Wonderful analysis, with which I find myself in total agreement. Since I did immigrate into the US based on my skills in high tech, I thought sharing my own personal experience might be a useful contribution to the discussion, too, so, here it is.
Around the mid-90s, as my first marriage (in my native country, Italy) failed, and my high-tech career was flourishing, I did consider moving to the US (where I was starting to have romantic relationships based on long-distance net connections, AND I had worked on and off based on employment on US firms with Italian subsidiaries -- Texas Instruments and IBM -- so I knew I'd really like the overall "vibe"... not to mention the detail that I knew my compensation, as an employee with high tech skills, would at least DOUBLE compared with Italy's meager ones).
However... I studied the possibilities available back then (30+ years ago) and decided H1B was way too restrictive and hard to get in, so I chose another career path. I wrote a couple books about the emerging tech I had luckily adopted (the Python programming language: a totally niche thing back then, THE most popular programming language in the world today -- lucky me for being on the best bandwagon SO early on!-).
My books became best-sellers, and got me great job offers from early-adopting companies all over the world -- and, together with my reputation as a very helpful poster to usenet's comp.lang.python, eventually got me prestigious awards from the Python community, too.
I eventually chose to drop out of the Italian employee treadmill and risk myself into freelance consulting in tech and tech-management matters -- quite successfully, but at the cost of flying all over the world to supply my services to various clients while still living in Italy... spending more time in airplanes, airports, and hotels, than I did at either my own office, or clients' ones, was a truly stressful life... but it did keep earning me kudos and recognition...
Eventually one of my US romantic connections gelled into perfect intensity -- my wonderful now-wife, US-citizen and veteran, Anna, moved to Italy to live with me (our connection also reinforced by fiery shared interest in this "Python" thingy...), we married, we started the uphill fight to get her a work and residence visa in Italy (even harder than the US makes it, believe it or not -- I wouldn't believe that, myself, if I hadn't experienced both!), she became my co-author on some of my best-selling tech books...
Eventually one of the companies I was offering my consultancy services to, a startup with a weird name, having put me to the wringer of more and harder interviews than any previous potential clients ever had, ended up offering me a tech manager job ("we'd rather not employ consultants for technologies crucial to the company" -- Python, in my case)... if I was to move to the US (or maybe to their nascent offices in Dublin or Zurich, which, visa-wise, would have been a no-brainer, BUT... I had previous experience with being at a nascent foreign office, and had decided that being at "the mothership" was really the only good choice for me).
On checking with the immigration lawyers they employed, it emerged that I probably qualified for an O-1 visa, "for the individual who possesses extraordinary ability in the sciences" &c, given the amount of awards, tech best-selling books, and likely endorsements from previous employers and clients, that I had accumulated by then.
So I ended up immigrating to the US in 2005 (after that startup's IPO) -- my offer had been extended in 2004, but visa proceedings (e.g., me asking so many people for endorsements) took too long for me to make it any earlier than early '05.
It all worked out very well in the end -- "I like to be in America", as Sondheim's wonderful lyrics to Bernstein's wonderful song "America" (in "West Side Story") say -- that startup (with only modest help from my own work) flourished, I naturalized as a US citizen, I contributed a fortune or two in taxes to the US and California over a couple of decades -- I'm now retired, but SO like it here that my wife and I still live in Silicon Valley (and visit Italy often rather than moving back there, though I do miss it and my many relatives when I'm here in the US).
BTW, that startup's weird name was... "google". Weird indeed, right? But, it DID work!
So if the US immigration system eventually worked so well for me, why do I so fully agree with the ideas in this post? Think about it -- with a saner visa system for high-skill tech-professional immigrants, I might have moved to the US maybe a decade earlier -- 10 more years of high earnings and high taxes (income, sales, SS contributions, Medicare ones...) paid by me (and my employers) to Federal and State systems -- to substantial mutual economic benefits for me, for California, AND for the US.
When I did move, I was about 50 years old -- so, just a couple decades of big tax contributions (then retirement, when I get some money back from SS and Medicare -- although my savings, productively invested in the US, yield enough that my Federal and State taxes are still pretty high). Had a sane, welcoming immigration system induced me to immigrate at 40 (as I considered back then, but rejected on deeply examining that system), that would have increased the time of my highest contributions to the US and State economies (and tax revenues) by fully 50% -- far from chicken feed...
Canada just had it's reckoning with diploma mills. A few schools had grown 10x with "education" programs that involved a couple optional lectures every week, but came with a 20 hour/week work visa. 100's of thousands of "students" came in through these programs and never learned a thing. They locked it down last year, and Canada's population actually shrank last year because of it. Hopefully going forward the student work visa program will help highly qualified students earn a degree, and then they can stick around :)
Just switch to a point system, and make things like wage potential, attended an American university (maybe extra points for masters/phds), speaks english, lived/worked in the US before and didn't break any laws (maybe under a temporary visa), etc give extra points.
At one point I had tried to design an ideal point structure and had come around to this:
wage offer / earnings potential: 25
education: 15
US degree: 10
English: 10
age / work horizon: 10
prior lawful US work and status compliance: 15
shortage occupation / national interest: 10
assimilation / civics / clean record: 5
Some would be a bit hard to find objective measures for, but overall seems better than a lottery!
The problem is that most Americans are at least as dumb on the topic of immigration as they are on housing, if not more.
The average voter's reaction to the mention of low-wage immigration would be to worry that they're leeching off the system by sucking up welfare benefits. The reality is the opposite: any immigrant working legitimately pays taxes and is not eligible for major welfare benefits. The typical solution to the voters' imagined problem is to make it harder to get a work permit, which just encourages under-the-table work and reduces tax intake.
When it comes to H1B, the prevailing negative sentiment about the program comes from right-wing tech bros on Reddit, emboldened by Silicon Valley CEOs' recent rightward turn to say what they really think. Of course, by saying that, they reveal the truth: 1. they are bigots, and 2. they're not being hired because of their own innumeracy, not because of H1B candidates who make up just a few percent of the overall tech workforce.
The only way to fix this is to sneak reform through Secret Congress style and force voters to live in the better reality until they move on to the next fake concern. There is no way to convince people of the truth on this topic.
Many Americans get confused when they hear immigrants are on "benefits." In immigration law, "benefits" just means they are awarded the right to be in the US.
