A Conversation With Abhijit Banerjee

Wednesday, March 6, 2019
Abhijit Banerjee

Ford Foundation International Professor of Economics and Codirector and Cofounder, Abdul Latif Jameel Poverty Action Lab, Massachusetts Institute of Technology

Isobel Coleman

Chief Operating Officer, GiveDirectly; Former Representative to the United Nations for Management, Reform, and Special Political Affairs, U.S. Department of State

Although the global rate of extreme poverty is at a historic low, the pace of poverty reduction is slowing and the World Bank estimates that more than 700 million people still live on less than $1.90 a day. The 2019 Robert B. Menschel Economics Symposium discusses the ways behavioral economics can inform development policy to create effective solutions to poverty at the international, national, and local levels.

COLEMAN: Hello, everyone. Good afternoon. I’m Isobel Coleman. I’m the chief operating officer of GiveDirectly, and I will be presiding today.

It’s my great pleasure to host this event. It is the Robert Menschel Economics Symposium. And as you can probably guess, unfortunately, Esther Duflo is not able to be with us here today. She had an emergency at home and really does send her regrets. And for me it’s a real disappointment. I have been citing her work for fifteen years now on the incredible research she’s done women’s economic empowerment.

But it’s fantastic that Abhijit Banerjee is here with us. He is the Ford Foundation International Professor of Economics at MIT, and he’s also the co-founder of JPAL, the Jameel Poverty Action Lab, which has done really, really innovative work over the years since 2003, when it was founded. Abhijit was educated in Calcutta, received his Ph.D. from Harvard, and has been off and running doing really, really interesting work in economics—development economics ever since. So we’re in for a treat to have this terrific discussion today.

And also, this symposium is presented by the Maurice Greenberg Center for Geoeconomics and is made possible through the generous support of Robert Menschel. So thank you all for being here.

So I thought we might start with just what is behavioral economics. It’s a little bit of the zeitgeist these days. Lots of people talk about it. Richard Thaler has made it quite in the mainstream, as has Daniel Kahneman with his book Thinking, Slow and Fast (sic; Fast and Slow). It’s sold seven million copies, I think, and going. And Michael Lewis has even written a book about behavioral economics.

So a lot of people are talking about it. A lot of people are talking about nudges and things like this in a policy perspective. But why should we care about behavioral economics? First of all, what is it? How do you think about it? And why should we care about it?

BANERJEE: That’s a small question, no. (Laughter.) So I think that—in a sense I think maybe if you asked me what it meant ten years ago I would have said it meant economists have a quite specific definition of what is rational behavior. And this is everything that we have learned from psychology about the ways in which people are not rational being brought into the field.

I think that’s probably no longer a good working definition. And the reason why it’s no longer a good working definition is that, in fact, I think what we have realized over the last, I don’t know, twenty years or so is that it’s not just that economics get the, you know, specific psychological models wrong sometimes. They don’t—they don’t even model the—don’t even apply their own models with any degree of transparency. So, in other words, they are often kind of—there are prejudices that are just prejudices that are built into the way economists kind of react to facts. And then you say that, well, why isn’t that consistent with “rationality,” quote/unquote, and the answer is usually it is but we don’t like that rationality.

Let me give you an example, an example that I’ve written about. So if you sort of take the—kind of the big-picture talk in economics, there’s this idea that, you know, tests are sovereign. People have their own preferences. They do what they want. You should not question people’s tests. De gustibus non est disputandum is how it was put by I think Gary Becker, who won the Nobel Prize in economics many years ago. And yet, if you see somebody who doesn’t have enough food and owns a TV, I think the first reaction of an economist is that that’s irrational.

Now, and one of the things that I think we’ve been—where we’ve been pushing back and sort of learning a lot is in questioning that sense of some sort of morally imposed rationality, if you like. It’s not actually anything to do with rationality; it’s about our sense of what these people should be doing.

And why does that matter? Well, let me say two things. One is why do you think, in fact, these people were buying TVs? For a very good reason. It wasn’t that they were crazy. It was because, you know, they realized that as poor people—and this particular example I remember also Morocco, but you could—it could be anywhere in the world. Lots of poor people own TVs. The reason why they want to own a TV is because life if you are a poor villager is pretty boring. There’s not much to do. Once you’ve gone to the tea shop, you go—I mean, I spent lots of evenings when doing field work in a tea shop in a village, and what happens is people come in and say one word, they sit down, they stare at their cup of tea, and then they leave. And that’s—it’s not extremely exciting. So when this guy explained to us that the reason why he owned a TV was TV is more important than food, he was telling his felt truth. His felt truth was that a TV was more important because his life was very boring otherwise. And boring—it’s OK to be bored for three weeks, but it’s not OK to contemplate being bored for the rest of your life. So that’s the sense in which what I think this person was telling us was that, you know, I’m making the right choice. You may not think it’s the right choice, but that’s because you don’t live my life.

