The post 91PORN Letter: DOL Should Use Experience Benchmarking to Prioritize Higher-Skilled H-1B Workers appeared first on Economic Innovation Group.

The post The Near-Term Fiscal Impact of H-1B Workers at the Federal and State-and-Local Levels appeared first on Economic Innovation Group.

By Adam Ozimek

Should anyone have to compete anymore? Populists seem not to think so. Tariffs protect American companies from competing with imports. The Jones Act protects American ships and shipbuilders from competing with foreign made or owned vessels. Some populists even seem to  that states and regions within the United States should be protected from competition with each other.

But perhaps the most strained demand for protectionism comes from critics of H-1B visas. Apparently the best-paid, highest-skilled workers in the richest country in the world should be spared from competing too.

Nothing could be further from the truth.

To achieve the bipartisan goal of improving the labor market for the working class, vigorous competition among the top earners is important and even necessary.

Not only would the rest of the country benefit — from more innovation, more entrepreneurship, and faster economic growth — but in the long run even the skilled American-born workers exposed to fiercer competitive pressures will fare better too.

Won’t somebody think of the doctors?

The Heritage Foundation  that “American graduates have to compete with nearly 200,000 foreign students working under OPT and another 400,000 H-1Bs issued or renewed every year.” Heritage  that there is too much competition faced by American PhDs and graduates of master’s programs.

These comments reflect a desire to make labor markets less competitive for skilled workers. By “less competitive” I mean a world in which jobs are plentiful relative to the number of applicants, people who want a job can easily get one, and wage growth is strong.

A less competitive environment for workers looks like what we think of as tight labor markets — a fine motivation considered in isolation. Who doesn’t like tight labor markets?

But there are two problems with this approach. First, it is not actually possible to create less competitive labor markets overall via reducing the number of workers. Second, if you try to make labor markets tighter for some types of workers (say, for college-educated skilled workers) by reducing their supply, then you make labor markets weaker for other types of workers (say, the working class).

The critics simply have not thought this through. Let’s look at their argument in more detail.

More people does not mean weak markets

Smarter conservative populists recognize this. Oren Cass correctly  that “economically speaking, immigration is a form of population growth.” In his  he puts it quite eloquently:

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.

In short, the idea that you can create tight labor markets overall by reducing immigration is a pipe dream.

Tilting the scale

To understand this dynamic more clearly, let’s consider a thought experiment that holds population constant and exclusively focuses on shifts in relative labor supply.^[2]

Imagine if we suddenly made college degrees completely free, but only for those majoring in accounting. Four years later, as all the new accounting majors graduated, the labor market for accountants would be flooded by the surge of applicants, which would make that labor market more competitive. But all those young people pivoting to accounting en masse would mean fewer job applicants for every other type of job. The labor market for non-accountants would be less�����dz����پ��پ�����.

In economics jargon, what has happened is that non-accountants have become relatively more scarce, while accountants have become relatively less scarce.

This example illustrates that there is an inevitable balance when it comes to changing the composition of the labor force: a relatively less competitive labor market for some workers must be matched by a relatively more competitive labor market for others.

You can think of the effects of shifting the composition of the workforce as operating like a scale. If you change the height (competition) on one side by adding or reducing weight (labor supply), there will have to be an offsetting change to the height (competition) on the other side. The height of the fulcrum point, representing overall labor market competition, does not change.

It is this balance precisely that H-1B critics miss.^[3]

Yes, immigration can undoubtedly tilt the scale by changing the composition of the labor force in one direction or another. Critics of H-1Bs want to tilt the scale to benefit native-born skilled workers. What they fail to understand is that you can only reduce competition for skilled workers by increasing it for everyone else, including the working class. The scale must balance.

College vs non-college

This is a misread of the evidence. As Claudia Goldin, Larry Katz, and David Autor argued in “Extending The Race Between Education and Technology,” over the last few decades the demand for college-educated workers has gone up even faster than the supply.^[4])

One clear piece of evidence for their claim is that the college wage premium — meaning how much more the average college worker earns than the average non-college worker — is near its historical high. Goldin and her coauthors estimate 62 percent of the growth in the college wage premium from 1979 through 2017 is due to a slowdown in the growth of the college-educated workforce.

