[See Part Two—Borjas On Open Border Libertarians: " I Don't Really Miss Them, Actually"]
[An abridged version of an interview published in Immigration and the American Future. See also Heaven's Door After A Year, By George Borjas, June 10, 2001]
Peter Brimelow writes: Everyone knows, or concedes, that immigration is good for the economy—except economists. Amazingly, since the early 1990s, a consensus has existed among labor economists that the current unprecedented influx into America is of no particular economic benefit to native-born Americans in aggregate. I reported this consensus in my 1995 immigration book Alien Nation: Common Sense About America's Immigration Disaster and it was confirmed by the National Research Council's 1997 study The New Americans, the survey of the technical literature on the economics of immigration done at the behest of the Jordan Immigration Commission. Equally amazingly, this consensus has been totally ignored in the public discourse on immigration—one of the most startling failures of democratic debate of which I am aware.
No-one has more to do with the new consensus about the economics of immigration than Professor George J. Borjas, Professor of Economics and Public Policy at Harvard University's John F. Kennedy School of Government and a Research Associate at the National Bureau of Economic Research. Borjas first began to depart from the optimistic orthodoxy with his 1990 bookFriends or Strangers: The Impact of Immigrants on the U.S. Economy. His most recent full-length treatment of the subject is his 1999 book Heaven's Door: Immigration Policy and the American Economy. Borjas, himself a Cuban immigrant, has every emotional reason to favor immigration. That he does not is entirely a function of the data—and his scrupulous scholarship.
I spoke to him in his Cambridge office and began by asking him to summarize the findings of the NRC's The New Americans.
Borjas: Basically the "immigration surplus"—the net gain to people already living in the U.S—was very small—one tenth of one percent GDP, or at that time around ten billion dollars in a 5-6 trillion dollar economy. It's trivial.
Brimelow: That's not net of transfer payments like welfare, education, right?
Borjas: That's correct. There's another chapter where the authors estimate that, on net, native-born households in the U.S. paid about $200 a year more in taxes because of immigration. That would add up to something like $10-20 billion dollars a year. So net, it's basically a wash. Whatever the "immigration surplus," it's eaten away by the cost of providing services to immigrants.
Brimelow: But the redistribution impact within the native-born community is very large.
Borjas: Yes. Let me make that very clear. At the time I wrote that initial paper, I was basically taking a relationship out of the labor demand literature—a X% increase in labor would lower wages by Y%.
That meant current immigration had lowered the total wage of natives by about 2%. And all that goes straight to the employers, to the capitalists. In the long, long run, some of that would filter down to the consumers also. But I didn't do that in my paper. Nobody knows what the breakdown is between consumers and employers.
So the way we freeze the argument is: immigration redistributes wealth from people who compete with immigrants—namely workers who have the same jobs as immigrants—to people who use immigrants. For example, a California family—gardener, the maid, all this stuff.
Brimelow: Why do you think the National Research Council findings had absolutely no effect on public debate?
Borjas: I don't know. Certain parts seem to be cited over and over. For example, the Wall Street Journal—they often cite that there is no wage impact based on the National Research Council. They'll also cite another chapter, that if you were to follow immigrants and their children for 300 years, and assume a tax increase, then immigrants could be a huge benefit to the U.S. Even though a 300 year projection is complete nonsense.
They choose and pick what they want.
Brimelow: But the fundamental conclusion, that there are no substantial aggregate benefits for the native-born from current immigration, was completely buried?
Brimelow: And, for example, the microstudy showing that immigration means a $1,100 fiscal transfer from every native-born family in California was completely buried too. Why?
Borjas: [Laughs, throws up hands] I don't really know the answer. I think part of it has to do with a sort of an implicit bias in the media. Not just in terms of the National Research Council study, but in terms of the general immigration debate. At least until recently, the main stream press, when it covered immigration, would begin with a very sentimental kind of story about a particular immigrant. And then they would proceed to describe how good immigration is. That was more true for some papers than for others, but it generally describes the typical immigration story for a long time. It is only more recently that people have begun to discuss whether in fact there could be a negative impact on wages, on social programs and so on.
Maybe it's because of the way they frame the problem. But I'm not a journalist—I really don't know, okay?
Brimelow: It was interesting to see Paul Krugman cite you recently in the New York Times. [ North of the Border, by Paul Krugman, New York Times, March 27, 2006]
Borjas: That's right, Paul Krugman cited the paper on immigration's wage impact I did with Larry Katz. [The Evolution of the Mexican-Born Workforce in the United States". George J. Borjas and Lawrence Katz. NBER Working Paper 11281, April 2005] It's only about a year old, but the paper it was based on was published in 2003. [PDF]Nobody talked about that in the media when it was published.
Brimelow: Do you find that the Krugman column made a difference?
