In this episode, Shruti speaks with Lant Pritchett for a second time. They discuss internal and external migration, the concept of open borders, definitions of poverty, the flaws of randomized controlled trials and much more. Pritchett is a development economist from Idaho. He is currently affiliated with Oxford’s Blavatnik School of Government as the research director of the RISE Programme, is the Research Director at LaMP (Labor Mobility Partnerships) and is a fellow at the London School of Economics. He previously worked with the World Bank from 1988 to 2007, living in Indonesia 1998-2000 and India 2004-2007. His publications span a wide range of development topics including economic growth, state capability, education, labor mobility and development assistance.
SHRUTI RAJAGOPALAN: Welcome to Ideas of India, where we examine the academic ideas that can propel India forward. My name is Shruti Rajagopalan, and I am a senior research fellow at the Mercatus Center at George Mason University. Today my guest is Lant Pritchett, who is the research director of the RISE Programme affiliated with Oxford’s Blavatnik School of Government, is the research director at LaMP [Labor Mobility Partnerships] and a fellow at the London School of Economics. We talked about his research on labor migration and its benefits, NIMBYism, a global prosperity line to replace the dollar-a-day poverty line, his reading habits and more.
For a full transcript of this conversation, including helpful links of all the references mentioned, click the link in the show notes or visit mercatus.org/podcasts.
Lant, thank you for coming back on the show. One of the reasons we’re having you back is because we couldn’t get through all the papers that you’ve written last time, especially some of my favorites. One of my favorite papers by you is one that you’ve co-authored with Michael Clemens and Claudio Montenegro. It’s called “The Place Premium.”
The Place Premium
RAJAGOPALAN: What is a good way of thinking about the place premium within a country? Because what you really look at is barriers to movement across international borders. You find that this is one of the largest remaining price distortions in any global market. In fact, from some places like Nigeria, you can have up to an 8X difference in wage premium. If you think about a country like India, which is fairly large in size, India has effectively had open borders for the entirety of its history.
In the last census that was conducted in 2011, about 50 million economic migrants were counted. I think that’s an underestimate, but it’s still useful. In the 2017 economic survey, they placed it at about 100 million. Now, if you think about India, and you’re very familiar—there are places like Bihar, which are at the GDP per capita level of a country like Haiti. Then there are places like Goa, which are at the GDP per capita levels of Jordan. Why do so few people move around, even within the borders of India where clearly these kinds of international barriers don’t exist, but there is a wage premium?
LANT PRITCHETT: I think there’s a couple of solid things and a couple of conjectures. One is, people much prefer to live where they were born, around people they grew up, around their family. Even in equilibrium, we would expect a pretty substantial wage premium, particularly across a place like India that’s culturally heterogeneous—in the sense that the culture of Bihar is not the culture of Goa, it’s not the culture of Kerala, it’s not the culture of Tamil Nadu. I think of India like I think of Europe; it’s at least as culturally heterogeneous across India as it is across Europe.
Even if you think of India as having a common Hindu religion, by living in Tamil Nadu for a year, I learned it’s not exactly the same Hinduism in a lot of ways. First of all, I think there’s bias. And when we tried to look at culturally different but legally integrated markets, you get wage premium differences of identical individuals as large as 50% in apparent equilibrium. You should expect some pretty large equilibrium.
Second, to get from the comparison you’re talking about of GDP and average incomes—the whole point of “The Place Premium” was to essentially, if you think of a wage surface or a wage function that relates wages to individual characteristics, you need to drill down to identical-productivity individuals. To some extent, areas like Bihar are lagging in part because they have underinvested chronically in education, and places like Kerala are in part rich because they have invested in education.
Well, a large part of the difference in average incomes is differences in the characteristics. You’d have to drill down and do the econometric exercise like we do of estimate the wage function of the two places, and then equate the individuals, and that’s going to account for something. That’s pretty solid.
The conjecture I had, and this is part from research I did together with Chandra Bhan Prasad, who as you may know is a very flamboyant Dalit activist, who I enjoy very much. I actually went out with him, while I was doing the research, to visit his house in Uttar Pradesh.
The second thing is that I don’t think land markets in India work very well, if at all, in two senses. One—and I think Mark Rosenzweig a long time ago had a paper like this—the heterogeneity inland is very high in hard-to-observe ways. People own given parcels, and the parcels depend on really, really specifics. I think people are reluctant to sell their land.
Second, given the weakness of the legal environment in India, I think people feel they need to stay on their land to protect their claim to it. I think there’s a lot of partial migration to where somebody needs to stay on this plot that this family owns because if we leave for a year or two years, we can’t really expect that someone won’t essentially expropriate it from us. When we looked at, for instance, Dalit migration, there was a lot more migration among Dalits than among the land-owning castes.
Janis Joplin has an old rock song, “Freedom is just another word for nothing left to lose.” I think a lot of people in India in the rural areas have something to lose. Hence it freezes them in place because they can’t get a good return out of it and take the lump sum and move to the city. They can’t, as a family unit, as a census-counted unit, move to the city without just the risk of losing their property. Somebody has to stay at home.
I think you get a lot of fractional migration, meaning parts of the household move. My conjecture is the census radically undercounts that. I had heard when I was living in India in 2005 or so, there was a prominent Indian sociologist who was basically saying, “Look, if you go into villages in India, there’s nobody there.” The census might be undercounting migration by a factor of two.
If you had 200 million people in India that had moved, out of a population of 1.3 billion, that seems consistent with the true gains to mobility. They would be moving on a rotational basis because they wouldn’t give up in some sense their claim to the family household, the family plot.
It does mean then this leads to a whole different set of conjectures about India’s response to the changes since, say, the liberalization in 1991. I think the people who bet on staying in the rural areas made a big mistake. A lot of the divergence in inequality within India is the result of people just above, who owned a little bit of land and thought, “Well, I can’t give that up to move to the city.” Then just owning a little bit of land in rural India did not turn out to be a great long-term bet.
RAJAGOPALAN: I agree with you on the land market. Actually, it’s much worse than what you described. There are a lot of restrictions on the sale of land, and some of this goes back to colonial fears—people who will have to sell their land in a fire sale when there’s a drought. Actually, there are lots of rules about how farmers cannot sell their land to non-farmers or rich farmers.
There’s also all kinds of land ceiling laws, which prevent a family from owning more than X hectares of land. In most states, that number is high single digits or in the teens. You can’t quite have a rich family or a rich firm just buying up swaths of land and freeing these rural farmers who would’ve otherwise been urban migrants. Your land point is well taken.
PRITCHETT: They would’ve been household migrants, meaning as opposed to individual migrants. Like I said, with Chandra Bhan Prasad, we did a survey just of Dalits, a universal survey of districts. All Dalits in a district, not a random sample. 60% of the households had someone in the household in an urban area elsewhere in India. I think the amount of fractional household—I don’t know what to call it exactly—partial migration is very much higher. And then those don’t necessarily get counted by the census because it’s, “What’s your usual residence?”
Anyway, I’m 100% confident that you know more about the land market restrictions than I do. It’s very hard to get a fair price. It’s hard to sell it at all, hard to get what you feel is a fair price, hard to then take that lump sum and make a big bet on, and it’s also hard to acquire urban land, move to a city. The land markets in cities don’t work that well either. Let’s pull up our stakes, sell our stake in the rural area, move to the urban areas. It’s made really complicated by existing markets and regulations in those markets.
Have We Reached a Migration Equilibrium?
RAJAGOPALAN: It seems to me like you’re saying that whatever premia there may have been has been exhausted, given all these various frictions in land markets and housing costs in urban areas and informal sectors, transaction costs and things like that. Whatever premium there exists is now gone. We have the equilibrium number of people who would migrate, given the regulatory circumstances. Is that a fair way of thinking about what you’re saying?
PRITCHETT: I think that’s a fair way of thinking about what I’m saying. Because a lot of research I do is around migration, and with migration, it’s just coercion. If you try and move from India to the United States, someone with a hat and a gun and a badge stops you at the border, or tries to, asks for your appropriate documents and will physically turn you back and send you back if you don’t have that.
The premiums that can exist in international markets are just—or we would expect them to be huge because we actually see the process whereby these premiums are being maintained of highly restrictive, coercive restriction on movement across borders. Whereas within regions, there’s frictions in the labor market, friction in the land market. I would guess India is pretty near equilibria, and there aren’t these huge gains to releasing Biharis to go work in Tamil Nadu.
That said, there are huge flows. When I lived in Chennai, I was there as a trailing spouse, and my wife was working at the American school when they were building new buildings at the American School of Chennai. And a big question was what language was the job site language, because basically nobody on the job site spoke the local language.
