Immigration and Border Enforcement

OPT Work Permit Fee

Colleges would pay 70,000 dollars before an international graduate can start the work permit that most of them use after their degree.

8 arguments 18 Sources
  • 194,554 students received an OPT work permit in 2024 [1]
  • 142,000 new international graduate students enrolled in 2024/25, 25,000 fewer than a year earlier [5][6]
  • 10,000 OPT participants worked at highly suspect employers, by the department's count [1]

On 8 October 2026 the Department of Homeland Security proposed a fee on colleges for each recommendation of Optional Practical Training, the work permit most international graduates use [1]. The administration says the fee fights fraud and protects American workers, while colleges and international educators say it would drive students away [12][15]. Comments close on 9 November 2026, and the fee would apply 60 days after a final rule.

The proposal

Colleges would pay 70,000 dollars before an international graduate can start the work permit that most of them use after their degree.

The key question

Does the fee open more first jobs to American graduates without losing the work international graduates do here?

The debate

The department argues that the program has been used to get around the H-1B cap and that thousands of participants worked for suspect or sham employers. Colleges and international educators answer that the fee would push students to other countries and that these graduates create jobs rather than take them. Researchers add that a fee of this size would cut the use of the program by a third or more. The change comes by rule, so Congress does not vote on it.

  • “DHS believes that this fee is necessary to place the United States' interests first and protect American workers. The fee is necessary to ensure that schools, students, and employers participate in OPT for its intended purpose instead of a means to circumvent the H-1B visa program.”

    argues for: Department of Homeland Security, Trump administration Proposed rule, Federal Register, 8 October 2026, page 64572
  • “OPT enables international students to gain the same hands-on experience as their domestic peers while filling labor shortages in high-demand STEM fields and generating jobs for U.S. workers on U.S. soil. The research is clear: international students don’t take opportunities from Americans, they create them.”

    argues against: Fanta Aw, NAFSA: Association of International Educators Statement on the proposed rule, 7 October 2026

The analysis

OPT Work Permit Fee

37% of the weight is for, 63% 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 open more first jobs to American graduates without losing the work international graduates do here?

No on both counts. One study finds some entry-level posts going to American graduates, but a better-identified one finds the opposite. The fee deters many international graduates, and most of the work they would have done is not taken by anyone else. Jobs for experienced Americans around them and the tuition margin colleges earn shrink as well.

On the current analysis, the fee on work permits for international graduates is a poor proposal.

Very likely worse

Very likely worse: Under the assumptions used, the proposal brings more than it costs in 6 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, for much the reason its critics give: the graduates it deters add work here rather than taking it. The fee itself is money changing hands from colleges, students and employers to the Treasury, and neither side does more good with it, so it cancels. What decides is the work that no longer gets done, together with the jobs for experienced Americans around these graduates and the margin colleges lose. The department's two arguments, first jobs for American graduates and fewer sham employers, point the right way but weigh little, and the better-identified study contradicts the first. The graduates' own lost earnings are left out for now; counting them would widen the gap.

In detail

All arguments

Arguments for3 arguments

Fee money for the Treasury

9.5points

Every OPT recommendation that still goes ahead brings 70,000 dollars into the Treasury, and every extension 30,000 more. Even if most students stop using the permit, those who remain would pay about 3.7 billion dollars a year. The same money leaves colleges, students and employers, which is the first argument against.

