What Don’t We Know about Microfinance

Another great summary blog on last week's Microfinance Impact & Innovation Conference by Laura Starita over at Philanthropy Action:

What We Don't Know About Microfinance

by Laura Starita

Both days one and two of the conference were largely dedicated to communicating the results of recent impact studies—i.e. what we “know” in microfinance. Yet the first session of Day 2 took a step back to discuss what we don’t know, but should. Panel participants were Chris Dunford from Freedom from Hunger, Rich Rosenberg from CGAP and Abhijit Banerjee from JPAL.

Chris Dunford spoke first, giving a thoughtful presentation that effectively amounted to a distillation of the disadvantages of RCTs for practitioners. He called for a broader focus on evaluation and not just impact studies. Evaluation, in his mind, is the practice of answering definitional questions (What do we mean by microfinance?); value questions (Are the poor overindebted? Are interest rates too high?) and decision making questions (Why don’t policy makers and donors consider the results of RCTs when making funding decisions?) as well as impact questions (Does microfinance improve the lives of the poor?).

His view was not that impact studies are useless, but that they are limited. More relevantly, they have limited applicability to the people making funding decisions. [If this conference reinforced anything it was the difficulty that both the researchers and practitioners alike had in translating research results into practical actions.]  One major limitation of RCTs that he highlighted was the fact that the MFIs that work with researchers to measure an intervention are often rolling out a product or service that is either new to them or offered in a different way from how the bank might do it on its own. This newness means that the results may be affected by how well the program was implemented, clouding the feedback on whether it “works” or not.

The larger issue, however, is that massive amounts of funding is already going toward microfinance, after all, so in a way it is almost too late to be asking “if” microfinance works. In fact, the “if” question is tacitly critical to the people who deployed the money in the first place (and will deploy still more in the future). More useful is to help the momentum of investment alight on products and approaches that work better.

Dunford’s bottom line? Keep using impact evaluations to answer the “does it work” question, but don’t overlook the value of the kind of narrative, qualitative work like the financial diaries that underlie Porfolios of the Poor. They offer, he said, a “less expensive and more intuitive ways to get non-intuitive results.“

Rich Rosenberg from the Consultative Group to Assist the Poor (CGAP) offered the second set of comments on what we “don’t know” about the poor. His talk focused on over-indebtedness. In his view, regulators are “flying blind” because they do not know how pervasive over-indebtedness is in poor communities, and experience on overindebtedness in the developed world shows the potential for extensive, long-term economic damage on a regional, national and even international scale. The challenge begins with the fact that there is no working definition of what it means to be “over-indebted.“ Rosenberg offers some thoughts from the work of a researcher named Jessica Schicks on the circumstances that might arise in an individual case, such as falling behind on payments or making unacceptable sacrifices to stay current. Beginning with a useful definition, he urges the identification of easily measurable proxies for overindebtedness that can be used by banks and policy makers to identify whether an individual, or indeed a region, is struggling with too much debt. Ideally, he wants something that shows up before the “tailing indicators” of default and delinquency.

Abhijit Banerjee from the Latif Jameel Poverty Action Lab at MIT followed with a presentation that asked why microenterprises do not grow. He opened by presenting a set of facts that are empirically true about microentrepreneurs, but that do not seem to fit together. Specifically, that the marginal returns of their business are relatively high (in the double digits) and that many have access to micro-credit. Yet few avail themselves of formal credit even when they use more expensive money lenders, their businesses never grow to take on additional employees, and they never decrease the amount of debt they carry. How could this be?

Banerjee offers three possibilities:

1) Micro-credit is limited, and therefore does not allow for growth
2) The marginal returns to micro-enterprises are high up to a very limited point, after which there is a huge hurdle to jump before the returns rise again. Poor business owners may see the hurdle, feel they can’t jump high enough and give up
3) The businesses were never intended to grow

Banerjee’s comments align with some of the work of Scott Shane, author of The Illusions of Entrepreneurship. Women entrepreneurs around the world, including in the developed world, enjoy lower profits and returns to capital than men. Shane’s research showed that a certain percentage of female entrepreneurs start their businesses in order to contribute to their household in a limited way on their schedule. Entrepreneurship for these women is a lifestyle choice—the businesses were never intended to grow.

We will be writing up a more detailed version of Banerjee’s comments in the coming weeks.