> But actually it’s good! Adam Ozimek and Sarah Eckhardt show that the median H-1B worker earns $130,075 per year (the 91st percentile nationally), meaning that the average H-1B household makes a net federal fiscal contribution that’s 2.6 times as much as the typical American household’s. This is a big win for the federal budget
The program's intention is that companies only hire H-1B if an American isn't available IIUC, and here you've implicitly assumed it works as intended. But my impression is that much of the anti-H1B rhetoric says it doesn't work as intended, so you'd best argue that point specifically. Also, by leaving it implicit, anti-H1B people will think you amazingly didn't comprehend that An American Could Have Had That Job! (What a *****!)
Hoped to hear more of a steelman here. The way it's been argued, a $100k fee is a strong selection mechanism to only select for the highest of the high skilled. Not only in terms of wages alone, but also as a form of immigration loophole. India gets mentioned and then not expounded upon - it's my understanding there really is an aspect of nativist/racial resentment fueling this, with some even making allegations that it's a caste-based form of chain migration (this almost certainly overstates the case). You see the same thing with a China-hawk flavour too, especially concerning employment for graduates, where vague fears over national security are used to justify xenophobia. Not entirely unjustified, of course, China does historically play fast and loose with IP and espionage. But it's difficult to make a meritocratic-flavoured argument while throwing shade at this cohort of academic overachievers (who are also very lucrative for universities, so a double-dip fiscal win for us). I guess this is in line with Matt's general "use race-neutral framing whenever possible" popularism ethos.
Plus the fee is, of course, just straight-up upfront revenue for the government, taken from unsympathetic groups, the same way one can view tariffs as free yuge revenue with enough blinders on. Business As Usual from this admin, not that one should encourage it, but it's very understandable.
I mean... why isn't it "free [not terribly] yuge [but still not nothing] revenue"? The same way datacenter builders pay off the locals to compensate them for the imagined hardship of living near a datacenter, we should be able to pay off the losers who experience a taco truck on every corner as a personal hardship. This becomes much harder for asylum immigration, family sponsorship, etc. - for H1Bs the arrangement is just too perfect.
But this is not compelling either - are you claiming that academics are paid at exactly the cost that universities can absorb to employ them, such that paying them any more would make the transaction nonviable? Do we really not believe that at least some hiring universities might be able to absorb at least some of the cost of this fee? Hard to square that story with the story told 8 billion times over the past 10 years of an academic job market that is absolutely swamped with candidates, and universities able to sign them up for indentured servitude as adjuncts despite the substantial value they create for the university. Furthermore it's already substantially more expensive to hire an H1B holder for legal reasons; my impression is that the legal fees are on the same order of magnitude as this $100k Trump fee. If the fee clears out the lottery so that employers aren't spending tens of thousands a year on lawyers to fight for the visaholders to stick around, this could net out to neutral.
What a desert of policy analysis debate we live in these days.
On the one side, here's a long-standing policy which despite some flaws is net good and here are some ways to improve it.
On the other hand, here's Trump who destroys everything he touches and everything just gets bad and the US looks into the abyss of becoming dysfunctional.
I guess sure we can talk about improving programs in the hopes of a Democratic trifecta in 2028 but it's hard to have one's heart in it.
Often lost in this debate is that there are well-vetted solutions to the main problems in the H-1B and OPT programs waiting to be picked up. People don't want you to know this!
Worried that employers are paying visa holders less than similarly qualified Americans? We could just fix that by requiring employers not pay less than the median American in the same occupation with the same combination of education and experience.
Worried that an H-1B lottery advantages employers hiring lower-skilled workers instead of using scarce visas on the highest-skilled, highest-paid workers? We could just fix that by allocating visas according to compensation.
Worried that OPT participants' exemption from FICA taxes is unfair to American workers? There is a Republican-led bill in Congress that could just fix this for good.
"Worried that employers are paying visa holders less than similarly qualified Americans? We could just fix that by requiring employers not pay less than the median American in the same occupation with the same combination of education and experience."
You could also just make it easy for visa holder to job switch and this would go away on its own
"We could just fix that by requiring employers not pay less than the median American in the same occupation with the same combination of education and experience."
Why median if the goal is to import talent that can't be found in the US than we'd want 90th percentile, right? The other huge issue is "median salary" is very gameable. A software developer can make $100k or $10 million with the same generic corporate title. Are they working at Anthropic or supporting the billing department of a midwestern auto insurer?
The $100k fee is just designed to be punitive, not to make money. If you charged $30k a year, for example, I think that would be fine! A lot of tech companies would just pay that! That's about how much you pay a recruiter for a new position anyway. And since an H1B can last for 5 or more years, the government could actually make *more* money that way.
The $100k fee is hard to justify, even for a worker who is going to be worth more than that, because you have to pay it all up front. $100k for someone that could just switch to a new job a few months later.
I think the best reform would be to eliminate the sponsorship requirement of H1-B.
Most of the resentment I’ve seen comes from the belief that companies prefer H1-B workers because of they have a unique level of control over them. This seems like yet another place where normal liberal sentiments and good policy align.
The reason this is a problem is only because of the per country quota in EB green cards. I got my green card in less than 5 years but it takes decades now if you’re from India and 5+ years if you’re from China. Fixing the green card backlog fixes this issue. There are other simple fixes like if the PERM and I-140 are approved, then you don’t need to file them again when you switch jobs. It’s possible to switch jobs easily on H-1B, at least prior to the Trump administration, but it involves redoing step 1 and 2 of the green card application. Once you get to step 3, you’re relatively safe.
Yes, this is such a widespread misconception. H-1B workers can absolutely change jobs. The employee leverage comes when the employee sponsors the green card application and you wait in the queue. Which, as you described, is a very solvable problem.
Importing skilled workers is a little too convenient compared to hiring Americans for several reasons, but the most important is sociological. There are in fact huge numbers of smart, talented, highly educated people in India. But hiring a room full of these people in California gets you bizarrely WAY less shit than hiring a room full of high SAT math score Americans.
This works even though they are a similarly privileged strata of their own (much more unequal) society, and even though this in no way benefits the descendants of American slavery who are owed representation in the logic of DEI.
It’s not surprising that this arbitrage would eventually be closed.
