But most such studies had a significant problem: If some households with higher borrowing also display higher earnings, can we really know which is causing which? Does credit make households less poor, or does being less poor make them more apt to borrow money in the first place?
In recent years, a new generation of development economists has addressed this problem by experimenting with microcredit programs, randomly offering loans to some people and not others. Just as in the best drug trials, this has allowed researchers to measure cause and effect more precisely. If, one year later, those who received loans earned more or enrolled more of their girls in school, what factor other than the loans could explain those happy outcomes?
The first randomized studies of microcredit appeared in 2009. MIT economists found that in the slums of the megalopolis of Hyderabad, India, small loans caused more families to start micro-
businesses such as sewing saris. Existing businesses saw higher profits. But over the 12 to 18 months the researchers tracked, the data revealed no change in bottom-line indicators of poverty, such as household spending and whether children were attending school. Perhaps those who made more from their own businesses earned less in wages outside the home. A study in Manila by American economists Dean Karlan and Jonathan Zinman also found no effect on poverty for families one to two years after they received a loan.
On the heels of such studies came more bad news. Microcredit bubbles began to reveal themselves as financial bubbles do — by popping, in places such as Nicaragua, Bosnia, Morocco and Pakistan. In Bosnia, for instance, the microloans outstanding shot from $275 million in 2005 to $1 billion in 2008, before slumping to $830 million in 2009 on defaults and write-offs.
In 2010, amid reports of suicide among overindebted borrowers, the government of the Indian state of Andhra Pradesh ambushed the microcredit industry there with a harsh law that all but shut it down. Microlenders must now register with the governments of the districts in which they operate and must seek approval for each loan. This puts much power in the hands of local officials, who in some cases are known for their aptitude in converting such leverage into delays and graft.
I believe that Andhra Pradesh overreacted in quashing rather than reforming the industry. But in talking to borrowers there in late 2010, it became clear to me that, even if the state was overreacting, it was doing so in response to a real problem. The proliferation of fast-
growing microlenders had made it easy for poor men and women to get in over their heads in hundreds of dollars of debt. Not all were wise enough to avoid the trap.
These bubbles may be the first in history fed more by charity than by greed. Almost all the large-scale financing for microcredit has come from private donors, socially minded investors, public aid agencies and, in India, private banks complying with legal quotas for lending to the poor and minorities. Ironically, almost all were motivated by the idea that microcredit was a sure-fire aid to the poor.