Value 5 · Public financesImpact 3.2Plausibility 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 agency that collects the fee does not keep it: the rule sends it to the Treasury as general receipts [1][13], so it pays for nothing in particular and 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 colleges and passed on to students and 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 rule charges 70,000 dollars for a first OPT recommendation and 30,000 dollars for each later one, including the 24-month extension for science and engineering graduates [1]. About 150,000 first recommendations a year are taken as the starting point, between the 108,000 graduates a year that one model counts and the 194,554 permits granted in 2024 [1][3]. International enrolment is already falling, so 85 percent of that is used for the years ahead [5][7]. Of those, 40 percent are assumed still to go ahead at the new price, the mirror of the fall in use in the arguments against [4]; that is 51,000 a year. Each brings 70,000 dollars plus 30,000 more for the 42.5 percent who later take the extension, 82,750 dollars on average, or 4.2 billion dollars a year. The first year counts half because the rule applies only 60 days after it is final, which gives 3.7 billion dollars a year over the four years. The department's own figure, which assumes almost no fall in use, is 8.4 to 16.5 billion dollars a year [2]. The range here runs from 1.2 to 8.9 billion dollars, depending above all on how many 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
150,000 recommendations a year
First OPT recommendations a year without the fee
× × 0.85 Setting, range 0.75 to 0.95, the same as in every argument here
127,500 recommendations a year
Share left after the fall in enrolment already under way
× × 0.4 Setting, range 0.2 to 0.67: the mirror of the fall in OPT use used in the arguments against
51,000 recommendations a year
Share that still goes ahead at the new price
× × (70,000 + 0.425 × 30,000) dollars 70,000 dollars for the first recommendation, 30,000 for the extension, which about 0.425 of first-time users later take
4.22 billion dollars a year
Fee per graduate
[1]
× × 0.875 the rule applies 60 days after it is final, so the first of the four years counts half
3.69 billion dollars a year
First year only half in force
[1]
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 1.01 to 7.75 billion euro with the settings at their bounds
3.18 billion euro a year
In euro
× × 1.0 Treasury, colleges and the students and employers they pass it to are all counted at the weight of median income
3.18 billion euro a year
Weight of a euro on both sides
÷ 1 billion euro a point scale of this evaluation
3.18
Normalised Impact
Score 3.18 Impact × 5 Value × 6 Plausibility ÷ 10 = 9.5 of 100

Plausibility

The amount depends on one behavioural link: how many recommendations still go ahead once a college must pay before it can enter them. The counterfactual is the current programme with no fee. No fee of this size has ever been charged on a student work permit, so the number rests on estimated price responses rather than on a precedent [4]. The confounder is the fall in international enrolment that was under way before the rule, which is handled by a separate trend factor [5][7]. 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 first jobs go to American graduates

0.81points

The rule's stated purpose is to protect American workers. With fewer international graduates on OPT, some entry-level posts go to an American instead. One study finds this for recent master's graduates in the same field; a better-identified one finds the opposite.

Value 5 · Household budgetsImpact 0.5Plausibility 3
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is earnings in the hands of young American graduates who get a post they would not otherwise have got. What is counted is not the salary, since the post exists either way, but the gap between that post and the next best one open to them. It is money and is priced at the middle of the scale. That these are early-career earners below the middle of American incomes raises the weight of each dollar, and that is set in the Impact. What an employer loses when a post goes unfilled is not counted here; it 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 99,000 fewer international graduates work on OPT than without the fee, the same figure the arguments against use [1][3][4]. Whether an American takes each of the vacated posts is the open part: the employer may hire another foreign worker, move the work abroad, or leave it undone. Three in ten is used here, in a range from one to six in ten, the setting this site uses for the shift in entry-level H-1B selections; that gives about 29,700 posts. Each is worth the gap between this post and the graduate's next best option, set at 17,400 dollars a year, again as for the H-1B rule. These are early-career earners between the second fifth and the middle of American incomes, so a dollar is counted at 1.2. The result is 621 million dollars a year, in a range from 0.06 to 2.8 billion. The Impact is about a third of the output the arguments against see lost, because only some of the vacated posts are filled from here.

▸ Show calculation ▾ Hide calculation
150,000 recommendations a year
First OPT recommendations a year without the fee
× × 0.85 Setting, range 0.75 to 0.95: new international graduate students fell 15 percent in 2024/25 and graduate enrolment is projected to fall again in 2026
127,500 recommendations a year
Share left after the fall in enrolment already under way
× × 1.85 194,554 first permits and 165,524 people on the 24-month extension in 2024
235,875 graduates on OPT at any time
Years each graduate spends on OPT
[1]
× × 0.6 Setting, range 0.33 to 0.8: OPT starts fall by at least a third in every scenario of a 100,000-dollar fee and by about four-fifths where colleges or employers pay
141,525 fewer graduates on OPT
Fall caused by the fee
× × 0.7 Setting, range 0.6 to 0.8: the rule applies 60 days after it is final and the extension cohort empties over two years; averaged over the horizon, about 99,000
99,068 fewer graduates on OPT on average
Build-up over the four years
[1]
× × 0.3 Setting, range 0.1 to 0.6, the same as this site uses for entry-level posts in the H-1B selection rule
29,720 posts
Posts taken by an American graduate instead
× × 15,000 euro a year Setting, the same as for the H-1B selection rule: the post exists either way, only the step up counts
0.45 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 0.05 to 2.37 billion euro with the settings at their bounds
0.54 billion euro a year
Weight of a euro for early-career earners
÷ 1 billion euro a point scale of this evaluation
0.54
Normalised Impact
Score 0.54 Impact × 5 Value × 3 Plausibility ÷ 10 = 0.8 of 100

