Of the many distinctive attributes possessed by the three most recent Nobel laureates in economics—Abhijit Banerjee, Esther Duflo, and Michael Kremer—one of the most noteworthy is their relative youth. Most Nobel economics laureates receive the award on the back end of their careers—the 2007 winner, Leonid Hurwicz was 90 and the 2012 winner, Lloyd S. Shapley 89—and are often recognized for contributions made many decades ago. Duflo, however, is only 47 years old, which makes her the youngest economics prize laureate ever. (Banerjee, who I interviewed for my previous article, is 58 and Kremer is 55.) This means that all three recipients, and Duflo in particular, are not only currently active in their research and policy outreach, but also likely to have twenty or thirty more years to leverage the prize and impact the future.
So now the world—especially those of us who work in the social sector—look to this Nobel triumvirate to get a sense of how they envision that future, and how they will leverage the prize to shape it. Indeed, they were awarded the prize in part for anticipating and shaping the future by applying randomized evaluations to poverty alleviation in developing countries beginning two decades ago, a time when no one else was doing this.
Banerjee and Duflo are the co-founders, and Kremer a long-time affiliate, of MIT-based Jameel Poverty Action Lab (J-PAL), a social sector organization which aims to reduce poverty by ensuring that policy is informed by scientific evidence. J-PAL offices worldwide have implemented much of their field research and been part of many efforts to scale the programs that were effective, reaching hundreds of millions of people. I am fortunate to have a strong working relationship with Banerjee and Duflo, and also with Iqbal Dhaliwal, who is the global executive director of J-PAL. Dhaliwal, indeed, ranks among the most thoughtful and effective nonprofit leaders I have encountered and his leadership includes the self-appointed task of developing the vision and strategy for “J-PAL 2.0.” Adapting to forces affecting our future is on my own radar overseeing the philanthropy of Bob and Dottie King. Bob King’s success as an investor was informed by his analysis of future change forces; we seek to apply a forward-thinking lens to our philanthropic decisions. Over the past few weeks I have had the opportunity to ask probing questions as to what Banerjee, Dhaliwal and Duflo see as the next frontier—a topic the prize has spurred them to consider, too. In fact, Dhaliwal told me, on the very morning of the day Banerjee and Duflo were to give their Nobel Prize lectures, the three had breakfast together in Stockholm and agreed to start actively considering what more they should be doing in light of the prize.
Two of the themes that consistently bubbled to the surface were climate change and leveraging administrative (admin) data for social impact. Or, as Dhaliwal put it, “Perhaps the biggest challenge of our generation, and probably the next few, is climate change. And one of the biggest opportunities we have to improve the outcomes of social and development programs is to use large digital data to inform decision making and test innovations.”
It is very telling that Duflo, in the Nobel Banquet speech she made on behalf of all three laureates, deviated from her published transcript to conclude with the exhortation, “There is a lot of work left to do against poverty and to fight climate change—and we better get back to it.”
While J-PAL’s focus is on global poverty, climate change is squarely on its radar because the world’s poorest are most vulnerable to, and will be disproportionately affected by, rising temperatures, warming oceans, rising sea levels, and an increasing number of extreme weather events. In fact, these have already begun to jeopardize food security and agricultural livelihoods, displace coastal communities, and undermine economic development and advancements in human well-being around the world. Because of this, said Dhaliwal, “There are a few key themes that will define our research in the future and climate change is right there among the top.” He explained:
I believe this is the biggest challenge we face and we are wholly unprepared for it. What is the point of focusing only on quality of health care in hospitals in Sierra Leone or coastal Bangladesh if they are going to be inaccessible due to rising sea levels in just a few decades from now? Or of improving learning outcomes for children in India if the classrooms are going to be baking at 50 Celsius for a quarter of the school year, which makes it impossible for teachers and kids to go to school?
