuristically without proving it without showing it to you, the side which is relatively more elastic pays relatively, share relatively less burden of the tax. The side which is more inelastic, shares the larger burden of the tax. Is it clear? Again look at the word elastic and inelastic, what does elasticity mean? perfectly elasticity represents the price responsiveness basically. So, elastic means that you have extreme high responsiveness. So, if price goes up by even little bit, your response in quantity is great. So, here demand is a downward sloping curve, so let us look at from the demand side, price goes up by little bit your demand goes down considerably, that is supplier would not prefer. So, what would they do? They would share the burden higher burden of tax. So that consumers do not decrease their consumption significantly. So, the tax has, who would, the incidence of tax in economic sense, in real sense it depends on the relative elasticity of demand and supply, fine. |
Microeconomics | Lecture11_Factors_Affecting_Demand.txt | Factors affecting the demand function. Can you think of any factor that affect the demand function? Some we have already covered, some we have covered in today’s class. The first session we have covered, can you think of some of the factors? Income income. Income how does it impact, what happens? Income increases demand increases. Income increases demand increases. So, what you are saying that we have here price on y axis Q on x-axis, let us say this is the demand curve, downward sloping function, and then if income goes up, this is what we get, is it true? It is true in most of the cases, but not always, let me say there are two kinds of goods; normal goods or normal good let us write and. Inferior. Inferior good; when quantity demanded rises with an increase in income, then we call that good as normal good. Remember what we are saying, when we were talking about demand function, we said that everything else is held. Constant. Constant and we talked about change in the price of that particular item but hear me what I am saying now I am saying the price of an item is fixed and income is changing. So, of course, if income is changing what would happen? Shift. We would not have movement along the curve, but we will get shift in the curve. And there are two types of shift possible, one that quantity demanded at the same price would go up or go down or would remain fixed, you can say that is another time, but let us consider these two; either it goes up or it goes down. So, when quantity demanded goes up at the same price with an increase in income, then we call that good as normal good. When quantity demanded at the same price decreases as your income goes up, then that good is called inferior good for yourself. Can you give me any examples of normal good and inferior good? When let me say again it depends on your income level and again it depends on parameters. So, I am just creating a case, that may not be always true, but this particular case illustrate the inferior good in the market. Let us say you are a very poor person, probably you spend a significant proportion of your income on potato, you buy potato. As your income goes up you become capable of buying some meat. So, as your income would go up your consumption of potato would decrease. So, in this case, potato is clearly an inferior. Inferior. Good, but that is a possibility that your income goes up and you like potato so much, that your consumption for potato goes up. In this case, if this is the case then potato would not be an inferior good for yourself, fine its clear, and normal good is simple that you know, more income you have more meat you will buy. So, income definitely affects your demand. So, one factor is income, how about the second factor? So, income we discussed today, even before we started discussing about factors affecting the demand function. So, think about it we discussed some other factors also, in course of discussing demand function. Alternative. This is for, let us say here income goes up, then this is the case of normal good. Availability of Fine the second one is availability of. Alternatives. Prices of substitutes and compliments. So, let us say the price of a substitute goes up. What will happen to the demand for good x? Increase. Increase increase. So, it will increase and if y is complement. Decrease. It will decrease. It is clear? Just to understand what we are talking about that when price, let us say we earlier we learned the tea and coffee are. Substitutes. Substitutes. So, price of coffee goes up, now remember we are not talking about individual demand we are talking about market demand. So, price of coffee goes up, more and more people would opt for tea, at the same price. So, at the same price quantity demanded for tea would go up. So, the increase in price of coffee would shift the demand curve for tea rightward. So, substitute, how about for compliment that you can generate on your own, that is part of the homework that what happens when the prices of a compliment goes up, any other factor? So, what we had discussed income prices of substitute and complement, any other thing? Diminishing value, we already right now we remember we are talking about effects on demand function or demand curve value, you know when we talk about change in price of that good, we get movement along the curve. Right now we are not talking movement along the curve we are talking about change in demand function; it means we are talking about shift in demand function, the factors which would cause demand function to shift and we have figured out income as well as complement and substitute prices. What else think about. Let us say we are talking about demand for today, let us say we are talking about quantity demanded of petrol today in the market. If we expect that government of India is going to increase petrol prices by tomorrow, what do you think, what would happen to the demand function today? Increase. It would increase. So, your expectation about future also impacts your demand function. Expectation is also quite important, what else? Or rather not just preferences change in preferences. Although in later chapters I would say that preferences would not change, but right now that is again an assumption, I told you right in the beginning we make lot of assumptions in order to learn something about complex world. So, later on we will assume that preferences do not change over the period, but we all know that I may like something today and tomorrow I may not like, and typically it happens in context of fashion goods. If something is in fashion today the demand is quite high, it goes out of the fashion, demand decreases. So, change in preferences and last or you can also think of more, but the last one that I am going to discuss is number of buyers. One thing you should notice that all other factors above, I described number of buyers they impact not only market demand, but also impact individual demand, but here now we are talking about number of buyers it does not impact individual demand, it impacts market demand. So, if we have more and more people in the market buying the same good, what would happen to the demand curve? As the number increases demand would. Demand would increase. Increase. So, it would shift, it would shift again right ward. Let us say n is the number, as n increases. Is it clear? So, that brings closer to the concept of, at least we had described the concept required to understand demand schedule in adequate detail, now let us move to supply any question you have about demand? Sir just one question. [FL]. Like we can discuss about improvement in technology like if technology improves like say a processor technology will improve, that will change in production function. So, would that, that would also effect the demand function. We will see, let me let me understand your question clearly what you are saying that if for some reason technology improves in one particular sector, would it impact our demand? It sometimes, it does impact our demand, but through a particular channel that we have already described probably, it changes our preferences, for one example earlier you did not had LCD TV, you had CRT TV, Big TVs. Your demand for LCD TV in other word was equal to, equivalent to 0, that is because of the introduction of LCDs you learnt about it and your now you incorporated it in your tastes, in your preference and of course, now there is a huge demand for LCD TV. So, what is basically happening, it impact your demand, but we already took care of it here, change in preferences. Sometime technology brings change in preferences. But if you just say, because of change in technology, demand does not change. Supply. We will learn later on even just now that change in technology brings change in supply, not change in demand. If it brings change in demand through change in preferences, your taste changes over time because you see new technology in the market. But demand is changing not because of change in technology but change in your preferences, that is the way we will take it. sir change in technology does it mean that we were producing the same good by a efficient technology or like you told LCD tv and big tv. So, sir if change in technology then we will consider that big tv is only produced with different technologies. You see again we are using this term technology vaguely, typically you know what you are saying typically when you are saying technology here, what we mean is that the same thing can be produced by better technology, and in that case that it does not affect demand at all. Yes sir. It does not affect demand at all. Yes sir. So that is what I said. So, if you just say technology, it does not affect the demand, but if it changes your preferences somehow then it may or may not affect your demand function. |
Microeconomics | Lecture41_Preferences_as_a_Mathematical_Construct.txt | We talked about three properties and those three properties were completeness, reflexivity and transitivity, and what did we say that if our preference satisfies these three properties then we can talk about our choice, our selection of a particular bundle of good in consistent fashion. And I had also given you an example if preference does not satisfy a particular property transitivity, then in a way you will be losing money that I had shown you. So, before we begin new thing I want to satisfy that a mathematical, I want to describe a simple mathematical construct that what is preference. We talked about it in economic sense. let us talk about it in mathematical sense. For just few minutes can we call preferences a kind of relation. What is the relation by the way, if you do not know about it, it would not hamper your understanding of next topic, but if you understand this it is fine. Relation is nothing, but a subset of Cartesian product. What is Cartesian product? X X multiplied by Y; where X and Y are two different sets. Here what we are doing? We are picking two elements from consumption set. These two elements represent two different. Consumption bundle. Consumption bundle and. So, I can say that we are talking about a very particular kind of relation that is defined on X multiplied by X, or in other word this preference is defined on set X, and what does it mean? Defined on set X means this is a subset of X multiplied by X. Relations are very simple like Mohan is a brother of Sohan. So, its a relationship, Mohan is related to Sohan or Romeo loves Juliet, its a relationship ok. So, we are trying to represent those relationship using mathematical notation. So, here we are talking about preferences in terms of particular mathematical relation. What kind of relation that a set, an element of a set is related to the same element or the another element of the. Same set. Same set. We are not changing the set. So, consumption set is remaining the same fine, and we talked about 3, that this particular relation should satisfy three different properties; completeness, reflexivity and transitivity. So, we say just a definition that I am going to use. We say when a preference relation satisfies these three relations, then we call this preference relation rational, when a preference relation or of course, we are talking about, or any relation which satisfies these three accents, then we call that relation. Rational. Rational. So, to satisfy rationality in economics when I say that Mr. X has rational preferences or in short Mr. X is rational, what do we mean that Mr. X preference satisfies completeness, reflexivity, and transitivity axioms, these leads to rationality. Let me ask you, forget about economics for a moment, when we talk about rationality in day-to-day life, what do we mean when we say that Mr. X is rational? Every decision we take is for our benefit. I am not I think, for. Our benefit. For. you are talking about yourself. So, you are rational when you take if all the decision for your own benefit I would say that term would be people would say that you are a selfish person not a rational. Sir he is logical. He is logical Someone who is logical then he is called. Rational. Rational someone who is sensible someone uses reasoning to reach to a decision or someone who has sound judgment then we call that person rational in our day-to-day life, but in economics what we are saying Mr. X is rational if his preferences satisfy completeness, reflexivity and transitivity properties. So, are these two the same thing or they are different? |
Microeconomics | Lecture46_Properties_of_Preferences_Continuity.txt | Now, as we have seen that when we have finite number of consumption bundle in the consumption set, it is very easy to rank, we do not need any more we can rank them. But what we have remember go back to your, the definition of the properties of consumption set. What does it imply? That consumption set would have infinite number of. Bundles. Bundles. How long will you go on ranking? Therefore, we have utility functions. So, we know, but utility function again even from if you want to look at the prospective of utility function then again we should have some particular properties in the utility function. Properties such as continuity, you might you have heard about these in mathematics continuity and differentiability. Why do we need these properties? So, that we can use the optimization technique that we are going to learn shortly even in this course. So, you can utilize them because using calculus people have developed optimization technique. So, utility function should be continuous in nature. But so far we have not talked about the utility function should be continues and so in that sense we have not talked about the ranking of infinite number of bundles, it would be herculean task. Remember when we are talking about preferences we are not talking about the budget set, we are not talking about the feasibility set. So, whenever we are talking about preferences we want to rank all the bundles in the consumption set and if you have an infinite number of bundles then just using the earlier three properties it would be impossible to rank all the bundles, we can try, but we will keep on going and going because we have infinite number of bundle. So, we look for certain specific properties. We are not saying that everyone should have, everyone’s preferences should satisfy this particular property but if it satisfies, the tools we are going to learn in the future would be applicable. If this property is not satisfied that some of the tools that we will learn in the future cannot be used. We will have to go back to the first principle. I am not saying that we cannot solve the problem, if you have done calculus and calculus we are going to use extensively remember whenever you have to differentiate you have certain thumb rules. But whenever you get a new function what do you do go back to the first principle. Principles So, similarly something we do not call something first principle here, but the idea is very similar. So, we are developing we are trying to develop certain tool kits which would be applicable when your preference satisfies these certain properties. I am not saying that there in the world everyone whoever exist has this kind of preference, no I am not saying that. But if you have a preference which does not satisfy these property then either you will have to go back to the first principle and if they are inconsistent then you cannot even talk about it fine. So, the new properties that we are going to talk about is continuity. And what is continuity? You can define a function at each point. You can define a function at each point. There is no break with the fixed value, at a fixed value. That