One of the most articulate critics of expanded immigration into the United States is Harvard economics professor–and immigrant–George J. Borjas. Because I wanted someone who knew the facts about immigration to the United States and could present a balanced view of immigration, noting costs and benefits, I commissioned him to write the articles on immigration in both editions of The Concise Encyclopedia of Economics. See here for the article in the first edition and here for the article in the second edition.

I still think he did a good job in both. In retrospect, though, I was, as editor, too accepting of his implicit nationalistic way of judging costs and benefits. Notice that in neither article does Borjas discuss benefits to the immigrants.

Co-blogger Bryan Caplan has ably criticized the sections of George Borjas’s recent book in which he expresses skepticism about how many low-wage people would move from poor countries to richer countries in response to wage differentials if immigration restrictions were removed. I lean strongly toward Bryan’s critique.

But here’s what I don’t get. If Borjas were right that relaxing immigration restrictions substantially would not lead to much immigration, how is that an argument for keeping the restrictions? Borjas’s argument is that moving costs on their own would heavily limit the number of people moving. So why would Borjas have the government impose what he seems to suggest is close to a non-binding constraint?

Incidentally, this view that moving costs heavily constrain moves, even within the United States is new to Borjas. Here’s what he wrote in his article in the second edition of The Concise Encyclopedia:

The early empirical evidence on this subject seemed to indicate that immigration did not have a substantial impact on the labor market opportunities of native workers. This evidence was based on the speculation that if the services of natives and immigrants are interchangeable, natives should earn less in cities where immigrants are in abundant supply, such as Los Angeles or New York, than in cities with relatively few immigrants, such as Nashville or Pittsburgh. Although natives do earn somewhat less in cities with large immigrant populations, the correlation between the native wage and the presence of immigrants is close to zero.

Further evidence of the weak correlation comes from the Mariel boatlift. In April 1980, when Fidel Castro declared that Cubans wishing to emigrate could leave from the port of Mariel, 125,000 people accepted the offer and Miami’s labor force suddenly grew by 7 percent. Yet, the trends in wages and unemployment rates in Miami between 1980 and 1985, including those of black workers, resembled those observed in comparable cities.

Such evidence, however, is not conclusive. This approach to measuring the labor market impact of immigration ignores the fact that labor and capital are mobile between cities. If an influx of immigrant workers reduced wages substantially in a particular city, native workers and some immigrants would leave that city and find work elsewhere. And natives who contemplated migrating to that city would choose another destination. Also, capital would “migrate” to cities with large numbers of unskilled immigrants, where capitalists can earn a greater return on their investment. Large-scale immigration, therefore, may not drive down wages in particular cities. Rather, its depressing effect on wages is nationwide.

Notice that in the crucial third paragraph, Borjas does use the word “may” and the evidence on moving costs that he cites is his latest book came well after he wrote the piece for the Encyclopedia. So this is not a “gotcha.” Rather, I’m pointing to a tension between his earlier and more recent views. Assume, as Borjas now believes, that moving costs within America are very high. Then his reasoning in the above quote does not follow. In other words, if moving for native Americans is very costly, then we would not see the movement of native workers between cities that he suggests is responsible for the apparent absence of negative effects of immigrants on native workers’ wages. Then the apparent absence of negative effects within a city that has a lot of immigrants would suggest that immigrants really do not have a substantial negative effect.

P.S. I note that Tiago makes my point in a comment on Bryan’s post.