Dhaliwal further noted that this academic year alone in Delhi, schools have been closed for over two weeks due to some combination of severe air pollution or extreme temperatures; on several days the air quality index was 472 out of a maximum 500, nine times the level recommended by the World Health Organization. Losing so many days of school will have considerable impact on education—indeed, it was probably enough to effectively wipe out the annual impact of costly educational interventions. “So, we invest millions of dollars into programs to improve learning outcomes that can be measured as the equivalent of a few extra days of schooling,” said Dhaliwal, “and then the air quality plunges and in one shot you’ve basically gone negative on the benefits…of programs to improve educational outcomes.” Climate change thus presents a huge problem to everyone in the field of development economics. “If we don’t pay attention to climate change,” he added, “just about everything else that we are trying to do is going to be quite meaningless pretty quickly because we will be playing in the little league while the major league is somewhere else.”
Since, according to the Intergovernmental Panel on Climate Change, “scientific evidence for warming of the climate system is unequivocal” and the impact of this will be greatest on the world’s poorest, J-PAL aspires to play a leading role in the movement to apply randomized evaluation to the realm of climate change so that leaders can make data driven decisions. This is of urgent importance because, Dhaliwal said, most of the opinion is essentially divided into two camps who agree on almost nothing when it comes to the fundamentals of climate change—but nonetheless share one fatal flaw. He explained:
There are people who believe climate change is happening and it’s manmade and we can and should do something about it right now, and a different set of people who believe the whole thing is either greatly exaggerated, or part of a natural trend that cannot be slowed. And very rarely does the conversation advance to consider scientific evidence about what strategies do or do not mitigate climate change, slow it down, or help vulnerable communities adapt to it.
This refusal to use evidence is unfortunate, unnecessary, and increasingly unconscionable because, he continued, “The tools are there; we now have some seminal RCTs which have shown that you can use the tools of randomized evaluation to understand the impact of programs in environment and climate change.”
Although there can be inherent tension between preserving environmental quality and achieving poverty alleviation through rapid growth, J-PAL has identified numerous strategies to manage this. These strategies can be grouped within four pillars: carbon reduction (mitigation); adaptation, including support to households and farmers in coping with climate change; access to clean energy; and pollution reduction. Duflo, for her part, warned that dealing with climate change “will require a change in behavior, particularly in the rich countries” that were big consumers of goods and energy. Though some argued it would not be necessary to reduce consumption if it was fueled by renewable energy, Duflo begged to differ. “It would be great if that were the case but I don’t think we can count on it necessarily,” Duflo said at a Swedish Academy press conference.
Leveraging digital data for social impact
Increasing the use of digital admin data by governments, non-profits, and private firms to enable evidence-informed decision-making presents the opportunity about which Banerjee told me he is most excited right now because it is “the new frontier.” This admin data includes any information that governments, large NGOs, and businesses are already collecting on program participants and outcomes—and there is more (and better quality) data being gathered than ever before. However, many stakeholders have previously been either unaware of this data’s potential or unable to use it to inform anti-poverty programs, as so doing requires advanced data analysis skills, highly technical guidelines, and partnerships for collaborative thinking to design, run, measure, and scale interventions. Examples include geo-coded, time-stamped data collected on mobile devices and transmitted in real time, like scans of a patient’s unique ID card and QR code on vaccines administered, or household level data on electricity or water usage available in real time from smart meters.
Dhaliwal echoed Banerjee’s excitement, explaining that the relatively recent existence of so much admin data collected digitally at its source with numerous checks for reliability and accuracy (including GPS coordinates and biometrics), available in real time and refreshed continuously, presents a tremendous opportunity. He explained:
We are really at the cusp of a new wave of experiments. In the last 15 years, we managed to do 900 RCTs where each one of these involved going into the field with an army of fieldworkers collecting this data manually and investing significant time and effort in back checks and data validation. And still often have a sample size limited by budgets rather than the full needs of the underlying question. If, instead, we had access to outcomes data that is regularly collected and accurate, we could have probably tested 9,000 interventions instead of 900.
To successfully use this data will require two things. First, J-PAL believes it necessary to create what Dhaliwal called “rules of engagement,” or a standard approach to accessing data from different sources. “Just about every nation, state, department or organization has its own set of rules for access that range from no rules all the way to no-access-ever,” he said. “Navigating this maze is a huge deterrent for greater data-driven decision making.” J-PAL is currently using a Sloan Foundation grant to create a best practices Handbook of Using Administrative Data for Research and Evidence-Based Policy intended to help researchers and data providers work together. Building on those standards, J-PAL is launching a new initiative called Innovations in Data and Experiments for Action, or IDEA.