is not continuity. But the definition, sir in Maths what we learnt is that we draw its graph without lifting a pen without that is. That is very basic definition that you learnt in class eight class ninth that when you can draw a function. Two value. Of course, then in that case function has to be a two dimensional function without lifting your pen when the function is continuous. That I am pretty certain that you must have come across more rigorous definition that is not the idea right now, but what here what we are talking about these two are relative the definition of continuity in function. And here what we are saying let us take two bundles x and y and let us say x and y are ranked in a particular fashion either x is ranked above y or y is ranked above x, does not matter. Now, if we pick a bundle that is sufficiently close to x and let us call x 1 and a bundle that is sufficiently close to y, y 1. So, then your ranking of x 1 y 1 should not be different from ranking of. x and y. x and y fine. Yes sir. That is in terms of preferences. , ok. If we are now we have learnt the concept of utility it means x is assigned a number and y is assigned a number. And if we pick x one that is sufficiently close to x then the number assigned to x 1 should be very very close to. x. Number assigned to. x X. So, what it does that it preserve the, continuity is about preserving closeness bundle which are close to each other should be ranked near each other. Let me give you an example. So, let me say here, let us take again two-dimensional world good 1 and good 2 and let us take a bundle here and a bundle here and a bundle here. Let us say this bundle is assigned a number 5, this bundle is assigned a number 100. So, then this bundle should have a number nearer to. 5 5, because it is closer to. 5 5, let us say 6 or it may be four that is what that is not what we are talking about here whether it could be 6 or 4 should be close to each other. But if it is 110 then definitely. It would be. This preference is not. Continuous. Continuous fine, is it clear. Yes sir. So, now just one thing I want to add here. We talked about utility function and what we said, then when your preference relation satisfies completeness reflexivity and. Transitivity. Transitivity, then we get utility functions such that it preserves the rank. Now we talked about one more property, that is continuity. (Refer Slide Time: 07:23) So, here the thing is that any preference which satisfies rationality axioms. What does it mean? Satisfy rationality axioms means all 3. All 3 those properties. Completeness, reflexivity and transitivity are satisfied rationality axioms and continuity axiom can be represented by a utility function that is continuous in nature or you can say transitivity represented by a continuous utility function, fine. What we just talked about is that any preferences that satisfies continuity axiom or if you want to call it assumption then assumption is also fine continuity axiom and rationality axioms, there are three axioms if all these three axioms are satisfied then we say that preferences preference of a person satisfies rationality axioms. If rationality axioms and continuity axioms are satisfied then preference of a person can be represented by a continuous utility function. And when we have already learnt that when we figure out one utility function to represent someone else preference, then what we can do? We can do we can take any monotonic transformation of that particular utility function and we will get a new function that will also be a utility function in the sense that it would represent the same person’s preference, same person’s taste. So, utility function is not unique whether it is continuous or not fine. So, earlier we talked about you know continuity is nice mathematical property we can if a function is continuous and differentiable, then we can use optimization technique as I told you that we will be using some optimization technique. But now let us say you got a utility function that is continuous, but that is not differentiable. What can we do? We can always transform a utility function using some monotonic transformation into a differentiable function a continuous function can always be transformed into differentiable function using monotonic transformation. So, we are not worried about differentiability of the utility function, is it clear, fine, clear? |
Microeconomics | Lecture34_Budget_Line_and_Budget_Set.txt | So, we talked about 4 basic building blocks of consumer theory we talked about consumption set or choice set that is all about what you can conceive in your mind. It is not limited by any of the constraints that you face, it is all about what you can conceive in your mind and whatever bundle or coming that is bundle is a combination of goods whatever combination of goods that you can conceive in your mind would be part of this consumption set and then from conceivability we move to the affordability. And affordability has something to do with the either monetary constraint because of monetary constraint you cannot access certain bundles. So, the bundle that you cannot afford because of monetary constraint that bundle cannot be a part of your feasible set or if we are not if we can also call it the budget set what whatever bundle that fits in your budget would be part of this budget set and here the key word is affordability as opposed to conceivability in consumption set. And third we talked about preference relation that that is about ranking of all the bundles. So, that given your constraint you can choose the bundle that you like most and lastly the 4th block that we talked about is the human behavior that here you choose the highest ranked bundle. Now, conceivability of course, it depends on what you know of if I talked about a world which has only one good available then you cannot conceive of any other bundle, you can conceive of bundles, different bundles which contains different amount of the same good. But that is not a very interesting example because there is a typical cutoff point let us say it is about money, then let us say if I have all the fish that you can buy and you have let us say if you are this is the only one good available in this world. So, of course, you can buy up to the certain point x and then you cannot afford beyond x. So, this is affordable, and this is not affordable fine this is very simple. But one thing I want to mention that the bundle here, the fish bundle here at this point, and fish bundle here the unit here there they are the same they are interchangeable they are perfect substitutes of one another this is a big mistake that we make in our day-to-day life. Let me give you an example from cricket because that is the most popular game in India. So, when let us say India needs to win, India needs to score at least 254 run to win a game against Pakistan and then we say that person who scores the 254 run he says that he contributed to India's win. But just think about it the person who contributed the first run or the second run those run have equal value as the last run. So, those two runs are interchangeable they are the same, it does not matter. Similarly, let us say if you want to buy shirt and you are short of let us say 100 rupees and someone gives you the 100 rupees then you say I bought shirt using his 100 rupees. It does not make any sense if you look if you think about economics perspective because if you are using 100 rupee you could use that particular 100 rupees for some other work and the 100 rupees you are going to use for the other work, you can pull it in for buying the shirt. So, all these 100 rupees are interchangeable. So, in other words they are the word use is called fungible. So, never ever get confused about it. You can exchange although why I am mentioning it because time and again I am going to talk about the last unit, but in terms of your the gain that you get from either the first unit or the last unit day you can exchange them the order can be exchanged anyway that is for the one good world. Now, let us talk about two goods world and there we can talk about the budget set affordability because it is not as trivial as the earlier one just to make it interesting. Let us imagine an island economy and this island is populated by Robinson Crusoe and he is maroon on this island. He has access to either, let us say coconut, he has access to coconut, or he can use his time to catch fish. Let me denote coconut by C and fish by F. Let us say that in 24 hours he cannot work for more than 8 hours. So, total time available to him is, to Robinson 8 hours, and let us say that it takes 1 hour to catch 1 unit of fish and it takes 2 hours to gather 1 unit of coconut. So, the all the combinations that he can consume, what is happening let me say here I may reverse it but let us say 1 hour for 1 unit and 2 hours for 1 unit. Let us say if he gathers C units of coconut how many hours, he spends on collecting coconuts 2 C and if he catches F units of fish then how many hours, he spends catching fish F. So, 2 C plus F because this is the total time, he is spending in gathering C units of coconut and catching F units of fish. So, in this bundle we have C comma F, C denoting the number of the first number denotes the number of coconuts and the second number denotes then a unit of fish and this has to be less than or equal to 8. And this is his budget constraint. Although it is not in terms of money it is in terms of time but you can very well imagine a scenario where a person is interested in only two things coconut and fish or it can be cloth and fish does not matter any 2 items. And then 2 rupees is the cost of 1 unit of C and 1 rupee is the cost of 1 unit of F and he may have 8 or 4 I if you want to want to change 8 to some abstract number you can say it is time T, but this time here because 1 is in hours. So, it is in hours. But if it is in rupees then T will be in rupees fine and this is the constraint. Can we draw it? Let us say we put C on x-axis and F on y-axis. How would it look like? Downward sloping straight line just. So, so that it becomes clear to you let us forget this less than sign and just let us focus on 2 C plus F is equal to T. So, if F is one unit how many let us put here for this example it is equal to 8, how many F you can get? three point five You can get 6, oh sorry I you can get 3.5 and if you have 2 units of F how many units you can get. 3. 3 and similarly 2.5 and 2 and so on, and if you plot you will get a curve that is how does it looks like. Downward sloping straight line Downward sloping straight line 8 and here you will have 4. How about a point here? Let us say this point 0 comma 0, is it affordable, it is affordable if we again bring the less than sign. So, if we are talking about 2 C plus F is equal to 8 then we get the straight line and when we talk about 2 C plus F less than or equal to 8 then we get this straight line and also this triangular area and any bundle here is inside this triangle or on this triangle is affordable. And how about a bundle outside this triangle? Let us take a bundle here. This green point what does it indicate which bundle does it indicate? It indicates a bundle 5 comma 0, 5 units of coconut and 0 units of fish. To gather 5 units of coconut how many hours do you need? 10. 10 and that is not feasible given here you can say very well how about we include the next day, but that would change the problem we are taking, we are here the basic assumption although I did not mention it explicitly. The basic assumption is that you cannot store these products and we are talking about 1-day budget, you gather it, and you spend it on the same day fine. Now, what can we do we can rewrite this equation: 2 C plus F is equal to 8, it means we are taking the straight line only. This straight line is called the budget line and the whole triangle is called the budget set. How the budget line is different from the budget set? The budget line is only the maximum we can take we can. Maximum in which sense. In terms of. Let us say for this problem the budget set indicates the maximum amount of fish you can get given the amount of coconut you have already gathered or you can reverse it also the maximum amount of coconut you can gather given that you want to catch this much amount of fish, while any point inside the triangle it gives you all the possible bundles. So, let us look at it just on the line what do we get we can write it like this also let me write it in terms of F what do we get 8 minus 2 C and what is this 8? This is a math problem, not an economics problem. What is this 8? 8 is a vertical intercept 8 is the vertical intercept and also it gives you the maximum amount of fish you can catch, of course, when we want to catch the maximum amount of fish you have to decrease the amount of coconut to 0 fine. We should also think here if you look at pay attention to this table if you increase your fish catching by 1 unit what happens to the coconut? It decreases by 0.5. It decreases by half a unit. So, here one unit implies 0.5 units at any level you see here from 3 to 4 you go in fish gathering fish catching then coconut goes down from 2.5 units to 2 units, so 1 0.5. So, can I say the cost of 1 unit of fish is half a unit of coconut or price of one unit of coconut is 2 units of fish, can I say that? This is the relative price in way, relative prices of coconut in terms of fish or fish relative price in terms of coconut. This also indicates the opportunity cost of catching fish in terms of coconut what is opportunity cost by the way. Just a revision we have talked about opportunity cost during the first chapter, but what do we mean by the opportunity cost. By substituting one good we can the amount of other good that we can get. But here it is simple you are talking in this context. So, what basically we mean by opportunity cost is in general that it is the cost of it is the value of best alternative for gone. Like say you devote one more hour to catching fish, you devote 1 more hour to catching fish it means you are devoting 1 hour less to gathering coconut and in this economy you have only two activities catching fish or gathering coconut. So, 1 more hour devoted to catching fish means 1 less hour to gathering coconut. So, what is the best alternative of catching fish here? Gathering coconut. Gathering coconut. So, when you catch fish, this is the alternative forgone. In a more complex world when you can do more than to act you can have option of choosing among choosing from a list of a large number of activities then your opportunity cost is the value of best alternative forgone. For example, here this person can do also apart from catch catching fish gathering coconut, he can also gather let us say some other fruit apple that is on the other side of the island although we are not considering for this example what will be the opportunity cost of catching 1 unit of fish. Apple or coconut. Which one, how would we figure out it depends, it will depend on that particular person what he likes most after fish. Since he has decided to catch fish, of course, he likes catching fish most, he prefers catching fish most. So, you will have to figure out what is the best alternative among these two. So, this is also giving you the opportunity cost fine. And also, we are talking here in terms of production just to give you some again that concept work that we learned in the first chapter. This budget line is also a production possibility frontier PPF, production possibility frontier how it is production possibility frontier what does it give. What is the production possibility frontier? Combination of output that can be produced in an economy given the limited supply of input and that is how you get the production possibility set. And what is the frontier? The frontier gives you the most efficient combination of outputs that you can produce. So, here, for example, the points inside they are in production possibility set, but on the straight line it is production possibility frontier. I also just a moment I want to take to emphasize that budget set is not always a straight line it is not always a straight line. For an example think let us come out of this island economy let us talk about real world where someone is concerned about only fish and coconut another think is that if you buy more than 5 units of coconut in the market then the seller gives you discount that is the general practice, in that case, budget curve is not going to be a straight line what is going to happen. Still 5 units or could be a. Still 5 units you will have this particular line and beyond 5 units. If the price will decrease so. Price will decrease so. It would. So, let us focus on this what happens when price decreases or income or amount of resources increases, that is how we will focus on the factors affecting the budget line. |
