Russ Roberts

Clemens on Aid, Migration, and Poverty

EconTalk Episode with Michael Clemens
Hosted by Russ Roberts
PRINT
Morris Fiorina on Polarization... Michael Lind on Libertarianism...

Michael Clemens of the Center for Global Development talks with EconTalk host Russ Roberts about the effects of aid and migration on world poverty. Clemens argues that the effects of aid are positive but small. But emigration has the potential to have a transformative effect on migrants from poor countries who emigrate to richer ones. The discussion concludes with the impact of migrants on the host country.

Size: 27.3 MB
Right-click or Option-click, and select "Save Link/Target As MP3.

Readings and Links related to this podcast

Podcast Readings
HIDE READINGS
About this week's guest: About ideas and people mentioned in this podcast:

Highlights

Time
Podcast Highlights
HIDE HIGHLIGHTS
0:33Intro. [Recording date: July 12, 2013.] Russ: Our topic for today is world poverty, and how aid and migration might play a role in reducing that poverty. We're going to draw on two academic papers you've written on these topics and try to weave them together a bit. I want to start with aid. There are studies that find that aid increases growth and presumably reduced poverty. There are studies that find that aid doesn't increase growth. I assume there are studies that find that aid actually decreases growth. What are some of the problems in those studies in terms of measurement? What are some of the challenges with trying to measure the impact of aid on growth--which are quite numerous, I assume. Guest: Yes. So, first of all I appreciate this framing of the issue. I think that we'll be going from the less important to the more important topics. I just want to make it clear why I say that. Right now, globally, foreign aid is a little over $120 billion a year, if I've got the latest numbers right. Just the portion of international migrants' earnings that they send home is about 4 times that. And that doesn't even account for the gains to themselves, to the destination countries, etc. So really we are starting with the sideshow, and then we're going to move on to the big stuff. Russ: But that sideshow is where a lot of people have focused over the last few decades: the need for more aid, or should we start cutting aid. Your point is that, you are saying that, migrants, temporary or not--I guess either way--that people who arrive, all migrants, send home roughly $500 billion to their country of origin as a transfer back to their families. Guest: It's between $400 and $500 billion a year. The World Bank has some estimates. And those are imprecise, but by any measure they are several times foreign aid flows. And as I said, that's just a part of the economic [?] to migration. Russ: Correct. So, but sticking at least for now with aid. So, what are the challenges with trying to figure out whether aid is a good thing or not for the countries that receive it? The people in those countries? Guest: Yes. So, there are many layers of the onion here. We have to start with, what you say: is it a good thing or not? What we've studied, in some depth is whether or not it causes economic growth over a certain time horizon. That is certainly only one part of both descriptively and normatively what people mean by a 'good thing.' Descriptively a big part of the purpose of lots of foreign aid is geopolitical concerns, and one could argue that whether or not it's a good thing should be judged on those criteria. A lot of foreign aid--for example, the foreign aid that went to eradicating smallpox in the late 1970s, was not intended to promote economic growth, certainly not in the short term. So, within the question or not of whether it's a good thing there's the question of whether or not it promotes economic growth. And that's extremely difficult to assess, just like assessing whether or not anything causes economic growth. Russ: But that doesn't stop people from assessing it, does it? Guest: No, it doesn't. Long ago, back in 1992 I think it was, one of my former professors and someone I deeply admire, Greg Mankiw, wrote a paper in which he urged policy makers to pretty much ignore all growth regressions, and I am not sure that he was very far off in that assessment. But it is probably better to do it than not do it, and as you say, many people do it. Many people since the early 1970s have regressed growth on aid in various forms. And therewith come all kinds of problems. The biggest problem is attributing causation. If you see a poor country getting aid, it could be that the country was getting aid precisely because it is poor relative to other countries. It could even be that it's getting aid relative to what it once was: that is, it had a negative economic shock and then aid was ramped up within that country over time. If you see aid flowing and then a country is doing very well, say, Korea got lots of foreign aid after the Korean War and is not pretty much a rich country: is it that it was attracted by that success; did they cause that success? It's extremely difficult to sort out in macroeconomic data. Russ: There's also the issue, as you raise in your paper, of timing. And why is that relevant? Guest: Yes. So, this is all very commonsensical. If you have a large aid project to, let's say, build a dam, build roads into an agricultural area that didn't have market access before, you might expect some increase in economic activity around those projects. When would it arrive? Well, not the day that you disburse the aid. Probably some time before a century later. But there isn't a lot of guidance from economic theory about exactly when that would happen. A lot of economic theory is built around comparative statics, and you might shift a curve, endow a country with more capital, and expect at some rate within some time period a new equilibrium to arrive. But really there is not much guidance from economic theory about when those might arrive. For some types of aid--I mentioned health aid, to eradicate smallpox or the U.S. government gives aid to promote democracy; there are aid projects to promote women's' rights--if those things ever affect economic growth--and they might--it's not clear when that would happen. So if you are looking for the growth effects of aid there is some art and some feeling around in the darkness that has to happen to find them.
7:38Russ: And of course when you said, by the way I just want to mention to our listeners: comparative statics is just fancy economics for what happens when one thing changes and everything else is held constant. Which we try to measure with statistical techniques, with a lot of difficulty, as we'll be talking about. Guest: Yes. Russ: But one of the other points is that these statistical analyses that try to tease out the independent effect of aid on economic growth--aid is just measured in dollars, correct? It's a very crude aggregate measure. Guest: Yes. It's a variable that we use and that most studies use, a measure developed by the Organization for Cooperation and Development (OECD), and it is calculated through a pretty complex process to include the 'grant component' of loans and grants. So it is the concessional component of loans, in dollars as you said, and it is also grants in dollars, lumped together. Russ: And then the other problem I want to just raise before we go on is the work of Morten Jerven, who was a guest on EconTalk fairly recently, who points out that in many poor countries, income, or Gross Domestic Product (GDP) or whatever measure you are going to be using going into the thing that aid is affecting, is often mismeasured in systematic ways. Guest: Terribly mismeasured. Yes. You might have recently seen that the size of the economy of Ghana as estimated by the World Bank underwent a discontinuous change. That's happened about think twice in recent history for China. This isn't because anything actually changed in the real world. It's that people got arguably better at measuring what's happening what's actually going on. So, yes, it's an incredibly noisy measure. Russ: And before we get to what you found in your study, I want to ask one more theoretical question, which is: Why would I ever think that there's a relationship between aid and growth? Aid and income--the level of income--I could have imagined. It's a possibility, if it's money well spent. But it would be difficult, it would seem to me, to consistently change the path of income. Which is what you are usually trying to do here. Correct? So what kind of projects--you mentioned road-building. I could imagine. Of the range of stuff that foreign aid goes to, what are some of them that might plausibly lead to aid and what are some of them that we wouldn't expect at all? Guest: As you say, aid goes to all kinds of things. Aid purchases weapons. Aid purchases vaccines. Aid purchases nebulous training seminars for low-level functionaries of ministries. Aid purchases building roads and airports. And as you said, some of those things it's much more intuitively clear that it's very hard for agriculturalists in internal Cameroon to get their products out and if you build a road that is accessible most of the year, by normal vehicles, so that they can sell stuff in urban areas. You'd kind of be surprised if that didn't affect economic activity in a broader sense. On the other hand, a training seminar for low-level functionaries of the finance ministry in Monrovia--it depends a lot on who is there and what information is being conveyed and what the other bottlenecks in the system are. And then as I said there are large swaths of aid, I'm not sure it makes much sense to evaluate their economic impact on economic growth. The eradication of smallpox is arguably on the top-10 list of human achievements, and I'm not sure if that's because it affected economic growth. It's a wonderful thing.
12:06Russ: So, I'm not a big fan of foreign aid of most kinds. Yet it would seem to me that given this measurement issue, that much of these transfers, much of this aid, we wouldn't expect it to affect economic activity or growth. Does that mean that many of the studies are biased downward in what they measure, that they are going to underestimate the impact of growth because they are including in aid things that you wouldn't expect to be, and therefore that would bias the coefficient, the measured impact, to be smaller than it actually is? Or do some of the studies that are important in the literature, do they correct for that? Guest: Some do and some don't. In our work we eliminate humanitarian aid, which really--distributing bottles of water after the Asian earthquake in Haiti was not intended to promote medium- and long-term economic growth. It was a humanitarian measure. Aggregate aid includes that. We in some of our specifications we included a project to promote primary education. Again, in parts, that's inherently valuable; in parts that might affect economic growth. But probably not within the next few years. Maybe decades from now, if ever. Russ: It might reduce it. The kids who were otherwise working are now going to be investing; and measured income could go down. Guest: It could go down, in the short run. And the long term. Russ: That, too, because sometimes they don't learn anything. There's a lot of issues. Guest: Lant Pritchett has some fascinating works documenting how little learning is going on in schools. He argues against measuring education by what he calls 'rear ends in seats,' and instead by learning. And if you look at the learning metrics we have from Programme for International Student Assessment (PISA) test scores, etc., there's so little learning going on in a lot of extremely low-quality school systems that it's not immediately obvious that an additional year of schooling in some countries would promote economic growth, ever. Russ: