Michael Kremer on Economists as Founders (Ep. 107)

Why economists should build more than a body of research.

Michael Kremer is best known for his academic work researching global poverty, for which he was awarded the Nobel Prize in 2019 along with Esther Duflo and Abhijit Banerjee. Less known is that he is also the founder of five non-profits and in the process of creating a sixth. And Kremer doesn’t see anything unusual about embodying the dual archetypes of economist and founder. “I think there’s a lot of relationship between the experimental method and the things that are needed to help found organizations,” he explains.

Michael joined Tyler to discuss the intellectual challenge of founding organizations, applying methods from behavioral economics to design better programs, how advanced market commitments could lower pharmaceutical costs for consumers while still incentivizing R&D, the ongoing cycle of experimentation every innovator understands, the political economy of public health initiatives, the importance of designing institutions to increase technological change, the production function of new technologies, incentivizing educational achievement, The Odyssey as a tale of comparative development, why he recently transitioned to University of Chicago, what researchers can learn from venture capitalists, his current work addressing COVID-19, and more.

Watch the full conversation

Recorded September 9th, 2020

Read the full transcript

TYLER COWEN: Hello, everyone. Welcome back to Conversations with Tyler. Today, the guest is Michael Kremer, recent Nobel laureate in economics. Michael, welcome.

MICHAEL KREMER: Thanks. Happy to be here.

COWEN: Obviously, you’re famous as an economist, but I’d like to start with a notion of Michael Kremer as founder. In my notes, I have it that you founded at least five nonprofits with the sixth to come. What is it about founding that, for you, is so enjoyable?

KREMER: Oh, I do like doing that. I guess it’s great to feel like you’re having an impact on people’s lives. I think that some of the things that I’ve founded have gone on to have that impact, and that’s very exciting. It’s obviously also an intellectual challenge to think about how to put things together and how to grow a team and get something moving.

COWEN: What comparative advantage of yours is it using? A lot of economists will produce ideas and let other people do all the founding. What is it that you bring to founding that’s special to you?

KREMER: There’s a lot of relationship between the experimental method and the things that are needed to help found organizations. I think that’s something that, historically in economics, many of us have seen research as something that’s divorced from doing things in the real world and, in some cases, even antithetical to it. I think that’s really different in many other fields.

Nobody thinks it’s strange that people who are in computer science go on to found firms, or that people who are researchers in biology go on sometimes to found firms. Obviously, there are plenty of people in computer science or in biology who aren’t interested in that, but people don’t think it’s strange that there’s some people in those fields who do. We’re now getting to a time in economics where I don’t think I’m unique in this.

Many economics researchers — researchers in development economics, in particular — have gone on to found organizations, and those have had real impact. I don’t think this is exclusively tied to development economics since the experimental method, but I could explain a little bit more about why I think the experimental method and contemporary development economics research does have a natural affinity for that.

COWEN: But that I get. Why do you have a natural affinity? If I think about your O-ring model — as you know, that predicts the less productive people in an organization, in a way, drag down the productivity of the more productive people.

Now, you’re one of the most productive people. You might think, according to O-ring theory, maybe you should found with Abhijit and Esther, but not with anyone else. There’s not a certain comparative advantage in that model. How do you think about the economics of you as founder? How should I model that?

KREMER: I’ll just give an example. The most recent thing that I’ve been involved in cofounding was an organization called Precision Agriculture for Development. What that organization does is, it works on mobile phone–based agricultural extension. At least that’s where it started. It’s now broadening out.

That’s something that requires we use tools from experimental economics. We do A/B tests. We use ideas from behavioral economics. So it’s using a lot of tools from economics, but it also takes a bunch of many other skills to found things and to operate an organization. I don’t think it’s a matter of those other people being less skilled. It’s a matter of people being skilled in different areas.

COWEN: Do you think you will ever found a for-profit in developing economies?

KREMER: I don’t have any plans to, but I don’t always have plans in advance. Who knows?

COWEN: What’s the first organization you founded or cofounded?

KREMER: The first organization was an organization called WorldTeach, which placed volunteers. I spent a year teaching a secondary school in Kenya. After teaching in Kenya, when I was living in Kenya, I’d be asked to teach in schools frequently. At that particular time — I think the situation’s really changed — there was a lot of demand for teachers, and there were lots of people in the US who were interested in volunteering. It seemed there was a role for something to match them together.

That was one organization. More recently, I guess a lot of people who found NGOs or found organizations sometimes go on to be funders or go from being founders of for-profit tech firms to being VCs. One organization I’ve been involved in for 10 years now is Development Innovation Ventures, which is part of the US Agency for International Development.

What Development Innovation Ventures does is, it helps invest in innovations for development. Sometimes those are nonprofit organizations; sometimes they’re for-profits. Sometimes they’re started by development economics researchers; sometimes they’re started by others.