Plausibility

Whether international graduates displace young Americans or work alongside them has been measured, and the results disagree. The counterfactual is the same labour markets with OPT available as before. Demirci (2020) compares fields and degree levels more and less exposed to the 2008 OPT extension, a controlled comparison, and finds more international graduates went with lower employment among recently graduated American master's holders in the same field [9]; the department also cites a 2017 study by which wages and employment of American computer scientists would have been several percent higher without foreign workers [1]. Kwon, Lee and Monràs (2026) compare commuting zones heavily exposed to the 2016 extension with unexposed ones in a synthetic difference-in-differences design and find employment of young American graduates rose rather than fell [8]. The confounder in the field comparison is that fields attracting international students may have cooled for other reasons, which the commuting-zone design handles better; that design in turn measures local effects, part of which may be jobs drawn from elsewhere. Reverse causation is addressed in both by the timing of a policy change. The claim is therefore supported for one group in one design and contradicted in another. The Plausibility is low: only part of the claim is supported, and about three tenths of the gain claimed here is likely to materialise.

Evidence: Study · plausibility at most 7 Evidence of effect: Controlled · at most 7 Placement: Partial aspect supported · plausibility 2.5–3

Open: A study of employment among recent American graduates in the fields that lose OPT workers once the fee applies would lift P towards 5 if it confirmed substitution.

Fewer students working for sham employers

0.033points

The department cites more than 10,000 OPT participants at highly suspect employers, and a news investigation found over 4,000 who used shell companies to fake employment. A 70,000-dollar fee prices most of those arrangements out. The gain is real but small next to everything else in this debate.

Value 6 · Law enforcementImpact 0.01Plausibility 5.5
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is the enforcement of a condition: OPT is granted for real work in the field the student studied, and a permit used to stay in the country while working for a shell company, or not working at all, breaches it. That is priced with the working of legal and administrative systems, one step above money. It is not counted as a harm to American workers, which is the subject of a separate argument, nor as a security risk, for which the rule gives no evidence [1]. What is counted is the breach removed. The value sits in the middle-upper part of the scale, at the level this site uses for the working of public systems.

Impact

The department cites more than 10,000 F-1 students on OPT at highly suspect employers, and a news investigation found over 4,000 who used shell companies to fake employment records [1][12]; one analysis puts the first figure at about 3 percent of participants [3]. Those using a sham employer have less reason to pay 70,000 dollars than those with a real job offer, so eight in ten are assumed to drop out, in a range from six to ten in ten. With the same build-up over four years as in the other arguments, that is about 5,600 fewer such cases on average. No price exists for a breached permit; this evaluation sets 2,300 dollars a year for each, in a range from 580 to 5,800, a setting rather than a measurement. The result is 13 million dollars a year, in a range from 2 to 46 million. The Impact is the smallest in this debate: the abuse the rule names is a few percent of the programme, and its price is a setting.

▸ Show calculation ▾ Hide calculation
10,000 people
OPT participants at highly suspect employers
× × 0.8 Setting, range 0.6 to 1.0: a sham arrangement has less reason to pay 70,000 dollars than a real job offer
8,000 people
Share priced out by the fee
× × 0.7 Setting, range 0.6 to 0.8, the same as in every argument here
5,600 cases a year on average
Build-up over the four years
× × 2,000 euro a year Setting, range 500 to 5,000 euro: no price exists for a permit used against its conditions; range 0.002 to 0.04 billion euro
0.01 billion euro a year
Value of a breached permit removed
÷ 1 billion euro a point scale of this evaluation
0.01
Normalised Impact
Score 0.01 Impact × 6 Value × 5.5 Plausibility ÷ 10 = 0.0 of 100

Plausibility

The chain is short: a large fee falls on every recommendation, and arrangements that exist only to keep a student in status are the least able to carry it. The counterfactual is the current programme with the same oversight, about 60 field representatives for some 6,000 schools. Nothing measures how sham arrangements respond to a price, so the chain is mechanistic, with no study or precedent behind it. The confounder is that some employers counted as suspect are legitimate small firms, which would make the number smaller without changing its direction. Reverse causation does not arise. That the fee removes such cases is close to certain; how many there were and what each is worth sits in the range. The Plausibility is just above the middle: the chain is closed, but nothing has measured it.

