Immigration and Border Enforcement

H-1B Petition Fee

Employers would pay 103,265 dollars on top of today's charges for every new H-1B worker hired under the annual cap.

7 arguments 13 Sources
  • 85,000 new H-1B visas a year fall under the cap, 20,000 of them reserved for holders of American advanced degrees [1]
  • 211,600 registrations competed for those places in the season of 2026 [7]
  • 8.8 billion dollars a year is what the department expects the fee to raise, assuming no employer stops filing [1][2]

Employers who hire a new H-1B worker under the annual cap pay a few thousand dollars in fees today. On 25 August 2026 the Department of Homeland Security proposed an extra charge per cap petition, presented as protection for American workers and opposed by business groups and researchers [1][2]. Comments closed on 24 September 2026, and courts have twice struck down an earlier 100,000-dollar charge on new entrants from abroad, most recently on 30 September 2026 [4][5].

The proposal

Employers would pay 103,265 dollars on top of today's charges for every new H-1B worker hired under the annual cap.

The key question

Does the fee protect jobs for American workers without costing the country skilled workers that employers need?

The debate

The department says the fee recovers the cost of the immigration system and makes employers less likely to pick a foreign worker over a qualified American. Business groups and immigration researchers answer that a charge of this size would price most employers out of the program, push work abroad and keep out graduates of American universities. They also dispute the legal basis, since most of the money would be spent outside the agency that collects it. No final rule has been published.

  • “DHS believes that U.S. employers, if required to pay an additional $103,265 fee when filing an H-1B cap-subject petition, would be less likely to hire an H-1B worker over a qualified and highly-skilled American worker unless the need is legitimate and they have no alternative for obtaining the specialized skills of the employee.”

    argues for: Department of Homeland Security, Trump administration Proposed rule, Federal Register, 25 August 2026, page 54832
  • “DHS's demand projections rest on a single computer simulation, not observed employer behavior, that assumes values for productivity gains, hiring costs, and employer willingness to pay, then concludes that filing volumes will hold steady at 85,000 visas per year.”

    argues against: Patrick Shen, U.S. Chamber of Commerce Comment letter on the proposed rule, 24 September 2026

The analysis

H-1B Petition Fee

40% of the weight is for, 60% against.

An argument's weight combines three questions: how big the effect is, how certain it is and how much the good at stake matters. The larger the slice, the more the argument weighs.

For

Against

Answer to the key question

Does the fee protect jobs for American workers without costing the country skilled workers that employers need?

Hardly, and it costs more than it protects. Some posts go to Americans, but the studies disagree on how often, and many vacated posts go to other foreign workers, move abroad or stay empty. The workers deterred would have paid more in taxes than they use and filled posts employers value. The fee itself only moves money from employers to the agencies.

On the current analysis, the new H-1B petition fee is a poor proposal.

Very likely worse

Very likely worse: Under the assumptions used, the proposal brings more than it costs in fewer than 1 of 100 model runs.

Overall effect: large

How strongly the proposal works overall, in five steps from very small to very large, comparable across all debates. The ring is always full, even when the overall effect is small; this step says how large it is.

The analysis comes out against the rule, largely for the reasons its critics give. The fee itself is money changing hands from employers to the agencies, and neither side does more good with it, so it cancels. What counts are the hires that no longer happen: people who earn far more here than elsewhere, pay more in taxes than they use and fill posts employers value. The department's own argument, posts going to Americans, is real but small, because many vacated posts go to other foreign workers, move abroad or stay empty. The largest loss, the earnings of the workers kept out, rarely comes up in the debate; without it the rule would still fall clearly behind.

In detail

All arguments

Arguments for2 arguments

Fee money for the immigration agencies

15points

Every cap petition that still goes ahead brings 103,265 dollars to the immigration agencies. Even if a third of hiring under the cap stops, the fee would raise about 5.7 billion dollars a year. The same money leaves the employers who pay it, which is the first argument against.

Value 5 · Public financesImpact 4.9Plausibility 6
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is public revenue, priced at the middle of the scale like any other dollar. The proposal sends the money to the immigration fee account and from there to immigration courts, enforcement, vetting, visa processing and labour inspection in four departments, about 8.8 billion dollars in all [1][2]. Those budgets are funded today from other receipts and appropriations, so the fee relieves the public purse rather than buying anything new in particular, and it is valued as general public money. Whether a payment of this kind is a gain or a loss depends on where it comes from and where it goes, and that difference is booked in the Impact of the pair rather than here. Its counterpart, the same amount paid by employers, is the first argument against. The value is the middle of the scale, the level this site uses for public money whatever it is spent on.