Microeconomics | Lecture61_Perfect_Complements.txt | So, the example I just gave had 2 goods, they are perfect substitute for each other and what does it mean when I say they are perfect substitute for each other. It simply means that 2 goods are that, that 2 goods are perfect substitute for a consumer. Of course, we should not forget that we are talking about a particular consumer ok, if which is a consumer the example, I gave you tea and cola it might not be perfect substitute for the other consumer. So, of course, we have to keep in mind that we are talking in context of a particular consumer. So, 2 goods are perfect substitute for a consumer if he is willing to exchange one good for the other good in the fixed ratio and also what is important that no matter what how much of unit of good 1 and good 2 he has. So, at all level he is willing to exchange good 1 to good 2 in the same proportion, then only these 2 goods would be perfect substitute. Now, we are going to study another case where goods are perfect complement of each other ok. So, take example of let us say that good 1 and good 2, let us take some examples; on good 1 we have left shoe and good 2 we have right shoe ok. You cannot use left shoe without right shoe of course, I am talking about a person who has who has both his legs intact ok. Ah similarly let us take example of bread and butter you can eat bread and butter on its own, but let us say for the example sake that whenever you have 2 slices of bread you will take 1 unit of butter. Similarly, you can say that tea 1 unit of tea requires 1 unit of milk, you consume them in the fixed proportion and if 1 good is absent the second good is not valuable at all for the consumer. So, in this case, let us say, let us take an example where you have of course here when I talk about shoe these are all the same type of shoe. So, we have 10 left shoes and 5 right shoes how many pair of shoes you can use? 5 only pairs, 5 left shoes are useless. Similarly, let us say you have just for example, 12 slices of bread and you have only 5 units of butter, how many slices of bread you can use only 10, 10 slices and plus 5 units of butter that you will use and 2 would be 2 slices would be wasted. Similarly, rather than naming them you can think of such examples let us say you need good 1 a units of good 1 and b units of good 2 to have 1 perfect combination. Now, if let us say you are living in a world. Of course, the world for the example’s sake it is made of only 2 goods and let’s say in that good, that world both the goods are perfect complement, this these are the examples of perfect complement, what do I mean by perfect complement that 1 good is valueless if the other good is not present in the fixed predetermined proportion. Not just in any fixed proportion because this ratio is important how would you draw the indifference curve and how would you describe the utility function. So, let us try to describe, let us try to describe the indifference curve first starting with 10 slices of bread and 5 units of butter and let us say here we have bread and here we have butter and here we have 10 and this is let us say for example, this is 5 units of butter. Let us say your level of satisfaction is this much, given by red point it gives the red point denotes 10 comma 5 bundle fine. Now, let us say that the amount of bread remains the same, but amount of butter goes up now from it is 5 to 6. So, you are moving from here to here, does this person experience any change in the level of satisfaction or in other words change in level of utility no same. It goes up 6, 7, 8 level of utility would remain same as this particular bundle 10 comma 5 bundle, similarly let us say if we have, if we keep the amount of butter fixed and we increase the amount of bread from 10 to 11 or 11 to 12, what would happen amount of bread butter is fixed at 5. So, again the utility will not increase level of satisfaction will not increase. So, all these red points they are at the same utility level, it means a curve passing through all these points would represent the same indifference, would be on the same indifference curve. So, let us see what happens if we decrease from 10 comma 5, if we decrease the amount of butter to 4 what will happen. Now, if we move amount of its now we are here, what will happen the utility level will decrease. So, this point is no it is not on the same indifference curve. So, let us erase this is not on the same indifference curve and similarly here if we keep the level of butter same and if we decrease the amount of bread again it will not be on the same indifference curve, utility will decrease. So, this part also we should erase. So, what we will get is basically this will be the indifference curve just quick deviation, little, small deviation. Do you think here this is the preference of this person is convex or it isn’t convex? Similarly, let me let me draw a few more indifference curve and they will be like when we move from 10 to let us say here 12 and then here 6 and then again, we get another indifference curve and so on. We can get things like that we will get is it convex is it is strictly convex. Let us look at it let us take this one if we pick any 2 bundle, let us say the bundles are this is one bundle this is another bundle and we draw a line, at all these points utility level is higher ok, but how about when we pick this point and these this point and we draw a line utility level is same. So, it means it satisfies convexity, but not strict convexity, is it clear? Remember earlier I talked about it that the all the axioms rationality axioms, continuity axioms, monotonicity and strict convexity all these are satisfied then we can use our optimisation technique, but here one strict convexity is not satisfied. Now, we have drawn we have drawn our indifference curves, how can we give the utility what will be the utility equal to let us say Br denotes bread and Bu denotes butter , minimum of Br comma let us check, Br comma to b a means let us take the example here is the point, here is the point 12 comma 5 what do we get minimum of 12 comma 2 multiplied by 5. So, that is equal to 10, is this fine. So, this one is also what we what we mean to say here is that to increase utility by 1 unit although I am using here in the cardinal sense, but one unit is meaningless I am just talking about the ranking. So, increase utility by 1 unit what do we need we need increase in 1 unit butter, 2 unit bread and 1 unit butter. That is what we need or in other word one unit of bread will increase half unit of utility if we have one unit of butter, I am saying that if you have one more unit of butter, if bread is going up by one unit then and if we have one more unit of butter then in ordinal sense let us say that utility will increase by half. So, in other word we can say utility is Br by 2 multiplied by Bu, but here these 2 are different are they, but if you notice. The second one is the monotonic transformation of the first one or the first one is the monotonic transformation of the second one, you multiply the whole thing by 2 what we will get minimum of Br multiplied by 2 bu. So, they represent the same utility level or we can say that if you require, let us say to increase in utility by 1 unit you require a units of good 1 and b units of good 2. So, how many one unit of good will increase the 1 by a. So, utility is going to be represented by minimum of x 1 by a and x 2 by b and if you multiply by whole thing you can also say this is equivalent to minimum of bx 1, ax 2 fine this is the utility representation. Now, how about the optimal consumption bundle, about optimal consumption bundle always be on the corner point that we have ok, fine. So, if it is always on the corner point what does it mean, how can we solve the problem. Now, the problem is let us say maximise minimum of let us say this is the problem bx1 comma ax2 with respect to x1 comma x2 such that p1 x1 plus p2 x2 is equal to I. Again, we do not have to worry about less than or equal to sign because this does is satisfy monotonicity, it does not satisfy monotonicity, it satisfy the another version of monotonicity the version that we have talked about in the class is not satisfied. What we have said in the class that same amount of all the other goods and at least the person has some more of one good then his utility will be higher than the previous situation scenario that is not true here. So, that version of monotonicity is not satisfied. Student: Only on the line. Yes. Student: Joining corner. Only on the joining corner this is satisfied, but fortunately here the optimal bundle lies on the on the corner itself, what, how can we solve from here what we need to maximise utility what we need is bx 1 is equal to ax 2, if this is not equal then you are wasting at least 1 of the you are wasting one of these 2 goods because both goods are expensive ok. So, now you have 2 equations, 2 linear equation and it is very easy to solve. So, let us solve it what will you get x 1. Is equal to ax 2 by b so p 1 ax 2 by b is equal to p to x 2 i. So, basically what we have a p 1 plus bp 2 x 2 by b is equal to i. So, x 2 is. So, x 2 is b I a p 1 plus bp 2 and x 1 is if you solve it you will get a i, a p 1 plus b p 2. So, x 1 and x 2 are in a is to b ratio that is how we started a that the description, it is clear and solution will be on kink. |
Microeconomics | Lecture79_More_on_APL_and_MPL.txt | Ok let us take this graph again on y axis we have and on x axis we have labour input and of course, we are taking capital as fixed. So, we are not putting capital on any axis; so, what we have here is that output is going up as labour is increasing till this point and then its starts decreasing. That is why when we say when we talk about production function we draw here this function in dotted line; why? Because production function gets the maximum output possible given the amount of input. Input. And we are allowed to freely dispose of as one of the assumptions and one of the properties that we have talked about earlier. That if let us say we are here of course, if we are here and if we take this function it gives the production level as this much while we can produce higher products, higher output by freely disposing of some of the labour input. So, that is why if we want to talk about the production function it should be like this ok, but that is not the point. Now here we are talking about if we take this point just arbitrarily I am taking a point just to remind you if we draw line from origin to this point on the output curve; the slope of this line is slope of this line is equal to ? Average product of. Average product of labour; why? Because this is Q at this point and this is L an average product of labour is. Q by L. Q by L, but if we take the slope of this output curve at this point then what is this the slope is equal to? Marginal. Partial derivative of Q with respect to L and this is nothing, but marginal product of labour. So, on this curve at any point if we have this curve; we can figure out APL as well as MPL. Although here the process is reversed, we already knew APL and MPL an using that we had drawn this curve, but what I am saying when you have an output curve with respect to one of the inputs; we can always figure out average product with respect to that input and that input here is labour. And of course, also we can figure out marginal product of labour fine. And if you notice at, till this point that average product of labour is increasing ok; it is increasing and after this point it starts decreasing fine ok. Now, what we can do we take another graph where L is still on x axis, but on y axis we are not plotting q, but we plot APL and MPL two different curves. So, what is happening? Till this point if we just take of course, it is rough, it is not very exact what is happening till this point average product of labour is increasing and after this point it starts decreasing. And if we see the value of MPL at this point if you look at the value that is labour input is equal to 1; APL and MPL both are equal. But at this point rate of increase of MPL is higher than rate of increase of APL. Or in other word MPL is increasing at faster rate than APL; so, it goes up and then it starts coming down again its not to the scale it is very rough what is remarkable is here at the point where APL is highest the maximum at that point MPL curves; MPL curve cuts the. APL curve APL curve from above. That is also very important from above ok. So, is it because the way I have given the I have put the value; the way I have given the table or its true all the time that wherever APL is maximized MPL is equal to APL. It is true all the time It is true all the time why? if you just look at it let us say average product of capital is 2 and marginal product of capital at that point is 3; what does it mean? That if you add one more unit of labour increase in output because of this new labour would be equal to. 3. 3 while average is 2. 2. Of course 3 is greater than 2. So, in numerator you are adding 1 and as a result in denominator you are adding 1, but a numerator as a result 3 is being added and that is of course, more than? 2. 2. So, average per person would go up ok. So, this what you can say 3 minus 2; 1 will be distributed over all the labour. So, for every one average, for every average contribution would go up from 2 to 2 plus. 1 by L. 1 by L. Sir can also say this by the by the way that when average is maximum; it means we are adding something do it be makes it maximum and after that it will decrease. So, at that point of time MPL that is the del change in labour, should it be less than average, should be less than the average else it else it will increase again. No, but it I am talking about in general at all the point. I am saying for all the points whenever average is maximum. Not whenever average is maximum, So, when you are saying average is maximum average is maximum at only one point let us look at it here average is maximum only at this point. Yes sir. Not everywhere. So, sir. But we are talking about all the positions where MPL that is marginal product of labour is greater than average product of labour. Sir what I am saying is that MPL should cut the APL graph at that point I am saying the reason for that. It is because the average is maximum; that means, we a we a we are adding something in that average which is maximum which has ablated maximum and the rest of the things are lesser than that term that is why the average decreasing after that. For maximum what you are saying is true. But in general it is true for all the point that whenever MPL is greater than APL on average it is adding average contribution of all the labours are going up. That is why average product of labour will increase. Yes sir. And to. Every new labour is better than the previous labour. Not better than that incremental Contribution. Incremental Contribution Is more than the average contribution of all the previous workers. And that is why the new workers contribution would add on the average contribution of all the other workers and that is why average product of labour will increase ok. And similarly, here marginal product of labour is whenever you have marginal product of labour less than; whenever you have marginal product of labour less than average product of labour; what will happen? Again, the incremental contribution is less than the average, that is why it would decrease the average of all the workers. Workers. And thus, average product of labour will starts decreasing; so, where when maximum what does it mean? Maximum is a kind of a stationary point; at maximum you know when at which point the average product of labour is maximized where this is one of the requirement that average product of labour should be stationary. So, when it can be stationary? When it will not change where it is equal to marginal product of labour that is why at maximum marginal product of labour inters curve intersect the average product of labor curve. Now another point it why should cut the average product curve from above not from below. What I mean see there are two intersection points one is here and one is here to green. So, at this point also average product of labour is equal to marginal product of labour. And at this point also average product of labour is equal to marginal product of labour that is why we need to a specify what is the difference between at this point and this point. Here marginal product of labour curve is cutting average product