A global anti-immigration movement is gaining momentum. In the United States, ICE carries out mass raids to deport migrants. Across Europe, a number of countries have introduced harsh restrictions on newcomers. In Russia, thousands of migrants are about to lose their jobs in taxi services and food delivery. Politicians offer a variety of unjustified claims about migrants driving up crime or pushing down economic indicators. Yet numerous studies point to the opposite: migrants do not lower wages or reduce employment opportunities for locals — not even in low-skilled sectors — nor do they hold back innovation. And notably, technological progress does not shrink the overall demand for labor — it creates new jobs, even for workers with lower levels of qualification.
Labor markets have recently turned into a battleground for political rhetoric. Donald Trump warns Americans that migrants are “stealing their jobs.” His vice president, J.D. Vance, insists that an abundance of cheap labor is slowing down technological innovation. Trump’s point is a classic — but outdated. Not only experts, but even the general U.S. population (64%), no longer believes it, understanding instead that newcomers tend to fill jobs that locals often avoid due to low pay, hard working conditions, or a lack of prestige. In Russia, however, the myth of migration as a threat to the labor market remains strong: 44% of respondents believe migrants are taking jobs away from locals, while only 37% disagree.
Over the past few decades, there has been a large number of studies across dozens of countries aimed at measuring the impact of mass migration on local populations — both on society as a whole and on groups whose skills and qualifications are closest to those of migrants. When it comes to the general population, there has never really been a debate among economists: migration makes countries richer, and average living standards rise. The controversy, for a time, centered on specific groups of locals — typically low-skilled workers. In 1990, economist David Card published a landmark study analyzing the consequences of the so-called Mariel boatlift — the mass exodus of Cubans to Miami in 1980 that followed Fidel Castro’s sudden announcement that anyone wishing to leave the island could do so through the port of Mariel. About 125,000 Cubans, most with little formal education, crossed to the United States by boat, swelling Miami’s labor force by 7% almost overnight. Card compared the evolution of wages and employment among Miami’s native workers before and after the migrants’ arrival with trends in four similar U.S. cities — Atlanta, Houston, Los Angeles, and Tampa. He found no evidence of declining wages or a rise in unemployment — and the positive trends held not only for Miami’s population as a whole, but even among earlier waves of Cuban migrants whose skills and background were expected to be hit the hardest by the sudden surge in competition.
In 2017, Harvard economist George Borjas attempted to challenge Card’s findings. Expanding the sample to include more cities, he focused narrowly on non-Hispanic men who had dropped out of high school — arguing that they were the group most directly exposed to competition from Cuban migrants. For this sampling segment, Borjas claimed, local incomes fell sharply. But in 2018, economists Giovanni Peri and Vasil Yasenov revisited the data. They widened the sample to include Hispanic high-school dropouts (who by education and background shared a demographic profile that was much closer to that of the arriving migrants) and also women, whom Borjas had inexplicably excluded. Their recalculation refuted Borjas and upheld Card’s original conclusion: there was no decline in wages or employment, even among these more economically vulnerable groups.
Other studies have reached similar conclusions. Economists have analyzed, for example, the repatriation of Algerians of European descent to France after Algeria’s independence in 1962, the mass immigration from the Soviet Union to Israel after exit restrictions were lifted in 1990 (which increased Israel’s population by 12% in just four years), the surge of European migration to the United States between the 1910s and 1930s, and long-term research on the effects of contemporary migration to Denmark. In all these cases, researchers found no significant negative impact on local populations. In fact, in the case of European migration to the U.S., employment among locals actually rose.
By now, a broad consensus has emerged among economists. In 2017, the U.S. National Academy of Sciences convened all the leading experts on migration — both advocates and skeptics, including none other than George Borjas. The resulting report delivered a clear conclusion: “The empirical research of recent decades indicates that, when measured over a period longer than 10 years, the impact of immigration on the wages of native-born workers is very small.”
So why don’t local incomes fall when the supply of labor rises? Economists Giovanni Peri and Chad Sparber maintain that the reason lies in the way migration deepens the division of labor. Migrants typically take on more physically demanding jobs — in construction, cleaning, harvesting — while locals move into roles that require language proficiency and communication skills, such as management, administration, or sales. And since these communication-heavy jobs tend to pay more, the earnings of locals actually rise.
The positive effect is particularly striking for women. As World Bank analysts note, the arrival of low-skilled female migrants enables highly skilled local women to spend more time improving their careers. Locals can delegate household tasks to affordable domestic helpers and focus on their professional lives. In turn, overall prosperity rises thanks to a more efficient allocation of labor.
Overall prosperity rises thanks to a more efficient allocation of labor
History offers many examples of this deepening division of labor. In the 1930s, the United States ran the Bracero program (Spanish for “manual laborer”), which brought 76,000 Mexicans into the agricultural sector. During the Great Depression, the program was shut down in the hope that limiting the labor force would free up jobs for Americans. But as another study by Peri and his colleagues shows, the absence of Mexican workers actually reduced the number of managerial and coordinator positions and pushed Americans into less prestigious, lower-paid manual labor.