Now, why is that important? You—go ahead.

COLEMAN: No, go ahead.

BANERJEE: That was—that was the other half of your question, why do we.

Well, because that then frames the way we make policy. We make policy about, you know, people—we should not give people cash, because if we gave people cash they will buy TVs and in fact they should buy food. That particular idea is a judgment about how people not only actually live their life, but how they should live their life, about both. And then in both cases I think we are imposing our suppositions, which are not necessarily true.

COLEMAN: So let’s stick with that example for a moment. And full disclosure, GiveDirectly, the organization that I work with, does unconditional cash transfers, and Abhijit is one of the academic leaders of the impact evaluation of programs that we’re doing. But stick with the cash example, because you’ve done a lot of research—in fact, you have a paper about the myth of people not working when you give them cash. So talk a little bit about that, and how you’ve done that research, and what you found.

BANERJEE: So we’re part of an organization called JPAL. And what JPAL does—it’s the Abdul Latif Jameel Poverty Action Lab—what we do is we do randomized controlled trials that is a large-scale equivalence of medical trials of social programs. So we try to see what is the impact of particular interventions on lives of people. We work in, I think, eighty-something countries, including the U.S. And our sort of tool—you know, our basic tool is this idea of what I’m going to call our RCT, a randomized controlled trial like a medical trial. Basically, you randomly give some people a particular intervention, you give others either some other intervention or nothing, and you see what it does.

One particular randomized controlled trial that has been quite popular is the—popular meaning there’s been lots of examples of it—is this—of some version of we give people some cash without working, so increase their income; what does that do to various things they do, including their labor supply. Do they stop working? Basically, if you get welfare, do you get lazy? And I think the basic fact that comes out of these is that there is no evidence—and I say this without any caveats—no evidence that people become lazy when you give them cash. They sometimes change what they do. They may very well stop doing something that’s extraordinarily painful or boring and do something else with it. This gives them options. They often invest the cash in exploring opportunities of the—but there is no evidence that they—if you take the total leisure that they enjoy, that that goes up. In fact, I’ve just been working on a study in Ghana where we show clearly that any intervention of this class basically reduces people’s, you know, leisure. They spend less time, you know, going to church or chatting with other people and more time working. So whether that’s a good thing or a bad thing is a separate question. I don’t have a strong view on the fact that, you know, going to church is worse than working. But whatever it is, they’re doing—whatever it is, they’re doing less of it and working more, if anything.

COLEMAN: For what it’s worth, we find the same thing in our cash transfer programs, too, that when you give very poor people cash they actually invest it productively and tend to work more in many cases.

But stick with the policy piece of this. You’ve done some really interesting work looking at nudges, or more than a nudge you could say, for immunization among children ages one to three. When you provided a small gift, in effect—I think it was lentils and a metal plate in India—immunization rates went dramatically up. What’s going on there?

BANERJEE: Yeah, so I think that this is—I mean, expanding, as you say, this is connected to this general theme of, I think, what does getting some extra resources do to a particular family, that’s related to the question of what happens if you don’t have those extra resources. And one of the things that happens is that your attention becomes extremely valuable, meaning you really—you have a stressful life. You have a bunch of problems. You’re worrying about them all the time. So, you know, anything that demands attention is actually a difficult task.

And one thing about the immunization is that it’s something that demands immediate attention. You have to sort of plan it. You have to—your immunization center is two kilometers away. You have to get up in the morning, make sure your child is ready to go with you, take your child there on the day when it’s available, and come back. That—

COLEMAN: And maybe miss a paid day of work to do that.

BANERJEE: Maybe miss a—but I want to focus on even the attention piece of it. Just you have to be attentive.

So one thing that we realized is that if you actually make the occasion more memorable, then it doesn’t demand as much attention. If you think that I’m—there’s something to look forward to—it’s not just that I’m going to take the child there, this child’s going to cry a lot because she’s going to get a shot, and then I’m going to come back with a crying child, that’s the kind of thing you’re trying to push out of your mind—even I try to push out of my mind if I can. So it’s not the kind of thing that naturally attracts attention. And if attention is a scarce resource, then we have to think of designing social policies that are responsive to the fact that attention is a scarce resource. One thing that—I think the reason why these might just work is because they kind of make the moment attractive. I don’t have to think about—the child might come home crying, but I’m going to come home with some free food, and that’s nice, and you know, I don’t have to worry about money for a few days and that’s nice. So I’m going to look forward to it.

And that’s—I think that transformed—the reason why tiny pieces of money—these are like, you know, even for these people it’s not a lot of money, the people who were subject to these interventions. Why they have large consequences—and people have found this repeatedly—is because I think they grab your attention. They’re attractive.

COLEMAN: What about from a cost perspective? As a policy person designing a program like this, is it cost effective to give out those small bits of money or goods or services?