of labour curve from bellow. So, when L is increasing it is moving like this average product of curve is like this and marginal product curve is like this. So, it is cutting from below while here it is opposite. It is cutting from above. So, that is another requirement why should be cutting from above? Sir because initially like we said that diminishing marginal product of labour. If the marginal product of labour diminishing and keep on adding labour then marginal product of labour could. Let us look at it little bit mathematically, see what is happening where it is cutting. Like you said including the criteria of free disposing when we have obtain. It has nothing to do with free disposing. Sir when you obtain when you obtain like we using the computer example when we have 7 people for 7 computers and when we No I am not saying I am not saying for this specific example; what is happening in general. Let us look at where its cutting from below here it is cutting from below; what is happening it here? That beyond this point beyond this point MPL is above APL ok. So, it means that what will be its impact? We have just discussed that APL is APL will increase. (Refer Time: 11:37). APL will increase. So, if APL is increasing by increasing labour; it cannot attend maximum here. Fine while what is happening here? Here it is cutting from it. Is it possible for MPL to cut this from below? Here it is cutting from here it is. But it is APL is just sum of MPLs at end. So, if MPL is cutting from below that is mean it is some will. MPL is not some of all the APLs. Sir it is just. It is not; it is not let me just tell you clarify it to you; here is total quantity curve here is the total, when we what we are doing is here we have Q, here we have L fine. And what is APL? APL is let us say APL at this point APL at this point is this angle because here we have Q, here we have L at this level. So, at L naught APL is Q naught by L naught; in other word the slope of slope of this line and what is MPL? MPL is slope at this point. So, MPL is not integration of APL MPL is at this point at L is equal to L naught ; is it clear? Fine now it is clear why it has to cut from a buff. Yes sir. Because at that is the stationary you cannot get the stationary point if it cuts the APL from below. Fine and that is the reason. So, two requirements one that is where APL is maximized max APL; APL should be equal to MPL fine and MPL curve should cut APL curve from above. If you know calculus you would notice that this is basically first-order condition and this is second-order condition, let us look at it let us do it mathematically what is APL? Q by L. Q by L fine let us differentiate the APL; what do we mean by that? If L increases APL increases, what does it mean? It means that the derivative of APL with respect to L is? Positive. Positive that is what is means. So, let us say we have only one variable what can we right it? What can we right it this is nothing, but. Q by L square minus Q by L square plus d Q by d L into 1 by L. This is Q by L square; you can say would partial derivative sign thinking that there may be more than 1 input fine is it clear? Take 1 by L common. (Refer Time: 15:47). And as a result what would you get? 1 by L and what is this? MPL. This is MPL at that point and what is this? APL. APL just leave it here. So, now, it is very clear if MPL is greater than APL what it means? That partial derivative of APL with respect to labour is? Positive. Greater than 0; it means it means L goes up. APL goes up. APL goes up fine is it clear ok? Now, if MPL is less than APL what it means? The partial derivative of average product of labour with respect to labour is negative and it implies L goes up APL. Goes down. Goes down and third would be MPL is equal to APL and that is how we get the? Stationary point. L goes up momentarily not at all level; just at that point that L goes up APL remains the same no change. Remember if you know calculus, although we have already discuss the graphically, but using calculus just helps you get little more information. So, what is really happening? That this is one of the requirements that MPL is equal to APL then APL may have maximum or. Minimum. Minimum or one of the critical points the boundary points fine. So, we cannot say that it would be the maximum; what we have learnt? We have learnt that there are two points where they are equal this point and this point, but only here we have maximum. The second requirement that I am not going to discuss, but if you know calculus we have obtained the first-order condition you can obtain the second-order condition. And how can you obtain the second order condition? Just let me tell you how you can obtain the second order condition; differentiate average product of labour with respect to labour twice not just once. So, now, what you need to do is you need to obtain this and this will give you the second order condition. The interpretation; the graphical interpretation would be the requirement that it gives you would be just that MPL curves should cut the APL curve from. Above. Above fine it is clear; it is very very important. And also, you should not get confuse because this is the common mistake when people say what if MPL is decreasing? What would happened to the APL? Can you say that let us say that MPL is decreasing what would happened to APL? We cannot make any comment, why? Because let us look at this zone here, in this zone MPL is decreasing. But? APL is. APL is increasing. Increasing. And in this zone, MPL is decreasing and APL is also decreasing. So, we cannot say anything about it; we can comment about APL given the information about MPL only when we can compare the values of MPL and APL ok. So, be very wary about this, this is typical question that you get in exam. |
Microeconomics | Lecture14_Factors_Affecting_Supply.txt | Let us look at the factors affecting the supply curve. We have learned what do we mean by supply curve. Supply curve, just to remind you that supply curve is a curve that gives the combination of market price and at that price seller’s willingness to supply certain quantity of good. So, now let us look at the factors affecting the supply curve and the first is technology. So, here we have again, I remind you again and again, then whenever in economics whenever we are drawing the graph, quantity is always on x axis and price on y axis. Even though quantity is dependent variable and price is independent variable and roughly this is the supply curve. It need not be a straight line, it can be any upward sloping curve, but just for illustration I am drawing this straight line. So, now let us see what happens if there is an improvement in the technology, what can you get, what does it mean if there is an improvement in the technology. It means, it can mean two things either you need less amount of input to produce. Student: the same The same amount of output, that can be one meaning or it can be that there is absolutely new technology, that produces the same output more efficiently, but it does not mean what you take the meaning of technology. It if there is an improvement in technology it means the marginal cost to produce one more unit is little less and let me remind you what is marginal cost? Marginal cost at any particular level is the cost to produce one more unit. So, if seller is producing that unit, he needs to recover at least that much of amount from the market. If his marginal cost is 5 rupees and the market price is 4 rupees, he should not produce this particular unit. Because by producing it he would incur five rupees of a cost and he will get only 4 rupees. So, he will make a loss that is why he would not produce. So, technology leads to reduction in cost. So, what does it mean in terms of supply curve? Student: at the same price At the same price, the producer will be willing to produce more. So, we will have a shift in the supply curve, it will shift outward. So, let me draw it here, this is the shift. So, technology, an improvement in technology means, an increase in supply. The second factor is the prices of raw material or more broadly speaking prices of inputs. So, take example of one input which is not raw material labor. You need human resources to produce something. Now, let us say wage goes up wage that you need to pay to a worker goes up, what will be its impact on the supply curve? Let me again draw the basic supply curve what will happen? Student: Inward shift inward shift why? Because wage has increased, it means to produce one more unit, you need to pay a little bit more to the worker. So, marginal cost has increased, it means the supply curve would shift inward; it means supplier is willing to supply less amount of good at the same price this will be the movement. Now, let us look at the third factor; third factor is expectation about future. Let us take an example from petrol industry. When a petrol pump owner, although he is not producing, but he is supplying the petrol, later on you will see that in economics, we sometime we use goods and services interchangeably; he is basically not producing a good, but he is producing some sort of service. So, if he expects that the price of petrol, it will be the input price for him in a way, that it is going to go up in the near future, let us say in 2-3 days what would he do? Student: Less His willingness to supply petrol today will decrease. He would supply less why? Because he thinks that in the future he will be able to sell his petrol at higher price. So, what will be the shift in the supply curve? Student: Inward. Inward it will be inward shift. Now, fourth. Is it clear? Similarly I just talk, I gave you only one example from future, you will have to analyze depending on the scenario. If you feel that tomorrow onwards no one is going to buy petrol because of some reason probably you will be willing to sell at less lower price. So, your supply would increase fine . Now, let us look at the fourth factor, that is the prices of substitutes and complements in supply not in consumption, in supply. So, I gave you an example from airplane industry Boeing, and I discussed that civilian aircraft and military aircraft they are substitutes in production. So, let us say because of some reason the price of military aircraft goes up, what will happen to the supply of civilian aircraft by this Boeing company. Student: It will decrease. It will decrease; why because Boeing will devote its resources to produce. Student: more. More military aircraft and because they are substitute the military and civilian aircrafts are substitutes in production, it means Boeing is using the same sort of, same kind of facility, same facility to produce either military aircraft or civilian aircraft. It means the willingness to supply civilian aircraft goes down if price of military aircraft goes up. So, supply of civilian aircraft will decrease. And similarly let us say I gave you another example complements in supply milk and cow dung, if price of cow dung goes up. Student: Supply of milk. Supply of milk will? Student: Increase. Increase fine. So, let us move to the last factor, at least in my list you can of course, think of more factors by the way all these factors that we have discussed; they affect not only the market supplies, but also individual supply. Now I am going to talk about a factor that affects only market supply not the individual supply and that is number of suppliers. If number goes up what will happen to the market supply? Student: Goes up. Market supply will go up. Is it clear? |
Microeconomics | Lecture90_Diminishing_Marginal_Product_Vs_DMRTS.txt | So, we studied about diminishing marginal rate of technical substitution and we also studied diminishing marginal product. So, let us see do not get confused between these 2 terms, because it is very easy to get confused. So, let us look at graph, and try to understand how these 2 things are different. So, first, let me say let us take these are isoquant fine and what we have let us, let us say on y-axis we have capital and x-axis we have labour let us say. So, these are three isoquants Q 1, Q 2, Q 3 fine here we have L 1 here let us say its L 1 plus 1, here it is we have L 1 plus 2, can we say anything about Q 1, Q 2 and Q 3? If let us say when we talked about diminishing marginal product, what did we say? That eventually may not be in the beginning, but eventually, the marginal product of an input starts diminishing that is what we talked earlier. Let us say that diminishing marginal productivity has already set in at this point ok. So, this is not the case that the marginal productivity is increasing when we increase labour beyond L 1. So, what I am saying? We are moving from L 1 to L 1 plus 1, L 1 plus 1 to L 1 plus 2 and capital is fixed at K naught and these are the three isoquant Q 1, Q 2, Q 3. Can we say anything about Q 1, Q 2, Q any functional relation we can talk here? That we know that this is given the way we have talked about isoquant, we have talked about monotonicity. So, that is given that Q 3 is greater than Q 2, is greater than Q 1 that we have already talked about. But if we bring the concept of diminishing marginal productivity can we say anything more, that is the first of all diminishing marginal productivity has nothing to do with the slope. Sir; Q 3 minus. That is what we are talking about now, but you what you have said is not correct, Q 2 minus Q 1 is greater than Q 3 minus Q 2 because it is diminishing what we have done here. We have fixed the K naught and what we are doing? We are adding labour by one unit. So, of course, as labour is being added Q 2, Q 2 will increase output will increase ok, but by how much? We can quantify the increase by Q 2 minus Q 1 now again at L 1 plus 1, we are adding one more unit of labour what we are getting is now Q 3, but this addition is because capital is fixed, I have already fixed capital at K naught. So, only we are increasing the amount of labour. So, what does the marginal product diminishing marginal product says? That. That each addition is; Less than the; Less than the previous edition is subsequent addition is less than the previous addition that is what you know the diminishing marginal product means. So, the subsequent addition is Q 3 minus Q 2. So, what we get? Q 2 minus Q 1 is greater than Q 3 minus Q 2 it has nothing to do about nothing to do with the slope is it clear fine ok. Now let us look at it, again, what do we mean by diminishing marginal rate of substitution? DMRTS Diminishing Marginal Rate of Substitution what we are saying? Here we take we are taking one isoquant and we are not fixing the capital or labour what we are saying? Of course, when we increase the amount of labour let us say from L to L 1 to L 1 plus 1, we need to decrease the amount of capital to produce the same amount of output. So, what is important here, that we remain on the same isoquant that is let us say Q naught. So, what we are talking about here is that the marginal rate of technical substitution keeps on decreasing as we keep on increasing L. So, here the slope is this is the slope. Now, the slope as l increases slope decreases and what does the slope represent? Slope represents the marginal rate of technical substitution. So, as we increase labour the slope keeps on decreasing this is this slope is steeper than this, this slope is steeper than this. So, MRTS is diminishing simply means that this slope is decreasing, the shape of isoquant is like this ok. So, do not get confused when we talk about diminishing marginal product, we jump from one isoquant to another isoquant, keeping one of the; Input fixed. All other inputs fixed in our case we have only 2 inputs. So, we keep the capital fixed and we are talking about changing the label. So, marginal product of labour we are talking about ok. So, we move from one isoquant to another isoquant but here when we talk about diminishing marginal rate of technical substitution, we are at the same isoquant; we do not keep inputs fixed what we keep fixed here we keep the output fixed. Output. That output is equal to Q naught and as we are as of course, we are increasing labour, it is decreasing. Another way to look at it is because here we are defining MRTS a s d K by d L. One can change the terminology and one can define it as d L by d K in that case we will keep on increasing k. K. And d L by d K will keep on decreasing. Decreasing. Because we say marginal rate of technical substitution diminishes fine is it clear? So, do not get confused between these 2, and it is very very easy to get confused between these two, fine. So, theoretically clear graphically clear. |
