The Immigration Shimmy

Immigration expansionists face a difficult challenge: they support high levels of immigration—including many more less-skilled immigrants—for a variety of legitimate reasons, but the less-skilled immigration has detrimental economic effects on vulnerable American workers, including earlier generations of immigrants already in the country. That tradeoff is one the expansionists would make, but it is understandably unpopular. One option available to them would be to argue their case on the merits: Here is why we should pursue this approach to immigration despite its costs, here are the ways we can compensate those who are harmed, etc. That would be an honorable and interesting debate. Unfortunately, most have chosen a different option, which is to insist that no such tradeoff exists.

It’s not easy to avoid the obvious issue that flooding the labor market’s low end will harm those already there. As I am fond of noting, shorn of its political implications this reality is entirely uncontroversial. Reporting on the return of immigrants from the United States back to Mexico, the Washington Post wrote in 2017, “More returnees means lower wages for everybody in blue-collar industries such as construction and automobile manufacturing, where competition for jobs is likely to increase, economists say.” Libertarians are happy to acknowledge a comparable dynamic when it comes to occupational licensing in particular segments of the labor market, which “restricts the supply of labor. … As the supply of labor is reduced, the price of their services will increase.” In health care, “less restrictive licenses reduce physician salaries.”

Expansionists sidestep the inconvenient truth with a subtle shimmy, responding to concerns about the effect of their policy on particular classes of workers by making a general claim about how immigration will eventually affect average wages across the economy as a whole. To watch the Immigration Shimmy in action, contrast my recent post on the tradeoff between liberal immigration policy and tight labor markets with responses from the Cato Institute’s Alex Nowrasteh and Bloomberg columnist Noah Smith. Alex warns that I “reject the economics of immigration,” but while I question the effects of “unconstrained immigration into the labor market’s weakest segments,” he assures us that “an increase in the supply of workers doesn’t lower wages in general.” I suggest that “immigration places downward pressure on wages of comparable workers,” while he says that “the overall long run wage effect of immigration to the U.S. economy is zero.” I question the wisdom of a policy approach that holds “the more unskilled workers willing to accept low wages the merrier,” while he defends the approach with a theory that holds “overall in the long run” (all emphases added).

Confusingly (though to his credit), Alex pounds the table for his maximalist but unresponsive claims, but then concedes the salient debate about effects on the particular labor-market segments in which less-skilled immigrants most likely compete. In the short run, he writes, “an increase in the supply of workers might push down wages.” Even in the long term, he acknowledges that his theory sometimes “doesn’t hold for workers in certain occupations or in specific regions of the country.” He quotes from a National Academy of Sciences report that acknowledges negative impacts occur and notes, “they are most likely to be found for prior immigrants or native-born workers who have not completed high school—who are often the closest substitutes for immigrant workers with low skills.”

Noah does something similar, boldly titling his post “why immigration doesn’t reduce wages” but then qualifying his claim in the very first sentence—he’s only talking about “native-born” workers (he never says why foreign-born Americans don’t count for him) and immigration does lower native-born wages (though he describes the magnitude as “possibly a little bit, in a few special circumstances”). He concludes, “we find that immigration can occasionally have some small negative impacts on labor markets.” Commenting on the situation, Matt Yglesias observes, “everyone can be right here. The idea that immigration restrictions can increase relative wages in immigrant-heavy occupations is not the same as saying that they can increase overall wages and incomes — relative wages will go down elsewhere as the economy adjusts — and understanding the difference is critical.” Everyone can be right if they state their views clearly. When Noah writes “immigration doesn’t reduce wages,” on the other hand, he is wrong.

Theorizing the Shimmy

A delightful feature of the Shimmy is the handwaving that accompanies the backward shuffle. To argue with a straight face that an influx of labor supply does not affect the affected workers, one must have a theory as to why. Alex and Noah offer starkly different theories, neither of which holds water. That’s not surprising—recall, they both admit that immigration does affect wages, so we shouldn’t expect a theory to the contrary to work very well. Connoisseurs of the Shimmy will appreciate the effort.

Alex contends that wages are determined solely by worker productivity, regardless of market dynamics. Large influxes of less-skilled immigration cannot affect wages, he writes, because “demand for workers is determined entirely by the marginal value product (MVP) of the worker. The MVP is the quantity of goods or services supplied by a worker multiplied by the market price for those goods (marginal physical productivity times market price)” (emphasis in original). Put simply, employers will pay workers what they are “worth,” and adding millions of new workers can’t change that. Thus, in Alex’s labor market, “The overall long run wage effect of immigration to the U.S. economy is zero and it can’t be any other way.”

This idea has the simplistic and superficial appeal characteristic of market fundamentalism, which copes with the disconnect between crisply abstract models and real life by clinging to the former. As a result, it is demonstrably obtuse in too many ways for a single blog post to cover, but to run briefly through a few, it assumes: that worker power does not matter, that employers do not respond to supply constraints with productivity investments, that end-markets do not respond to supply constraints with increased prices, that entrepreneurs and capital always stand ready to put any available workers to their best use, that short-term labor shocks do not have long-term effects—oh, and that firms even know the MVP of their workers and attempt to set wages accordingly. None of this is true. For a discussion of how firms do set wages, I recommend this excellent post by Michael Lind.

