No one likes failure. That’s especially true when failure leads to terrible consequences: a family not being fed, constant power outages, widespread death from a pandemic.
The financial and human consequences of failure are even bigger for decisions at the level of a region or whole country. Growing up in Bangladesh, Ahmed Mushfiq Mobarak, an economics professor at Yale, witnessed how the actions of the government and NGOs could impact lives for better or worse. Untangling how to make these programs work for as many people as possible is what Mobarak has done every day for the last 20 years.
Mobarak has worked in areas from air pollution to sanitation, but perhaps the most famous example from his research, illustrating the importance of scaling up such work, is seasonal migration. In widely cited experiments, Mobarak’s team found that giving low-income people in Bangladesh some money to temporarily move to the city to work during the “lean season” increased their family incomes.
This experiment was part of an effort to make the international development field more evidence-based. Through randomized controlled trials (RCTs), researchers can test the effectiveness of programs — in this case, giving people money to temporarily migrate to find a job — with some randomly selected people, and comparing the outcomes of the people who got the program with those who didn’t.
The findings sounded great, but the story doesn’t end there. The migration paper inspired the nonprofit initiative No Lean Season, which gave a travel subsidy to people in Bangladesh who wanted to seasonally migrate for work. But when the program was scaled to millions of people, it had no effect. Based on the evidence at scale, No Lean Season stopped in 2019.
That’s part of the learning process for research. And while finding a program doesn’t work at scale might seem like a setback, being able to test whether policies work at different levels and adjust accordingly is a very good thing and something more governments and policy-making organizations should do.
Lessons on scale are vital to consider for policy based on randomized experiments, which is a focus of Mobarak’s work at the Yale Research Initiative on Innovation and Scale (Y-RISE).
Y-RISE works in areas as varied as microfinance, refugee integration, and childhood development. Bringing programs in areas like health and agriculture to scale carries with it lots of challenges — changes in government behavior, effects on macroeconomic growth, implementation difficulties — but if done well, it can improve the lives of millions of vulnerable people.
Y-RISE aims to do just that.
“The world becomes vastly more complicated when a program is scaled up,” he said at the launch of the initiative in 2018. A program that might work with a couple of villages might not work on a national level; a program that works in Bangladesh is not guaranteed to work in Kenya. And when a program with promising results is tested multiple times and scaled in the real world, we can gain a lot of insight into what parts might work in the future and how to reform a program — or stop funding it altogether.
Another area necessary for large-scale policy that Mobarak studies is “spillover effects” — the positive or negative effects of programs on people who don’t receive them. In a study on latrine subsidies, Mobarak and a team found latrine usage — which decreases the spread of pathogens — increased not only among subsidy recipients but also their neighbors, and that these effects were higher among poorer people in denser neighborhoods. While this hasn’t yet been tested at a larger scale, information about who programs will have the largest impact on is part of the learning process governments can use when deciding who to reach with limited resources.
In keeping with his work on scale and behavior, Mobarak’s current projects include ongoing research into pandemic response in low- and middle-income countries, and how to promote technology adoption. The world is ever-changing, and his research continues to provide insight on the varied policies that allow people to thrive and catalyze economic growth.
“If we better understand what constraints people are under,” he said in an interview with Yale, “we’re able to design policies to address those constraints.”