Last week, the Nobel Prize in Economic Sciences was awarded to three worthy development economists, including partners in life and work Abhijit Banerjee and Esther Duflo. Their study of poverty alleviation, with heavy reliance on randomized control trials, has helped to bring more scientific discipline to a cause historically clouded by politics, dogma, and good intentions.
As deserving as this recognition is, however, even Banerjee has cautioned against overgeneralizing their findings. The randomized control trial fieldwork he and Duflo led in India helped shed light on how to get teachers to show up for school, a key ingredient for achieving better educational outcomes for low-income populations.
But when they tried the same approach with nurses, it failed. What changed? Not their methods. The context changed, and with it the usefulness of the randomized control trial results.
This is where the debate about randomized control trial evidence in development economics heats up. Given the complexity of any particular environment, scientific results from one place may not apply elsewhere. Earlier this year, the University of Oxford’s Lant Pritchett even went so far as to call it “madness” to pin our development hopes on randomized control trial research arguing that “none of the common domains of RCTs are plausibly important determinants of the level of income or of growth.” In short, randomized control trials might give us clues about what works relatively well under specific conditions, but they can’t provide a big, universal solution to poverty that we can implement at scale.
This cuts against our collective impulse to “do something” to end poverty once and for all, but it’s a healthy reminder that in our zeal to make the world a better place we must recognize the limitations of our own knowledge, as well as the proper limitations of our own role as outsiders when it comes to intervening in other people’s lives.
How to account for those limitations is a growing concern among international development circles. In early October, Population Works Africa and #BlackWomenInDev cohosted a forum at the Center for Global Development in Washington, questioning the historical role foreign outsiders play in leading development projects on behalf of local populations. Repurposing well-known terminology concerning appropriate sexual behavior, presenters discussed whether aid organizations truly get “consent” from local communities when they bring interventions to bear on local issues.
Caring about consent is about more than wanting to respect local people and culture. It’s also about recognizing that outsiders do not, and cannot, know enough to make important decisions on behalf of locals, no matter their good intentions, resources, and technical expertise. An unconscientious faith in the universal applicability of randomized control trial results threatens to galvanize, not reform, our historical habit of aid without consent.
Perhaps taking context and consent seriously means decentralizing the power and leadership structure of our government-led economic development efforts. Recent headlines about President Trump’s call to Ukrainian President Volodymyr Zelensky make clear the fundamental power dynamics and politics inherent in the process of foreign aid.
As an alternative, there are plenty of local organizations in developing countries working on questions of poverty and prosperity that we can support directly. They understand better than outsiders what their priorities are and, from within their local context, they can best consider whether and how to apply research results from Nobel Prize winners such as Banerjee and Duflo.
Matt Warner is president of Atlas Network and editor of the forthcoming Poverty & Freedom: Case Studies on Global Economic Development, available on Amazon.com Nov.1.
