Trump’s H-1B lottery policy… kind of works?
Clumsy, imprecise, but directionally right. Let's take our wins where we can get them
In September the Trump administration announced, without warning, that a new $100,000 fee would apply to H-1B petitions, effective immediately1. Employers panicked.
The rollout was a mess. The proclamation dropped late on a Friday afternoon, and different government agencies scrambled to issue different interpretations of the rule (sometimes by tweet). It was initially unclear whether the fee would apply retroactively to everyone currently on H-1B, or only to new applicants. Many employers instructed their employees not to leave the country.
To some extent, the confusion is part of the point2. Trump wants to discourage people from coming to the U.S., and uncertainty achieves this goal.
But the Trump Administration also had a straightforward and uncontroversial goal here: To fix widely-acknowledged issues with the H-1B program3. In this - perhaps surprisingly - Trump also succeeded.
Problems with the H-1B
The H-1B has several extremely well-attested problems, which frankly previous administrations (including Trump 1.0) have done next to nothing about. Two major problems:
Outsourcing exploitation: The structure of the program is widely seen as easy for large outsourcing/staffing companies to game (e.g., by flooding registrations). This shifts the program away from “best-and-brightest” hiring and toward labor-arbitrage staffing models that undercut wages for Americans. In some years, something like 20% of the H-1B lottery slots go to these outsourcing firms.
Wage protections don’t work as intended: In theory, the H-1B requires employers to pay enough that hiring H-1B workers shouldn’t undercut comparable U.S. workers. But in prevailing-wage rules, job classification choices, location/wage-level selection, and the worker’s limited ability to switch employers can still allow effective underpayment and weaker bargaining power—particularly in the outsourcing-heavy segment of the program.
Trump decided to take a sledgehammer to these problems ahead of this year’s lottery.
Trump’s fixes: $100k fee and wage-weighting
The Trump administration made two key changes to the H-1B lottery this year:
The $100,000 fee. This is a prohibitively high fee for most employers, but it only applies to consular petitions. Roughly speaking this means:
It applies to people who are outside the U.S. and want to use the H-1B as their initial visa to get into the U.S. Importantly: This is the category that most outsourcing firms fall into.
It does not apply to people who are already in the U.S. (e.g., on a student visa).
A weighted lottery that favors higher-wage jobs. Historically, every entrant into the lottery had an equal chance of getting selected4. This was the first year of a kind of wage-based weighting, where those who are earning higher wages have better odds of getting selected.
Both of these changes are conceptually pretty ugly.
As mentioned above, the $100k fee rollout was confusing. It’s not clear the $100k fee is even legal5. And it serves as essentially a blanket ban on H-1Bs for people not already in the US, which is an incredibly imprecise way to target outsourcing companies.
Wage-weighting sounds great in theory6. But the details matter a lot, and Trump’s immigration policy has not been very detail-oriented, to say the least. The actual method used to implement wage-weighting actually advantages outsourcing firms7.
The impact: So crazy it actually works?
So the changes are undoubtedly clumsy and imprecise policy. But the translation of economics into policy is always clumsy and imprecise. The question is, does the policy move us towards our desired goals?
We won’t get the full answer until national H-1B lottery data is released later this year. But the initial findings:
There seem to have been roughly 30% fewer lottery entries than previous years. Notably, that’s close to the share of cap-subject filings that historically go through consular processing. Essentially it seems like employers (like outsourcers) who wanted to bring people directly in from outside the US simply gave up because they didn’t want to pay the $100k fee. The fee filtered out roughly who it was supposed to.
Among applicants already in the US, the wage weighting meaningfully shifted lottery selections toward higher-paying companies. At Lighthouse our clients saw 50% selection rates in the H-1B lottery - and as high as 67% for the highest-paying companies. Historically, odds have been closer to 30%. The most innovative tech companies benefited from the changes to the lottery.
The net effect: The $100K fee that blocked out outsourcing firms. This is a major positive result. And as a secondary benefit, wage-weighting the lottery shifted selections towards the highest-value workers.
Seems pretty good, actually! And it’s worth re-iterating that over the past 20 years, Republican and Democratic congresses, and the administrations of Obama, Trump 1.0, and Biden all failed to move the needle significantly on the improving H-1B lottery. This time around Trump took a swing and hit, if not a home run, at least a single.
Real negative impact: Universities and hospitals
The positive impact is offset to some extent by negative impact on universities and research hospitals.
These employers are typically exempt from the lottery, able to apply for H-1B visas for any qualified candidate at any time of year. The new $100k fee applies to these employers, exactly as it does to for-profit employers.
This means the hospital at the University of Iowa would have to shell out $100k if they want to hire a pediatric oncologist from Toronto8. Rural hospitals already can’t staff their wards, so restricting the supply of international medical talent is a totally unforced error9. And do we really want to tell universities that they can’t hire the best researchers in the world?
As it happens, the $100k fee proclamation actually offers a provision that seemingly would address the university and research hospital issue: The Secretary of Homeland Security has discretion to waive the $100k fee for workers, companies, or industries deemed to be in the national interest. As of now, it seems no such waivers have been issued.
What’s next?
More changes are yet to come. The $100k fee has been set to only be in effect for one year, and government departments are considering various proposals for more permanent solutions.
One change being considered by the Department of Labor would replace the wage-level system with experience benchmarking. This would, in IFP’s analysis, “permanently end employers’ ability to get visas for foreign workers who they will pay less than similarly qualified Americans.” This would be a major step forward.
But this year was a decent start all things considered. We can’t let the perfect be the enemy of the good. In trying to improve the H-1B lottery, the Trump administration moved fast, broke things, and — on net — improved how it works. We can build from here.
—
Thanks to Alexander Kustov and Jeremy Neufeld for comments on an initial draft. All opinions and mistakes my own, not my employer’s.
Technically it was announced on September 19 (a Friday), to go into effect that Sunday. Not a fun Friday evening for the Lighthouse team and our customers.
To summarize, for those blessedly not in the weeds of the H-1B visa program:
The H-1B is the most common immigration pathway for people who go to undergrad in the U.S. and then transition to full-time work. A lot of my friends who went to MIT were/are on H-1Bs.
The H-1B is a U.S. temporary work visa that lets employers hire foreign workers. The H-1B is capped at 85,000 visas per year, and there are always 300,000 or so people who want to get H-1Bs. The way that they determine how to allocate these 85,000 visa slots among 300,000 interested applicants is via lottery which happens each year in March.
With an exception for those who got master’s degrees inside the U.S., who historically had slightly higher odds.
It’s technically structured as a “restriction on entry” rather than a fee, since its implementation did not follow the proper process that would be required to implement a new fee. It is being currently challenged in the courts, though initial rulings have upheld the fee so far.
Once you resign yourself to the fact that we are using a lottery (random!) in the first place. Sigh.
Jeremy Neufeld has written all the detail you need here. Since the $100k fee screened out the outsourcing firms this year, wage-weighting had a net positive effect. But the effective of wage-weighting probably flips negative if you don’t have the $100k fee.
This is important because the $100k fee is set to expire in September. So if no further action is taken, next year’s lottery could be a huge step backwards.
In fairness, some of these jobs getting filled from overseas are relatively low-paying lab technician jobs. Experience benchmarking would solve this problem.
In case you haven’t noticed, we are nowhere near equilibrium here, and these rules are still being challenged. The healthcare sector is making the loudest noise. The American Healthcare Association is behind a bill would exempt physicians, nurses, and other healthcare workers from the fee outright.