Impact

The fee is 103,265 dollars for each petition under the cap of 85,000 a year, of which 20,000 are reserved for holders of American advanced degrees [1]. The department expects all 85,000 to be filed, which gives its figure of about 8.8 billion dollars a year [1][2]. This evaluation does not follow that assumption: five analyses of a fee of this size expect demand to fall by 63 to 91 percent, four of them expect the cap to go unfilled, and after the 100,000-dollar charge on new entrants of 2025 consular H-1B filings fell by more than 90 percent while employers paid it for only about 700 people [2][3][4]. A fall of 35 percent in cap petitions actually filed is used here, in a range from 12 to 70 percent, the middle of the published estimates of cap visas no longer used. That leaves 55,250 petitions a year and 5.7 billion dollars. Every one of the four cap seasons in the horizon counts in full. The range runs from 2.7 to 7.8 billion dollars, depending on how many employers still pay. The Impact is large in money and small in welfare, because the same sum appears as a cost on the other side.

▸ Show calculation ▾ Hide calculation
85,000 petitions a year
Cap-subject H-1B petitions a year without the fee
× × 0.65 Setting, range 0.30 to 0.88: the mirror of the fall in filings used in the arguments on the people deterred
55,250 petitions a year
Share still filed at the new price
× × 103,265 dollars the amount in the proposed rule, on top of existing fees
5.71 billion dollars a year
Fee per petition
[1]
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 2.27 to 6.66 billion euro across the range of filings
4.92 billion euro a year
In euro
× × 1.0 the agencies and the employers who pay are both counted at the weight of median income
4.92 billion euro a year
Weight of a euro on both sides
= 1 billion euro a point scale of this evaluation
4.92
Normalised Impact
Score 4.92 Impact × 5 Value × 6 Plausibility ÷ 10 = 15 of 100

Plausibility

The amount depends on one behavioural link: how many employers still file once a petition costs about thirty times what it costs now [2][3]. The counterfactual is the current weighted selection with no fee. No fee of this size has been charged on all cap petitions, so the number rests on estimated price responses and on one precedent, the 2025 charge on entries from abroad, which exempted people already in the country and so covered a narrower group [2][4]. The confounder in that precedent is that the charge arrived together with other restrictions, which may have deterred filings on their own. Reverse causation does not arise, because the fee is set by rule. That money is collected once the rule applies is not in doubt; how much is, and that doubt sits in the range. The Plausibility is just above the middle, the ceiling for a projection of a behavioural response, and the argument it is paired with carries the same figure.

Evidence: Projection · plausibility at most 6 Evidence of effect: Mechanistic · at most 6

Some posts go to American workers

0.39points

The rule's stated aim is to protect American workers. With fewer H-1B hires under the cap, some posts go to an American instead, mostly at the lower end of the wage table where the fee bites hardest. The visa-lottery studies disagree on how often that happens.

Value 5 · Household budgetsImpact 0.3Plausibility 2.5
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is earnings in the hands of American workers who get a post they would not otherwise have got. What is counted is not the salary, since a post at a similar wage exists for them either way, but the gap between this post and the next best one open to them. It is money and is priced at the middle of the scale. The posts the fee removes are mostly at the lower wage levels and held by early-career workers between the second fifth and the middle of American incomes, which raises the weight of each dollar; that is set in the Impact. What an employer loses when a post goes unfilled is the second argument against. The value is the middle of the scale, because the stream is a difference in earnings.

Impact

Averaged over the four years, about 56,500 fewer people work on cap-subject H-1B visas than without the fee: 85,000 petitions a year, 35 percent fewer of them, with about 1.9 yearly cohorts in the workforce at once, the same count this site uses for the weighted selection rule [1][2]. Whether an American takes each vacated post is the open part: the employer may hire another foreign worker who needs no cap visa, move the work abroad, or leave it undone; multinationals hire about 0.4 people abroad for every H-1B refused [10]. Three in ten is used here, in a range from one to six in ten, the setting this site uses for entry-level H-1B posts; that gives about 17,000 posts. Each is worth the gap between this post and the worker's next best option, set at 17,400 dollars a year, in a range from 9,300 to 29,000, again as for the weighted selection rule. These are early-career earners, so a dollar is counted at 1.2. The result is 354 million dollars a year, in a range from 0.09 to 1.4 billion. The Impact is a small fraction of the fee money, because only some of the vacated posts are filled from here and only the step up counts.

