Book Reviews

Letting the Evidence Speak

Abhijit Banerjee and Esther Duflo

Poor Economics

A Radical Rethinking of the Way to Fight Global Poverty

Public Affairs, New York, 2011, 336 pp., $26.99 (cloth).

The most significant increase in understanding in development economics in recent years has come from the growth of randomized control trials (RCTs) to learn about the behavior of individuals in poor countries. The intellectual entrepreneurs and founders of RCTs, Abhijit Banerjee and Esther Duflo, have synthesized a large number of results and proceeded to draw inferences for policies designed to lift the poor out of poverty in this eminently readable and important book.

As their name implies, RCTs study the responses of various groups of individuals or firms, controlling for other characteristics that influence behavior, when confronted with a new set of circumstances. For example, researchers in Kenya offered mosquito nets at prices ranging from somewhat subsidized to free, to ascertain how sensitive the use of the protective nets was to price (very sensitive to price but not sensitive to users' income).

Behavior in the areas of public health, teacher school attendance, household saving and borrowing (especially microcredit), setup and expansion of small businesses, children's school enrollment, and much else are reported. In examining these and other issues, Banerjee and Duflo often start with an anecdotal report of the behavior of (or circumstances and issues confronting) a poor individual. They then draw inferences about behavior and report on the results of RCTs, which shed light on poor people's responses to different incentives, and draw conclusions as to the most effective policies for fighting poverty.

The book's analytical framework asks whether emergence from poverty is relatively linear: As people get less poor, are they more able to continue to improve their situation or does something akin to a "poverty trap" call for a "big push" to propel them across a threshold beyond which they can progress on their own? The authors view the results of RCTs as generally supporting the big push approach.

Based on the evidence accumulated through their results and inferences about behavior, Banerjee and Duflo provide a large number of policy prescriptions. For example, regarding the health sector, they conclude that "inexpensive" medical technologies are "low hanging fruit." These include "getting children immunized, deworming drugs, tetanus shots for would-be mothers, provision of vitamin B to fight against blindness, iron pills and iron-fortified flour against anemia."

The book's overall conclusion about health care is worth quoting: "The primary goal of health-care policy in poor countries should be to make it as easy as possible for the poor to obtain preventive health care . . . . Free Chlorin dispensers should be put next to water sources; parents should be rewarded for immunizing their children; children should be given free deworming medicines and nutritional supplements at school and there should be public investment in water and sanitation infrastructure."

Even if their only concern were health care, the authors' wish list is so long that it raises the question of cost. But Banerjee and Duflo have many more policy prescriptions for other aspects of social policy. They point out the high risks to which the poor are subject, and the absence of insurance options for them, and conclude "there is thus a clear role for government action. The government should pay a part of insurance premiums for the poor." The authors also endorse cash transfers to encourage staying in school, regulation of banks to require lending to "priority sectors," universal school enrollment, more infrastructure investment (especially in villages where the poor live), provision of "good jobs" to enable escape from the poverty trap, and much more.

Because good jobs tend to be in cities, the authors call not only for the creation of such jobs (although beyond requiring banks to extend credit to medium-sized firms, they do not say how), but for subsidized migration to urban areas. Given that most places have more migrants than available good jobs, the macroeconomics of this prescription is questionable.

There is no doubt that many of the programs Banerjee and Duflo advocate, if effectively implemented, would be worthwhile. But two major questions, and some...