Book review by Dr. Dale Adams
The recent book by Abhijit V. Banerjee and Ester Duflo (Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty, New York: PublicAffairs, 2011) joins Karlan/Appel in the food fight with the appologists for the microdebt industry. On the one hand, the RCT folks argue that microlending isn't doing much to alleviate poverty. On the other hand, Accion International, FINCA, Grameen Foundation, Opportunity International, Unitus, and Women's World Bank trot out a handful of anecdotes to "prove" that more debt is transforming the lives of millions of poor people in their "Measuring the Impact of Microfinance: Our Perspective." Everyone who is interested in the microfinance industry ought to take a reasoned position in this important debate, and reading B and D should help in that reasoning.
As was the case with Karlan/Appel, the B&D book reports on the recent results of random controlled trials (RCT). About a third of B&D focuses on microfinance issues: debt, savings, and risk/insurance. I enjoyed the fact that some talented economists have gone to the boondocks and collected large amounts of primary data, including talking with lots of poor people. I also appreciate their use of an evaluative technique that eliminates the selectivity bias, the attribution problem, and the representativeness problems that plagued earlier types of evaluations of microfinance, including anecdotes. I also like the fact that they focused their research on ways to improve development projects.
A few quotes from B&D gives a flavor of their controversial findings on microfinance:
[Referring to the results of an RCT study in India] The fraction of families that started a new business over the fifteen-month period went up from about 5 percent (in the control group) to just over 7 percent (in the treatment group)--- not nothing, but hardly a revolution (p. 171).
In our eighteen-country data set, the majority of businesses operated by the poor have no paid staff, with the average number of paid employees ranging from essentially zero in rural Morocco to 0.57 in urban Mexico (p.213).
Perhaps the many businesses of the poor are less a testimony to their entrepreneurial spirit than a symptom of the dramatic failure of the economies in which they live to provide them with something better (p. 226).
. . . [w]e are kidding ourselves if we think that they (microloans) can pave the way for a mass exit from poverty (p. 234).
In order not to appear fawning, I have five minor quibbles with B&D. The authors get lost in the weeds in their discussion of savings/deposits. I remember hearing Dick Patten, the Titan of deposit mobilization, say that capturing deposits was easy. Offer an attractive and flexible deposit product, locate the financial facility close to clients to keep their transaction costs low, provide security, offer a positive reward for depositors, and adjust the incentives of bank employees to reflect their work in capturing deposits. When these simple measures are in place, get out of the way so you won't be trampled by people rushing in to deposit.My second quibble with B&D deals with the displacement that occurs in microlending, a point that Milford Bateman has made on various occasions. The economic gains realized by one microborrower who wedges her new (or expanded) business into an already crowded market takes business away from other entrepreneurs in the market. Her gain may be another's loss. If B&D added displacement to their analysis, it would likely reinforce their negative conclusions.
Quibble number three relates to the transaction costs that clients of microfinance incur. B&D expend quite a bit of energy on how interest rates do or do not affect clients' decisions, but interest is only part of the clients' costs of transacting a loan or a deposit. In some cases transaction costs dominate their decisions and help to explain why informal finance continues to be so popular.
The fourth quibble is the same as one raised by William Easterly. RCT's are extremely useful in helping to make decisions about variants of a development program. For example, how to design the best set of incentives to induce people in malarial areas to place mosquito nets over their beds, or how to promote the use of de-wormning medication for children. RCTs, however, don't shed much light on larger policy issues such as selecting among a range of policy alleviation programs: education, health, nutrition, law-and-order, financial services, and land titling, for examples. RCTs also have limitations when it comes to evaluating programs that improve the quality of life for the poor gradually over time and don't leave a very large economic footprint in the short term, typically covered by an RCT study. I suspect this latter point holds for microfinance. The proponents of microfinance may be right that their activities take the edge off poverty and enhance the quality of life of millions of people, even though it doesn't lift many of them out of poverty as the RCT crowd is increasingly asserting.
My final quibble regards the cost of RCTs. The heart of economics is comparing costs with benefits. B&D and Karlan/Appel have exposed us to the benefits of RCTs, but they haven't shared information about the costs of these obviously expensive studies. One thing to be said for anecdotes is that they don't cost much, even if they can be seriously misleading. The profession needs some ideas about the costs of these studies before passing judgement on their overall usefulness and how many more RCT credit-impact studies out to be done . . . jane austin.