Microeconomics | Lecture94_Cost_Minimization_Few_Examples.txt | So, now let us solve some mathematical we have looked at it graphically, some problem we will solve graphically and then we will derive the mathematical way to solve this cost minimization problem. What we have done? Let us say where production technology is, let us say production technology is given by Y or Q that is what we have been using Q is equal to K plus L ok, price of capital is r; rent for capital is r and wage for worker is w. The unit should be corresponding to each other. If you are using capital here per day, then here r should be in per day term. If L is here in the per hour term, we cannot have these two in the different terms first of all, if it is in per day this should also be in the per day term then only we can add fine. So, how can we solve? We have solved similar problem while when we discussed consumer theory. Straight line Isoquant is straight line. Isoquant is a straight line. So, this must come on the end, either on the x-axis or on the y-axis. Ok, that is one way to think about it and right, you are right you are not wrong. But I will say the more intuitive way to describe and what is more intuitive? What you are interested in? You are interested in Q amount of output and how you can produce. Using capital and labour and what it does not matter, how many units of capital you are using and how many units of labour you are using. What matters is that the sum of units of capital plus unit of labour is equal to the amount of output that you need ok. So, now it is very clear that if rent is if capital is expensive than labour, then it is good idea to use only labour and if labour is more expensive than capital then it is you use only. Capital. Capital. So, you will use only one of these two inputs in the rare case where capital and labour both have equal prices. Then only you can use the combination of two, otherwise you will use the cheaper of these two. So, let me make this little more complicated. How about when Q is equal to a k plus bL, where a is a positive real number and b is also a positive real number, rent is still for capital r and wage for labour is still w. What will you do? Is it clear? Are you guys following me? Sir, if first we would check on, we have to check on both simultaneously on prices as well as output. The simple thing that you should. W r by a is greater than w by b. Then we use all labours. So, right logic let us see here, what is happening? You are in the right direction to produce 1 unit of quantity, let us say if you are using only capital. If you are using only capital to produce 1 unit of quantity, 1 unit of output how many units of capital do you need? 1 by a. Q by a Q by a. I am saying Q is equal to a, so you need 1 by a unit of capital and to produce this output using only labour, how many units of labour do you need? 1 by b. How much is the cost of this capital? R by a. R by a. How much is the cost of this labour? R w by b. W by b. Now it is very clear from here. If r by a is greater than w by b; it means it is expensive to produce this output using only. Capital. Capital. So, you will use only labour. So in this case what will be your cost? Your cost is to produce units of output is going to be. W by. W by b by multiplied be Q or Q naught if you whatever you want to see. Why w by b? Because 1 unit of labour would produce b units of output. So, total if you are using L you will produce w by b and that is for 1 multiplied by Q. So, what you are going to get? Q b by b and if r a is less than w by b, then you are going to use only capital and to produce Q naught amount of output how many units of capital do you need? Q by a and what would be it is price? R. Multiplied by r. So, in we can write it like this r a multiplied by Q. So which one will you use? Minimum of these two, so, cost to produce Q amount of output is going to be minimum of r by a comma w by b whole multiplied by Q and this is called cost function; this is cost function. What is cost function? When cost function is not any cost to produce Q amount of output, cost function gives the minimized, the minimum cost that you would incur to produce Q amount of output. So, cost function is always going to be a function of how many units you want to produce, let me write it here. How many units you want to produce? Rate of rental rate of capital and wage rate for workers. So, it is always going to be a function of Q r and w. The earlier problem that I gave you where we a and b were equal to 1, what would be the cost function? It will be simply minimum of r comma w multiplied by Q. Let me change the problem little bit. Now let us say both inputs are used to produce some output and these two inputs are perfectly complement each other. In other way what it means is that Q; to produce Q amount of output what you need is; a amount of capital; a units of capital and b units of labour. Can you give me the production function first. (Refer Time: 07:38) Minimum of. Think about it, think about it for a moment. Q by comma a by b. Yes. Minimum of. So, minimum of minimum of. K. K by a comma L by b. L by b. This is your production function, fine. It is very similar to what we have done in the consumer theory, and we had derived it there in detail. So, let us not waste time about it. So now tell me the cost function. Minimum of k by. First, let me solve make this problem easier for you where a and b are equal to 1. Into minimum of k comma L plus b r into. Think about it again, we are not using any mathematics, any calculus. A simple that, you have to think about, you are little bit in the right direction. R and. What we are doing here? Let me write what is our purpose here. What we are doing? Cost minimization, let me write the complete problem cost minimization. In other words what we have is minimize rk plus wL with respect to k and L, such that minimum of k comma L is equal to Q, that is what we are doing. W plus r into Q. W plus r into Q that is fine; that is you do not have any minimum or something there. Same thing minimum of. You may be right; I did not pay that much attention but say that is what will happen. Now, to produce let us say to produce 1 unit of output what do you need? You need at least 1 unit of capital and 1 unit of labour. If you have more than 1 unit of capital and just 1 unit of labour, what you will have? Some unit of capital you would be wasting. So, we are talking about minimizing the cost, so we should not be wasting anything. So, what is the most cost-minimizing way to produce it to have exactly 1 unit of capital and 1 unit of labour and how much is the cost of 1 unit of capital and 1 unit of labour? R and. R plus. R plus w. W and of course, if you are producing Q then what do you need rather than 1 unit of capital, and 1 unit of labour, what you need is Q amount of capital and Q amount of. Labour. Labour. So, cost is going to be Q multiplied by r plus Q multiplied by w or it is going to be r plus w multiplied by Q. So, C of let us say write it here r w comma Q is r plus w multiplied by Q, you can solve the more difficult problem on your own. |
Microeconomics | Lecture71_An_Application_Subsidy_Vs_Direct_Benefit_Transfer.txt | Now let us that is the last topic in the consumption theory. Let us do an application a simple application. See in the budget our finance minister announced that slowly at least for some particular goods government would move from specific subsidy to keys cash transfer. Ok and Bihar government is already doing it for several product. So, let us see what are the differences in these 2 scenarios. So, let us say that; we have good 1 and good 2 again what we can say we are talking about good 1 on which there is subsidy and a good 2 is the composite good, it represents all other good like for example, let us say good one is LPG gas cylinder ok. Now, let us say that this is the scenario, this is the indifference curve that this person is able to achieve of course, I should draw more indifference curve. Now what is happening that government if government give subsidy on gas cylinder what will happen to the budget line? Budget line shift or budget line will. Rotate. Rotate in which direction it will rotate? Anti clockwise. Anti-clockwise pivoted at x 2. So, it will be something like this fine of course, in this case the utility level of consumer that utility level that consumer can achieve will go up what is happening let us say here maximize x 1 u of x 1 comma x 2. And what we have ? P 1 x 1 plus p 2 x 2 should be less than or let me put just equal to I. Now, what is happening if government gives subsidy on good 1 and let us say it is the specific subsidy then, what will happen to the budget line? Equation tell me the equation. P 1 what did we do in case of taxes subsidy here is just opposite of taxes and it is a specific subsidy means; you why once cylinder gas instead of paying p 1 you pay p 1 minus t because t amount is given by the. Government. Government. So, it is p 1 minus t x 1 plus p 2 x 2 it is clear. Now, let us see. That optimal level we figure out is x 1 star, x 2 star and this is the optimal level. We are of course, interested in x 1 how much will be the subsidy government subsidy. T. T x 1 t x 1 star and that is fixed ha that is given, that is that does not vary with x 1 because x 1 star is a fixed number. Now there is another way of giving this subsidy, the another way could be that rather than giving it on gas cylinder government can do cash transfer equivalent to this much amount. So, rather than telling you that when you go and buy gas cylinder rather than paying the full price you pay something less. Now government figures out that this is the optimal level of consumption in this case and government says rather than giving you subsidy on gas cylinder I give you subsidy on income. In a sense that I gave I make some cash transfer to you. So, in that case let us see what would be the problem? maximize u x 1 comma x 2 such that; p 1 x 1 p 2 x 2 plus. I plus t x. I plus t x star 1 and this is fixed. t x how can we get t x star 1? What did I say? . X star x 1 star is not unknown, let me give you in this example because government is moving from a specific subsidy to cash transfer on some matter government knows x 1 star, how many cylinders people are buying in generally ok. So, this is known. So, what government can do rather than giving this subsidy on cylinder, we will just make some cash transfer to you. And how this cash transfer will change the problem, now let us see let me use the red color. Now, what will happen? Now instead of writing tx 1 star to wide confusion let me write it here as R. This is the or S would be probably the better word. This is the subsidy Now, what will happen to the budget line? Parallel shift. It will be a parallel shift that we know that it will be parallel shift can I say anything more other than parallel shift. It will pass through this x 1-star x 2 comma x 2 star. Let us see whether this budget this new budget line satisfies whether it will pass through x 1-star comma x 2 star or not, what we can do ? P 1 x 1 star plus p 2 x 2 star is equal to I, and what is S ? T x 1 star this is p 1 minus t x 1 star plus p 2 x 2 star is equal to I. That means this line this line is passing through x 1 star comma x 2 star. So now we can draw a line draw this line, what we do? That there is a parallel shift and it will pass through x 1 star comma x 2 star. So, what we will have if and we also know that our new budget line is cheaper than this line. Thin line. And it passes through this point. So, what is happening it will be something like this. Of course, these 2 lines will be parallel ok. Now tell me that on this budget line the new budget line the utility the maximum level of utility that this consumer can achieve would be higher or lower. Would be higher. Would be higher why? Because this line is not tangent to the the highest possible utility level that can be achieved in case of a specific subsidy it is crossing it. it means; clearly what we have learned earlier during optimization? That if it crosses. Then. The budget line then utility can be. Higher. Increased, how it can be increased my moving in this direction. So, what will happen that? Buy less of x 1. Let me use green color. This will be the new utility level let us start with u one u not u 1 and u 2, u not is the utility level achieved without any subsidy u 1 is the utility level achieved with a specific subsidy and u 2 level is u 2 is the level of utility that can that this person can achieve in case of cash transfer. Now, see what is happening that when a person is given cash transfer, he can decide optimally whether he should be using this transferred money to buy some gas cylinder or something else. That is up to him. Now here government is forcing in a way that because you get subsidy only on. Gas. Gas cylinder. So, it is not optimal for the consumer. So, what we see here that of course, cash transfer will make the consumer. Better. Better off or government is the same scenario, why it is the same scenario ? In the earlier case also, the expenditure is tx 1 star and in the new case is also expenditure is tx 1 star subsidy remains the same. Same. In both the cases, but in this case what happens that the utility level that a consumer can achieve goes up and is not it this is the aim behind the subsidy. . It is like you know we talked about poor people cylinder is very expensive that is why it should be subsidized, but what we are saying equivalent amount we give you as rupee in as cash transfer and you can do whatever you want to. sir just a question sir what would happen if in the second kind of subsidize they give like the cylinder would be of 700 rupees and for each cylinder bought they would transfer 400 rupees. Now we do not know x 1 star. So, how would it matter, it would just convert with the in the case like. The earlier subsidy know. It will converse like the earlier case fine ok. So, again let us see it is not perfect let me tell you it is not perfect there are some problems with this ok. What are the problems? Again, like it is it is very easy to implement that when you go to buy gas you pay the fixed price. Now, transfer would be very, very you know due to a ascertain the transfer would be difficult. And also let us say here people who use more gas cylinder typically they would prefer specific subsidy. So that their gas cylinder becomes cheaper fine that is one problem the second problem is that we are talking it is this kind of analysis is called partial equilibrium analysis, because we are studying only one side of it, tax will also change the supply pattern of suppliers that we are not taking into account fine. So, it how suppliers would be effected supplier the supplying on the same price ? Supplier how supply will be affected because see what happens these days that this subsidy is given to the in company. Ha. Subsidy is given to the company. Now, subsidy will be given to the consumer. Consumer. So, that will of course, depending on the shape of supply curve it would affect the supply. But some marginal cost is same. Marginal cost will change marginal cost will change, because of various other factors transportation and other factors that are that those are involved there. Ok. It would change. So, so can this increase the price of cylinder also? It can increase the price of cylinder to some extent. ok. So, again we can know this is very you know one side of the story economists are famous to talk about the both side of the story and that is where the problem is, but I just wanted to say one thing here that here the cash transfer, cash transfer in cash the consumer is making the choice about his consumption, that he gets this much of extra money what he should do about, what he should by using this money, but in the other case the government is already making the choice for the consumer, it is in a way paternalistic. It is like saying you know that if you study medical you will be given some extra money. So, government is making this choice at least imposing government is imposing his choice on you. I am not saying it is anything against it, but I am saying how it is working. So, in this case I say in a way that there is more personal freedom individual freedom. But again, how economics these 2 are very similar kind of scenario. On onset they look very similar, but they lead to different kind of result. So, the idea is that the theory that you are learning you should be applying to different day to day problem, to think about them in more rational and in more meaningful manner fine. |
