Datasets:

---

baobabtech

/

decision-models-evaluation-docs

Like 0

Follow

Baobab Tech 9

ERROR: type should be string, got "https://doi.org/10.1093/jae/ejac012\n\nAdvance access publication date 3 June 2022\n\nArticle\n\n## Supporting Micro-enterprise in Humanitarian Programming: Impact Evaluation of Business Grants versus Unconditional Cash Transfer\n\nArays Abdullahi$^{a}$, Mohamed Ali$^{b}$, Elijah Kipchumbac,* and Munshi Sulaiman$^{c,d}$\n\n- $^{a}$Save the Children International, Mogadishu, Somalia\n- $^{b}$Dalarna University, Falun, Sweden\n- $^{c}$BRAC International, Kampala, Uganda\n- $^{d}$BRAC Institute of Governance and Development, Dhaka, Bangladesh\n- *Corresponding author: Elijah Kipchumba. E-mail: kipkechgm@gmail.com\n\n## Abstract\n\nHumanitarian programming in fragile economies often use unconditional cash transfers (UCTs) to offset food insecurity. However, there is an increasing focus on using cash transfers to boost household incomes beyond the short-term through micro-enterprise start-up and growth. This paper conducts a randomised control trial to measure the impact of three different sizes of business grants against UCT in Somalia. We find that giving the same amount of money as a lump sum business grant results in higher likelihood of business ownership and income compared with UCT in the short run (3-4 months after the transfers). However, the impacts are larger and persist 3 years later only for those who received larger amount of grants. The results indicate our 'medium'-sized grant being more cost-effective.\n\nKeywords: micro-enterprise, cash transfer, Somalia\n\nJEL classification: 016, 017, I32\n\n## 1. Introduction\n\nGenerating income from micro-enterprises is one of the prominent avenues of reducing poverty being pursued by development agencies in low-income countries. Initiatives to support micro-enterprise typically include transfer of capital, access to credit and skills training. While anti-poverty programs have a wide range in the comprehensiveness of the support packages to promote micro-enterprise, the use of unconditional cash transfer (UCT) has been on the rise due to its simplicity. With UCTs being a common feature in humanitarian programming, there is also growing interests to leverage on these transfers for building 'resilience' among the beneficiaries in fragile and conflict affected countries. This study looks at the long-term (3.5 years after transfer) impact of variations in cash transfers in the prolonged humanitarian crises context of Somalia. UCT interventions in Somalia typically gives small amount of money in monthly instalments to improve food security, whereas transfers intended to promote income-generating activity (IGA) are usually done as one-off grants. Besides testing the impact of lump sum 'business grant' vis-à-vis monthly UCT, we also varied the amount of cash given out as business grants.\n\n<!-- image -->\n\n<!-- PAGE_BREAK -->\n\nSeveral studies have investigated the impact of one-off small (USD 100 to 200) cash transfers to existing businesses to find positive results on their income in Ghana (Fafchamps et al. , 2014), Mexico (Mckenzie &amp; Woodruff, 2008) and Sri Lanka (De Mel et al. , 2008). On the other hand, small cash transfers to micro-business owners in Uganda (Fiala, 2014) and Tanzania (Berge et al. , 2015) did not find significant impact on income or business profit. Despite such differences in evidence, small cash transfers are generally expected to facilitate short-term business growth, especially for male entrepreneurs (Baird et al. , 2018). Studies involving lump sum cash transfers that have looked at the impact on business startup include, among others, Brudevold-Newman et al. (2017) and Hicks et al. (2017) assessing UCT of about USD 250 to youth in Kenya, Blattman et al. (2014) assessing the average transfer of USD 382 to youth in Uganda, Fafchamps &amp; Quinn (2016) assessing a transfer of USD 1,000 dollars to winners of business competition called aspire in Africa, De Mel et al. (2014) assessing a business training with a cash transfer of USD 130 to potential female entrepreneurs, Macours et al. (2012) assessing the marginal impact of a business grant of USD 200 with a business training layered on conditional cash transfer in Nicaragua and Beaman et al. (2014) assessing a cash transfer of USD 140 to farmer households in Mali. All these studies find positive impacts on business ownership and/or income in the short run, within a year after the transfer. Fewer studies that have longer term results generally indicate that the impact dissipate within a few years. Both Brudevold-Newman et al. (2017) and De Mel et al. (2014) found such declining trend within 2 years after the transfer. The study in Nicaragua shows a shift towards non-farm enterprises 2 years after the interventions and a weak effect on total income due to lower income from wage employment. Beaman et al. (2014) find that the positive effects on income from crop cultivation observed up to 2 years after the transfer did not persist in their 7-year follow-up survey. Blattman et al. (2020) found that positive effects of the grants 4 years after the transfer, but no impact after 9 years. One notable exception is De Mel et al. (2012), who find positive effects of cash transfer on the likelihood of business survival of male entrepreneurs 5 years after the transfer.