BANERJEE: So I think—I guess you know the answer. I’ll reveal the—the answer is yes. It was actually—it was—this indisputable study, we showed that if you give people this small gift, then the cost per child immunized, including the gift, goes down, because the main cost is not actually the gift; it’s the time of the person who’s sitting there doing nothing because nobody’s coming to get immunized. That’s a much more expensive resource because you have to have a call chain. You have to bring the immunizing material all the way down to the village. All of that’s costly. Once you’ve got it there, the marginal cost of giving, you know, a little bit of some gift to somebody is actually small. So basically, the cost per immunized child is much lower when you—half; I think we cut it by 50 percent—when you actually gave out—when you give gifts to people, including the cost of the gift, the immunization cost less when you do through this mechanism for a gift because just so many more people get immunized.

COLEMAN: So have you seen this—RCT was done in India. Have you seen the Indian government revise its policies to increase immunization by matching it with small gifts?

BANERJEE: So I guess the answer’s a little bit yes and no. So we have now a new RCT, which we finished a year and a half ago, where we worked with the government of the state of Haryana, which is one of the low-immunization states. And they actually—now the technology has improved, so you don’t have to give lentils. You can actually give, you know, cash transfers on their cellphones. And we, again, find that that increases immunization. So the answer is—now, has it become universal policy? Sadly, no.

COLEMAN: No? Yeah. OK, another example—

BANERJEE: I mean, there’s a reason for that, by the way. The reason is that people so hate the idea of giving gifts for—to people for something they should be—ought to be doing—it’s literally—we get this message from UNICEF in particular, was dead-against giving gifts to people for something they ought to be doing. I don’t know what ought to be doing means, but we’re in that—in that particular prejudicial frame, you get this—you get this response, that this is a bad thing. Don’t give them gifts. Well, the children will die, but you know we didn’t give there any gifts.

COLEMAN: Another paper that intrigued me was the work you’ve done on fertilizer use in Africa. Farmers should buy fertilizer. It certainly increases their productivity. And yet, even with subsidies, which can be quite expensive, they don’t. And then you did a project where I think it was you offered—

BANERJEE: I didn’t. Some of my colleagues in—

COLEMAN: Not you. Somebody—you used it—yes, you evaluated it. Where there was free delivery offered. And it increased uptake. What—talk a little bit about that project.

BANERJEE: That’s a project where I think we—I think we started from a premise that was kind of behavioral economics 1.0, which is that these people, the reason why they want—why they can’t do it, is because they can’t—you know, they can’t save. So they get the money at the end of the harvest, and the next planting is three months later. So they can’t actually hold onto the money between one of those. So we promised them that they’ll get the fertilizer, you know, just before the planting so that they didn’t have to save money. So we’ll take your—we’ll take your payment now, at harvest, and then we’ll give you the fertilizer in three months.

Turned out, they actually didn’t want that. So we started with our own behavioral economic 1.0. It turned out that was the wrong behavior economics. They were perfectly happy to take the fertilizer the day after they paid the money, they just wanted—and they could hold onto it. They didn’t sell it to buy trinkets or anything. They just—what we were solving was mostly a problem of attention. They just couldn’t get to the—you know, this whole sequence of things, of going and buying it, and stocking it up, that’s something they couldn’t do. And it was a problem of attention that we managed so solve, I think, there. And that’s why I think these had big consequences.

COLEMAN: And have you seen uptake of that intervention, rather than providing bigger and bigger subsidies, instead trying to solve that attention problem?

BANERJEE: No, I have not, I must say. I think there are reasons why. I mean, people—it’s not entirely just that, you know, bad economics has an unbroken sway. I think there is some good reasons for doing this. I think this requires a certain amount of credibility, because you are—what you’re basically offering people is that, you know, we’re going to get—instead of you going and buying fertilizer, we are going to intermediate. And maybe we don’t provide the right quality. So I think there are reasonable arguments that people provide. But I think—fundamentally, I think this mode of thinking, where you think of, you know, people not automatically taking the right decision, that not just actually useful in making them take the right decisions, this mode of thinking is not yet at the universal.

COLEMAN: One other area that I’d love to just ask you about is some of the work you’ve done on women’s empowerment and women’s leadership, which has been very important to my work over the last fifteen years. And Jay Powell has done some very interesting studies in this vein. And I think back to work that I’m familiar with, with when women hold leadership positions it begins to breakdown stereotypes. So when there have been no women in leadership positions, men and women rate women leaders poorly, even though on an objective basis it’s the same performance. And yet, once you’ve had a woman in a leadership position, then it doesn’t negatively impact how they view men, it just positively impacts how they view women. Which is fascinating. It’s just a change in the prejudices that we’re talking about. How—talk a little bit more about some of that work, and what the influence of it has been.