But, putting all that to the side, there are studies that try to do this, that try to measure this impact; they do the best they can. And of course there's a big problem with confirmation bias, which you allude to in your paper. Your paper, by the way, is called "Counting Chickens when they Hatch: Timing and the Effects of Aid on Growth." And your co-authors are Steven Radelet, Rikhil R. Bhavnani, Samuel Bazzi. And that was published in the Economic Journal and won the Royal Economics Society Prize for the best paper in the Economic Journal for the last two years. So, well done. And I really like what you try to do in this paper. Guest: We did not much else but submit it. Russ: Well, I think they probably filled out the journal as best they could. But I really admire what you tried to do in the paper, which is an issue that comes up often in EconTalk. Which is: there's a bunch of advocates on one side of a policy issue; there's a bunch of advocates on the other; they both claim that they have empirical support for their position; and they just yell at each other as if they are scientists and the other guys are hacks. But you actually tried to figure out why they get different results. Which I applaud. So give us an idea. It's a technical paper. We'll put a link up to it. It's available online. Tell us what you tried to do to try to adjudicate. You are kind of a referee between these two competing arguments that aid is worthless and aid is phenomenal. Guest: This paper evolved a lot. Economists justifiably like to complain about the peer-review process and how there's no accountability and it takes forever. But this is a case where peer review really made this paper a lot better. We started by doing statistical analysis on the relationship between aid and growth in our own way. And one of the things that the referees complained about was, look, is this mutually commensurable with other estimates, because there's all kinds of regressions out there; each of them is doing it a different way, different time period-- Russ: Different assumptions-- Guest: Embodying all kinds of different assumptions. And we read those and saw that they were right. So one of the things that we changed over years that this paper was floating around was to throw away our original analysis and start with the three most-cited papers in the literature. Just as a proxy for what's been influential to economists' thinking, start with their datasets, start with all of the assumptions that they had made, and see if we could understand why they were getting massively different results. On the one hand you had an American Economic Review paper by Burnside and Dollar finding a positive effect of aid on growth. On the other hand you had a paper by Rajan and Subramanian finding the opposite, and again with different assumptions, different time periods, different measures of aid, etc. So, the strategy was to try to peel away one by one the differences between them so that we could find the common link. Russ: There was a third paper, right? Is it Boone? Guest: Yes, so maybe a highly seminal paper in this literature, by Peter Boone, was about the relationship between aid and investment, so closely related to economic growth but not exactly growth. Russ: And what did you find? You actually found something that could lead to knowledge, at least possibly. What was that? I'm very hesitant to make a claim. Go ahead. Guest: I have to say, again I'm quite sympathetic with Mankiw's view of this whole literature, it's very difficult, inherently difficult, to tease out true causal relationships in cross-country data, so this is an area from which it is impossible to remove uncertainty. But what we found was that in a few of the cases the results were highly sensitive to the data sets. So, for example, in Boone's analysis, he truncated the data sets, so that if you removed a bunch of observations--and he had removed all of the data points in which a country had received more than 15% of GDP in aid. He argues in a footnote that those are not informative because aid is fungible when it gets to be a large fraction of GDP. We didn't think that was a reasonable thing to do. Russ: What did he mean by 'fungible' there? Meaning it could be used for other things--what was the--do you understand the intuition even behind his claim? Guest: Yes. So, what he was interested in was whether aid increases investments. And you can imagine a world in which low levels of aid that are intended to promote investments do in fact promote, are in fact spent on the thing that they were intended to be spent on. Whereas when a country is awash in aid and it's 50% of GDP, there's no accountability for what happens to it; it starts to leak into other areas or start to displace local government expenditures on things that would have happened anyway and be less well targeted toward the projects that would have increased investments, like building a [?] or a road. Russ: I'd go the other way, though. I'd say the larger it is the harder it is for the fungibility to offset the intention of the aid. Right? If the aid is relatively small it is really easy for me to substitute what I'm already spending on investment and just use the aid. I--I guess it doesn't matter. Never mind. But the point is, it's not a compelling reason for cutting out those other people. Truncating the data. Guest: I like your theory. It's perfectly plausible to me. And, as you say, once you get into this game of I can come up with a plausible footnote which justifies cutting whatever 10% of the data set I want, that's a route toward putting the 'con back in econometrics', as Ed Leamer once wrote. So, what we did was just reconstruct the data base and put those back in, and that made the results much more harmonious with the other papers. In the case of the paper by Rajan and Subramanian, they had begun their analysis in 1980 and had not included observations from the 1970s, even though we couldn't find a compelling reason to do that because observations for the 1970s for all of the relevant variables were available. So we added those back in. And when you do this and make other changes that I could go on and on about if you'd like me to, you find much less conflicts between the results of the various papers.
22:04Russ: But the most important thing that you did was you tried to harmonize the timing of when the aid might have an impact, to try to hold that constant in the analysis, right? Guest: Yes. So what a lot of what these studies had done was chop up time into periods, usually 4- or 5-year periods, and ask: What is the relationship between aid flows during that period and growth during that period? And there's something a little unsatisfying about that. Let's say the period is 1970-1974. Well, obviously aid that arrives in 1973 can't affect growth in 1971. So you are introducing some noise into the calculation just by the way that you are comparing the data. Maybe it makes more sense to ask the question: Does aid that arrives between 1970 and 1974 affect growth between 1975 and 1979? And we did the latter. And that also, we argue, is an important reason why you find different results between different studies. And that's exactly what you'd expect. If you are testing contemporaneous relationships versus short term relationships versus long term relationships, those could all be different. As I mentioned, economic theory doesn't guide us very much about the relative strengths of the relationships we would find in those three kinds of analysis. Russ: So when you tried to make the same assumptions with the same data, what did you find? Guest: We find that in 5-year periods, let's say, divide the time into 5-year periods, that aid arriving in one period is positively correlated with growth in the subsequent period, after removing the effects of characteristics of countries that don't change over time--that is differencing the data, in technical terms. Russ: That is, trying to hold constant things like the culture of the country-- Guest: latitude-- Russ: institutions, the climate, etc. Guest: Exactly. But not holding constant things like wars and political regimes that do change over time within a particular country. Russ: So, you found a positive effect of aid on growth. But small. Guest: A positive correlation that we think is most plausibly interpreted as an effect in this context but really quite modest. We are talking about a 10 percentage point increase in aid as a fraction of GDP in one period being associated with about a 1 percentage point increase in growth per year in the subsequent period. What that means and what that doesn't mean: It doesn't necessarily mean that each dollar of aid is giving you back a dollar or more of growth. That depends on what the full time path of the effect: does it last 2, 3, 4 5-year periods into the future or does it peter out after the first period. We didn't do that analysis. Russ: A 1 percentage point increase in growth rates that persisted would be phenomenally important. Guest: But one that didn't would actually be giving you less economic value than the dollar value of the aid that was disbursed. Russ: Right. Guest: So, this is not a cost-benefit analysis. We were just trying to answer a very simple question, which is, in this difficult empirical environment with very dirty data, is it possible to detect any effect whatsoever? Because a 30-year literature had been very equivocal about that question. Russ: And when you published that paper, you were, I guess poking your finger a little in the eye of those who said it doesn't have any effect. But you probably were also attacking implicitly some folks who thought it was much larger than what you found. So what kind of reaction did you get from the profession and people in the growth error? Guest: As you say, there's a lot of confirmation bias--in economics, in all science, in fact. And I don't exclude myself from it. Unfortunately, science is done by human beings. So, people who had priors in one direction liked the paper; people who had priors in another direction didn't. I think the thing that should shock us and surprise us as economists would be to find that all of the aid funded, dams, roads, malaria eradication campaigns, children's vaccination campaigns, etc. that have gone on since the Truman Administration had absolutely no effect whatsoever on any economic decisions on anyone in the entire developing world. At the same time there is just no evidence that I've ever seen, and certainly not in the data we look at, that foreign aid per se can be anything resembling a growth strategy, per se, for any developing country. We are talking about very modest effects, effects that we find diminish as the level of aid to GDP rises. And fundamentally the thing that is going to affect people's economic opportunity throughout the developing world is going to be other forces; that the reason that the First Millennium Development Goal for halving world poverty was met years ago is primarily because of growth in China and India. No observer of those two countries could claim that foreign aid played a major role in those two experiences.
