Letter to the Editor

To the Editor of the New Yorker

Dear Sir,

As I read the excellent May 17 article on Esther Duflo and her randomized evaluations of social policy tools, I was surprised to find CGAP (a World-Bank affiliated research and policy center for microfinance) lumped together with the microfinance networks that have been discounting the results of recent studies by her and other randomistas. On the CGAP blog, I have been criticizing—tactfully, I hope, but clearly—the defensive reaction of these networks. Far from discounting randomized trials, CGAP has been supporting them to the tune of $1.2 million of our own money and funds we raised from our members.

Prof Duflo dismissed a line I had written (Does Microfinance Improve Their Lives? The Poor Say Yes) as “the moronic revealed-preference argument,” that is, the argument that if people keep using something it must be good for them. The offending line occurred in a position paper I wrote for CGAP. It was the title of a section that included the statement: “Of course, repeated use does not by itself prove that a service is benefitting users. No one would make this argument about repeated use of heroin, for instance. People do not always borrow wisely.”

The CGAP paper clearly acknowledges that randomized studies by Prof. Duflo and others have raised fundamental questions about the claim that microcredit raises incomes and consumption, lifting poor people out of poverty—so much so that funding solely on the strength of that claim would not be justified at present. The paper then argued that there were reasonable (not conclusive) grounds to think that microfinance at least helps poor people to cope with poverty, for instance by smoothing irregular consumption and cushioning against shocks. Consumption gaps and uncontrollable calamities are serious problems for poor people. They place a high value on financial tools that help them cope with these problems. And well-executed microfinance can create such tools at very low cost in terms of public subsidy. Finally, the paper recognized that these arguments need to be tested by further research.