The H-1B fee at the center of the Massachusetts ruling was never struck down because a judge disliked the immigration policy behind it. It was struck down because of what it was, structurally, as a matter of constitutional classification: a tax, imposed by the executive, that the Constitution permits only Congress to levy. That single classification did all the work. Once the court concluded that the $100,000 charge on certain new petitions functioned as a tax rather than as a regulatory fee, the outcome followed almost mechanically, because the power to tax sits with the legislature and a president cannot manufacture it by proclamation. Understanding precisely why a payment can be a tax even when the government insists it is a fee, and the test that decides the question, is the difference between reading the headline and understanding the law.

This article is about that test and nothing else. It is not a recap of who won and who lost, which the June 8 ruling explained in full is the place to start, and it is not a walk through the reasoning chain from the tax finding to the nationwide remedy, which has its own detailed treatment. It is a constitutional and statutory deep dive into one question that controlled everything else. When is a government charge a tax, and when is it a fee? The distinction sounds technical, even semantic, but it is one of the oldest load-bearing lines in American public law, and it is the reason a payment labeled a fee in a presidential proclamation could be vacated as an unlawful tax. A reader who finishes this piece should be able to explain the functional test in plain terms, apply it feature by feature to any charge, and see why that classification, not immigration politics, decided the case.

The constitutional text that puts taxing power in one place

Begin with the document, because the whole dispute is anchored there. Article I of the Constitution opens by vesting all legislative powers in Congress, and Section 8 of that article begins its enumeration of congressional powers with the taxing power itself: the authority to lay and collect taxes, duties, imposts, and excises. The placement is not accidental. The framers, having just fought a revolution whose rallying complaint was taxation imposed without the consent of the governed, put the power to extract money from the public in the branch closest to the public, the one that stands for election most frequently and most directly. The Origination Clause reinforces the point by requiring that bills for raising revenue start in the House of Representatives, the chamber the framers tied most tightly to popular accountability. The structural message is consistent and deliberate: if the government is going to reach into private pockets to fund itself, the decision must be made by the people’s elected representatives, debated in the open, and recorded in a statute.

What the Constitution does not do is hand any slice of that power to the president. The executive enforces the laws Congress writes and exercises the authorities Congress delegates, and in the immigration field the president holds significant power over who may enter the country, much of it delegated by statute and some of it rooted in the executive’s role in foreign affairs. But none of that is a power to raise revenue. A president may, where Congress has authorized it, suspend the entry of a class of foreign nationals, attach conditions to entry, set processing requirements, and direct agencies to administer the immigration system. A president may not, on his own signature, impose a levy whose function is to raise money for the federal treasury, because that is taxation, and taxation is a legislative act. The Massachusetts court’s holding rests on this elementary separation. The question it had to answer was not whether the president has broad immigration authority, which he does, but whether the specific $100,000 charge was an exercise of that authority or an exercise of the taxing power dressed in immigration clothing.

That framing matters because it explains why the label on the payment could not save it. The proclamation that created the levy described it as a fee, a condition tied to the entry of certain nonimmigrant workers, justified by a stated national interest in protecting the domestic labor market. If the courts took labels at face value, that description would end the inquiry: the document says fee, the president has fee-setting authority in the immigration system, case closed. But constitutional classification has never worked that way, and for good reason. If the executive could convert a tax into a permissible charge simply by calling it something else in the operative document, the careful allocation of the taxing power to Congress would collapse into a drafting exercise. The protection the Constitution gives taxpayers would be worth exactly as much as the executive’s willingness to use the right vocabulary, which is to say nothing. So the law asks a different question than what the surcharge is called. It asks what the levy does.

Does the Constitution let only Congress impose taxes?

In substance, yes. Article I vests the taxing power in Congress and requires revenue bills to originate in the House, and nothing in the Constitution grants the president an independent power to tax. The executive may administer and enforce taxes Congress enacts and exercise delegated authority, but it cannot create a tax by proclamation, which is the line the H-1B charge crossed.

A fee, a tax, and a penalty are three different things

Public law recognizes several kinds of mandatory payments to the government, and they are governed by different rules because they do different work. The three that matter here are the regulatory fee, the tax, and the penalty. Keeping them distinct is the entire ballgame, because the administration’s defense depended on the measure being something other than a tax, and the challengers’ case depended on it being a tax and nothing else.