BANERJEE: So I think that’s a—that’s a work which I think has had legs, I think, in the sense that I think lots of—I mean, in the last fifteen years, basically you’ve seen lots of countries extending their—the—you know, some way of getting more women into positions of power, basically. And usually using some kind of quota. And the first order reaction from, I think, both the economics community and the policy community was that, you know, you are not going to be able to change this with quotas because essentially if there are deep-seated prejudices, you know, if these women—these women might actually get elected. If you force a quota, they will. But that won’t change anything, because their husbands will still run the show.

In fact, it’s not true. I mean, it turns out that when you actually change the kind of gender of the person in place, both policies change and changing the direction that is, you know, more in the—you know, more consistent with the preferences of women. And second, the preferences of the voters change. So it’s not that you—what happens after a while is that, you know, first you put a woman in place who is—who happens to be quota-driven. So meaning she would not have been elected without the quota. But ten years later, there’s been a bunch of women being elected. The women who get elected are not just getting elected in places where the quota is in place. They’re getting elected in places where there’s no quota in place, because people have seen that these women can be competent.

And it’s not the same women necessarily. It’s the next generation of women. In fact, we see parents investing more in the education of their girls where the leaders were women in a previous generation, because they are—they feel like these are all—you know, women’s—these careers are real. So in some sense, the transformation is much broader than what the quota would have just implied. The quota would just change the identity, but in fact it changes the entire perception of the players.

COLEMAN: It reminds me of a story that I’ve heard, which was probably apocryphal or created by a behavioral economist, that Angela Merkel visiting a kindergarten class and asking all the little children what they wanted to be, and all the girls saying they wanted to grow up to be a politician, or a president, or this or that. And all the boys wanted to be firemen or policemen. And she asked one of the little boys: Do none of you want to be president? And they say, are boys allowed to do that? (Laughter.) Anyway, it’s the reverse stereotype.

BANERJEE: Excellent.

COLEMAN: OK. Last question from me. So we’re talking about prejudices and, you know, real sort of blinders in some ways, among economists designing programs before policy people. What about donors? Donors have a lot of prejudices in terms of how they think about the recipients. You know, there’s the idea—I know I see it all the time in the work we do—you can’t just give unconditional cash to people. They’re going to waste it. They’ll drink it, smoke it. Although now we have nearly two hundred randomized control trials that show exactly the same thing, which is that they don’t. So how about a little behavioral economics in the donor community?

BANERJEE: Well, I mean, I think it’s almost pre-behavioral economics. In the donor community, I think what is really remarkable is the strength of beliefs about, A, what’s happening in the world and, B, how people behave. And those are—it’s not that—I don’t know whether I would call that behavioral economics or not. It’s just—I don’t have to know anything. I already know how the world works. And I think one of the examples is the one you gave, but I think there’s a broader misperception. I mean, one of the things you keep hearing is that there’s no point trying to do anything because we know that all the money goes down the drain. In fact, the experience of the last twenty years is the world—every single kind of social policy indicator has improved. You know, most children are in school now. Girls participation in schools is crossing boys participation in high schools in many countries.

In Bangladesh, to take an Islamic country, girls are much more likely to complete high school than boys. It’s—if you look at infant mortality rates, they’re falling. In fact, they had kind of plateaued for a while, and they’re falling again. If you look at life expectancy at five, it’s growing. You know, take any single measure—malnutrition rates are falling. In fact, they were again plateauing for a while and have started falling again. So it’s not that the—I think the first order fact about the world, I have—in other audiences, at other times, I can rant about all the things that are wrong with the world. The first thing you have to say is that there’s something incredible right about what’s happening. That, you know, good things are happening.

And this is not because, you know, somehow, you know, some magic is spreading. It’s because we are moving to right policies. The example—for example, one of the policies that our work, and Jay Powell I think has supported a lot, is the delivery of free insecticide-treated bed nets, which basically prevent malarial deaths. And this is—was something that saved several million deaths a year. And so one of the reasons why, you know, child mortality rates are falling is because they are not dying of malaria. And that’s not—there’s no magic there. It’s very simple. If you have money, you can give away bed nets. And if you have bed nets, then you don’t get malaria.

And I think that particular sense—so I think the first order, I would say, prejudice that’s damaging is this idea that we’re in hopeless world where we can’t do anything. It just doesn’t seem to be true. The world is just—things—good things are happening all the time, in a sense.

COLEMAN: And I would just give a shout-out for Jay Powell and the work that you’ve done to really bring scientific methodology into evaluation, so that donors can figure out how best to apply a dollar. Does it have an impact? And what’s the benchmark? And using the scientific methodology of randomized control trials, and control groups, to really measure that impact over time has been, I think, a revolutionary development in the global development space. So thank you.

BANERJEE: Very kind.

COLEMAN: OK. So let’s turn to our members now and take your questions. If you could remember, this is on the record. And if you could wait for the microphone, stand, state your name and affiliation, and ask a question. And we’ll start right here.

Q: Professor Banerjee, Robyn Meredith from JPMorgan.