▸ Show calculation ▾ Hide calculation
85,000 petitions a year
Cap-subject H-1B petitions a year without the fee
× × 0.35 Setting, range 0.12 to 0.70: the middle of five published estimates of cap visas no longer used at a fee of this size
29,750 fewer petitions a year
Fall caused by the fee
× × 1.9 averaged over the four years, the same count this site uses for the weighted selection rule
56,525 fewer H-1B workers on average
Yearly cohorts in the workforce at once
[2]
× × 0.3 Setting, range 0.1 to 0.6, the same as this site uses for entry-level posts under the H-1B selection rule
16,958 posts
Posts taken by an American instead
× × 15,000 euro a year Setting, range 8,000 to 25,000 euro, the same as for the H-1B selection rule: only the step up counts
0.25 billion euro a year
Gain over the next-best job
× × 1.2 between the American second fifth (1.4) and the median (1.0), taken half and half; range 0.077 to 1.212 billion euro, factor ranges combined
0.31 billion euro a year
Weight of a euro for early-career earners
= 1 billion euro a point scale of this evaluation
0.31
Normalised Impact
Score 0.31 Impact × 5 Value × 2.5 Plausibility ÷ 10 = 0.4 of 100

Plausibility

Whether H-1B workers displace Americans has been measured with the visa lottery itself, and the results point in different directions. The counterfactual is the same employers with the visas they lost in the draw. Doran, Gelber and Isen (2022) compare firms that won and lost the 2006 and 2007 lotteries, a quasi-experimental design, and find that each additional H-1B worker crowds out other workers at the firm, about one and a half on their estimate [8]; that supports the claim that a missing H-1B hire is replaced. Mahajan and co-authors (2024), with the same lottery design on Census employment records for 25 states, find no difference in the employment of American-born college graduates between winning and losing firms: the losers replaced part of the missing hires with other immigrants [9]. The confounder in both would be firm growth that differs after the draw for other reasons, which the random assignment handles. Reverse causation is excluded by the lottery. The claim is therefore supported for replacement in general and not found for replacement by Americans. The Plausibility is low: only part of the claim is supported, and about a quarter of the gain claimed here is likely to materialise.

Evidence: Study · plausibility at most 7 Evidence of effect: Quasi-experimental · at most 8 Placement: Partial aspect supported · plausibility 2.5–3

Open: A study of employment of American graduates at firms that stop filing cap petitions once the fee applies would lift P towards 5 if it found them hired into the vacated posts.

Arguments against5 arguments

Employers pay it

15points

The employers who still hire under the cap pay the fee, about 5.7 billion dollars a year. It is the same money the agencies receive, counted at the same weight, so the two arguments cancel. They are shown so that the size of the payment is visible.

Value 5 · Business budgetsImpact 4.9Plausibility 6
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is money paid by businesses, priced at the middle of the scale. Cap-subject petitions are filed by companies, from large technology firms and consultancies to hospitals and small firms; universities and non-profit research institutes are exempt [1][3]. Whether part of the charge reaches workers through lower pay offers is open; those workers earn around the fourth fifth of American incomes, so either way the payment sits close to the weight of median income. The payment is therefore booked at the same weight as the receiving side. The value is the middle of the scale, the level for money paid by businesses and households.

Impact

The amount is the one the revenue argument derives: about 55,250 petitions a year still filed at 103,265 dollars each, 5.7 billion dollars a year over the four cap seasons [1][2]. The department itself estimates that for more than three quarters of small filers the fee would exceed one percent of annual revenue [3]. The payers are counted at the same weight as the agencies. The two arguments therefore cancel in the balance; what firms lose by no longer hiring is the second argument against, not this one. The range runs from 2.7 to 7.8 billion dollars, as for the revenue. The Impact is as large as the revenue and cancels it, because a fee moves money without creating or destroying it.