A regulatory fee is a exaction tied to a specific government service or to the cost of regulating a specific activity. When you pay a few hundred dollars for an agency to adjudicate a petition, that payment is a fee: it approximates the cost of the work the agency performs on your behalf, it is collected in exchange for a discrete service, and it is calibrated, at least roughly, to what that service costs to provide. The classic markers of a regulatory fee are that the payer receives a particular benefit or service in return, that the levy is reasonably proportioned to the cost of providing that benefit or regulating that activity, and that the money raised is used to defray those costs rather than to fund the government generally. A passport application fee, a permit charge, an inspection fee, the ordinary filing fees in the immigration system: these are user fees, exchanged for identifiable services, sized to the cost of the service. Because a fee is the price of a transaction the payer chooses to undertake, the executive can set fees where a statute gives it authority to do so, and doing so raises no taxing-power problem, because a fee is not an exercise of the taxing power at all.

A tax is different in kind. A tax raises revenue for the support of government, and it is not the price of a particular service rendered to the payer. When a payment is untethered from the cost of any service the government provides, when its function is to produce money rather than to recover the expense of a transaction, it is a tax regardless of the noun attached to it. The amount a tax raises bears no necessary relationship to the cost of anything the government does for the payer, because raising revenue, not recovering a cost, is the point. This is the category the $100,000 charge fell into, and the reason is arithmetic before it is doctrine: the levy bore no relationship to the cost of adjudicating an H-1B petition, which runs to a few thousand dollars at most, and instead functioned to extract a very large sum into the treasury on a lawful transaction.

A penalty is a third thing, and the distinction between a tax and a penalty turns out to be central to why the court reached the conclusion it did. A penalty is a payment imposed to punish or deter unlawful conduct. It attaches to behavior the law condemns, it often carries a scienter or culpability element, and its purpose is to enforce a prohibition rather than to raise money. A fine for breaking a regulation is a penalty. The reason the difference matters here is that the administration, at points, leaned toward characterizing the surcharge as a regulatory measure aimed at conduct rather than a revenue measure, and the court answered that framing directly. Hiring an H-1B worker is not unlawful. The program is a lawful one that Congress created and that employers use within rules Congress set. A charge imposed on a lawful activity, that raises money and does not aim to brand the activity as forbidden, is not a penalty for wrongdoing. It is a tax on a permitted transaction. The court’s observation that the payment was not a penalty, because it did not seek to establish that hiring these workers was illegal, was not a throwaway line. It closed off the one off-ramp that might have let the measure survive as something other than a tax.

What is the difference between a tax and a regulatory fee?

A regulatory fee is the price of a specific government service, sized to the cost of providing it, paid in exchange for a benefit to the payer. A tax raises revenue for general government support and is not tied to the cost of any service. The H-1B charge functioned to raise money far beyond processing cost, which made it a tax.

The functional test: substance over the label

The rule that decides which of these categories a measure falls into is the functional test, and its core instruction is simple to state and consequential to apply: courts look at what a exaction actually does, not at what the government calls it. The substance controls the form. A charge that operates as a tax is a tax for constitutional purposes even if every official document calls it a fee, and a surcharge that operates as a fee is a fee even if a statute happens to use the word tax. The label is evidence of nothing that matters, because the entity imposing the exaction has every incentive to choose the label that makes its action lawful, and a constitutional protection that could be defeated by word choice would be no protection at all.

The most important modern statement of this functional approach came in the Supreme Court’s 2012 decision on the Affordable Care Act’s individual mandate. That case is worth understanding precisely, because the Massachusetts court drew on the same mode of analysis and because the comparison illuminates how the functional test cuts in both directions. In the health-care case, Congress had required most people to maintain insurance coverage and had attached a payment, collected through the tax system, for those who did not. Congress called that payment a penalty. Opponents of the law argued it was a penalty and therefore could not be justified as an exercise of the taxing power, while the government argued that whatever Congress called it, the payment functioned as a tax and was a valid use of the taxing power. The Court agreed with the functional reading. It held that the label Congress chose did not control the constitutional question, and that the payment operated as a tax: it was collected by the revenue authority alongside ordinary taxes, it produced revenue for the government, it applied to a broad class of people, and it carried no finding of wrongdoing for those who paid it. Because it functioned as a tax, it was sustained under the taxing power even though Congress had named it a penalty.