\n\nSome of these experimental studies in the cash transfer literature have also tested the marginal effects of varying the frequency and size of the transfers. For example, in their study in Kenya, Haushofer &amp; Shapiro (2016, 2018) found that lump sum transfers are more likely to be spent on household durable goods and building assets compared with monthly transfers. Comparing the transfer of the same amount in monthly versus quarterly disbursements in Northern Nigeria, Bastian et al. (2017) found similar effects of both interventions across food security and assets, and they concluded in favour of lumping transfer to improve cost-effectiveness by lowering administrative costs. From the studies that compare the effects of different cash sizes, the evidence from Sri Lanka (De Mel et al. , 2012) and Kenya (Delius et al. , 2020; Haushofer &amp; Shapiro, 2016, 2018) shows an expected pattern of large cash grants generating larger impact on investments in assets, livestock and non-farm micro-enterprises. Recently, Kondylis &amp; Loeser (2021) found increase in transfer sizes reduces cost effectiveness both in the short-term and medium-term.\n\nDespite the prominence of cash transfer in humanitarian programming (ODI, 2015), their evidence is generally focused on food security and nutrition and do not measure impact on labour market outcomes (Baird et al. , 2018). For example, Hidrobo et al. (2014) compare the impact of cash transfer with food and voucher for Colombian refugees in Ecuador on food consumption only. Hoddinott et al. (2018) compare cash and food in Niger, and the indicator related to livelihood is expenditure on agriculture inputs. Lehmann &amp; Masterson (2014) find evidence of reduced labour supply in their study in Lebanon. The study was implemented in Somalia, which is known for its fragility and ongoing conflict. While humanitarian support has remained a key life-line in Somalia, there has been increasing focus on accelerating recovery and systematic resilience building. Other than humanitarian support, small enterprises are the main way for people to generate income, especially in\n\n<!-- PAGE_BREAK -->\n\ninternally displaced persons (IDPs) camps, and are largely managed by women. Programs aimed at supporting micro-entrepreneurs in the country typically combine business skills training with vocational or skills training, followed by small business grants.\n\nThis paper looks at the sustainability of short-term effects of business grants (over UCT) 3.5 years after the interventions. The study compares four treatment groups, who were randomised through public lottery, to receive (a) small UCT of $175 in two monthly instalments, (b) small one-off business grant of $175, (c) 'medium' business grant of $500 and (d) 'large' business grant of $1,000. Participants of the business grants groups also received a short business training after the transfers. Using three waves data, we examine sustainability of short-term impact on employment, micro-enterprises and income.\n\nOur estimates show over 3 years later, both medium and large business grants have positive effects on the likelihood of owning non-farm businesses, by 14 and 16 percentage points, respectively. The short-term impact of small business grant did not persist. By comparing the net worth of these enterprises, we find that the impact of the medium and large business grants converged in the long-term. Examining profits also shows similar pattern, whereby the point estimates of the impact on profits earned from micro-enterprises are similar between the medium and large grant arms, at around $20 in the past 1 month and $110 in the past 6 months, and statistically not different. Therefore, there seems to be a decreasing return to the size of business grants, and medium grant is more cost-effective.