28:26Russ: So let's turn to migration. And the paper we're going to be talking about is "Economics and Emigration: Trillion Dollar Bills on the Sidewalk?" So, start by talking about why you reference sidewalks. Some of my listeners will know that reference; we've probably mentioned it before. But tell why you gave it that title. Guest: Yes. So, it comes from a bad joke. You've discussed it in previous episodes, maybe with Mike Munger. Russ: I think that's true. Guest: You questioned whether a joke was right for this [?]. Anecdote. Russ: There you go. Guest: The term you preferred, and since it's not very funny: an economist and someone from another discipline, often a sociologist or anthropologist, walking along the street; one of them spies a $20 bill and the economist argues that it couldn't be there lying there on the sidewalk because somebody would have picked it up. This is often a joke that is used to poke economists for their assumptions. Russ: But the underlying assumption behind it that is roughly true is that money that's lying around gets picked up quickly. The larger it is, the quicker it gets picked up. Doesn't mean there's never a $20 bill, because you could be the first one to come across it. But if a lot of people are walking along that street, the odds are good it's not really there--it's an illusion. So when your buddy asks you to go in on a great stock buy because he just knows this company is poised for greatness, he's probably not the first to think of that, and the stock price probably already reflects the information. And that is one of the more useful things you might hear from EconTalk. But remember, past results are not necessarily a guarantee of future results. So, what you are arguing in this paper is that barriers to the movement of people, the important of that dwarfs the importance of barriers to the movement of goods--tariffs and quotas. And not only does it dwarf it; it's enormous. It's much bigger, and it's really big. So, what's the argument and how do you begin to try to measure this? Obvious it has to be very crude. Talk about the underlying economics of emigration. Guest: Yes. So, now we're getting to the really first-order effects on the world economy. So, there is an astonishing fact about the economy. I can see it from my Washington, D.C. office right now. I am looking out at Massachusetts Average., which is full of taxis. A lot of those drivers are Ethiopian, Eritrean men. And they earn hundreds of percent more, in real terms, doing the exact same job, driving a cab, here than they would if they were in Addis Ababa, Ethiopia. Indistinguishable. It's the same car driving around doing the same thing. And that is, from different points of view it's an arbitrage opportunity, the same goods, the labor of a person, selling for vastly disparate prices in two markets. It is a wedge in the graphs of economic supply and demand that everybody's used in Econ101, a wedge between supply and demand meeting. And, it is an opportunity to expand the economic welfare of the world. And what I did in that paper was just look through the handful of estimates of how big that welfare gain opportunity is. If we could somehow harness that potential welfare gain by eliminating that wedge to some degree, moving across that arbitrage opportunity to generate economic value for the world, how big would that change in welfare be? And the long and short of it is, it comes out to something in the tens of hundreds of trillions of dollars. All of these amount to fancy, back-of-the-envelope calculations. Obviously nothing like free labor mobility has happened in recent history around the world, and so they are all modeling exercises. They all embody lots of different assumptions. But I looked through what's out there. I saw that those numbers are very large. And then I discussed a research agenda around that issue: understanding better the size of the opportunity, the kinds of institutions that would be needed to realize that opportunity, addressing research that challenges the size of those estimates, and so on.
34:10Russ: So, a cab driver in Ethiopia is currently making $4 a day; and they come to Washington, D.C. and they make $100 a day. They are obviously a lot better off. But if every cab driver in Ethiopia came to the United States, a bunch of things would start to happen. So one of the challenges of these kind of back-of-the-envelope calculations is, you usually assume that the impact isn't linear. You can't just keep getting that 25-fold increase in well-being. But it probably doesn't go to 0, either. So, how do you deal with that? What are some of the assumptions you have to make in trying to assess those potential gains? Because we assume--I assume you assume--that not every Ethiopian is going to move to the United States, but a lot of them might. So you have to make a guess about that, too. Guest: Yes, yes. You are talking about the difference between the gain to the marginal migrant and the gain to the average migrant in areas of large migration. What would be the gain to the next person who comes, versus the typical person who comes, under a mass-movement scenario? And those could be very different. In a paper with Lant Pritchett and Claudio Montenegro called "The Place Premium," we tried to estimate the marginal effect--what is the gain to the next migrant who comes to the United States from each of 42 different countries. And it is gigantic. It is typically 300% of that person's real income in the country they start from. In some countries it's over 1000% in real terms. But as you say, those are the gains to the next person. If there were to be a mass movement, how far down the labor demand curve might the destination country move as there's a greater and greater supply of labor--that can only be an area of speculation. I think that is an area of critical importance to economic research, and that's one of the reasons I wrote that Trillion Dollar Bills paper, to nudge economists towards studying these issues more. As I talk about in that paper, there are a lot of reasons to question the automatic assumption that comes to people's minds that somehow we would be driven all the way down the labor demand curve to a Dante-like scenario of death and destruction. One of those is U.S. history. So, in 1905 this was a country with just over 75 million people in it. A hundred years later in 2005 it was over 300 million people. Russ: And look how poor we are. Guest: Absolutely. So, first of all we had an astonishing increase in the real living standards of essentially everybody in the country. We had exactly 0 change in unemployment between those two years--unemployment in 1905 and 2005 were both about 5%. So there's something wrong with a simple model of the long-term change in countries in which more and more labor just drives you down the labor demand curve. All kinds of things are going on. But even in much shorter term scenarios, that mental model just shatters against reality. And one that I often mention is the end of apartheid in South Africa. There were parts of the Republic of South Africa (RSA), what people call the Republic of South Africa now, that were not considered by the RSA, part of the RSA before the end of apartheid. So, both Botswana for example, Siskai, Venda, were considered at different points in history separate nations. They had their own institutions, they had judicial systems, stamps. Other countries as far as I know, nobody who accepted the RSA recognized those countries. But certainly the RSA did. Suddenly that distinction was eliminated, through a series of epochal legal changes in the early 1990s. Suddenly you have this population that was about 7 times as large as the white population, depending on how you define non-white. It was about a sixth the average income of the white population. And suddenly all of decades-long barriers to labor mobility were eliminated. There had been a complex system of past laws, who could work in the white areas, in what occupations, even during what hours of the day. In fact they went way, way beyond eliminating those barriers. People got entitlements, people got the vote for President--way, way more than just labor mobility. And if that were to drive whites in white areas down a labor demand curve, we would have seen now, after 19 years and the process of opening finished, we would see now that whites had been terribly impoverished by this huge increase in labor supply. And not just a huge increase in labor supply--a huge increase in low-skilled labor supply in white areas. And the bottom line is absolutely nothing like that happened. There are a lot of studies of what's happened to workers in South Africa since these barriers went away. Their usual concern is to try to figure out how much convergence between white and black living standards there has been. But a notable fact in all those studies is that there is no decline at all in physical white living standards across the board. It's not that poor whites have gotten worse off or is compensated by some sort of gains to the rich. It's that across the entire distribution of white incomes you don't see any sort of downward convergence towards the level of black income. Apparently there are multiple labor markets in South Africa. They are segmented for reasons other than geography and just a complete geographic opening did not mean that whites and blacks were competing against each other in exactly the same labor markets. These kind of examples--I don't mean to suggest that we know anything about what would happen under greater labor mobility in the world definitively from the South African example. I raise it just to try to urge people to question any simple model that might be in their mind of, hey, there's one labor market in the United States and the more people from Vietnam and Peru who come here are just going to drive us ever downward into a low-wage, high unemployment hell.
41:53 Russ: And I think that is a--I won't say it's a common view. I would say it's a common argument. I don't know how widely the view is actually held, but that argument is often invoked as an argument against increases in immigration. And I understand the argument in the following way. Let's think about it--and I sometimes get criticized for being a free-marketer in immigration because, well, after all there's not as much of an impact on me. And that could be because I have tenure--which I don't right now, so that argument doesn't matter. But they'll say you'll have more competition. And a lot of people say, then, okay, you free marketers, you need to advocate for the free movement of not just low-skilled labor, which doesn't compete with you, but high-skilled labor. And I certainly support increases in the mobility of both low- and high-skilled labor. But having said that, I assume that if there were a huge influx of economists from foreign countries who could compete with what I do that that could affect my wage. My answer to that is: I don't think I have the right to keep them out. There's a moral issue here, which I want to put to the side, but I want to mention it and then we'll put it to the side. But on the economics issue, the key claim is that by letting in lots of low-skilled workers, for example--there are people who say, well, I'm in favor of more immigration as long as it's the right kind. And I'm in favor of all kinds. But for those who make the distinction, I'll say I'm in favor of high-skilled immigration. It's okay for Andy Grove to come here, the founder of Intel, or for the parents of Sergei Brin, who founded Google--they came here and Sergei was able to go through the American education system for a little bit and then create Google with Larry Page--they're okay, but these low-skilled workers, we can't have them because all they do is they compete with other low-skilled workers; those workers are already having a tough time, and therefore it's wrong to let these immigrants come in. And I understand that argument--I don't understand it on moral grounds, but on economic grounds, I understand the potential for it to be true. What evidence for it--you gave the South African example, it's a fascinating example. What evidence do we have for it that it's true or not true? There are again people on both sides of this issue, empirically, how would you begin to think about which one's right? Guest: Let me start by saying that in my Ph.D. cohort of, I believe it was 30, there were 3 Americans. And when you look at the other 27, when you find someone like Justin Wolfers, who I understand is a friend of EconTalk; Stefano DellaVigna and Marios Angeletos and other geniuses, you might think that there should have been 2 Americans rather than three. Yes. I don't only talk about low-skilled immigration; I talk about high-skilled immigration as well. I think it's quite obviously beneficial to the economics profession and others, that my class and my profession is drawing from the whole planet. Let's talk about low-skill immigration. So, it is fascinating to see the descendants of almost exclusively low-skill migrants in the United States debate whether or not low-skill immigration is a good thing. In 1940, the U.S. Census records what fraction of the white foreign-born had a high school degree, and it was 12%. Twelve. 