When you talked, you gave some examples of relatively small amounts and generally developing countries. But of course, I couldn’t help think about the argument now for universal basic income. Can you walk us through what you research shows, or what you can apply from your research to help us think about whether universal basic income in a developed country like the United States is a good idea or a bad idea, given income inequality and technological disruption?

BANERJEE: So, sadly—let me—let me—I’ll try to answer the question. But there is very little, partly informed by the political prejudices—there have been very few unconditional cash transfers in the U.S. Everything is kind tied to being poor, et cetera. So it’s hard to see very strong, direct evidence on this.

Now, so let me use the indirect. The indirect evidence is that if you look at the measured—what people—economists call labor supply elasticity, which is what happens when you kind of—if you—if you lower the wage. So imagine that what I’m doing is sort of dampening people’s incentives. Do they work less? There’s almost no evidence of that, even in the U.S. So the labor supply elasticity, even in the U.S. is very low—much lower than people thought it was. In fact, I think there is the current—the current view is it’s about zero. So I don’t think there is a strong view that, you know, you need to provide incentives for people to work. They work or not, but it doesn’t seem to be very correlated with incentives. So that’s a starting point, I would say.

There is clear—I think in the U.S. there are a bunch of studies which show that this kind of general model I was sketching of scarcity—which is people behaving in irrational, quote/unquote, ways, because they are under stress, that—there’s actually nice experiments run in New Jersey showing that when you make poor people think about money, they take worse decisions. So I think—I think that—so in reverse, if you give them money, they might take actually good decisions. So all of—I wouldn’t say that there is a slam-dunk here, because I think partly because—since the 1960—after the 1960s, basically this direction was shut down in the U.S. And essentially the idea—the strong prejudice that if you try to give people—poor people money they’re going to become lazy and retire and, whatever, drink it all, has such—taken such powerful hold of our imagination that we don’t actually even have many tests.

Most of the evidence from developing countries—because their this prejudice was weaker, in a sense. And so many countries tried some version of giving cash to people. So we know a lot more about what happens if you give cash in some version, or assets to poor people in poor countries, than in rich countries. We haven’t really tried it very much. If you look at the Finland experiment, we see the—as the first example of kind of what they call a negative income tax experiment, where—you know, if you try to bring people up to a particular income. In recent years, the results just came out. And the results suggest that people don’t stop working. They also don’t get richer. They don’t get poorer. But what happens? They’re much happier. They’re much more content with their life. And I think that’s a big deal, given the political context in which UBI is being talked about in the U.S.

COLEMAN: Right here.

Q: Hi. Peter Osnos of Public Affairs Books, your publisher. (Laughter.) Among other things.

COLEMAN: You’ve met before.


Q: Haven’t actually. So you never mentioned microcredit, either of you. And yet, obviously microcredit is one of the more extraordinarily important parts of the economic changes that have taken place, particularly of Bangladesh, of course. But so what do you make of microcredit? Is it as good as it appears to be, as we would like it to be?

BANERJEE: Those are two separate questions. So I think we’ve done actually a lot of work on microcredit. And I will say three things about it. So good or bad depends on what you think is the objective. If you think the objective is getting people to start millions of small businesses that increase incomes, the answer is it doesn’t work. And I think there’s resounding evidence suggesting that.

Second, does that mean it’s not a good thing? Not necessarily. If people use it to buy—mostly people use microcredit, whatever they say, they use it to pay for their daughter’s weddings and to buy refrigerators and televisions. Am I happy that they’re able to do that mostly? Yes. There’s no evidence that’s making people—you know, driving them into poverty either. They seem to be able to handle it. They seem to be able to take loans. I take a loan to buy my house. They might take a loan to be able to buy a refrigerator. It seems to me to a perfectly good thing in people’s lives.

Third thing, there’s a small—I think this is more—maybe more recent. There’s a small group of these people—these poor potential entrepreneurs, 5-10 percent, who seem to get very large benefits from getting microcredit, because it—but there’s a small fraction, so they don’t really show up in the mass data. So it’s not the case that everybody’s sitting there wanting to be an entrepreneur and if you just give them the credit they’re going to have these wonderful ideas and transform the world. I think that particular vision I think is spurious. I do think that entrepreneurship is hard, and that only a few—a small, selected group of people do it. That doesn’t mean that they don’t exist among the poor. But they’re no more likely—I wouldn’t be able to run a business to save my life. I don’t think of it as being necessarily prejudicial. I think some people are meant to be professors as well, or whatever, do some other job. (Laughter.)

So I think the evidence is that the poor are just like everybody else. There’s a small fraction of people who are ready to be entrepreneurs, who actually take their business forward. Most others, they start a business and the business goes nowhere and doesn’t make money.

COLEMAN: What if your goal is poverty alleviation?

BANERJEE: So it doesn’t change their incomes. So that was my first point was on average it doesn’t generate extra income. It does generate potentially happiness because they buy televisions, they buy refrigerators, they repair their roof. They do all kinds of interesting stuff. But this doesn’t add to their come. Small minority of these people actually do gain income. But a small meaning 5 percent, rather than 25 percent. So it’s—I think that’s the—my sense of the bundle of evidence that we have at this point.

