Repost: Why immigration doesn't reduce wages
Not that anyone is going to listen to the evidence...
Still traveling for another two days, and people seemed to like yesterday’s repost from the early days of my Substack, so I think I’ll do one more.
With the return of the border crisis and Biden’s plans for a big new immigration reform, it looks like the immigration issue is heating up again. In the coming year we’re probably going to hear lots of very heated arguments from a lot of different angles, and I’m inevitably going to be blogging about some of these. But one aspect of the immigration debate that I feel like I’ve already dealt with pretty well is the question of wages.
A lot of people think immigration lowers wages for native-born Americans. This is true of both immigration opponents, but also of some supporters, who think an influx of workers will help restrain inflation. This is a natural thing to think, since everyone knows that immigration increases labor supply. But what fewer people realize is that immigration also raises labor demand, which tends to cancel out the downward pressure on wages. In late 2020 I wrote a post called “Why immigration doesn't reduce wages”, in which I explained how this works and listed a bunch of the relevant evidence. I wanted it to be a useful and comprehensive guide that people could come back to again and again.
In this post, I’m going to explain why immigration doesn’t lower wages for native-born people (except possibly a little bit, in a few special circumstances). But before I do that, there’s one thing you really have to understand: No one is going to be persuaded by this post. There are two reasons for this.
First, people don’t really believe social science evidence. It took years and years of empirical research — solid results, almost all pointing in the same direction — just to shift academic economists’ opinions on the effects of minimum wage. The average person, not being an academic economist, has even less of an idea of how reliable social science research is.
Second, in my experience, anti-immigration people are completely set in their belief that immigration should be restricted. It’s their fixed north star. The justifications change — Lower wages! Environmental destruction! Brain drain! Rule of law! Cultural change! — but the policy conclusion never wavers. They know what they want to do.
So this post isn’t going to persuade anti-immigration people to change their stance, and it’s probably not going to persuade many normies to be up in arms about the need to let in more immigrants, either. But it’s still important to write, and not just because I’m a stubborn S.O.B. who will die face-down in the muck fighting for The Empirical Evidence. It’s because in another 20 years or so, when America’s current freakout over identity and nationhood has passed, we’re hopefully going to be ready to let in a bunch of immigrants again. And when that time finally comes, these arguments will matter.
Most people think of labor markets as determined by supply and demand. This is actually not a great model of the labor market in general, but for the purposes of this post, it’ll do (see note at the end for more details).
Basically, most people think of immigration as an increase in labor supply. Labor supply is the number of people willing to work at a given wage. So, more people, more workers for any given wage. As a result of the labor supply increase, wages go down.
OK, but working isn’t the only thing that immigrants do. They also buy stuff. They rent apartments. They buy food. They get haircuts. They go to the doctor.
All that stuff takes labor to produce. Food takes labor. Haircuts take labor. Doctor visits take labor. Building new apartments takes labor. And so on. Even if the immigrants don’t start spending their money on day 1, businesses can see the immigration wave coming and they know there will be increased demand for their products. So they hire more people. To hire more people they have to…raise wages.
So immigration increases labor demand as well as labor supply. It looks like this:
A positive labor supply shock pushes wages down. A positive labor demand shock pushes up wages. Maybe one of those effects is a little bigger; maybe the other. But they’re going to mostly cancel out.
And to see why this is true, just think about babies. Each new generation is bigger than the one that came before it. If those young people were just a labor supply increase, then as population went up, wages would go down. But obviously that’s not what happens, because young people also buy stuff, which pushes up labor demand, which pushes wages back up.
Immigrants are just babies from elsewhere.
As you might expect, economists have done quite a lot of research on the question of whether immigration lowers wages. It’s not the kind of thing where you can just wave your hands and say “Oh, immigration is down, wages are up” and conclude that the former causes the latter.
Here are just a few reasons you can’t just look at overall correlations:
Immigrants probably get drawn to economically booming places, where wages are rising on their own. In fact, they may come in the first place because of economic booms.
Recessions and pandemics can cause wages to rise, as low-wage jobs are eliminated first, even as they cause immigrants to leave.
