The H-1B Wage Gap Is Real. The $100,000 Fee Is Still the Wrong Answer.
A wage ranking, not a flat fee, is how to fix the H-1B program and raise revenue at the same time.
Last fall, the Trump administration imposed a $100,000 fee on new H-1B visas with the stated goal of ending the negative effects of H-1B visa workers on the wages of American workers while raising substantial revenue for the federal government. From an economist’s perspective, the H-1B fee could be thought of as a price to ration a limited number of visas but the fact is that the policy is counterproductive to both goals of the Trump administration, it will actually lower wages for native-born Americans in the long run and reduce federal revenues. If the Trump administration seeks to select truly high-skilled immigrants and root out the fraud and abuse of the H-1B visa program, it should instead select H-1B visa recipients purely based on their age-adjusted wages, rather than charge a fee.
What is the H-1B program
In short, the H-1B program allows employers to sponsor a college-educated immigrant to work for up to 3 years and renewable for another 3 years and also to sponsor them for permanent residency if they qualify while they are on this visa. Private for-profit employers can sponsor up to 65,000 H-1B workers, with an additional 20,000 visas for immigrants with U.S. advanced degrees, for a total of 85,000 new visas every year. Non-profit and government employers are allowed to sponsor an unlimited number of H-1B visas and they tend to sponsor some 50,000 annually, mostly consisting of universities sponsoring professors and non-profit hospitals sponsoring physicians and researchers.
Given that there is an 85,000 cap on the number of H-1B visas and hundreds of thousands of workers are requested by U.S. companies, the federal government conducts a lottery to assign visas to those who apply because the Immigration and Nationality Act states explicitly that slots shall be awarded in the order of application. The lottery structure is ripe for abuse, as many immigrants and employers abused the system by submitting duplicate applications to increase their selection chances and the lack of wage requirement allowed some specifically Indian companies to create a business model where they essentially only hire H-1B workers and their applications are all essentially interchangeable with one another.
The first Trump administration improved the H-1B program by simply switching the lottery order between the 65,000 regular visas and the reserved 20,000 U.S. advanced degree visas. Until 2019, the H-1B lottery first selected 20,000 applicants among those with a U.S. masters degree and then 65,000 among everyone else, including those not selected in the first round. Beginning in 2019, the Trump administration switched the order and increased the chances of immigrants with a U.S. advanced degree.
For instance, if employers were seeking 200,000 H-1B workers with a bachelor’s degree or a foreign advanced degree, and 100,000 workers with a U.S. advanced degree, the two systems would yield the following selection rates:
Pre-2019: 20,000 out of the 100,000 U.S. advanced degree applicants selected, then 65,000 out of the remaining 280,000 selected, for a selection rate of 23.2% for applicants without a U.S. advanced degree and 38.6% for those with a U.S. advanced degree (20,000+80,000*23.2%=38,571).
Post-2019: 65,000 out of the 300,000 applicants are selected, then 20,000 for the remaining applicants with a U.S. advanced degree not selected in the first round, lowering the general selection rate to 21.7% and increasing it for those with U.S. advanced degrees to 41.7% in this hypothetical example (100,000*21.7%+20,000=41,667).
Then, to deal with the problem of multiple applications, the Biden administration further reformed the H-1B program by ensuring only one application per immigrant counts for the lottery, reducing the number of total applications by over 50% in two years by ending the practice of multiple registrations, thus, increasing the selection rate of legitimate applicants.
Where does the $100,000 fee come from
The main justification for the $100,000 comes from a recent paper by my esteemed colleague and dissertation advisor Dr. George Borjas, who finds that H-1B workers earn about 16 percent less than comparable Americans after adjusting for education, occupation, and location. Other researchers find a smaller gap, but the basic finding holds. From there, Dr. Borjas argues that employers should be willing to pay a fee equal to the wage surplus they capture from each H-1B worker over 6 years, yielding a revenue-maximizing fee of somewhere between $97,000 and $154,000 for the available 85,000 H-1B visas every year.