PRITCHETT: Nobody spoke Tamil. Tamil wasn’t really even an option. It was between Hindi—because many of the construction managers were from the North—and the language of Orissa, where most of the workers were. And it was a big safety issue because imagine being on a job site where there’s four or five mutually unintelligible languages.
So there is massive migration going on, but I don’t think it’s in huge disequilibrium is my gut take—other than, again, the disequilibria that are sustained by both regulatory and market failures in the sense that the market really doesn’t work well.
RAJAGOPALAN: Now, I want to bring you to the United States, which is another huge subcontinent with open borders internally. In this case, there is far less heterogeneity. I know Massachusetts is very different from Arkansas, but is it? On matters of language and other things, there’s also far fewer transaction costs involved in something like selling a house in Massachusetts, buying in Arkansas or the other way around.
Yet in the United States, the mobility, especially among the millennials’ generation and the one after, has been declining compared to previous generations. Once again, has any premium from movement been completely exhausted, or are there other reasons why people don’t move anymore?
PRITCHETT: I really don’t have anything useful to say on this. I live in the United States—born and raised there—but I haven’t read up on this literature enough to really know. I know that it hasn’t been true of my family. Every generation has moved across states for five generations. My children don’t live where I grew up. I didn’t grow up where my father grew up. My father didn’t grow up where his father grew up. His father didn’t grow up where his father grew up.
We’ve been part of the mobility. And I understand that there’s research showing it’s gone way down, but I don’t understand it and why it is. The only thing I’ve read on this that I find very interesting is that Ed Glazer has a really interesting paper about: Suppose you’re not in the labor market, and you’re living on transfers of some kind. Then your main incentive is not to move to a high-productivity environment but to stay in a super cheap environment.
I think a lot of the really depressed areas in America—if you’re dealing with people whose labor market prospects are relatively low because they’re disabled in one way or another or on disability, and they’re living, in one way or another, off transfers—then the incentive is to stay in these really depressed places because housing is super cheap, and not move to a high-productivity area because high-productivity areas are naturally associated with high land prices.
The United States really does have some embarrassingly depressed regions of the country where economic prospects have declined. And yet I think people are rationally not moving away from them, in part because they might be trapped by housing prices, if your own only asset was your house and housing prices were depressed in your area. But second, they might be making optimal choices to respond to, if I’ve got a fixed income transfer, then I want to minimize living costs.
RAJAGOPALAN: I think another might be the economic homogeneity that set in post-globalization in the sense that a lot of the industry clusters moved out of the United States. And people tend to move to clusters, and there’s a particular skill which is associated with the wage premium in that particular cluster. And now we have very few clusters like that remaining in the United States.
Silicon Valley is an obvious example. Of course, it’s not an industrial hub, it’s more a tech hub, and even that benefit might be dissipating post the pandemic and work from home and so on. I wonder if that has something to do with it, an economic homogeneity that’s set in since firms moved to China and so on.
PRITCHETT: Yes, that sounds like a plausible and interesting conjecture, but like I say, I don’t have any special insight on this.
Open Borders and NIMBY-ism
RAJAGOPALAN: Going back to open borders globally, one of the things that your paper shows is that—one is the gaps are just extraordinary, what you find after you measure them. Of course, they’re different depending on the origin country and the skills of the individual in question, and you control for all of that. But one of the main takeaways from the paper is, allowing greater movement of labor can in some sense be one of the greatest anti-poverty programs that ever existed.
Now, let’s say we’re in an ideal world and we do allow for completely open borders, or more open borders than what we have right now. Are you worried about the level of NIMBY-ism that already exists is going to set off all the gains that we get from open borders?
PRITCHETT: I have lots of friends like Bryan Caplan and others who talk about open borders. I never talk about open borders; I talk about greater labor mobility. And the reason I talk about greater labor mobility is I think there’s a big tension facing the world. The rich societies are just aging in really extraordinary terms. So people talk about low population growth, but low population growth isn’t the issue. The issue is the inversion of the demographic pyramid.
And the inversion of the demographic pyramid is creating societies that just have way too few workers relative to the number of retirement-aged. And it gets worse and worse inevitably over the next 30 years. Now, the difficulty is that most of the way the regime for mobility of persons around the world has worked since the 1920s is that people who are allowed to work in a country are either citizens or on a path of citizenship in the country.
I’m actually a big advocate of separating those two things and saying the needs of U.S. or Germany or France for labor are not being met. Because if the only way in which a person can come and work in France—to take care of the elderly or perform relatively low-skilled services—is by allowing that person to become a French citizen, the political consensus is no. We’ll prefer not having the service.
I don’t know if you remember that scene from “The Matrix” where Neo encounters the architect, and he says, “Look, we’ll wipe humans out.” And Neo says, “Well, if you wipe humans out, then you won’t have all the service you get from these humans.” And the architect says, “There’s levels that we’re willing to go.” And I feel like increasingly, the rich world is saying, “There are sacrifices we’re willing to make if our only choice for having people work in our country is putting them on a path to citizenship,” which, given the magnitudes of the flows, is inevitably going to change everything about the society and the politics and everything else.
People like Paul Collier write saying, “Look, people just want this sense of national identity. And hence if you force them to choose between preserving national identity and meeting the labor needs that exist, they’ll make hard choices in favor of national identity,” which I think is where we are. My big thing is if we actually had rotational mobility, in which people could come and perform the labor services but not necessarily instantaneously be on the path to citizenship, this could be a big thing that would be a win-win-win. It would be a win for the countries that need the labor. It would be a win for the workers that move. It would be a win for the sending countries.
Open borders implies that these concerns of national identity are going to go away or be weak. I don’t look at any rich country and see those concerns getting weaker. The fact that twice in a row—and I’m not sure I’m pronouncing her name right because I don’t speak French—but Marine Le Pen has been the presidential candidate in the run-off twice in a row, and she’s gone from 34% to 42%. Doesn’t sound like the world’s getting friendlier to open borders.
That said, the needs for this labor are going to get so huge, in my view, that there needs to be some intermediate solution. I think a well-regulated industry that does rotational mobility is a massive, massive opportunity. I just wrote a paper that I think I made everybody angry because I pointed out the gains from rotational mobility are bigger than the losses from climate change.
RAJAGOPALAN: Let me focus just on the gains from rotation mobility. Now, just the way you said there’s open borders, which is different from an increase in labor mobility, one kind of NIMBY-ism is the political NIMBY-ism that you’re talking about. “We don’t want people to get citizenship.” The other kind of NIMBY-ism is just, “We’re not going to let people build homes, and we’re not going to allow people who are working in these communities to live in these communities.” Is that kind of local level NIMBY-ism going to wipe out the gains from labor mobility if it continues? Because look at San Francisco. The whole thing is bananas.
PRITCHETT: Hey, I agree. We can stipulate that San Francisco is bananas, but I think San Francisco is a super special case. Because I think lots of places in America, it’s “get immigrants or die”—places with meat-packing industries, places with agriculture. A large part of the political tension in the United States is the fact that there are just chronically declining places and chronically booming places.
I think you’re seeing NIMBY-ism in places that have had pressures for growth for a very long time, like the Bay Area and Silicon Valley and San Francisco. Whereas I did a paper about mobility and geographic mobility over time. There are large swaths of the Midwest that from 1930 to 1990, over the 60-year period, well, the population was a third as large as it would’ve been had you had the rate of natural increase.
There has been massive out-migration from rural areas of large swaths of America for a very long time. I don’t see Cleveland or Toledo or Mississippi or a lot of these places having super strong NIMBY-ism pressures. If you could reassure people that these people are going to come and work, they’re going to be part of the local economy, but we’re not pre-committing from the minute they arrive that they’re on the path to citizenship. Again, I’m not saying every rotational mobility doesn’t have some path to citizenship, but it’s not immediate and automatic.
I think a lot of regions, their big problem is, if you think the way urban and spatial agglomeration works is, they’re on the downside of spatial agglomeration economies. I don’t know what the opposite of NIMBY-ism is; is it YIMBY-ism?
RAJAGOPALAN: YIMBY-ism, I believe.
PRITCHETT: YIMBY-ism. I think there’s a lot of pressures among the declining areas for YIMBY-ism. Just as a weird, weird anecdote, right? One time my wife and I were driving across the country, which we like to do. We happened to be driving through Nebraska at the time of the Nebraska State Fair. It was amazing. Really, one of the truly amazing applications of science to a field is agriculture in the United States. It’s unbelievable how scientific modern agriculture in the United States is, and yet this is a dying lifestyle.
The Nebraska State Fair, it’s about agriculture as an industry, and you have this unbelievably rich land, unbelievably high-productive agriculture, and you’ve got nobody doing it. Sooner or later, Nebraska has to choose: Are we going to have the land occupied by people that are going to work it and sustain a potentially high-productivity industry, or are we going to decline as they have been? Again, large slots of Nebraska have fewer people there absolutely than they had 100 years ago.