Immigrants may compete with some groups of native-born workers more than others; for example, less-educated immigrants might compete mainly with native-born workers without a high school degree.
An expansion in low-wage industries will cause average wages to fall, even though it really just means there are more jobs than before and those jobs are concentrated in low-wage industries.
Immigration might push native-born workers out of an area, or draw them in.
…And so on.
Economists try to get around these problems using several different tricks. First of all, they try to find episodes when big waves of immigration were caused by things that probably had little to do with economics — for example, waves of refugees from war or political persecution. Second, they try to compare places that get immigrants with places that don’t, but which are otherwise similar in all economic respects. Third, they try to be very careful about looking at the appropriate comparison group, taking note of whether native-born workers moved in or out.
So here are some of the papers that study the effect of immigration on wages. I got most of these papers from literature reviews, to make sure I’m not simply selectively citing one slice of the literature. The listed papers are from the last 20 years; earlier studies are summarized in the survey papers.
Studies on refugee waves
Refugee waves make for great immigration studies, for three reasons. First of all, you know why the immigrants left their home country, and it wasn’t for economic opportunity. Second, since refugees tend to go either to the nearest geographical country, or to a place that makes a point to welcome them in. And third, refugees tend to be a random sample of people from poor countries, which tend to have low education levels, meaning they’re mostly the kind of low-skilled laborers that restrictionists are most afraid of.
“Is It Merely A Labor Supply Shock? Impacts of Syrian Migrants on Local Economies in Turkey”, by Doruk Cengiz and Hasan Tekguc
This paper analyzes the effect of refugees from the Syrian War who went to neighboring Turkey. They find no negative effect on Turkish workers, even at the lower education levels. They find, increases in local demand and investment, consistent with the idea that immigration is a positive labor demand shock.
“The labor market impact of refugee immigration in Sweden 1999–2007” by Joakim Ruist
Wages are set by sectoral bargaining in Sweden, so instead the author looks at the effects of refugee waves on employment. He finds immigrants don’t cause unemployment for the native-born.
“The Impact of Syrian Refugees on the Labor Market in Neighboring Countries: Empirical Evidence from Jordan”, by Ali Fakih and May Ibrahim
Another Syrian-refugees-to-Turkey study. They do find some slight negative impacts of the immigration — slightly higher housing and food prices, and a slight reduction in internal migration to the areas with lots of refugees. But no labor market impact.
“The Impact of Mass Migration on the Israeli Labor Market”, by Rachel M. Friedberg
This study is about Jews who left the former USSR and moved to Israel when the USSR fell. It looks at occupational categories that had lots of refugees vs. those with fewer refugees and finds no negative impact of immigration.
“Immigrants’ Effect on Native Workers: New Analysis on Longitudinal Data” by Mette Foged and Giovanni Peri
This study finds that when refugees fled to Denmark in the 80s and 90s, less-educated Danish workers got more education and ultimately had higher wages — in other words, immigration increased native-born wages in the long run.
“The Labor Market Effects of a Refugee Wave: Synthetic Control Method Meets the Mariel Boatlift”, by Giovanni Peri and Vasil Yasenov
This paper studies the Mariel Boatlift, in which Fidel Castro kicked a bunch of mostly low-skilled people out of Cuba to Miami. They carefully compare Miami’s labor market to that of other cities, and find no negative effect, even for high school dropouts.
“The Effect of Internal Migration on Local Labor Markets: American Cities During the Great Depression”, by Leah Platt Boustan, Price V. Fishback, and Shawn E. Kantor
Internal migrants are pretty similar to immigrants. Here, the authors look at migrants who were forced to move across country by bad economic conditions in their hometowns — for example, the “Okies” of the Dust Bowl. They do find some negative impacts on locals, though not on wages.
“The Migration Accelerator: Labor Mobility, Housing, and Demand”, by Greg Howard
This paper also looks at internal migration in the U.S., and uses traditional migration patters to try to isolate periods in which internal migrants were forced to move (e.g. by a bad local economy). The author finds that inflows of internal migrants cause a boom in housing construction, which supports local labor markets — just as the simple theory above would predict.
“The effect of immigration on wages in three European countries”, by Aslan Zorlu and Joop Hartog
This paper looks at immigration waves to the Netherlands, the UK, and Norway, and finds no labor market impact for the native-born.