▸ Show calculation ▾ Hide calculation
85,000 petitions a year
Cap-subject H-1B petitions a year without the fee
× × 0.65 Setting, range 0.30 to 0.88: the mirror of the fall in filings used in the arguments on the people deterred
55,250 petitions a year
Share still filed at the new price
× × 103,265 dollars the amount in the proposed rule, on top of existing fees
5.71 billion dollars a year
Fee per petition
[1]
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 2.27 to 6.66 billion euro across the range of filings
4.92 billion euro a year
In euro
× × 1.0 the agencies and the employers who pay are both counted at the weight of median income
4.92 billion euro a year
Weight of a euro on both sides
= 1 billion euro a point scale of this evaluation
4.92
Normalised Impact
Score 4.92 Impact × 5 Value × 6 Plausibility ÷ 10 = 15 of 100

Plausibility

The amount depends on one behavioural link: how many employers still file once a petition costs about thirty times what it costs now [2][3]. The counterfactual is the current weighted selection with no fee. No fee of this size has been charged on all cap petitions, so the number rests on estimated price responses and on one precedent, the 2025 charge on entries from abroad, which exempted people already in the country and so covered a narrower group [2][4]. The confounder in that precedent is that the charge arrived together with other restrictions, which may have deterred filings on their own. Reverse causation does not arise, because the fee is set by rule. That money is collected once the rule applies is not in doubt; how much is, and that doubt sits in the range. The Plausibility is just above the middle, the ceiling for a projection of a behavioural response, and the argument it is paired with carries the same figure.

Evidence: Projection · plausibility at most 6 Evidence of effect: Mechanistic · at most 6

What the deterred workers lose

3.4points

People who would have come on an H-1B visa earn far less elsewhere. A lottery among programmers at one Indian firm found that working in the United States raised pay by about 55,000 dollars a year. This evaluation counts that loss, as it does for the people removed under the detention buildout.

Value 5 · Household budgetsImpact 1.2Plausibility 5.5
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is the earnings of the workers the fee keeps out: what they would have been paid in the United States after tax against what they earn in their next best place. It is priced at the middle of the scale, like earnings anywhere on this site. Their American earnings would place them in the fourth fifth of the income distribution, so a dollar is counted at 0.75, the American weight of that band; that is set in the Impact. Whether the earnings of people who do not yet live in the country count is a choice; this site counts them in its evaluation of the detention buildout and does the same here. Taxes are left out, because they are the third argument against. The value is the middle of the scale, the level for household money, and the weight for higher incomes carries it below face value.

Impact

Averaged over the four years, about 56,500 fewer people work on cap-subject H-1B visas [1][2]. Clemens (2013) compared programmers at one Indian firm who won the 2007 and 2008 visa lotteries with those who lost, and found that working in the United States raised pay by about 55,000 dollars a year [11]. Many of the people now affected are graduates of American universities whose next best option may be Canada or Britain rather than India, so 45,000 dollars is used, in a range from 25,000 to 75,000. After the quarter that would have gone in tax, that is 33,750 dollars a year each and 1.91 billion dollars in all. At the weight of 0.75 for the fourth fifth of American incomes, the result is 1.43 billion dollars a year, in a range from 0.51 to 4.0 billion. The Impact is the largest real stream in this debate, because the gap between American and foreign pay is wide and falls on every worker who does not come.

▸ Show calculation ▾ Hide calculation
85,000 petitions a year
Cap-subject H-1B petitions a year without the fee
× × 0.35 Setting, range 0.12 to 0.70: the middle of five published estimates of cap visas no longer used at a fee of this size
29,750 fewer petitions a year
Fall caused by the fee
× × 1.9 averaged over the four years, the same count this site uses for the weighted selection rule
56,525 fewer H-1B workers on average
Yearly cohorts in the workforce at once
[2]
× × 45,000 dollars a year Setting, range 25,000 to 75,000: below the 55,000 dollars measured for Indian programmers, because many applicants have other destinations
2.54 billion dollars a year
Pay gap against the next best place
× × 0.75 the tax share belongs to the argument on public revenue
1.91 billion dollars a year
Share kept after tax
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars
1.65 billion euro a year
In euro
× × 0.75 the American band of the fourth fifth of incomes, where their American earnings would place them (until v1 the German 0.7); range 0.436 to 3.487 billion euro, factor ranges combined
1.23 billion euro a year
Weight of a euro in the fourth fifth of incomes
= 1 billion euro a point scale of this evaluation
1.23
Normalised Impact
Score 1.23 Impact × 5 Value × 5.5 Plausibility ÷ 10 = 3.4 of 100

Plausibility

The counterfactual is the same workers with an H-1B visa. Clemens (2013) uses the visa lottery among one firm's employees, so winners and losers differ only by chance, a quasi-experimental design; the measured gap is about 55,000 dollars a year in exchange-rate terms [11]. The confounder that could produce the gap without the visa would be selection into who was sent, which the random draw removes. Reverse causation is excluded by the lottery. What is transferred is the size: the study covers Indian programmers at one firm more than a decade ago, while the people the fee deters include graduates of American universities with other destinations, which is why the figure used is lower and the range wide. The Plausibility is just above the middle: the gap is measured by a lottery, and what is extrapolated is who is affected and where they go instead.