One of the reasons why I was saying I don’t think this is something unique to me, is that when we’ve looked at our track record, we see a really strong record of success, and a lot of that is development economics researchers.

On advance market commitments

COWEN: Now, your current work on advance market commitment — why do you think that advance market commitment is more politically popular than patent buyouts?

KREMER: Maybe it’s useful for me to give a little bit of background on this. I had written an academic article many years ago, proposing a particular way that the public sector or the government could buy out patents and then put the intellectual property in the public domain. The idea is, this would provide a way of rewarding innovators without some of the distortions accompanying monopolies.

I liked the idea academically, but it didn’t really go anywhere politically. I was thinking very much about pharmaceuticals. This is speculation, but I think the pharmaceutical firms — they’re obviously very attached to the current system of intellectual property. They may have felt that if they sold their patents and put them in the public domain, that would weaken the legitimacy of that system, where it was just very different than what they were used to, whereas the idea of selling their products wasn’t so much.

You can avoid some of the distortions of monopoly. One of the key distortions of monopoly is that when the price is high, when there’s monopoly pricing, less of this gets sold. When less of this gets sold, that’s harmful for many products, but in the case of pharmaceuticals, that can be very harmful to human health.

So the approach was, commit in advance to buy the efficient quantity for some products, in particular vaccines. It’s fairly easy to know roughly what the efficient quantity is in advance. Commit to buy that in advance, and then that can both create an incentive for vaccine development and the efficient quantity to be purchased.

In the case of many childhood vaccines, the efficient quantity is enough to vaccinate pretty much the entire birth cohort, obviously. In practice, you can’t reach quite everybody.

That idea did turn out to be politically acceptable, and a number of countries, together with the Gates Foundation, pledged $1.5 billion to finance the purchase of a vaccine against a disease called pneumococcus. In particular, there was already a vaccine for the strains of pneumococcus that were common in the developed world. But no firm had yet developed a vaccine against the strains of pneumococcus that are common in the developing world.

So they pledged that if firms did develop that, they would purchase it or they would help finance the purchase. That led to three different pneumococcus vaccines that have now been developed against the strains that are common in the developing world. Hundreds of millions of people have been vaccinated, and 700,000 lives are estimated to have been saved by those pneumococcus vaccines. I think that’s a tremendous example of a success.

Often, when we think about founding, we think about founding organizations. That wasn’t the case of founding an organization, but it was a case where a lot of issues had to be thought through — the economics issues, but many other issues as well. I don’t want to say the policy took off after that. The policy was adopted in the case of pneumococcus, and then there’s been additional work to try to use this in other settings as well, and a few other examples.

COWEN: The political economy and legitimacy issues aside, is the main micro-difference between patent buyouts and advance market commitment simply price versus quantity and certainty? Or is there some other strategic behavior that might make advance market commitment better? How do you think about that tradeoff? Let’s say you didn’t know what was the quantity that you needed.

KREMER: I think it’s very hard to use advance market commitments for innovations that you don’t even know that you need ahead of time. In the case of, let’s say, a malaria vaccine, we can say, “Look, we know we need a malaria vaccine.” And you could potentially say we have a process — FDA trials or the European equivalent or others — for deciding when something has passed that goal. Then you can use an advance market commitment.

You think about the opposite case — let’s say the Post-it Note. We didn’t know we needed the Post-it Note until it was developed. So it’s very hard to have an advance market commitment or prize or something like that to incentivize the development of innovations that we never knew we needed till they were invented.

I personally think it would be worth seeing if we could explore patent buyouts as well. That’s a different tool for encouraging innovation that’s worth considering.

COWEN: Do you think your ideas on advance market commitment can improve, say, how military procurement is done today? Or do you view that as another example of advance market commitment? Is there something special to doing something in very limited doses? So, malaria vaccine, COVID-19 vaccine — there’s a hit-and-run approach. You enter; you leave.

It’s hard for outside parties to game the system very much, but if you do it over decades, there’s Eisenhower’s military-industrial complex and the big mess that you get with high costs, a lot of overruns. Is that AMC? Or is that somehow people not heeding your work?

KREMER: Let me break that into pieces. One question that you raised is, if you do this over time, maybe people learn to game the system. I think that could be true, but I also think that the system designers will learn over time how to do it. That’s one big insight from the experimental method. It’s not just the experimental method, but we’ve rediscovered it there, which is something that every founder, every innovator knows, which is it’s not a one-time event.

You need a continual cycle of testing, adapting, coming up with a new version, et cetera. There are a lot of details on how you implement the advance market commitment. I don’t want to claim that the pneumococcus advance market commitment got everything right. It probably could be improved over time. Hopefully, they’ll be learning on both sides of the table.