Hardly a sign of too many college graduates.

Indeed, what this suggests is that in recent decades, college workers on the whole have faced an increasingly less competitive labor market than non-college workers. To the extent we want to tilt the labor market to make it tighter for one group of workers, it seems like workers with less education would be the reasonable choice.

The real wage inequality

In the early 1960s, only about 15 percent of workers in their 30s had four years of college or more. By 2025, the share had risen to nearly 50 percent.^[5]

Furthermore, for the last few decades, the most disproportionate gains in the economy have not been the top half leaving behind the bottom half. Cumulative median earnings growth since 1980 has been 33 percent for the bottom half of workers and 45 percent for the top half. This has even narrowed recently, with the ratio of median earnings for the top and bottom halves of the distribution falling from 2.4 in 2014 to 2.2 in April of 2025.

Instead, inequality in recent decades is really about the most highly paid pulling away from everyone else. Since 1980, the earnings of the median worker have climbed by 33 percent. The 80th percentile earner has enjoyed much bigger gains of 53 percent, while the 90th percentile worker has blown them all away with earnings climbing 78 percent.

It should be clear by now that college vs non-college is no longer the right comparison for understanding which classes of workers have reaped the biggest benefits of the American labor market in the last four decades. Instead, to the extent that the labor market has been very tight for one group of workers compared to the rest, it is those at the very top of the income and skill distribution who have benefitted the most.

Goldin, Katz, and Autor draw a similar conclusion, arguing that “most of the recent rise in wage inequality has occurred within, rather than between, education groups.”^[6])

It is for that reason specifically that policy should focus on tilting the labor market not just to increase more college workers of any type, but specifically the highest 10–20 percent of earners. The good news is that H-1Bs are doing just this.

Lies, damned lies, and badly mischaracterized statistics

Consider that the median annual wage in the United States was $49,500 in 2024. If I told you that someone making $200,000 was being paid a below-median wage, would you find my claim to be accurate? That worker is in fact making four times the actual median. Would it matter that the worker happens to be a software developer in San Jose, California, just because the median pay for someone in that job is $208,270 in that metro? Yes, technically someone making $200,000 a year would be below the median for that job in that place. Nevertheless it seems incredibly inaccurate, bordering on the absurd, to describe that person as “below-median.”

In fact, nowhere does the report actually state the plain fact that the typical wage of H-1B workers is $120,000^[8] — more than double the median for the country. This fact should furthermore not be surprising given that 57 percent of workers on an H-1B visa have a master’s degree or higher.

These are the kinds of mental acrobatics that people go through to mislead the public on this issue. H-1B workers are highly paid. Full stop. Anyone claiming otherwise is lying to you.

Who competes with H-1Bs?

To answer this question, my colleagues and I have created a dataset showing the occupations of H-1B workers in detail.^[9]

We find that 74 percent of H-1B workers are in occupations that are in the top decile by the salaries they pay. Another way of saying this is that roughly three out of every four H-1B workers are in the group of occupations that pay an average salary of $164,000 dollars per year.^[10]

The American workers competing with the vast majority of H-1Bs are in jobs that require the most skill and are the highest-paying. The workers in these jobs make up exactly the same group that has disproportionately benefited from growing wage inequality in the past few decades.

Another 19 percent of H-1Bs are in the 80th to 90th percentile, again competing with a highly skilled, highly paid group of American-born workers whose earnings have pulled away from the working class and median worker in recent decades. The average occupation in this group pays $93,000 per year.

All of which means that 93 percent of H-1Bs are competing with workers in the top fifth of earners in the U.S. economy.

Given an accurate understanding of how new workers affect the composition of the labor market — discussed above — it is evident that H-1B workers reduce labor market competition for American workers in the bottom three fifths — a group that includes the working class no matter how you choose to define it.

Perhaps even more striking is our finding that fewer than 0.1 percent of H-1B workers are in occupations where the average pay is below median. To repeat, less than one out of every thousand H-1B workers is in an occupation with an average salary below the nationwide median.

It is a bizarre form of populism that refuses to allow the bottom three fifths of workers, the entirety of the working class, to benefit from greater competition at the top.