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

Arguments against5 arguments

Colleges, students and employers pay it

9.5points

The rule makes the college pay before it can enter the recommendation, and lets it pass the cost on. One model expects colleges to carry a little over a quarter and students most of the rest. Whoever pays, about 3.7 billion dollars a year leave their budgets for the Treasury's.

Value 5 · Household budgetsImpact 3.2Plausibility 6
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is the same sum seen from the paying side: money that colleges, students and employers no longer have. It is priced at the middle of the scale, as money always is on this site. The rule names the college as the payer and leaves it free to pass the charge on [1][18]; one model expects colleges to absorb a little over a quarter and to pass most of the rest to students [3]. Who carries it changes the weight a dollar carries, and that is set in the Impact. The value is the middle of the scale, the level this site uses for money in private and public budgets alike.

Impact

The amount is the one the revenue argument derives: about 51,000 first recommendations a year still going ahead at 70,000 dollars each, plus 30,000 dollars for the share who later take the extension, 3.7 billion dollars a year over the four years once the half first year is allowed for [1][3][4]. The payers are counted at the same weight as the Treasury. Colleges are institutions, and the students who carry most of the charge earn around the middle of American incomes once they work, so neither side is weighted up or down. The two arguments therefore cancel in the balance; they are shown so that the size of the payment is visible. The range runs from 1.2 to 8.9 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
150,000 recommendations a year
First OPT recommendations a year without the fee
× × 0.85 Setting, range 0.75 to 0.95, the same as in every argument here
127,500 recommendations a year
Share left after the fall in enrolment already under way
× × 0.4 Setting, range 0.2 to 0.67: the mirror of the fall in OPT use used in the arguments against
51,000 recommendations a year
Share that still goes ahead at the new price
× × (70,000 + 0.425 × 30,000) dollars 70,000 dollars for the first recommendation, 30,000 for the extension, which about 0.425 of first-time users later take
4.22 billion dollars a year
Fee per graduate
[1]
× × 0.875 the rule applies 60 days after it is final, so the first of the four years counts half
3.69 billion dollars a year
First year only half in force
[1]
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 1.01 to 7.75 billion euro with the settings at their bounds
3.18 billion euro a year
In euro
× × 1.0 Treasury, colleges and the students and employers they pass it to are all counted at the weight of median income
3.18 billion euro a year
Weight of a euro on both sides
÷ 1 billion euro a point scale of this evaluation
3.18
Normalised Impact
Score 3.18 Impact × 5 Value × 6 Plausibility ÷ 10 = 9.5 of 100

Plausibility

The amount depends on the same behavioural link as the revenue: how many recommendations still go ahead once a college must pay. The counterfactual is the current programme with no fee. The number rests on estimated price responses rather than on a precedent, since no fee of this size has been charged on a student work permit before [4]. The confounder is the fall in enrolment already under way, handled by the trend factor [5][7]. Reverse causation does not arise. Who ends up paying, college or student, does not change the total. The Plausibility is just above the middle and equal to that of the revenue it mirrors.

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

Work that no longer gets done here

5.2points

A fee of this size cuts OPT starts by between a third and four-fifths, depending on who pays. Most of the posts these graduates would have filled are not taken by someone else. What the country loses is the employer's margin on that work and the income tax the graduates would have paid.

Value 5.5 · Output and taxImpact 1.6Plausibility 6
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

Two streams come from the same lost post: the margin an employer earns on the work above the wage, which is output, and the income tax the graduate pays, which is public money. Output belongs to the class this site uses for economic prosperity, one step above money; tax is money. Together they are priced half a step above money, because the margin is the larger part. The graduate's own earnings are not counted here; whether and how to count them is the open argument that follows. The value sits between money and economic output, closer to output because the margin is the larger part.