Evidence: Study · plausibility at most 7 Evidence of effect: Quasi-experimental · at most 8

Taxes these workers would have paid

2.8points

H-1B households pay more in taxes than they use in public services. Each post that is neither filled from abroad nor by an American takes that surplus out of the public purse, about 30,000 dollars a household a year.

Value 5 · Public financesImpact 1.0Plausibility 5.5
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is public revenue net of the public services the households would have used, priced at the middle of the scale like any other dollar. It is counted only for posts that no American fills; where an American takes the post, the taxes are paid either way. The fee money is a different stream and is the pair of arguments on the payment. The worker's own loss after tax is the fifth argument against, so no dollar is counted twice. The value is the middle of the scale, the level this site uses for public money.

Impact

About 39,600 of the 56,500 vacated posts are not filled by an American, as in the argument on lost output [1][2]. An estimate of the Economic Innovation Group puts the taxes a typical H-1B-headed household pays above the public services it uses at about 30,000 dollars a year, and at about 48,700 dollars for households selected by pay [2]. The lower figure is used, because the hires the fee removes sit at the lower end of the wage table; the range runs from 20,000 to 48,700 dollars. That gives 1.19 billion dollars a year, in a range from 0.41 to 3.5 billion. For comparison, the Institute for Progress, on its own assumptions of a fall of three quarters, puts the federal revenue lost at about 59 billion dollars over ten years [2]. The Impact is about a fifth of the fee money, because it counts only the households that no longer come and only what they pay above what they use.

▸ Show calculation ▾ Hide calculation
85,000 petitions a year
Cap-subject H-1B petitions a year without the fee
× × 0.35 Setting, range 0.12 to 0.70: the middle of five published estimates of cap visas no longer used at a fee of this size
29,750 fewer petitions a year
Fall caused by the fee
× × 1.9 averaged over the four years, the same count this site uses for the weighted selection rule
56,525 fewer H-1B workers on average
Yearly cohorts in the workforce at once
[2]
× × 0.7 complement of the setting in the first argument for the rule, range 0.4 to 0.9
39,568 households
Posts not taken by an American
× × 30,049 dollars a year estimate for a typical H-1B-headed household, range 20,000 to 48,697 dollars
1.19 billion dollars a year
Taxes above public services used
[2]
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 0.352 to 2.981 billion euro, factor ranges combined
1.03 billion euro a year
In euro
= 1 billion euro a point scale of this evaluation
1.03
Normalised Impact
Score 1.03 Impact × 5 Value × 5.5 Plausibility ÷ 10 = 2.8 of 100

Plausibility

The counterfactual is the same households working here under weighted selection. The surplus per household is a fiscal accounting estimate, not a measured effect: taxes on observed H-1B wages set against the average cost of the public services such a household uses [2]. The confounder is that some of these workers would have left within the horizon anyway, which the count of 1.9 cohorts in the workforce allows for. Reverse causation does not arise, because the fee is set by rule. That deterred workers pay no American taxes follows directly; the size rests on one estimate from an organisation that favours high-skilled immigration, and that is reflected in the range. The Plausibility is just above the middle, the level for a projection built on observed wages.

Evidence: Projection · plausibility at most 6 Evidence of effect: Mechanistic · at most 6

Posts that are no longer filled

1.3points

Where an employer decides a hire is not worth 103,265 dollars and no American takes the post, the work is not done here. What is lost is what the employer would have earned on the post above the wage. Since only hires worth less than the fee drop out, that margin is small for each post.

Value 6 · OutputImpact 0.3Plausibility 6
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is output: the value an employer would have produced with the worker above what it would have paid. It belongs to the class this site uses for economic systems and prosperity. The wage itself is not counted, because it is a payment inside the economy; the worker's own loss is the fifth argument against and the taxes the post would have paid are the third. Small firms, which the department expects to be hit hardest, are part of this stream and are not counted again [3]. The value sits in the middle-upper part of the scale, at the level this site uses for economic output.