Borjas: Huge. It was amazingly influential. The minute it came out, the emails start flowing in.
Brimelow: So what's happened in academe since 1997?
Borjas: In 1997, at the time of the NRC report, it was generally thought that immigrants had a minimal impact on wages for the following reasons: studies tended to focus on comparing how natives do in cities that have large immigrant populations like San Diego or Los Angeles, with cities that have few immigrants, like Pittsburgh or Oklahoma City. When you do that, you tend to find very weak impact.
People were aware of two potential problems. One was that immigrants gravitate to cities that pay higher wages. That could build a positive correlation between high wages and immigration which could easily swamp anything in the real market.
Reason number two was that if immigrants are going to a city like San Diego, both native workers and native capital will respond. Native workers will move out. Native capital will move in, where wages are low. These equalizing flows would tend to take away negative impacts.
So now there is a consensus that cross-city correlations don't really matter. To gauge the wage impact of immigration, you have to move to the national level. And that's eventually what I ended up doing in the paper that was published in 2003, that I wrote with Richard Freeman and Larry Katz.
The key insight there was, look, immigrants have come in over the last 30-40 years at different education levels. But the age structure of immigration varies a lot over time. So what I ended up doing was using data from 1960-2000, and looking at the wage from each group defined by education and age to see how the evolution of the wage of a given education group over forty years responded to immigration.
The minute you do that, the negative wage impact of immigration becomes very apparent. And remarkably enough, it was about what we estimated before. The 10% increase in supply reduced wages by 3%. It confirmed what we thought we knew about labor demand in the context of immigration. [The Labor Demand Curve Is Downward Sloping: Reexamining the Impact of Immigration on the Labor Market, The Quarterly Journal of Economics, November 2003,(PDF)]
This was not a paper that was picked up in the newspapers. The Wall Street Journal didn't grab it, the New York Times didn't run an op-ed about it. I mean, it was just terrible. Economists recognized that it was important, but the media did not.
I suspect very strongly that had I come out with a different answer, it would have been picked up.
Actually, I'll give you the best piece of evidence for why wages must drop with immigration. There have been a ton of hearings in Washington regarding the guest worker program. Who exactly is lobbying for guest workers? Is it you and me? No, it's employers, right? Why would employers tend to go to Washington and expend their resources lobbying for something that doesn't benefit them?
Brimelow: It can all be explained in rather crass Marxist terms, can't it? The class analysis works.
Borjas: Of course! Of course! The Marxist analysis works.
Brimelow: The thing that is interesting is that neither the labor unions nor the leaders of the minority groups are opposed to it.
Borjas: But I have a feeling that will change soon. You're beginning to see a breakdown in the model of political correctness.
Brimelow: What else is happening in the field?
Borjas: You heard about the Ottaviano-Peri paper [PDF]that was cited in The Economist, right? [Myths and migration, The Economist, [Pay archive] April 6, 2006] My 2003 paper assumed is that the low skilled group immigrants and native are what we call perfect substitutes. They tried to relax that. And they end up with a result that basically says that in the long run the average native wage goes up, not down, by one or two percent as a result of immigration.
Now let me tell you why that is not completely sensible. Somebody else's wage must have changed in the opposite direction. Who was that somebody else? Immigrants. For native wages to go up by 1-2%, the average wage of immigrants must have gone down by like 15-20%. That's just so outside what we know about labor demand that it puts the whole paper into question. They chose to focus on natives knowing the fact that whatever you do for natives, the immigrant wage must have fallen by like 15-20%, which would make the whole thing impossible. So they chose to ignore that.
But, you know, some people like the theory that immigrants "increase wages"—even though we know from theory that immigrants cannot increase wages.
Brimelow: The Economist, of course, is fanatically pro-immigration. In the article you just mentioned it also attempted to downplay your critique of current immigration policy. It said about you that
"Immigration's critics therefore count Mr Borjas as an ally. But hold on. These figures take no account of the offsetting impact of extra investment. If the capital stock is assumed to adjust, Mr Borjas reports, overall wages are unaffected and the loss of wages for high-school drop-outs is cut to below 5%."
That's actually purely hypothetical—mathematical theory, isn't it? It has nothing to do with the data?
Borjas: Yes. All these results are based on a theory of labor markets. It states the following: If the U.S. economy has constant returns to scale, namely, you double inputs, you double outputs, which is the key assumption we all make in this, then immigration cannot have an impact on the average wage, in the long run. That's the mathematical theory.
The way an economist finds the short run is, he holds everything else equal. So holding everything else equal, 10 million that are let into the country, what happens to wages?
The long run, basically, is the complete opposite. We know that when immigrants come in, wages go down in the short run. Then capitalists build factories to exploit the cheaper labor and so on. So in order to find the long run, we suppose that every expansion that could take place, actually does take place.
