The Case That Immigrant Labor Doesn't Hurt Low-skill U.S. Workers

The heated debate over immigration reform continues in Washington. Much of the heat comes from one claim: immigration hurts low-skill US workers.

The idea is that immigration causes overall economic harm to low-skill US workers, because many immigrants are low-skill themselves and compete for the same jobs. This is a reasonable belief if all one has to go on is the kind of day-to-day experiences that everyone has—such as applying for jobs and competing with other applicants.

It is also wrong.

Many of the world’s best labor economists have spent the last quarter century exhaustively looking all over the world for negative effects of immigration on low-skill workers. They cannot find such effects. This is one of the most robust findings in the labor economics research literature.

These economists take two main approaches. The first is to look for sudden, large increases in immigration to a particular area, and track what happens to native workers in that area. These include massive inflows of Cubans to Miami in 1980, Algerians to France in 1962, Russians to Israel in the early 1990s, all immigrants to different regions of Germany in the 1980s and to regions of the United Kingdom in the 80s and 90s, and Former Yugoslavians to the rest of Europe. Not one detects substantial effects of immigration on wages or employment.

In the second approach, economists chop up the data differently. They divide workers in the migrant-destination country not into different geographic regions but into different groups of broadly similar people—for example, 25–29 year-old males with a high school degree, 30–34 year old females with a college degree, and so on. They then watch what happens to wages and employment among native workers in each group, across the whole country, upon the arrival of immigrants with those characteristics. With this method, the leading-edge research is by Gianmarco Ottaviano and Giovanni Peri. In a paper published last year, they show that total, cumulative immigration to the United States between 1990 and 2006 had almost no net effect on the wages of American workers—including those with no high school degree.

How did our common-sense intuition get this crucial fact so wrong? Our best understanding now is that two broad forces yield this counterintuitive result. For the low-skill jobs that exist, there is extremely little competition between native workers and immigrant workers. And the indirect economic effects of low-skill immigrants’ work are the reason that many low-skill jobs exist in the first place, including the ones filled by native workers.

First, the vast majority of low-skill immigrants are not competing with any U.S. worker at all. The problem is not that Americans entering the labor force aren’t willing to take certain jobs. It’s that there aren’t enough Americans to do them.

In the next decade most of the fastest-growing occupations will be jobs that require less than high school education. The US Bureau of Labor Statistic projects that between 2010 and 2020 there will be 3.6 million new, additional jobs in the US economy in home health care, basic childcare, and food service, and other low-skill industries. That is, the Census Bureau reports, is more than double the total American workers age 25–54 who will enter the labor force in the same period: 1.7 million.

This point bears repeating: even if every single American entering the labor force in the next decade dropped out of high school and college to do low-skill work, they wouldn’t be able to fill half the labor positions America needs in these areas. And of course that isn’t going to happen: In that period roughly 30 percent of Americans entering the labor force will have high school only, and under 10 percent will have less than high school. That means we can reasonably expect somewhere around 5–15 percent of these very low-skill jobs to be filled by American workers. For the rest, no American will be competing with immigrants.

Second, low-skill immigrant labor has numerous economic effects that end up, indirectly, creating new jobs for low-skill workers—including American workers. These effects are difficult to observe precisely because they are indirect.

Economists have identified these and other indirect ways that low-skill immigrant workers create jobs for other low-skill workers:

· Low-skill immigrants are not just sellers of their own labor; they spend money, meaning that they are consumers of other people’s labor—across all skill groups.

· Very low-skill immigrant workers tend to specialize in different tasks than moderately low-skill American workers. For example, on many American farms, native workers with some high school education often perform very different jobs than immigrant workers with no high school education (paper here, ungated here).

· Low-skill immigrant workers make other workers’ paychecks go further by making basic services like restaurant food and landscaping more affordable (paper here, ungated here) and thus driving up the demand for low-skill workers.

· Low-skill immigrants cause more high-skill U.S. women to enter the labor force, by making childcare and eldercare more affordable (paper here, ungated here).

· Low-skill immigrants encourage U.S. workers to stay in high school, so that they don’t have to compete with immigrants for low-skill jobs (paper here, ungated here).

· Low-skill immigrants also keep entire U.S. industries alive: many parts of U.S. agriculture would have had to close up shop years ago if not for low-skill workers who pick the crop, and that would mean the loss of all U.S. jobs in those sectors.

All of these forces tend to raise the productivity of the whole economy, generating jobs en masse, including low-skill jobs for both American and immigrant workers. It is understandably hard for people to believe in effects like these, because they are indirect and hard to discern. How, a reasonable person might ask, can a Guatemalan nanny changing diapers have broad effects on US economic productivity? You have to think for a moment about what you don’t see: the economic activity that the child’s mother is thereby freed up to perform, and how the ripple effects of that mother’s work end up bolsters whole industries and creating new jobs.

These indirect effects are the key to understanding the difference between the repeated, robust findings of this literature and the better-known research of George Borjas at the Harvard Kennedy School. In a paper published ten years ago, Borjas measured the effect of additional immigrant labor on native workers when many of these indirect effects do not occur. For example, the Borjas paper assumes that low-skill natives and lower-skill immigrants do not complement each other by specializing in different tasks, and it assumes that lower-skill workers do not raise the productivity of higher-skill workers (as when a low-skill immigrant nanny allows a high-skill US woman to enter the labor force). If what you want to do is measure the pure, idealized relationship between the amount of labor and its price, the ways that other workers react to and accommodate inflows of immigrant labor are a distraction. If instead what you want to do is measure what really happens in an economy when immigrants enter it, it is critical to account for the ways that other workers react and accommodate the flows. Workers within immigrants’ skill groups specialize in different tasks to avoid competing with them, and workers outside immigrants’ skill groups depend on the labor of immigrants to make occupation and education decisions that raise their own productivity. Borjas’ work removes these effects, Ottaviano and Peri’s work includes them. Both are high-quality economic research, but only the more recent work of Ottaviano and Peri addresses the full effects of immigrant workers on US workers.

The closest thing that science has to “laws” are findings that pass empirical tests again and again, in all kinds of settings. Immigrant labor is still an active area of research, but the absence of substantial negative labor-market impacts of immigrants on native workers may be one such broad, robust finding. This discovery will certainly take a long time to enter the public consciousness, if it ever does. In the meantime, public discourse will likely go on assuming that low-skill immigrants are somehow economically harmful. In fact, low-skill workers are now and have always been the backbone of the American economy. A substantial majority of the ancestors of today’s Americans arrived here as low-skill immigrants, not PhDs. Low-skill immigrants are not“cheap labor”: they typically earn vastly more in the United States than they could at home, andthanks to migration their labor is receiving proper compensation for the first time in their lives. Low-skill immigrants are foundational to the U.S. economy: every job, from janitors to biochemists, depends on a legion of low-skill workers that made that job possible. Throughout history that economic contribution has been contested and ignored. Whether recent discoveries in social science can alter that pattern remains to be seen.