Noah has a different (and better) theory, which emphasizes that immigrants expand not only labor supply but also labor demand:

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.

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.

This is, again, a reasonable description of immigration—in the aggregate, if the skill distribution of immigrants mirrors that of the existing population. It does not hold for less-skilled immigrants in particular. As I explained in The Once and Future Worker:

The skill-based distinction is critical to understanding immigration’s aggregate effect on the labor market. The economy has no fixed number of jobs that immigrants might “take” from native workers, nor do higher or lower levels of immigration necessarily lead to higher or lower wages. Immigrants are both producers and consumers, so their presence increases both the supply of labor and the demand for it. But we should expect immigrants who compete in a specific segment of the labor market to suppress wages of existing workers in that segment.

This is because each immigrant’s economic footprint is asymmetric. All people have somewhat similar consumption patterns, regardless of their skill and income levels. High-income households consume more overall, and more of luxury goods in particular, but everyone uses housing and transportation, food and health care, education and entertainment. So every new immigrant adds demand for many types of goods and services provided by many types of workers. On the other hand, each immigrant is himself only one type of worker. A worker with only a high school diploma contributes far more to the supply of such workers than to the demand for such work; a highly skilled worker, by contrast, adds at least somewhat to the demand for less-skilled work while adding to its supply not at all.

Defenders of “immigration” as a general phenomenon miss this point when they argue that recent waves have not been responsible for wage declines. They may be right about that, but they are answering the wrong question. The issue is not whether immigration has been or will be better than no immigration; it is whether we are admitting the appropriate mix of immigrants. The counterfactual to ponder is how different things might look had we welcomed highly skilled immigrants but more tightly restricted the entry of those who compete directly with less-skilled native workers—or even had we simply enforced the immigration laws on the books, which are flouted predominantly by less skilled immigrants and their employers. Going forward, regardless of the policy adopted for immigration in the highly skilled segment of the labor market, lower-skilled American workers will do better if their segment of the labor market has relatively fewer of them.

This last point is the critical one for getting beyond the Shimmy to the policy debate. If the United States has a goal of improving labor-market outcomes for less-skilled American workers, restricting the in-flow of less-skilled immigrants is a potent policy tool that we have at our disposal. This is a point that both Alex and Noah implicitly accept when they limit their argument to immigration overall and its effect on the labor market overall. It is also, interestingly, a point that Professor Ethan Lewis makes in a recent Cato Institute symposium for which Alex penned the introduction. “The Right Model” for immigration economics, writes Lewis, recognizes that “immigration affects the wages of one type of worker when it affects the ratio of the number of that type of worker to other types of workers.”

The Study Flutter

Those most skilled in the Immigration Shimmy accompany it with the Study Flutter, in which they Shimmy through numerous studies that they say prove immigration does not affect wages, while the studies mostly make weaker or unrelated claims. Alex points to more than twenty studies and Noah cites nearly that many, some but not all overlapping.

The findings are all over the map. Some find negative wage effects, others do not. Some don’t even look at wages but, for instance, find negative effects on employment. As far as I can tell, the only ones to look at the American labor market and find no negative effects from immigration on less-skilled workers already here pertain to the infamous Mariel Boatlift, and even there the literature is mixed. From Alex and Noah’s cornucopia of regressions, I would direct readers toward three studies that I find helpful in distilling the field’s contours:

1. “The Economic and Fiscal Consequences of Immigration,” from the National Academy of Sciences. Alex starts here as well, though I’m not sure why. It does say that long-run estimates tend to find no wage effect but, as the report explains in a footnote, this is because the models assume no effect: “this result is built in by theoretical assumptions.” Meanwhile, it says in bold, “there are larger negative effects on native wages from immigrant inflows in the short run.” The report’s Table 5-2 summarizes the findings from a large number of studies, nearly all of which find a negative wage effect.

2. “The Effect of Internal Migration on Local Labor Markets: American Cities During the Great Depression.” This is the first paper that Noah cites after various studies of refugees. I find it helpful both for its focus on the American labor market (albeit a long time ago) and its thoughtful review of existing literature. Its conclusion, from the abstract, is that “Migration had little effect on the hourly earnings of existing residents. Instead, migration prompted some residents to move away and others to lose weeks of work and/or access to relief jobs. Given the period’s high unemployment, these lost work opportunities were costly to existing residents.” I want to quote at length from its discussion of the various labor-market dynamics at play:

Goldin (1994) documents that the mass migration from Europe at the turn of the 20th century led to a large reduction in the wages of native-born workers in high migration areas. Few studies using modern data – with the exception of Altonji and Card (1991) – have detected a wage response of this magnitude, and many find no effect on wages at all (for a survey of this literature, see Friedberg and Hunt, 1995). This disparity could reflect a true change over time in the response to a local shock – for example, a nationally integrated labor market may allow a local disturbance to dissipate more rapidly – or may simply be the result of differences in available measures and research design.