Reinterpreting the H-1B Wage Gap
The description of the H-1B wage gap is correct. However, the interpretation of the wage gap as employers’ willingness to pay exaggerates the revenue-maximizing amount because it misses some key aspects of how the H-1B visa program works. Most importantly, even when correctly calculated, the H-1B fee is suboptimal relative to a wage ranking because it allows the firms that most underpay to sponsor workers while shutting out the highest paid immigrants who pay most in taxes, thus a fee is fails to achieve the stated goals even when correctly measuring willingness to pay.
Adjusting the revenue-maximizing H-1B fee
After accounting for the fact that H-1B workers switch jobs and the wage gap shrinks over time as employers, the Trump administration’s $100,000 is already above the most accurate of the Borjas estimates: $97,000. Why pick $97,000 over the high-end $154,000? Because the lower estimate is based on H-1B visa demand similar to the one we are experiencing now rather than the high levels of the first year after the COVID-19 pandemic when tech firms were hiring en masse.
Furthermore, employers already have to pay a multitude of fees to sponsor an H-1B worker. Large employers pay a $215 electronic registration fee to participate in the lottery, a $780 fee for form I-129 if they are selected, $1,500 for a fund to train American workers, $500 for fraud prevention, $600 to fund Biden’s asylum disaster, and it is common practice to pay a $2,965 premium processing fee to guarantee a decision within 15 business days. Add to this, approximately $5,000 in lawyer’s fees and the cost to sponsor an H-1B worker absent the Trump fee is already close to $12,000. This lowers the additional revenue maximizing fee estimated by Dr. Borjas from $97,000 to $85,000.
While there is a turnover rate that Borjas accounts for in his thorough paper, this is just the low-end of the real H-1B turnover. Most non-Indian H-1B workers become U.S. permanent residents before the visa is over, without the need to switch jobs and thus are not on an H-1B visa but do remain with their employer and obtain full labor freedom by virtue of being permanent residents. Therefore, the actual revenue maximizing fee is much lower than the updated $85,000.
Problems caused by the H-1B fee
Because the H-1B fee is a one-time and equal fee for all employers, it creates several problems that supporters should be worried about:
Free rider problem: Immigrants can be sponsored by a company that pays a large sum of $100,000 and then switch jobs at no cost to other firms for no cost. The consequence is that firms will be unlikely to sponsor immigrants who other firms may seek to poach, which will likely be the most skilled, entrepreneurial and innovative. This risk further lowers the revenue-maximizing rate, by the way.
H-1B abusers benefit: Dr. Borjas shows that large Indian outsourcing firms underpay their H-1B workers more so than any other firm. These companies are: Infosys, Tata Consultancy Services, Cognizant Technology Solutions, Wipro, Capgemini, HLC, Tech Mahindra, and Larsen & Toubro Infotech. Thus, because the H-1B fee is equal for all and will be paid by employers who can reap the most benefits from hiring H-1B workers, it is these Indian tech firms that will be willing to pay the visas, not the most innovative firms with lower wage gaps. In fact, American companies such as Amazon, Google, Intel, Walmart, Qualcom, and McKinsey & Co. actually overpay their H-1B workers relative to similar Americans.
Cap-exempt visa issuance reduction: Because the H-1B visa fee affects cap-exempt petitions, it will reduce visa issuance outside the rationed 85,000 slots, this will reduce tax revenue paid by these immigrants, each of which I have estimated to reduce the deficit by $2.3 million over 30 years.
Practical inapplicability: Finally, the great problem of the H-1B fee is that, in practice, it can only apply to applicants abroad, and thus it can be skirted by the most problematic firms with the worst wage-gaps: The Indian outsourcing firms. The end result is that startup firms are shut out of bringing top talent while abusers receive a free ride. These firms can skirt the fee by transferring their employees from affiliates in India on L-1 intra-company transfer visas to the U.S. and applying for the H-1B lottery inside the country. L-1 visas are unlimited and not subject to any fee.