Particularly, by the way, if you could make rotational migration be region-specific, I think that would change the political dynamics a ton. If you could give a person a visa to come work in the United States, but they could only work in designated YIMBY areas, then, of course, the whole national dynamic that everybody worries, that all the migrants want to go to Silicon Valley or all the migrants want to go to New York, could be addressed.
The difficulty is these discussions aren’t even on the table. Nobody’s even talking about all of the various ways in which very clever rotational labor mobility could be designed and used, and the win-win-win that could be produced if we had this legally enforced, industry-engaged mobility. I think if we started to think hard about it, we could develop and design things that would slice through some of these political considerations. We have to be open to the—not all labor mobility is citizenship.
RAJAGOPALAN: I think Alex Nowrasteh has written about this, where he thinks that a lot of these visas should be issued at the state level, maybe even more local than that, right? That’s really where they’re trying to attract workers.
RAJAGOPALAN: I want to go back to declining fertility rates and aging populations. The global average fertility rate is just above replacement rate, I think 2.3. In India, in particular, for most of my growing years, it was that India’s overpopulated. We need to tame the population, this whole narrative. Now India’s population is at replacement rate, almost.
Developing countries were super obsessed with overpopulation being the main cause of endemic poverty. Various family planning programs were announced. Some of these were quite draconian, including forced sterilizations and things like that that we saw in India for a brief period. Most of it is, going back to our previous conversation, seeing like a state. We need to contrive it in such a way that families have fewer kids.
Now, going back to some of your previous work, there was a big discussion at one point in time. In large part, the feminist movement said that the fertility rates are not declining fast enough because of supply-side problems and contraception and family planning methods and this unmet need. Some of your work is, on the other hand, on women’s preferences, more about their individual choice, which is much more locally rooted, cultural, their economic circumstances and so on.
Where are we in the world on this demand-side versus supply-side debate on fertility issues? Second, is the entire point moot? Because it seems like economic development and lower poverty rates across the world seem to have solved that problem for us in a way. All of this doesn’t matter anymore. Where are we on this exactly?
PRITCHETT: Part of what you said I think isn’t quite right, which is, in this debate, if we go back, I think this is an example in which there was a huge key shift in mindset that was very dramatic, happened at an identifiable point. In 1994, there was the Conference on International Population Development held in Cairo. In the run-up to that, there were three really powerful groups—or there were two powerful groups and economists as one group. Economists were saying—me, for instance, and others were saying—“Look, it’s just not that plausible that people are having kids they don’t want because of a lack of ability to control their fertility.”
So I produced a paper that said, “Look, if we look at reasonable measures of how many children women want to have, and then that explains nearly all the actual desired fertility. And there’s no evidence that, conditional on women’s fertility desires, the availability or lack of contraception plays a huge role.” It plays some role, but relatively minor and modest relative to the population bomb needs.
Now, the interesting thing is, that split the feminists versus the demographers—there were demographers who were the population bomb types who said, “The reason we need to limit fertility in order to limit population growth is—”
RAJAGOPALAN: They were Malthusian.
PRITCHETT: Yes, they were Malthusian, and it was a popular view among—McNamara, when he was head of the World Bank, said, “The greatest threat to humankind besides the atomic bomb is the population bomb.” The demographers were perfectly willing, or at least willing to turn a blind eye, to the coercion of women to achieve what they required as the needed fertility.
They had fertility targets in their mind for social larger reasons, and they wanted to say, “If we just provided women with contraception voluntarily, it would dramatically reduce fertility.” Because it meant their demographically driven family planning programs could be perceived and projected as noncoercive. And at the ’94 conference, it was the feminists that said, “No, no, no. We’ve had it with this demographic stuff. We want a reproductive health approach that says women should have access to contraception because women should have access to contraception, not driven by this exogenous need to reduce fertility.”
I think this was an interesting thing where the coalition between, in some sense, the liberal economist view that people should have choices, and when they make choices they’re going to maximize their utility, and the feminist view aligned. And the feminists just wiped the demographers out in ’94 in Cairo, completely shifting the rhetoric around family planning from family planning programs as population driven to family planning programs as an integrated part of reproductive health.
This is a case where economists were right on the substance and right on what’s good for human freedom is actually going to be okay. And we don’t need all this coercive, top-down regulation of women’s fertility because if conditions get better, there’s a transformation in how many children couples and women and men want to have. And it actually put us on a path that, turns out, it’s not been so terrible.
Anyway, so really interesting story, though, because here the demographers were seeing a state view. And I think for a while, the feminists went along with them. But it was ultimately the feminists who revolted against the coercive and draconian population-target-driven family planning programs in favor of more choice for women about their own fertility, and completely transformed the field.
RAJAGOPALAN: Now, is all of this moot given where fertility rates are today? That’s one part of the question, and the other part is now, do we need a new target to increase fertility rates? Because fertility rates have dropped so much in so many parts of the world.
PRITCHETT: I do think it is mostly moot in the sense that, as you say, aggregate population worldwide—fertility rates are still quite high in some areas of Africa. But again, there’s no sense in which, if Africa were to continue to have a development path, it wouldn’t also, therefore, be on this declining trajectory. Now, that said, I’m just as nervous about targets on the other side as I am about targets on this side. No way do we want women coerced to have babies, just as we didn’t want women coerced to not have babies.
RAJAGOPALAN: That’s not what I meant by targets, but yes.
PRITCHETT: No, and demographers never had a model of human behavior, in the sense that demographers always had the problem with, if you did long-run projections, fertility had to go to replacement level, or population went to zero or infinity. So they were always like, “Well, we don’t need a model of why people have 2.1. They just have to, because otherwise our models are unstable,” which is not a very good position to be in.
I don’t know that we as economists have a particularly strong handle on why fertility has fallen so very low. I have a paper called “Why Demographic Suicide?,” in which I point out that an original economic innovation around Gary Becker and “A Treatise on the Family” was, well, this reducing fertility in spite of children being a normal good is quality/quantity shifts.
Now, if you look at European fertility, there is a very large spike of women finishing the reproductive years on zero. Analytically, quality-adjusted zero is zero. So if our theory were right that a lot of that or most of the shift or all of the shift was a quality/quantity shift, we should have seen a huge stack of fertility rates on one because there’s integer constraints. So we should have seen a lot of super-high-quality only children, but instead we’ve seen a lot more voluntarily choosing no fertility, zero children in their lifetime. And that’s not a choice that we can explain in the quality/quantity tradeoff.
I think there are some new papers recently trying to get at why is fertility going so low? And do we have a good handle on why, faced with hugely prosperous and productive economies, men and women are choosing—men and women, but especially the locus of control is with women—now are choosing to have so few and including fewer children?
RAJAGOPALAN: Well, you were talking about economists having a model of behavior. Not to make this too simple, but if through technology and immigration and cheaper labor and additional help, you can actually make it cheaper on the margin to have more children and raise more children, you could possibly move the needle on that. So in a sense, your previous work on labor mobility might be more connected to the fertility issue than just simply about aging populations.
PRITCHETT: That’s a good point. I think we might be headed there with our considerations that if it were the case that allowing rotational migration to handle the end of care too because the sandwich generation is caring for the elderly and for children. So given that double burden, you often will choose to not have as many children. Whereas there’s no reason why the rich world couldn’t have just radically more support to the care industry than it has from controlled and regulated labor mobility.
RAJAGOPALAN: Yes. Fertility is quite tightly linked to poverty and economic well-being. Even now, the world is very much on the trend of the places that are poorer are having more children that is higher than replacement rate versus lower. So it’s also weird in a global inequality sense that this is disproportionately shared by poorer countries, which always have fewer capabilities. There are, again, more and more reasons to have greater mobility than not because if the survival of the richer countries depends on this, they’ve got to make it easier to have and sustain higher-than-replacement rates of fertility.
Global Poverty/Prosperity Line
RAJAGOPALAN: There’s a lot of conjecture on whether poverty rates have reduced since the Modi government took office, whether they have increased since Modi government took office. And, of course, all of this is hugely dependent on how we measure, estimate this and which data source we use.
There’s a recent IMF paper that shows that India eradicated extreme poverty by 2019. Similarly, there’s a new World Bank paper that says that these estimates are higher than previously imagined at about 10.2%, something like that. But there’s still a big debate on this. But my question is—all of these are measuring around the $1.90 a day poverty line.
I have a slightly different question. Given that a lot of this is about measurement and about estimation and how some of this can even be rigged by local and national governments because they want to look better or worse in global rankings, should we just eliminate this problem and have a single global poverty line?
