Immigrants and native wages

I know I’ve had you riveted with the literature review in the last few posts about the effect of immigrants on native employment outcomes (i.e. wages and employment levels). I thought it was important, though, to point out that academic economists don’t really contest the existence of this zero correlation. That said, we have learned a lot by trying to poke holes in the landmark “experimental” studies summarized here, here and here.

While those famous papers dealt with important measurement issues, we’ve also learned that experimental studies are inadequate to deal with at least two other factors, indicated by economic theory, that may be biasing the simple correlation one way or another.

First, and this is the measurement issue dealt with by the experimental studies, we want to be sure that there is not something that makes a particular area or job a more attractive place to work causing both foreigners and natives to migrate to that area or job. If such a third factor existed, it would hide the effect of immigrants on natives. Suppose, for example, there was a technological advance making a particular job more lucrative (e.g. an acceleration in computational power makes the wages of computer programmers go up) or suppose there was a surge in demand for a job (e.g. nurses are in more demand as the population ages). Because immigrants and natives would be attracted to these jobs, the increase in immigration would be positively correlated with the increase in wages and the increase in native employment in these jobs. These omitted variables, to use economic jargon, might mask the potentially negative effect of immigration on wages.

Second, natives relocating to other areas as a result of immigration can spread out any potential negative consequences of immigration among a wider population of natives, making those bad effects hard to detect. For example, when immigrants move to California, natives may move to other states. These spillover effects would cause the labor supply in those other states to increase putting pressure on wages there, too.

Third, we need to take into account the total effect of immigration. Immigrants with a particular set of skills will compete with natives with those same skills and decrease their wages, but they will make natives with complimentary skills more productive, increasing those natives’ wages. An immigrant that puts up drywall on a construction site, for example, competes with native drywall installers, but they compliment the native foreman who may be able to spend more time managing the construction site rather than helping put drywall up. Furthermore, when we properly account for all these indirect effects immigrants have on natives, we will want to pay special attention to the possibility that natives and immigrants, even those with the same skills, do not perfectly compete with each other. To return to the example: among drywall installers, natives may specialize in particular tasks like reading blueprints or leading work teams while immigrants specialize in hauling materials or the actual installation. To use the economic jargon again, we need to check to see if natives and immigrants are perfect substitutes or not.

In the last few years, researchers have shown that when all of these factors are taken into account, the surprising result is that immigrants don’t seem to have a negative effect on native wages or employment levels. In fact, its likely that immigrants have even been a net positive for natives, a result that this paper has replicated.

So there is not much debate about this non-correlation. Academic economists have moved on to try to explain why no correlation exists.

7 thoughts on “Immigrants and native wages”

  1. Hey Will,

    Just ran into this, a very nice summary of the literature. Since I also find these issues fascinating I thought I’d add another channel to the potential “zero correlation” culprits. If immigrants are cheaper than natives then there is a very straightforward “productivity effect”. I.e., firms can produce the same output with fewer inputs. As a result, firms that hire more immigrants will do relatively better and will expand output more. This drives up demand for all workers.

    Obviously there are many caveats in general equilibrium, but I like this story and have found evidence for it. This story comes from the offshoring literature and makes me think, as usual, there’s a lot to be gained from comparing stories across the literature.


  2. This is from

    “Following the 1986 immigration reforms, for example, previously-undocumented immigrants experienced big pay boosts – as much as 15 or 20 percent – and immigrants who already had legal status saw hefty wage gains, too. But the reforms also led to higher wages for lower-skilled native-born Americans. ”

    This writer argues that immigration reform is good because it raises wages for the lowest tier of the workforce. He also argues that immigration does not lower wages.

    How are those things compatible?
    If immigration does not lower wages, immigration reform should not raise wages.

  3. Nice catch. I’m not familiar with the studies that author is referring to, but here’s my guess: The way to make those two facts compatible is to think about how natives and immigrants compete with each other in the labor market. Without reform, illegals would choose to work in a limited number of jobs because they may be afraid of being caught if they were in a job with too much exposure. Because there was so much supply in those few jobs, this would depress wages by a lot. I’m guessing those same jobs are where many low skill natives were trying to get too. After reform, immigrants felt more comfortable working in other jobs, reducing the competitive pressure on natives (and increasing wages).

  4. A very delayed response here: yes that is what Ottaviano, Peri, Wright is about, in part.

    A thought on the ’86 immigration reforms, which have got me a bit confused now. Clearly wages should rise for the formerly undocumented workers, since they can no longer be discriminated against so easily. As for lower-skilled native workers, it must simply come down to the elasticity of substitution between them and the formerly undocumented. If these two groups were complementary then higher wages for undocumented would result in *lower* wages for natives. But seemingly they are strong substitutes so increased wages for one group raises the wage of the other.

    Note that this is a bit in opposition to work suggesting that wages should not fall due to immigration because of complementarities between natives and immigrants, as possibly suggested in the comment above from lark. So I’m a bit stumped on this still, I guess. Your explanation is interesting, except that I’m a bit unsure of the idea that natives are competing to work in the fields of the central valley, or in food processing plants in laredo, texas. Perhaps the latter is reasonable actually. Sounds like room for a paper here in any case…

Comments are closed.