In January, the American Economic Journal: Applied Economics published a special issue devoted to impact evaluations of microcredit. You can see an overview and some of my thoughts here (and some other write ups here and here). Justin Sandefur, from CGD, titles a blog post on the issue, “The Final Word on Microcredit?”
I really hope the answer to Justin’s (rhetorical) question is “No.” Because I still have a lot of questions.
If you haven’t been keeping score on the microcredit evaluations, all of which have been circulating for a few years now, here’s the bottom line according to Esther Duflo (editor of AEJ:Applied and co-author of one of the studies) “These loans do help, but the changes are not transformative, certainly not transformative enough to justify charitable donations to the standard microcredit model.”
To provide a bit more detail, while the six studies are quite different—who got access to credit, loan amounts and terms, local context, time, metrics, etc.—none found significant increases in income, consumption or spending on things like health or education. To be clear, none found any significant evidence of harm either. But “no one got hurt” isn’t exactly the kind of conclusion that inspires investment in or further study of microcredit.
That’s a shame because there are still so many unanswered questions:
- Why is microcredit take-up generally low (in the 30 percent range)? The people targeted by microcredit have liquidity constraints and have a hard time building up lump sums. They do a lot to get around those constraints. So why isn’t microcredit more popular? One possible answer is that the typical microcredit contract is unappealing and limits the benefits of relaxing those constraints—we’ve seen some tantalizing evidence of that but there hasn’t been nearly enough work on altering the basic microcredit contract by changing the repayment terms, allowing more flexibility to reschedule payments or the timing and amount of borrowing, or even encouraging consumption loans rather than demonizing them.
- If microcredit’s benefits are so muted, why is take-up and repayment so high? Borrowers often make sacrifices to repay and maintain access to an on-going stream of loans. A variety of studies strongly point to continued access to credit being one of the main drivers of high rates of repayment. Indeed, even in the midst of the Andhra Pradesh crisis, SKS was reporting that only 90 percent of clients had stopped repayment. Why were those 10 percent of clients still repaying, at least nominally in violation of the law?
- Since there’s strong and growing evidence of (at least potential) high returns to working capital, why doesn’t credit expansion yield better results? When retail shops have more inventory, when microenterprises have more raw materials, even when they simply have change on hand, they can boost revenue and profits substantially. So why are the results of the impact studies so muted? Again part of the answer is probably the structure of the product, and certainly there are behavioral issues at play, but that’s likely not the whole story.
- Since impact studies look at average effects, who are the people that benefit most? By design, impact studies look for average treatment effects. But a finding of small average treatment effects doesn’t necessarily mean that no borrowers see large gains. Do the people who have larger than average gains have identifiable characteristics? Or conversely, can we identify the people likely to see minimal gains?
- What if we not only had better products, but we had better targeting of specific products to specific needs? Why can’t we seem to make much progress in improving the operations of microenterprises? There’s no question that there’s plenty of room to improve business practices from accounting, to inventory management, to operations, to marketing.
That’s just a sampling of the questions I still have. But I’m worried about getting answers. More on that in the next post.