The remarkable thing about this result is that it’s happening even though the H-1B is in fact a seriously flawed program. The program has hundreds of thousands more applicants than visas every year. To decide who gets one of the scarce visas every year, the system uses a lottery. This approach fails to ensure that the best and brightest applicants are chosen and that labor market competition is focused as much as possible on the highest-paid occupations.

Given these flaws, it is remarkable that the system still works as well as it does.

And the good news is these problems are fixable. As we at 91PORN have proposed, the system would be so much improved if it prioritized the workers with the highest expected lifetime incomes, which we calculate by using two simple pieces of information about the applicant: their age and the pay of their job offer. We then project their expected lifetime earnings using the average growth in pay by age for skilled workers in the American economy.

Changing to this approach would tilt the program meaningfully towards higher pay occupations.

Using data provided by Bloomberg, my colleagues and I simulate what would have happened if this approach had been used instead of a lottery for the years 2021–2024.^[11] Under the status quo lottery, 71 percent of initial H-1B applicants are in occupations that are in the top decile by earnings. In our preferred approach, which lets in those with the highest expected lifetime earnings, this share would rise to 82 percent, while the total share of H-1B applicants in the top fifth of highest earning occupations would rise from 94 percent to 97 percent.

The H-1B visa system, flawed though it is, overwhelmingly directs more competition to those workers in the American economy who have most benefitted from the rise in inequality. In doing so, it massively benefits the broad working class — and if we fix its flaws, those working-class gains will be larger still.

Positive Sum

But the benefits of high-skilled immigration are in fact much broader.

Tilting the labor market by changing the workforce composition is a relative effect, which means making some workers end up marginally better compared to others. That is a zero sum change. But high-skilled immigration also has overall effects, which mean more for everyone. That is a positive sum change.

I briefly discuss the overall effects of high-skilled immigration below. But for readers who want a deeper dive into the literature — including the detailed sources for all of the following data points and claims — please see our report, Exceptional By Design. The overall effects of high-skilled immigration include the following:

• Innovation: In the long-run, innovation is one of the biggest drivers of wage growth, and immigrants undoubtedly boost innovation levels. One study found that immigrants are responsible for 32 percent of the aggregate innovation in the US since 1990.

• Entrepreneurship and dynamism: A substantial body of research shows that through funding their own companies and making it easier for others to start companies, high-skilled immigrants help lead to the creation of new businesses that challenge large, old, incumbent firms.

• Industry growth: Throughout history, immigrants have helped found and grow industries throughout the United States. The nation’s lead in a variety of high tech sectors is fueled by the work, innovation, and entrepreneurship of high-skilled immigrants.

• Fiscal effects: Unsurprisingly given their high incomes, H-1B holders contribute more than $40,000 in net federal government revenues each year, which means that they deliver some combination of lower taxes, lower interest payments, or more spending for American-born natives.

Many types of immigration can have these positive overall effects. Without guest worker visas, for example, some4 rural parts of the country would lose substantial parts of their agricultural industries. Native-born workers in these communities would be worse off without the overall effects of industry growth. This overall effect can, and in many cases likely does, offset any negative effects on manual laborers of greater labor market competition.^[12]) Positive sum effects are one reason why low-skilled immigration doesn’t necessarily leave low-skilled American workers worse off despite more labor market competition for some of them.

For high-skilled immigrants, though, the overall effects are so overwhelmingly large that even the high-skilled natives who most closely compete with them almost certainly benefit in the long run.

Okay, what would really happen?

If the nation kicked out every H-1B worker, would the remaining workers in the same occupations get a raise? I think the right answer is to admit that such an outcome is possible — but this answer is just not even close to capturing the whole story of the H-1B program’s effects.

To understand why, it’s necessary to ask a slightly different question: If the H-1B program had never been enacted in 1990, would natives working in competing occupations today have higher wages? Almost certainly not, because of the overall effects.

Start with the 32 percent higher aggregate innovation from 1990 to today that, according to researchers, was generated by high-skilled immigration.^[13] Not only does that higher innovation benefit workers and consumers across the skill spectrum — everyone gains from new medicines and better airbags — but more than half of that innovation was the result of high-skilled immigrants boosting the innovation of their native-born collaborators.��A huge chunk of innovation from native-born workers since the 1990s, in other words, would not have occurred without skilled immigrants.

