Analysis, Interviews, and ReviewsArchive
Jul 13, 2011
An Interview with Banerjee and Duflo, Part 4
Abhijit Banerjee and Esther Duflo, co-founders of the Jameel Poverty Action Lab and co-authors of the recent book Poor Economics are at the heart of the movement to seek rigorous evidence about the lives of the poor and programs that aim to help them. As they write in Poor Economics, they believe that “we have to abandon the habit of reducing the poor to cartoon characters and take the time to really understand their lives, in all their complexity and richness.”
Recently I had the opportunity to sit down with Banerjee and Duflo for an extended conversation about the small and big pictures that emerge from their research, their critics and their plan to change the world. In fact, the conversation was so extended that I’ve had to break it up into pieces. We’ll be publishing it in four parts over the next few weeks—the full interview will be available soon via Amazon Kindle.
In Part 4, we discuss Banerjee’s and Duflo’s theory of change, the bad news for optimists and cynics, whether taking responsibility away from the poor aligns with Amartya Sen’s Development as Freedom and the ongoing debate over the usefulness of RCTs. In case you missed them here are Part 1, Part 2, and Part 3 of the interview.
TO: How much have you thought about change among the people who make the small changes in policy. You write a book like this with, I presume, the idea that it will change the way certain people behave and they’ll stop trying to make the big sweeping changes and just make the tweaks.
Have you thought about the mechanism of change? What is your hope that this changes people’s minds that are trapped in ideology and inertia? What are the tweaks that you can make to those people’s lives that can make them willing to hear and do and make changes?
AB: People talk about how hard it is to change policy or influence policy and that the institutions in developing countries are resistant to change. I feel like a lot of what we see as the culture of government is actually in some ways a very mechanical product of the last 50 years. Most countries in the world have created a whole bunch of institutions very very recently.
Developing countries inherited a kind of a state which was not particularly devoted to welfare of the people. So they saw these models of the developmental state elsewhere and decided, “We’re going to have this particular set of people implement this particular set of things.” So the institutions are not the product of a long accretion of incremental changes which made the bureaucracy particularly well suited to a particular kind of policy. This bureaucratic culture we have in so many places is just happenstance.
Most developing countries have only been independent since the late 1960s. So we’re talking about a period of 40 years. Forty years that began with a disaster. It’s hard for me to imagine that what we are observing in developing countries now is a steady state of some complicated process.
Most things are done with such amazing casualness that I feel like that can’t possibly be anybody’s reflective thought on how or what a process should be. People were given lots of power, lots of decision rights, little training, lots of ideology. I don’t feel that there’s a bureaucracy has reached a place, an approach, where it’s snugly sitting and then we’ll never push it out. Lots of bureaucrats and especially their political masters are very sensitive to the fact that somehow they lose a lot of elections and people don’t like them and they get thrown out. My sense is that they’re quite bewildered by that. I think in principle there’s a lot of demand for rethinking government.
When we talk to bureaucrats in India, obviously there are some who know everything and will tell you that they know everything and don’t need any help. But there are a lot who say, “Yes. I think we should do that. Can you help us with the that?” The reaction is often “Why can’t I get many more people like you to come in and help us with how to redesign this whole thing?” They know things are broken, too many programs don’t work, but it’s just too difficult in the middle of the job to fix it.
ED: I think we are reasonably internally consistent. If we want to be internally consistent, then we can’t be calling for a big revolution when one day everybody will start looking only at the evidence before making decisions. That’s not going to happen, let’s face it. There are a lot of reasons why things happen the way they do. So if we were trying to advocate to replace the world the way it is by some fully efficient technocracy that experiments with things before trying them out and then launches them with no error in the process, that would be inconsistent with what we are saying, and crazy. That makes our life easier. If a big change, a revolution is not what you are targeting, if you are targeting improvement at the margin, there’s lots of margin where you can start doing things. Maybe it’s a somewhat opportunistic thing to say. We don’t have a huge reason to fight people who are really resistant because you can always try with someone else. Like Abhijit was saying there are a lot of people who are willing to try things out. In the beginning of JPAL we worked a lot with NGOs because they were more nimble. Even there you have some who are never going to change their ways, and some who are more flexible and we worked the ones who were flexible and eventually the ones who weren’t going to change eventually came to talk to us.