Impact

Averaged over the four years, about 56,500 fewer people work on cap-subject H-1B visas, the same count as in the argument for the rule [1][2]. Seven in ten of those posts are not taken by an American, the complement of the setting there; part of such work moves abroad, about 0.4 jobs for every visa refused at the average multinational [10]. A new H-1B worker earns a median of about 103,000 dollars [6]. This site uses an employer margin of 20 percent above the wage for the average H-1B post, but here only posts the employer judged not worth the fee drop out: over a stay of about five years their margin is at most the fee spread over that stay, about 20 percent of the wage, and half of that on average. Ten percent is used, in a range from 5 to 20 percent. That is 10,300 dollars a post on about 39,600 posts, 408 million dollars a year, in a range from 0.13 to 1.4 billion. The Impact is small for the number of posts, because the fee removes exactly those hires that were worth least to the employer.

▸ Show calculation ▾ Hide calculation
85,000 petitions a year
Cap-subject H-1B petitions a year without the fee
× × 0.35 Setting, range 0.12 to 0.70: the middle of five published estimates of cap visas no longer used at a fee of this size
29,750 fewer petitions a year
Fall caused by the fee
× × 1.9 averaged over the four years, the same count this site uses for the weighted selection rule
56,525 fewer H-1B workers on average
Yearly cohorts in the workforce at once
[2]
× × 0.7 complement of the setting in the first argument for the rule, range 0.4 to 0.9
39,568 posts
Posts not taken by an American
× × 103,000 dollars median compensation of workers starting H-1B employment in fiscal year 2025
4.08 billion dollars of pay a year
Median pay of a new H-1B worker
[6]
× × 0.10 Setting, range 0.05 to 0.20: at most the fee spread over a stay of about five years, about a fifth of the wage, half of that on average
0.41 billion dollars a year
Employer margin on the posts that drop out
[1]
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 0.106 to 1.157 billion euro, factor ranges combined
0.35 billion euro a year
In euro
= 1 billion euro a point scale of this evaluation
0.35
Normalised Impact
Score 0.35 Impact × 6 Value × 6 Plausibility ÷ 10 = 1.3 of 100

Plausibility

The counterfactual is weighted selection with the cap filled, as in the season of 2026, when about 211,600 registrations competed for 85,000 places [7]. That a price about thirty times the current fees reduces hiring is a mechanism with a precedent: after the 2025 charge on new entrants from abroad, consular H-1B filings fell by more than 90 percent and employers paid it for about 700 people [3][4]. How far filings fall is carried in the range, not here. The margin of the hires that drop out is bounded by the fee itself and is not measured. The confounder in the precedent is that other restrictions arrived at the same time, which would overstate the fall; the range starts at 12 percent for that reason. Reverse causation does not arise, because the fee is set by rule. The Plausibility is just above the middle: the direction follows from the price and was observed once, and the size is carried in the range.

Evidence: Mechanism · plausibility at most 6 Evidence of effect: Mechanistic · at most 6

Fewer jobs for Americans around them

0.5points

Skilled immigrants can create work for the people around them. One study finds that each additional international graduate kept in a local labour market went with more jobs for experienced American graduates. The H-1B lottery studies find no such gain in total American employment inside the winning firms.

Value 6 · OutputImpact 0.3Plausibility 3
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is output lost where jobs of Americans depended on the presence of H-1B workers: teams that are not built and firms that do not expand. It belongs to the class this site uses for economic systems and prosperity. What is counted is production lost net of the time the worker gets back, not the wage. It is separate from the first argument for the rule, which counts Americans taking the vacated posts themselves, and from the second argument against, which counts the vacated posts. The value sits in the middle-upper part of the scale, at the level this site uses for economic output.

Impact

Averaged over the four years, about 56,500 fewer H-1B workers are employed [1][2]. In a commuting-zone comparison of the 2016 extension of practical training for international graduates, each additional high-skilled immigrant went with about 0.62 additional jobs for older American graduates, the ratio of the two measured coefficients [12]. That comparison is local and some of the jobs may have been drawn from elsewhere, so a quarter is counted as a national effect, in a range from a twentieth to a half, and the ratio itself in a range from 0.3 to 0.9; that gives about 8,800 jobs. What each represents in lost output, net of the time the worker gets back, is half a full-year wage of 75,000 dollars, the rule this site uses for jobs lost alongside foreign workers. The result is 329 million dollars a year, in a range from 0.07 to 1.5 billion. The Impact is below the output lost in the posts themselves, because only part of a local effect is counted as national.