COLEMAN: It’s why I’m working on unconditional cash transfers not microfinance. (Laughs.) We’ll go into the back.

Q: Hi. Thank you very much. My name is Ameina Tarana (ph). I’m with Visa, Inc.’s Social Impact Division, leading our work on policy and measurement.

So I have a follow-on question from your latest comment, and a more theoretical one. And the theoretical one is, how can—what is your advice to policymakers, to companies, to donors to take the findings from the behavioral research that you and others are conducting and actually apply it? And I ask that very specifically in the context of your latest comment around microcredit, because that evidence has been out for quite some time. And we’re all quite aware of it. And yet, the rhetoric is still very, very strongly: Businesses need credit to grow. There’s a shortage of loans. That’s what we hear all over, all the time. And so where’s the disconnect?

BANERJEE: Excellent question. So I think the disconnect is exactly on the point of—the last point I was making. Which is that there is a—there are businesses that need credit. That is exactly the point I was making. You just have to find a way of identifying them. It’s not the case that every business, a $200 or $300 loan to every business does a lot of good. What—I don’t think there’s a way to avoid the hard work of actually identifying the businesses that are talented. There is some interesting recent work. So I don’t think this work is—the work that I was taking about just a minute ago, about the fact that there’s 5 percent who seem to be talented. That’s actually very recent work. There was a bunch of papers, all of which seem to find that kind of thing, which is there is a small minority of talented entrepreneurs.

And I think that’s a—what that implies is that you can’t just sort of dump every—say the shop is open, anybody comes and takes away credit. You have to go back to the old-fashioned problem of identifying who are the good entrepreneurs. That’s what banks are supposed to be doing. And the fact that—I think part of the trick of microcredit was to say you can abdicate that responsibility, just give us the money and we’ll give everybody an equal amount, and we’ll collect it. And they were very good at collecting the money, but they were not particularly good at targeting it to the right people. What you want to do is give $50,000 loans to 5 percent, and not $200 loans to everybody.

And that—so I think the answer is all the evidence is consistent with what you said at the beginning, which is that there is—or in the middle—that there is huge unmet demand for credit. That’s absolutely true. I’m not disagreeing with that. I’m saying that comes from a small number of people. You have to find those people. They have—you have to find the people who are both reliable repayors and productive users of the loan. That takes some work. That’s what the financial sector should be doing. It doesn’t do enough of it. And microcredit didn’t solve that problem.

COLEMAN: In the back, right here.

Q: Hi, professor. Darshan Mody from EY.

Sir, have you seen instances where nudges in the opposite direction work as well? So, for example, you know, how do you get folks to smoke less, drink less, you know, pharmaceuticals, or social media, or any of those things? When—you know, because of course what you are doing is also being done by researchers at these companies. And how do you, you know, sort of counter that?

BANERJEE: That’s a great question. I don’t know the answer to how I counter that. That’s a—they seem to have—I mean, I think if we had a lot more money and a lot more resources, I think we could do a bit better. There’s certainly—the cigarette companies have much more resources than we do, so we’re kind of up against it. So the first-order answer is, you know, I think people should pour more money into this kind of work, trying to see how one plants those ideas early.

One thing that we find a lot of surprisingly strong effects from entertainment. We’ve been doing a lot of work on basically we made up—I mean, in—among—we, in J-PAL, we made up a sitcom in Bhojpuri, which is one of the Indian dialects of Hindi. And in that sitcom, we—the idea was to get people to take—use double-fortified salt—that is, iron-fortified salt to get them to have less anemia. And we do many interventions of different kinds, you know, give them advice, give them price discounts, et cetera. The only thing that works two years later, the only people who are taking it, are the people who saw the sitcom.

We did this even more, I think, remarkably, I think we have a—we worked with the MTV Foundation to do an RTC off of the intervention, which was a TV series in Nigeria called Shuga, which was—and this is actually HIV and on gender-based violence. And we find large effects on both of those things. So I do feel that there is actually scope for interesting work trying to tell good stories, get people into the story, to understand what’s really damaging here. And I think people are responsive. So that’s—the good news is I think people are responsive. The levers seem to be there. Then it will take much more, I think, money. And the money is on the wrong side sometimes, as you are implying.

COLEMAN: I wrote a piece years ago about women’s empowerment in Afghanistan, arguing that one soap opera, which focused on women in some very interesting roles and interesting family dynamics, did more to empower women than all of the State Department’s interventions. They were not very happy with me but I stand by it. I think it’s true.

BANERJEE: I will believe that.

COLEMAN: Right here.

Q: Thank you for your work and for your talk today. My name is Joel Cohen. I’m at Rockefeller University.