A room of high SAT Math scoring Americans would also mostly be Asian? Don’t think there’s that much of a difference.
Also, privilege hierarchies are global. Their society as a whole is less privileged than ours so they are lower in the global hierarchy. Someone who is the chief of an African village but still lives in a mud hut with no education is clearly less privileged than an upper-middle class American growing up in a McMansion going to an excellent public school even if the chief is higher in his own society.
Much of the problem with both the underlying diagnosis and the proposed fixes stems from MAGA’s seeming insistence that there is thing called a “job,” and it functions like a chair in an auditorium that either an American or a non-American sits in. But a job is much more like an opportunity to *build* a chair that then can get purchased by some auditorium somewhere, and eventually get sat in. Sometimes, the person who build the chair can even get that seat!
The first objection, around the $100k fee being too steep, is not convincing, because there’s no argument made that the fees can’t be financed. If a ~$200k payment over the next 10 years of the worker is really too risky a proposition for a lender, then what are we doing here? I suspect that there are similar just-one-step-on-the-chessboard arguments being made here.
An H1B is a 3 year visa. An OPT is a 1 year one. It's not a green card. At those time frames lenders would balk at that.
The usual pipeline is OPT after college > H1B > H1B Renewal > Green card application. Essentially American businesses have a 'try them out before you buy them' immigration system. But there's many things that can go wrong there to make a bank feel comfortable with that arrangement.
But those timelines aren't relevant - the question is what's the likelihood of the visa-holder transitioning into a green card, not whether they can pay back the loan in two years. Banks are not stupid, and the same way they will lend doctors money against their future salaries while they're in residency to buy a house, they will lend an AI PhD who hasn't even gotten a job yet against their future salary to pay a loan that will be a relatively small fraction of their first year salary.
The fact that each step has some likelihood of sending them back is a more interesting objection - but again, the premise here is that we're solving the problem of having a bunch of workers who we don't care to actually offer citizenship to, right? In which case, that likelihood drops to much lower levels.
I am pro-immigration. But I am anti the deterioration in the quality of analysis that I’m getting as a SlowBoring subscriber, and I would like Matt to address obvious points like this one if he’s going to bother critiquing a policy at all.
It’s illegal. The law specifically bans the government from profiting from immigration applications. The fee can only cover the cost of employing federal immigration workers and the cost of processing. The Congress is welcome to change the law but right now, it’s illegal.
I know academics who, while still first year PhD students, took out larger student loans. Seems like, all else equal, a bank willing to lend to the PhD student would be willing to lend to the employed academic.
All else might not be equal, but that's the discussion to have, not this caveman analysis "we have this arrangement that will generate millions of dollars of economic surplus but we can't find $100k to unlock it!"
Personally, the way I prefer it is we simply make H1B holders pay a flat surcharge until they're citizens. We have a bunch of loser Americans who experience having taco trucks on every corner as a hardship; we have a bunch of economically productive entities who would gobble up the purveyors of taco-truck-or-equivalent culture; it is simple to compensate the loser Americans for their pain and suffering and make it a win-win. Charging an upfront fee isn't a terribly efficient way to accomplish that, but banks have been making money off of exactly that kind of inefficiency for a very long time, and I don't see why they'd stop.
Companies who hire H1Bs generally have access to more than $100k, they don't need financing help. The problem is that you have to pay this $100k for someone who might not work out, or is free to leave the job whenever they want. Companies would generally be willing to pay even more than $100k (like they do already in taxes) if it was spread out over the lifetime over the worker.
The issue is the fall in demand for tech workers after all the tech layoffs. What needs are the H1Bs filling that can't be filled by recently laid off Google or Facebook employees?
Maybe they can go work as a lab tech at a land grant university except no they can't because the feds decided ag research was too woke and defunded it.
One of the things most misunderstood about H1B visas is that most of them don't go to tech at all.
50% of H1Bs go to "Professional, Scientific and Technical" sector, but that includes all professional fields^1. Many go to finance, consulting, etc. So in reality Tech is just about 1/3 of H1Bs.
I think this is a much larger part of what is going on than anyone admits to, at the moment. Not just in the job market, but a lot of these folks were highly paid for a while and are moving back to places with very strong purchasing power.
This further knocks on to the school situation. CS went from a degree you would be foolish not to try, to one that you probably still want to consider if you can do well.
I meant that this is impacting the full economy pretty heavily.
The complaints of H-1B were there. And I think most of the current backlash to immigration is that MAGA made it the cultural focus of their fight.
But it is hard not to think it is also true that the layoffs in tech have cooled the tech companies' needs for these. Though, I would expect these are highly concentrated in software and software management. Manufacturing and other industries that also drive H-1B will be different.
In particular, I would imagine we need those visas to really build out chip manufacturing quickly. Literally importing people that have done this before.
I don’t think the first sentence is remotely true. Tech layoffs started in 2022 and have been going on since then with little effect on the economy. High interest rates, tariffs and the Iran war this year are the more relevant factors affecting the current economy.
I'm seeing it is hitting Seattle pretty heavy. The surprise, to me, is that it does indeed seem to be limited to the tech areas. People that are getting laid off are, generally, not moving away.
The current consensus seems to be that a large portion of this crowd financed a house and doesn't want to give up their current interest rates. Some thought that they are hoping the job market rebounds where they are.
And I should have limited my assertion to the housing economy more than the overall economy. I would have still been wrong, mind. :D I just wasn't trying to compete with the obvious fallout from the tariffs and the war.
There’s a lot of hiring in AI startups so the overall market is not as bleak for experienced tech workers. The entry level jobs have dried up because big tech is not hiring fresh grads in the same kind of numbers as before.