On this issue of the political economy, without going to the military example, let me say one thing that I am concerned about with the advance market commitments, which is, I think there’s a tendency . . . When Rachel Glennerster and I wrote a book on advance market commitments, proposing the idea, we focused on malaria.

I was involved in co-chairing a working group at the Center for Global Development. That working group recommended both a technologically distant target like malaria and a technologically closer target, pneumococcus. What the political system went with was that technologically closer target, pneumococcus.

I think that partly is due to political economy reasons. The politicians want a quick win. The bureaucrats want a quick win, and their firms or other entities that are going to be pushing the things that they think they’ve got a technological advantage on — they’re technologically close.

I think that is an issue. Personally, in terms of the economics, it would make sense to really try a technologically distant product. There are some advantages of advance market commitments for technologically distant products. That’s the political economy issue that I would say is — I don’t want to claim it’s the only one — but that’s the one that’s most prominent for me.

COWEN: Is there something you would nominate that’s specific? Solar-powered space stations, a trip to Mars, or what do you want to do?

KREMER: Well, look, I would love to have, say, a malaria vaccine. I think it would be a great example. Let me give you another example. I’m going to give examples that are from the world of development economics, but I completely agree with you — this could be used in many other settings as well.

A lot of people in developing countries cook with a small fire in their house, and that creates incredible indoor air pollution. That’s estimated to kill millions of people a year. It also produces not just carbon emissions, but some chemicals that are actually much worse for global warming than that per unit volume. There have been efforts to develop cleaner stoves. Now, those haven’t gone that far, in part because they haven’t found something that consumers want to use.

That’s actually a great case for an advance market commitment because part of the idea of an advance market commitment is, you have to meet certain technological thresholds, but you also need to get customers willing to put up something. We could say that if somebody produced a cleaner stove and if consumers were willing to pay $10 for it, then there could be an additional $20 or $30 worth of subsidy, for example. Details would have to be thought out, but that would be another example.

Sure, we could have things to reduce recidivism rates. We could have programs to address educational challenges. There’s a whole host of areas in high-income countries or in the United States where there are technological challenges that, for whatever reason, are not attracting quite the level of effort that tech companies are putting into other fields, and if we want more innovation, we could promise to reward them.

COWEN: If you think the market risk premium is off for very large investments for biomedical research, do you think the same is true for smaller investments? That is, what are the micro-foundations of what goes wrong with the market risk premium and biomedical investment?

Does it plague all venture capital firms for biomedical ideas, or just the big stuff because it’s a liquidity issue? So for vaccines, we need advance market commitment, but for ordinary investments, venture capital will work just fine? What’s your view?

KREMER: There may be issues around risk. That’s not necessarily where I would go first in terms of the market failures associated with vaccines. The first market failure associated with vaccines is just a straight communicable disease.

If we thought about individuals buying vaccines on their own — when I take a vaccine that would benefit you if you were closer or nearby, but indirectly, it even benefits you. Because I don’t infect somebody, they don’t infect somebody, and eventually, they don’t infect you.

COWEN: If you could subsidize only the buyers, you could have vaccine vouchers. Let the production side take care of itself and capture the externality that way. But presumably, there’s something wrong with the generation mechanism.

KREMER: It turns out there’s a whole host of distortions here. We do subsidize vaccines. Governments do a lot of the purchasing, but then there’s a commitment problem, at least in some countries. Governments aren’t necessarily going to pay the full social value of the vaccine. They may pay less than that. If that’s the case, then firms might not be adequately incentivized. They may be incentivized, but maybe not adequately incentivized to produce the good.

If we think about something like malaria, for example, there are many different governments involved, for example, many small African countries. No one of those countries has an incentive to say, “Well, let’s pay a price that would generate the socially appropriate incentives for developing a malaria vaccine.” So there’s a global public-good problem there in addition to everything else.

There’s quite a few distortions in vaccine markets. In general, however, we don’t think that innovators capture that much of the value. It’s a fairly small fraction. People who’ve estimated the social rate of return to private investment and R&D often estimate numbers like 70 percent. I don’t think it’s limited to vaccines — that might be an extreme case.

But I would actually be in favor of trying to boost R&D incentives in a wide variety of fields, not just some of the ones that we’ve mentioned so far.

COWEN: I’ve seen estimates — they’re actually from one of the groups you founded — that a deworming pill could cost as little as 50 cents a year per person in many parts of Africa. So why isn’t deworming done much more?

KREMER: You could say the glass is half empty, you can say it’s half full, or you can say it’s almost three-quarters full. I think it’s about three-quarters full. When I first got involved in deworming, it was testing a small NGO program. We found phenomenal effects of that. The original work found health gains and education gains. Now we’ve tracked people over 20 years, and we’re seeing people have a better standard of living or earning more.

