IV. Returns to capital and access to credit

Many studies have looked at the impact of access to cash grants, loans, or in-kind transfers on microenterprises. In an influential paper, de Mel et al. (2008a) use the straightforward approach of providing cash grants to microenterprises in Sri Lanka to estimate their returns to capital. Treated firms received either an in-kind or cash grant, in the amount of 100 or 200 USD. Participants were told that it was not required that the cash grants be spent for business purposes. All types of grants significantly raised the capital stock of microenterprises compared to the control group. Pooled together, the treated arms enjoyed a significant increase in profits: a grant of 100 USD raised profits by 5 USD per month, which corresponds to a return of 60% per year. Note that this estimate is not exactly the marginal return to capital, as recipients might not have invested all of the money in their firm, or they might have complemented the grant with additional capital from other sources. One of the striking findings in the study is that the return to capital for female microentrepreneurs, who make up half the sample, is not statistically different from zero. I discuss this further in the subsection on gender below. Follow-up data collection shows that the benefits of relaxing capital constraints were persistent up to five years later (de Mel et al., 2012). The same type of intervention among small, male-owned businesses in Mexico also seemed to lead to large but imprecisely estimated increases in profits, with suggestive evidence of larger effects for the in-kind grants (McKenzie and Woodruff, 2008).

Fiala (2018) offers loans or grants to microentrepreneurs in Uganda, cross-randomized with business training. Surprisingly, loans have larger effects than grants; they increase profits by 50%. The author speculates that receiving a loan may have pushed men to make more productive investments due to the obligation to repay, which a grant does not entail. This interpretation requires some other friction or behavioral bias, as the same options were in recipients’ choice sets under grants. None of the treatments have positive impacts for female microentrepreneurs.

Fafchamps et al. (2014) vary whether grants offered to small businesses in Ghana are given as cash or in-kind. For the in-kind treatment, the business owner specified what capital he or she wanted (typically inventory) and the research team delivered it. Overall the in-kind capital has a larger impact on profits than cash. Another result, echoing other studies, is that the impacts of the cash treatment are larger for men than women.

Bianchi and Bobbba (2013) use variation in the receipt of Progresa conditional cash transfers in Mexico to study how cash transfers affect the extensive margin of entrepreneurship. They find that entrepreneurship increases by 1 percentage point from a base of 3 to 4 percent with the cash transfer. Through further analysis, the authors argue that the impacts are due to the program providing steady income and thus reducing risk aversion, rather than the easing of credit constraints.

Another intervention that, like Progresa, was not focused on entrepreneurship yet could have affected it is analyzed in Haushofer and Shapiro (2016). This study examines impacts on an array of outcomes from one-time cash transfers to poor individuals in Kenya from the non-profit GiveDirectly. The authors report a 33% increase in revenue from self-employment and a 50% increase in business expenditure, but no changes in profits or the likelihood of being a business owner.

Blattman et al. (2014) use cash grants to promote entrepreneurship among young men and women in post-war northern Uganda. Groups of around 20 to 30 people, most of them mixed gender, submitted applications to the government’s Youth Opportunities Program, in which they wrote up a plan for how they would use the grant to acquire skills in trades such as tailoring, hairdressing, or carpentry. Funding was provided to 265 groups, randomly chosen from the 535 eligible groups. Groups received about 7,500 USD on average, or around 400 USD per person. The grants were successful in promoting non-agricultural employment among recipients. The treated group also had higher earnings, worked more hours, and accumulated more capital than the control group. Receiving the grant caused the treated group to adopt formal business practices such as record keeping and paying business taxes. Women benefited more from receiving the grant, and their earnings grew much faster than those of women in the control group. The authors speculate that this heterogeneity by gender is due to women facing more stringent credit constraints. Despite these impressive initial results, a follow-up nine years after baseline found that the treatment group no longer had significantly higher income than the control group. However, those in the treatment group, particularly women, were still more likely to be engaged in a skilled trade at the time of the nine-year follow-up survey.