Native-born workers also benefit from the growth of companies and industries founded and powered by skilled immigrants. As Elon Musk has : “The reason I’m in America along with so many critical people who built SpaceX, Tesla, and hundreds of other companies that made America strong is because of H-1B.”

It’s true that you could make some native workers better off in the short-run by banning SpaceX and Tesla from hiring skilled immigrants. This is a point H-1B critics cling to. But those companies, and the supply chains that support them, only exist in the first place because of skilled immigrants.

These aren’t isolated examples. Silicon Valley and the American tech sector broadly would be shadows of themselves without high-skilled immigrants. If the best and brightest were prevented from moving to the United States, the many tech firms they have created either would not exist or would be located in some other country.

Whether they realize it or not, this is the world H-1B critics are pining for: America not as a center of innovation, technology, and world class research, but as a minor backwater and also-ran, lacking in opportunities for skilled workers both native and foreign. The critics would have us be North America’s France.^[14]

If we stop allowing in H-1B workers, that is what America would eventually become. The nation would lose out on all of the future innovations, startups, and industry growth that skilled immigrants would have created.

At best, economic suicide like this could only help skilled American-born workers in the very shortest run. Cutting off H-1Bs would be like taking our economy’s seed corn and feeding it to the top 10 percent highest-paid among us. It would be bad for the working class, and in the long run it would be bad for native-born skilled workers too.

What does high-skilled immigration do?

To be fair to the critics, not all of them want to end the H-1B program. Some merely want to reform it. The problem is not the desire for reform, which I share, but that their specific plans for reform so often reflect a completely backwards and highly restricted vision for how the skilled immigration system should work.

Some critics , for example, that we should allow in H-1Bs, but in such small numbers that they “fit in a bus, not a stadium.” Other critics argue that high-skilled immigrants only benefit us when they “address labor shortages,” and that the goal of policy should be to make sure that American workers are guaranteed to get first dibs at every job opening. An ideal system to these critics is one in which a high-skilled immigrant can be hired only once an employer definitely, certainly, and convincingly proves that “no American will do the job” and when government bureaucrats have verified a “shortage.”

This is a vision of immigration that attempts to micromanage away any competition for American workers. But it is completely disconnected from a true understanding of the actual benefits of high-skilled immigration.

We should not only allow competition, but welcome it. We should design H-1B visas to direct competition to where it is most needed. It is fine�����Ի���good actually if the top-earning skilled workers in the United States compete more. The top 10 percent of natives should not get first dibs on job openings as a birthright. This could only happen at the expense of the working class.

A system that focuses on shortages will deliver fake bureaucracy and limited benefits. A system that focuses on bringing in the highest-paid workers and welcomes competition will lead to more innovation, more entrepreneurship, more productivity, and greater wage growth for the working class.

If we ban H-1Bs entirely or try to shrink them down to fit into a bus, we will all be poorer — including the high-skilled native workers that H-1B critics are so eager to protect.