Following the early results, we presented the results of the government of Kenya to the World Bank. Kenya scaled this up nationally, in part with assistance from the World Bank, primarily just in conveying some of that information.

Indian states started doing that, and then the national government of India took this on. They’re reaching — a little bit harder to know the exact numbers — but probably 150 million people a year. Many other countries are doing this as well, so it’s actually quite widely adopted.

COWEN: But there’s still a massive residual, right?

KREMER: That is for sure.

COWEN: What’s your best explanatory theory of why the residual isn’t smaller? It would seem to be a vote winner. African countries, fiscally, are in much better shape than they used to be. They’re more democratic. Public health looks much better. The response to COVID-19 has probably been better than many people expected, say, in Senegal, possibly in Kenya. So why not do deworming more?

KREMER: The people who have worms are pretty poor people. The richer people are less likely to have worms within a given society. Richer people are probably more politically influential.

There’s also something about worms — they gradually build up in your body, and one worm is not going to do that much damage. The problem is when you’ve got lots of worms in your body, and even there, it’s going to take time.

I’ve had malaria. I don’t think I’ve had worms. I hope I haven’t. When you have malaria, you feel terrible. You go from feeling fine to feeling terrible, and then you take the medicine. You feel great afterwards. With worms, it’s much more like a chronic thing, and when you expel the worms from your body, that’s sort of gross. I don’t think, even at the individual level, do you have quite the demand that would be commensurate with the scale of the problem. That’s a behavioral economics explanation.

I think there are political issues and then there are behavioral issues. I would actually say that a huge, huge issue . . . This sounds very boring, but this falls between the Ministry of Health and the Ministry of Education, and each one of them has different priorities. The Ministry of Health is going to be worried about delivering things through clinics. They’re worried about HIV and malaria, tuberculosis, as it should be.

The Ministry of Education — they’re worried about teacher strikes. It’s very easy for something to either fall between the cracks or be the victim of turf wars. It sounds too small to be, “How can that really get in the way?” But anybody who’s spent time working in governments understands those things can very easily get in the way. In some ways, it’s surprising how much progress has been made.

Here’s one way the political economy works in favor. You mentioned democracy — I think that’s a factor. I actually find — and I don’t want to be necessarily a big fan of politicians — but in some ways, politicians hear how much this costs, and they think they can affect that many people for that small amount of money, and they’re like, “Hey, I want to get on that. Maybe this is something I can claim as an achievement.” We saw that in Kenya. We saw that in India.

On economic growth

COWEN: If I think of Michael Kremer on the issue of economic growth, I see at least three strands. The first is the very famous population paper, where there’s increasing returns with market size. The second is your — I think — 1993 JME paper with Larry Summers, “Good Policy or Good Luck,” where growth rates vary a great amount, and a lot of that is due to luck or positive real shocks. Then there’s the work on education, which would seem to imply growth is pretty stable over time because human capital doesn’t change that rapidly. The current Michael Kremer of 2020 — how do you put that all together and think about growth in the best possible model? Where do you stand now?

KREMER: One insight from the study of growth is, a fundamental driver of growth in the long run is technological change. One thing that we economists should be thinking about more is, what are the institutions to increase the rate of technological change and to try to direct it to human needs? That’s one of the reasons why I’ve been excited about advance market commitments, patent buyouts.

I think we’ve got a set of institutions, primarily intellectual property institutions, but also research-funding institutions. We should be experimenting and try new approaches to this and try to improve things over time. I would say that would be one key message about economic growth.

Another paper that I’m working on recently with Jack Willis and Yang You — it looks at something that . . . When I was in graduate school in the ’90s, I studied economic growth. Robert Barro was my adviser. One of the things I learned is that there’s not unconditional convergence. Some models predict poor countries would catch up to rich countries, and maybe they do if you take countries that have similar policies or similar human capital, but overall poor countries don’t catch up.

I was teaching my development economics class a few years ago and thought, “Well, I better update my slides.” We looked at that again, and that’s totally changed. Actually, after the ’90s, this pattern changed. Now, poor countries are catching up to rich countries.

This isn’t just a matter of India and China. This is a broad-based phenomenon. That’s something that I think is very exciting and very encouraging in a world where, within certain countries, there’s been increases in inequality. If you look at a global level, there’s also been this tremendous lifting of people out of poverty, and that’s very exciting.

COWEN: Don’t you find it puzzling that the global economy is really much larger than 40 or 50 years ago, but total factor productivity or measured rates of innovation seem to be down? The pharmaceutical pipeline, at least pre-COVID, is commonly described as drying up or slowing down, and yet there are so many more consumers out there for drugs.

Does this in some way invalidate or raise doubts about increasing-returns-to-scale theories, that maybe market size doesn’t matter very much?