▸ Show calculation ▾ Hide calculation
85,000 petitions a year
Cap-subject H-1B petitions a year without the fee
× × 0.35 Setting, range 0.12 to 0.70: the middle of five published estimates of cap visas no longer used at a fee of this size
29,750 fewer petitions a year
Fall caused by the fee
× × 1.9 averaged over the four years, the same count this site uses for the weighted selection rule
56,525 fewer H-1B workers on average
Yearly cohorts in the workforce at once
[2]
× × 0.62 ratio of the two measured coefficients (0.19 / 0.306), range 0.3 to 0.9
35,046 jobs, local
American jobs per high-skilled immigrant
× × 0.25 Setting, range 0.05 to 0.5
8,761 jobs
Share that is national, not drawn from elsewhere
× × 37,500 dollars half of a full-year wage of 75,000 dollars, net of the time the worker gets back, as for the practical-training fee
0.33 billion dollars a year
Lost output per job
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 0.06 to 1.334 billion euro, factor ranges combined
0.28 billion euro a year
In euro
= 1 billion euro a point scale of this evaluation
0.28
Normalised Impact
Score 0.28 Impact × 6 Value × 3 Plausibility ÷ 10 = 0.5 of 100

Plausibility

The counterfactual is the same places and firms with the H-1B workers present. Kwon, Lee and Monràs (2026) compare commuting zones exposed to the 2016 extension with unexposed ones in a synthetic difference-in-differences design and find that employment of older American graduates rose [12]. Mahajan and co-authors (2024), with the H-1B lottery, find no difference in total American-born employment between firms that won and lost, although high-productivity winners hired more American graduates and winning firms hired more workers without a degree [9]. The confounder in the local comparison is growth around universities, which matched pre-trends absorb; the firm comparison cannot see jobs created at other firms. Reverse causation is excluded by the timing of the extension and by the lottery. The claim is therefore supported in one design for one group of immigrants and not found in the other for this visa. The Plausibility is low: only part of the claim is supported, and about three tenths of the loss claimed here is likely to materialise.

Evidence: Study · plausibility at most 7 Evidence of effect: Quasi-experimental · at most 8 Placement: Partial aspect supported · plausibility 2.5–3

Open: Evidence from the H-1B lottery on employment of Americans outside the winning firms, in the same local labour market, would lift P towards 5 if it confirmed the local gain.

How robust the result is

Under the assumptions used, the proposal brings more than it costs in fewer than 1 of 100 model runs.

gains more than it costscosts more than it gains

Each dot stands for 200 of the 20,000 model runs.

Two choices outside the ranges shape this result. The earnings that deterred workers lose by working elsewhere are counted, as in this site's evaluation of the detention buildout; left out, the arguments against the rule shrink by about two fifths and still outweigh those for it many times over. And the fee is booked as a payment from employers to the agencies at equal weight, so it cancels; the department's own assumption that filings would not fall at all lies outside the range used here, and under it the rule would move money without changing anything else. The fee itself is mainly a distributional item: it carries most of the circle, but agencies and employers both count at the weight of median income, 1.0, with no band of their own, so it cancels. The verdict would turn only if a dollar at the agencies weighed about half again as much as a dollar at the employers.

The verdict does not depend on how incomes are weighted: if a euro at a low income counted somewhat less or somewhat more than assumed, it would stay “Very likely worse”. How incomes are weighted

Show the distribution of the model runs

20,000 simulated futures

Very likely worse · 0 %

In fewer than 1 of 100 simulated futures the proposal brings more than it costs.

No effect is known exactly; every argument has a range. The calculation draws a possible size for each argument from its range 20,000 times and counts how often the benefits exceed the costs.

Net effect in points per future. Left of zero the drawbacks outweigh the benefits, right of zero the benefits win. Eight in ten futures fall between −15 and −1.8 points.
Scale and assessment

Scale of this evaluation: Normalised Impact — unitless, calibrated to this topic. For comparison: one point here is worth roughly 1 billion euro per year.