Here what I mean by single global poverty line is not that the $1.90 in India is the same everywhere, but rather that whatever works for Denmark or whatever works for the United States also works for India and also works for Sudan. If the global poverty rate is at, say, the U.S. line of $15 a day, then that is the line. I know you’re chuckling, but you’ve kind of suggested this in a slightly different way.
My reason for suggesting this is more simply, it can’t be rigged by national governments, and they just simply have to aim higher. It can’t just be, “I gave someone a little bit extra rations and some extra transfers, and I pushed them further up or further down the line.” That’s my very public-choice motivation for this. What do you think of this idea, and is this a good way to go globally?
PRITCHETT: Well, as you might know, I had a paper I think from 2006 called “Who Is Not Poor” that made exactly the proposal of—and again, I think there are two ways to think about this. One is exactly the proposal of, let’s just have a global standard for poverty. It’s hard to come up with a justification for why, if you happen to be born in the United States or Germany, there’s one standard of what it means is to be poor. And if you happen to be born in Bangladesh, you’re not entitled or you shouldn’t have any expectation of having that; you should live by this other poverty line.
Now, I think my preferred solution to this would be to have a measure called global prosperity. And a measure of poverty is, you’re poor if you’re not in prosperity. That would be at least something like $15 a day, maybe $20 a day—which is lower than, by the way, European poverty lines—but just a single global poverty line, and call that the high poverty line. If you think about Type I and Type II error, you’re pretty convinced that someone who’s above that line really is not in dire poverty.
The other—we could continue to have this low-bar poverty line, but I find the low-bar poverty line just theoretically, pragmatically and morally obscene. I’ve been against the dollar-a-day poverty line since it was invented. I fought very hard against it when it was introduced in the WR, the World Development report in 1990.
Matter of fact, I complained to the head of the WR that this was just going to focus people on this absurdly low standard, and his response was, “Oh, Lant, you’re overthinking this. We just need some aggregate number of poverty. But of course, this poverty line’s too low for anybody to really take it seriously.” Well, turns out I wasn’t overthinking it, and people all of a sudden really did take this line way too seriously.
That said, I think we could say, “Look, we’re going to measure super low-bar poverty where each country is going to have its own national poverty line, and then we’re going to measure high poverty, and we’re going to report all three all the time.” Like you said, the high-bar poverty essentially becomes a national income average consumption measure because nearly everybody in India is poor by a high-bar poverty line even today. And that’s a prosperity goal that I think we could get on board with, but I could go on and on. The low-bar poverty line is just a terrible, terrible, terrible thing.
RAJAGOPALAN: Yes, but tell me why it’s terrible. Tell me all the arguments against the low-bar poverty line. For me, it is that it’s aiming too low and that it can be rigged. Those are my fundamental problems with it. What are some of the other problems?
PRITCHETT: Well, my fundamental problem is it’s claiming there’s this objective function, where the derivative of that objective function with respect to income gains goes to exactly zero above the poverty line. Now, you could say marginal utility of income is well approximated by a zero at some line because above that, we’re all into positional consumption and Veblen-esque excess positional consumption. We could say that there is a line above which marginal utility of income is so low that it’s well approximated by zero and have a line.
Angus Deaton has the famous line, he’s in favor of poverty lines as long as it’s infinity. We could have a high-bar poverty line. The thing is, to imagine that marginal utility income is exactly zero at $1.95 a day is just morally obscene. People at that level of income have dire material shortages in nearly every dimension of their lives, and to act as if additional income to those people isn’t enormously important and valuable is just obscene.
I don’t know how the world has gotten away with pretending we care a lot about the poor, but at the same time creating this completely paternalistic, completely unrealistic thing and saying, “Oh, by the way, if you’re at $2 a day, what the hell? Marginal utility of income for you, we don’t have a goal for reducing that.” I think my main objection is that $1.90 per day is not at a level at which marginal utility of income is well approximated by zero, just above that.
It puts us into stupid economics. There is no line. One way I put this is, if we take a $1.90 poverty line, nobody has ever held a $1.95-a-day party. Who’s ever, “Holy moly, I’m at $1.95 day with my latest raise. Oh, let’s have a big old party. I’m out of poverty.” It’s just absurd.
Imagine marginal utility of income really were well approximated by zero at $1.95 a day. Then why the hell people are working 45 hours a week at really demanding and dangerous and dirty occupations at those income levels? Why are people scrambling around to continue to improve their material circumstances at those levels?
We have this paternalistic superimposition of a poverty line. The consequence necessarily is that you treat increments to income above that line at zero with respect to this objective function as an objective of it all. I think it was just a stupid, evil mistake from the get-go. It’s time to get rid of it.
RAJAGOPALAN: I think there’s also other concerns. I think it also changes how governments and development economists and policy people think about this. What I like about the $20 poverty or prosperity line is, you cannot redistribute your way out of it.
RAJAGOPALAN: There is no option except economic growth. Whereas at the $1.90 line, you can make some minor adjustments and redistribute and give some transfers and rations, and manipulate that whichever way you want. I think it also changes how policymakers think about the problem. There’s a question of aiming high versus aiming low, and maybe increasing that line globally will bring the conversation back to growth and convergence.
PRITCHETT: Absolutely. Again, I don’t think it was ever the South politicians who were driving a focus on low-bar poverty. This was always a development Global North agenda, not a Global South agenda.
RAJAGOPALAN: They’ve taken it and they’ve run with it, right? It’s one of those things where this is the metric, they’ll game it.
PRITCHETT: Yes, no. You create a metric and you game it. But I think if we want to have a real development economics going forward, it has to be a South-based, South-driven economics. Again, if you go back to the vision of the first-generation postcolonial leaders like Nehru and Nkrumah, they didn’t have a vision of India as a slightly less poor place. They had a vision of India as being a global player at the global favor within equality of nations and fully prosperous. They wanted everything that the U.K. and the U.S. had, and they still do.
The weird thing is that we’ve been able to convince South politicians to take this on and start gaming it with programs that are aimed at poverty reduction when, again, there is no line. People aren’t radically happy. Because the analytical mistakes we get into—we get into the mistake of like, “Oh, we have a poverty program.” Great. “And we’re trying to target the poor.” Great, but therefore stuff that doesn’t go to the poor is leakage.
That’s just stupid because the marginal utility of a person just above the poverty line is exactly the same as the marginal utility of a person just below the poverty line. And yet the way we calculate leakage in poverty programs, it assumes that if a unit of transfer goes to somebody at $1.95 a day, it’s not benefiting poverty. And if it goes to somebody at $1.85 a day, it is benefiting poverty. That’s just crazy stupid. The marginal utility of those two individuals is exactly the same, and it creates all kinds of crazy politics and crazy obsession with how do we correctly target that aren’t good politics. They’re not good economics. They’re not good anything. It’s just wrong.
Who Should Fight Global Poverty?
RAJAGOPALAN: I think we agree. One question I had here: When you say “we,” in terms of fighting world poverty, who is the “we”? Is it the poor? Is it the government? Is it development economists? Is it the World Bank? I ask because I think it’s important to clarify what are the incentives at play here?
When we say “we” as in you and me, our motivation and incentives are quite different from, say, the governments of the Global South or the paternalism of the Global North, whether it comes from the Bank or some other institutions. So who is fighting world poverty at the moment? Who is the “we” that we keep referring to?
PRITCHETT: I have to say, having worked in and around the World Bank for 20 years, I sometimes, by “we,” frankly just mean the World Bank and people and my colleagues and people there. There is something called the development community that has an academic branch, people who are in universities and think tanks and claim development economics is their field. And then there are donor institutions that do development.
Often when I say “we,” I have the development community in mind. And I have a paper called “Can Rich Countries Be Reliable Partners for National Development?” that goes to the fact that it used to be that the development community was more aligned with the needs, wants and wishes of the South and has diverged. I think the Global North has adopted poverty because of its terrible economics, not in spite of its terrible economics. Because if you think what the poverty analysis says, if the poverty rate’s 40%, it claims that the marginal utility of the median person in a developing country and the marginal utility of a median person in a rich country are the same. They’re both zero.
That’s stupid. That’s just wrong. The marginal utility of somebody at $2 a day and somebody at $30 a day are radically different, and yet low-bar poverty analysis makes it look the same. And that way, the Global North countries can go, “Oh, well, we don’t want to give away money of our citizens to the rich because, oh, the rich in these poor countries, they don’t need it.”
They act as if the median citizen of France and the median citizen of a country like India are at the same level of income and there’s no altruistic reason to make transfers. Well, that’s just crazy stupid. It’s politically convenient because it gets you out from under a true commitment to broad-based development, which I think the Global North has been trying to do since the end of the Cold War.