Impact

Averaged over the four years, about 99,000 fewer international graduates work on OPT than without the fee. That comes from about 150,000 first recommendations a year, 85 percent of it after the fall in enrolment already under way, an average of 1.85 years on OPT once the extension is included, a fall of 60 percent caused by the fee, and a build-up over the first years [1][3][4][5]. Seven in ten of those posts are not taken by an American graduate, the complement of the setting in the first argument for the rule. The graduates would earn about 75,000 dollars, below the 86,908 dollars offered to former students sponsored for permanent residence, because these are first jobs [11]. The employer's margin above the wage is set at 20 percent, as for the H-1B rule, and federal and state income tax at 15 percent, since OPT wages carry no payroll tax [3]. That is 26,250 dollars a post and 1.8 billion dollars a year. The range runs from 0.10 to 10.7 billion dollars, because the fall in use, the margin and the wage are all uncertain. The Impact is the largest net effect in this debate, because the fee takes people out of productive work rather than moving them between posts.

▸ Show calculation ▾ Hide calculation
150,000 recommendations a year
First OPT recommendations a year without the fee
× × 0.85 Setting, range 0.75 to 0.95: new international graduate students fell 15 percent in 2024/25 and graduate enrolment is projected to fall again in 2026
127,500 recommendations a year
Share left after the fall in enrolment already under way
× × 1.85 194,554 first permits and 165,524 people on the 24-month extension in 2024
235,875 graduates on OPT at any time
Years each graduate spends on OPT
[1]
× × 0.6 Setting, range 0.33 to 0.8: OPT starts fall by at least a third in every scenario of a 100,000-dollar fee and by about four-fifths where colleges or employers pay
141,525 fewer graduates on OPT
Fall caused by the fee
× × 0.7 Setting, range 0.6 to 0.8: the rule applies 60 days after it is final and the extension cohort empties over two years; averaged over the horizon, about 99,000; range 30,000 to 219,000 with all settings at their bounds
99,068 fewer graduates on OPT on average
Build-up over the four years
[1]
× × 0.7 the complement of the three in ten used in the first argument for the rule
69,347 posts
Posts not taken by an American graduate
× × 75,000 dollars Setting, range 60,000 to 90,000: below the 86,908 dollars offered to former students sponsored for permanent residence, because these are first jobs
5.2 billion dollars a year
Wage of a graduate on OPT
× × (0.20 + 0.15) Setting: margin 20 percent as for the H-1B rule (5 to 40), federal and state income tax 15 percent (10 to 20), no payroll tax on OPT wages
1.82 billion dollars a year
Employer margin and income tax on that wage
[3]
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 0.09 to 9.19 billion euro with all settings at their bounds
1.57 billion euro a year
In euro
÷ 1 billion euro a point scale of this evaluation
1.57
Normalised Impact
Score 1.57 Impact × 5.5 Value × 6 Plausibility ÷ 10 = 5.2 of 100

Plausibility

The counterfactual is the current programme without a fee. That OPT graduates add to employment rather than taking existing posts is the finding of Kwon, Lee and Monràs (2026), who compare commuting zones heavily exposed to the 2016 extension with unexposed zones in a synthetic difference-in-differences design without differential pre-trends: the extension raised high-skilled immigrant employment by about 6 percent of the existing stock, and native employment rose with it [8]. The size of the fall in use is projected from estimated price responses for a 100,000-dollar version of the fee, with OPT starts falling by at least a third in every scenario [4]. The confounder in the commuting-zone comparison is that exposed zones host universities with their own growth path, which the matched pre-2016 trends absorb. Reverse causation is excluded by the timing of the policy change. What is transferred is an expansion measured in 2016 to a contraction of a different size ten years later, which costs a point. The Plausibility is above the middle: the direction is measured in a design that can carry it, and the size is a projection.

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

Fewer jobs for experienced Americans

1.7points

Where the 2016 OPT extension brought more international graduates, employment of older American graduates rose as well, and more small firms appeared. Teams need junior and senior people together. Removing many of the junior ones removes some of the senior work too.

