Esther Duflo: Social experiments to fight poverty

Esther Duflo: Social experiments to fight poverty | Video on TED.com (video, 17 min)
Esther Duflo, professor at MIT and the co-author of Poor Economics: A Radical Rethinking of the Way to Fight Global Poverty, argues powerfully for using scientific methods in public policy, an idea long overdue. Decision making in medicine changed from mostly guesswork in the past centuries, when the doctors used leeches to "cure" diseases, to more data driven methods, often based on randomized controlled trials, today. Amazingly, decision making in public policy is still stuck in pre-scientific stage full of opinions and angry rhetoric - in other words, in the "leeches" stage. Duflo suggests to add more science to decision making in public policy. She illustrates this on three examples from the poverty prevention in the developing world.

* Randomized, controlled trials in India showed that one can increase the rate of child immunization from 6% (no intervention, "control" group) to 17% by bringing monthly immunization camps to villages, which makes it easy for the parents to bring kids for immunization. One can increase the rate to 38%, when immunizations camps also offer 1 kg (about 2 pounds) of lentils to the parents for each immunization. Interestingly, the total costs are lower with lentils ($27 per one immunized child) than without lentils ($50 per immunized child), probably because you have to pay for the nurses anyway and lentils make the whole process more efficient.
* Bed nets are known to help reduce the rate of malaria, a disease transmitted by mosquitoes. Duflo and colleagues ran randomized, controlled trials, distributing bed nets to the populations with varying discounts ranging from 0% discount (people have to pay full price for the nets) to 100% discount (free bed nets). It turns out that, contrary to predictions of the critics of free handouts, people do not get used to free handouts. The researchers offered people to buy another bed net later, at the same price for all groups. People tend to buy bed nets at roughly the same rate whether or not they got them for free the first time or paid the full price.
* Duflo and colleagues used randomized, controlled studies to show that some methods of improving school education work much better than others. For example, investing $100 in either hiring extra teachers, buying school meals, buying school uniforms, or providing scholarships yields as a result 1-3 years of extra education years. However, investing $100 in deworming (curing worms in children who live in areas with parasitic worm infections) yields 29 extra years of education. Investing $100 in telling people about the benefits of education (providing them information on returns from extra schooling) yields 40 extra years. These are very counterintuitive results. And the data are strong enough to make a real difference: Deworm the World organization dewormed 20 million school-age children in 2009. Gradually, scientific thinking begins to penetrate the public policy decision making process - an idea long overdue.