Sorry. I just wanted to say again, the frustrating thing is that poverty analysis survives because it’s wrong in a way that’s politically convenient for powerful interests in the North. It doesn’t survive because it’s right. It survives because it’s wrong.
RAJAGOPALAN: It also survives because the people who are doing it are invariably rich. There’s very little say of the people who are actually anywhere on this line or below or just above it in part of the discussion.
RAJAGOPALAN: On this, I have a question on whether there are players in the Global South who can change the way we think of economic development. For instance, India has a fantastically influential diaspora across the world, right? Especially, it’s probably one of the largest country partners for even institutions like the World Bank. Is there room for specific countries like India to use their influence to change how we think about development economics? Or do you think now the Global South is completely compromised? The Indian government benefits as much from this kind of low-bar poverty measure as the North does?
PRITCHETT: No, I don’t think so. After all, when—and now it’s gotten really tricky—but when China went and started creating new development institutions, I think there was a lot of enthusiasm among the leadership in India and Indonesia and other Global South countries that are out of the low-bar poverty levels to create these new development institutions that are actually responsive to the needs of South.
I do think that South-based, South-born, South-living economists and think tanks can play a big role in trying to push back and just say, “Look, this isn’t our agenda.” The Indian government isn’t really reliant in any significant way. It’s more cosmetic to them than existential, that they be seen well by the rest of the world and the donors. It’s not a big deal for them. I think a time will come soon in which the mainstream development institutions will be taken over by the agenda of the Global South, or they’ll continue to die as they’ve been dying.
RAJAGOPALAN: They have lost their purpose.
PRITCHETT: Exactly. They lost their purpose because they lost their way in the sense that they lost their purpose. And once they lost the purpose—all kinds of countries like the U.K. have absorbed their formerly independent development organizations into their broader—if it became part of the Foreign Commonwealth and now it’s FCDO [Foreign, Commonwealth and Development Office]. One observation is that Australia, Canada, the U.K., the U.S.—all of them over time have seen their development organizations essentially disappear into the broader diplomatic trade organizations. I have never once heard any developing country complain about that.
RAJAGOPALAN: That tells us something.
PRITCHETT: That’s embarrassing, isn’t it?
RAJAGOPALAN: It is.
PRITCHETT: You’re out there helping all these countries, and yet, when the domestic politics turn against you, the leaders of the country aren’t calling the prime minister of the U.K. and say, “Oh, God, no, don’t do that. It would be terrible.” I think in part, that’s because they lost their purpose. And they lost the purpose of really being continued as seeming as truly valuable to the development purposes of the South population and South leadership.
Randomized Controlled Trials and Economic Growth
RAJAGOPALAN: Speaking of having lost our way in development, I want to talk once again about RCTs.
RAJAGOPALAN: We talked about this a little bit last time, and it was mostly about who really benefits from them. I think you were right, that it’s really philanthropists and donors who ought to insist upon RCTs and impact evaluations and so on. And a little bit about the external validity issues, but you’ve written a lot more on what are the problems with RCTs.
The reason I want to have a detailed conversation with you on what is the Pritchett critique of RCTs—and feel free to take your time and develop it—is every time I’m back in India, the number-one conversation among young people, either at university or about to start their master’s or graduate work, is that what’s really happening in India is RCTs. That must be what’s going on, and that must be the way to go. If there’s another way, Lant Pritchett, please show us the way.
PRITCHETT: Well, let’s go back to what is sometimes called the Pritchett test, which is things that are really good for development should have—or good for growth in particular. Let’s just talk about economic growth, because I think economic growth is linked, and reliably linked, to a whole bunch of good things. Let’s talk about economic growth for a second. It ought to have four basic facts about it.
Rich countries should have more of it than poor countries. If something’s really good for level of income, then rich countries should have more than poor countries. Countries growing fast should have a lot more of it than countries growing slow. Countries that are now rich—we should have a lot more of it in countries that are now rich than when those same countries were poor. And when you see substantial changes in it, we should tend to see accelerations.
Those are the four fundamental things. When looking for something that was really important for economic growth, I would expect Denmark to have more of it than Haiti. I would expect Korea to have more of it in 1965 than it had in 1961. I would expect Denmark to have way more of it now than they had in 1870. I would expect Korea to have more of it than Ghana. Those are the things that are important.
None of those are true either about RCTs themselves, or are they true about any of the things on which the research method of RCTs is amenable. Because the things that are important for growth are important at an ontological level of the market. And at the ontological level—and I say ontological—the things for which RCTs are well adapted, coming out of drug trials in medicine, are things for which the treatment is individualizable. I can do it for you and not for me. But the things that lead to development are at the level of the market, at the level of the country, at the level of the society, at the level of the political structure, not things that are individual treatment.
First of all, no country that is developed is doing radically more RCTs than the developing countries—Denmark, Sweden, France, Japan. There’s no movement to do RCTs in those countries, nor did doing RCTs play any significant role in the way they become rich and develop. Neither RCTs as a movement nor the individual topics, in which if you look at the 3,000 RCTs being done, meet the Pritchett test at all.
So my fundamental critique is, we’ve moved into putting enormous amounts of effort into things that we know ex ante by just crude empirics, by the simplest possible, “Does it make any logical sense?” won’t make a big difference to development. And again, I think we did that because we lost the purpose around broad-based prosperity and productivity in the first place, and we moved into this dollar-a-day poverty. Then that allowed you to think in a programmatic way about pushing people just above or below this line.
I see RCTs as a symptom, not the disease, in the sense that they’re an opportunistic infection. If you lose your immune system, you’re going to get some nasty strains of diseases, and they go, “They died of this disease.” But you didn’t die of that disease. You lost your immune system. I see RCTs as an opportunistic infection in the body of the development community.
RAJAGOPALAN: Why are they so popular if they have nothing to do with development? What is going on within the political economy of academia and the political economy of the development community that has made this the number-one thing that everyone seems to work on? Because they’re very smart people working on it. It always makes me wonder why smart people devote their lives to something which you are so obviously pointing out is not going to get us the results that we want.
PRITCHETT: Because they’re smart people. What do we as economists believe? We believe people maximize their self-interest, and smart people are going to be more effective at maximization of self-interest than non-smart people. That’s why we have super smart people working on RCTs, because it creates an attributable product that in the academic world counts for a lot. And when they manage to create an environment in which, methodologically, this became, in some sense, for a variety of history of economic and social scientific thought—it was easy to create an environment in which that was the easiest, most legitimate way to create an academic career for yourself. Why wouldn’t we expect smart people to do what’s in their best interest?
Now, you could ask how did it—and then we go back to my story of the loss of purpose, the fact that the development community was losing that commitment to broad-based development. Rather than gaining, it was interested in economizing and cutting back expenditures on development.
We had to be more targeted, because we’re going to devote less effort and less concern to the broad-based development issues, which led to an environment in which therefore you have this sharp-targeting, programmatic, projectized world in a competition for fewer resources. And that therefore then, just from an academic point of view, just producing an individually reliable piece of empirical work (independent of it actually being important) became possible.
Where Are the RCT Critics?
RAJAGOPALAN: On this, then, I have a similar question on the rest of what’s going on in the academy. Now, normally academics love to spot a weakness and go after it and prove other people wrong. If it is so obvious that RCTs have no value and they’ve lost their way, why are there so few critics of RCTs?
PRITCHETT: In part, you need the next big thing. Meaning RCTs came along at a period in which other previous methods, like general equilibrium modeling and growth regression empirics were becoming burnt-over districts, and it was hard to get that. I think part of the issue is if you go to a graduate student, I think a typical—first of all, you shouldn’t underestimate how much academic fields get driven by the needs to produce new Ph.D.s. An incredibly high fraction of all the research out there is produced at very early stages of academic careers, including Ph.D.s. What a graduate student says is, “Tell me what to do such that you’ll give me a Ph.D. and I’ll get a job.” RCTs provide a really wonderful answer to that, or at least have.
If I come in as a critic of RCTs, I don’t come in with the next greatest thing. Like, “No, no, no, don’t do an RCT. Do this specific, concrete, wrap-your-head-around-it, methodologically straightforward thing, and I’ll give you a Ph.D. and get you a job.” If I come along and go, “No, no, no. We need to worry about growth diagnostics, and we need to worry about deals versus development. And we need to worry about things that Daron Acemoglu and [Douglas] North and other great thinkers of various stripes are worried about.”
I’m a grad student, I’m like, “No, no, no. I’m not going to hitch my fortune on becoming Daron Acemoglu. It’s like, he’s that, but I’m not that. I want to do something I know for sure is going to produce a determinate outcome in a limited period that’s going to produce something that people will call a dissertation. Boom, they’ll stamp it and I’ll go.”