I completely agree with awarding the visas by salary and this reform would make H1B into a good program. But this piece misses many of the serious problems with the current program, which is in many ways a scam. The important thing to understand about the US tech industry is that pay is bimodal. The average company pays around 130K but the top tier firms pay double that. Because the visas are awarded by random lottery and because most jobs are at average firms, most visas go to such 130K workers. It is an uncontroversial fact from econ 101 that this increase in the labor supply for such workers drives down wages and employment prospects for US natives (while theoretically growing the total size of the pie). Any argument otherwise is pure sophistry. H1B is billed as a way of finding scarce skilled talent that doesn't exist in the US, while in reality displacing US tech workers at average firms. This makes it harder for talented US workers who lack connections and cultural capital to get on the tech labor ladder. In reality hiring decisions are made by humans, not profit maximizing firms, and so we end up with companies where their entire IT department is staffed by Indians on H1Bs. One suspects this has something to do with hiring manager preferences. The dishonesty does not stop there. The hiring companies are supposed to "prove" that there were no qualified US workers for the jobs they are filling by posting ads for fake jobs and conducting fake job searches, supporting make work jobs for HR specialists and employment lawyers. The immigrants are then put in a Kafkaesque bureaucracy where they constantly have to spend time and energy on prolonging their visas and are tied to their employer to accomplish this.
"It is an uncontroversial fact from econ 101 that this increase in the labor supply for such workers drives down wages and employment prospects for US natives (while theoretically growing the total size of the pie). "
Not if the entire industry is experiencing hyper growth like the tech industry did in the last 30-40 years. Tech salaries have grown faster than average inflation during this time. It's not consistent with BS theories that supply is depressing wages. 100-150k at public tech companies was a perfectly respectable salary for senior engineers till 2005 in major tech hubs. Now that number is at least 500k+.
"H1B is billed as a way of finding scarce skilled talent that doesn't exist in the US, while in reality displacing US tech workers at average firms."
Those are the same thing. If you think of the 70th percentile worker at an average firm, every company would like to hire more of those people, and then fire their lowest performers. But there are not enough of those high quality workers for every employer to hire as many as they want.
If the claim was that H1B benefits employers (and potentially has spill over benefits to workers) by making it faster to fill positions with qualified people at a given salary than they would be able to otherwise, while acknowledging that it hurt some US native tech workers, we could have a more honest conversation about it. Instead it is absurdly claimed that employers would not be able to find someone to work a 130K tech job in the US without H1B.
Mostly I think you're underrating the benefits. H1B visas are good for the average American tech worker, because it raises the skill level in American companies, and now they get to work with more competent coworkers. The company will now be more productive, and everyone's stock options will be worth more.
Obviously it sucks to get laid off because your employer has a low opinion of your skills and was able to hire an H1B who they think is better than you. But I think the honorable thing to do is to take the L, suck it up, and do better at your next job, rather than complaining about H1Bs.
I disagree it raises the skill level. Hiring decisions are made by individual hiring managers looking out for their own interests, not rational agents trying to maximize the firm's long term profits. They hire H1B workers because excluding them from the hiring pool would mean it takes longer to hire (a real benefit!), and once the company's IT department is mostly Indian for culture fit reasons. If the visas were only awarded to those earning a 75th percentile salary for software engineers I would agree with you, but they are awarded by random lottery! Awarding by random lottery for a skilled visa program is absurd.
I definitely wouldn't say that the program is a "scam," although it does have some problems, as Matt agrees.
I think it's important to emphasize that the H-1B program isn't just for tech workers. It also provides a critical pipeline of highly skilled workers for rural hospitals, universities, research laboratories, etc. That's one reason I would be a little concerned about Matt's proposal to prioritize absolute salaries, rather than occupation-relative wage levels - we would in effect be dramatically prioritizing analysts/traders for big banks rather than scientists, and I don't think that's necessarily in the national interest.
It is definitely not accepted economic fact that H-1B workers drive down wages and prospects for U.S. natives. The world is much more complicated than Econ 101, and the empirical data suggests otherwise.
On a couple technical points - only "H-1B dependent" employers are required to undertake efforts to recruit U.S. workers for the prospective position offered to the H-1B worker, so that recruitment requirement you mentioned isn't applicable in the majority of cases. I'm also not sure why you stated that H-1B workers have to "constantly spent time and energy on prolonging their visas" - typically, the renewals take place every 3 years and are solely the responsibility of the employer.
I certainly agree that the world is more complicated than econ 101, but what part of the supply and demand picture do you not agree with? H1B adds more workers to the labor supply who are willing to work at a given wage. This decreases wages for the people directly competing with the H1Bs. I am not claiming anything about US natives as a whole - only the subset directly competing with H1B workers and specifically my concern is with workers at average firms, not FAANG. The program itself is structured around this fact, making employers "prove" that there are no qualified US workers available, which only makes things worse by creating BS make work jobs. I worked at a company where I supervised H1B workers and have first hand knowledge of the process. I don't know if they were officially an "H1B dependent employer" or not, but if yes then many large companies heavily using H1Bs for their IT workers must fall into this category. I'm saying the immigrants have to constantly spend time and energy on it because I've talked to them and helped them jump through the hoops.
Employer: Good. You wouldn't believe the number of applicants we have had who were based in Guinea.
Me: You mean like New Guinea?
Employer: No no, the old ones. Is New Guinea nearer the new world? Anyway, we are so happy to finally meet a candidate who has already lived here. Ok, what else?
Me: I speak English.
Employer: Well, I can tell. Between you and me, thats a rapidly diminishing skill. Less than two billion speak English. That leaves like so many billions out. Anything else?
(ed: Fewer than two billion. The skill is indeed rare)
Me: I have other soft skills.
Employer: *gasp* no! not other soft skills! You have hit the trifecta. I resign. I am not worthy. You, yes you, must take my job right now.
It *is* slightly wacky that there are no fewer than six countries (kind of: one is technically part of France) with some permutation of "Guinea" as part of their name: Guinea, Guinea-Bissau, Equatorial Guinea, Papua New Guinea, Guyana, and French Guiana. It would be seven, but Dutch Guiana had the decency to change its name to "Suriname" when it became independent.
Wonderful analysis, with which I find myself in total agreement. Since I did immigrate into the US based on my skills in high tech, I thought sharing my own personal experience might be a useful contribution to the discussion, too, so, here it is.
Around the mid-90s, as my first marriage (in my native country, Italy) failed, and my high-tech career was flourishing, I did consider moving to the US (where I was starting to have romantic relationships based on long-distance net connections, AND I had worked on and off based on employment on US firms with Italian subsidiaries -- Texas Instruments and IBM -- so I knew I'd really like the overall "vibe"... not to mention the detail that I knew my compensation, as an employee with high tech skills, would at least DOUBLE compared with Italy's meager ones).
