I first learned about Esther Duflo and her approach to poverty in Ian Parker’s profile, The Poverty Lab, in the May 10th issue of The New Yorker. Subsequently I watched her presentation at the TED 2010 Conference where she spoke about the critical role of randomized field experiments in formulating social policy in the developing world.
Doing this kind of research doesn’t seem especially innovative but, according to Duflo, it is rare in economics, the field in which she was trained. She says, “I hated economics. I thought it was moronic.” In response she has taken economics out of the lab and its tradition of modeling into the field in trying to discover the sources of poverty and the means to eradicate it.
For her work at the Poverty Lab at MIT she was awarded a MacArthur fellowship last year and this year was the winner of the Clark Medal which is awarded by the by the American Economic Association to "that American economist under the age of forty who is adjudged to have made a significant contribution to economic thought and knowledge.”
I spent my professional life teaching the virtues of randomized, control group experiments. This was in the area of experimental social psychology where this type of design is the gold standard among researchers. So when I read about her work, at first, I was far from impressed. In my naïve way, I imagined that’s what everybody was doing in studying methods of overcoming poverty.
No, this was far from the case. Instead, most programs to combat global poverty form aid agencies, build schools, distribute medical information and supplies and above all distribute enormous sums of financial assistance. They do all this without clear evidence that their programs are working. In fact, a recent article in the Times reported that nearly half of the people in the world still live on less than $2.00 a day and a fifth survive on $1 or less.
To overcome this widely recognized failure of most aid programs, Duflo believes it is necessary to conduct the same type of randomized control experiments that are common in medical research, say in testing a new drug. “I have one opinion—one should evaluate things—which is strongly held….Randomization takes the guesswork, the wizardry the technical prowess, the intuition, out of finding out whether something makes a difference."
Duflo asks, for example, “When trying to prevent very poor people from contracting malaria, is it more effective to give them protective bed nets, or to sell the nets at a low price…?" Theoretically, I would have predicted that people would be more likely to use the nets if they paid a small amount for them rather than if they were distributed freely. However, in testing two groups in Kenya a colleague of Duflo’s found that the best (likelihood of using one) price for bed nets was free.
In another study, she addressed the problem of frequent teacher absenteeism in 120 schools run by an Indian nonprofit group. The teachers in half of the groups were asked to have their photograph taken with the students at the start and end of each school day. Teachers in the other 60 groups were not photographed. Teacher pay was based on their attendance record. The photographed teachers were much more likely to be present than those in the control groups.
Duflo comments: “Who do you care about? Lazy teachers who show up sixty percent cent of the time, or the kids? O.K., I care about the kids.” Because the teachers were more often present in class, the kids were taught more and they performed much better on tests.
In spite of her success in applying randomized experiments, doing this kind of research in the field is not without its difficulties. Anyone who conducts research like this immediately confronts practical problems that undermine the degree of rigor that is possible in the lab. There is also the problem of generalizing from one field setting to another.
As a case in point, Duflo has long wanted to test the effectiveness of microfinance programs. Parker reports, “As she saw it there was little beyond anecdote to support claims that the technique had any special power to combat poverty, gender inequality, and ill health.” In spite of several years of experimental research evaluating the impact of microfinance, the findings revealed that it was no “miracle.”
Parker reports “there had been no rise in average consumption (the best way to get a sense of economic well-being) and no evidence of improvement in levels of education, health or women’s decision making.” At this point we are left with mostly anecdotal evidence that the program helps some individuals to hold a steady job, expand their business, enable them to build a house, etc. But for many Duflo had to admit, “We tried to help them. They don’t want to be helped. Too bad.”
Still she remains optimistic: There is a lot of noise in the world. And there is a lot of idiosyncrasy. But there are also regularities and phenomena. And what the data is going to be able to do –if there is enough of it—is uncover, in the mess and the noise of the world, some lines of music that actually have harmony. It’s there, somewhere.
