Poor Economics

Take the oldest idea about poverty: people are poor because they are hungry, and they are hungry because they are poor. The logic is seductive. If you don't eat enough, you can't work hard; if you can't work hard, you can't earn; if you can't earn, you can't eat. A nutrition-based poverty trap. Give a poor person more calories and, the theory says, you'd see their income climb as their strength returned, each extra bite paying for itself. It's the kind of mechanism that justifies large food-aid programmes.

When Banerjee and Duflo looked at what poor families actually did when they had a little more money, the trap sprang open in an unexpected direction. People didn't spend the windfall on more calories. They spent it on better-tasting food, on a festival, on a television, on a wedding. In the Indian state of Maharashtra, they found families that could have bought more grain choosing instead to buy things that made life less bleak. The poor, it turned out, were not starving at the margin in the way the model assumed. Most were getting enough calories to function; what they lacked was not fuel but relief from monotony.

Some of the book's most striking experiments concern things that are nearly free, enormously beneficial, and still widely refused. Intestinal worms, for instance, stunt children's growth and keep them out of school; a deworming pill costs pennies and works. In studies run in Kenya by Michael Kremer and colleagues, treating children for worms produced some of the highest returns of any education intervention measured — more extra schooling per dollar than almost anything else. And yet, when the pills were offered at a small price, uptake collapsed. Made free, uptake soared.

This flipped a comfortable assumption. Many economists had argued that people value what they pay for, that giving things away breeds waste and dependence, that a token price ensures only those who truly want the product take it. The bed-net experiments told the same story: charging even a little for a net that prevents malaria dramatically cut how many families got one, without improving how diligently the nets were used. Free was not wasteful. Free was what actually got the life-saving thing into the house.

Step back from the lentils and the bed nets and a larger argument comes into focus — one about how we are entitled to claim we know anything about poverty at all. The randomised trial is, at bottom, an epistemological stance. It says: stop trusting the plausibility of your story, and start measuring. A theory can be beautiful, internally consistent, endorsed by Nobel laureates, and simply wrong about what a mother in Udaipur will do when offered a free vaccine. The only way to find out is to try it on one group and not another and look at the difference.

Banerjee and Duflo are careful not to oversell the method, and this restraint is part of its credibility. A trial tells you that this programme worked in these villages at this time. It does not automatically tell you it will work in Lagos or Lima, or at national scale, or ten years later. The book is a mosaic of findings, not a single master key, and the authors resist the temptation — so common in popular economics — to crown one idea as the answer to everything. Poverty is not one problem. It is a thousand specific problems, each with its own texture.

The book that came out of all those coin flips won its authors, along with Michael Kremer, the Nobel in economics in 2019, cited precisely for an experimental approach to alleviating global poverty. But the prize is almost beside the point of what Poor Economics set out to do. It took a billion people usually discussed as a statistic and treated them as what they are — decision-makers working with impossibly little room, making the kinds of trade-offs anyone would make in their place. The dollar a day turned out to be full of ordinary human reasoning, not mysterious pathology.