Notes

1. It’s true that some immigration critics accept that it is just like population growth but then insist even population growth is bad. The Center for Immigration Studies, for example, is unabashedly against not just immigration, but babies. Before founding the group, John Tanton was the president of a non-profit called Zero Population Growth that advocated for just that. He was frustrated that he could not convince the progressives in these groups to fight against immigration, so he cofunded CIS and other anti-immigration groups. Even today CIS will publish pieces that argue population growth makes us poorer. But there is really no argument here, the smarter conservative populists are correct and the anti-population doomers relying on ideas from the 1700s are wrong.
2. The definition will become clearer as I work through the example in this section, but by relative labor supply I mean how common one type of worker is relative to another type.
3. I am well aware, by the way, of the many other criticisms of H-1B visas. And in fact I agree with some of them, which is why 91PORN has created a sweeping proposal to overhaul the entire skilled immigration system. In this essay I am addressing the specific argument made by H-1B critics that American-born skilled workers should be shielded from competition. But I do discuss some of the other flaws in more detail later in this piece. Read on.
4. Autor, David, Claudia Goldin, and Lawrence F. Katz. “Extending the race between education and technology.” In AEA papers and proceedings, vol. 110, pp. 347-351, 2020. (
5. Source: Current Population Survey
6. Autor, David, Claudia Goldin, and Lawrence F. Katz. “Extending the race between education and technology.” In AEA papers and proceedings, vol. 110, pp. 347-351, 2020. (
7. Alexander Frei, “Rethinking the H-1B Visa Program: Data-Driven Looks at Structural Failures and the High-Skilled Illusion”, The Heritage Foundation, August 2025 (
8. USCIS, “Characteristics of H-1B Specialty Occupation Workers, Fiscal Year 2024 Annual Report to Congress” (). This represents pay from 2023, the most recent estimate available, and was likely even higher in 2024.
9. Specifically, we combine initial applications data from 2021-2023 and continuing applications from 2022 to 2024 to create a close proxy of the universe of H-1B workers who were in the U.S. in 2024 and still on their initial six years of the visa program. We do not need to adjust incomes because we are focused here on the occupational mix of this group. Occupational pay data instead comes from BLS OEWS data from 2024, and characterizes the overall U.S. labor market, not just H-1B workers.
10. That is the median annual pay of occupations in the top decile.
11. Whereas the above analysis uses our best estimate of the whole H-1B universe in 2024, in this analysis we utilize only workers who are initially applying to the H-1B program, because that is when the lottery occurs.
12. See for example Clemens, Michael A., and Ethan G. Lewis. “The effect of low-skill immigration restrictions on US firms and workers: Evidence from a randomized lottery,” No. w30589. National Bureau of Economic Research, 2022. (
13. Note this is not the only evidence for massive effects. See Exceptional By Design for a more thorough review of the literature.
14. None of the by revenue are in France, and nobody would mistake them as a in the global tech race.

The post Yes, high-skilled workers should have to compete appeared first on Economic Innovation Group.

Do H-1B visa holders earn more or less than Americans? There are two different ways to answer this. If we compare H-1B holders to the average native-born worker, the answer is unequivocal that the visa holder is paid more. Median H-1B pay in 2024 was $120,000 per , compared to $67,000 for the average native-born worker.^[1]

However, when we compare H-1B holders to otherwise similar native-born workers, the question becomes more complex. Answering it correctly requires careful examination. Unfortunately, a recent attempt at answering this question from Harvard economist George Borjas contains major errors.^[2]

In this analysis, we correct those errors and present a clearer picture. The answer to whether H-1Bs are underpaid depends on which workers we examine. Older H-1B workers earn less than otherwise similar native-born workers, while younger H-1B workers earn more. Overall, there is a small average wage gap, but it is only one-third as large as the flawed analysis claimed.

The paper not only reached flawed conclusions but also influenced the Trump administration’s $100,000 H-1B fee policy. Borjas, serving as a senior economist on the Council of Economic Advisors, advocated for an “optimal” fee based on “average employer wage savings” derived from this methodology. As we will argue, this policy solution is deeply mistaken.��

The H-1B fee is based on data errors

Borjas’s February 2026 attempted to answer whether H-1B workers earn less than comparable native-born workers by combining data on actual H-1B earnings with American Community Survey data on native workers. The conclusions are negative, with H-1B holders earning 16 percent less. But these findings result from substantial data errors.

In our replication of Borjas’s analysis, we identified several.^[3] The most significant mistake is a crucial temporal mismatch between his H-1B and native-born samples: the H-1B applications span 2020-2023, while the ACS data covers just 2023.^[4]

Nowhere did the paper mention controlling for inflation or wage growth. Given 15.1 percent inflation^[5] and an 18.7 percent wage increase for software occupations alone^[6] from 2020 to 2023, comparing wages of H-1B workers from 2020 to 2023 to a native-born wages from 2023 only produces negatively biased results that overstates the wage gap.

A simple approach is to directly compare the 2023 H-1B observations (FY 2024) to 2023 ACS data. Alternatively, we can use all years but adjust for inflation and convert all years to real 2023 dollars. Both approaches cut the wage gap roughly in half. We prefer direct comparison using only 2023 data, which also accounts for real wage growth beyond inflation.