About eight hundred million people are estimated to be chronically undernourished and about a hundred and fifty million children under the age of five are stunted—that’s about 23 percent of all of the world’s children—are stunted due to chronic under nutrition. Would you talk about what are effective interventions that are known to diminish childhood stunting and the stunting of adults, what are the obstacles to the spread of those innovations, and what are the economic consequences for labor force later of reduced childhood stunting?

BANERJEE: About three hours of talking. (Laughter.) I’ll say something. So I think that right now I would say I think for the—I think the kind of unfortunate reason that there is a—you know, the view is that, you know, this is one of these things where parents have the right judgment. Part of—part of—a place where there is very little research is in changing dietary preferences.

We haven’t done much, and I think part of the reason is—for example, the weaning foods are terrible. In many parts of the world children go from having, essentially, only being breastfed to eating adult foods and that’s—you can see exactly where you see the dip in the child, when the six-month-old in many parts of the world where there is malnutrition is still not malnutritionated. At two, he or she is—so I think that there is a—this is a big gap, I think, in terms of how to communicate to parents that, you know, the weaning period needs a different set of foods, and I think while it’s easy to say that, you know, there is—you know, there are traditional systems of providing nutrition, this problem seems to exist in some places and not others. So I suspect there is something to be done there. So I think—let me say one—and you asked many things. I’m not going to be able to answer all of them. But let me say one other thing. One good news is the flip side of the TV—the TV. The TV example was that very poor people buy TVs.

The good news is that if you give them a little bit more money they buy food. So the idea was that, you know, if very poor people buy TVs, then slightly richer people will buy better TVs. But, in fact, not true. The elasticity of food demand, meaning what happens to food demand when you increase income beyond the—through some of the interventions that give directly and others that do it, these are very large elasticities.

Often, if you increase income by 10 percent, food demand increases by more than 10 percent. So, in fact, people have this reversed ordering. They first buy the TV but after they bought the TV they are happy to—happy to buy more food. So I don’t—I think that’s the good news. The good news is that I think we can actually get people to eat more if we increase incomes. Maybe that’s—let me stop there.

Q: Thank you. I’m Allen Hyman, Columbia Presbyterian.

I want to bring us back to America. The response of many politicians or policymakers to deal with poverty is to raise the minimum wage and this has become quite an important question in the—as we go forward in our political campaigns. But the behavior response of employers is to buy technology to reduce the number of people they have to employ. So I wondered what’s—have there been any good studies to show the effects of raising the minimum wage on poverty?

BANERJEE: So lots have studied work on the effect of minimum wage on poverty. I think you are highlighting something important, which is the weakness of this work. The strength of the work is—and then the conclusion of the work is that in the short run there is no strong negative effect of raising the wage. You raise the wage, basically, you get a countervailing effect, which is the poor people spend more money and so it seems like there is no big negative response.

What might be true is that in the medium run you get substitution away from labor. I think that’s a general problem that is bigger than raising the minimum wage. I think, if anything, my sense is that right now the entire tax system subsidizes not employing people because it creates, you know, payroll taxes.

You don’t pay payroll taxes on a robot, and so I think there is a sense in which the entire structure of taxation is subsidizing substituting capital and we should be thinking about the broader question, which is how do we create incentives to employ people rather than employ machines. This is a very big question. Again, it’s another question—a three-hour question. But I think that’s where this connects up.

And so I think you’re right there is—this is a big worry that in the short run we’ll be able to get away with it but in the long run maybe this will help, even it’ll speed up the process of mechanization, which is already frightening me, at least.

COLEMAN: Right here.

Q: Darin Kingston with the Global Development Incubator.

I’m just wondering what are the, like, top two or three policy questions that you would love for JPAL to be able to contribute to but haven’t gotten funder interest or there’s not a line donor interest or whatever the case?

BANERJEE: Excellent.

COLEMAN: (Inaudible.)

BANERJEE: Yeah. So I think one of them is the one I already expressed, which is how do you design real programs for identifying talented entrepreneurs among the poor. I think this is—this is super important. It’s something we haven’t really—I mean, partly it’s expensive because, you know, people are happy (to do ?) our cities on $200 but on $50,000 they are less willing to do it. So we have sort of hit a brick wall whenever we have tried to do that, and I would love to have been—I know a lot of people in this space and when I broach that subject they are—they become less friendly. So I—(laughter)—so that’s one of them.

I think the second one, if I had to pick, I think getting—I think getting this—the nutrition frame right. I think that one—again, I don’t think we have done a lot of work there partly because I think it would require a different set of actors. It needs to bring together people who are willing to experiment with designing foods. In some sense, I feel like, you know—because I think you can’t tell people, like, you should have these dry biscuits and they’re better for you, I mean, or boiled eggs. Give your child a boiled egg every day.

It just doesn’t work. Somehow, you have to work with the particular cultural substrate that you’re working on and I don’t think we have made that much progress on this. I mean, I think, in general, there is good news. I think malnutrition is falling but I think, overall, it’s falling slower and we could speed it up. The third—and perhaps the place where, again, I think it’s the constraint, so isn’t so much money as much as ideas—is how to upscale sort of education at the secondary level.

