Abhijit Banerjee and Esther Duflo: Testing What Works

How Banerjee and Duflo, with Michael Kremer, transformed development economics by using randomised experiments to test anti-poverty programmes.

In 2019, Abhijit Banerjee, Esther Duflo and Michael Kremer shared the Nobel prize “for their experimental approach to alleviating global poverty”. Banerjee, born in Mumbai and raised in Kolkata, became one of the few Nobel laureates in economics of Indian origin. Duflo became the second woman, and the youngest person, to win the prize.

The approach

Instead of debating big questions like “Does aid work?”, they broke poverty into smaller, answerable questions:

Do free bed nets increase their use?
Does extra teaching help children who are behind?
Do small incentives raise vaccination rates?

They answered them using randomised controlled trials, randomly assigning some people or villages to a programme and comparing them with others.

J-PAL

In 2003, Banerjee and Duflo co-founded the Abdul Latif Jameel Poverty Action Lab, or J-PAL, at MIT. It has a large presence in India, where many studies have been conducted, and its network has run thousands of evaluations worldwide.

Examples of findings

Teaching at the right level: grouping children by learning level rather than age, as tested with the Indian NGO Pratham, improved basic reading and maths. Several states adopted the approach.
Vaccination incentives: in Rajasthan, offering a small bag of lentils for vaccination sharply increased immunisation rates.
Microfinance: studies in Hyderabad found microloans had modest effects, less transformative than claimed.

Books

Poor Economics (2011) explains how poor people make decisions and why small details matter.
Good Economics for Hard Times (2019) addresses migration, trade and inequality.

Criticism

Critics, including Angus Deaton, argue trials may not generalise and can overlook big structural issues. Supporters reply that rigorous evidence helps governments spend money better.

The lentil incentive

Many mothers in rural Rajasthan didn't complete their children's vaccinations. When camps offered a small bag of lentils for each vaccination and plates for completing the course, full immunisation rates rose sharply at a low cost per child.

Thinking randomised trials only work in laboratories

Banerjee and Duflo ran trials in real villages, schools and clinics, testing how programmes work in practice.

Key takeaways

Banerjee, Duflo and Kremer won the 2019 Nobel for experimental approaches to poverty.
They broke poverty into small questions tested with randomised trials.
J-PAL, founded in 2003, has run many studies in India.
Findings shaped programmes like Teaching at the Right Level.