88% of adults, white, foreign born, in 1940 were what today would be called high school dropouts. So, really talking about the immigrant ancestors of myself and nearly any American--I shouldn't say nearly any--the vast majority of Americans listening to this show, whose parents came several decades ago or more--we're talking about low-skilled migrants. They built this country; they created the entire economy of this country; their descendants are the labor force of the economy. And yet, as you say, there is an active public debate about whether anybody who doesn't have specialized knowledge in some field is really adding anything to the economy. And it always puzzles me. Adam Smith could have written a book about how French and Portuguese exports, when they come into Britain, just drive down prices and that's bad for the economy. He wrote something a lot more sophisticated than that. Ricardo expanded on his argument. Economists have been broadening people's view of the economic consequences of actions like these for centuries. And yet a lot of the discussion about immigration right now among top economists, really supremely brilliant people, focuses on this issue. One of the most influential papers in the literature, in the Quarterly Journal of Economics a paper by George Borjas, whose title is literally "The Labor Demand Curve is Downward Sloping." With 'is' in italics. When there's more labor, prices go down. That is a stunningly incomplete analysis of the impact of low-skilled labor on the country as a whole. I don't think that Borjas himself intended it to be a complete analysis of the economic impact of labor; in fact, a speech he gave at the World Bank in February of this year he definitively stated that the results of this literature on whether or not low-skilled labor drives down wages of low-skilled workers has no policy implication at all. Nothing whatsoever, I think were his exact words, because he's well aware there are many other considerations. What are some of those consideration?
48:48Russ: Let me stop you there for a minute. So, we've allowed in the last, oh, 25 years roughly, an increase in immigration from relatively low-wage countries, and the people who come here are predominantly low skill. And my argument has always been that's great for them; just like low prices of imports are good for most Americans, that's good for people who hire lower-skilled labor. Guest: Yes. Russ: Whether that's a customer who wants a ride in a taxi or somebody who wants their house repaired or cleaned or their lawn cut, things that require relatively low skills, that's been good for most people. But it's also true, it would seem to me, and this--I'm with you and I'm with George Borjas partially--this is not the crux of the matter, but it is one aspect of the matter that I could imagine that people who do compete in that niche of low-skilled workers that were already here could be harmed. Maybe in the long run. Certainly it's plausible that they could be harmed in the short run. And my thought about that is--and Borjas's findings that show that the demand curve slopes downward, which is again a fancy way of saying that when there's an increase in labor the wage goes down, holding everything else constant--I'm perfectly okay with that, in the sense that: his findings are quite small. The impact on the wages on the low-skilled workers who are harmed, the impact is quite small. It's mostly high-school dropouts. And it seems to me that instead of keeping out immigrants, we ought to try to keep people in high school. That that would be the appropriate policy response to that empirical finding. And yet that is not always the case. So, react to that point. Guest: Just as a footnote, you could even argue that it is antithetical to the goal of keeping people in high school to prop up the wages of people who do not finish high school. But let's get to the real question: What are the impacts? It's obvious that labor demand curves are downward sloping. As you say, that tells you something about the microscopic proximate impact on a worker of another worker competing for a job. This afternoon, if you and three other people are answering an advertisement and one of you gets it, well that might affect the wage that you would insist on for getting a job next week and might tend to drive down wages. But there's so much else that goes on in an economy. And this is where I get frustrated with the literature on immigration that is focused exclusively on labor economics. So, what are some of the things wrong with that? So, one is that native workers and immigrant workers, even people who look similar in statistical data, let's say people who have less than high school education, are not in competition with each other to anywhere near the degree that a lot of people assume they might be. And a completely separate thing is that the demand for the labor of low-skilled natives is bolstered by all kinds of economic forces to which immigrants contribute. That is, the jobs that a lot of low-skill immigrants have now exist because there has been low-skill immigration to this country. So, let's talk about each of those separately, because they are completely separate issues. I've been writing recently about a point that Lant Pritchett of Harvard and also the Center for Global Development brought up, which is the massive disparity between the supply of low-skill Americans and the Bureau of Labor Statistics (BLS) projections for the demand for low-skill workers in this country. So, if you take the U.S. Bureau of Labor Statistics projections for increase in employers' demand for certain occupations over the coming decade--the do it for 2010-2020, but it's more or less the decade to come--and you look just at the occupations which do not require a high school degree--we're talking about basic personal care, care for the elderly, basic childcare; we're talking about warehousers, back-of-the-restaurant food service workers, basic construction workers, landscapers--these are some of the highest growth occupations in the country. If you add up just those and the others in the top 20 which do not require a high school degree and ask, oh, for the next decade how many more jobs will there be, how many more people will employers be demanding in those areas, it comes out to about 3 million more people. And if you look at the Census Bureau's projections for how many working age Americans there are going to be entering the labor force in the same time period, it is 1.7 million people. So, think about that for a second. Three million is the demand for additional low-skilled workers; and the entire growth in the whole U.S. labor force is going to be about half that amount. Russ: But if I'm a low-skilled worker, that's great. And if we keep out those low-skilled immigrants, that means I'm going to get a much higher wage. And we're going to have a lower poverty rate in the United States. Again, I don't find this argument compelling but that's the argument that would seem to follow if you believe those BLS projections. Guest: Sure. What about the employers of the other half? Let's say hypothetically every single American entering the labor force, every last person, took a low-skilled job. No reason to go to college, no reason to even finish high school. For any of them. All of the new American labor force entrants over the next 10 years did these jobs. That would leave half of them unfilled. Those are people with nobody to take care of their grandparents, nobody to cook food at the restaurant. That's impoverishment of the whole country. That would kill economic activity. Russ: Well, I'm not so sure. I take the point that there would be an impact on the non-skilled population because the prices would be higher. But we assume there would be substitution in various cases. I don't find that argument to be a very compelling argument for immigration. Just like this argument you hear that if we didn't allow immigrants in, we wouldn't get our food picked. We'd get it picked; it would just be by people who were now paid a lot more. There might be less of it picked; it might be picked by machines now. We'd figure things out--the economy I think would adjust, just as it would if there were a plague, God forbid, or other things that restrict the labor supply that we didn't have policy control over. Guest: Sure. To a degree. Russ: So to me, I personally don't like that argument. I don't find it convincing. Give me something else. Guest: To a degree that's true. It is also true that if the price of U.S.-picked cucumbers went up by a certain amount, there wouldn't be any more U.S.-grown cucumbers. And that would be to the detriment of U.S. GDP. I's also true that if childcare gets to a certain point there are parents who are not going to enter the labor force. This is something that Patricia Cortes at Boston U. has shown in a couple of excellent papers. It's also true that--I recently had the experience of being an employer of elder care workers and there was a certain amount of retirement savings. And if the wages we were paying were doubled, that means that we would have gotten a lot less care and would have had to have gone through a lot of machinations to make that happen. So, yes, some substitution can happen to some degree, however-- Russ: And low prices are really good. I'm all for it. For the economy as a whole. So, we agree on that.