That’s where I think there is really—I think JPAL has done a lot of work on primary education. I think we have reasonably reliable answers to some questions. Where we are weaker is in secondary education partly because in secondary education, again, innovation has been slow.

So it’s not just that we are not able to do the research. The innovation is slow and it’s where, I think, innovation is needed because that’s where the population bulge is. That’s where I think the political dangers are, et cetera. So that’s off the cuff. If you ask me this question tomorrow, I’ll give you a more reliable answer.

COLEMAN: Right here.

Q: (Off mic)—like. Lots of interesting ideas. Lots of interesting whats, but how is always the more difficult issue. How do you gather all these interesting ideas together and make more of them happen? What are the institutional constraints that need to be overcome and how can they be overcome to do the things you’re talking about?

BANERJEE: I’m glad you asked that question. So one of the things we’re doing in JPAL, in fact, where we put a lot of effort—my own effort—I put a significant part of it—is in getting institutions to scale up things that evidence has shown works, which is a question you’re—and the answer, surprisingly, to me, was that it’s not—the reason why it fails is often not because there is, like, an unwillingness to take it on. That’s there, too. Sometimes that’s what stops us.

But, often, it is the case that it just—you explain an idea to some bureaucrat or politician and the action is, can you come here and help us do that. And my answer is usually, no, I have a day job. I can’t do it. But the real—the relevant answer is that could you provide us with a bunch of staff members who would sustain that effort, you know, would—because I think most of—most effective bureaucrats are overworked in the developing world and so, therefore, getting them to innovate is hard work. So we’ve actually done this.

We’ve actually planted—we have hired people and planted them in government not to—just to take the task of helping implementation of research and that’s been very effective. So we actually have a mechanism for it. We have a funding system for it. We raise money for it, not very successfully. But it’s very hard to explain to people that this a really important step, which is you go for—you have the research and then you have the people who are implementing it. You even have the willingness to do something but you can’t do it because there isn’t enough legwork happening.

And I often find it frustrating to explain to donors that actually this is an extraordinarily valuable way to use your money. The governments have tremendous capacity. They actually—relative to the—all the donors in the world, governments have—in the Third World have much more money—much, much, much money. There’s the volumes that are not comparable.

And so that’s where you get leverage. If you can get them to implement good policy rather than bad policies you get huge amounts of buy-in and leverage. And so we’ve tried and where we have done it we’ve often had big successes. I think this is where I’m—so I’m encouraged. I feel that, you know, we can get—if we can get—because sometimes you have the—all the ducks are aligned except the legwork duck, which is, you know, everything requires sixteen memos that have to go to a parliamentary committee that’s going to approve this policy change.

But who’s going to write those memos? Who’s going to make sure that those memos are, you know, all—you know, nobody finds a hole in it and it gets tossed up there. So this process is actually extremely underfunded at this point. I would say we are—we have the concept. We have been implementing it actually now for fifteen years now but not necessarily at a scale we would like to.

COLEMAN: But just following on that, I mean, don’t you find also that development has been done with such a lack of evidence for so long that there are very deeply vested interests in perpetuating what’s currently done, even when evidence is brought to bear that what they’re doing isn’t really effective? You know, you have—

BANERJEE: I think—I think that’s surely true. It’s also surely true that—but what I think I’m actually struck by is the opposite. I’m struck by the fact that I often encountered the willingness to try something different. People are also realistic, I mean, especially—I mean, it’s fashionable these days to beat up on politicians. But politicians often have a very clear idea that if I can show something that works then I can get reelected. They’re—and people are reasonably responsive if you give them good ideas. I’ve actually—I’m often struck by how little—how little resistance there is to ideas.

Well, you might think that, look, you know, why should we change because—just because you happen to say it. But, in fact, people kind of see that is not working. It’s not that they are blind to the fact that, you know, they’ve been doing something, it was based on some silly idea, and it’s not working. They see it. I mean, their eyes open.

So they’re actually open to the idea that you could bring some value here. You could try something different. We often don’t provide the product and the product has to be not just a concept but a fully designed thing operating on the ground and that’s where the failure is, often.

COLEMAN: I’m afraid we are out of time. It’s 2:00 and we do end on time. Can we—join me in thanking Professor Abhijit Banerjee so much. Thank you. (Applause.)

BANERJEE: Thank you. Thank you. Thank you very much.


Top Stories on CFR

South Korea

In seventy years, the U.S.-South Korea alliance has evolved from a patron-client relationship to a global comprehensive strategic alliance.


Far-right candidate Geert Wilders emerged as the biggest political force in Dutch politics in November’s general elections. What does his win mean for the Netherlands and Europe?


Myanmar's military has recently suffered a string of defeats—but the U.S. government seems unprepared to face the country's potential state collapse.