However... I studied the possibilities available back then (30+ years ago) and decided H1B was way too restrictive and hard to get in, so I chose another career path. I wrote a couple books about the emerging tech I had luckily adopted (the Python programming language: a totally niche thing back then, THE most popular programming language in the world today -- lucky me for being on the best bandwagon SO early on!-).
My books became best-sellers, and got me great job offers from early-adopting companies all over the world -- and, together with my reputation as a very helpful poster to usenet's comp.lang.python, eventually got me prestigious awards from the Python community, too.
I eventually chose to drop out of the Italian employee treadmill and risk myself into freelance consulting in tech and tech-management matters -- quite successfully, but at the cost of flying all over the world to supply my services to various clients while still living in Italy... spending more time in airplanes, airports, and hotels, than I did at either my own office, or clients' ones, was a truly stressful life... but it did keep earning me kudos and recognition...
Eventually one of my US romantic connections gelled into perfect intensity -- my wonderful now-wife, US-citizen and veteran, Anna, moved to Italy to live with me (our connection also reinforced by fiery shared interest in this "Python" thingy...), we married, we started the uphill fight to get her a work and residence visa in Italy (even harder than the US makes it, believe it or not -- I wouldn't believe that, myself, if I hadn't experienced both!), she became my co-author on some of my best-selling tech books...
Eventually one of the companies I was offering my consultancy services to, a startup with a weird name, having put me to the wringer of more and harder interviews than any previous potential clients ever had, ended up offering me a tech manager job ("we'd rather not employ consultants for technologies crucial to the company" -- Python, in my case)... if I was to move to the US (or maybe to their nascent offices in Dublin or Zurich, which, visa-wise, would have been a no-brainer, BUT... I had previous experience with being at a nascent foreign office, and had decided that being at "the mothership" was really the only good choice for me).
On checking with the immigration lawyers they employed, it emerged that I probably qualified for an O-1 visa, "for the individual who possesses extraordinary ability in the sciences" &c, given the amount of awards, tech best-selling books, and likely endorsements from previous employers and clients, that I had accumulated by then.
So I ended up immigrating to the US in 2005 (after that startup's IPO) -- my offer had been extended in 2004, but visa proceedings (e.g., me asking so many people for endorsements) took too long for me to make it any earlier than early '05.
It all worked out very well in the end -- "I like to be in America", as Sondheim's wonderful lyrics to Bernstein's wonderful song "America" (in "West Side Story") say -- that startup (with only modest help from my own work) flourished, I naturalized as a US citizen, I contributed a fortune or two in taxes to the US and California over a couple of decades -- I'm now retired, but SO like it here that my wife and I still live in Silicon Valley (and visit Italy often rather than moving back there, though I do miss it and my many relatives when I'm here in the US).
BTW, that startup's weird name was... "google". Weird indeed, right? But, it DID work!
So if the US immigration system eventually worked so well for me, why do I so fully agree with the ideas in this post? Think about it -- with a saner visa system for high-skill tech-professional immigrants, I might have moved to the US maybe a decade earlier -- 10 more years of high earnings and high taxes (income, sales, SS contributions, Medicare ones...) paid by me (and my employers) to Federal and State systems -- to substantial mutual economic benefits for me, for California, AND for the US.
When I did move, I was about 50 years old -- so, just a couple decades of big tax contributions (then retirement, when I get some money back from SS and Medicare -- although my savings, productively invested in the US, yield enough that my Federal and State taxes are still pretty high). Had a sane, welcoming immigration system induced me to immigrate at 40 (as I considered back then, but rejected on deeply examining that system), that would have increased the time of my highest contributions to the US and State economies (and tax revenues) by fully 50% -- far from chicken feed...
Canada just had it's reckoning with diploma mills. A few schools had grown 10x with "education" programs that involved a couple optional lectures every week, but came with a 20 hour/week work visa. 100's of thousands of "students" came in through these programs and never learned a thing. They locked it down last year, and Canada's population actually shrank last year because of it. Hopefully going forward the student work visa program will help highly qualified students earn a degree, and then they can stick around :)
Just switch to a point system, and make things like wage potential, attended an American university (maybe extra points for masters/phds), speaks english, lived/worked in the US before and didn't break any laws (maybe under a temporary visa), etc give extra points.
At one point I had tried to design an ideal point structure and had come around to this:
wage offer / earnings potential: 25
education: 15
US degree: 10
English: 10
age / work horizon: 10
prior lawful US work and status compliance: 15
shortage occupation / national interest: 10
assimilation / civics / clean record: 5
Some would be a bit hard to find objective measures for, but overall seems better than a lottery!
The problem is that most Americans are at least as dumb on the topic of immigration as they are on housing, if not more.
The average voter's reaction to the mention of low-wage immigration would be to worry that they're leeching off the system by sucking up welfare benefits. The reality is the opposite: any immigrant working legitimately pays taxes and is not eligible for major welfare benefits. The typical solution to the voters' imagined problem is to make it harder to get a work permit, which just encourages under-the-table work and reduces tax intake.
When it comes to H1B, the prevailing negative sentiment about the program comes from right-wing tech bros on Reddit, emboldened by Silicon Valley CEOs' recent rightward turn to say what they really think. Of course, by saying that, they reveal the truth: 1. they are bigots, and 2. they're not being hired because of their own innumeracy, not because of H1B candidates who make up just a few percent of the overall tech workforce.
The only way to fix this is to sneak reform through Secret Congress style and force voters to live in the better reality until they move on to the next fake concern. There is no way to convince people of the truth on this topic.
Many Americans get confused when they hear immigrants are on "benefits." In immigration law, "benefits" just means they are awarded the right to be in the US.
> But actually it’s good! Adam Ozimek and Sarah Eckhardt show that the median H-1B worker earns $130,075 per year (the 91st percentile nationally), meaning that the average H-1B household makes a net federal fiscal contribution that’s 2.6 times as much as the typical American household’s. This is a big win for the federal budget
The program's intention is that companies only hire H-1B if an American isn't available IIUC, and here you've implicitly assumed it works as intended. But my impression is that much of the anti-H1B rhetoric says it doesn't work as intended, so you'd best argue that point specifically. Also, by leaving it implicit, anti-H1B people will think you amazingly didn't comprehend that An American Could Have Had That Job! (What a *****!)