57:17Russ: I think you had another point you wanted to make. Guest: Yes. All I want to argue--I don't want to argue that this is some sort of mechanistic relationship. I am just arguing that there is a lot less competition between these people and U.S. low-skilled workers than people tend to assume. I don't think that most people realize the gigantic demand for low-skilled workers that's happening in this country and how much greater it is than the number of Americans willing to do low-skilled work right now. Russ: Fair enough. Guest: But more important is this second concern, which is that low-skilled workers have indirect economic effects on all kinds of other things. So, when low-skilled workers from abroad and low-skilled workers here work together, they often specialize in different tasks and make each other more productive. You see that on farms, in which you would have two people who might in a George Borjas regression look similar, having a high school or less education and be 24 years old, but they tend to specialize in different things, often because of language skills and for other reasons. You find that there is also complementary between the skill groups. I mentioned the work of Patricia Cortes suggesting that wherein childcare for low-skilled workers is cheaper, more high-skilled women, people with doctoral degrees and law degrees tend to partipate in the labor force, work outside the home and that grows the economy. Patricia Cortes has another paper in the Journal of Political Economy pointing out that low-skill immigrants have driven down the prices of some essential services, which raises the real wage of all other workers. You mentioned the relationship between education choices, the choice to stay in school, go to college, or not, and the price of low-skilled labor, the price of high-skilled labor. Jennie Hunt, who is currently the--she's at Rugters but right now she's at the Department of Labor, Chief Economist--has a recent and fascinating paper in which she argues that the availability of low-skilled labor and the consequent effect on the structure of high-skilled versus low-skilled labor have encouraged people, unsurprisingly, to stay in school more. I believe she looks across U.S. states, which states have gotten more and which states have gotten fewer low-skilled workers in recent years, and finds that in places where low-skilled labor is more available, U.S. workers, not wanting to compete in that market, have tended to get more educated. And that makes the entire economy more productive as well. And we also mentioned a few minutes ago that the existence of some subsectors of the economy depends on the availability of both skilled and low-skilled labor; yes, there are parts of U.S. agriculture that can be mechanized; in the short term there are many parts that there is no profitable way to mechanize right now. Lettuce picking, melons, cucumbers, apples--there are machines that help; fully mechanizing it and getting products that are of the quality that can compete with imports is not now possible. And at least in the short or medium term, those industries would not exist. There are lots of other sectors that really can't be mechanized. Most aspects of elder care are not going to be mechanized any time soon. They can't be traded. They are person-to-person interactions. Lant Pritchett called them hard-core non-tradeable goods. It's hard to envision how they are going to be substitutable by technological change or international trade any time soon. We really have a long list of ways in which low-skill labor through various channels, by creating sectors of the economy, by changing people's education choices, by encouraging female labor force participation, by making high-skilled workers more productive, by making other--through specialization--low-skilled workers more productive--all of these things are in economic terms shifting out the demand curve for those people's labor, even as they are being driven down that curve. Russ: Offsetting potential effects of the labor supply change. Guest: Exactly. And this is why, in the state of the art and the analysis of the overall effect of low-skilled immigration, which is right now a paper in the Journal of the European Economic Association by Ottaviano and Peri last year finds that even for low-skill workers, even for less than high-school educated workers, the cumulative--I believe it's 15 years, 15 or 20, I don't remember exactly--of immigration in the latest data has a positive effect on the wages of workers. And this is through this offsetting shifting of the demand curve, even as obviously workers are moving down it.
1:03:07Russ: Yeah, I have to say, again, as emphatically supportive as I am about allowing people to move freely across borders, the empirical literature trying to tease out these effects making such incredibly heroic assumptions about the nature of the production process--the only thing I would say about this literature: I think it's important to emphasize that even the people who are hostile to increased immigration, who presumably have a tendency to make the assumptions--again, there are many to make, trying to measure and tease out the effect of more immigration--do find a relatively small effect. And relatively small is--correct me if I'm wrong--that the increased immigration of low-skilled workers to America in the last couple of decades may have lowered high school dropout salaries by 5%. And that seems like an incredibly small price to pay to encourage people to stay in high school and to come here and make people have better lives. Guest: If I'm remembering correctly it's 8%, in the 2003 Quarterly Journal of Economics paper by Borjas. Russ: That's why I said it's 'like' 5%. That's my confirmation bias. Sorry. 8%. Guest: It is on the order of 5%. I think you'd find 8%. For the average American worker you'd find 3%. And really this is a great point you are making. This is worth pointing out that the big debate with Borjas and others on one side and David Card and Giovanni Peri on the other side is over numbers that are tightly circumscribed around 0. Is it plus 1%, is it -3%? Because, as you said, we are talking about the cumulative effect of decades of migration. This is not a per year effect. This is decades and decades of all immigration, authorized and unauthorized, in all sectors from all countries. Borjas's estimate is 3% decrease at the end of that period--I think it's .14% per year, whatever that works out to as an annual [cumulative?] change. These are just minuscule compared to the other things that are going on in the economy, like, I don't know, the Great Recession.
1:05:35Russ: So, we're short of time here. Let's try to finish. I want to just mention--we haven't talked about it at all: Obviously there are many non-monetary aspects to immigration and emigration. I actually think that people tend to underestimate the positive benefits of those and tend to focus on those they perceive as negative benefits. But we're going to keep that to the side. I just want to mention, we all understand that the monetary impacts, financial impacts, are not the only thing we care about. But on just the economic side, do you have a view on what the ideal policy should be toward emigration and immigration? Without getting into the current legislative issues, which I have personally not been following and are quite complex: Do you think we should distinguish between low skilled and high skilled immigration in our immigration policy? How much more open should we be? Do you think there's any politically plausible way we'll get to a more open set of borders? Guest: People often ask me if I am in favor of open borders. And I take an ignostic[?] approach to that question. That's kind of a strange term but by it I mean that I think the question is ill-posed. Russ: I like that. Guest: I don't understand what people are asking when they ask it. Do they mean anyone from everyone in the world should be able to freely move to every other spot on the world? Well, I don't have that right right now. I don't know of anybody who has ever had that right, actually. I can't walk into your house. I can't walk into a military base. I can't go sit on the street--police would remove me after a while. My movements are tightly regulated. Property markets are regulating where I can pitch a tent and live. If open borders means absolutely free movements then we certainly don't have that in this country. If open borders means anybody can come get immediate access to any public service no matter whether or not they've paid into the system, that's not something that I enjoy either. I don't get to take Social Security money out unless I put money in. That's also true for immigrants, by the way--you can't get any money out of Social Security until you have paid into it for at least 40 quarters, that is a minimum of a decade of work or more. If open borders means absolutely free movement of people without any sort of tracking of who they are or any sort of concern for free riding in public services or any concern for trespassing on private property, then, no. Open borders doesn't exist in any space that I've ever seen. I don't really want it to exist. Before we talk about open borders, I need to know what that means. Usually people mean something like a great relaxation to the policy barriers that people face right now. And I think a useful thought exercise is to ask what are the effects of those barriers relative to their absence. That is, if we didn't have them. Take a particular policy barrier--look at the world if we didn't have it and ask if we are better off from having it. We didn't have a Chinese exclusion act until 1883, when Congress voted to prevent the immigration of essentially all people of Chinese ethnicity--not just Chinese nationals but people who were ethnically Chinese. And that was pretty much the rule for the next 70 years. Now, we had a world without that law; we had a world with that law. Do we have any scrap of evidence that that law made us better off? Well, I've never seen one. And I think it would be reasonable to doubt that that made anybody in the United States economically better off. Russ: And it certainly made a bunch of Chinese worse off. Guest: It made all kinds of people worse off. You could ask the same about lots of provisions of legislation. I know you don't want to get into the weeds of legislation, but one aspect of a bill which has been passed by the U.S. Senate and which the U.S. House is now considering its response to is temporary work visas. And those visas, as outlined--there's a low-skill work visa that's been proposed called a W-Visa. There's a high-skill, a continuation of the current high-skill work visa called the H-1B. Both of those in the legislation have both caps on the number of visas that can be given each year and a labor market test that potential employers are required to pass. That is they must prove to the Department of Labor that they could not find Americans to fill the position in question. And if you think about that for a minute, let's say that as proposed in the Senate bill, there's a cap of 200,000 maximum low-skill W Visas that can be given in any given year, and I am the restaurateur who wants to hire the 200,001st worker. And I duly advertise the position; I prove to the Department of Labor that there wasn't any American willing to do this job, even though I as an American can add value to the U.S. economy by employing somebody in that job. And then I'm told that actually, this is the 200,001st, so you can't hire them. That doesn't make any logical sense. Nobody gains from that--no U.S. worker gains because I couldn't find one willing to do it. I, the employer, is hurt. The whole economy is hurt. The Department of Labor doesn't get anything out of the deal. The potential migrant certainly loses. If it were to be a typical low-skill Mexican, in the place-premium study I mentioned before, that person would have tripled or quadrupled his or her living standard by filling a job like that. Everybody loses in that scenario. So if more open borders means getting rid of provisions of legislation that don't help anyone and hurt everyone, then movements toward open borders fall into that class of things that economists love to search for--which are Pareto-optimal things that make some people or everybody better off without making anybody worse off. Russ: In your ideal world, would you make any distinction between low and high skill labor, in terms of immigration policy? Guest: I have never seen good evidence that planners in Washington are better than employers at doing that. Certainly distinctions need to be made. There could be very high demands for some kinds of low-skill labor relative to some kinds of high-skill labor. There could be high demands for some kinds of high-skill labor to some kinds of low-skill labor. That needs to be determined by some process. Then you have to ask: What is the best process to make that distinction? And as far as I know the people who are best able to determine how many cucumber, lettuce, and melon-pickers are needed are farmers. The people who are best able to determine how many Information Technology (IT) engineers are needed in Silicon Valley are the people who are at those firms day to day, knowing exactly what skills and type of person is required. And the ability of people who are not there to do it much better than them is not proven, to put it in the lightest possible terms. Russ: Well, we'll close on that Hayekian note about the particular Hayekian knowledge of time and place.