KREMER: The view that I took in my paper on the ’90s on this was . . . I was a near contemporary of Chad Jones, and I think Chad’s fundamental insight . . . Let me put it technically first. “Chad, what’s the production function for new technologies?” Clearly, having existing technologies helps you invent new technologies. Having calculus helps you invent spacecraft or aircraft.

But on the other hand, it’s not clear that it helps you, that an improvement in existing technology helps you, one for one, in developing new technologies. In some sense, it’s getting harder and harder to invent new technologies. Chad’s work at the time showed that, and more recent work that Chad has done.

In that world, the task of technological advance is getting harder and harder over time. We see that at the macro level. We see that at the micro level. We see that in the increasing size of teams needed to develop new innovations. The task ahead of us is getting harder, and that means that there’s a race between the total resources that are going into innovation and the increasing difficulty of the task. You have those two forces operating against each other.

I’ve no doubt that the increasing size of the market and the fact that there are now many people who are contributing to technological change is helping us, but the issue is, we’ve got a harder and harder task ahead of us.

COWEN: If there’s some constraint out there limiting the benefits from larger market size, what is indeed the constraint you would wish to target? You’re Michael Kremer. You’re put in charge of everything, and you can target the constraint that’s slowing down growth. What’s the first thing you look at?

KREMER: The first thing I would look at is the thing that we have the most control over, which is the institutions we use to try to advance technology. If you believe Chad — and I do — there’s just something fundamental about the progress of technology, which means that it’s getting harder over time to have a constant growth rate of technology. We invent more technology. We don’t increase the growth rate over time. That’s just something inherent about technology. Obviously, that could change.

It could be that when we get AI working well enough, we actually get to an incredibly virtuous cycle, and then we head off to a singularity where the AI gets smarter and smarter over time. We don’t know whether that’s a fundamental constraint, but whether that’s a constraint or not, there’s not much we can do about it. What we can do is, we can think about the myriad institutions.

We talked about IT institutions. There’s also research-funding questions. Should we be funding research the way we do? Or should we be saying, “Look, it’s not a matter of writing the grant proposal. Once you’re five years past the PhD, we’re going to fund you based on what you’ve accomplished, not based on your grant proposal.”

That’s one example. I have no idea what the answer is to that. Should the committees be a different size than they are now? Should we say that we’re going to fund something as long as somebody really likes it rather than take the average of the committee view when we’re thinking about government grants?

There’s a bunch of questions about what the optimal way to do research funding is, and I think we should be experimenting with this, trying to learn more about how do we set up institutions. There’s questions about research publication process. That’s the thing that we can control, and that should be a priority for anybody who cares about economic growth.

COWEN: Let’s say the current Michael Kremer sets up another high school in Kenya. What is it that you would do that the current high schools in Kenya are not doing? What would you change? You’re in charge.

KREMER: Right. We’ve learned a lot in education research in recent years. One thing that we saw in Kenya, but was also seen in India and many other places, is that it’s very easy for kids to fall behind the curriculum. Curricula, in particular in developing countries, tend to be set at a fairly high level, similar to what you would see in developed countries.

However, kids are facing all sorts of disadvantages, and there are all sorts of problems in the way the system works. There’s often high teacher absence. Kids are sick. Kids don’t have the preparation at home, often. So kids can fall behind the curriculum.

Whereas we’ve had the slogan in the US, “No Child Left Behind,” in developing countries, education system is focused on kids at the top of the distribution. What’s been found is, if you can set up — and there are a whole variety of different ways to do this — either remedial education systems or some technology-aided systems that are adaptive, that go to where the kid is . . . I’ve seen huge gains from this in India, and we’re starting to see adoption of this in Africa as well, and that can have a very big impact at quite low cost.

Another thing is — I think it’s called structured pedagogy, which is just providing very detailed lesson plans for teachers. That can be done in a variety of ways. We’re seeing big effects from that as well. The good news here — as in a lot of development — is, if we try a variety of things, and if we think about it using theory and using the tools of economics and tools of other disciplines, we can really make progress.

COWEN: Does it work just to pay the students for better grades?

KREMER: I’m not an expert on that type of work. Let me give you two contrasting results. I think some of Roland Fryer’s work on that — if I remember right — in the US wasn’t so encouraging.

But I’ll tell you one thing that I am reasonably excited about. I’ve done some work with Rebecca Thornton, Willa Friedman, and Ted Miguel on scholarships, where a lot of developing countries — you have to pay something, at least to go to secondary school if not primary school. Even primary school — it’s officially free, but there are often fees.

There is a program in Kenya that said if kids did well at sixth grade, their fees would be covered for seventh and eighth grade. These were in poor schools, so you know this was going to poor kids. What we saw was, obviously, there’s a direct benefit of allowing more kids to go on to the higher grades of school.

But we also saw that kids were putting in more effort in response to this, and not just the kids at the top of the distribution, but throughout the distribution. It also looked like teachers were putting in more effort, as well, because they were motivated to try and help their students do better.

