Immigration and Innovation: What’s the Connection?
Perhaps no social-economic policy issue is more rife with contention and confusion than immigration. In recent days, a conflict has broken out in Trump circles. While certain tech leaders like Musk and Ramaswamy are stressing the importance of immigration, especially high-skill immigration and H-1B visas, many “America Firsters” like Steve Bannon are decrying H-1B visas.
At one level, this debate boils down to values and goals. Democrats generally want more immigration across the board, in part for partisan political reasons (they believe immigrants are more likely to vote Democratic), social policy reasons (they think national origin and racial diversity benefit the nation), and moralistic reasons (they believe it is morally right to let the underprivileged from other nations into America). Republicans generally want less immigration for pretty much the opposite reasons.
This most recent spat is specifically about H-1B visas—visas that allow companies to bring workers to the United States temporarily in hard-to-fill occupations, especially in technology fields.
Before looking at H-1Bs, it’s worth first discussing the effect of immigration on innovation. To do that, it’s important to distinguish between two kinds of innovation: product and process. Product innovation is the development of a new product (or service) or the improvement of an existing one. Process innovation is the adoption of a new way of producing a good or service. The latter is critical to boosting productivity. For example, workers using pneumatic hammers are more productive than workers using regular hammers.
When it comes to immigration, the relationships are opposite. By keeping wages low, more low-skill, low-wage immigrants reduce process automation and productivity growth. When wages are lower, the payback from investing in machines is longer. In other words, greater numbers of low-wage workers reduce capital intensity as it becomes easier for employers to substitute workers for machines. Why invest in apple-picking robots when low-wage immigrants can do the job?
Napasintuwong and Emerson estimate that limits on low-wage immigration to the United States would stimulate the adoption of additional available labor-saving technology with an increased substitution of capital for labor. Assuming less low-skill immigration leading to a hypothetical 10 percent increase in the wage rate, their estimates suggest an 18 percent increase in the capital-to-labor ratio. By contrast, an even less restrictive policy in the United States toward lower-skill foreign workers would reduce the incentives for adopting new technology, reducing the extent of substitution of capital for labor.
But what about higher-wage immigration, especially STEM workers? In this case, more STEM workers would mean less wage pressure, meaning that companies, all else equal, will employ more STEM workers and, by definition, innovate more. This, by the way, is the logic behind the R&D tax credit (most of which goes for credit on R&D labor expenses): By lowering the after-tax rate on R&D labor, policymakers encourage companies to spend more on R&D.
Multiple studies have shown that high-skilled immigration, particularly STEM-based immigration, boosts innovation. Just like having Nikola Jokić and Luka Dončić in the NBA makes the game much better (even if it means two fewer native-born Americans will have an NBA career), having the best foreign STEM talent in America strengthens American innovation. However, unlike the NBA, which is fixed in size (at least in the short run), more American innovation means higher demand for STEM workers in the United States because the technology companies gain global market share.
So, what about H-1B workers? Some argue that employers rely on H-1Bs to cut wages. However, various scholarly studies show this is not the case and that they actually boost overall firm employment and wages.
Either way, if Congress is worried about H-1Bs being used to limit wages, rather than eliminate the program, as some MAGA advocates want, it could change the program so that visas are allocated not by lottery, but by an auction process where firms that are willing to pay more get first crack. These market forces would mean fewer firms would use the program to get lower-wage workers. Moreover, the revenues from the auction could then be dedicated to a National Science Foundation program that supports STEM education in the United States.
However, even though the H-1B program is a convenient whipping boy, for many, the real issue is immigration per se. It is notable that at least some of the opponents of the H-1B program also oppose green card STEM immigration, even though these workers would have labor mobility rights and presumably the same ability to negotiate for higher wages with American citizens. These advocates just don’t want any more competition for their STEM members, just as the doctors’ lobby often opposes the expansion of medical schools. Who can blame them? No one likes competition. But others, particularly on the right, just don’t want any immigration, especially from non-majority-white nations, even if the immigrants have PhDs in quantum physics. Just like free-market ideologues on the right say, “Potato chips, computer ships, what’s the difference,” these nativists say, “Foreign law graduates, foreign computer science graduates, what’s the difference.” Hint: The latter makes America great.
The anti-H-1B advocates also fixate on wages. Even if STEM immigration—H-1B or green card—lowers wages in an occupation due to more competition, that is not really the point. Lower wages in particular occupations in companies are usually passed along to customers in the form of lower prices. So, while some occupations might pay less, other American workers make more because the cost of goods and services they consume goes down.
