The 2006 Nobel Peace Prize winner and founder of the Grameen Bank (the first microfinance institution) Muhammad Yunus famously defined credit as a human right.
Today, microfinance providers reach more than 200 million clients globally. In Bangladesh, country of birth of microcredit, the Grameen bank has about 8.5 million borrowers, of which 96% women. When the New York Times interviewed Mr. Yunus in 2013 during a ceremony at the United States Capitol (where he received the Congressional Gold Medal for his efforts to combat global poverty), he explained that “long before the crisis of 2008, when financial institutions were crumbling all over the world, many of us had been saying that we need to redesign the financial system which only serves the top one-third of the world; two thirds are left out. Microcredit has shown how you can reach out to people that conventional banking cannot. It has demonstrated that it’s a doable proposition”.
Microfinance lending aims to overcome the high administrative costs and asymmetric information issues that conventional lending institutions face. In particular, the group lending model (a cornerstone of the Grameen methodology), is based on small loans to groups of self-selected people, who are jointly responsible for the loans of their group. If all members of the group repay, they are eligible for further loans. This model uses social networks to address what economists call “hidden information” (the inability to assess the risk type of borrowers – namely, high risk borrower vs low risk borrower) and “hidden action” (a.k.a. “moral hazard”, which happens because the actions of borrowers are impossible or extremely costly to monitor for lenders, which leads to less effort being put into using the money on loan). Group lending allows people who have better information about their community members to self-select (presumably preferring low risk type borrowers) and creates an incentive for self-monitoring (given the joint liability).
Despite its rapid diffusion and noble purpose, whether and to what extent microfinance reduces poverty is still the subject of intense debate. According to the Cato Institute, a public policy research organization, the potential for the world’s poor to be lifted out of poverty by microcredit institutions is grossly overestimated. Indeed, “the average poor person in the past (and today) is not an entrepreneur and when he or she has access to credit it is largely for consumption or cash flow smoothing”, explains Thomas Dichter, a development expert. Also, 30 years after the introduction of microfinance in Bangladesh, poverty statistics are worse than ever, suggesting that microfinance might not be as miraculous as it seems.
Not only the outcomes of microfinance are controversial, they are also hard to evaluate. Indeed, when assessing a microfinance program and attempting to model relationships within the data (for example, using regression analysis), one might face a couple of biases. Firstly, a so-called “selection bias”: comparisons between borrowers and non-borrowers cannot be trusted if the two groups systematically differ from each other. For example, if borrowers tend to be more entrepreneurial on average, then they are going to be more likely to set up a new business in any condition. If we claim that this is a consequence of microfinance programs, we are overestimating the effect of microcredit because a difference in the likelihood of starting a new business would have been observed even in the absence of it. The other main issue is the so-called “programme placement bias”: areas that receive the program might differ systematically from areas that do not receive it (for example, if I set up a program in one of the poorest London borough and then compare its poverty level with Chelsea, I might conclude that microfinance actually increases poverty).
A paper by Banerjee et al. published in 2010 addresses these biases by conducting a randomized evaluation. Spandana, a microfinance organization in India, selected 104 areas in Hyderabad (India) and paired them based on minimum difference according to different factors (e.g. per capita consumption, level of indebtedness) and randomly selected one of each pair to receive the program. This way, the authors made sure that the two groups (the one which received the program – the “treated” group – and the one which didn’t – the “control” group) were comparable. The study found that the effects of microfinance are heterogeneous for different kinds of people: households with an existing business at the time of the implementation of the program tended to invest in durable goods to expand it, while not changing non-durable goods consumption (e.g. food, clothes); households without an existing business but with a high entrepreneurial propensity increased spending on durable goods (investment for a new business) and decreased non-durable goods consumption (due to the need to pay a start-up fixed cost to enter entrepreneurship); finally, people with low entrepreneurial propensity increased spending on non-durable goods (using loans for consumption smoothing). These findings suggest that in the short-run microfinance succeeds in increasing investment on new and old businesses. Nevertheless, it has no statistically significant impact on education, health or women empowerment.
Even though microcredit might not generate broad-based economic growth, it can surely improve lives. Nevertheless, its long-run impact remains difficult to assess. As the late British development economist Peter Bauer puts it, it might be that “to have money is the result of economic achievement, not its precondition”.