So I think having some sort of merit scholarship program is something . . . There are some flaws in our paper. I don’t know whether we should scale up. Maybe we should, because saying that kids who do well in school can go on to the next level of education doesn’t seem that controversial, so maybe we should scale it up.

I don’t want to claim the evidence is totally dependent upon it. We could test this more, learn about how to make it work, but this is the type of program that also deserves . . . Probably, it should be part of a system and certainly deserves more testing.

COWEN: One strand of your work concerns peer effects. There’s your paper on the Hajj. Influence on college roommates — your paper on alcohol use. If peer effects are so strong in the United States — again, say, in high school — how should we restructure how we treat honors classes? Should we have more of them? Fewer of them? What’s the right way to think about what you might call intellectual segregation in this country?

KREMER: Yes. It’s a great question. Great question. I don’t know the answer to the question, but I think we should try and find out. I believe there have been some results. I’m not super familiar with the papers, but I believe there’s been very positive results from peer mentoring programs. I believe that some results in Italy very recently for this — if you can create incentives for students to help each other out, that can be very positive.

This does not seem to be a problem in most developing-country education systems I know, but certainly, the US can have a system where it’s not cool to do well in school, can have a culture like that — or not cool to do too well in school. I don’t see that in Kenya, and I don’t hear about that in India. Thinking about how can we set up a system of incentives and rewards to try to address that is a really interesting question.

Despite saying this isn’t even a problem in Kenya, two students, Ronna Jane and Brandon Tan, have some, I think, very preliminary . . . More work needs to be done on this, but they have some very exciting results for Kenya.

You were talking about paying kids to do well as individuals. They’ve been trying something at that level in the classroom as a whole, where they reward the classroom as a whole for doing well. You can imagine that can possibly create a culture in the class as a whole, where you want each other to do well. The initial results on this are actually extremely promising.

COWEN: Here’s a few questions from outside of economics. Ugali is arguably the Kenyan national dish. Does it taste good?

KREMER: No, [laughs] not to my taste. There are Kenyan dishes that do. There’s a mixture of corn and beans that I like. I have had ugali that was delicious in fancy restaurants in Nairobi that Isaac Mbiti took me to, but in general, it’s a little bland.

COWEN: Why doesn’t Harvard Square have food as good as that you can get in Kenya? A lot of purchasing power, obviously high level of education, but it’s terrible.

KREMER: I don’t know if this is the case, but I think transient populations that turn over a lot may be associated with worse-quality food, but I’m not sure.

COWEN: Out of Africa, the movie — do you like it? Is it any good?

KREMER: Oh, gosh, so long ago since I saw it. I don’t remember it. I don’t remember the movie at all.

COWEN: What’s the ideal Kenya trip? Say I come to you. I say, “Michael, I’m going to Kenya. I have two weeks. Obviously, safari, but what else should I do?”

KREMER: I’m very happy to be a tourism ambassador for Kenya. It’s really a great country to visit. If you’re into mountains, there’s Mount Kenya. There’s the coast. I think you’re totally right that, obviously, the top thing for Americans would be a safari because here you can get mountains and you can get beaches, other places.

The coast of Kenya — it’s not just the beaches. It’s the Swahili culture that’s a very interesting blend of cultures there. But game parks have to be very high on the agenda for people.

Nairobi right now is a cosmopolitan, exciting city. It’s a happening place. I didn’t feel that way about it in the late ’80s or early ’90s, but now it’s a very exciting place to be. If you’re talking about where you can get good food, you can get good food of a lot of different types in Nairobi.

COWEN: Not Harvard Square? [laughs]

KREMER: No. [laughs]

COWEN: What’s interesting about reading James Joyce?

KREMER: Yes. I was just rereading Dubliners recently.

COWEN: It’s also about development economics.

KREMER: It is. It is. It’s about a place that was a colony, and trying to get out of that situation, and with people living very constricted lives. Look, I don’t want to claim this is the most interesting thing about reading that. I was talking about Nairobi being a happening place. Well, I was in Ireland recently. That’s a happening place as well. It’s not the Dublin of Dubliners anymore.

I was talking about convergence recently. Things are really improving in a lot of the world, and it can be very easy to forget that in the time of COVID, a time of all the other problems that we’re having right now. People are freer, and people are living more exciting, more open, less constrained lives.

COWEN: What’s your favorite novel?

KREMER: I don’t know if it’s a novel, but maybe The Odyssey.

COWEN: Why The Odyssey?

KREMER: There are so many rich, archetypal stories in there. It’s hard for me to say why, but when I’ve gone back to it, I really enjoy it.

COWEN: It’s, again, a comparative development tale, right? Autarky doesn’t work; war is bad for economic growth; don’t take too many drugs.