Hoped to hear more of a steelman here. The way it's been argued, a $100k fee is a strong selection mechanism to only select for the highest of the high skilled. Not only in terms of wages alone, but also as a form of immigration loophole. India gets mentioned and then not expounded upon - it's my understanding there really is an aspect of nativist/racial resentment fueling this, with some even making allegations that it's a caste-based form of chain migration (this almost certainly overstates the case). You see the same thing with a China-hawk flavour too, especially concerning employment for graduates, where vague fears over national security are used to justify xenophobia. Not entirely unjustified, of course, China does historically play fast and loose with IP and espionage. But it's difficult to make a meritocratic-flavoured argument while throwing shade at this cohort of academic overachievers (who are also very lucrative for universities, so a double-dip fiscal win for us). I guess this is in line with Matt's general "use race-neutral framing whenever possible" popularism ethos.
Plus the fee is, of course, just straight-up upfront revenue for the government, taken from unsympathetic groups, the same way one can view tariffs as free yuge revenue with enough blinders on. Business As Usual from this admin, not that one should encourage it, but it's very understandable.
I mean... why isn't it "free [not terribly] yuge [but still not nothing] revenue"? The same way datacenter builders pay off the locals to compensate them for the imagined hardship of living near a datacenter, we should be able to pay off the losers who experience a taco truck on every corner as a personal hardship. This becomes much harder for asylum immigration, family sponsorship, etc. - for H1Bs the arrangement is just too perfect.
Like I said below, a $100k fee is a de facto ban on H1B visa holders in academia, not any kind of “selection mechanism.”
But this is not compelling either - are you claiming that academics are paid at exactly the cost that universities can absorb to employ them, such that paying them any more would make the transaction nonviable? Do we really not believe that at least some hiring universities might be able to absorb at least some of the cost of this fee? Hard to square that story with the story told 8 billion times over the past 10 years of an academic job market that is absolutely swamped with candidates, and universities able to sign them up for indentured servitude as adjuncts despite the substantial value they create for the university. Furthermore it's already substantially more expensive to hire an H1B holder for legal reasons; my impression is that the legal fees are on the same order of magnitude as this $100k Trump fee. If the fee clears out the lottery so that employers aren't spending tens of thousands a year on lawyers to fight for the visaholders to stick around, this could net out to neutral.
What a desert of policy analysis debate we live in these days.
On the one side, here's a long-standing policy which despite some flaws is net good and here are some ways to improve it.
On the other hand, here's Trump who destroys everything he touches and everything just gets bad and the US looks into the abyss of becoming dysfunctional.
I guess sure we can talk about improving programs in the hopes of a Democratic trifecta in 2028 but it's hard to have one's heart in it.
Often lost in this debate is that there are well-vetted solutions to the main problems in the H-1B and OPT programs waiting to be picked up. People don't want you to know this!
Worried that employers are paying visa holders less than similarly qualified Americans? We could just fix that by requiring employers not pay less than the median American in the same occupation with the same combination of education and experience.
https://ifp.org/prevailing-wage-benchmarking/
Worried that an H-1B lottery advantages employers hiring lower-skilled workers instead of using scarce visas on the highest-skilled, highest-paid workers? We could just fix that by allocating visas according to compensation.
https://ifp.org/the-wage-level-mirage/
Worried that OPT participants' exemption from FICA taxes is unfair to American workers? There is a Republican-led bill in Congress that could just fix this for good.
https://ifp.org/opt-fica-exemption/
"Worried that employers are paying visa holders less than similarly qualified Americans? We could just fix that by requiring employers not pay less than the median American in the same occupation with the same combination of education and experience."
You could also just make it easy for visa holder to job switch and this would go away on its own
"We could just fix that by requiring employers not pay less than the median American in the same occupation with the same combination of education and experience."
Why median if the goal is to import talent that can't be found in the US than we'd want 90th percentile, right? The other huge issue is "median salary" is very gameable. A software developer can make $100k or $10 million with the same generic corporate title. Are they working at Anthropic or supporting the billing department of a midwestern auto insurer?
The $100k fee is just designed to be punitive, not to make money. If you charged $30k a year, for example, I think that would be fine! A lot of tech companies would just pay that! That's about how much you pay a recruiter for a new position anyway. And since an H1B can last for 5 or more years, the government could actually make *more* money that way.
The $100k fee is hard to justify, even for a worker who is going to be worth more than that, because you have to pay it all up front. $100k for someone that could just switch to a new job a few months later.
I think the best reform would be to eliminate the sponsorship requirement of H1-B.
Most of the resentment I’ve seen comes from the belief that companies prefer H1-B workers because of they have a unique level of control over them. This seems like yet another place where normal liberal sentiments and good policy align.
The reason this is a problem is only because of the per country quota in EB green cards. I got my green card in less than 5 years but it takes decades now if you’re from India and 5+ years if you’re from China. Fixing the green card backlog fixes this issue. There are other simple fixes like if the PERM and I-140 are approved, then you don’t need to file them again when you switch jobs. It’s possible to switch jobs easily on H-1B, at least prior to the Trump administration, but it involves redoing step 1 and 2 of the green card application. Once you get to step 3, you’re relatively safe.
Yes, this is such a widespread misconception. H-1B workers can absolutely change jobs. The employee leverage comes when the employee sponsors the green card application and you wait in the queue. Which, as you described, is a very solvable problem.
Importing skilled workers is a little too convenient compared to hiring Americans for several reasons, but the most important is sociological. There are in fact huge numbers of smart, talented, highly educated people in India. But hiring a room full of these people in California gets you bizarrely WAY less shit than hiring a room full of high SAT math score Americans.
This works even though they are a similarly privileged strata of their own (much more unequal) society, and even though this in no way benefits the descendants of American slavery who are owed representation in the logic of DEI.
It’s not surprising that this arbitrage would eventually be closed.
A room of high SAT Math scoring Americans would also mostly be Asian? Don’t think there’s that much of a difference.
Also, privilege hierarchies are global. Their society as a whole is less privileged than ours so they are lower in the global hierarchy. Someone who is the chief of an African village but still lives in a mud hut with no education is clearly less privileged than an upper-middle class American growing up in a McMansion going to an excellent public school even if the chief is higher in his own society.