At the same time, migrant workers don’t just expand the supply of labor — they also generate demand through their consumption of local goods and services. History even provides a telling counterexample. After the fall of the Berlin Wall, Czech citizens were allowed to work in German border towns without being granted residency permits — forcing them to return to the Czech Republic and remain consumers there. The program had a negative impact on German workers, as a 1% increase in the share of Czech workers in the local labor force reduced German wages by 0.13% and employment by 0.9%.
Finally, migrants themselves often become entrepreneurs, creating additional jobs. In the U.S., a migrant is 30% more likely than a native-born person to found a company that grows to at least 10 employees within five years. These businesses also play a crucial role in absorbing migrant labor: in 1992, 43% of migrants in the U.S. were employed by firms founded by fellow migrants.
Vance’s critique of immigration also leans on the idea that cheap labor reduces the incentive to innovate. There is, indeed, a link between innovation and labor shortages — but it is not as straightforward as the vice president suggests. Sometimes, a scarcity of workers can spur technological progress. For example, the catastrophic Mississippi Delta flood of 1927 drove large numbers of Black residents out of the region, and as a result, landowners who had relied on their labor were forced to invest in mechanizing their farms. Soon, agriculture in the flood-affected areas became more technologically advanced than in neighboring regions that remained untouched by the disaster.
A similar story unfolded after the termination of the Bracero program. Farmers who had depended on Mexican workers had no choice but to adopt new technologies. This shift even triggered a rise in patents registered in the affected regions. Before the program ended, tomato harvesting equipment was virtually absent in California — unlike in states such as Ohio, where braceros never worked. Once the program was shut down, Californian farmers rapidly adopted tomato-picking machines.
But even this case comes with complications, as Shmuel Sand of the Hebrew University of Jerusalem points out. The mechanization wave did not benefit California’s consumers or its farmers. On the contrary, their incomes fell over the long run, while tomato prices rose. Farmers also abandoned a wide range of crops — such as asparagus, strawberries, cucumbers, celery, and lettuce — for which harvesting machines had not yet been invented. In other words, relying on migrant labor rather than technology had been the more rational economic choice at the time, and the main winners from the end of the Bracero program turned out to be the companies selling tomato-picking devices. Something similar happened under Donald Trump, who was ultimately forced to halt deportations of farm workers simply because the U.S. had neither the domestic labor supply nor the technologies to replace them.
Of course, the effects of migration and innovation vary across economic sectors. A study comparing patent data across 41 countries found that “an increase in low-skill wages leads to more automation innovation,” while “an increase in high-skill wages tends to reduce automation innovation.” The differing effects of wage increases in low- and high-skilled segments of the labor market stem from a simple fact: low-skilled work is easier to automate, since it typically involves repetitive tasks that do not demand much creativity. Rising costs for high-skilled labor, on the other hand, tend to push firms to cut back on investment in innovation altogether.
Technology and skilled workers tend to complement one another, argues 2024 Nobel laureate Daron Acemoglu. Technological development, he notes, depends largely on the type of labor available at a given moment. When the market is flooded with low-skilled workers, technology evolves to suit them, but once a sufficient pool of skilled workers emerges, technologies become more sophisticated. In short, when low-skilled labor becomes relatively expensive, firms are pushed to invest in skill-intensive technologies, which in turn increases demand for a more educated workforce.
However, without a sufficient supply of educated workers, such investments make little sense. This is why countries like China — or other developing economies — could not simply import Western know-how until they had first built up a large enough pool of skilled workers to operate them.
Technology and skilled workers need each other
In the 19th century, technological development was aimed both at boosting the efficiency of unskilled workers and at automating the labor of highly paid craftsmen. This is why the mass influx of migrant laborers to the United States between 1840 and 1920 fueled industrial growth and the rise of new factories. However, as overall education levels increased — driven in large part by rising demand for workers who could operate new machinery that required basic literacy — the number of skilled professionals increased, and the industrializing trend reversed. For most of the 20th century, technologies were designed to complement skilled workers, enhancing their productivity while automating the work of the unskilled. During this period, wages for skilled workers rose in line with growing demand for their expertise, while the low-skilled saw a decline in wages as their labor was increasingly displaced by machines.
It should be noted that the same principle applies to more than just labor. Scarcity drives up prices and pushes companies to invest in more accessible alternatives. High hydrocarbon prices, for instance, have spurred the development of alternative energy sources. Firms are always resource-constrained and must decide where to allocate funds first. This means that investing heavily in automation — replacing people with machines — can lead to underinvestment in other critical areas.