Value 6 · OutputImpact 0.5Plausibility 5.5
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is output that does not happen when experienced American workers lose work that existed because of the people around them. It is priced with economic prosperity, as in this site's evaluation of the detention buildout, where the same kind of loss is counted for workers alongside removed immigrants. What is counted is the production lost net of the time the worker gets back, not the wage, which is a payment inside the economy. Young American graduates are not part of this argument; what the fee does for them is the first argument for the rule. 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 99,000 fewer international graduates work on OPT [1][3][4]. In the commuting-zone comparison of the 2016 extension, each additional high-skilled immigrant went with about 0.62 additional jobs for older American graduates, the ratio of the two measured coefficients [8]. That comparison is local, and some of the jobs may have been drawn from other places, so a quarter is counted as a national effect, in a range from none to half; that gives about 15,400 jobs. What each represents in lost output, net of the time the worker gets back, is half a full-year wage, the same rule this site uses for jobs lost alongside removed workers: half of 75,000 dollars. The result is 575 million dollars a year, in a range from zero to 3.0 billion. The Impact is about a third of the output lost in the posts themselves, because only part of the local effect is counted as national.

▸ Show calculation ▾ Hide calculation
150,000 recommendations a year
First OPT recommendations a year without the fee
× × 0.85 Setting, range 0.75 to 0.95: new international graduate students fell 15 percent in 2024/25 and graduate enrolment is projected to fall again in 2026
127,500 recommendations a year
Share left after the fall in enrolment already under way
× × 1.85 194,554 first permits and 165,524 people on the 24-month extension in 2024
235,875 graduates on OPT at any time
Years each graduate spends on OPT
[1]
× × 0.6 Setting, range 0.33 to 0.8: OPT starts fall by at least a third in every scenario of a 100,000-dollar fee and by about four-fifths where colleges or employers pay
141,525 fewer graduates on OPT
Fall caused by the fee
× × 0.7 Setting, range 0.6 to 0.8: the rule applies 60 days after it is final and the extension cohort empties over two years; averaged over the horizon, about 99,000
99,068 fewer graduates on OPT on average
Build-up over the four years
[1]
× × 0.62 ratio of the two measured coefficients in the commuting-zone study, 0.19 native jobs against 0.31 high-skilled immigrants per 100 residents
61,422 jobs locally
Jobs for older American graduates per international graduate
[8]
× × 0.25 Setting, range 0 to 0.5: a local comparison partly measures jobs drawn from other places
15,355 jobs
Share that is a national effect
× × 37,500 dollars half a full-year wage of 75,000 dollars, the same rule this site uses for jobs lost alongside removed workers
0.58 billion dollars a year
Output lost per job, net of the time the worker gets back
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 0 to 2.64 billion euro with the settings at their bounds
0.5 billion euro a year
In euro
÷ 1 billion euro a point scale of this evaluation
0.5
Normalised Impact
Score 0.5 Impact × 6 Value × 5.5 Plausibility ÷ 10 = 1.7 of 100

Plausibility

The counterfactual is commuting zones without exposure to the 2016 extension. Kwon, Lee and Monràs (2026) use a synthetic difference-in-differences design that matches pre-2016 trends and show no differential pre-trend; employment of older native graduates rose by 0.19 per 100 residents against 0.31 more high-skilled immigrants, and the number of small and mid-sized firms grew [8]. Demirci (2020) finds a related pattern in a different design: more international graduates in a field went with higher earnings for experienced American master's holders [9]. The confounder that would produce the association without the effect is local growth around universities, which the matched trends absorb. Reverse causation is excluded by the timing of the extension. The study is a working paper and measures an expansion, not a contraction; that the effect runs back when graduates leave is assumed, and the national share is a setting carried in the range. The Plausibility is above the middle: the direction is measured twice in two designs, and what is extrapolated is the scale.

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

Colleges lose what these students pay above cost

1.1points

International graduate students usually pay full tuition, and colleges use the margin to fund other places, including for American students. The fee makes an American degree less attractive, so fewer come. One estimate puts the margin at 13,700 dollars a year for each master's student.

Value 5 · University budgetsImpact 0.4Plausibility 5.5
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The stream is the margin colleges earn on international students above what teaching them costs, money that public and private universities use to fund other places, including for American students [16]. It is priced at the middle of the scale as money. What the margin pays for, more places for domestic students, is not counted a second time; the money is counted once, where it arises. The students' own loss from not studying here is not part of this argument. The value is the middle of the scale, the level this site uses for money in budgets.

