The original promise of microcredit was to reduce poverty by fostering self-employment in low-income communities, an idea first promoted at mass scale in Bangladesh (Yunus 1999). But critics of Muhammad Yunus and the Bangladesh microcredit model argue that supporting larger businesses (small and medium enterprises or SMEs) may instead create more and better jobs for poor individuals (e.g., Karnani 2007, Dichter 2006). That’s only possible, however, if those larger enterprises employ poor workers in large numbers. In a newly published paper, NYU economists Jonathan Buachet and Jonathan Morduch argue that that can’t be assumed.
Most studies of SMEs implicitly or explicitly compare them to large firms (e.g., Beck and Demirgüç-Kunt 2006). In this paper, Buachet and Morduch instead compare the employment and poverty outreach of SMEs to that of microenterprises. There is surprisingly little data on the profile of microcredit borrowers, and even less that might be matched to comparable surveys of SME employees. They draw on a series of surveys of both microcredit borrowers and SME employees, building from a 2008 survey of Bangladeshi SMEs which are customers of BRAC Bank, a for-profit arm of Bangladesh’s largest NGO.
The survey is particularly valuable in including questions that can be used to predict the likelihood that the employees’ households are below global poverty lines. They then compare those household-level predictions to similarly-constructed likelihood scores taken from independent data on microcredit borrowers in Bangladesh. Bauchet and Morduch relate those results to data from an additional microcredit survey with detailed consumption data. In focusing on customers of BRAC Bank, they narrow attention to SMEs that are most likely to align with BRAC’s broader imperatives of development, social welfare improvement, and poverty reduction.
Together, the three data sets show that the average employee of a small enterprise in their sample is a 26 year old male with almost five years of formal education and who is semi-skilled, while Bangladeshi microcredit
borrowers are mostly women, about half have no formal education and most have few professional skills. On average, microcredit borrowers are far more likely to be female (91% versus 7% of SME employees). Analysis
of the average likelihood that employees live in poor households shows a similar bifurcation.
The data show that SME employees work long weeks (on average 11 hours a day, six days a week), which creates employment barriers for women with primary child-raising responsibilities. As a result, SMEs in Bangladesh
are not typically creating jobs that reach the kinds of workers supported by microcredit, nor are SMEs employing many family members of microcredit customers. The findings here thus align with the lack of a robust
correlation between SME growth and poverty reduction found in crosscountry data (Beck et al 2005a).
The argument that “micro is too small” rests on the assertion that supporting larger businesses might be a more efficient way to achieve similar ends to microcredit. Bauchet and Morduch show that the proposition is only half right in Bangladesh. SME finance is more profitable and can create larger financial multipliers than investing in microcredit institutions, but they do not find that patterns of job creation (and, by implication, the distribution of social benefits) are likely to be similar.