As Daron Acemoglu and Boston University economist Pascual Restrepo note in a series of studies on automation, the process can certainly displace workers, but it can also create new tasks, thereby increasing demand for labor. This is why the industrial revolutions did not result in the replacement of human work. Two hundred years ago, most people worked in agriculture — a century ago, in industry. Today, in advanced economies, the majority of people are employed in services, and most of GDP is generated there.
Automation in some areas has consistently created new tasks and demand for labor in others
Economists point to three main reasons why automation can actually increase demand even for low-skilled labor.
- Automation reduces costs, boosts productivity, and makes goods cheaper. In turn, this raises demand for those goods. As a result, automation often drives the expansion of production — and with it, the demand for the non-automated labor that production still requires. Take ATMs, for example. Their invention reduced the number of bank tellers needed per branch, but it also lowered the cost of running branches, which encouraged banks to open more of them. The result was an overall increase in bank employment. In 1985, the U.S. had 60,000 ATMs and 485,000 bank tellers. By 2002, those numbers had risen to 352,000 and 527,000 respectively.
- Automation fuels demand for machinery, which in turn boosts employment for the workers involved in its production.
- In sectors where technology generally dominates, further automation — for example, the replacement of horses with tractors in agriculture — does not displace human labor. However, technological improvements in such areas still leads to a rise in productivity and increases demand for machinery, thereby boosting demand for labor through the two effects mentioned above.
In addition, automation usually creates demand for the personnel needed to operate and maintain machines. Over time, these tasks themselves can become standardized — making them suitable for less-skilled workers. This is exactly what happened during the First Industrial Revolution, when the tasks of craftsmen were standardized and mechanized to the point that they could be carried out by lower-skilled factory workers.
Still, there are cases in which automation backfires — when low-skilled workers are replaced by expensive but less efficient machines that also require maintenance from employees whose skills could be put to better use elsewhere. Elon Musk’s attempt to automate as many tasks as possible in Tesla’s factories is a telling example. As a result of his efforts, productivity actually fell, and by 2018 the company had missed its production targets. Musk was forced to admit that the push for automation had gone too far — and that he had underestimated the productivity of human workers.
A good example of a technology that both requires substantial labor input and boosts efficiency is the rise of taxi platforms like Uber or delivery services like Amazon. These technologies could not exist without a large pool of low-wage, low-skilled workers whose main requirement is the ability to use a smartphone. At the same time, they have significantly lowered the cost of such services, making them widely accessible — thereby stimulating demand for countless goods and improving the allocation of time and effort for millions of workers who use them.
Excessive automation — not automation itself — leads to lower decreased employment and lower wages
In recent decades, the United States has seen a move toward technologies that displace labor without creating new tasks, note Acemoglu and Restrepo. This is largely because labor is taxed and capital is not — in fact, capital is subsidized by the state. The result has been stagnant productivity and growing inequality, as firms cut back on both labor and technology rather than advancing innovations that would raise labor productivity.
The right combination of policy prescriptions, however, would solve this problem, and taking a more sensible approach can also ensure that future technological developments have the same overall positive effects that past innovations have. The rise of artificial intelligence is unlikely to wipe out human labor altogether, since AI is still capable of performing only narrow, specialized tasks, while in many fields the human mind remains more effective. Yet AI will almost certainly displace many workers, and that reality will need to be balanced by the creation of new tasks for people — tasks that can raise overall economic productivity.
Researchers Acemoglu and Restrepo see real potential for augmented reality technologies to create exactly these types of new tasks in the fields of education, healthcare, and manufacturing. First though, it will be crucial that government action is taken in order to facilitate the retraining of workers whose jobs were lost.
Migrants, in many ways, function in the economy much like machines: they serve as a substitute for some tasks while creating others through a more efficient division of labor. Now though, even many developing countries face declining birthrates as people live longer, move to cities, and face rising costs of raising children. As a result, the recent surge in migration from the developing world into advanced economies is likely to taper off. In this context, investments in automation — even if they do not immediately boost productivity — can still be justified.
Countries that have long-since confronted labor shortages due to limited migration and aging populations offer interesting case studies. The number of industrial robots per 1,000 industrial workers in the United States is a comparatively small 8 — much lower than in Japan (14), Germany (17), or South Korea (20). Those latter three countries have been investing heavily in labor-substituting automation for decades, yet their economies have grown much more slowly in recent decades than that of the U.S., which has drawn on foreign migrants or of China, which has relied on internal migration.
Still, as demand for such technologies rises — driven by population aging and the rising cost of human labor — advanced economies will likely benefit from exporting the solutions they have developed, but countries that today profit from an inflow of cheap migrant labor will probably remain well-positioned in the future as well. Catch-up growth is always faster and easier: there is no need to reinvent the wheel when ready-made, efficient technologies can be adopted from more advanced economies without the need to cover the costs of failed R&D.
Moreover, the use of migrant labor in these countries frees up the work capacity of educated women and encourages locals to move into jobs requiring more advanced qualifications. This, in turn, makes the process of automation easier, since a larger pool of skilled experts becomes available.