Until we solve that problem, I think RCTs are here to stay and independent of the critiques, because the critique that they’re no use to people in the world—when have academics really ever cared a lot about that? I think the difficulty is that they’re infecting the development community from academia. I think the development community’s pushing back against it, but the problem in academia can’t be solved unless we say, “No, no, no, you can’t just do a silly little field experiment or pilot RCT of something that nobody’s ever cared about in the first place.”
Let me attack a specific paper, which I’ve written a blog about that’s in the public domain. These people did this RCT about whether, if you build community schools in Afghanistan, enrollments rate will go up—literally 11 villages in one region of Afghanistan. And what they found was if you build schools, kids are more likely to go to school. Their paper had literally three citations to the previous research.
Now, the idea that if you build schools, kids are more likely to go to schools has been the foundation of education policy around the world for a hundred years. RCTs have created this truly weird mindset in which showing something with an RCT, even if everybody already knew and believed it, still was meritorious and new academic research.
RAJAGOPALAN: Are RCTs the backdoor solution to what became the identification mafia in the 2000s, right? This is basically the escape route, right?
PRITCHETT: No, it was an escape route to—I don’t ever actually think the critique was really ever well founded, in the following sense. The game that got played in graduate seminars and economics is, I’m going to disbelieve what you say unless my prior—my Bayesian or otherwise prior—let’s not even say that it’s Bayesian; it’s just a prior. I’m going to play this academic game in which my prior is that what you’re claiming, that whatever you’re talking about has zero impact. And then the dynamics of my prior or subset—I’m not going to change my prior unless you present super gold-standard evidence.
If you think about it, it’s like, whoa, whoa, whoa. Where did your prior come from? The prior that was tightly centered on zero, that I would only be willing to move off it with completely pure identification, wasn’t a well-formulated prior based on the understanding of theory, evidence and the reality of the world. It was an arbitrarily chosen prior of zero.
Having a prior based on an understanding accumulated of reconciled evidence, and then not being willing to move that prior without record evidence, that’s a reasonable way to be. Picking an arbitrary prior with no justification, Bayesian or otherwise, and then not moving off that prior unless you have perfect evidence, that’s being an asshole.
There’s no side scientific justification for arbitrarily choosing a prior and then being stubborn about changing your mind. That’s just being a goof. “I’ve chosen to believe X, and I won’t stop believing X. The burden is on you to prove that I’m wrong.” It’s like, no. X was just a goofy belief.
The idea that you could say, “Oh, we’re going to create the pretense,” I call this feigned ignorance. “We’re going to pretend that no one really knows that if you build more schools, kids will go to school because there hasn’t been rigorous evidence of that. So I’m going to choose that my prior about the impact of building schools on enrollment zero. And then I’ll only move off that prior if you show me an RCT.” Well, no, no, that wasn’t ever a really scientifically coherent approach to the problem.
And so, the problem with the identification mafia was—it wasn’t that they wouldn’t move their priors without rigorous evidence. They never had any coherent rationales on how they were choosing the prior off which they had to be moved by rigorous evidence. Without that, they never really had a scientifically coherent set of claims, or approach to a field in the first place.
The Bad Science of RCTs
RAJAGOPALAN: The farce is almost worse than that. Because, okay, let me explain myself. It’s not just that their prior was incoherent, right? That “Oh, we have no idea if building schools helps or not.” Now they’re saying that because we know that it helps in these 11 different villages or districts in Afghanistan, we know this as a fundamental truth everywhere. Whereas now there’s a mountain of evidence that the external validity of these RCTs is actually pretty poor, and it’s highly contextually driven.
And you know all these arguments better than I do. I think it’s both; it’s a particular kind of incoherence at the level of forming a prior, but it also completely suspends all rational thought of being persuaded by the evidence everywhere universally. Because there’s a very narrow result, which is the gold standard in measurement and attribution nevertheless, but it’s still a very narrow result.
PRITCHETT: Let me just say—I’m going to say it one more time—that there’s this critique that I’ve been making for a long time that I can just never get people to really wrap their head around, which is, suppose you had a prior distribution about—and I’ve written a paper called “Let’s Take the Con Out of Randomized Control Trials,” building up the famous “Let’s Take the Con Out of Econometrics.” They’ve pushed the con out of econometrics, but they’ve pushed it under randomized control trials.
The problem is, let’s say I have done 150 observational studies about the impact of building schools on child enrollment. That’s going to have a distribution of the impact, with some places having high impact, some places having a median impact—there’s going to be a distribution. Now I’ve got 150 studies; here is the distribution of estimated impact from observational studies.
Now you come along, and in one context you do an RCT. Now, what that RCT should give you is, it gives you two things. It gives you an estimate of the impact of building schools on enrollment, the kind of impact effect. But it also, if you do it right (and it’s not that hard to do it right), gives you an estimate of the bias relative to the observational estimate.
The problem is they never—don’t yet to this day—have a coherent answer to how do I move my prior about the overall distribution in response to these two new pieces of information from one specific context? The problem is, without loss of generality, you could easily have a situation where the estimated impact effect is higher than the mean median of the existing observational studies—I should move my prior distribution up, meaning it has bigger impact. But at the same time, it could well be that the result of the estimate of the bias is that the bias of observational is negative, and so I should move my distribution down. It isn’t just an external validity problem.
People keep acting as if the only thing produced by an RCT is the impact estimate. It also produces an estimate of the bias. And there’s been no coherent answer to the question, should I move my prior of the existing observational studies toward the impact effect or toward the bias in the effect? There’s no guarantee that those turn out the same way.
Second, there’s never been any coherent argument as to why we should change our variance. Let’s say we have a variance of our estimates, and in the paper I have variance of the estimates of the learning premium of being in a private school. We pick a kid, we observe their score. We observe what kids are at private schools or public schools. And we say, what is the causal effect of a private school? That has a variance.
You come along, and now you have a rigorous estimate of the causal impact of a kid being in private school in a given context. But then am I supposed to reduce the variance? Well, if I reduce the variance, it means I have to say some countries’ context that my observation lesson was higher than that, I move lower. And some countries’ contexts that were lower than that, I moved higher. That makes no sense. It’s just logically incoherent.
Again, as science, the RCTs haven’t met even the minimalist standard of having coherent answers to pretty easy questions about methodology, about what am I to make of this new evidence you’ve presented me with? What people are saying about external validity is actually, once you did enough experiments, you just reproduce the variance of the distribution of the observational studies.
RAJAGOPALAN: The thing about “What do I make of the new evidence?” To me, I think that’s the most interesting question not just on the bias and the variance, but also on what is being chosen as the experiment. If you look at some of these experiments with like mosquito nets for malaria, you would think that the solution to malaria is mosquito nets. Whereas, the world over, we know that the solution to malaria is better water and sanitation services. I have not seen mosquito nets in the rich and developed world. I have never used one myself, because I’ve always lived in neighborhoods which are relatively prosperous and don’t have standing water and sewage and things like that.
Yet what is being measured? Though I’m sure they are correct in measuring that if you gave away free mosquito nets or maybe for a few cents you gave away mosquito nets, the malaria infection rates go down, it still, for me, doesn’t solve the bigger development problem. Because I’ve never heard a poor person saying, “My main problem is I can’t afford a mosquito net for my child.” It’s usually, “I can’t move my child out of this horrendous neighborhood where there are all kinds of infectious diseases. I think that is my more fundamental problem.” At the scientific level, it is what is being chosen for measurement.
RAJAGOPALAN: And I fully appreciate that it might be very difficult to do an RCT of do we provide sewage services, or are we going to have contaminated water in a particular area and endanger people? I understand that these things can’t be easily studied in an RCT, but it still is relatively the more important question.
PRITCHETT: Absolutely. This is another good example of, look, you’re biasing what’s being looked at toward treatments that are individualizable, away from systematic or systemic—America eradicated malaria, we got rid of it, even though it was endemic in the South for a long time. We shouldn’t be choosing policy levers to address a problem based on the fact that one of them is easier to generate rigorous evidence, because we can solve the identification problem to the standard of the identification perfection. Again, good example of “Why are we looking at this versus that?” as a prior and important question to “How can I write a dissertation that meets the needs of the seminar room about identification?”
Education in Developing Countries
RAJAGOPALAN: To move the question back to the nonbiased question, so this is going back to a lot of the RCTs in India on education. And education does seem to be one of the most important areas where investments need to be made and where both government and private philanthropy have a big role. If you walked away from this attribution game, what have Vietnam and Peru done right that India got terribly wrong in education?
I picked these countries because we see that richer countries typically have better test scores and better outcomes, even if the models of schooling and education vastly differ both between countries and within countries and so on. But if you are richer you tend to get better education, which is pretty straightforward.