With the government, and with international institutions it’s a bit like that as well. You can start having conversations with people and trying things out with people and trying to improve something somewhere. So we’re really talking about: can we make your school committees work better or something “unambitious” like that and from there you can demonstrate that it can be done and eventually it can become bigger and bigger.
So for example together we started working with the police in Rajasthan. We describe this in the book. It happens to be that at least some people in the police in Rajasthan were interested in improving performance and tried some things and it had some effect. Then we could use that as an example when we went to talk to the government of Gujarat about improving their pollution inspection and their audit business and in the process of that got the attention of the central government who are thinking about should we start a cap and trade scheme for carbon. And now Rohini Pande and Michael Greenstone who were working with me on the pollution issue are now working on cap and trade, which if they manage to do it will be really big. But it didn’t happen by us knocking on the door of the minister and asking, “Can we design your cap and trade system?” It happened by taking the path of least resistance.
Another thing is the example of the experience we had in France where you start by doing one project then another project and eventually the right politician at the right time will see it and say let me put in $250 million. It’s just creating the space and a fund and a technical capacity for anybody to try things out in the poverty space. And if it can happen in France then it can happen anywhere in Europe because people don’t like being left behind.
TO: While I was reading, it occurred to me that your book is bad news for optimists and for cynics.
ED: [Laughs] Who is it good news for?
I think it’s good news for patient people, for people who don’t have a hugely strong view about how anything in particular should work and are willing to apply that mindset but are willing to be proven wrong by the data.
TO: One of the things that you say toward the end of the book that leapt of the page at me is, to quote directly, “The poor bear responsibility for too much of their lives.” I can imagine a lot of people reading that sentence and being taken aback. It seems to be a stark contrast with the Sen idea that escaping poverty means having more freedom, more control of your life.
AB: I think we choose our words carefully. Control is not responsibility. In a sense I think having a lot of responsibility undermines control…
ED: …and freedom.
AB: And freedom. When you want to exercise control you need to have the psychological freedom to actually exercise it actively, rather than passively reacting to many many things. Control is not passively reacting to many many things. It’s agency. I think responsibility, lots of responsibility for very difficult things undermines agency. It’s not that you choose to do those things, it’s that you have to do those things. So it’s not choice, responsibility is not choice. It’s things that are dumped on you that you have to struggle with and in the process I think your agency is undermined because you can’t reflectively decide that this is the life choice that I want and this the life choice that I don’t want. Buffeted by a hundred things that I may or may not have chosen and many of which I don’t want and I somehow need to negotiate through that morass. I wouldn’t call that being in control or being free. I think taking responsibility off people’s hands and giving them a domain of freedom which is uncluttered by that where they can make choices without being constantly frightened by all the things that can happen, all the risks around, just the intellectual challenge of balancing 500 things, I think it is giving them freedom. Maybe it’s a disagreement about the nature of freedom. Prima facie I would not say that responsibility is control
ED: I don’t think that’s against Sen. It’s going further in the Sen path. The whole argument of Sen is that freedom means nothing unless you have the capability to exercise it. There are many examples of that, but one very striking example of that which he cites is the family in Bengal which is completely free to buy grain but they have no money so they can’t do it. The point he is making is that freedom is meaningless unless the possibility to exercise it, the capability to exercise it. We are not saying anything different. I mean, we are not free in the US to drink water that is contaminated with e. coli because it comes to us clean. Is our freedom reduced in comparison to the person in Kenya who is free to do that because if they don’t want to drink the water with e. coli they have to put chlorine in it. I don’t think we are less free, I think we are more free.
TO: Oddly there are people in the US now who are making that argument. [laughter]
ED: I’m not saying that we are not making a political statement on some level on what is the nature of freedom, but I think it’s an argument that’s pretty naturally in the Sen line.
TO: Is the war over Randomized Control Trials (RCTs) in economics being won? Are you having to spend less time explaining, justifying, fighting the internecine battles within the economics profession about the limits of RCTs?