Impact

About 142,000 new international graduate students enrolled in 2024/25, 25,000 fewer than a year earlier [5][6]. Graduate enrolment is projected to fall again in autumn 2026 [7], so 85 percent of that is used, the same trend factor as for OPT. Analyses of a 100,000-dollar version of the fee expect master's enrolment to fall by at least 35 percent where colleges pay [17][4]; for a first fee of 70,000 dollars a fall of 25 percent is used, in a range from 5 to 45 percent. That is about 30,000 fewer new students a year, each enrolled for about 1.7 years, building up over the horizon to about 33,000 fewer students on average. Each brings colleges a margin of about 13,700 dollars a year [3]. The result is 457 million dollars a year, in a range from 0.06 to 1.14 billion. Undergraduates, for whom OPT matters less, are not counted. The Impact is a quarter of the output lost in the posts, because only the margin is counted, not the tuition.

▸ Show calculation ▾ Hide calculation
142,000 students a year
New international graduate students a year
× × 0.85 Setting, range 0.75 to 0.95, the same as in every argument here
120,700 students a year
Share left after the fall in enrolment already under way
[7]
× × 0.25 Setting, range 0.05 to 0.45: at least 35 percent fewer master's students are expected where colleges pay a 100,000-dollar version of the fee
30,175 fewer new students a year
Fall caused by the fee
× × 1.7 most are in one- to two-year master's programmes
51,298 fewer students at any time
Years each is enrolled
× × 0.65 Setting, range 0.5 to 0.8: admissions respond from the next intake on
33,343 fewer students on average
Build-up over the four years
× × 13,700 dollars a year net of what teaching the student costs
0.46 billion dollars a year
Margin a college earns per international master's student
[3]
÷ ÷ 1.16 exchange rate used throughout this evaluation, 1 euro = 1.16 dollars; range 0.05 to 0.98 billion euro with the settings at their bounds
0.39 billion euro a year
In euro
÷ 1 billion euro a point scale of this evaluation
0.39
Normalised Impact
Score 0.39 Impact × 5 Value × 5.5 Plausibility ÷ 10 = 1.1 of 100

Plausibility

The counterfactual is the current programme without a fee. That international students choose where to study partly by the chance to work afterwards is supported by Demirci (2019), who finds that the 2008 extension of OPT for STEM graduates raised the likelihood that they stayed and started work [10], and by a 2026 survey in which 70 percent of STEM-eligible students said they would not have enrolled under the fee [3]. The size of the fall is projected from estimated price responses rather than measured [4][17]. The confounder is the fall in enrolment already under way for other reasons, handled by the trend factor. Reverse causation does not arise, since the fee is set by rule. That the margin funds places for domestic students rests on Shih (2017), who uses the swing in foreign graduate enrolment between 1995 and 2005 and finds domestic enrolment rose with it [16]. The Plausibility is just above the middle: the direction is well supported, and the size is a projection.

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

Graduates lose their first job here

0points

The people most directly affected are the graduates themselves. A student who cannot start OPT goes home or to another country and usually earns less there. This site has not yet decided how to count what visa holders themselves gain or lose, so the argument is recorded but not scored.

Value 5 · Household budgetsImpact 0Plausibility 5
▸ Show reasoning & sources ▾ Hide reasoning & sources

Value

The good is the graduate's own livelihood: the earnings from a first job in the field they studied, in the country where they studied it. It is money and would be priced at the middle of the scale, with the weight of a dollar set by the graduate's income. What is unresolved is not the value but whether and how this site counts the welfare of people who are not residents when a rule decides whether they may stay [1]. The evaluations of weighted H-1B selection and of the detention buildout leave the same question open, and this one follows them. The value would be the middle of the scale, but the stream is not yet counted.

Impact

The fee removes about 99,000 graduates from OPT on average, the same figure as in the argument on lost work [1][3][4]. Each would otherwise have earned about 75,000 dollars a year [11]. Their next-best option is a job at home or in another country that admits graduates; at 60 percent of the American wage, and after the income tax that stays with the Treasury, the loss is about 18,750 dollars a year each. A survey of students shows how much they value the option: the share of STEM-eligible students likely to stay falls from 80 to 48 percent when only colleges may pay the fee [3]. The argument is recorded and not counted. The Impact is open: the number of people is estimated, the rule for counting what they lose is not decided.

