Examining the world through randomized control trials, Esther Duflo and Abhijit Banerjee's new book summarizes findings on what works, and what really might not, in development policy and practice.
The book's conclusions are based on two premises. First, use data to try and understand the real drivers of behaviors like school attendance and condom use. Second, assume the "poor" are making choices that are just as rational or irrational as anyone else - just that they have different information and constraints. If it isn't clear what those choices are or why they'd be made, keep looking.
The true excitement of the book bubbles from the small diamonds of insight that could dramatically affect human development outcomes (on either a micro or macro scale).
For example, in certain countries, could providing free school uniforms be a key to reducing teen pregnancy and HIV infection rates in teen girls?
Check it out. http://pooreconomics.com/
Then check out some criticism on Chris Blattman's blog here.