clear
set mem 500m
set more off
set logtype text
set matsize 3000

cap log close
log using "$Log/Table7.log", replace


use "$AnalysisData/bills_analysis_101-111_replication.dta"
keep if private==0

bysort v2 sponsor_state_abbrev sponsor_district sponsor_term_served: egen tot_bills=count(v2)

gen byte sponsor_democrat = sponsor_party==100 if sponsor_party~=.

gen byte NE = sponsor_state_icpsr>=1 & sponsor_state_icpsr<=14 if sponsor_state_icpsr~=.
gen byte MW = sponsor_state_icpsr>=21 & sponsor_state_icpsr<=37 if sponsor_state_icpsr~=.
gen byte SO = sponsor_state_icpsr>=40 & sponsor_state_icpsr<=56 if sponsor_state_icpsr~=.
gen byte WE = sponsor_state_icpsr>=61 & sponsor_state_icpsr<=82 if sponsor_state_icpsr~=.

replace sponsor_age=. if sponsor_age>100
gen sponsor_rookie=(sponsor_tenure_run==1) if sponsor_tenure_run!=.
gen byte leader = (comc==1) | (comr==1) if comc~=. & comr~=. 

gen lnpden=lnpop-lnarea
gen absMV=abs(MV1_democrat)

egen tag_congressmen=tag(v2 sponsor_state_abbrev sponsor_district sponsor_term_served)
egen group_congressmen  = group(v2 sponsor_state_abbrev sponsor_district sponsor_term_served)
egen tag_sponsor = tag(v2 sponsor_state_abbrev sponsor_district sponsor_term_served)
egen group_sponsor = group(v2 sponsor_state_abbrev sponsor_district sponsor_term_served)

cap drop pct_cosponsors_opposite
gen pct_cosponsors_opposite=100*numb_cosponsors_opposite/(numb_cosponsors+1)
replace pct_cosponsors_opposite=. if pct_cosponsors_opposite>100 

compress

egen tag_bill = tag(v2 HRnumber)

gen int decade = 1980 if v2<=102
replace decade=1990 if v2>=103 & v2<=107
replace decade=2000 if v2>=108

egen district_id = group(decade sponsor_state_abbrev sponsor_district)
 
local controls1 = "i.v2 "
local controls2 = "i.v2 MV1_female femaleXMV1_female "
local controls3 = "i.v2 MV1_female femaleXMV1_female "
local controls4 = "i.v2 "
local controls5 = "i.v2 "

local other_controls = "i.minor house_* sponsor_democrat sponsor_rookie sponsor_tenure_run sponsor_age leader ivycoll black occ0 occ1 occ2 occ3 occ4 borninstate tot_bills NE MW WE pct_black pct_urban pct_for_born pct_age_over65 lninc lnpden"
local billchars = "i.minor house_*"
local indivchars = "sponsor_democrat sponsor_rookie sponsor_tenure_run sponsor_age leader ivycoll black occ0 occ1 occ2 occ3 occ4 borninstate tot_bills"
local districtchars = "NE MW WE pct_black pct_urban pct_for_born pct_age_over65 lninc lnpden"

local if1 = "if tag_bill==1"
local if2 = "if sponsor_party==100 & tag_bill==1"
local if3 = "if sponsor_party==200 & tag_bill==1"

gen tenured=sponsor_tenure>=5 if sponsor_tenure!=.

foreach depvar of varlist numb_cosponsors { 
      forvalues l = 0/1 {
         forvalues j = 1/5 {
            forvalues k = 1/3 {
	           qui reg `depvar' sponsor_female `controls3' `other_controls' `if`k'' & mixed_gender_election==1 & tenured==`l', robust cluster(group_sponsor)
	           qui gen sample=e(sample)
	     
 		       di ""
	           di ""
	           di "j = ", `j', "k = ", `k'
	           di ""
	           di ""
	     
   	           di "rdbwselect_2014 `depvar' MV1_female `if`k'', c(0) kernel(uniform)"
	           rdbwselect_2014 `depvar' MV1_female `if`k'', c(0) kernel(uniform)
	           local band = e(h_CCT)