Size class:medium

Effect large (−7.6) · 5 % of the effect is positive

Areas show individual argument weights. The positive share of the effect nets out money transfers between the two sides; the probability of a positive net effect is determined separately.

How we score →

Review and sources

AI evaluation · not yet approved by a human

This evaluation was produced, sourced and reviewed by an AI model; approval by a human is still pending. Figures and conclusions may still change. The review log is at the foot of the page. How review works →

Last reviewed by Claude Opus 5.5 · November 10, 2026 · 5 reviews since 11/2026 · 4 corrections · not yet approved by a human
  1. November 10, 2026AI reviewClaude Opus 5.5re-scored

    Wave 2 of the rebooking (decisions of 31.10. and 06.11.2026): the deterred workers move from the German fourth-fifth weight 0.7 to the American 0.75 (con-5), pro-2 now mixes the American bands; very likely worse unchanged, P(D > 0) 0.000.

  2. November 10, 2026AI reviewClaude Opus 5.5revised

    Kernfrage von Julian Ohm im Chat freigegeben; Zahlen nicht geprüft.

  3. November 9, 2026AI reviewClaude Opus 5.5revised

    Key question and its answer added to the page texts (draft); numbers unchanged.

  4. November 1, 2026AI reviewClaude Opus 5.5revised

    Page texts (layout G) added before first publication.

  5. November 1, 2026AI reviewClaude Opus 5.5First evaluation

    KI-Entwurf aus der Themen-Automatisierung übernommen; Anzeige-Maßstab gesetzt, Rechenweg und P-Deckel an validate angepasst.

Every evaluation is reviewed again and again: an AI model looks for errors, gaps and new developments and corrects them. Rounds approved by a member of the editorial team carry their name. Each round is listed here with its date, and earlier versions stay readable.How we review →

Show all 13 sourcesHide sources
  1. Federal Register (Department of Homeland Security): Fee for Certain H-1B Petitions, proposed rule, 91 FR 54817, 25 August 2026. federalregister.gov
  2. Institute for Progress: Explaining the proposed 103k fee on H-1Bs (2026). ifp.org
  3. FWD.us: H-1B fee of 103,265 dollars: policy brief (2026). fwd.us
  4. Forbes (Stuart Anderson): Judge strikes down latest version of 100,000-dollar H-1B immigration fee (1 October 2026). forbes.com
  5. University of Colorado Boulder, International Student and Scholar Services: 100,000-dollar H-1B fee: current status (7 October 2026). colorado.edu
  6. U.S. Citizenship and Immigration Services: Characteristics of H-1B Specialty Occupation Workers, Fiscal Year 2025. uscis.gov
  7. SHRM: USCIS reaches FY2027 H-1B visa cap (preliminary agency figures, May 2026). shrm.org
  8. National Bureau of Economic Research (Doran, Gelber and Isen): The effects of high-skilled immigration policy on firms: evidence from visa lotteries, Working Paper 20668 (published 2022). nber.org
  9. Federal Reserve Bank of Richmond (Nicolas Morales): How do H-1B lotteries affect firms and workers? Economic Brief 24-15 on Mahajan, Morales, Shih, Chen and Brinatti (2024). richmondfed.org
  10. National Bureau of Economic Research (Britta Glennon): How do restrictions on high-skilled immigration affect offshoring? Evidence from the H-1B program, Working Paper 27538. nber.org
  11. Foreign Policy: For computer programmers, it's all about location: Clemens on randomised H-1B processing (21 May 2013). foreignpolicy.com
  12. Universitat Pompeu Fabra (Kwon, Lee and Monràs): The effect of retaining high-skilled international graduates: evidence from the STEM OPT extension, Working Paper 1940 (2026). econ-papers.upf.edu
  13. Forbes (Stuart Anderson): DHS immigration rule proposes a 103,265-dollar tax on new H-1B visa holders (25 August 2026). forbes.com

More debates in this area

All debates in Immigration and Border Enforcement →

Ask the bot

The bot knows the evaluation of this debate and explains it and the method. It answers only from Argupedia's content and does no research of its own.

Before you ask

  • The answers are written by an AI model (Claude Haiku by Anthropic). They may be wrong or incomplete. The content of the pages is authoritative.
  • The bot gives no legal, financial, medical or voting advice.
  • What you type is sent to Anthropic in the USA. Do not enter names, contact details or other personal information.

The terms of use (section 9a) and the privacy notice for the chat apply.