Microeconomics | Lecture84_Elasticity_of_Substitution.txt | Now, what we have is the elasticity of substitution? What we have done; let us look at MRTS. What is a MRTS? MRTS is basically slope of isoquant and this is MPL divided by MPK. Although, we have not done, but look at this isoquant; when you want to produce we will do this in more detail later on that the concept that I am talking about, what we are talking about is here; that to produce Q naught amount of output all these combination of inputs are fine. All these combinations of inputs are efficient; in the sense that you would not be wasting any input to produce Q naught amount of output, but it does not mean that we can pick any one of these; these are technically feasible we are talking about technical constraint, but how about the economic motive; you want to maximize your profit or you want to minimize your cost, to produce the same amount of output. We have not talked about cost minimization or profit maximization that we will do shortly, but this is very simple that you should be able to understand. So, a point you will pick a point such that the cost of producing Q naught amount of output is minimized; so how will you pick; at that point, this MRTS will play a very important role. If you remember the concepts from consumer theory, what did we do? There, the counterpart of MRTS; that is marginal rate of substitution should be equal to the ratio of market prices, here it is the same thing. So, what we are talking about is that it has something to do with the MRTS should be equal to the price ratio at the optimum level. Of level Fine; although, I have not discussed it in detail, but you do not need to know this. So, MRTS can proxy for the price ratio at optimal level at all the optimal level fine and what is elasticity? I am going to talk about it in little differently also without using the concepts from profit maximization, what is elasticity? How much of the value deflects corresponding to something else. When we say go back to, if what we have learned price elasticity of demand. What did we talk about? The percentage change in the quantity due to 1 percent change in price. Price that is what we have; or it is rate or proportional change in demand with respect to proportional change in. Price. Price that is; what we have talked about? So, what we are talking about is elasticity of substitution. We are not talking about; we are not saying here price elasticity of substitution, but what we are talking about is; some sort of elasticity of substitution. So, how the substitution, the proportional substitution changes as the proportional price; proportional price change in the changes in the market, and we do not have proportional price, what do we have; to proxy for proportional price, we have MRTS. So, what we can say in the proportional change in substitution with respect to proportional change in MRTS and that is elasticity of substitution. So let me write it here in this particular context it is; this is proportional change in the inputs that you are using, this is instead of using this partial derivative sign, what you can do; you can write delta proportional change in the ratio of KL with respect to proportional change in MRTS. And instead of taking MRTS, I am taking the absolute value of MRTS, because MRTS is negative. So, does not matter I am taking the proportional value of a MRTS. So, in a sense it is very similar to the; what we had learned earlier; what we can; if we can rewrite it, what will we get; you can rewrite it like this K by L divided by fine; that is one way to write it. Another way to write it is take log, and you put here and of course, we are missing the minus sign typically ok, what we have here is basically delta ln K by L divided by delta ln MRTS what is ln? Natural log. Natural log ok. So, this can be written in the form of natural log and that is what we get. And the denominator here in this part can be written as this fine ok. Let me let me not use the concept of profit maximization, because we have not used it yet ok. Without using the concept of profit maximization, I will again try to explain; what is the elasticity of substitution; and let us look at it here graphically. We have two graphs fine and let us take one, where we have Cobb Douglas function this is isoquant in the case of Cobb Douglas function. And let us take here isoquant, in case of perfect complement K inputs. Now, let us say this is the MRTS at this particular point, which point will you choose again we are again using concepts from profit maximization, but not explicitly used to you would not choose a point here, why; because anyway you are wasting this much of labor and labor is costly. So, you do not want to use that point. So, you will always produce here at the corner, because you are not wasting any of the inputs. So, how can we define a MRTS over there it will be any line. Good point I should not say it is MRTS. What I should; let me change it little bit let me change it a little bit, what we have instead of steep corner sharp corner what we have is a. Curve. Curve good point my mistake. Now, we do not have a sharp corner ok. It is differentiable everywhere fine. Again, roughly whenever you figure out where you want to produce you will probably prove you will be producing probably here in this zone fine. Now let us say this is the MRTS here fine, if MRTS changes this is MRTS old and this is MRTS new. Again, K by L will not change significantly; in this particular case; how about here. In this case, let us say the change is same, I am trying to you know old MRTS new; it is not (Refer Time: 09:18) it is this is here K by L would change significantly in comparison to the previous case, because from here you will move to this point. Again, one thing that I am silent about; why I am choosing this particular MRTS? Why this MRTS?. Because what I am talking basically talking about; that earlier the price ratio was this, because MRTS is in your control you change the point the combination and your MRTS will change is not it; here you can say that I will produce here I will produce here and accordingly MRTS will change, but wait which where you will produce actually where MRTS is equal to the price ratio. So, again I have gone back to the previous concept that we have just talked about fine. So, in that case you are producing here, because this MRTS represents the price ratio in the market. So, if MRTS you need to change, because of price changes in the market and this is the new MRTS, what will happen; the K by L will change significantly. Here, K by L will not change significantly because we are talking about substitution remember these two are per near perfect compliment ok. So, substitution is not taking place that much, but here substitution would take place fine is it clear. So, basically if we go back and look at the geometry, what basically happening; when we are talking about MRTS, we are trying to measure the slope of this ISO Quant, but when we talk about elasticity of substitution, what we are trying to measure the curvature. Curvature of this ISO Quant, it represents the curvature of ISO Quant, how the different combination of how the combination of input to produce the amount of output would change with changing prices in the market and that is what we talk about in the elasticity of substitution is it clear. If you want, we can come back to it again, when we talk about profit maximization and there we will figure out that MRTS is nothing, where the production will take place it is equal to P what would it be equal to. P. L by PK. minus 1. Fine ok. So, it is doing this the proxy MRTS is a kind of a proxy for PL by PK and that we will see shortly. So, now, let us calculate MRTS, we had calculated MRTS for a particular case of Cobb Douglas function and do you remember what did we get? Minus. Minus of course, it is always minus MPL by MPK, but what did we get in case of Cobb Douglas function check. K by b L bK by a L. BK by a K. Fine; can we calculate now elasticity of substitution, how can we calculate the elasticity of substitution? You take here first the absolute value, what do we get? B a by KL and what we have learned that elasticity of substitution is nothing, but. delta p by L then This is what we have learned fine. So, we have everything, we can calculate and let us see what do we get? From here, if we take log both side we get log b by a plus log K by L and this is the property of. Log. Log a multiplied by b is equal to log a plus log b fine, and that is what we have used. We differentiate both side with respect to log of absolute value of MRTS, what do we get here we get one this, if we differentiate with respect to log of MRTS, what do we get; 0 and then here what do we get l n K by L. MRTS. MRTS; so, what we have got; and this is our elasticity of substitution. So, in the case of Cobb Douglas it is equal to. 1. 1 and of course, based on it we can figure out a new production function where; because this is quite useful, that a special class I am not going to discuss it in detail a special class of production function for which, elasticity of substitution is always constant not just one always constant some constant value. And of course, then Cobb Douglas would be a special case of that kind of production function, because it for Cobb Douglas also it is a constant value 1, fine. So, that that particular class of production function let me write it here, why do not you change it; why do not you check it; calculate let me write it here a K b L or let me write it this way this is of course, a production function Q is equal to a K to the power rho plus b L to the power rho and whole to the power 1 by rho and of course, there should be some restriction on rho, what should be that restriction? Greater than 0. Rho is greater than 0, probably check rho is greater than 0 and what you need to do now; calculate the elasticity of substitution in this particular case. And then; so that Cobb Douglas functions is nothing, but a special case of this production function; that is the homework you need to do fine ok. |
Microeconomics | Lecture35_Factors_Affecting_the_Budget_Line.txt | The example that I just gave you there also budget line you know is combination of 2 straight lines, but its it is not always true that the budget line is a line, it can be a curve. So, in that case we call it a budget curve rather than budget line for example, the price of coconut is changing with each unit of coconut fine ok. So, let us talk about what happens, this is the line that we have obtained 2 C plus F, 2 C plus F is equal to 8 and here we have C here we have F, and it is like this, 8 to 4. Now, what happens if the price of coconut goes up or in this world what happens that it becomes difficult to gather coconut. Now, instead of devoting that instead of you need to devote 2 hours to get 1 unit of coconut now you need to devote 4 hours to get one unit of coconut, what will happen to the budget line, it will rotate? Anticlockwise clockwise. Inward or outward, from where? Inward from 8. From 8. It would be convex outside is. Convex outside. So, right now what is happening from 2 C plus F, now we are talking about 4 C plus F , we are talking about 4 C plus F. So, of course, it will rotate right now how much maximum quantity of fish that you can catch still 8. So, this will not change, but how about the maximum amount of coconut that you can get 2, just 2. So, here it is going to rotate like this clockwise from 8. It would become more clearer if you pay attention to this, let me write it in the abstract form, the 2 items are C and F and price of C or time that you need to gather 1 unit of C is p c, p c, C plus p f, F is equal to I denotes the income P c denotes the price of coconut either in terms of time or in terms of money does not matter, C is the amount of coconut p f is the price of fish and F is the amount of fish gathered. So, if you can write it again earlier we wrote it you can write it what do you get or let me write it in terms of F because if is on y axis. So, F is equal to I divided by p f minus P c by P F C and what is this minus P c by P F is the slope, is the slope and what is slope. Slope is typically. In account also. Now, just mathematically, slope is rise divided by run that is the slope ok. So, when you move from this point to this run is 4 and how much is the rise in that case it is 8. So, how much is the slope 8 divided by. 4. Four and its moving in the negative direction. So, negative minus 2. But what does this minus 2 indicate, I just described it using economic concept, what does it indicate? That for 1 unit of coconut you have to substitute 2 units of fish you have to give away 2 units of fish. Very true. So, let us write it mathematically, you want to let us say you want to increase your consumption of coconut from let us say right now you are consuming c comma F. Right now you are consuming C comma F and you are changing from C comma F to C plus delta C and F plus delta F you are changing. Now, but you still want to satisfy the this constraint either the time constraint or the monetary constraint and you want to exhaust all your time or all your money whichever the resources you are using. So, what do you get in that case P c C plus delta C P F plus delta F and this is again equal to I. I. So, what do we get P c delta C plus P F delta F is equal to 0, why because from here we know P c multiplied by C plus P F multiplied by F is equal to I. So, in other word what we can write delta F by delta C is equal to. Minus P c by pf. Minus P c by P F and is it the slope, delta a by delta C this is the slope why because delta F is the rise and delta C is the run. Run. So, this is the slope and slope is equal to minus P c by P F that is what we got here and here we again get it using mathematics and what does it indicate. MRS Right now we have not studied the MRS we will study later on. So, we will come back to the that this indicates slope this indicates also relative price of. price of coconut with respect to fish. Price of coconut with respect to fish and it also indicates the opportunity cost, why it indicates the opportunity cost because here you have only 2 options either you catch fish or gather coconut. So, this is the opportunity cost for 1 unit of. Fish, fish or coconut. One unit of fish or one unit of coconut? Coconut. Think about it again. The fish because is in the denominator. Do not use mathematics, think logically you think about it. Opportunity cost of one unit of coconut. 