KREMER: [laughs] Good point. I hadn’t thought of it that way, but yes.

COWEN: Literature as a whole, arguably, makes one believe more in what would sometimes be called a cultural backwardness theory of economic underdevelopment.

There’s something in Dubliners about the perspectives people have on situations. They’re remarkably local, and they’re keen, and they’re perceptive, and they’re highly emotional, but the arc at which people make comparisons, typically, is a very small circle. Ireland, when it joins the EU and even before then — it’s changing that, I think.

KREMER: Yes. It’s a great point. As an undergraduate, I didn’t study economics. I did a program called social studies. We read Weber among other people. Weber talked about the role of culture in influence and economics. A lot of modern development economics was a reaction to that, perhaps a warranted reaction.

I think Weber himself, obviously, may be wrong, but very sophisticated, not simplistic. But there was a lot of simplistic analysis of “Confucian societies will never grow.” Or “Hindu rate of growth is inherently low.” So people were making, clearly, statements that were just wrong. It turned out to be wrong.

Even before the head economist said, “Look, we don’t need to” . . . the founding of modern development economics would say that the people are poor but rational. We can use the same tools of economics to analyze people in low-income countries as high-income countries. I think there’s a lot to be said for that.

But what we’ve realized with behavioral economics is that people in high-income countries are not completely rational either, and there are certain patterns in the way that they’re not rational. Those things can play out differently in low-income countries and high-income countries just due to different institutions.

But what you’re suggesting is something much stronger than that — that there can be certain ways of thinking embodied in culture. Those are interesting ideas, and people like Nathan Nunn or Alberto Alesina are doing very exciting work on culture and economics. We have a lot more to learn about how far we can take that, but I think there’s exciting work going on in that area. We have to be careful and avoid some of the simplistic approaches of the past.

I also think — just to come back to this point about convergence — one of the things a lot of the work on economic history, and on culture in particular, has emphasized is just how long-lasting things are. Whether societies used to plow thousands of years ago affects society today. That’s the claim.

I’m very open to those ideas, but what we’ve also seen is low-income countries really catching up — and remarkably quickly — to higher-income countries. That suggests that we shouldn’t take talking about culture as indicating that there’s no possibility for development or for change. There’s clearly that possibility. We’ve seen that in country after country.

COWEN: In your own intellectual development, what did you learn from your mom?

KREMER: My mother wrote about literature on the Holocaust. One thing that brings out is that when there are issues that we can do something about, we need to do something about them. We shouldn’t wait. Obviously, there were people who did things at very great cost to themselves in that situation.

And they, right now, in COVID, are doing things. They’re risking their lives to help people get through the COVID crisis. I work on international development. That’s at least in part because I feel like this is something — at actually fairly low cost to ourselves — we can help address these problems. The case of deworming — this is something that costs 50 cents a year to address. I guess those values did influence me.

COWEN: At this point in your career, why move to University of Chicago, as you’re doing this fall, basically as we speak?

KREMER: Right, I’m speaking to you from Chicago.

COWEN: They let you into your office?

KREMER: I’m in my office.

COWEN: That’s remarkable. I didn’t know that was allowed.

KREMER: Yes, I know. We have to go through certain precautions, but I’m speaking to you from my new office at the University of Chicago. Chicago, obviously, has a wonderful tradition in economics. One of the things that I was very excited about was to talk to President Zimmer here, president of the University of Chicago.

You talked about experience as a founder, and I’ve talked about technological change and the need for changes in technology, by which I mean gadgets, but I don’t just mean gadgets. I mean institutional change, innovation, and policies as well. I think what we’ve seen is, there’s a lot of potential to generate that.

University of Chicago made a big commitment to development economics in general. I hope we’ll be hiring more people over time, and certainly the slots will be there for that, and to create a development innovation lab, which will use the tools of economics to create practical innovations that can improve people’s lives, and to work closely with partners to develop and refine and iterate on those innovations so that they can be scaled up.

What I’ve been seeing in my own research, what’s true with a lot of research of J-PAL Poverty Action Lab at MIT, and what I’ve seen in the USAID’s Development Innovation Ventures is, when you invest in that type of technology — anytime you’re investing in innovation, most innovations don’t succeed. They don’t take off, but you get a few.

If you’re a venture capitalist and you invest in a Google or Facebook, it’s fine if a lot of your other investments don’t work. I think the same thing in investing in development and innovation.

When we did an analysis of the rate of return, we tried to get a lower bound on the rate of return for the early investments at Development Innovation Ventures at USAID, and we’re getting a 5:1 lower bound. It’s a super conservative lower bound. I think we’ll have a much higher rate in our next iteration of this.

Just to be clear what a lower bound this is, we had 41 innovations in our early portfolio. We couldn’t quantify the benefits for all of them, but of the 9 or 10 that reached over a million users, we could quantify the benefits for 4.

