Interview with the 2019 economic sciences laureate Michael Kremer on 6 December 2019 during the Nobel Week in Stockholm, Sweden.

How was your interest in science sparked?

Michael Kremer: As a child, I was interested in science and I was fortunate enough to be mentored by a physics professor, Larry Weaver. He taught me the beauty of elegant theory, but he also taught me another lesson. One day he told me about a paper he was working on, and a colleague had found a mistake in it. I thought he was going to be upset by this, but instead he was happy about it. That made a big impression on me. He really was an example of a commitment to science as a collective search for truth. That’s a lesson that I hope has stayed with me. 

My father is an architect. He taught architecture at Kansas State University. Watching him work on something as simple as a poster, advertising someone coming to speak, and seeing him design and redesign taught me the value of craftsmanship. My co-authors have noted that I sometimes want to keep revising things. So now they know who to blame.

My mother taught literature and she emphasised our responsibility to address issues of injustice and suffering if we can. I think it was thanks to her influence that I chose the field that I wound up in of development economics. The one wonderful thing about this working on this experimental approach to address issues of poverty is it allows both the connection to science, not physics, but science in some way or at least aspire to that. But also the more immediate and direct connection to human beings and a feeling that we’re working to try to address these very important issues.

How did your experimental approach to economics begin? 

Michael Kremer: I felt it was important to work and on address these issues for a long time. But after college, I had the opportunity to teach secondary school in Kenya. Then I went back to the US. I was involved in starting an NGO, then went to graduate school, eventually finished and got a real job with a real salary, and then I could afford to go back to Kenya. My wife and I went back to Kenya, and we visited some old friends of ours, and one of them Paul Lappe had just been hired by an NGO to help them start activities in Western Kenya.

The NGO worked on education issues and with school children. Paul and I were talking, and it turned out that he was supposed to choose some new schools to start working in. They hadn’t worked in the area before, so they would have to phase things in over time. They had a particular program, a child sponsorship program that they had done before. But they weren’t sure what the impact was. And I mentioned that if they wanted to understand the impact, they could phase this in so that there was an initial group of schools where it would be phased in, were comparable to a later group. By doing so, they could have the equivalent of, or something close to the equivalent of, a medical trial where there’s a treatment group, at least during the early phase in period, only some of them would’ve been treated, and then later more would be. But if they observed after the early phase, then they could try to understand the impact.

This was just a conversation with a friend. I didn’t expect them to go ahead with us. So I flew back to the US after my vacation, and then I got a call. He’d spoken to his boss, and his boss had decided they wanted to do this. I flew back to Kenya and met with him, and we got involved in the evaluation that way. That’s how we started. It was really a discussion with a friend who was working for an NGO, and then they decided that they wanted to understand the impacts of their programs, which I would really applaud them for, because not every NGO, not every government is truly interested in understanding the impact of what they’re doing, but they were.In fact, when they learned about the results, they then decided they wanted to try new things to try to find how they could best serve the people that they were intended to serve.

I think that turned out to be a very fruitful process, and they were able to identify more effective ways of fulfilling their mission that would improve lives for more people. Ultimately, of course, that had a big impact, not just on their operations, but most importantly on what the Kenyan government and other developing country governments were doing with their own resources. That’s how I got involved.

Can you tell us about your relationship with your co-laureates, Abhijit Banerjee and Esther Duflo?

Michael Kremer: I used to teach at MIT. I was a junior faculty member there, and Abhijit was a colleague of mine. I’ll tell one story. Abhijit was trained as a theorist. I was trained as a macro economist. At the time I remember I was doing some work very related to mathematical epidemiology. I was doing some work modeling the HIV epidemic. It turned out that I found an unexpected result that there were some circumstances in which if people had more partners, that could actually lead to the reduction of the epidemic, or in some cases to the end of the epidemic. So very counterintuitive. I was very excited about this result, because mathematically it was a surprise. It was nice. Very simple natural assumptions lead to a very counterintuitive result. I remember talking to Abhijit about it, and he mentioned something about how the best papers are the ones where, after you read it, you think, how could we have not known that this was true? That made me think, and think back to the story I told about Larry Weaver.

This was something that was a mathematical possibility. It could theoretically occur, but, you know, how important was it in the real world? Was it just a mathematical curiosity, or was it something important? I think that was another experience that maybe pushed me in the direction of, tried to focus on what I thought was actually important for the real world. That’s part of why I wound up focusing more on this type of work. Esther actually came as a graduate student to MIT when Abhijit and I were both teaching there. We’ve been collaborators on a number of projects, and that’s been a wonderful experience. 

Do you enjoy teaching?

Michael Kremer: Yes, that’s really a huge perk of the job of getting to work with such wonderful students.