1 unit of. Coconut. Coconut in terms of fish. Fish. Let us look at it, what you can do delta F is equal to minus P c by P F multiplied by delta C and put delta C is equal to 1, change in coconut is one what do you get. Delta F is equal to P c by P F. Pc by P F and then can you say then, then can you say what is minus P c by P F. They (Refer Time: 08:28). It is clear it should be clear that either it is opportunity costs of fish in terms of coconut or it is opportunity cost of coconut in terms of fish, but which one is this think about it, think about it ok. Do not get confused about it because if you are confused about it you will not be able to clear the next concept, the later concept that we will learn later fine. Now, we have seen the effect of change in price ok. Let me write it again here F is equal to I by P F minus P c by P F multiplied by I and if price of fish changes what happens this is the vertical intercept and this is the slope. So, a price of fish changes what happens, your vertical intercept changes and as well as the slope changes. Slope changes. And it is not clear from this equation, but in this case horizontal intercept remains the same. Same. Just this is the vertical intercept and this is the horizontal intercept, fine. So, that we understand now the effect of change in price. What happens when you have more resources at your disposal? Effect of change in income or time from our the example case what happens in that case of course, when I am talking about this I am talking ceteris paribus that the price of fish and price of coconut they are the same, they do not change from, a let us say now Robinson Crusoe can work for 9 hours. Shift outwards. It will shift. Outwards. Outwards and there will be a parallel shift. Yes sir. Fine, it is clear, nothing fancy, now think about it let’s think here just for a moment what’s happening to your budget set when the income is going up. Budget. Set. Set is increasing. Budget set is increasing in size the new budget set contains all the bundles available to you earlier, but also some more bundles. So, when this is the situation you cannot be made worse off probably you will be. Better off. Better off or you will remain the same ok, because earlier the choice you had those choices are still available to you with some more options, new options and these all are affordable to you fine. So, either you pick someone from the old budget set or the some new available set, if you are picking from the old budget set then you remain the same and if you are picking something new from the newly available bundles, why you are picking it probably they are making you better off, that is why you are picking something from the new of available options ok. So, either you would remain the same or you will be made better off fine. Similarly, here in the earlier case when the price of fish goes down, what happens when price of fish goes down? Vertical intercept rises. Vertical intercept rises, there is clockwise shift pivoted at the maximum possible coconut, in this case also you will be made either better off or you will remain the same. You cannot be worse off because of a decrease in price of fish fine. Now, also let me tell you why do you we use an example of, example of 2 good world. Can you think of it why do we use the example of 2 good world why not 3, why not 4, 1 is trivial that I had shown you. So, mapping we can map 2 goods in a 3 d 3 d of. 1 1 is of course, it makes life simple ok, it is easier to deal with 2 good world 2 dimensions in 2 good you get 2 dimensional graph. Yes. In 3 goods you get 3 dimensional graph and in 4 goods you get more than 3 dimensional graph that you cannot describe graphically up to 3 goods you can describe them graphically fine. But of course, 2 d diagram is lot easier than the 3 d diagram, but more important reason is that typically whenever you are making a decision. Right now most of the time, we talk about our decision about the particular good. Like for example, we can talk about when we are thinking that our decision about buying milk or buying cloth then what we can do, what we are really worried about is that our expenditure on milk and all other item. So, what we can do we can club all other item in a composite good all. So, what we can do right. Now, let us say if we are talking about our decision about buying milk, what we can do on x axis we can put milk and on y axis we can put a complex good, a made up good or just money, left over money ok. So, that takes care of the problem see in economics we always try to make our simpler and here if decision is we are talking about only milk then is good enough for us fine and what do we say of course, what would be the price of a money one unit of money 1. So, that is why when we talk about milk and the complex good we say the price of 1 unit of complex good is equal to 1. That represent because the money here presents your buying capacity of all other good the complex good, fine. Also let me tell you let us look at this example px or not px P F fish and P c coconut is equal to I, F and C you are controlling to certain extent as long as this equation is satisfied you are controlling. How about P F P c and I, they are beyond you, they are the parameters you take them as given. Now, let us see what happens when all the parameters goes up by 100 all the parameters go up by 100 percent. Nothing will change. Nothing will change, you still get in our earlier example 2 C plus F is equal to 8 and you get this line 8 here 4 and you have coconut here and here you have fish. Now, let us double everything what do we get. 4 c. 4 C 2 F. 16. 16, we can take 2 common 2 C plus F 2 multiplied by 8, 2 and 2 will get canceled, again we come back to 2 C plus F. Yes. You understand. So, in this sense what we can do if we take this logic what we can say we can declare 1 good as a numeraire good and we express everything in the economy in terms of that good ok. The term is numeraire ok, we say that here money does not make sense, even when you have large number of goods. Hm. We can still express that money in terms of a particular good. So, what we are basically saying, let us say if fish is our numeraire good what we will say that we will take the price of fish as 1 and how we can do it because we can take P F common that is what we did here if we take P F common what do we get F P c by. Pf. Pf C plus. I by. Is equal to I by P F. So, the everything in the economy is expressed in terms of price of in terms of fish. Sir I by P F. Multiplied by P F multiplied by P F and then P F P F will get cancelled. So, what do we get F plus P c by P F in to C is equal to I by P F this equation this equation is same as this equation these 2 equations are the same , but what is the difference, in the earlier example the price of fishes P F. P F. Price of coconut is P c and the income is I , but when we take here the price of fish as one what is the price of coconut. P c by P F. P c by P F. So, now, here you go what is the price of coconut in terms of fish P c by P F and what is the income in terms of fish I by P F understand. So, it is not necessary, that we make the first good as the numeraire good, you can do it with the second good also or you can do it with the money also. You can say that the money available to me is 1 and then how will it go, what we are doing P F by I we are taking I common. Hm. Multiplied by F p c divided by I multiplied by C and I multiplied by 1, I and I will get cancelled the new equation that we get here the income of this person is equal to 1. 1. So, here in this case we are you are making we are basing the economy in terms of money, here we are basing in terms of fish ok. It makes our life simpler how it makes our life simpler; remember here how many variables we have. 3. How many no, how many parameters we have. 3. Pf P c and i. 3. Three parameters and how many parameters we have in this P F by I and P c by i. 2. So, this problem is mathematically simpler than the earlier problem. That is why we take a good as numeraire. |
Microeconomics | Lecture98_Cost_in_Short_Run_TC_FC_and_VC.txt | What we are going to do now? For time being, we are going to move from long-run to short-run cost curves in short-run and how short-run is different from long-run. Sir at least one of the variables does not change. One of the inputs. Input varies input. One of the inputs cannot be varied and so far, we have been talking about that capital is that input which we cannot vary in the short run, but just keep in mind it can be any input. Varied. So, what we can do again we can start with the total cost, remember, this total cost; it can be divided into 2 components; one is fixed cost and second is variable cost. What we have learned in the long run the total cost is a function of Q, anyway, if you look at this graph what we have here is p and L, even if we use different combinations it will depend total cost will depend on the K amount of capital used amount of labour used and of course, amount of capital and labour would also determine the quantity produced, so total cost is a function of Q. How about the fixed cost? Sir, it would be By name, it should be clear that the fixed cost is the component of cost that does not vary with the output. Output capital. So, let us denote it by FC, in our case, we can say that if capital is fixed at 5 then fixed cost is the cost of having these 5 units of capital. But even further I am not purging it that much, but you can think of that fixed cost can be of 2 types, one is one you can say one component is I do not know what name to give just, just think about it that one component of fixed cost is that you incur it even if you do not produce anything. For example, you have decided to start a firm to produce let us say pen and you have already rented a building you have already hired some machinery ok, you are not producing anything you are not operating anything because you have not yet hired the workers. But still, you have to pay for the rent of that building and the rent for that capital ok. So, even if you are not producing anything you are incurring that cost, so that is one kind of fixed cost. Second fixed cost can be that it depends on only your decision to produce and no production if you do not product produce anything you do not incur any cost, but if you decide to produce even one unit then you incur that cost. But why I am keeping it in the fixed trust because that cost would not change with the level of production. For example let us say that government has this regulation that, if you produce you will have to get your come your output audited or your cost audited. So, it does not matter whether you put it and the cost of audit is let us say independent of your output, it is about if you hire an auditor you have to pay, does not matter whether you produce 5 units or 10 units. So, that is a different if you do not produce anything you do not have to that auditor. So, that cannot be a cost to you. But if you even if you produce just 1 unit you will have to hire that auditor. So, it is a different kind of fixed cost, but for theoretical purposes, so we would not distinguish between these 2 different kinds of fixed cost. Even the example that I gave you of the building we can say that again if we have a building and building can accommodate let us say production up 2000 units, but if you go from 1000, 1001, you will have to rent another building. So, now it becomes here dependent on the output, but here we will keep it if as long as we are producing in this zone 0 to 1000 of course, for this particular example and for theoretically as long as we are in the zone, we will keep the cost for that building in the fixed cost is it clear and what is a variable cost. Sir the cost of labour. In our example the cost of labour because cost of labour we can change, cost of cost of labour would change with the labour that we would change with different level of production. So, in other words variable cost is the component of total cost that varies with the output. So, I can say VC is a function of Q, so total cost that we have here is again a function of Q which has 2 components, a component that depends on Q and the second component that is independent of Q. So, we can try to draw it here this is very simple. let us say, here we have cost here we have output and this is fixed cost by definition, it is not it does not change with the output fine and here, we have let us say variable cost; typically variable cost would start at variable costs will start at the origin because it varies with the output, variable cost and by adding these 2 what we can get is total cost. So, from variable cost how can we get total cost it is just parallel shift and how much is the sift equal to. The fixed cost. The fixed cost this is the amount of fixed cost here it gets added, fine. Yes sir. So, just let me just write it here what we have done so far is that we have taken Q is a function of K and L and we have said in the short run K is equal to K naught. So, what vary, fixed cost we have r multiplied by K naught this is the fixed cost and how much is the variable cost. W l. W l because remember this is fixed this is not fixed, so what we get is total cost as fine. Yes sir. What we can do we have earlier defined the average cost that is basically average total cost, we can define also average using the concept of average; what we can define is average fixed cost and average variable cost. In short average fixed cost is AFC and it is nothing, but fixed cost divided by total quantity can you draw this; how would it look like? Sir, it would shrink to a low level. So, starting from really infinity because, when you have almost 0; like let us move in the direction of 0 direction towards 0, then what do we get really very high fixed cost on average, and we keep on producing more and more the fixed cost gets distributed over large amount of output. So, average fixed cost starts decreasing as the output increases and eventually, it becomes equal to 0 ok. Average fixed cost and here output more technical would be to say that asymptotically it becomes equal to 0 because in reality it never is equal to 0 and how about average variable cost. We cannot say with certainty, how would it look like, but let us say this particular case that we have taken, and Q is equal to let’s say K 0 is equal to 5 and Q is equal to 1 by 5 KL. So, in the short run what we are getting 1 by 5 multiplied by 5, multiplied by L, so Q is equal to L. So, in that case, variable cost is simply variable cost and Q output is simply a line passing through origin. So, in this case, average variable cost is going to be fixed. So, this is what I want to emphasize that do not get confused average fixed cost is decreasing with output, while average variable cost is fixed with the output, and this is common mistake that people make. But this is for only the given case. Even for the given case, but what we can have what I am saying there is a possibility that average variable cost is fixed, we may get let me use a different color we may have something like this or more appropriately if you think about it remember when we were talking about production function, what did we say that output increases in the beginning at the fast rate. So, in that case, if it decreases at the fast rate, it means average variable cost is decreasing and then rate of increase become. More. It starts decreasing of outputs the rate of increase in output starts decreasing and eventually, it becomes flatter and flatter. In. So, it so here we are getting the opposite result, and that we will see immediately. So, sir in this graph if L is varying, we got the final Q as L, if you shift it Let me erase this, so this is the case. sir, but if L is varying then this won’t be a straight line. Would not be a straight line, it would be something different that is what I am saying, so 1 case would be that it would be something like this. Ok sir. And there is a reason that I am picking this one, remember when we were talking about. Sir, but if labour is fixed; how will it be a variable cost, it will not come into the category. Labour is not fixed. Sir, like you the conversion is you did just before the graph. What we have learned here in this graph just look at this graph again, what is happening rate of increase in again we do not have the same production function as Q is equal to L; it is a different production function the functional form is not given what is happening here, L is increasing Q is increasing here rate of increase is increasing with L, but eventually, rate of increase in Q starts decreasing. Decreasing. So, what we can we have learned it means what is happening here what we have learned just in the context of return to scale, in this zone what we are getting is basically increasing return to a scale and in this zone, we are getting decreasing return to scale. Decreasing. Now coming back to your question here. Sir, the calculation you have done at the top right if you shift the screen a little lower, yes sir. Yes. Sir like you did K naught equals to 5, so Q equals to half 1 by 5 into K into L it is it Q so Q is equal to L. Ah And then you took that L is fixed so it is a straight line. L is not fixed because remember here what we have here is it is Q and here we have variable cost if we move in this direction Q is increasing. So, Q is increasing what does it mean that L has to increase that is why cost is linearly increasing. Ok in that way yeah that term. So, this is a straight line passing through origin. Origin. Now, what we have here we are talking about average variable cost ok. So, average variable cost is one way to say is that you get a point from on origin and get a point on the variable cost and you draw a line again you will get the same line slope is going to be same everywhere. So, that is why you get a straight line, but remember why it is decreasing here because, in the beginning typically we have already discussed that in the beginning we get some kind of increasing return to scale, and later on we get. Increasing return to scale. Decreasing return to constant return increasing first then constant and then decreasing return to scale. So, that is what is reflected here ok. So, but here it is decreasing first. So, the return to notice this we are talking about return to scale that is increasing that is what I am saying there is inverse relation between the cost function and return to scale if it is increasing return to scale average cost will be decreasing. Yes sir, according it will fix. So, that is why I am saying here in this zone what we are assuming that we get increasing return to scale even in the short run, but you can very well say; how can you talk about increasing return to skill when you are not varying the capital ok. So, what I mean to say is that rather than taking Q is equal to f of K naught comma L as the production function we take Q is equal to f of L as the production function you understand, fine. |
