Evidence on micro-lending effectiveness

Well, that's disappointing. Micro-lending was still relatively new when I was majoring in International Economics with a focus on Economic Development way back 35 years ago. It promised a lot and there was a natural logical appeal to it. Over the years, I have remained marginally interested in its progress as it spread to other countries, became more formalized, etc. I read some skeptical assessments but also some very positive ones.

Now there's this report, Microloans Don’t Solve Poverty by Ben Casselman.

What the mailing didn’t mention was that Morduch’s quote was from 2005, before a decade of research called into question the effectiveness of so-called microcredit as a tool for fighting poverty. More recently, a series of six independently conducted randomized controlled trials found that a variety of microlending programs had little to no effect on participants’ income or financial well-being. Morduch now says that the studies, along with earlier research that reached similar conclusions, suggest that the impacts of microcredit have been, at a minimum, “overhyped.” (A FINCA spokesman said Morduch’s quotation would be removed from future mailings.)

Researchers say the recent studies carry a broader message about the need for rigorous research in charitable programs. Microcredit has grown into a $60-billion-plus industry reaching 200 million borrowers worldwide despite limited evidence that it actually achieves its goals. Anecdotes like Chikaluma’s are powerful, they say, but only data can reveal what programs work and, just as importantly, how to make them more effective.

The idea behind microcredit is compelling in its simplicity: In many parts of the world, the best pathway out of poverty is entrepreneurship — selling firewood or food or clothing. But the poor often can’t cobble together even the few dollars necessary to get such businesses off the ground. By giving people access to small loans — often in the hundreds of dollars or even less — microcredit organizations allow would-be entrepreneurs to buy, for example, chickens to produce eggs or a refrigerator to keep food cold. And because the help comes in the form of a loan, not a grant, a single dollar of aid can be recycled to help many people.

[snip]

But even as microcredit drew attention, there was relatively little evidence that it succeeded in reducing poverty. Some critics even claimed the opposite: that microcredit left the poor worse off by saddling them with debt. News reports went so far as to link suicides in India to the stress of struggling to pay back microloans.

The recent research rejects both extremes. The studies find no evidence that borrowers are, on average, hurt by the loans. But they don’t appear to be helped much either. In a paper introducing the six randomized studies, economists Abhijit Banerjee, Dean Karlan and Jonathan Zinman walked through the findings: None of the six studies found statistically significant increases in household income or spending. Four of the six found no change in food consumption; one found a modest increase and the sixth found a significant decrease.

Some of the studies did find ancillary benefits. Borrowers got more of their income from their businesses, suggesting that they displaced other sources of income such as wages or government benefits. Those businesses also appear to have become more profitable. But the studies didn’t find any significant increases in school attendance or women’s empowerment in local communities, two commonly cited benefits of microcredit. The findings echoed those of earlier studies that also revealed minimal impacts.

That won't be the last word. There will be further studies. There will be debate. Perhaps there will be tweaks that might make it effective. But it seems like this might be one more logically seductive social policy that doesn't actually make a difference. Fortunately though, it does not, as is the case with some well-meaning social policies, actually cause negative consequences.

Evidence-based decision-making is harsh, holding good and well intentioned people to hard standards. All data has to be viewed skeptically but there are too many people in the well who eschew data completely if it stands in the way of making decisions which are emotionally satisfying rather than effective.

But some microcredit leaders reject the studies’ findings outright. Rupert Scofield, who runs FINCA, one of the largest and most prominent microlenders, said randomized controlled trials work well in medicine but don’t make sense in evaluating development aid. He said that for years, his response to claims like Duflo’s was to say he didn’t need data when he could see the effect firsthand.

“I would just say dismissively to them, ‘I don’t have time for that,’” Scofield said. “It’s perfectly obvious what our impact is … I don’t need proof. As long as the clients are coming to us, that’s all I need to know.”

FINCA has now developed its own measure of success, a step Scofield now says he wishes he had taken earlier. But he said he already knows that the outside researchers’ findings are flawed. “The fact that they would reach these conclusions that I personally know to be false really discredits them in my eyes,” Scofield said.

Damn the evidence! Full speed ahead. One more departure from reason into faith.