I mentioned that, in some sense, Esther was a student. I wasn’t her advisor, but I met her when she was a student at MIT. There’ll be some other former students here joining us in Stockholm, including Ted Miguel. Ted and I worked together on a study of the impact of treating children for intestinal worms such as hookworm, whipworm, roundworm, schistosomiasis. We found big effects of that, and then we found that those effects spilled over onto neighbouring to others because of reduced disease transmission.

We followed this group of children for many years. What we saw was improved educational outcomes. The girls were more likely to finish primary school, go on to secondary school, and ultimately improved living standards. So working with Ted over these many years has been a great pleasure. And Pascaline Dupas will also be here. Pascaline’s done extraordinary work on number of areas, but in particular on pricing of preventive health goods. One of the areas of greatest impact of the experimental approach has been, you know, there’s some very specific things like deworming, but there’s also some general principles. And Pascaline and his work has quite comprehensively examined the issue of pricing of preventive health goods.

Ted and I had an earlier result showing that pricing of deworming medicine had a very big impact on the number of people using it. If you charged even a small amount there was a huge drop off in usage rates. But then Pascaline launched a series of very nice projects that looked at many dimensions of the issue. She looked at mosquito nuts and found similar results as we had, on the one hand, but then added a very clever bit of design. When we started working on this issue, not only multilateral organisations like the World Bank and governments, but also most NGOs had the view that it was actually very important to charge people at least something for preventive health goods. One of their rationales was that if you don’t charge for it, people won’t value it and they won’t use it. There was a lot of stories about things like this, and people felt it very deeply, but as far as I know, nobody had ever tested it. And Pascaline and her co-author, Jessica Cohen, came up with a very clever way of actually testing that. And they found no evidence for that story whatsoever. Subsequent research is quite consistent with that. When not talking about a laboratory setting, but rather in in a real world health setting. Then she went on to examine are there benefits in terms of motivating the providers? A whole series of other questions. Does it affect takeup later on? If you get used to paying for something, you’ll continue paying.

In fact, it turns out that the free sample effect is stronger. If you give a free sample, then people learn about something and they’re more likely to buy it later. She looked at a whole series of questions, and I think it’s fair to say that the weight of the evidence from the work that she did, and the work many others have done, is consistent with the idea that free provision of preventive healthcare goods like mosquito nuts or water treatment solution or deworming medicine has a lot of advantages. And many of the hypothesised disadvantages seem not to be real.

Are women and girls more affected by poverty?

Michael Kremer: Gender is a very important area of research and development economics, and obviously a very important area of policy.

You’re exactly right that when we studied the impact of deworming we saw effects really differed by gender. Both girls and boys stayed in school longer, but the girls actually passed more grades and were more likely to graduate from primary school and go on to secondary school. The boys stayed in school longer, but they tended to repeat grades. In that sense, the program had a bigger impact on girls. Then we also traced out the labor market impacts. We initially saw it looked like stronger earnings effects for men. But over time, it seems that gap is not so clear.

But let me mention things a little bit more broadly. This is one example. I think more broadly, in many settings we’ve seen big impact from programs to increase access for girls. Another example is work on a scholarship program for girls. This was a program that was put in place when there were fees for primary school in Kenya. Those fees have now been abolished. But the program paid said that for girls who were successful in school at up to grade six, that the fees would be paid for grade seven and eight, and also some school supplies would be provided. One might think that the big impact would be on girls going on to seven and eight.

But what we saw was actually that there was not only an improvement there, but there was an improvement in sixth grade scores as well, because we looked at the mechanisms and it looked like students were studying more, teachers were in the classroom more. The fact that people knew that there was an opportunity to go on to seventh and eighth grade if they were performing well in school, led to more investment in school, not just by the girls, but clearly by their families and by the larger communities, including the teachers as well. We saw benefits that wound up benefiting not just girls, but the entire class and boys as well. Presumably in part because if teachers are there teaching more that’s going to help boys as well.

So more broadly, I think that’s a finding in the field. Esther Duflo, who’s also recognised this year, has done a lot of work on this issue. One very nice paper she’s done is on reservations in the local government for women in India. What she found is that this led to improved provision of water. Water is an issue that women care about. A lot of people had thought that these reservations of positions for women had no impact whatsoever, because they thought, well, the women are under their husband’s control. Yes, a woman’s officially sitting there, but it doesn’t make a difference. And she showed that, in fact, it does make a difference and has some very nice follow on work showing that once one woman has done the job, then that opens up the path for other women and people’s attitudes start to change a bit.

What do you like to do in your free time?

Michael Kremer: Right now I like to sleep! But I enjoy reading and my mother taught literature and that’s something I enjoy. On the plane over, I was just looking at James Joyce’s stories.

Watch the interview