Banerjee, Duflo and Kremer together have launched a movement within development economics that seeks to ensure that clear, unambiguous answers can be found to the question of whether a particular policy intervention is effective. This is extremely relevant when it comes to framing policy in low- and middle-income countries, where state capacity is quite limited and it is particularly necessary to be able to prioritise more effective policies over less. Duflo and Banerjee’s book, Poor Economics, is an argument for such evidence-based policy, particularly in the Indian context.
While Duflo and Banerjee — together with another Indian-origin economist, Sendhil Mullainathan — set up the influential Poverty Action Lab in MIT in the early 2000s, it was Michael Kremer who, with his studies of schools in Kenya, first popularised RCTs within the economics profession. The idea behind such policy trials is to treat an intervention similarly to, for example, a medical treatment or new drug that is being scrutinised.
If a particular group can be randomly divided into a control set and an experimental set, and then the latter is the one exposed to the policy treatment, then by comparing the outcomes on the control and experimental group the analyst can be more certain that other factors are not confounding the results. For example, if half the schools in a district are randomly chosen for a special trial — for example, biometric attendance records for teachers — and the other half are not, then comparing school outcomes afterwards makes a pretty unarguable case that such records do (or do not) help improve learning outcomes.
The effect of such rigour on policy analysis is considerable, and as a consequence the RCT movement has almost taken over the development economics field. This change is not without its critics. Some argue that by reducing problems to manageable sizes for an RCT, the development economics profession is avoiding the really big, hard questions about whether systemic change is needed for growth and development. But for countries like India where resources, particularly in the least developed states and areas, are hard to come by, the results of RCTs are vital input into decisions. They ensure that vital years and budgets are not wasted.
For Duflo and Banerjee, an important part of their work has been ensuring that the agency of the “beneficiaries” — usually, in developing countries like India, poorer individuals — is put at the centre of any policy design. This is a crucial way in which experimental results are often better than large scale data-based inference. The latter is run through with assumptions about individual and group behaviour that may be how economists assume the poor act, not how they actually do. For example, Duflo and Banerjee often point out that even income-constrained people who may not appear to be getting enough to eat will, if they are given more money, spend it on televisions or mobile phones rather than better-quality or more food. This is, they argue, not in fact that surprising: it merely shows that the desire for an escape from boredom is as much of a motivator as hunger. For well-off economists who have never truly been either bored or hungry, figuring out how people who are regularly both may be difficult in the abstract; but an RCT examining how extra money is spent helps solve that conundrum.
A related insight, therefore, has been on the limited propensity towards entrepreneurship shown by the poorest individuals. A common refrain about India has been that its entrepreneurial culture is unmatched, and that poverty has only enhanced it. In fact, these economists have shown that, for example, lending from micro finance may have no real link to return-bearing investment decisions. It does, however, increase welfare appreciably by allowing people to avoid big ups and downs in their consumption. There are important implications for policy today given many hope that the “personal sector” and micro-sized enterprises will, thanks to easier capital availability through government schemes, invest and grow the economy. Among the other insights from their research is that poor people in fact spend a large amount of money on health, even at the cost of other purchases; that individuals find costly ways to discipline themselves into saving and protecting their money from other family members; that Indian primary education tries to do too much too early, and winds up doing little.
Banerjee and Duflo — married to each other — are frequent visitors to India, and the former has written extensively on the Indian economy’s current problems.