COMMENTS (26 to date)
SaveyourSelf writes:

Complex questions. Complex answers. Interesting conversation.

Nice touch at the end when Michael Clemens equates immigration policy to centralized-labor market-planning. In Clemen's words, "The ability of people who are not there to do it much better...is not proven." Economist humor. Funny on so many levels.

Bogwood writes:

My agenda,reciprocal immigration,is not yet mainstream so will minimize the comment. There is an optimistic ecological view to the podcast for those worried about AGW,species extinction,and resource depletion. Reduced migration would have a leveraged effect on global growth and an ecological benefit. It is nice to see proof of this even with all the precautions about the data.

There should be more migration but only if reciprocal. You cannot add to watts per person by bringing people to the USA without sending some one to the country of origin.(no knee jerk reaction unless you've read Clugston on Scarcity).

Jane writes:

Brilliant.

Thanks.

Jonathan writes:

I loved the talk overall. It was very insightful and thought-provoking. I am not a great test because it confirmed my priors but I think he gave compelling reasons.

Another good thing is that he was nuanced and recognized, generally, that while opening immigration policy was a good thing, there were limits to the benefits and some offsetting effects.

Until the end when he said

there's a cap of 200,000 maximum low-skill W Visas that can be given in any given year, and I am the restaurateur who wants to hire the 200,001st worker. And I duly advertise the position; I prove to the Department of Labor that there wasn't any American willing to do this job, even though I as an American can add value to the U.S. economy by employing somebody in that job. And then I'm told that actually, this is the 200,001st, so you can't hire them. That doesn't make any logical sense. Nobody gains from that--no U.S. worker gains because I couldn't find one willing to do it. I, the employer, is hurt. The whole economy is hurt. The Department of Labor doesn't get anything out of the deal. The potential migrant certainly loses. If it were to be a typical low-skill Mexican, in the place-premium study I mentioned before, that person would have tripled or quadrupled his or her living standard by filling a job like that. Everybody loses in that scenario. So if more open borders means getting rid of provisions of legislation that don't help anyone and hurt everyone

It may be that 200,000 isn't the right cap but he overstates the case saying the provisions "don't help anyone and hurt everyone" for two reasons.

First, he said that the restauranteur "proved" that he couldn't find an American who could fill the position. But, that is not really true.

I sponsored someone for an H-1B visa (which he received) so apparently I "proved" that I couldn't find an American. But, of course if I kept looking surely I could have found someone else (perhaps not AS capable but still capable) or PERHAPS AT A HIGHER SALARY.

So, I agree, at Clemens said for most of the talk, that the benefits seem to outweigh the costs by a large multiple but part company when he says that limits to those sorts have NO costs.

Great talk. As said above, complex issues with complex answers (for the most part).

David writes:

What I don't understand about immigration visas is the part about employers proving that they cannot find someone to fill the job. I don't see how with a U3 unemployment rate of ~7.5% and U6 rate about twice that how any company can claim they cannot fill jobs and need to import workers to do so. Shouldn't the question be why are you either a) not willing to pay the 14.7% a decent wage or b) not willing to participate in training the 14.7% to meet your labor needs?

John Berg writes:

First, Imagine the US as a closed system and new job-seeking people(and new citizens) resulting only from births to existing citizens. Shouldn't our policies ensure jobs for our children?

But in fact, isn't the game for those in power to figure out creatively how to gain rents from the US commons, like renting out the White House's Lincoln bedroom?

But what's the point of sweating "public policy" if the Executive Branch can choose to enforce, the Legislative Branch are unaccountable because they merely authorized experts to make the laws, and the Judicial Branch can take control from the people(voters) and declare laws.

Besides, I going to have to budget time over the next few months to have an "honest dialogue" on the White/Black thing.

John Berg


Russ Roberts writes:

John Berg,

We don't live in a closed system. And there is no evidence that keeping out or letting in foreigners has anything to do with unemployment rate. I want my children and other people's children to have rich and fulfilling lives. I think letting others come here helps that goal.

rhhardin writes:

The economic issues are simple except partial economic theories are used as cover for multicultural problems.

The problem is that non-assimilative immigration gives an unstable political system, once political spoils are to be had easily.

Not to mention hostile cultures, not merely non-assimilating ones.

It's our country, and we get to say who comes here, who we trust and who has to leave.

That's a political decision, with bad economics as a cover for various interest groups with stakes in the game.

So being enthusiastic about the economics may be too quick to a conclusion.

Greg Pandatshang writes:

I felt that the discussion of immigration made a some interesting points but I believe it also had a hortatory tone which was inappropriate in context. For example, one thing that’s clear about South Africa is that it’s a highly atypical country. That fact by itself certainly doesn’t vitiate the findings about wages in the South African labour market(s), which is indeed fascinating. On the other hand, I’m skeptical of generalizing this result. What’s called for is more study, both of a wider variety of cases and in more detail. However, it seemed to me that the implicit message I got from Russ and Michael is that what’s called for is more immigration. I remain doubtful of that.

Another example, Michael finds it "fascinating to see the descendants of almost exclusively low-skill migrants in the United States debate whether or not low-skill immigration is a good thing". However, his brief description of the topic is quite one-sided and not very nuanced, which leads me to believe that Michael has not actually thought about it in depth. Not so fascinating after all. This word choice to turns out to be little more than snarky sarcasm. Your audience could have done without this snark and certainty.

Michael's answer about "open borders" was odd and frustrating. For one thing, this wasn't actually the question that Russ had asked. However, Michael brought up a different question, "do you want open borders" and proceeded to answer that instead. Most of his answer took the form of a lengthy straw man in which he explained what he does not favor. And then he talks about greatly relaxing the existing rules. Why have any immigration rules at all, beyond same rules of law and private property that govern everyone in a country? Back when I was a hardcore libertarian, that's exactly what I was in favour of. None of the arguments presented in this episode imply a limit to how much immigration is good for a country. If this were philosophytalk, I would be quite curious to know on what grounds Russ or Michael could support any limits to peaceable immigration at all.

Michael said several times that existing immigration policies benefit no one. As a rule of thumb, when an economist tells you that something people do has no beneficiaries, my advice is to back away slowly, avoiding eye contact, and then, when you've reached a safe distance, turn around and run for it.

It's to Russ's credit that he expressed his skepticism of certain arguments for immigration and his frank assessment that the studies involved are unreliable because they require researchers to make a lot of assumptions (I would have liked to hear stronger push back on the concept of employers being "unable" to find any workers for a certain position). Immigration policy is essentially a political decision based on a variety of political preferences and no small amount of handwaving. It's to Borjas's credit that he understands that his studies are far from dispositive about policy decisions. Do Roberts and Clemens understand the same thing about the facts and ideas presented in this conversation?

