This year’s Nobel Prize in economics was awarded to three scholars who revolutionized the effort to end global poverty: Abhijit Banerjee and Esther Duflo of MIT and Michael Kremer of Harvard are essentially credited with applying the scientific method to an enterprise that, until recently, was largely based on gut instincts.
But how does their approach work in practice? Why is it considered so ground-breaking? And how much has it actually changed life for the world’s poorest? Here’s an explainer.
Don’t Assume – Test!
For decades, programs and policies to help the poor have largely been designed around what seemed like reasonable assumptions:
Kids in poor areas can’t afford new textbooks — so surely giving them free ones will improve their test scores.
Impoverished women have a hard time finding jobs – so surely giving them a “micro-loan” of a few hundred dollars to start a small business will boost their incomes.
But Kremer, beginning in the 1990s, and joined by Banerjee and Duflo some years later, helped lead a growing movement of economists who were concerned by what they saw as a key flaw in that approach: Assumptions are just that. Without evidence you can’t be sure they are true. After all, just because peoples’ lives improved after they were given a particular form of aid doesn’t mean it was the aid that made the difference. Some totally unrelated factor could have been at work – say, a concurrent rise in overall economic growth or a different aid program distributed at the same time. Also, the money available for aid programs is finite. So even if a particular intervention does help, it might not be the best use of resources if there’s a different intervention that would deliver bigger results. The only way to find out is to put your assumptions to the test.
So how do you test?
The Nobel winners helped develop and popularize the application to poverty research of a method hitherto more common to “harder” sciences like biology: the randomized controlled trial – or RCT. In a nutshell, to see if a particular aid program works for a given population, you compare its impact to the results for an otherwise identical “control” group of people to whom you did not give the aid. Poverty researchers who focus on the value of RCTs are so committed to the concept that their colleagues have dubbed them “randomistas.”
What have the randomistas revealed?
Some sacred cows have certainly been gored. Returning to the example of microloans for would-be women entrepreneurs who can’t qualify for financing from a bank because they are too poor: A raft of RCTs (including some by members of this year’s Nobel team) have disproved the once popular notion that microloans can substantially boost the incomes of poor people. (As this article explains, it’s not that microloans are never useful for poor people. They can be. It’s just that the evidence suggests their broader impact is muted.)
Similarly, RCTs by members of the Nobel team and others have shown that when it comes to education, there’s a limit to the effect of intuitively reasonable measures such as reducing the student-teacher ratio, providing free lunches and … you guessed it, distributing textbooks. Instead, some of the biggest boosts to student educational outcomes come from less obvious fixes such as providing kids with cheap de-worming pills that dramatically reduce the number of days they have to miss school due to tummy trouble. Also in the most impactful category: targeting assistance to the lowest performing students and making teacher contracts contingent on their students’ performance.
The randomistas’ reach …
Over the last two decades there’s been an explosion in the use of RCTs. Some are aimed at uncovering the underlying factors that keep people trapped in poverty. Others are effectively “impact evaluations” – as they’re called in the movement — of specific policies and programs for the poor.
Many of these studies originated from a network of more than 180 researchers around the world that Banerjee and Duflo helped found at MIT in 2003. It’s called the Abdul Latif Jameel Poverty Action Lab, or J-PAL. Less than 15 years ago it was running about 70 RCTs worldwide. Today the total completed or currently has underway is just under 1,000. Colleagues of the Nobel team have also set up a number of nonprofits with similar missions. This includes Innovations for Poverty Action, which has conducted more than 830 impact evaluations since it was founded in 2002.
The randomista movement has also inspired various governments to use RCTs to inform and tweak the design of major programs. This includes the “Teaching at the right level” (TaRL) – an approach pioneered by an Indian NGO that has been tested with the help of J-PAL researchers. It has been applied by local governments in India and Africa to improve elementary school instruction for millions of children.
Similarly, a team from J-PAL has worked with Indonesia’s government to test and then roll out measures to curb corruption in a rice distribution and subsidy program. (As this NPR article describes, the program sent out a special identity card to more than 15 million Indonesians to let them know how much rice they were entitled to and at what subsidized price, so that unscrupulous officials couldn’t get away with keeping some of the rice for themselvescheating them out of their full share.)
… and the limits of their influence
Even as they tout the successes, many champions of this evidence-based approach to fighting poverty say it’s not being used widely enough. For instance, at the World Bank and the U.S. Agency for International Development only a fraction of projects are subject to impact evaluations, according to an analysis by the think tank Center for Global Development. The group, based in Washington, D.C., did an exhaustive review to identify large-scale health programs that made a big impact through 2016. Of about 250, they found that only 50 used rigorous methods to establish the attributable impact. Similarly, some staff at the U.S. Agency for International Development have been pushing an experimental effort to use RCTs to test whether any given aid program is more impactful than simply handing out an equivalent amount in cash. But that effort has also run into resistance, at least partly say former staffers there, because there’s an inherent bias against trusting poor people to spend cash wisely – regardless of what the evidence shows.
Meanwhile, there’s also debate among economists as to whether the focus on RCTs obscures deeper, more fundamental drivers of poverty – such as systemic inequality – that must be addressed if there’s any hope of truly improving lives on a mass scale.
What’s the impact of this Nobel Prize?
Of course it’s too soon to tell. Duflo says she and her fellow recipients plan to figure out how to make use of the $915,000 prize to further their research. But perhaps the greatest impact will be the attention this brings to the approach they’ve pioneered. “It’s a prize not just for us, but for the whole movement,” said Banerjee at a joint news conference at MIT.