It is worth isolating the specific factors the health-care decision treated as marks of a tax, because they map onto the H-1B analysis with unusual precision. The Court looked at how the payment was collected, finding that it ran through the ordinary revenue machinery rather than through a separate enforcement apparatus. It looked at the magnitude, finding that the payment was not so high as to leave no realistic choice, which kept it from operating as a pure command backed by punishment. It looked at whether the payment carried a finding of unlawfulness, finding that it did not, since those who paid were not treated as having broken the law. And it looked at whether the levy produced revenue, finding that it did. Several of these factors translate directly. The H-1B levy raised revenue, the dispositive factor, and it carried no finding of unlawfulness, since hiring through the program is permitted. The factor that pointed the other way in the health-care case, modest magnitude, pointed even more strongly toward a tax here, because the visa charge was anything but modest relative to the cost of the service. The factor analysis, in other words, did not merely permit the tax classification of the visa charge; it compelled it.

The lesson is that the functional test is neutral as to which way it points. In the health-care case it converted a payment Congress called a penalty into a tax in order to uphold it. In the H-1B case it converted a levy the executive called a fee into a tax in order to strike it down. The mechanism is identical: ignore the label, examine the operation, classify by what the levy does. What changed between the two cases was not the test but the identity of the party doing the imposing and the source of authority claimed. Congress may tax, so calling the health-care payment a tax saved it. The president may not tax, so finding the H-1B charge to be a tax doomed it. The same analytical move, the same substance-over-form principle, produced opposite results because the constitutional question on the other side of the classification was opposite.

What legal test decides whether a charge counts as a tax?

The functional test does. A court examines what the payment actually does rather than its label: whether it raises revenue or recovers a service cost, whether it is tied to the cost of a benefit, and whether it punishes unlawful conduct. If a surcharge raises revenue and is untethered from service cost, it is a tax whatever the government calls it.

Why a tobacco comparison belongs in the reasoning

One feature of the analysis that can look strange on first reading is the appearance of tobacco. Why would a decision about a visa charge invoke cigarettes? The answer reveals something fundamental about what a tax is, and it answers the administration’s most creative argument, which is treated in full in the next section. The short version is that taxes are routinely used to discourage activity, and the fact that a measure discourages the very thing it is imposed on does not make it any less a tax. Congress taxes tobacco heavily, and one acknowledged purpose of those taxes is to reduce smoking. A high tax on cigarettes is designed in part to make people buy fewer cigarettes. It is still unquestionably a tax, and no one suggests that the deterrent purpose somehow converts it into a regulatory fee or strips it of its character as revenue. It raises money on every pack sold even as it pushes total consumption down.

The tobacco comparison does analytical work because it isolates the principle that a tax remains a tax even when it suppresses the taxed activity and even when, in the aggregate, suppression might reduce total collections compared to a world with no charge at all. The purpose of a exaction and its incidental effect on behavior are simply different questions from its constitutional classification. A charge can aim to discourage, can in fact discourage, and can still be revenue-raising in its operation and therefore a tax in its nature. The relevance of all this to the H-1B charge is that the administration tried to use the levy’s deterrent effect as a sword, arguing that because the levy might reduce the number of petitions and therefore reduce net revenue, it could not really be a tax. The tobacco analogy is the clean answer to that move. We tax things we want less of all the time, and they remain taxes.

How the test applied to the $100,000 charge

With the framework in place, the application becomes straightforward, and it is worth walking through feature by feature, because the strength of the holding lies in how cleanly the payment fits the tax category on every marker that matters.

Start with the relationship between the measure and the cost of any service. A regulatory fee approximates the cost of the government work it pays for. Adjudicating an H-1B petition is administrative work, and the ordinary fees associated with the program reflect that, running from a modest base filing fee up through various add-on charges to a total that, depending on the petition, sat in the low thousands of dollars before the proclamation. Court filings in the case described the pre-proclamation cost of seeking a worker through the program in figures spanning roughly a thousand dollars to several thousand, depending on the employer’s size and the specific filings involved. Against that backdrop, a flat $100,000 charge is not in the same universe as the cost of processing. It is not a fee that drifted a little above cost. It dwarfs the cost of the service by a factor that is not close, and a levy that exceeds the cost of the underlying government work by more than an order of magnitude cannot be explained as recovering that cost. The gap between the levy and the expense it supposedly defrays is itself powerful evidence that the charge was doing something other than paying for a service. It was raising revenue.

Next consider purpose and design. The charge was structured to produce money on a lawful transaction. It applied to covered new petitions across the board, it was set at a fixed and very large amount, and it operated to extract that amount into federal coffers from employers using a program Congress authorized. A charge designed in this way, a uniform and substantial levy on a permitted activity, set without reference to the cost of administering that activity, has the architecture of a tax. The court’s characterization that the payment was designed to raise revenue from a lawful program captures exactly the feature that controlled the classification. The money, not the service, was the object.