J. Hohenauer writes:

Very interesting and insightful.
Concerning the liberalisation of people moving / working transnationally the EU could be worth studying. From my personal point of view this seems to be a success for receiving economies and the people moving. And for those living in the receiving regions. I wonder if (or at which price) you would someone for plumbing et. al. in Vienna, would there not have been such an important influx of people from the eastern EU countries.
It would be interesting, to cross check any publications on this with Michael Clemens findings.
And then I just wanted to add one additional aspect. How do the countries really fare from where people emigrate. I understand that the cheques sent home are a benefit. But how do countries cope if an important share of their highly qualified population moves abroad. This is for example the case in the field of health care, where there develops a dramatic scarcity of health professionals in the eastern countries due to migration.
Just to give an anecdotal idea of the magnitude of the movement of people: I met a Romanian (about 35 years of age and in Austria for more than 10 years now), who told me that out of his high school class mates only one had stayed permanently in Romania. All the others had left Romania at least for long periods or focussed on living and working abroad. One could fear that only the less qualified and less motivated stay in their home countries.

Robert Reeve writes:

Greg Pandatshang’s comments on 18 July covered most of my concerns with this episode of Econtalk. However, I would like to offer some personal observations about the outcomes of the changes in the South African labor market regulations in the mid 1990s consequent on the change from the minority white (apartheid) government to the present majority rule government. My comments are not intended to make any moral points.

By way of background, I qualified as an accountant and worked in public practice and in industry in Durban (the second largest city in South Africa) from 1961 to 1973. From 1973 to 1980 I worked as an accounting academic at the University of Natal in Durban. I emigrated to Australia in 1981 and worked at two major Australian universities for 25 years. The precipitating event in my decision to emigrate was the death of Steve Biko under interrogation by the security forces in 1977.

The changes in the South African labor market regulations in the mid 1990s involved the removal of regulations that strongly favored whites to regulations that strongly favored non-whites (and women). I have visited Durban five times since the mid 1990s. On these visits I have observed an increasing number of homeless whites (sleeping in old cars) and whites begging (including young women with children) or semi-begging (washing windscreens at intersections). Also whites involved in informal trading (selling small items at sidewalk stalls or at intersections) and as car guards (semi-formal parking attendants). Under the apartheid regime these whites (especially the men) would have been able to find menial (but regular) employment in one of the state enterprises (for instance the railways). This suggests these low skilled whites are now not able to find any employment and doing whatever they can to get by.

Also, a large proportion of the white employees that worked in the public sector and state enterprises (police, military, hospitals, schools, universities, railways, ports, power generation) have been replaced by non-whites. At middle and higher management levels this effect is very pronounced. To a slightly lesser degree this has also happened in large corporations (banks, insurance, mining). Example: an ex South African acquaintance of mine, now in his early 40s, has emigrated to Australia and works for a major international oil company. He worked for the same company in South Africa in the marketing area from the time he graduated. He then went on an overseas tour of duty with the company working in Thailand and Vietnam for six years. When the tour of duty was completed he was advised that if he returned to the company in South Africa it would have to be in a technical position because paths to management positions were not open to white male employees for the foreseeable future. Fortunately for him the company found him a marketing management position in Australia.

In the light of the above I consider that Michael Clemens’ claim that there has been “no decline at all in typical white living standards across the board” very difficult to accept.

John Berg writes:

My first comment to this podcast made several assumptions which I would explicate.

A nation has secure borders and a government which enforces formerly defined policies arrived at in a democratic way in order to control changes to that nation's demographics.

Citizenship is a formerly defined state of residence which attributes equal rights and responsibilities to citizens.

Citizens have a right to vote and to require their constitution and laws to be enforced by their government.

Policies motivated by compassion always have unintended consequences. I am still proud of the Marshall Plan.

John Berg

Don Rudolph writes:

My first question would be do you think it is better to live in a more egalitarian society than a less egalitarian society?

My next question would be are you looking for what is better for the world or what is better for the USA?

My personal preference would be for a more egalitarian society. Those who command a high wage must be doing something highly valued by society that can't be done by many other people. Adding someone like that would add wealth to the community. The valuable thing this person does would become more affordable to the society as a whole as the wage for that job would slowly fall.

When a farmer says I can't find anyone in the USA who wants to pick my crops, he is saying I have a preconceived notion as to what this kind of labor is worth. I am baffled why I get no takers when I offer low wages for back breaking work in the hot sun. I say hurray, we have reached a level of wealth in this country where we don't have to live a life little better than the former plantation slaves.

I question the wisdom of open borders. Part of America's wealth is in the commons we all share. When questions like shall we keep these trees for recreation, or cut them down to build houses, came up I think we know what the answer would be if the USA had a billion people.

John Berg writes:

In my July 19th comment to this podcast, both uses of "formerly" should be replaced with "formally." My South Philly accent betrayed me.

John Berg

Gandydancer writes:

Clemens' arguments for trillion dollar returns on open borders (he evades the term, going off on a spiel in which he tramples various strawmen, ending with a denial that he is in a favor of providing benefits to immigrants that they haven't paid for that is unaccompanied by any suggestion as to how this can plausibly be avoided) reminds me of the arguments for the stimulative effects of building sports stadiums. Import Somali herdsmen, their incomes go up by an average of $7500 (that's the number in the first linked paper), and the world is a better place by that cumulative number. I mean, he complicates that a little in his paper to give an imitation of analysis, but that's basically it. Sheesh.

And Clemens admits, "An increase in migration... leads to a relatively small decrease in the wage rate for the high-income country... Migrant workers gain; nonmigrant workers gain in the low-income country and LOSE IN THE HIGH-INCOME COUNTRY" (emphasis added)

Note that he claims the loss for American workers would be small, implausibly (see letters above) offering South Africa as proof of this, but the justification for an experiment on your fellow-citizens' well-being when the expected result is bad is completely lost on me. I don't really care if the teeming masses of the world are made better off by giving them even more open access to the United States' already deteriorating social capital. The job of a government is to look after ITS citizens well being, and the rent seeking of those who profit at their fellows expense (the cheap labor lobby, in this case) ought to be excoriated. Roberts' idea that this is somehow a moral imperative leaves me utterly repulsed.

Michael Clemens writes:

J. Hohenauer: Thanks for this thoughtful comment. I wanted to offer some evidence on whether or not the migration of health professionals is a substantial cause of health worker shortages in poor countries.

Jean-Christophe Dumont and his colleagues have calculated the number of African physicians in OECD countries, in a massive project using raw census data. They show that the number of African physicians abroad is very small relative to the number of physicians Africa needs.

Suppose, for example, that OECD destination countries were to slash by half the immigration of doctors and nurses from Africa, as well as somehow force the return of half the African doctors and nurses who have immigrated in the past. Even this extreme and draconian act would only eliminate 6% of Africa's shortage of doctors and nurses, as estimated by the World Health Organization. The same figure for Southeast Asia is 5%.

You can find these estimates here: OECD. 2007. International Migration Outlook: SOPEMI 2007 Edition. Paris: OECD, p. 178.

Inside and outside the health sector, the roots of skill shortages lie in complex problems of economic and institutional development largely beyond the reach of migration policy.

On the other hand, it's clear that the policy outlined above--forcibly restricting African doctors' movement en masse, a policy few of us would accept to be applied to ourselves--would greatly harm them, and their families. There would also be numerous indirect negative effects of such a policy, which I discuss at length here.

John Popham writes:

Our (US) national immigration policies should in no way be based on considerations of the welfare of non-US citizens. Acting as a population escape valve for ineffective governments means we encourage (and subsidize) their bad behavior, postponing necessary political and/or economic adjustments in the home countries. Though some number of Chinese may have been less well off under US immigration policies of the past than they would have been under a US immigration policy that allowed Chinese immigration, earlier Americans rightly determined that was not a concern of the US government. That was a matter to be addressed by the Chinese government (and, by extension, the Chinese people themselves). The same could be said of current day citizens of any country with a standard of living lower than the US. That point seems lost on Mr Clemens. When asked at the end of the program to address the policy implications of open borders per se, he obfuscates, offering straw men and false choices (e.g., social security is not the only government provided benefit; pitching a tent wherever you want? really?).

Here is a fact: There are approximately 116,000,000 individuals in Mexico, a substantial portion of whom would better their condition by moving to the US. Should we allow 1% to immigrate each year? 10%? 30%? The US could conceivably take in a half-billion people over the next half century--there is a lot of ruin in a country--should we? If no limit on immigration were set, equilibrium would eventually be reached...what would be the state of the average low wage American worker in that eventuality? No academic study will convince me their standard of living will benefit from the competition of hundreds of millions of similarly skilled immigrants. It is not the role of the US government to beggar US workers in the interests of an economic theory (or global humanitarianism).

