Small ideas, big solutions
A new look at how to reduce poverty, one experiment at a time
A GOOD number of the world’s most challenging problems might be a lot easier to solve than anyone imagined. That’s what I couldn’t help thinking as I left a recent meeting with a young French economist.
Her name is Esther Duflo. Her focus is the planet’s truly poor, the more than 865 million souls who scrape by on 99 cents a day or less. But it’s her method — using rigorous experiments to solve social programs — that could be truly transformational.
For decades, international development has been a graveyard of grand schemes — massive dam projects, micro-financing, incandescent celebrities. But when Duflo and I sat drinking tea at the farmhouse table in her Beacon Hill kitchen, what she wanted to talk about was worms.
Worms, a.k.a. intestinal parasites, are a problem across the Third World. They are particularly damaging to children, whose growing bodies and brains they rob of nutrients. Harvard economist Michael Kremer wanted to know the true impact of worms. He hooked up with a Kenyan program which gave de-worming medicine to kids, and then followed them into adulthood.
He found, for example, that children who received pills for two years earned 20 percent more each year as adults than those who were treated for just 12 months. A treatment that costs $1.36 yields a lifetime benefit of $3,269. It is stunning that something so small, so easily scaled up, can have such a dramatic effect.
But what’s really interesting is that there is an answer at all. Most development efforts end without a precise measure of effectiveness. What Kremer did was an experiment. He compared the outcomes of kids, showing a particular treatment (cheap prescriptions) led to a specific change (dramatically increased wages).
Experimentation is an idea so simple — do something, compare it with doing nothing — that a child can understand it. Yet it is the engine that has driven the scientific revolution. Experiments have revealed our genetic codes and the nature of matter. Millions of experiments over centuries have built modern medicine: from “bloodletting does not appear to help’’ to “this anticancer drug will only affect tumors with particular genetic profiles.’’
The power of the experiment is what inspired Duflo and Abhijit V. Banerjee — both economists at MIT — to form the Abdul Latif Jameel Poverty Action Lab (J-PAL) in 2003. Banerjee, Duflo, and their colleagues are sometimes referred to as the “randomistas’’ — a play on the “randomized controlled trials’’ which they use to test anti-poverty ideas.
The randomistas have become a genuine movement. Banerjee and Duflo have a fascinating new book, “Poor Economics,’’ which serves as a kind of randomista manifesto. Another member of the J-PAL clan, Yale’s Dean Karlan, has a similar book, “More Than Good Intentions.’’ Harvard recently started an innovative effort, Ideas42, to solve social problems at home as well as abroad.
The randomista findings are aggressively particular: An extra year of schooling for an Indonesian child yields, on average, a wage boost of 8 percent. But each experiment ends with real facts we can all agree on, not loose statements of faith, hope, or political ideology. “They are,’’ said Duflo, “actionable.’’
Why aren’t our own governments running experiments all the time? Think of all the money the federal government spends on social programs, and how little is known about their effectiveness. Think of the new ideas out there, waiting to be discovered.
Imagine a Social Innovation program, paid for with some very tiny percentage of the total budget, which funded small experiments in social improvement — testing an idea for getting elementary kids to miss less school, say, or different methods to help people into new jobs? (The French government has established such a fund.) Hundreds of experiments would run every year. Some experiments would challenge the orthodoxies of our ruling parties. Others would suggest shockingly effective alternatives.
Government could then enter a virtuous cycle of learning. This is, after all, the story of modern medicine. Sometimes big problems don’t demand big solutions, they demand many small ones.