\n\nThe business profits directly fed to household income that increased by nearly same margin. We also find a significant portion of the profits were held by household as savings 3.5 years later. The larger income gains, however, did not yield substantial effect on food security both in the short-term and the long-term. Point estimates on household food security indicators (food stock at home, coping strategy index (CSI) and household dietary diversity) are positive in both follow-up surveys but not statistically significant when accounting for multiple hypothesis testing. With this introduction, the study design and contexts are presented in Section 2. Results are discussed in Section 3; first, the impacts of the three business grants versus the UCT arm. Section 4 presents some robustness checks with alternative specifications and some additional analysis. Section 5 concludes the paper.\n\n## 2. Evaluation Design\n\nInterventions under this study leveraged on two projects implemented by Save the Children in 2016-17. The projects, Humanitarian Support and Re-Integration of IDP and Returnees in Mogadishu supported by German Federal Foreign Office and Building Resilient Communities in Somalia supported by International Cooperation and Development of European Commission and DFID, included supports for micro-business establishment or expansion. These projects delivered varied cash grants sizes and thus provided an avenue to measure impact of lumped UCTs.\n\n## 2.1 Randomisation and treatments\n\nThe study was designed to understand whether the cash transfers can be utilised for generating impact on livelihoods by lumping the transfers, labelling them differently and including a business training. The interventions were randomised at household level to measure the marginal effects of lumping and varying cash grants with UCT as the control group. To implement the randomised interventions, field officers of the two projects identified eligible beneficiaries following standard selection process 1 . The beneficiaries\n\n1 A village relief committee is formed in each community consisting of clan leaders and people respected in the community. This committee organises open meetings attended by Save the Children field officers. They collectively identify the characteristics of vulnerable households and determine the eligible beneficiaries. Verification of the eligible beneficiaries is done by the project team based on project specific criteria.\n\n<!-- PAGE_BREAK -->\n\nFigure 1. Randomised Treatment Arms by Site\n\n<!-- image -->\n\nwere informed that the amount of transfer they will receive is going to be determined at a later stage and all the selected households will receive grants. We do not have a pure control group due to the concerns from the implementation team about potentially causing community backlash by not giving any support to some households after selection. After the selection was completed, the amount and method of transfer was determined through a public lottery attended by all selected beneficiaries. In each community, beneficiaries were divided into four groups with equal number of beneficiaries in each group.\n\nAs Figure 1 shows, the size of transfer for the households of Groups 1 and 2 varied between the two study sites (i.e., Mogadishu and Hiraan). This was due to budget constraint of the two projects and the programmatic necessity of disbursing the amounts within the time-frame. In Mogadishu, the small transfer was $250, whereas in Hiran, it was $100. However, in both cases, the same approach was adopted to make the transfer in two monthly instalments (Group 1) and as one-off grant (Group 2). Grant size for Groups 3 and 4 were $500 and $1,000, respectively, in both sites. The transfer took place between January and February of 2017 after the baseline survey.\n\nBesides the variation in transfer size, there are two important differences for Groups 2-4 compared with Group 1. Firstly, the transfer was framed as UCT for Group 1, whereas the three other groups were told about the transfers as 'grants for supporting IGA'. Beneficiaries of these three 'business grant' groups also participated in a 5-day business training conducted by a consultant for 1-2 hours per day. The training covered generic modules on planning, accounting, costing and marketing. This training took place after the business grants were disbursed, and there was no condition attached to the grant. Cost of training was about $38 per participant that included consultant fee and transport, refreshment for the participants and other logistics.