ED: Abhijit and I disagree on that. I think it’s been completely won in that I think it’s just happening. A lot of people are doing it without us. It’s being used. I think it is now understood to be one of the tools. The argument within the economics profession had two main consequences, both good. First it raised the profile. If something was debated, people began to believe it must be significant.
Second, it did force us to answer the challenges. There were a lot of valid points that were raised and it forced us to react to that. We’ve become more intelligent as a result. I think Abhijit somewhat disagrees. I think he sees very prominent people like Angus Deaton who came out so strongly against RCTs have provided cover for people who were against evaluation, to have a very quick answer, “Rigorous evaluation is not all it has cracked up to be.” My view is that if people were not enthusiastic they were not going to be enthusiastic one way or the other. It’s just a front that argument takes.
I think on balance RCTs are a useful tool and people realize that and therefore it will continue its life.
I don’t think you can do something that is important and changes the world without meeting some opposition. I also think that’s a good thing. We wouldn’t want to replicate the microcredit arguments, but our whole view is that things should be questioned and I think that applies to us well.
AB: I am less certain that it has been won. The acid test of whether an idea has come to stay is that it becomes something that no one needs to justify using. This has happened first to game theory and then to behavioral economics during my years in the profession. RCTS aren’t there yet: it is true almost everyone is doing them, but many of them are taking the trouble to explain that what they do is better than a “mere RCT”. We need to get to the point where people take RCTs to be the obvious tool to use when possible to answer a particular class of empirical questions
TO: There are a lot of people, though, still saying things like “The important things can’t be measured” or arguing that too many important things are beyond the reach of scientific study.
ED: Part of the debate has been what is the nature of proof, what is the nature of scientific inquiry. That debate has been had. It was Kant and Hume, why do we need to have it again?
TO: I’ve seen other people make that the point that people can accept that small changes create big changes over time in biology but not that a similar process works in economics. They believe that in economics big changes can only come from big policy initiatives and grand strategies. Do you see that? Why do you think that is? Why do people want to believe that the big levers are the ones that matter?
ED: Maybe because the science in economics is a bit more shaky and it’s younger. We understand a lot less about interactions between small things in a complex system.
TO: I was pretty discouraged about the idea of small changes and making progress when the Opportunity NYC program was canceled and the reaction to Roland Fryer’s experiments. Both showed improvement, but small improvement, and the programs have been abandoned. No one seems to believe that the impact was big enough to invest in.
ED: I think that’s a bit of problem that we encounter as well. We have that problem with the immunization incentive program. That’s something I care deeply about just from a human perspective. I want to get that expanded. But even though we’ve never seen anything better ever in terms of lives saved per dollar spent, there’s a lot of resistance. There are many reasons people are reluctant about things like giving people an incentive for immunization. Some are ideological, you shouldn’t pay people to do something that’s good for them. But some people say, “Oh that’s good you increased the immunization rate from 6% to 38% but there are still 60% who are not immunized.” My response is, “Well yeah, do you have something better?” We published that work in the British Medical Journal. One of the referee responses was, “That’s not enough to guarantee herd immunity, so why do we care?” Well we care because it’s still good even it’s not herd immunity. We care because they 30% of people aren’t going to die. And even if you don’t have herd immunity you still get contagion effects. All of the evidence on vaccinations is that you get effects [of lessening epidemics] at all levels.
Of course if you get to 100% you get the big prize of eradicating a disease like Gates is trying with polio, which is a completely legitimate exercise. But it’s not like that you shouldn’t do it if you can’t fully wipe out a disease. There’s no reason not to improve measles immunization from 6% to 38%. But people don’t get all that excited about it. So it’s very big, it’s still a lot of lives, but it’s not a game changer.
By measuring well and showing what the effects are you get these kind of realistic answers. When you don’t measure you can always claim something big. You can hope that it’s going to be something big and that every child in New York City will be reading at grade level in a few years. But when you actually measure you realize it’s not going to get there, but it’s still something.
Starting from a context where most things are at least a little bit overblown, then the sobriety of measurement puts any program at a bit of a disadvantage. But I think that’s a culture that could change, that people could get used to using a different lens. The honest but modest number that has been established from a trial or in another rigorous way could replace the rosy description of what you wish was happening but that change might take a little longer.