Microeconomics | Lecture96_More_on_Cost_Minimization.txt | So, we saw these three examples and what we try to do, we try to calculate the minimum cost to produce Q amount of output, given the price of 1 unit of capital is r and price of 1 unit of labour is w and of course not just 1 unit, it is 1 unit per unit of time. It can be per day per hour, just we have to be consistent about it and we have learn what we have derived that the cost function, let us say this is cost; cost function is a function of w r and output. So, notice this w and r; these two are given by market typically, it is a market determined and Q, Q is decided by firm. So, for a firm which is not able to influence w and r, cost function is function of output, I just want to show you to more things here. First, of course, we can take this not necessarily a Cobb-Douglas example, but isoquant we have different isoquant Q naught, Q 1, Q 2; In the consumer theory we talked about income expansion path. Similarly, here we can talk about output expansion path. Now, what we can do here is we can take see how much is the minimum cost to produce this Q naught amount of output and then we can say that we can try to figure out the cost and this is this and then what we have here something like this. So, optimal point is here, here and here. So of course, I have picked only 3 different level of outputs Q naught, Q 1 and Q 2. One can put many more levels of output and try to figure out the minimum cost to produce these respective outputs and here we have L and here we have K and if we try to draw a curve passing through all these optimal points, what we get is basically output expansion path or output expansion curve. And what we can do? We can again here what we have done? We have taken r as well as w; r is the rental for per unit of capital per unit of time and w is wage for per unit of labour per unit of time. So, we have kept r and w fixed and what we are doing? We are changing Q naught to Q 1 to Q 2 and we are seeing that where do the optimal bundles lie and we are drawing a curve passing through all those optimal bundles and that is how we are getting the optimal expansion path. But remember here this optimal expansion path that we are getting is on graph K L or L K to be precise. So, what we can do? We can take these optimal paths and we can draw here, we can say here, here we have output and here we have total cost C of Q and it is not exactly to the scale what we have here is and this is what we get here is C of Q; this is again here we are taking r and w fixed and what we are varying? We are varying Q and we are seeing it is impact on total cost and what here we have? This is total cost curve. Remember this total cost curve, this gives cost to produce respective amount of output, remember this has already been minimized, this is not any cost like Q naught can be of course, produced at higher cost also if you take this combination and this combination. Why, we are selecting this particular combination? Because this is the least cost required to produce Q naught amount of output. So, here we all the least costs are represented not just any cost. So, in the three examples that we saw, we can put we got two types of solutions. One is interior solution and what do we mean by interior solution? That all the inputs are used in the positive quantity, in the least-cost combination of inputs. For example, we talked about Cobb-Douglas production function ok. What did we get here the solution we had at this point; here we had L here we had K. At this one, this is the least-cost combination, here K is greater than 0 and L is greater than 0. So, what we have is interior solution. As opposed to interior solution what we have is something called boundary solution and what do we mean by boundary solution that at least there is one input which is not used in the which is not used in the least-cost combination of input to produce Q naught amount of output. So, at least we are not using one input, then we get boundary solution. For example, when we say we took a production technology where both the inputs were perfect substitutes of one another and in that case, we got the boundary solution. What happened, that at the least cost combination, only one of these two inputs was used. So, one thing if here if you pay attention to this interior solution, if isoquant is differentiable if isoquant is differentiable at all points, then the tangency criteria that we had learned is applicable. We can use tangency criteria to get the least-cost combination of inputs or the optimal solution of the cost-minimization problem. Remember even when the inputs were perfect complements, we got interior solution, but there we cannot apply the tangency criteria. Why? Because isoquant is not differentiable at all the points ok. So, if we want to use tangency criteria, we have to be certain about two things, one that isoquant is differentiable at all the points that we can check in advance. But the second is that all the inputs are being used in the positive quantity in the least-cost combination. So, this is a bit contradictory if you think of it. Why it is contradictory? We want to know about a solution that is why we are using this technique, but what we are saying that we can use this technique only if solution is of this kind. So, how would we know that solution is of this particular the interior solution kind ok so how can we proceed in such cases? Sir first we can differentiate. That is fine, that there is no problem with that. We assume that. We assume, we assume, and we proceed and then we check our result with our assumption. Contradiction yes sir. If there is contradiction, then there is problem. So, we do not have to worry about it. |
Microeconomics | Intro_An_Introduction_to_Microeconomics_Prof_Vimal_Kumar.txt | welcome to this lecture on microeconomics my name is doctor vimal kumar and i am an associate professor of economics at iit kanpur we are going to study what is microeconomics in this introductory video i will talk about three basic issues first what is microeconomics second why study microeconomics and third how are we going to study microeconomics ok so to understand what microeconomics is we should know what economics is and in hindi economics is known as [FL] or [FL] earth meaning material well being so economics as the name indicates very apt name in hindi that it is a study of material well being now when we talk about material well being the issue is that we all face scarce resources in our life what do we mean by scarce resources that limited amount of resources not just you and me but even if you think of bill gates who happens to be the richest person he also faces a scarcity maybe not a scarcity of money but a scarcity of time so everyone all human beings face scarcity an economics is important because we all have this scarcity of resources so in this course we are going to talk about how to allocate this scarce resources now economics has two major branch which is one microeconomics another is macro economics micro means at individual level that we study economics and in macro we study at aggregate level but you should understand you must have done botany or geology at your class eighth ninth tenth level that in botany if i remember correctly when i was a student we started with a mustard plant ok so must through the mustered plant we learned about the plant world through a frog dissecting a frog we learned about the animal world its somewhat the approach is similar but more fundamental in economics we study economics whether its macro economics or micro economics through the lens of microeconomics so it is the most fundamental for economics ok ah i should also clarify that some of the important topic which comes to mind whenever we talk about economics such as economic growth inflation unemployment all these things we will not study in microeconomics because these are aggregate phenomena and they are topics in macroeconomics so let me give you a very basic example that another way to talk about economics is that economics is nothing but smart common sense plus and let me clarify what i mean let us say you were unexpectedly given a free ticket to sonu nigam concert that is going to happen day after tomorrow let us say the market price of this ticket happens to be seventy five hundred but let us say if you dont want to go you can sell it for at most five thousand rupees now in the evening let us say you have already the sonu nigam ticket or in the evening you heard that at the same time shreya ghoshal is going to perform and ticket for her so is still available for seventy five hundred rupees if you had known this you would have bought shreya ghoshals concert ticket and not the sonu nigams ticket what should you do and this has now if you want to of course you should go which you should go for the one which gives you the maximum pleasure also taking into account the cost this is what we do in economics at you know basic you dont need to solve such problem you dont need any deep theoretical science but as you progress you would realize that problems become much more complex and it helps when you have theoretical foundation to understand all these things so welcome we are going to talk about choices we are going to talk about consumers we are going to talk about producers we will talk about what do we mean by competition what is monopoly and so on and so forth i hope you will join this course and enjoy this thirty hour lecture on mooc on microeconomics thank you |
Microeconomics | Lecture48_Indifference_Curve.txt | So, what we are getting or let us say by if we abuse this notation little bit what we can say that indifference set is given by x 1 plus x 2 is equal to k. All the bundles which if we add the quantity of good 1 and good 2 we get k would be on the same indifference curve. If we change the value of k we will get a different indifference curve, is it clear. Now, what we are saying that rather than describing it using set we can use graph to describe it. So, what we are saying, what do we mean by a indifference curve? Curve passing through all the consumption bundle such that they all give same level of satisfaction or same level of utility or same level of happiness to a particular consumer. And we can draw it here, here is x 1, here is x 2. So, x 1 plus x 2 again it is a straight line and when x 2 is equal to 0 it means we are on the horizontal axis, in that case x 1 has to be equal to k. Fine; and then when x 1 is equal to 0 we are on vertical axis and then we get x 2 is equal to k. And we know for a straight line this is an equation of a straight line, we know that we can draw a line, a straight line using only to two points, knowing only two points on this straight line. Will just, and this is the, this is one of the indifference curve for this person. Now, tell me one thing do you think it is sufficient to draw just, one indifference curve will it give you the idea of this person’s preferences this person’s preference. One in this case it is may, but that is not always true. So, what do we need? We need at least a family of indifference curve rather than drawing just one we should draw few more of course, number of indifference curve required would depend on the different person. But typically what we do if there is regularity in the pattern what we do that we draw at least 3 4 indifference curve if there is regularity in the pattern. And then a family of such indifference curve here we can rather than calling it indifference curve we can call it indifference line also, but typically it is not always a line. So, we call it indifference curve and now we take a family of indifference curve when your when someone’s preference exhibit some regular pattern and then a family of indifference curve would be called indifference map. And then you change the value of k, you will get this, this, this. Sir, indifference curve is a utility function. Indifference curve is not. I mean it represents the. What it represents is that all the bundles, all the consumption bundle on the same indifference curve give the same level of utility, it does not represent the utility function. But when you have a family of indifference curve now it indicates utility, it indicates what you want you do not need utility basically you need preference, you need to describe preference. So, now, if you have a family a large, some you know some indifference curve you are able to describe the preference of this particular person. So, you are able to describe his utility and so you are able to describe his preference or other way fine. But do you think it is sufficient? Rather than if I do not tell you this I just tell you this, do you think it is sufficient rather than giving you x 1 plus x 2 is equal to k. I give you these diagram. Do you think by looking at this diagram you will be able to express the, you will be able to describe the preference of this particular person. One more thing we need because we do not know the direction of its increase and decrease. So, sometime you should put an arrow that it is decreasing or it is increasing in this particular direction. Now you fully know, is it clear? So, now, we have learned about indifference curve and indifference map. Now, let me ask you can we have, let us take an individual can his two indifference curves intersect each other. So, what I am saying do you think it is a possibility of course, the theorem is more general than what I am describing because I am describing here only in two-dimension, its true for n dimension. Here we have x axis good for good 1, here it is for good 2 and let us have two diagram. Do you think it is possible to have something like this? Why? Because that implies that at one point x 1. So, let us say here we have x let us not put 1, just x, here we have y and here we have here we have z. Because sir, these are two different curves. But 1 has higher utility then the other. So, like if we take three points 1 of 2 on the x and z. And 1 over there. So, a point which is above that cannot have utility less than that want to say that. So, basically you are right. So, what we are doing here is that let us say we are talking about that these two curve are different, these two curve they are not representing the same level of utility same level of satisfaction they are representing different level of satisfaction. So, what we can say without loss of any generality that let us say that it represent k 1 level of utility and this represent k 2 level of utility and without loss of any generality k 1 is greater than k 2. It does not matter when you take k 2 is greater than k 1. What we are saying is that k 1 cannot be equal to k 2 because we are talking about two indifference curve. So, now what is happening, let us look at on let us say this is indifference curve 2 and this is indifference curve 1. Let us look at indifference curve 1, what we get that this person is indifferent between x and z. Now, let us look at the indifference curve 2. This person is indifferent between x and y, because x is on both of these indifference curve. So, can we say can you prove it of course, you know just by looking at it can you prove it that now because of this x this person is indifferent between x and. y. y indifferent between y and. z. z. Can you prove it? Transitivity. Transitivity, but transitivity z does not say anything about transitive what does transitivity say. That when x is at least as preferred as y and y is at least as preferred as z then x has to be at least as preferred as z. This is the transitivity axiom. So, that is what I am just going to show you for one example. How do we get this? This implies that x is at least as preferred as z and z is at least preferred as x. Both of these are true. Then we get this person is indifferent between x and z. Similarly what we get here is let me write here little differently, this person is indifferent between y, y is at least as good as x and x is at least as good as y. Now, combine these two what will you get? That this person prefers y; at least as much as z. And from here what we get? This person prefers z at least as much as y and when we combine these two what do we get this person is indifferent between y and z. So, from here we can reach to this person is indifferent between y and z. So, what does it mean? That in that case utility associated with y and utility associated with z should be equal, but what we have assumed that k 1 is greater than k 2; that is not feasible. So, why we are getting this sort of result? Because what we have assumed is wrong. So, using contradiction we prove that two indifference curve representing different level of utility cannot intersect each other, it is clear. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.