WHAT effect do immigrants have on native wages? It’s perhaps one of the most important questions of labour economics. It’s also one that is largely unanswerable. The problem is that it’s almost impossible to separate cause and effect. If a country with high rates of immigration also sees strong wage growth, we can’t assume that immigrants are boosting wages—it may well be the case that the migrants are choosing to move to places with stronger economies.
One approach to getting around this problem is to find a natural experiment in which either the supply of or demand for labour changes exogenously. Perhaps the most famous example of such an event in labour economics is the Mariel Boatlift. In 1980, Fidel Castro, then president of Cuba, eased emmigration restrictions. Some 125,000 Cubans moved to the United States that year. Almost instantaneously, the labour supply of Miami increased by 55,000.
The Mariel migrants were overwhelmingly low-skilled workers—less than half had high-school degrees. In 1990, David Card, now an economist at the University of California, Berkeley, wrote a highly regarded paper examining the effects of the Mariel supply shock on American-born Miamian’s wages, and found no evidence of any adverse effect. The debate surrounding Mariel seemed largely settled until last year, when George Borjas of Harvard published his own study on the topic (a working version of which was circulated in 2015). As we noted last May, Mr Borjas’ study found that wages for low-skilled, native-born Miamians fell noticeably following the Mariel boatlift.
Mr Borjas’ work has sparked plenty of controversy–his blog notes at least three challenges to his work, see here, here and here. After a few months of relative tranquility, Mariel has once again come into the limelight thanks to a new paper by Michael Clemens of the Centre for Global Development, a think-tank, and Jennifer Hunt of Rutgers University. Mr Borjas has since responded to this critique; their correspondence appears to have proceeded adinfinitum. (A working-paper response to Mr Clemens and Ms Hunts’ critique can be found here.)
While the Mariel studies are of great political interest, the fact that they’ve shown such different results have come down to fairly boring technical decisions by the papers’ respective authors. Mr Borjas’ results differed from Mr Card’s in part because he chose a different definition of low-skilled worker, and in part because he chose a different set of cities to compare Miami with. Mr Clemens and Ms Hunt argue that much of the statistical power behind Mr Borjas’ findings is an accident of a methodological change in the survey behind the Mariel studies. In some years, Mr Borjas’ paper feature sample sizes as small as around 20 people.
By economics 101 reasoning, the short-run, partial-equilibrium effects of a large influx of migrants are clear. Given a downward-sloping labour demand curve, a sudden increase in supply should be expected to lead to lower wages. It shouldn’t be surprising—or indeed controversial, that a study like Mr Borjas’ should find that evidence of a wage decrease. But even if large-scale migration hurts native workers in the short run, it should have little bearing on public policy.
Far more relevant to lawmakers are the long-run effects of immigration on wages. In theory these depend on how immigrants change the skill composition of the workforce. If lots of unskilled workers arrive in a country, unskilled workers’ wages should fall relative to everyone else’s. But economists usually have to squint hard to find a negative effect of immigration on wages for any native workers. In the long-run, immigrants tend to reduce the wages only of past generations of immigrants (whose skills presumably overlap strongly with those of the newcomers). One study found that while immigration between 1990 and 2006 had little effect on wages of native-born Americans, it lowered the wages of previous immigrants by 6.7%. American-born workers, perhaps because of their language skills, were better prepared to move on to different jobs.
I suspect that few who have participated in the Mariel debate actually care about the true short-run wage effects were—it instead serves as a proxy war for the broader immigration debate. It seems that proponents of immigration feel compelled to respond to Mr Borjas not because they believe Mariel boatlift was important, but because they are afraid of ceding any ground to immigration hawks. The problem is that Mariel isn’t nearly meaningful enough to warrant such attention. Suppose Mr Borjas’ findings are entirely correct. Then what?