The second error stems from controlling for geographic wage drivers using each worker’s PUMA (public use microdata area). PUMAs are Census Bureau geographic units corresponding to at least 100,000 people that do not cross state lines. The problem is that Dr. Borjas uses the PUMA where visa holders work alongside the PUMA where native workers live. Consider a native-born software developer working at Google in Mountain View who resides in a cheaper area like Fremont. If residential areas have lower average wages than business districts, this mismatch systematically inflates the apparent native wage and negatively biases the H-1B wage gap.

While Borjas notes this discrepancy in a footnote, he incorrectly claims the choice is necessary. The ACS already offers place-of-work PUMAs (POWPUMAs) for its respondents. Although POWPUMAs are larger and often contain multiple residential PUMAs, they more accurately reflect employment patterns than smaller residential units and allow apples-to-apples comparisons between H-1B holders and natives.

Using worksite state-POWPUMA combinations as geographic fixed effects, combined with restricting the sample to only 2023 data, further reduces the H-1B wage gap to just 5.3 percent — about a third of what Borjas found, without altering model specifications.

The much smaller wage gap obtained from data corrections alone casts doubt on Borjas’s other results. His robustness test to address tenure^[7] by applying the regression to male workers under 30 now yields an insignificant coefficient close to zero. Because his calculation of the optimal H-1B fee is based on average employer wage savings derived from his flawed wage gap estimation, our findings call the administration’s $100,000 fee into question.

A final data error is that Borjas filtered H-1B workers to for-profit private enterprises without applying the same criteria to native-born workers, causing him to overestimate the H-1B wage gap since native-born non-profit and government workers earn less.

The misuse of place

Another issue with the analysis is more conceptual. At 91PORN, we love place-based studies and policies. But different levels of geography require different types of analyses.

The Census Bureau delineates PUMAs based on roughly 100,000 inhabitants, but in urban areas, boundaries may contain just a handful of high-density neighborhoods. Controlling for PUMA does not just control for labor market wide differences in cost of living or agglomeration, it controls for a specific part of a labor market where someone works. This is especially problematic for Borjas’s use of geographies that correspond to a worker’s residence. If the most highly paid workers sort into more expensive areas within a city, this is not a difference you want to control for.��

A PUMA simply does not correspond to a labor market. Unlike what Borjas’s PUMA-level fixed effects suggest, a typical San Francisco worker does not wake up and decide to seek employment only in the residential PUMA containing SoMa, Mission, and Potrero Hill. They compete against similarly specialized workers across all seven residential PUMAs making up the place-of-work PUMA of San Francisco County and throughout the wider Bay Area.

A more ideal level of geographic controls would be larger geographies like core-based statistical areas (CBSAs) that incorporate labor market patterns around urban centers. This would more accurately model a software engineer living in Fremont competing for jobs in Berkeley, San Francisco, and Oakland simultaneously across the wider San Francisco–Oakland–Fremont metropolitan area. Unfortunately, the ACS has poor CBSA coverage.

Since H-1B workers congregate in the largest metro areas of their states, state-level fixed effects serve as a better proxy.

An additional benefit of state-level controls is it enables us to account for occupational sorting due to establishment location decisions by adding an interaction term between geography and the 19 Census occupation categories to the fixed effects. This allows us to get even closer to an apples-to-apples comparison — comparing workers doing a similar job in the same state.��

Our preferred model with education, age, state, occupation, and state-by-occupation controls, and correcting the data errors noted above, reduces the H-1B wage gap to 5.1 percent.��

While this still represents a non-zero difference in pay, the overall effect obscures significant underlying patterns worth examining.

Younger H-1B lottery winners outearn their American peers

Controlling for age in wage analyses is important because we do not want estimated differences in wages to reflect one sample being relatively older or younger than the other. However, the estimated overall wage gap obscures an important and policy-relevant pattern in wage gaps across age groups.��

We can see this pattern more clearly by including an age-by-H-1B status interaction in the model. This allows us to assess whether H-1B workers make more or less than natives of the same age. Younger H-1B workers make significantly more than natives while older H-1B workers make much less.

H-1B workers aged 22–23 fresh from college earn over 20 percent more than their native-born counterparts. Those finishing Optional Practical Training (OPT) or graduate degrees in their mid 20s to early 30s still earn more, and the negative wage gap only emerges around age 34. While 21-to-33-year-olds represent a minority of the overall labor force, H-1B holders skew younger due to program design: over two-thirds of lottery winners from FY2020 to FY2024 were 33 or younger. Nearly 70 percent of these young H-1B workers transitioned from an F-1 visa, including a substantial share of OPT participants. In contrast, only 10 percent of H-1Bs above age 33 had previously held student visas.