I am all for an immigration policy that promotes economic productivity, provided that policy is balanced against the interests of American citizens (with the scale tilted toward Americans). No sane, responsible populace would desire any other type of policy.

Free labor and capital markets provide the ideal environment for free citizens to realize their full potential. But free markets and economic efficiency are not the be all and end all of either nations or their individual citizens. Individual human lives and the cultures they inhabit are more complex, and interesting, than that. As an economic philosopher, Russ usually forcefully makes that point.

Huge Econ Talk fan, by the way.

Russ Roberts writes:

John Popham,

Glad you're a fan, and yes, many things matter more than efficiency. Or money.

Do you think the large increase in foreigners arriving in the US between 1880 and 1920 was bad for the country both culturally and financially? Do you think it was a mistake to let them in?

Laurie writes:

This is a very interesting topic! In regards to the discussion about ratios of the number of low-skilled jobs to low-skilled workers, I'd be curious to hear thoughts about a large number of Americans - myself and my peers included - who have graduated within the past 5 years or so with higher-ed degrees and can't find anything BUT a low-skilled job. This may raise other questions on other issues such as upward mobility, geographic location within the US, etc., but I find it compelling that many high school-only jobs are being filled by quite a lot of degree-holders.

John Popham writes:

Thank you for the opportunity to expand, Professor Roberts. I believe the U.S. was wise to allow generous immigration during the late 19th and early 20th centuries. Robust immigration is appropriate under certain conditions; that does not mean unlimited immigration ipso facto leads to a better life for all affected parties (which I take to be Mr. Clemens’ position).

Like many Americans, I am a beneficiary of that earlier, permissive immigration policy: My maternal grandparents emigrated from Sweden during that timeframe and, being poor, my grandfather worked as a logger, farmer, and carpenter. My paternal ancestors came over a few generations earlier, from the British Isles, and were also humble farmers.

Unfortunately for today’s foreign farmers who would like to emigrate to the U.S., the era of cheap land and independent farmsteads is long past. The sod busters have been replaced by mono-culturists. Farm work for today’s immigrants largely consists of manual labor for a depressed wage rather than the homesteading opportunities available to my ancestors (with wages being depressed due to the desperation of the workers). Immigrants can do this work because most Americans will not perform hard labor, exposed to the elements, for minimal pay. There are no jobs Americans won’t due, of course, only jobs most Americans won’t due as cheaply as some non-Americans. Limiting immigration will either raise agricultural wages to the point Americans will do the work, drive automation as a substitute for labor, or lead to greater produce imports. I suspect none of those outcomes would be catastrophic to the economy as a whole. To be sure, not all immigrants of old were farmers, but similar changes have affected all areas of our economy. There are no vast manual labor opportunities in today’s economy analogous to the fields of agriculture, mining, shipbuilding, steelmaking, etc., circa 1900.

Immigration will always raise cultural issues; fortunately our nation has a record of handling those challenges reasonably well (Know Nothings and other nativist flare ups notwithstanding). Any citizen who embraces our creed of individual liberty, personal responsibility, and republican government is the accepted equal of any other, no matter the circumstances or location of their birth. The key is effective assimilation, and I do have concerns that we may be less effective at driving assimilation than we once were. That could be due to a loss of confidence in our culture, a fear of giving offense, or some other factor. It seems reasonable to expect that the more immigration that occurs within a span of years, and the greater the percentage of the population that is therefore foreign born, the more challenging assimilation is likely to be.

I am not an economist, only a Navy lifer, but my gut tells me that during this time of persistent high unemployment and low labor force participation, allowing high levels of immigration (legal or illegal) will harm precisely those Americans left most vulnerable by the “Great Recession.” If wiser economic policies lead to rapid growth and labor scarcity, let us then open the door a bit further. Now may not be the time.

Kurt Luckenbill writes:

I'm an avid fan of Econtalk, but I found this podcast to be rather annoying. As others pointed out, your guest stood up a variety of straw men and easily knocked them down. Both of you also fell victim to the Libertarian Curse of thinking that everything would be great if only we lived in a truly free libertarian society. Our current society is structured so that immigrants as well as low (reported) income Americans get an enormous array of benefits whether or not they paid into the system.

Your guest mentioned Social Security, stating that immigrants must pay into the system in order to get benefits. The benefit side of Social Security is enormously progressive, with lower income individuals, and those that work for a shorter number of years getting much higher benefits relative to their taxes than higher earners that work more than 35 years.

I also think that many of the costs of immigration are intentionally obscured. Recently in PA there were some state payments made to schools, about 1/3 of which went for English Language training. this is necessary for immigrants but I suspect these kinds of costs rarely show up in the official estimates.

Russ, we are the same age, so ponder this; thirty years ago when we travelled and stayed I a hotel, the cleaning people were usually local African-American women. This was an entry level job that enabled people with low skills and a poor education to get on the first rung of the job ladder. Today hotels are almost universally cleaned by illegal immigrants. Today's local African-American women are just as capable and willing to take those jobs as their mothers were. The difference is that the hotel finds it more profitable to hire (through an agency to keep it at arms length) persons who won't complain if they are treated badly. Where they don't have to pay social security or medicare or workmens comp or unemployment or overtime. Perhaps none of these things would exist in a truly libertarian society, but they do exist here and they keep low skilled Americans from getting jobs.

John Berg writes:

If I remember the quote correctly, a mathematician suggested to a poet that while his poem stated that "a man died and a man was born," it would be more correct to say that "one and a third was born." The correction conformed to demographics, a factor that that may be at the basis of all economics. Way back, it was said that Social Security allowed a man to retire so a younger man could have his job. Still, a job also had to created for each young man who reached employable age.

We know that currently births increase our population of employables annually but the number of jobs hasn't been restored to the number extant in 2008. The current growth of jobs is not keeping up to the newly employable.

Isn't the best incentive for remaining in high school the likelihood of employment?

I agree with John Popham about our generation's duty to our children and grandchildren, but I wonder if he knows that the State Dept. was to produce a definition of a citizen but the courts "disallowed" it and we have no definition of what a citizen is?

John Berg

Gandydancer writes:

Popham: "Robust immigration is appropriate under certain conditions; that does not mean unlimited immigration ipso facto leads to a better life for all affected parties (which I take to be Mr. Clemens’ position)."

No, Clemens admits there will be American losers. But screw them. See my earlier post.

Btw, my maternal grandparents came through Ellis Island and my father was, I believe, an alien. I do not fell obliged to assert that their arrival was good for the offspring of the prior population and have no fixed opinion on the subject. Conditions have changed. Roberts knows this. What, then, do you suppose he is thinking that causes him to pose that question to you?

Kendall Ponder writes:

I must be missing something since nobody else commented on it. How can we only have 1.7 million people entering the workforce over the next decade when we will graduate around 3 million people from high school ever year. That would mean only 1.7 out of 30 people graduating from high school enter the work force. Those who go to college will still enter the work force later. It doesn't change the average number of people who enter the work force. What am I missing?

Robert Reeve writes:

This is a brief addendum to my comment submitted on 18 July 2013. Yesterday I was talking with my brother who has lived in South Africa all his life. I asked him about the situation regarding the increase in destitute/unemployed white people since the mid 1990s. He confirmed the observations summarized in my previous post and noted another related phenomenon. This is the emergence in recent years, under the auspices of church and community organizations, of communal farms for destitute/unemployed white people, especially families. These are effectively rural homeless shelters where residents are given food and accommodation in return for them working on the communal farm.

Seth Martin writes:

I understand that these discussions are supposed to be purely academic in nature, but the economic discussions around immigration have been thrust into the middle of policy debate. It seemed to me that Mr. Clemens was presenting evidence that was mostly positive around introducing greater numbers of unskilled labor into the U.S. economy. Lost on this discussion, however, is the very real interference of the U.S. Government in our labor markets. The area of interference that I am most concerned with, deals with the lower cost labor individual's ability to access the welfare and safety net programs which are growing very rapidly in our country. It seems to me taxpayers could be subsidizing groups that benefit from ever decreasing labor cost. To discuss the economics of these policies without addressing this very real interference to me seems pointless. In my mind this is just another form of crony capitalism that is ruining the country. Whatever benefit could be achieved by importing lower labor cost into the U.S., is lost on the inescapable reality that in the end, the taxpayer will lose, even though there will be winners and losers on both sides of the political argument.

Comments for this podcast episode have been closed
Return to top