We see the same dynamic and debate regarding automation. If, for example, we could overcome union resistance to automating U.S. ports more, there would be fewer dock workers in the future (although any labor contract could manage the decline through attrition). But, prices for Americans would be lower, and current and future America would be better off.
There are a few other points of confusion about STEM immigration. First, opponents assume that getting a STEM degree is no different than getting an art history or sports management degree. If we let in fewer foreigners, more Americans would naturally go into these fields because the wages would be higher. But, wages in STEM are already very high, and there are still domestic shortages. Moreover, STEM is not like other fields. To succeed in STEM, one needs a relatively strong IQ. The narrative in the United States, especially among those on the left, is that not only is the very concept of IQ fallacious, but with just more focus on inclusion, we can fill the STEM gaps with more women and underrepresented minorities. However, this has been a national goal for at least 25 years, and little progress has been made. While more efforts should be made, to assume, without evidence, that this strategy will work, especially in the short-term, and that America does not need to compete for the best STEM talent in the world is wishful thinking at best.
Next, the choice to enter a STEM field is not principally determined by lifetime compensation. It is determined, at least in part, by a willingness to work alone on mathematical problems and a personal orientation toward introversion. Not everyone wants to do this, regardless of salary. I have made it a practice to ask younger people I meet who have STEM graduate degrees and are not working in STEM why. Without exception, the answer is that they did not enjoy the day-to-day work of being a scientist or engineer. Some people do; some people don’t. Boosting salaries by 20 percent will not move the needle for most people.
Moreover, this is a chicken-or-egg issue. The United States desperately needs to expand its advanced industry output; many of these sectors employ relatively high rates of STEM. But it can’t do that without more STEM jobs. And students will be less likely to go into STEM if they think the United States will continue to lose global market share in STEM-based industries. We see this chicken-or-egg dynamic in semiconductors where the CHIPS Act will boost production, and hopefully, the student supply of semiconductor engineers will respond.
Also, while some anti-immigration advocates say that immigrants take native-born jobs, this assumes the lump of labor fallacy (a fixed number of jobs). It’s important to note that immigration does not mean fewer overall jobs for Americans. Anyone working in America—a native-born worker or an immigrant—spends the money they earn, creating demands for additional work. This is why more immigration cannot solve the so-called problem of low overall employment. Immigration is only an answer to particular occupational shortages.
Some argue that there is no shortage because if there were STEM wages would “soar.” There are two things wrong with this argument. First, most STEM jobs are in globally traded sectors without the luxury of not worrying about cost. If STEM wages were to soar, companies in the United States would be less globally competitive. And related to that, most production systems are global, for better or worse. If U.S. STEM wages soar, companies would just hire workers in other lower-cost economies. STEM is not like trucking or nursing, where there is no international competition and higher wages do not price the industries out of the market. Finally, according to the U.S. Bureau of Labor Statistics, software developer jobs pay 111 percent more than the average job, which is hardly low wages.
Everything above gets to the central point: U.S. immigration policy, like policy in so many other areas such as science, tax, education, and regulation, should focus on advancing strategic skills in strategic sectors. There is no reason financial services companies or hospitals need workers with H-1B visas. For the most part, these are not traded sectors and are certainly not strategic in terms of U.S. competition with China. Plus, there is no evidence that Americans don’t want to go into nursing or banking. However, it is in America’s best interest to have more immigrants working in key areas like materials science, engineering, life sciences, and computer science and key industries like aerospace, computing, robotics, biopharmaceuticals, and AI.
And even if having more of these workers means that some Americans don’t get into these fields, which is unlikely given the stagnation in the number of Americans entering these areas, who cares? The nation as a whole will be stronger.
It was clear that Bidenomics meant more immigration—legal and illegal, high-skill and low-skill—and that was and is bad for America. However, it would be a tragedy if the Trump administration could only respond in a Hegelian way: no immigration of any kind. I get that the United States has had too much immigration. I get it; the United States needs to crack down on illegal immigration, especially by punishing companies that employ illegal immigrants, and I can even go along with the idea of reducing overall legal immigration rates. But limiting the immigration of highly skilled physicists, chemists, computer scientists, mathematicians, and engineers would be a huge mistake. That would not make America Great; it would make China Great.
This means the MAGA movement needs to decide what Trumpian economics is: Americans first, or America first. Putting Americans first (as individuals) means no automation, no layoffs, and no STEM immigration. Putting America first means taking steps to restore America as the most powerful nation on earth. A key step toward realizing that goal is redesigning our immigration system to significantly reduce low-skill immigration while allowing the world’s best scientists and engineers to work here in strategic sectors. I support America first.