Much of the problem with both the underlying diagnosis and the proposed fixes stems from MAGA’s seeming insistence that there is thing called a “job,” and it functions like a chair in an auditorium that either an American or a non-American sits in. But a job is much more like an opportunity to *build* a chair that then can get purchased by some auditorium somewhere, and eventually get sat in. Sometimes, the person who build the chair can even get that seat!
Something something lump of labor fallacy
The first objection, around the $100k fee being too steep, is not convincing, because there’s no argument made that the fees can’t be financed. If a ~$200k payment over the next 10 years of the worker is really too risky a proposition for a lender, then what are we doing here? I suspect that there are similar just-one-step-on-the-chessboard arguments being made here.
An H1B is a 3 year visa. An OPT is a 1 year one. It's not a green card. At those time frames lenders would balk at that.
The usual pipeline is OPT after college > H1B > H1B Renewal > Green card application. Essentially American businesses have a 'try them out before you buy them' immigration system. But there's many things that can go wrong there to make a bank feel comfortable with that arrangement.
But those timelines aren't relevant - the question is what's the likelihood of the visa-holder transitioning into a green card, not whether they can pay back the loan in two years. Banks are not stupid, and the same way they will lend doctors money against their future salaries while they're in residency to buy a house, they will lend an AI PhD who hasn't even gotten a job yet against their future salary to pay a loan that will be a relatively small fraction of their first year salary.
The fact that each step has some likelihood of sending them back is a more interesting objection - but again, the premise here is that we're solving the problem of having a bunch of workers who we don't care to actually offer citizenship to, right? In which case, that likelihood drops to much lower levels.
They won't lend to them. Try getting a mortgage as an H1B holder, even a 10 year one. A green card or citizenship is required.
Are you in favor of other regulations that make it harder to fire employees, such as requiring approval from external bodies as they have in Europe?
I am pro-immigration. But I am anti the deterioration in the quality of analysis that I’m getting as a SlowBoring subscriber, and I would like Matt to address obvious points like this one if he’s going to bother critiquing a policy at all.
It’s illegal. The law specifically bans the government from profiting from immigration applications. The fee can only cover the cost of employing federal immigration workers and the cost of processing. The Congress is welcome to change the law but right now, it’s illegal.
A lot of H1B holders work in academia, where paying a $100k fee is absolutely punitive and is an effective ban on hiring H1B workers.
I know academics who, while still first year PhD students, took out larger student loans. Seems like, all else equal, a bank willing to lend to the PhD student would be willing to lend to the employed academic.
All else might not be equal, but that's the discussion to have, not this caveman analysis "we have this arrangement that will generate millions of dollars of economic surplus but we can't find $100k to unlock it!"
Personally, the way I prefer it is we simply make H1B holders pay a flat surcharge until they're citizens. We have a bunch of loser Americans who experience having taco trucks on every corner as a hardship; we have a bunch of economically productive entities who would gobble up the purveyors of taco-truck-or-equivalent culture; it is simple to compensate the loser Americans for their pain and suffering and make it a win-win. Charging an upfront fee isn't a terribly efficient way to accomplish that, but banks have been making money off of exactly that kind of inefficiency for a very long time, and I don't see why they'd stop.
Companies who hire H1Bs generally have access to more than $100k, they don't need financing help. The problem is that you have to pay this $100k for someone who might not work out, or is free to leave the job whenever they want. Companies would generally be willing to pay even more than $100k (like they do already in taxes) if it was spread out over the lifetime over the worker.
The issue is the fall in demand for tech workers after all the tech layoffs. What needs are the H1Bs filling that can't be filled by recently laid off Google or Facebook employees?
Layoffs disproportionately hit low performers and good luck getting them to move to eg Ohio.
Maybe they can go work as a lab tech at a land grant university except no they can't because the feds decided ag research was too woke and defunded it.
One of the things most misunderstood about H1B visas is that most of them don't go to tech at all.
50% of H1Bs go to "Professional, Scientific and Technical" sector, but that includes all professional fields^1. Many go to finance, consulting, etc. So in reality Tech is just about 1/3 of H1Bs.
1: https://ssti.org/blog/useful-stats-look-h-1b-visa-program-industry-employer-and-state
Whenever you want to hire the best people, some of those people will be H1Bs.
Which brings up the other issue a lot of H1Bs were doing low level accounting and IT tasks when the program was designed for more rare skills.
They’re a lot cheaper overall- contractors and consultants are cheap to hold and when MVP is all you need, they’re easy to rotate through.
I think this is a much larger part of what is going on than anyone admits to, at the moment. Not just in the job market, but a lot of these folks were highly paid for a while and are moving back to places with very strong purchasing power.
This further knocks on to the school situation. CS went from a degree you would be foolish not to try, to one that you probably still want to consider if you can do well.
It is a factor but there were lots of losers complaining online about H-1B and Indians even when hiring in tech went nuts.
I meant that this is impacting the full economy pretty heavily.
The complaints of H-1B were there. And I think most of the current backlash to immigration is that MAGA made it the cultural focus of their fight.
But it is hard not to think it is also true that the layoffs in tech have cooled the tech companies' needs for these. Though, I would expect these are highly concentrated in software and software management. Manufacturing and other industries that also drive H-1B will be different.
In particular, I would imagine we need those visas to really build out chip manufacturing quickly. Literally importing people that have done this before.
I don’t think the first sentence is remotely true. Tech layoffs started in 2022 and have been going on since then with little effect on the economy. High interest rates, tariffs and the Iran war this year are the more relevant factors affecting the current economy.
I'm seeing it is hitting Seattle pretty heavy. The surprise, to me, is that it does indeed seem to be limited to the tech areas. People that are getting laid off are, generally, not moving away.
The current consensus seems to be that a large portion of this crowd financed a house and doesn't want to give up their current interest rates. Some thought that they are hoping the job market rebounds where they are.
And I should have limited my assertion to the housing economy more than the overall economy. I would have still been wrong, mind. :D I just wasn't trying to compete with the obvious fallout from the tariffs and the war.
There’s a lot of hiring in AI startups so the overall market is not as bleak for experienced tech workers. The entry level jobs have dried up because big tech is not hiring fresh grads in the same kind of numbers as before.