If you take the benefits of those 4 and set them against the cost of our investments in all 41, the ratio is 5:1. And I think that that will grow over time. I just think there’s a huge, huge reward to invest in innovation. There’s also a scientific reward. University of Chicago’s making a big commitment in that area, and that’s a big part of why I came here.

COWEN: You were tempted by the ability to found a new institution?

KREMER: In part, yes. That was a huge motivation for me, and I’m very excited about it.

COWEN: What is it you think that you — as someone who studied organizations — know about managing institutions that maybe just regular managers would not know?

KREMER: I think that if you’re working in modern development economics — particularly if you’re using the experimental approach — some of the things that the stereotype of economist is not paying attention to, you’re forced to pay attention to. Let me give you a few examples.

Part of the reason why I originally got into doing experiments was to isolate causal impact, to disentangle “Is it a particular program that’s causing something, or particular intervention? Or is it something else?” Experiments are very good for that, but when you’re doing experiments, you’re typically working on a concrete problem because that’s what organizations do. They’re doing the experiment because they want to solve that problem.

You have to get to know the people who work in that institution. You also have to get to know the ultimate customer. One of the things that business books will tell you is, know your customer. That means, yes, you want to learn the statistics on the customers, but you also want to talk to the customers.

Development economists will typically — if it’s an education program they’re studying — they’ll talk to kids. They’ll talk to the teachers. They’ll talk to the school principals. You get in the habit of doing that, and that’s very useful for innovation more broadly.

Another thing is, as a researcher — and this would be true for any type of researcher — but you’re presenting your research in seminars, and you’re getting feedback on it. You’re getting criticism on that, and you’re going back and revising things. I think that’s actually very useful for management.

One of the problems that managers can have is, you can convince yourself — and this might be particularly true in development — you can convince yourself that what you’re doing is the right way to do things, the only way to do things, and you may not get feedback otherwise. In research you do, and that’s a great thing about academics.

One thing about modern development economics and experimental economics is, it’s iterative. You try one thing. Then either the same researchers or other researchers try to improve on it again, and it’s a constant cycle. That’s, again, something that your management books will tell you is very important.

You also get used to working with people. You have to raise money. You have to work with NGOs or governments or firms to talk to them about randomizing. There’s a whole bunch of skills that actually are applicable beyond just writing a paper.

If I think about GiveDirectly, an organization — it’s a very successful NGO, but founded by a development economist, Paul Niehaus. But there are many, many others who are doing things like this. Again, in many other fields, this wouldn’t seem like a strange combination, being a researcher and also being a founder.

COWEN: Final question. When it comes to advance market commitment, what is the very next thing you will be doing?

KREMER: I’m doing a lot of work on the problem of COVID. Some of this work is with people who are involved in the pneumococcus advance market commitment — Chris Snyder, Jin Lee, others. But we’re also working with some other industrial organization economists, for example, Susan Athey.

When we think about this problem, the problem of getting a COVID vaccine is different in some important ways. That’s, again, something that economics and being close to research, in theory, will tell you. Just to be clear — sometimes people have a stereotype of experimentation as opposed to theory. I think they’re strongest when they’re linked.

There’s a number of different things we’ve suggested, but one of them is — and this is very much with your colleague, Alex Tabarrok — very simply is, the world economy is losing $500 billion every month due to COVID, and that means that it’s worth investing a lot in trying to develop a COVID vaccine, even though there’s a pretty high risk of failure, and it’s also worth getting lots of shots on goal.

Now, for a variety of reasons, which I could go into, the private returns from building out a very large factory for a vaccine before we know whether the vaccine is going to work — before testing is complete — are much less than the social returns.

In this particular case, it probably makes sense for governments to offer to cover — you probably want a mix of push incentives and pull incentives. Some reward for success, but also just some up-front financing for building those factories ahead of time.

That’s what we’ve been arguing for, and a bunch of countries are doing things like that — so, the US Operation Warp Speed — but many other countries are putting in orders for vaccine even ahead of the time the vaccine’s been proven to work. There’s obviously risk involved, but it’s very much a risk worth taking.

There are other lessons that are coming out of the industrial organization of this, like it’s very important to invest in supply-chain capacity, not just in the production of the ultimate vaccine. We might want to consider ways to get some of the early information from the trials out, so we can use it even before the trial is finished. We can use it to invest optimally in manufacturing.

So a lot of insights from studying the economics of this, and that’s obviously a critically important problem. And policymakers seem . . . Again, you could say it’s a glass half full or half empty. I would say the glass is at least partially full. I’d like policymakers to be investing more in this, but they are starting to invest, and they’re starting to do these contracts well in advance of availability of the vaccine.

COWEN: Michael Kremer, thank you very much.

KREMER: Great, thank you.