Now, let me emphasize, neither the short run or the long run has really been proved. In the real world, things do adjust, but they don't adjust completely. So the best way to look at these extremes is as bounds of what the effect could be. You know, the short run is going to happen, and then, who knows when—Keynes said that in the long run we're all dead—everything adjusts completely. Nobody knows exactly how long the process takes.
But economic theory predicts unambiguously that if you have constant returns to scale production function, the average impact of immigration on wages in the long run has to be zero. Because everything adjusts completely. Capital adjusts enough to account for the extra labor.
But that doesn't mean that every group's wage is unaffected. It just means that the average wage effect is zero.
Now, I would say that the short run assumption is not completely plausible in the real world because it holds everything else equal. When you're living in the real world, people adjust. On the other hand, the long run assumption, that everything adjusts completely, is also not very plausible in the real world either, because not everything adjusts completely.
I will give you an example. Take Puerto Rico. It's part of the U.S., right, with very sizeable labor outflows, and very sizeable capital inflows. It still hasn't had everything completely adjusted after 40-50 years of migration. The wage in Puerto Rico today should be the average wage in the U.S. today. Is it? No. There is still a huge wage gap between Puerto Rico and the U.S.
Borjas: A good question. That's one of the problems with economics. Things don't adjust completely. There are frictions. In theory Puerto Rico should be empty now, because the wage is much higher in the U.S., right? But, it's not. In theory, capital flows would have equalized in Puerto Rico with the average in the U.S. market. It hasn't. There are these frictions that we don't really fully understand.
What I'm saying that, we can do it mathematically, we can look at the marketplace and we can look at the extremes, we can look at the short run and we can look at the long run. But it is very hard to tell where the truth is in the middle. We can say, however, that right now, immigration is impacting wages.
Brimelow: How does this paper by Davis and Weinstein fit in? [Columbia economists Donald R. Davis and David E. Weinstein estimated in 2002 that, by sharing their technological base with immigrants, U.S. natives suffer a loss of some 0.8 percent of GDP][ Technological Superiority and the losses from immigration, by Donald R. Davis and David E. Weinstein, NBER Working Paper No.. 8971, June 2002(PDF)]
Borjas: It doesn't really fit in. They have a different kind of argument that is much more familiar to trade economists than to labor economists. There are sort of field divisions in economics. People who study immigration tend to be labor economists. Labor economists tend not to be very well trained on international trade theory.
Brimelow: How important is their argument?
Borjas: It's pretty important, but the problem is that there's no evidence. They create a model and get a number out. If the model is correct, then the number is very important. But nobody is going the extra step of trying to see whether the model is correct or not.
Brimelow: But that's equally true for the labor economists, isn't it?
Borjas: No, because, for example, all the wage impacts that we've talked about actually come from data. That was the major contribution of our 2003 paper. I actually went out and looked at the data on wages and wage impact.
The Davis and Weinstein thesis needs more empirical work and nobody's doing it as far as I know.
Brimelow: No PhD students feel compelled to do it?
Borjas: They had a very hard time publishing the paper. People don't like the result. Even ignoring the fiscal impact, they find a huge negative loss to the U.S. And that's not a kind of result that most trade economists like to hear. Most trade economists argue that free trade is great, that labor mobility's great and so on. This goes very much against the grain.
Brimelow: What is the distinction made between trade and immigration from a theoretical standpoint?
Borjas: From a theoretical standpoint, there is actually very little distinction in the sense that both are importing resources into the country.
From a broader standpoint—people are not machines. We can import a car from Japan and not have a particular effect on the economy. We could import 25 Japanese who make that car, and that has a very different impact on the economy, because they're people. The machine will not incur schooling costs for the children, the machine will not incur Medicaid costs when it gets sick, right?
It goes back to the Milton Friedman quote you got in your Forbes interview, right? It's not possible to combine mass immigration and the welfare state.
Brimelow: What about claim we couldn't run the economy without illegal immigrants?
Borjas: It's complete nonsense. If there were no illegal immigration, and the demand was still there for gardeners and nannies, the price of those things would go up.
Illegal immigration is very highly regionalized, just like legal immigration. Does that mean that the rest of the country, between the two coasts, doesn't function because there are no immigrants? That's ridiculous. There are cabs in the middle of the country, believe it or not, there are gardeners.
It's like the argument that immigrants do jobs that natives won't do. That's complete nonsense too. It's a question of price. And one good by-product of all the current immigration controversy, I think, is that people now are getting that it's complete nonsense.
Peter Brimelow is editor of VDARE.COM and author of the much-denounced Alien Nation: Common Sense About America's Immigration Disaster, (Random House - 1995) and The Worm in the Apple (HarperCollins - 2003)