Now, Vietnam and Peru are not that much richer than India is, or at some point were (not that long ago) at the same income levels as the median Indian. And yet they have done something fantastically right when it comes to education in a way that India has gone horribly wrong. So what is going on in that story?
PRITCHETT: This is a good example where the right answer is not very attractive from an identification police point of view, which is the way—I’m head of this research project called Rise, and we had seven country teams. One of the country teams was in Vietnam, and the reason for choosing Vietnam was to answer this question: What did Vietnam do right?
After pushing our Vietnam team to say, “What was the answer? Why did Vietnam do so well?” in the end, one of our researchers, who I have a lot of respect for, he said, “Look, it’s just—they wanted to. They wanted to, and because they wanted to, they found a way to do it.” So you’re pressing for proximate determinant causes that aren’t the ultimate causal driver of this. If you want to know why Vietnam has really high learning performance among the students, it’s because they consistently, coherently wanted to.
If you don’t want to, knowledge of the type of this program versus that program, it’s just not necessarily going to work as designed when you implement it. Because it’s not going to get implemented, or it’s not going to be implemented as well. We crudely sometimes joke that the big answer from this long-term Rise project about what leads to higher learning achievement are the three A’s, not the three R’s. The three A’s are: Ya gotta wanna.
Yes, you got to want to, and if you don’t want to—and I think what India got wrong is India, as a society, as a government, was never really (and still to this day isn’t truly) committed to the belief that every child can and should achieve a relatively high level of learning performance. They’ve never really committed to it, still aren’t. There’s still the belief that education is a process of choosing the elite few who are good at it and devil take the hindermost, even inside Indian classrooms today.
I think India is in the process of coming around to the “you got to want to” stage where there is generating a lot more social concern over this. But until India gets there, as we saw with SSA [Sarva Shiksha Abiyaan] was this massive, massive investment. During that whole period, as best the evidence can tell, overall learning per year of schooling of children was on a stagnant at best, but probably declining trend during that whole period.
RAJAGOPALAN: When you say you got to want it, who are we talking about? Are we talking about parents, are we talking about students, teachers, society at large, government, policymakers, economists?
PRITCHETT: I think everybody is the answer. Meaning you got to want to at the system, society level. You can’t have these people talking about, “Oh, we need to have a dissemination program where we reach ministers of education with the right messages.” And it’s like, you got to be kidding me. When have you ever met a powerful minister of education? They’re completely responsive to the pressure from the rest.
When you ask—I think the World Development Report, which was written by friends of mine at the World Bank, of 2018 on education—they said learning for all requires all for learning. I think that’s—again, a nice little phrase is, when I say you got to want to, “you” is everybody. You need to have parents on board. You need to have the education community, broadly speaking, on board. You got to have the government on board, and they have to have enough of a commitment to a clear set of coherent goals that they’re willing to do what it takes to make it happen.
This isn’t the kind of endeavor in which some few reform champions are going to make a long-term difference, unfortunately. And I do think reform champions are going to make a difference potentially in the long run by creating the overall you got to want, I think. For instance, Rukmini Banerji—she’s a huge hero of mine—I think her and Madhav Chavan and their continued attempt to raise these issues in factually informed ways has been part of the shift that’s happening in India on this topic.
It wasn’t their individual interventions and project designs that were approved by an RCT to be effective in raising learning outcomes that made the difference. It was their recognition that they needed to change the overall environment and the concern about these issues. Last interview, did I tell a story about Madhav?
RAJAGOPALAN: You told us the story about Chandra Bhan and the village where they’re talking about whether you should, if there’s a snake—
PRITCHETT: That there’s a snake. Madhav would tell the story, but I’m going to tell it better than Madhav would. Maybe it’ll be less true but more to the point, which is: After they had done the RCT showing that this Balsakhi program of putting tutors in the schools really led to substantial gains and learning achievement and reading outcomes, he took it to the secretary of education of the place in which they done the RCT. And he said, “Oh, by the way, I have the solution to your problem of low learning levels, or at least part of the solution. Look, we’ve got this powerful evidence that this works to improve leading outcomes by putting these volunteer tutors and pulling their low learning kids out.”
The response of the secretary of education was, “What do you think my job is? Why do you think that this is a solution to a problem I have? Look around my office. See these piles and piles of files that keep me busy 60 hours a week and not one of these files is about a child not learning. I’m under no pressure about that problem. If I try and transfer a teacher, I’ve got a court case on my hand. If I try and close a school, I got a court case on my hand. My job is to administer the existing education policy such that there’s policy compliance. Super kudos to you for this cute little study you’ve done. It has nothing to do with my job as secretary of education.”
That is what led Madhav to realize that the problem wasn’t rigorous knowledge about what works. The problem was getting anybody to care about the problem, and hence ASER and the kind of advocacy of getting people aware of and concerned about the low levels of learning was an instrument to get people to adopt the effective thing. Again, back to the RCT thing, the idea that it was the RCT that convinced people to have an uptake of this improved innovation would’ve assumed that people in the system were out there looking for more effective ways to achieve an already shared consensus target, which was already a strong purpose of the system.
Until you change the agreed-upon purpose and understanding of the purpose of the system, all of that would’ve had zero impact. What’s changed in India is not the RCT about Balsakhi. It’s the changed advocacy environment making people aware of and concerned about—which has taken 15 years to really get the momentum it needs. It’s not like you drop in and go, oh, look, by the way, you needed this fertilizer on your crops.
RAJAGOPALAN: No, no need to apologize. It’s a little bit heartbreaking because what’s happening with poorer parents is, they spend a very large proportion of their disposable income exiting their kids from free public schools to these budget private schools, where education’s not great, but it’s still better than the free public school. I mean, what can be more damning than the evidence—than the fact that very poor parents are actually spending half their monthly income on pulling their kids out of these schools?
Exit vs. Voice and Loyalty
RAJAGOPALAN: But one thing that worries me about education, which is different from other kinds of welfare programs if you think at a political economy level, is if you think about exit-voice-loyalty frameworks.
RAJAGOPALAN: The mechanism has been exit.
RAJAGOPALAN: It’s understandable because voice and loyalty, the opportunity cost is just too high. A parent can’t stand and advocate for their child for years on end while their child can’t learn how to read and write or basic numeracy. The best way to deal with it is just to pull their kids out of school and send them to an okay school. Not that much better than the one they’re already at, just marginally better. They’re not going to have a dramatically better life, but they’ll be somewhat better off.
But what that ends up doing in the larger scheme of things is that the only people advocating in front of the secretary of education or the ministry of education are schoolteachers’ unions, people who are getting construction tenders for building new schools and stuff like that. It’s not really the people who are using the services. Unlike healthcare or unlike gram panchayat, there’s so many other government programs where the voice and loyalty overpower the exit. But in education, that doesn’t seem to be the case. What’s a good way to fix this all?
PRITCHETT: That characterization of health—there’s also been a massive exit in health. If you look at Jishnu Das and others that have studied, they did a study of first to contact. Who do people, when they have a health condition, who’s the first point of contact? Maybe there were a lot of things about that study about the quality of care, but the really, in some sense, impressive thing was that of first-care contact, the doctor at the PHC—3%.
People say, India’s healthcare system—sometimes when I say “Describe India’s healthcare system,” people describe this top-down chain of suppliers. But 3% of first point of care contact was an MBBS doctor. It’s not the system. The system is this huge. I think there’s been a massive exit from that too.
But I think this is maybe the deep problem about state capability in India is that—and Arvind Subramanian just had an interesting interview in which he was pointing out that what the government’s done is not try and fix the hard-implementation things like health and education. They’ve improved the direct targetable here. You’re getting gas here. You’re getting concrete benefits of cash or in-kind transfers that are logistically simple. That, again, it’s not going to re-create a coalition about fixing the core problems of the Indian state.
RAJAGOPALAN: I call it the upside-down Indian state. They don’t provide public goods and quasi-public goods. But what they provide is private transfers and private entitlements, which could be done very well by the private sector, except now it’s the government in the business of—it’s a little bit upside down if we think of it from a development point of view. It might still be better than nothing.
PRITCHETT: Yes, it is worrisome if you’re providing subsidized gas but you don’t have a police force that’s reliable.
RAJAGOPALAN: That’s exactly the situation.
PRITCHETT: I know, and that’s what I’m saying. That is upside down if you think of what functions are core to the state. A lot of the functions that are core to the state are being allowed to deteriorate over time, or at least stagnate over time, versus attention to providing what are just purely private goods in a redistributive way. Which I’m not super against, in some sense.
But on the other hand, the idea that using high tech to identify individuals, to make sure that people are getting the right this or that private good as a redistributive thing, is a reform of the state—that’s your main improvement—that’s a very twisted notion of what the fundamental responsibilities and capabilities of advanced developed states need to be.