Blattman et al. (2016) study how a large cash grant affects the entrepreneurial activities of marginalized women in war-torn northern Uganda. The intervention provides 150 USD in cash, equivalent to two years of income, plus 5 days of business training. Each of the 120 study villages was asked to provide a list of marginalized individuals; from these lists 10 to 20 were selected to participate in the study. Randomization was at the village level. Half of the treated villages also received group dynamics training. The treatment doubled entrepreneurial activity and income compared to the control group. The group dynamics training, which consisted of advice on how to collaborate in business ventures and instructions on how to form a ROSCA savings group, led to significant increases in income.

Another study of cash grants is by Klinger and Sch¨undeln (2011), described in the previous section, which analyzed TechnoServe’s business plan competitions in Central America. That study found that receiving a large grant as one of the winners of a business plan competition had a large impact on the likelihood of starting a business.

The studies above suggest that in many settings, giving grants to microentrepreneurs can have large returns. This evidence that small firms are capital constrained provides prima facie support for the hypothesis that, if the capital were in the form of loans rather than grants, recipients could expand their businesses and pay back the loans through higher profits.

Inspired by Muhammed Yunus’s Grameen Bank in Bangladesh, many organizations around the world specialize in providing loans to microentrepreneurs. The microcredit lending model varies across microfinance institutions, but often has elements aimed at encouraging repayment and achieving a broader social mission. For example, the loan structure often requires repayment to begin almost immediately. Several organizations also focus on female clients and use a group lending model with liability for the loan jointly shared by group members.

Several studies examine how expanded access to microcredit affects microenterprises. The thrust of the recent literature is that microcredit helps some microenterprises, but access to credit is not a silver bullet to transform most of them into thriving, fast-growing businesses. The title of a prominent recent study, “The Miracle of Microfinance? Evidence from a Randomized Evaluation” exemplifies the tamping down of excitement about microcredit associated with this recent literature (Banerjee et al., 2015a). That study uses the randomized roll-out of Spandana, a for-profit microfinance lender in India, across neighborhoods in Hyderabad, India that began in 2007. Spandana lends to female microentrepreneurs using a group lending model. Take-up of the loans is fairly low, at 18% in the treatment group (compared to 5% in the control group), indicating that demand for microcredit is not universal. The authors find no statistically significant increase in business profits, but they do find improvements for households with preexisting businesses.

Six other randomized evaluations of microcredit published at the same time as Banerjee et al. (2015a) find broadly similar results that access to credit only helps a subset of firms (Banerjee et al., 2015b). For example, Angelucci et al. (2015) worked with Credito Mujer, which offers group-lending joint liability loans to women in Mexico. The study uses an encouragement design, randomizing in-person marketing of Credito Mujer. Take-up of the loan product is 19% in the treatment group, compared to 6% in the control group. The study finds an increase in revenue but not profits or household income. Attanasio et al. (2015) randomly offer microcredit to women in Mongolia and find that it increases self-employment. There is no impact on self-reported income or profits, with the loans mostly going towards consumption. Meager (2019) presents quantile estimates that pool the seven studies that were published in tandem, using a Bayesian framework. She shows that the impacts are quite precisely zero for most of the distribution, with imprecise estimates pointing to positive impacts for the right tail.

Several papers on microcredit examine how the contract structure of microloans influences their impacts. Field et al. (2013) use a randomized experiment in Kolkata, India to study whether the requirement of high repayment amounts early in the loan period discourages profitable opportunities that entail a high upfront outlay but greater returns in the longer run. They find that adding a grace period to individually-liable microfinance loans, during which no interest payments are due, increases business investment and profits and decreases non-business loan use. The trade-off is that the loan default rate was three times as high in the treatment group as in the control group. Fischer (2013) points out that the joint liability structure that is common in microcredit could lead to inefficiently low risk-taking. The paper’s theoretical insight is that because fellow group members bear the downside risk but do not share in the upside risk, they might block entrepreneurs from taking advantage of high-risk, high-return opportunities. Other work compares group lending to individual lending (Gin´e and Karlan, 2014; Attanasio et al., 2015).