The weakness of the relationship between immigration and wages in port-of-entry labor markets has prompted an on-going discussion about other margins along which local areas may adjust. Borjas, Freeman and Katz (1997) point out that if new arrivals to a city induce some of the existing workforce to relocate, these out-migrants will spread the economic costs of immigration to other markets. More generally, with the free flow of factors between cities, the initial wage and/or employment response to a labor supply shock might be tempered by the outmigration of labor or the in-migration of capital (Blanchard and Katz, 1992). Thus, the downward pressure of immigration on wages at the national level might be larger than comparisons of local labor markets would suggest (Borjas, 2003; Ottaviano and Peri, 2006).

The empirical evidence on such factor mobility is mixed. Filer (1992) finds that immigrants crowd out existing workers on a one-for-one basis. However, more recent studies have not detected an appreciable out-migration response to international arrivals (see Card, 2001; Wright, Ellis and Reibel, 1997; and Kritz and Gurak 2001).

On the capital side, Lewis (2003, 2004) proposes that local labor markets have adjusted to low-skilled immigrants through slower adoption of skill-biased computer and information technology.

3. “Immigration Restrictions as Active Labor Market Policy: Evidence from the Mexican Bracero Exclusion.” Noah and Alex both discuss this paper, which I think provides a fascinating Rorschach test. It studies the effect on the wages of seasonal farm workers when the United States excluded nearly half a million Mexican “bracero” farm workers in the 1960s. The authors find no effect on the wages of seasonal farm workers because, as Alex summarizes, “Farmers turned to machine harvesting and planted less labor-intensive crops to take account of the new dearth of workers.”

What should we make of farms shifting to higher-productivity, capital-intensive production in the face of a labor shortage? Alex takes the lack of wage gains for seasonal workers as proof that immigration restrictions do not work. Noah is more blunt, writing, “places that had relied on guest workers saw no labor market impact.” But that’s not true. Whoever was making, maintaining, and operating the more productive equipment likely earned much higher wages than the supplanted seasonal workers. This sort of shift in the labor market’s underlying structure, the business models employed by firms, and the trajectory of capital investment, is one of the pathways we should be most enthusiastic about when considering constraints on labor supply for the benefit of those workers already here. To quote from The Once and Future Worker again:

When employers argue that, even if some native workers are on hand, they nevertheless need foreign labor, the question always looms, what is the alternative? One might be to offer more attractive wages and working conditions to lure enough native workers. Another might be to mechanize processes and employ fewer but more productive native workers. Illustrating its reporting in Arizona, the Wall Street Journal described a farmer struggling with the loss of illegal-immigrant employees who would pick jalapeños by hand for $13 per hour. He invested in machinery that he hoped (many fewer) native workers would operate at $20 per hour. This would lower total employment, yes, but if those illegal immigrants previously working the lower-wage jobs have left the country, the labor-market outcome is suddenly sunnier. People in the market will more likely have jobs, and those jobs will be better ones.

That response to a farmworker shortage is just one example of a broader principle: workers in the labor market benefit when employers have no alternative to them. A strange facet of the modern immigration debate is the tendency of the same pro-business interests that typically wax lyrical about free-market dynamism to lament in apocalyptic terms the prospect of a limited supply of labor. But the result of such limit is not, generally, economic ruin. Instead, businesses reorient themselves toward using the labor that is available. Investment goes toward making jobs more attractive and boosting productivity. In 2018, fearing that a shortage of construction workers would slow spending on new homes, Home Depot announced a ten-year partnership with the Home Builders Institute to invest $50 million in training twenty thousand additional workers. The shortage and its emerging solutions are not evidence of a market failing; they show how the market can work.

Unlike performers of the Shimmy, who would have us believe that these studies decisively prove their studiously ill-defined position, my point is not that some regression analysis of some natural experiment has vindicated me, QED. To the contrary, I think the literature is plainly inconclusive, and is useful largely for the wide variety of dynamics and considerations that it highlights, as is usually the case in academia.

That leaves policymakers where they are invariably left on most questions: faced with conflicting theories, analyses, and experience; risks and tradeoffs; and no choice but to do the best they can. Given the challenges we face as a nation and the information available, we should strongly prefer higher-skilled immigration and restrict less-skilled immigration, which will also require a robust capacity and willingness to enforce our chosen policies. This does not mean reducing the total immigration level; one might wish to increase that figure or leave it unchanged, independent of reconsidering its composition. The question of how many immigrants is a political one. The question of which immigrants has important implications for worker power and the tightness of labor markets, which should guide any decision on the economic merits.

Of course, some may see other concerns as trumping the economics. They should make that case forthrightly, not pretend they can have their dinner cheaply prepared by a housekeeper and eat it too.