	           if `j'==1 {
	              di "reg `depvar' sponsor_female `controls`j'' `other_controls' `if`k'' & tenured==`l', robust cluster(group_sponsor)"
	              reg `depvar' sponsor_female `controls`j'' `other_controls' `if`k'' & tenured==`l', robust cluster(group_sponsor)   
	           }
	           else if `j'==2 {
	              di "reg `depvar' sponsor_female `controls`j'' `if`k'' & sample==1 & abs(MV1_female)<=`band' & tenured==`l', robust cluster(group_sponsor)"
	              reg `depvar' sponsor_female `controls`j'' `if`k'' & sample==1 & abs(MV1_female)<=`band' & tenured==`l', robust cluster(group_sponsor)   			
	           }
	           else if `j'==3 {
                  qui probit sponsor_female i.v2 `districtchars' `if`k'' & tag_sponsor==1 & abs(MV1_female)<=`band'
                  predict pscore if sample==1 
                  gen wt = 1/pscore if sponsor_female==1
                  replace wt =1/(1-pscore) if sponsor_female==0
	              di "reg `depvar' sponsor_female `controls`j'' `billchars' [aw=wt] `if`k'' & sample==1 & tenured==`l' & abs(MV1_female)<=`band', robust cluster(group_sponsor)"
                  reg `depvar' sponsor_female `controls`j'' `billchars' [aw=wt] `if`k'' & sample==1 & tenured==`l' & abs(MV1_female)<=`band', robust cluster(group_sponsor)
                  drop pscore wt
               }
	           else if `j'==4 {
                  qui probit sponsor_female i.v2 `districtchars' absMV `if`k'' & tag_sponsor==1 
                  predict pscore
                  gen wt = 1/pscore if sponsor_female==1
                  replace wt =1/(1-pscore) if sponsor_female==0
	              di "reg `depvar' sponsor_female `controls`j'' `billchars' [aw=wt] `if`k'' & tenured==`l', robust cluster(group_sponsor)"
                  reg `depvar' sponsor_female `controls`j'' `billchars' [aw=wt] `if`k'' & tenured==`l', robust cluster(group_sponsor)
                  drop pscore wt
               }
	           else if `j'==5 {
                  qui probit sponsor_female i.v2 `districtchars' `indivchars' absMV `if`k'' & tag_sponsor==1 
                  predict pscore
                  gen wt = 1/pscore if sponsor_female==1
                  replace wt =1/(1-pscore) if sponsor_female==0
	              di "reg `depvar' sponsor_female `controls`j'' `billchars' [aw=wt] `if`k'' & tenured==`l', robust cluster(group_sponsor)"
                  reg `depvar' sponsor_female `controls`j'' `billchars' [aw=wt] `if`k'' & tenured==`l', robust cluster(group_sponsor)
                  drop pscore wt
               }

		     if `j'~=2 & `j'~=3 {
	            scalar b_col`j'_row`k' = _b[sponsor_female]
	            scalar se_col`j'_row`k' = _se[sponsor_female]
	            scalar n_col`j'_row`k' = e(N)
			    sum tag_congressmen if tag_congressmen==1 & e(sample)
			    scalar ni_col`j'_row`k' = e(N_clust)
	            scalar ob_col`j'_row`k' = .
		     }
		     else if `j'==2 | `j'==3 {
	            scalar b_col`j'_row`k' = _b[sponsor_female]
	            scalar se_col`j'_row`k' = _se[sponsor_female]
	            scalar n_col`j'_row`k' = e(N)
			    sum tag_congressmen if tag_congressmen==1 & e(sample)
			    scalar ni_col`j'_row`k' = e(N_clust)
	            scalar ob_col`j'_row`k' = round(`band')
   		     }

               drop sample
            }   // closes j loop    
         }      // closes k loop

         preserve

* Display results

forvalues i = 1/5 {
   gen var`i' =.
}
gen str20 var6 = ""
		  
forvalues j=1/5 {
   forvalues k = 1/3 {
      replace var`j' = b_col`j'_row`k' if _n==6*(`k'-1)+1
      replace var`j' = se_col`j'_row`k' if _n==6*(`k'-1)+2
      replace var`j' = n_col`j'_row`k' if _n==6*(`k'-1)+3
      replace var`j' = ni_col`j'_row`k' if _n==6*(`k'-1)+4
      replace var`j' = ob_col`j'_row`k' if _n==6*(`k'-1)+5
   }
}   

keep var1-var6
keep if _n<=18

order var6 var1-var5

ren var6 sampletype
ren var1 OLS_all
ren var2 RD_bwidth
ren var3 RD_match_bw
ren var4 PS_match
ren var5 PS_match_indiv

foreach var of varlist OLS_all RD_bwidth RD_match_bw PS_match PS_match_indiv {
   gen str3 `var'_stars ="*" if abs(`var'/`var'[_n+1])>=1.645 & mod(_n,6)==1
   replace `var'_stars ="**" if abs(`var'/`var'[_n+1])>=1.96 & mod(_n,6)==1
   replace `var'_stars="***" if abs(`var'/`var'[_n+1])>=2.58 & mod(_n,6)==1
}

replace sampletype = "All" if _n>=1 & _n<=5
replace sampletype = "Democrats" if _n>=7 & _n<=11
replace sampletype = "Republicans" if _n>=13 & _n<=17

order sampletype OLS_all OLS_all_stars RD_bwidth RD_bwidth_stars RD_match_bw RD_match_bw_stars PS_match PS_match_stars PS_match_indiv PS_match_indiv_stars 
         export excel using "$Output/Table7_`depvar'_T`l'", firstrow(var) replace
         restore	
      }   // closes l loop
}		  

log close