Plausibility

The fall in OPT use rests on the same estimates as the argument on lost work [4]. That graduates earn more in an American job than in their next-best option is not seriously in doubt; Demirci (2019) shows that longer OPT terms raised the share of STEM graduates who stayed and started work [10]. How much more they earn is the open quantity, and it depends on where they go instead. Once a rule for counting it exists, this argument is scored on the same people. The Plausibility is not yet scored, because the stream is not yet counted.

Recorded, not yet scored: Recorded, not yet scored: about 99,000 fewer graduates on OPT on average, each earning less abroad than here. If their next-best job paid 60 percent of the 75,000 dollars used in this evaluation and the income tax is left with the Treasury, each would lose about 18,750 dollars a year, about 1.9 billion dollars a year in all, the largest stream against the rule after the paired fee. The evaluations of weighted H-1B selection and of the detention buildout leave the same stream open. Once this site decides how what a visa holder gains or loses is counted, this argument is scored on the same people.

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

How robust the result is

Under the assumptions used, the proposal brings more than it costs in 6 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 graduates' own lost earnings are recorded but not counted, as in this site's other evaluations of visa rules; counted at the earnings gap described in that argument, they would add about 1.9 billion dollars a year against the rule. And the fee is booked as a payment from colleges, students and employers to the Treasury at equal weight, so it cancels; if what foreign students pay were left out as well, the part they carry, roughly 2.6 billion dollars a year, would become a gain for the rule and bring the two sides close to even.

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 · 6 %

In 6 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 −16 and −0.7 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 (−7.1) · 10 % 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 5.5revised

    Wave 2 of the rebooking (decisions of 31.10. and 06.11.2026): pro-2's mixed weight 1.2 now named on the American bands (second fifth 1.4, median 1.0); the fee pair stands at weight 1.0 on both legs and needs no band; no number changed.

  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 18 sourcesHide sources
  1. Federal Register (Department of Homeland Security): Optional Practical Training Fees, proposed rule, 91 FR 64566, 8 October 2026. federalregister.gov
  2. InternationalStudent.us: Proposed F-1 OPT fee rule: the department's estimate of annual fee payments. internationalstudent.us
  3. Institute for Progress: DHS's $100k fee on STEM OPT would break high-skilled immigration (October 2026). ifp.org
  4. IZA Institute of Labor Economics (Michael Clemens): Taxes on international students, Discussion Paper 18933 (2026). iza.org
  5. Institute of International Education: Open Doors 2025 press release. iie.org
  6. ApplyBoard: New international student enrolment down in the US in 2024/25, but participation in OPT continued to grow. applyboard.com
  7. NAFSA: Association of International Educators: Fall 2026 international enrollment outlook. nafsa.org
  8. 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
  9. Canadian Journal of Economics (Murat Demirci): International students and labour market outcomes of US-born workers (2020). ca.vlex.com
  10. Economic Inquiry (Murat Demirci): Transition of international STEM students to the U.S. labor market: the role of visa policy (2019). ideas.repec.org
  11. Cato Institute (David J. Bier): The facts about Optional Practical Training for foreign students. cato.org
  12. Inside Higher Ed: Trump administration expects institutions to pay 70K OPT fee (8 October 2026). insidehighered.com
  13. Forbes (Stuart Anderson): DHS immigration rule proposes 100,000 dollars in practical training fees (8 October 2026). forbes.com
  14. American Action Forum: Week in Regulation: DHS kicks off October with costly proposed rules (13 October 2026). americanactionforum.org
  15. American Immigration Council: Trump administration proposes 70,000-dollar fee on international students seeking post-graduate employment (9 October 2026). americanimmigrationcouncil.org
  16. Journal of Public Economics (Kevin Shih): Do international students crowd-out or cross-subsidize Americans in higher education? (2017). ideas.repec.org
  17. Economic Innovation Group (Peak and He): A 100,000-dollar tax on foreign students (September 2026). agglomerations.eig.org
  18. Wolfsdorf Rosenthal: Fifteen things to know about DHS's proposed OPT fees. wolfsdorf.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.