 

Gender

Explaining gender differences in returns to capital

A follow-up paper to de Mel et al. (2008a) further probes the gender differences in returns to grants in Sri Lanka (de Mel et al., 2009). The authors argue that women’s grants were “captured” by other household members, but that larger grants were more difficult to capture. This interpretation is related to the more general point that there is often no rigid separation between business and family for microentrepreneurs. One way this materializes is that pressure from family to share income can lead to income hiding and decrease the likelihood of starting a business or investing in an existing business (de Mel et al., 2009; Jakiela and Ozier, 2016). This challenge might be especially large for women, who typically have less power in the family than men. Another way home and business are entangled is that, with access to capital, an individual faces a choice to spend on household expenses or the business, and women might put more weight on or have more responsibility for spending for the household.

Consistent with this conjecture, Fafchamps et al. (2014) find that in-kind transfers are more effective than cash transfers in increasing profits for female business owners. Women are more likely to use cash transfers for household expenses and to make transfers outside the household. Friedson-Ridenour and Pierotti (2019) find that spousal and societal pressure can push female entrepreneurs to invest less in their business, save more for the household, and hide income so as to not have to pay for their spouse’s share of expenses. However, not all studies find support for this idea that pressure from others explains the gender patterns: Fafchamps et al. (2014) find that external pressure does not play a role in women’s lower business profits.

Bernhardt et al. (2017) systematically examine whether the pattern that women have a lower return to capital than men can be explained by these intrahousehold dynamics. They re-analyze data from prominent studies and find that returns to capital are lower for women in households with other entrepreneurs than for men who live in households with other entrepreneurs, but household-level profits do not differ. That is, when women receive grants or loans, the husband’s or other family members’ profits increase, pointing to the money being used for an enterprise other than the woman’s own enterprise. Consistent with this story, women in single-enterprise households have high returns to capital.

Klapper and Parker (2011) offers a different explanation, which is that the divergent returns to loans between men and women is due to the industries they work in, and that adjusting for the different industry composition, returns to loans do not differ by gender. Industry can explain much of the raw difference because women tend to work in industries that have smaller, less efficient firms, with lower potential for growth. An area for further work is to better understand why women work in industries with lower growth potential or, as suggested by Hardy and Kagy (2019), more competition.

Impacts of access to credit on women’s empowerment

Another gender-related question in the literature is whether business success improves women’s empowerment and, in turn, children’s outcomes. One motivation for microcredit being focused on women is that women are excluded from other credit channels. But an additional motivation is that helping women, and specifically giving them earning power, might increase their decision-making power. Women having a greater say in household decisions is a desirable end in itself for equity reasons if decision-making is dominated by men, and is also hypothesized to lead to better outcomes for children, for example if mothers put more weight on children’s health and education than fathers do. Thus, the literature on microcredit has examined if one of the downstream effects of access to credit is greater women’s empowerment. Many studies such as Banerjee et al. (2015a) find no evidence that access to microcredit improves women’s empowerment. An exception is Angelucci et al. (2015) who find some impact on women’s decision-making in the household.

 

Recap

Several studies that offer grants to small businesses find tantalizingly large impacts on profits. This suggests that, at a minimum, this type of philanthropic intervention could be helpful to many small firms. The high returns also suggest that this intervention could pay for itself if the capital could be given in the form of loans, with a high enough repayment rate and low enough administrative costs. The microfinance industry aims to do this, but several studies reach a similar conclusion: microcredit only has meaningful impacts on business performance for a small share of recipients.

There are several potential ways to reconcile the findings on grants versus microloans. The firms targeted in the cash grant studies tend to be somewhat larger and more established than the typical firms receiving microcredit loans. This is consistent with the finding that microcredit increases profits for only a subset of recipients and being an established firm is a predictor of being in this subset. In addition, the term structure and other requirements of microcredit might inhibit the sort of high-return investments that grant recipients undertook. These potential reasons suggest that more research on how to identify high-potential microentrepreneurs (as discussed in section 6) and on how loan terms influence the way the capital is invested would be fruitful.

A stark pattern across several studies is that grants often improve business outcomes exclusively for male-run businesses. Recent work identifies one important reason: money given to female entrepreneurs is often not invested in their businesses, whether by their choice or not. In light of this explanation, more work is needed to test if grant and loan programs can be redesigned to enable women to invest capital they receive in their businesses.

Previous Chapter III. Business Training
Back to top
Next Chapter V. Barriers to hiring