\n\n## 2.2 Context and participants' profile\n\nWith decades of conflict and recurrent droughts, Somalia has a chronic humanitarian crisis. The country is consistently ranked among the lowest in most social indicators. Provision of basic social services such as education, health and nutrition is predominantly reliant on humanitarian agencies. Humanitarian supports in form of UCTs are common phenomena in the country, largely justified on the basis that markets are strong and responsive (Goodman &amp; Majid, 2017). At the tail end of transfers in this study, a nationwide drought was declared in February 2017 following consecutive failures of the rainy seasons. While the\n\n<!-- PAGE_BREAK -->\n\n2017 drought affected all sectors of the economy, it exacerbated the vulnerability of the urban poor and small traders.\n\nThe study was conducted at several IDP camps in Mogadishu and in urban setting of Beledweyne district in Hiran region. While both sites are affected by cycles of drought, parts of Beledweyne are affected by periodic flash floods from Shabelle River. Apart from reliance on humanitarian support, small enterprises are the main way for people to generate income and meet their daily needs in Somalia. Baseline data from the study showed limited economic opportunities for the population. About one-third of the households had any wage income, predominantly casual labour, in the month preceding the survey. Only 20% of the households earned income from micro-businesses, and over 40% of the households did not have any cash earning in past 1 month. With average household size of 6.4 members and 0.7 earners per households, these households live in extreme poverty. However, the urban residents from Hiran were relatively better-off compared with the beneficiaries from IDP camps in Mogadishu at baseline. Table A1 in the appendix shows the baseline characteristics of the survey households across the four treatment groups. Despite the randomisation being done through public lotteries, we see the groups are generally balanced with F-statistics of five of the 27 variables showing statistical significance at less than 10% level. None of those are related to indicators of business performance, which is the primary focus for this paper. Nonetheless, we control for some of these variables as part of robustness check of our key findings.\n\n## 2.3 Data and empirical strategy\n\nBaseline data were collected between December 2016 and January 2017. The public lottery took place immediately after the baseline survey. Cash transfers were done in January 2017 followed by the roll out of business training for Groups 2-4 in February and March. Beneficiaries of the UCT group (Group 1) received their second instalment transfers in February 2017. Follow-up surveys were conducted in May/June 2017 and August/September 2020. In this study, we present both the short-term (referred to as 'midline', conducted 3-4 months after cash transfers) and long-term ('endline', conducted around 3.5 years after the transfers) effects of the interventions.\n\nWe managed to survey 760 of the baseline sample at both midline and endline. This represented an overall attrition rate of 4.5%. All the impact results presented in the paper uses only these 760 households with panel data. Attrition was 7-percentage points higher in Mogadishu compared with Hiran, which is understandable given that most of the participants in Mogadishu are IDP. More importantly, we find attrition rates being statistically similar across the four study arms after controlling for baseline characteristics (Table 1). There appears to be some differential attrition between UCT and small business grant (F-stat 1.21, p-value 0.201 for the joint significance of the interaction terms) although the rates are the same across all four arms. Therefore, we conduct robustness checks of main outcomes by using inverse probability weights.\n\nTo measure the marginal effects of the business grant arms over UCT, we use the following Analysis of Covariance (ANCOVA) specification:\n\n$$Y _ { i t } = \\alpha + \\beta _ { 1 } ( M i d \\times T _ { 2 i } ) + \\beta _ { 2 } ( E n d \\times T _ { 2 i } ) + \\beta _ { 3 } ( M i d \\times T _ { 3 i } ) + \\beta _ { 4 } ( E n d \\times T _ { 3 i } ) + \\\\ \\beta _ { 5 } ( M i d \\times T _ { 4 i } ) + \\beta _ { 6 } ( E n d \\times T _ { 4 i } ) + \\delta _ { 1 } E n d + \\delta _ { 2 } Y _ { i 0 } + \\delta _ { 3 } S _ { i } + \\delta _ { 4 } X _ { i 0 } + \\epsilon _ { i } ,$$\n\nwhere Y$\\_{it}$ is the outcome of interest of household i at time t while Y$\\_{i}$$\\_{0}$ is the corresponding baseline values. Mid and End are dummies representing midline and endline, respectively. T$\\_{2}$$\\_{i}$ is a dummy of being randomised to receive small one-off business grant, T$\\_{3}$$\\_{i}$ is a dummy of being randomised to receive the medium business grant (USD 500) and T$\\_{4}$$\\_{i}$ is a dummy of being randomised to receive the large business grant (USD 1000). S$\\_{i}$ is a site dummy taking\n\n<!-- PAGE_BREAK -->"