I completely agree with awarding the visas by salary and this reform would make H1B into a good program. But this piece misses many of the serious problems with the current program, which is in many ways a scam. The important thing to understand about the US tech industry is that pay is bimodal. The average company pays around 130K but the top tier firms pay double that. Because the visas are awarded by random lottery and because most jobs are at average firms, most visas go to such 130K workers. It is an uncontroversial fact from econ 101 that this increase in the labor supply for such workers drives down wages and employment prospects for US natives (while theoretically growing the total size of the pie). Any argument otherwise is pure sophistry. H1B is billed as a way of finding scarce skilled talent that doesn't exist in the US, while in reality displacing US tech workers at average firms. This makes it harder for talented US workers who lack connections and cultural capital to get on the tech labor ladder. In reality hiring decisions are made by humans, not profit maximizing firms, and so we end up with companies where their entire IT department is staffed by Indians on H1Bs. One suspects this has something to do with hiring manager preferences. The dishonesty does not stop there. The hiring companies are supposed to "prove" that there were no qualified US workers for the jobs they are filling by posting ads for fake jobs and conducting fake job searches, supporting make work jobs for HR specialists and employment lawyers. The immigrants are then put in a Kafkaesque bureaucracy where they constantly have to spend time and energy on prolonging their visas and are tied to their employer to accomplish this.
"It is an uncontroversial fact from econ 101 that this increase in the labor supply for such workers drives down wages and employment prospects for US natives (while theoretically growing the total size of the pie). "
Not if the entire industry is experiencing hyper growth like the tech industry did in the last 30-40 years. Tech salaries have grown faster than average inflation during this time. It's not consistent with BS theories that supply is depressing wages. 100-150k at public tech companies was a perfectly respectable salary for senior engineers till 2005 in major tech hubs. Now that number is at least 500k+.
"H1B is billed as a way of finding scarce skilled talent that doesn't exist in the US, while in reality displacing US tech workers at average firms."
Those are the same thing. If you think of the 70th percentile worker at an average firm, every company would like to hire more of those people, and then fire their lowest performers. But there are not enough of those high quality workers for every employer to hire as many as they want.
If the claim was that H1B benefits employers (and potentially has spill over benefits to workers) by making it faster to fill positions with qualified people at a given salary than they would be able to otherwise, while acknowledging that it hurt some US native tech workers, we could have a more honest conversation about it. Instead it is absurdly claimed that employers would not be able to find someone to work a 130K tech job in the US without H1B.
Mostly I think you're underrating the benefits. H1B visas are good for the average American tech worker, because it raises the skill level in American companies, and now they get to work with more competent coworkers. The company will now be more productive, and everyone's stock options will be worth more.
Obviously it sucks to get laid off because your employer has a low opinion of your skills and was able to hire an H1B who they think is better than you. But I think the honorable thing to do is to take the L, suck it up, and do better at your next job, rather than complaining about H1Bs.
I disagree it raises the skill level. Hiring decisions are made by individual hiring managers looking out for their own interests, not rational agents trying to maximize the firm's long term profits. They hire H1B workers because excluding them from the hiring pool would mean it takes longer to hire (a real benefit!), and once the company's IT department is mostly Indian for culture fit reasons. If the visas were only awarded to those earning a 75th percentile salary for software engineers I would agree with you, but they are awarded by random lottery! Awarding by random lottery for a skilled visa program is absurd.
That has changed this year. It's not by random lottery anymore.
I definitely wouldn't say that the program is a "scam," although it does have some problems, as Matt agrees.
I think it's important to emphasize that the H-1B program isn't just for tech workers. It also provides a critical pipeline of highly skilled workers for rural hospitals, universities, research laboratories, etc. That's one reason I would be a little concerned about Matt's proposal to prioritize absolute salaries, rather than occupation-relative wage levels - we would in effect be dramatically prioritizing analysts/traders for big banks rather than scientists, and I don't think that's necessarily in the national interest.
It is definitely not accepted economic fact that H-1B workers drive down wages and prospects for U.S. natives. The world is much more complicated than Econ 101, and the empirical data suggests otherwise.
On a couple technical points - only "H-1B dependent" employers are required to undertake efforts to recruit U.S. workers for the prospective position offered to the H-1B worker, so that recruitment requirement you mentioned isn't applicable in the majority of cases. I'm also not sure why you stated that H-1B workers have to "constantly spent time and energy on prolonging their visas" - typically, the renewals take place every 3 years and are solely the responsibility of the employer.
I certainly agree that the world is more complicated than econ 101, but what part of the supply and demand picture do you not agree with? H1B adds more workers to the labor supply who are willing to work at a given wage. This decreases wages for the people directly competing with the H1Bs. I am not claiming anything about US natives as a whole - only the subset directly competing with H1B workers and specifically my concern is with workers at average firms, not FAANG. The program itself is structured around this fact, making employers "prove" that there are no qualified US workers available, which only makes things worse by creating BS make work jobs. I worked at a company where I supervised H1B workers and have first hand knowledge of the process. I don't know if they were officially an "H1B dependent employer" or not, but if yes then many large companies heavily using H1Bs for their IT workers must fall into this category. I'm saying the immigrants have to constantly spend time and energy on it because I've talked to them and helped them jump through the hoops.
Employer: So, why should we hire you?
Me: I have already lived in the United States.
Employer: Good. You wouldn't believe the number of applicants we have had who were based in Guinea.
Me: You mean like New Guinea?
Employer: No no, the old ones. Is New Guinea nearer the new world? Anyway, we are so happy to finally meet a candidate who has already lived here. Ok, what else?
Me: I speak English.
Employer: Well, I can tell. Between you and me, thats a rapidly diminishing skill. Less than two billion speak English. That leaves like so many billions out. Anything else?
(ed: Fewer than two billion. The skill is indeed rare)
Me: I have other soft skills.
Employer: *gasp* no! not other soft skills! You have hit the trifecta. I resign. I am not worthy. You, yes you, must take my job right now.
It *is* slightly wacky that there are no fewer than six countries (kind of: one is technically part of France) with some permutation of "Guinea" as part of their name: Guinea, Guinea-Bissau, Equatorial Guinea, Papua New Guinea, Guyana, and French Guiana. It would be seven, but Dutch Guiana had the decency to change its name to "Suriname" when it became independent.