“The labor market impact of immigration in Western Germany in the 1990s”, by Francesco D’Amuri, Gianmarco I.P. Ottaviano, and Giovanni Peri
This paper looks at the influx of East Germans to West Germany after the fall of the Berlin Wall. It finds no negative impact for native-born German wages, but it does find that the East Germans competed with existing immigrants in West Germany and drove down their wages.
“Migration and Wage‐setting: Reassessing the Labor Market Effects of Migration”, by Herbert Brucker and Elke Jahn
Another study on Germany that finds much the same as the D’Amuri et al. study
“Immigration Restrictions as Active Labor Market Policy: Evidence from the Mexican Bracero Exclusion”, by Michael A. Clemens, Ethan G. Lewis, and Hannah M. Postel
When the U.S. ended a program for Mexican guest workers, places that had relied on guest workers saw no labor market impact.
“Closing Heaven's Door: Evidence from the 1920s U.S. Immigration Quota Acts”, by Philipp Ager and Casper Worm Hansen
This paper examines what happened to U.S. cities when the U.S. decided to ban most immigration in 1924. Manufacturing industries suffered and native-born workers were pushed into lower-wage occupations, though Black workers did benefit a little bit.
“Economic Impacts of Immigration: A Survey”, by Sari Pekkala Kerr and William R. Kerr
A comprehensive overview of papers on a large number of immigration-related topics. The section on wages reviews a large number of studies and meta-analyses, and finds that “the documented wage elasticities are small and clustered near zero”, meaning little or no wage impact from immigration.
“How to Measure Labour Market Effects of Immigration: A Review” by Liesbet Okkerse
A meta-analysis of immigration studies. The studies find very small or no labor market impact.
“Joint impacts of immigration on wages and employment: review and meta-analysis”, by S. Longhi, P. Nijkamp, and J. Poot
A meta-analysis of immigration studies. The studies find very small or no labor market impact.
“The labour market impact of immigration”, by Christian Dustmann, Albrecht Glitz and Tommaso Frattini
This paper surveys the evidence on immigration to the UK, and finds very small labor market impacts if any
No Card? No Borjas?
There’s a famous study of the Mariel Boatlift by David Card, in which he found no labor market effect. But George Borjas, an economist whose papers somehow always end up contradicting nearly every other economist in existence and finding negative effects of immigration, wrote a paper claiming to find negative effects for a very small slice of less-educated minority workers. Borjas’ paper was very bad. It had to slice the data incredibly finely to find a negative effect, narrowing the sample size to just a handful of observations — not enough for a statistical analysis. Other economists found that even that drop was actually the result of a change in measurement. A very poor showing by Borjas, and a reason I didn’t include his studies above. But out of fairness I also excluded the Card study.
So from the papers above, we find that immigration can occasionally have some small negative impacts on labor markets. In the middle of an economic catastrophe like the Depression, when jobs are scarce, it can bump a few people out of jobs. New immigrants can compete with existing immigrants.
But overall, immigration — even of the lowest skilled variety — has very little or no impact on native-born wages. And sometimes even a positive impact. The most probably reason is that, as explained above, immigration boosts labor demand, not just labor supply!
Of course, as I said, don’t expect 17 papers — or 1700 papers — to change the mind of the anti-immigration people. They’ll stroke their chin sagaciously, cite Borjas, or wave their hands at some case when immigration fell while wages were going up, or a few years before wages went up, or whatever. Or they’ll shift the discussion to environmental impact, or brain drain, or cultural change, or whatever, only to switch it back to labor markets again as soon as you address those other concerns.
But now YOU know that immigration doesn’t hurt wages. And someday, you might be able to do something with that knowledge.
A note about models: Supply-and-demand is actually not a great model of labor markets; they are a lot more complicated. One other factor is monopsony power, which means employers’ power to set workers’ wages. That can increase the negative impact of immigration on labor markets, though apparently not enough to make a difference in most cases, as the results above show. Also, several of the papers listed above lay out more complex and (hopefully) realistic labor market models, so if you’re interested you should browse through them. None of them change the basic result.