Abhijit Banerjee: Measured Success versus Failed Miracles

An alumni of the University of Calcutta and Jawaharlal Nehru University, he received his Ph.D from Harvard in 1988. He is currently the Ford Foundation International Professor of Economics at MIT. He is co-founder of the Abdul Latif Jameel Poverty Action Lab (J-PAL).

Banerjee is a fellow of the American Academy of Arts and Sciences and has been a Guggenheim Fellow and an Alfred P. Sloan Fellow. He has authored two books and made his first documentary film, ‘The Name of the Disease’, in 2006.

Here is an entirely banal idea that I think has the potential to change the world: Take evidence seriously.

Taking evidence seriously does not mean privileging numbers over all other forms of knowledge — theories, narratives, images. Nor does it mean the kind of radical scepticism that questions everything to the point where no action is possible.

What it does mean is being very conscious of the quality of evidence, about the danger of naively interpreting the patterns that we see in the world. It means being willing to piece together and think through the whole story — how is it supposed to work, what are the possible pitfalls, what do we know that could help us avoid potential problems. It means having the humility to know that our instincts can be (and often are) wrong and the modesty to admit that we don’t know the answer to something. And it means having the patience to try to learn before jumping in. Common sense stuff.

Common sense, but not easy. Consider the recent flap about the efficacy of microcredit. Until recently, the entire public conversation about microcredit was between those who believed that microcredit was about to more or less single-handedly save the world from poverty by unleashing the entrepreneurial talents dormant among the poor, and those who saw it as the next usury. When, recently, proper evaluations of microcredit came out, it became clear that while there was a clear positive impact — the number of new businesses that got started went up from 5 percent to 7 percent in one instance, and people bought more durables for their businesses and homes — it did not have quite the transformative impact that its most vocal supporters had claimed for it. The reaction, in the public domain, for the most part, was either to deny the evidence or, entirely unfairly, to write off the whole idea of microcredit.

This cycle of hype and then disappointment, is immensely costly. When things were going well, it took attention away from other important interventions. Now, as a defensive mood takes over the microcredit community, it threatens the process of innovation within microcredit — towards better financial products, but also towards more generally helping the poor to be more effective in their own lives. Yet the results were mostly what one might have expected, given what we already knew about the poor. They need credit for their businesses, but also to fix their homes and pay hospital bills — and microcredit, far from being usury reinvented, is the cheapest source of credit that they can reliably get. On the other hand, while they do start a lot of businesses, it is mostly because they are still looking for a job that is rewarding enough or because they want their wives to have some way to add to the family pot without having to leave home for too long. As a result, the business that the average poor person sets up is unlikely to be transformative for their lives. In the circumstances we should have expected more or less what came out — definite gains, but modest, at least for the time being. Indeed this is exactly what many of the more clear-sighted people in the microcredit business, who know how things are on the ground, had predicted.

Why was this evidence never used to set our expectations for microcredit? Why couldn’t we wait till the evaluations were done before it started to dominate so much of the conversation about anti-poverty policy?

The same unwillingness to pay attention to the evidence and to try things out before they get etched in stone is everywhere in the world of social policy. The National Rural Employment Guarantee Scheme was expanded to the whole country without any proper evaluation and before any serious experimentation with its design, despite the fact that there was no lack of evidence pointing to the various dangers it faced. Many of those warnings are now reality, but now changing the scheme or experimenting with it is immensely harder. The National Rural Health Mission massively expanded spending and employment in the government health sector, despite consistent evidence that health workers do not come to work and most patients, including the very poor, have switched their allegiance to private healthcare.

Such slippage is not hard to explain. Politicians prefer not to be told that they need to wait three years before they can launch their flagship programme; journalists prefer stories about grand successes to qualified endorsements; donors want to eliminate poverty today, and being told that we don’t know how to, puts a pall on the proceedings. But in the end these are our resources that are being wasted, our hopes that are being betrayed. And all these people are meant to be our agents. We have to convince them that we want evidence rather than emotions, measured success rather than failed miracles, trial rather than error.