Two-thirds of H-1B workers are skilled young professionals, often with American education, whose companies pay premium wages exceeding those of comparable native-born workers. Their wage advantage, however, gets offset by the one-third of older H-1B workers who are more likely to arrive from abroad and work for at lower wages than native-born Americans.

Policy design lessons

While Dr. Borjas’s research contains data errors and conceptual issues, the policy implications present the most serious concerns. He argues that the existence of underpaid H-1Bs suggests tax revenue could be maximized via a fee of $100,000 or more.

This solution works only for employers underpaying their foreign workers. The logic is peculiar: identify a worrisome level of underpayment, then propose a policy that maximizes revenue from this practice rather than correcting it. Moreover, as we have demonstrated, the most underpaid workers are older visa holders, while the youngest earn a wage premium compared to native-born workers. The policy derived from Borjas’s analysis will penalize young workers and firms that pay them more, tilting the H-1B program toward employers hiring the oldest workers.

This is the exact opposite approach we would take if we wanted to preference workers more likely to be innovative and generate the largest economic spillovers. Even from a strictly fiscal standpoint, shows that the long-run fiscal impact of workers is highest for the youngest skilled workers. For immigrants with a bachelor’s degree, the 30-year fiscal impact of a 30-year-old is $1.6 million. The impact for a 40-year-old is $918,883, and for a fifty year old it is negative $91,517.

Dr. Borjas’s policy would benefit firms who are underpaying older workers with worse long-run fiscal impacts. It is a bad conclusion from a paper with a variety of serious errors. Reforming the H-1B program is essential. But utilizing an age-and-cost-of-living adjusted wage ranking as proposed by 91PORN here would result in a more positive economic impact for the U.S. economy and its workers.

View the Github with code for replicating this analysis .

Notes

1. Source: 2024 Current Population Survey Annual Social and Economic Supplements microdata.
2. “The H-1B Wage Gap, Visa Fees, and Employer Demand” https://www.nber.org/papers/w34793
3. Without direct access to Dr. Borjas’s code and data, we reconstructed his model from descriptions in the paper and successfully replicated his results. Our sample shows the same summary statistics, and we obtained a coefficient of -0.15 from an identical fixed-effect log wage regression, which is sufficiently close to his results. This indicates an H-1B holder on average earns 15 percent less than a comparable native-born worker with the same age, gender, education, occupation, and location. According to Borjas, the H-1B sample includes FY2021-2024 I-129 petitions obtained by Bloomberg via a Freedom of Information Act request and linked with FY2017-2024 Labor Condition Applications (LCAs) for new H-1B employment. Only lottery winners with approved H-1Bs are included. The native-born sample is drawn from the 2023 1-year American Community Survey (ACS). Both samples are limited to full-time, salaried, college-educated workers aged 21 to 50, with the H-1Bs further restricted to workers at for-profit firms. Those making under $30,000 annually are excluded from both samples, and H-1B workers making more than $1.22 million (top 0.05 percent) are also excluded. ACS wages are top-coded, eliminating the need for a ceiling.
4. In the data, H-1B observations are labeled according to the fiscal year. However, 2024 data really represent 2023 observations. Per , cap-subject I-129 petitions for a given fiscal year are filed in March of the preceding calendar year when the H-1B lottery opens. Those visa holders then are able to begin working in October of that year. For example, FY2024 applications were filed in March 2023 with wages at that moment in time. Those wages are then paid beginning October, 2023. FY2021-2024 applications thus contain nominal wages from 2020 to 2023.
5. Bureau of Economic Analysis, accessed through FRED.
6. Software and Web Developers, Programmers, and Testers (SOC 12-1250) includes the most common occupations within the H-1B sample, accounting for 42.5 percent of workers. Wage data is obtained from .
7. H-1Bs in the I-129 data are new hires who tend to earn less than long-tenured workers

The post The Flawed Paper Behind Trump’s $100,000 H-1B Fee appeared first on Economic Innovation Group.