I do think the exit problem is a very, very serious problem because once you’ve exited, that often constitutes investments in another modality. My friend Junaid Ahmed, who was head of the World Bank office and has been in and out of India for the last 20 years, his big sector was urban water. He was like, “Why do Indian cities not have 24/7 water?” Then wanted to reform and municipal and corporatize. It was like, I always just said to Junaid, “Everybody who really is politically powerful does have 24/7 water. The way they’ve created 24/7 water is, they put a tank on the top of their house that is filled by the public water when it’s available, filled by a private tanker when it’s not.”
Once they’ve made the capital investment in a purely privatized system of providing 24/7 water to them, getting them to exercise voice and loyalty in a system in order to have a municipal provide reliable 24/7 water universally just isn’t on their agenda. Moreover, it’s like, well, wait a second. Once I’ve made this capital investment in a public sector capability-substituting investment, until that investment’s amortized, it doesn’t benefit me in the short run.
RAJAGOPALAN: Alex Tabarrok and I had a paper on Gurgaon as India’s first private city. What we meant was that it obviously grew awfully quickly, and there were huge gaps in public sector provisioning. As economists, one way to think about this is great, the private sector stepped in, in everything from law and order—Gurgaon had the first private fire brigade, and it was much—DLF provided far superior.
I went and spoke to the fire brigade, the top people there. I don’t know how much Hindi you understand. One of the things he said was, “Oh, we have no competition. The government municipality fire station, they were fighting with pichkaris,” which means water pistols, “whereas we actually have high-rise buildings, so we need a state-of-the-art equipment to actually protect our investment.”
On the one hand, the private sector did step in, but on the other hand, you had all these problems of the commons, which is groundwater depletion. And it’s massively inefficient for each condo building to generate its own electricity or generate its own water. You know all these problems. India’s functioning in this complete exit and quasi-private for the rich state, which creates much bigger problems, because there’s no real pressure from the bottom for governance among the most privileged class. That’s always a problem.
Role of an Economist
RAJAGOPALAN: Before we end, I want to ask you, what is the role of an economist? Is the role of an economist to be primarily preachers the way George Stigler meant it, as persuading other people to adopt these goals that maximize efficiency and a particular kind of outcome? Is it to just be academics and truth seekers, is it to measure, is it to model? What’s the role of an economist?
PRITCHETT: I’m not quite sure what George Stigler meant by being a preacher. The thing I like most about the field of economics is that it is still mostly people who are open to empirically grounded discussions of problems in which you’ll acknowledge what the facts are and alternative approaches to modalities of making the facts different.
Then the other thing is what you said in the last interview, is that the other is that I think economists—I don’t really teach undergraduates very often, but I was invited to give the opening lecture to a development economics course of undergraduates. My take was that economics is the social science of love. It’s the truly loving social science, and what I meant—and they were, of course, like, “What? Economics and love? That’s crazy.” But think about what economists do. We take individuals—objective functions are objective functions. We don’t start with any premise about what would be good for society or good for X or good for Y.
But I think economists, when they’re doing it right, they start from, what is it that people want to accomplish with their lives? Okay. Let’s think about what the actual outcomes are. Let’s think about modalities at the society, political, market level that would facilitate individuals achieving their objectives more or less. And what could be a better description of love than “I’m going to take—what you want is what I want for you, and I’m going to help you achieve that.” Economics is the loving social science, is my take on what economists do best.
RAJAGOPALAN: Oh, that’s wonderful. Well, certainly been true for me. I always say that once you see the world as an economist, it’s very difficult to see it in any other way.
PRITCHETT: That’s right.
RAJAGOPALAN: It’s quite wonderful. Last time I was asking you, what are you binge-watching? This time I want to ask you, what are you reading?
PRITCHETT: Embarrassingly enough, I’m reading a David Baldacci novel, which—it’s just awful. I’m reading that. In part I’m reading that because I’m also reading “Noise,” which I highly recommend. I thought, wow, here’s some people talking about what I do for a living, like how econometrics works, and I’m not insulted by it. That’s impressive. It’s a good book. I’m enjoying it.
Then I was reading David Graeber and somebody else’s [David Wengrow’s] “The Dawn of Everything” book. Have you seen this? Then I had to read David Baldacci to just contain my ire of that. It’s completely, totally insane. I would recommend that everyone read it, and then afterward we can have a discussion of, what have we come to in which this is regarded as an advance in the field of social science?
RAJAGOPALAN: What are your reading habits? How do you pick what to read? I can see that you’re trying to trade off between irritation and some kind of mindless pleasure at the moment, but more generally, what are your reading habits? Also, what’s your split between reading papers and books? This is a question I like to ask economists.
PRITCHETT: One of the problems with being my age and position is that I have to read a lot of stuff for a review, comment, so most of my reading is dictated by my having agreed to be on this committee, or a peer reviewer of that, or a referee of that. Rise, the research project I’m the research director of, has produced 158 or 159 working papers. I read every one of them or nearly all of them in a review stage. Anyway, a lot of my reading is determined and therefore is mostly papers.
And to the extent I read books, I often read books well before. I’m much more likely to read a manuscript of a book than a book. I was at one time talking to somebody about, “Oh, yes, and this book may do this and that.” They were like, “I just read that book, and they don’t do what you’re saying.” I’m like, “Oh, shoot. Well, I read the manuscript two and a half years ago, so they were doing that.”
You never have time to actually read the published version. I read Chris Blattman’s excellent new book in manuscript. I read most of Bill Easterly’s books in manuscript. Most of my reading is dictated by professional obligations. Often when I go to read at night, I can’t read at night before I go to bed. I can’t read professional stuff because I get irritated.
RAJAGOPALAN: You’ll stay up.
PRITCHETT: Well, it’s not that I stay up; it’s that I stay up reading it. I read something, and then I get obsessed with why it’s wrong. Then I spend the first hour I’m supposed to be sleeping, composing an essay or a blog on why it’s wrong. I’m like, “Just go to sleep.” I can never read professional stuff after 8:00 or 9:00 at night. Then I have to binge-watch something. That’s why I binge-watch rather than binge-read because if I read something professional after 8:00, it’s like eating pizza just before bedtime. My mental indigestion can’t handle it.
RAJAGOPALAN: What are some of your favorite economics books that you’ve read over the years? It may not be anything contemporary or something you read at manuscript stage. It could be one of the greats.
PRITCHETT: I think it depends on what you count as economics, I guess.
RAJAGOPALAN: Whatever you count as economics is good enough for me.
PRITCHETT: Okay. James Scott’s “Seeing Like a State” is a great book. It’s a great book. It’s like I was reading—and I think I said this before. I was reading it, and I’m like, “How has he been in the World Bank meetings I’ve been in? How does he know this is exactly how my organization works?” That was a great book. I actually read, and it’s hard to not talk about what you’re reading and not sound snobby, but I actually read Schumpeter’s “History of Economic Analysis”—600 pages.
RAJAGOPALAN: Another great book.
PRITCHETT: I read a lot of it, in part just because I enjoyed how incredibly snotty Schumpeter was. I came away with great lines that I can’t find again since I read it on paper. Anyway, so I really enjoyed that because I enjoyed the appreciation. I think anybody who’s really into a field appreciates the giants that came before and the evolution.
I really enjoyed “History of Economic Analysis” and came away with two things that I use all the time, one of which was Schumpeter describing an economist as having constructed a simple model that didn’t work to explain the world. Rather than concluding correctly, “I am a fool,” they concluded, “The world is mad.” That’s like, look, when the world doesn’t conform to your model, the solution is not to conclude that the world is mad. The solution is, maybe my model or maybe my understanding needs to be deepened.
And the second thing, from Schumpeter, is what he calls the Ricardian vice, which is having an understanding of a narrow little aspect of a phenomenon, and then insisting that that understanding is what should be driving the discourse. To a very large extent, coming back to my bête noire, the RCT movement is completely a Ricardian vice. We have this one little tool, this one little thing that augments our understanding of certain dimensions of certain things. Because we have that, that should dominate the public discourse. I think the Ricardian vice is alive and well.
RAJAGOPALAN: The first part that you mentioned—are you foolish or is the world wrong—I think you just described all of behavioral economics in one sense. I don’t know if you meant to. It’s an interesting thing, how behavioral economics has made an entire field of calling the real world not good enough because it doesn’t live up to the models of homo economicus, whether it is savings rates or whatever is the question that is being asked, whether we are eating too much or eating too little.
But thank you so much for doing this. This was such a pleasure.
PRITCHETT: I’m always happy to talk about India. I’m always happy to talk about economics and always happy to talk with an interviewer and interlocutor who’s so well informed. Pleasure is mine.