The Right Answer: Age-Adjusted Wage Ranking
If the goal is to allocate visas to the most productive workers, raise federal revenue, and stop the worst abusers of the program, the cleanest tool is the wage itself. Replace the lottery with an age-adjusted wage ranking. Allocate the 85,000 visas to the highest-paid applicants, weighted to favor younger workers who will pay taxes for more years.
This policy solution dominates a fee on every dimension that matters. The Trump administration is already in process of enacting a modified wage ranking (proving that they believe they have the legal authority to do so, though I doubt it).
An ideal wage-ranking would select workers based on their wage relative to those in their age group. This age-adjustment recognizes the reality that a 25-year-old with a $200,000 salary is more impressive than a 45-year-old and that his or her younger age means more years left to contribute to the American treasury and economy. How it would work is simple: Classify workers into age ranges and then pick the top percentiles based on demand. If demand is twice the number of visas, then pick those above the 50th percentile of wage offers for their age range. If demand is three-times the 85,000 H-1B visa cap then pick those above the 67th percentile, and so on. The law reserves 20,000 visas to those with U.S. advanced degrees so two such rankings would take place.
The benefits of this approach are threefold:
Efficiency and Simplicity: A wage ranking selects workers whose skills command the highest market value and observes characteristics that cannot be cheated or changed: Age and salary. If companies pay less than what the salary offer selected later, they would be in violation of the law under current regulations already.
Fiscal impact. Higher-wage workers pay more in federal taxes, year after year, for the duration of their working lives. The Institute for Progress estimates that a pure wage ranking would raise the average salary of new H-1B holders by about 50 percent. Adjusting for age compounds the gains by increasing the number of years of tax contributions.
No fraud: A wage ranking automatically excludes the firms whose business model depends on paying H-1B workers below the market rate. The wage rule does what the fee promises but cannot deliver: it ends abuse of the H-1B program.
However, the Trump administration proposed an abuse-prone “weighted wage ranking” rule to modify the H-1B program which gives more “lottery tickets” to applicants with a higher prevailing wage. While good in theory, the problem lies in the details of how prevailing wage levels are defined. These are based on both location and occupation, which companies can and will game to get their workers a higher selection chance. Furthermore, it assumes that a lower-paid occupation is just as valuable for the market than a higher-paid occupation. Occupations are arbitrarily defined and the government shouldn’t be in the business of privileging low-paid occupations but allowing firms to sponsor the most productive and innovative immigrants.
The administration is right that the H-1B program is broken, that firms capture a wage gap that ought to be eliminated, and that the federal government could raise serious revenue from a program with this much excess demand. The diagnosis is right but the $100,000 fee is the wrong response.
There is a simpler, more powerful, and already-available alternative that doesn’t require a bureaucrat deciding which occupations are to be favored or how they are defined, an alternative that roots out the corruption and fraud because it is objective: Allocate visas to the highest-paid applicants, adjust for age, and welcome talented people who can contribute to making America more prosperous.



Great piece. But I'd argue that what's important for an age adjustment is the expected lifetime contribution, not where they compare to the other people of the same age.
Bringing in a 95th percentile 64 year old right before retirement is much worse for America than getting an 90th percentile 30 year old who still has most of their career ahead of them.
EIG has an age-adjustment proposal that I think gets a much better fiscal impact: https://eig.org/eig-letter-dhs-should-revise-proposed-h-1b-weighted-lottery-to-prioritize-top-talent/
I had stopped caring about the $100,000 fee from a policy perspective (legally, it’s an atrocity and I’m surprised by the success the administration has had in court thus far) because the H1B visa is capped regardless, and it did not look as though it would cause the annual inflow to dip below the 85,000 annual cap. I had forgotten about the uncapped H1B categories!
Really enjoying these posts. Keep up the good work!