“We want to create IDEA Labs in the field in half a dozen strategically chosen countries,” he explained. “We see their role as working with local governments and large NGOs like Pratham, which operate at scale and are collecting a lot of data, to try and map what data they have to the decisions they make.” IDEA would support governments, companies, and non-profit organizations—data providers—willing to make their administrative data accessible. It would analyze this data to improve decision-making and then partner with researchers to design innovative programs, evaluate their impact through randomized experiments, and scale up those programs that proved successful. IDEA would also help data providers build the capacity to do this on their own, with their data, and create enduring institutional partnerships. Duflo explained:
A criticism that we sometimes hear of randomized evaluations is that they can be expensive and time consuming. But often what drives cost and time is not the experiment itself, but the field surveys to collect good quality data on baseline characteristics of the treatment and control groups, and to track outcomes of interest. Using available credible admin data can allow us to run cheaper and faster experiments, greatly increasing the potential for evidence-informed policy.
There is an exciting link between these two themes of climate change and admin data. Climate change is a pressing problem right now, but to conduct large-scale, high-quality randomized evaluations is expensive and can take a long time. However, with the use of different data sources—from kwH of energy produced in a given area on a given day to air quality sensors in different areas—there will be a new ability to run a lot of mini, real-time experiments. For example, say one wanted to inform people about their own energy efficiency, or energy consumption, and to evaluate if this then encouraged heavy users to consume less. It would be relatively straightforward to then look at the data from power utilities to track whether usage declined in the treatment households that received the information campaigns, without needing to undertake expensive and time-consuming household surveys. This underscores the exciting power for the future of admin data—real-time data that does not have to be manually collected by surveyors (which is expensive and slow) because it already exists.
The role of leadership
Essential to any vision for the future is the ability to marshal the leadership needed to effect real change in that future. J-PAL is distinctive in its leadership abilities and the experience of its leaders holds lessons that may be useful for others. Among these, said Dhaliwal, is the need “to find people who complement you rather than mirror you or say the same things as you.” Dhaliwal’s own conviction of the importance of this was informed by his six years spent working in India’s civil service entrenched in a government bureaucracy that placed great value on seniority and often functioned as an echo chamber. The obligation to freely state one’s views and to dissent, he said, was essential. He continued:
J-PAL would not be what J-PAL is if Abhijit [Banerjee] and Esther [Duflo] to begin with did not have, and express, strong opinions. And that’s also true for the colleagues I work with. Everything goes back to the same team—this team of countervailing ideas… My colleagues who have contributed the most to the organization and been the most successful are the ones who know the value added to the organization, and to me, of freely expressing their opinion and ideas even if they were not fully sure about these. And also know when they have been heard and it has been a good discussion and, based on the merits of the point, they might now have to put their energy and effort in implementing the opposite viewpoint that won the day.
It is also essential, he said, for a leader “to create a space for people with diverse backgrounds, ideas and style of working to do what they do best. This diversity has helped us extend the boundaries of what was possible and to think differently in a way I would not have earlier. So, this is not just embracing ‘diversity’ in background, but also diversity in ideas and styles.”
Dhaliwal further noted that he had learned a great deal from Duflo, who he said has “really great management instincts.” These learnings include the wisdom of hiring talented people when you have the opportunity to do so, even if it is not clear exactly how you will use them. “When you find someone who is really talented and who is interested in your mission,” he said, “I think it is in fact the responsibility of the CEO to think seriously about how they can create a position for that person that would be mutually rewarding. This is especially true at senior management level where good talent is so scarce and one needs to be opportunistic.”
Duflo echoed the sentiment and added:
Since we got the news of the Nobel prize, Abhijit, Michael, and I have said many times that this award is not just for us, but for the movement. And J-PAL—with its incredibly talented affiliated researchers, outstanding leadership, amazingly committed staff worldwide, and visionary implementation and funding partners—has been at the center of it all. We have a lot of work to do, but thanks to J-PAL, I feel very confident that we will be able to make even bigger strides in the fight against global poverty and climate change using the power of scientific evidence.
Banner image courtesy of Alexander Mahmoud/Nobel Media
Originally published in Forbes