Immigration Isn't That Bad for Native Workers
Does immigration cost native workers their jobs or drive down their wages? This is one of the most contentious issues in economic policy right now. Fortunately, a lot of academic economists are doing some very smart, careful and thorough empirical work to figure out the effects of immigration on local labor markets. For a survey of the literature, see this 2011 paper by Sari Kerr and William Kerr. For a meta-analysis of the effects of immigration on wages and employment, see this 2008 paper by Simonetta Longhi, Peter Nijkamp, and Jacques Poot.
These and other surveys and meta-analyses all reach one overwhelming conclusion: Immigration has at most only a small harmful effect on the native-born. If this were biology or astrophysics, that would be that -- the media would accept the scientific consensus, until new research came along and overturned it. But this is economics, and so politics and ideology inevitably get in the way. There will always be people who are in favor of immigration restriction, and they will always have reason to question what would otherwise be a well-accepted consensus.
To some, this means that true consensus is impossible. Russ Roberts, host of the popular podcast EconTalk, sounds a pessimistic note in a recent interview:
A famous…study [concluded] that the influx of Cubans from the Mariel boatlift didn’t hurt prospects for Miami’s native workers…Harvard’s George Borjas re-examined the Mariel data last year and insisted that the original findings were wrong. Then Giovanni Peri and Vasil Yasenov of the University of California, Davis retorted that Mr. Borjas’s rebuttal was flawed…To Mr. Roberts, this indicates something deeper than detached analysis at work. “There’s no way George Borjas or Peri are going to do a study and find the opposite of what they found over the last 10 years,” he says. “It’s just not going to happen. Doesn’t happen.”
To Roberts, empirical questions in economics will always degenerate into a politically motivated back and forth of he-said-she-said. But I disagree. Although complete agreement may never be possible -- there will always be the research outlier whose results differ radically from the norm -- there are two factors that allow conclusions in economics to be more than 50-50. Those are 1) weight of evidence, and 2) quality of methods.
We’ve already seen that the weight of evidence comes down on the immigration-is-mostly-harmless side of the debate. But what about the quality of the empirical methods used? Here too, I think one side is clearly winning.
In the academic debate over immigration, the standout is Borjas, a professor of economics and social policy at the Harvard Kennedy School of Government. Borjas may disagree with most of his colleagues, but his output is voluminous and he works at a top university. A number of business publications have called him “America’s leading immigration economist.”
Now Borjas has written a book on the topic, called "Immigration Economics." The book summarizes much of Borjas’ own work, all of which purports to show that immigration is very bad for native workers. However, University of California-Berkeley professor David Card and University of California-Davis’ Peri have written a paper critiquing the methods in Borjas’ book. It turns out that the way Borjas and the economists he cites do immigration economics is very, very different from the way other researchers do it.
One big difference is how economists measure the number of immigrants coming into a particular labor market. The natural measure of this seems like it would be the change in the number of immigrant workers, measured as a percent of the total workers in the labor market before this batch of immigrants arrived. This is the measure Card and Peri like.
But instead of using the change in the number of immigrants, Borjas just uses the number of immigrants itself, as a percentage of the current total number of workers. This creates a number of problems.
Let’s think about a simple example. Suppose there are 90 native-born landscapers in the city of Cleveland, and 10 immigrant landscapers. Suppose that demand for landscapers goes up, because people in Cleveland start buying houses with bigger lawns. That pushes up the wages of landscapers, which will draws 100 more native-born Clevelanders into the landscaping business. But the supply of immigrants is relatively fixed. So the percent of immigrants in the Cleveland landscaping business has gone down, from 10 percent to only 5 percent, even though the number of immigrants in the business has stayed the same.
Borjas will find that the percent of immigrants in the business goes down just as wages go up. But to conclude that native workers’ wages went up because immigration went down would be totally incorrect, because immigration didn’t actually fall! In fact, Borjas’ method is vulnerable to reaching exactly this sort of erroneous conclusion. Card and Peri point out that if you use the more sensible measure, there’s not much correlation between immigrant inflows and native-born workers’ wages and income mobility.
In another section of his book, Borjas uses a much more complex model to estimate the impact of immigration on the native born. But Card and Peri point out that Borjas assumes that immigrants and native-born workers are identical in terms of job ability. In reality, immigrants have different language skills, and have other skill differences from native-born workers of the same education level. Card and Peri find out that when you allow for even a small discrepancy between the two types of workers, immigration looks pretty harmless.
This isn’t the first or the most noteworthy example of Borjas using data analysis methods that look a little dubious. In an earlier case, he used a data set that was so finely sliced and diced as to be rendered useless. Again, Peri was on the team that uncovered the oddity.
So the debate between Borjas and his opponents shouldn’t make us throw up our hands and say that “opinions on the shape of the Earth differ,” or conclude that the debate is hopelessly corrupted by politics. I'm not claiming that Borjas is a bad researcher, or that his conclusions are wrong -- those are things for the economics research community to decide. But the public should know that the weight of evidence is against Borjas, and many of his methods also tend to look a bit shaky when subjected to careful scrutiny.
To not point this out would be to do the public a disservice, given how often Borjas is cited by immigration opponents. Of course, I'm pretty strongly in favor of immigration, so if you don’t view me as a reliable source because of that, you can check out the relevant papers yourself! But as I see it, Peri, Card, and others are winning the empirical debate.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
To contact the author of this story:
Noah Smith at firstname.lastname@example.org
To contact the editor responsible for this story:
James Greiff at email@example.com