Then consider the penalty question, which closes the last escape route. As discussed, a penalty punishes unlawful conduct and often presupposes culpability. The H-1B program is lawful, hiring through it is lawful, and the payment made no attempt to declare the activity forbidden. It simply attached a price to a lawful choice. That makes it a tax rather than a penalty, and since it is plainly not a regulatory fee given the cost gap, the only category left is tax. Every door but one was closed by the measure’s own design.

The contested figure on how many employers actually paid the surcharge before it was struck down does not change this analysis, and it is worth noting both because it appears in the record and because honesty about the data matters. The administration represented in a court filing that, as of a date in February, a relatively small number of payments of the $100,000 requirement had been received, in the dozens, producing a total in the single-digit millions of dollars. That figure is contested and has been characterized differently at different points, and it should be treated as a contested count rather than a settled number. But the size of the collections is beside the constitutional point. A charge does not become a fee because few people paid it, and it does not become a tax because many did. The classification turns on the nature and operation of the exaction, on whether it is tied to service cost and whether it functions to raise revenue, not on the receipts it happened to generate in any particular window. The dollars collected speak to the economic stakes, which are analyzed in their own right elsewhere in this series; they do not speak to the legal classification.

Why is the H-1B charge a tax and not a fee?

Because it raised revenue far beyond the cost of processing a petition, which runs to a few thousand dollars, and was designed to produce money on a lawful activity rather than to recover a service cost or punish wrongdoing. A charge untethered from service cost and aimed at revenue is a tax, whatever the proclamation called it.

The no-net-revenue rebuttal and why it failed

The administration’s most ingenious argument deserves a careful answer, because at first hearing it has a certain logic, and because the court’s rejection of it sharpens the whole functional test. The argument ran roughly as follows. A tax, the administration suggested, is a levy that raises revenue. But this payment was so high that it would deter employers from filing, suppressing the number of petitions, and a sufficiently steep charge might actually reduce the government’s net take compared to the lower fees that came before, since far fewer petitions would be filed. If the charge could reduce net revenue, the argument went, it cannot really be a revenue measure, and therefore it cannot be a tax. It must be something else, a regulatory restriction on entry justified by immigration policy.

The court did not accept this, and the reasons it did not are instructive. The first answer is the one the tobacco comparison supplies. Whether a measure, in the aggregate and over time, increases or decreases the government’s net collections compared to some baseline is not the test for whether it is a tax. Taxes routinely suppress the activity they fall on, and a tax that shrinks its own base by discouraging the taxed conduct is still a tax. If the no-net-revenue theory were correct, Congress’s heavy taxes on cigarettes would cease to be taxes the moment they succeeded at their public-health goal of reducing smoking enough to lower total receipts, which is an untenable conclusion. The character of a levy as revenue-raising is assessed by what it does when it is paid, not by speculative aggregate effects across a changed market.

The second answer is more direct still: every payment of the payment that was actually made indisputably raised money for the government. When an employer paid the $100,000, that sum went into the treasury. It was revenue. The fact that the measure might cause other employers not to file, and thus not to pay, does not unmake the revenue character of the payments that were made. The functional test looks at the operation of the charge on the transaction it governs. On every covered petition where the levy was paid, it operated as a revenue extraction far in excess of service cost. That is taxation in operation, and the possibility that the levy chilled other transactions is a fact about market behavior, not a fact about the legal nature of the charge.

The third answer goes to the structure of the argument itself. The no-net-revenue theory tries to convert a purpose argument into a classification argument. Even granting that the administration also wanted to deter hiring, a surcharge can have a deterrent purpose and a revenue character at the same time, exactly as a cigarette tax does. The presence of a regulatory motive does not subtract the revenue function. To treat the existence of a policy goal as proof that a exaction is not a tax would let the executive escape the taxing-power limit on every revenue measure simply by announcing a policy reason for it, which would swallow the rule. The court declined to let purpose launder the charge out of its category.

Can a charge be a tax even if it reduces overall revenue?

Yes. A charge is a tax based on how it operates when paid, not on its aggregate effect on collections. Every payment made raises money, and taxes routinely suppress the activity they fall on, the way tobacco taxes reduce smoking yet remain taxes. A deterrent effect does not strip a revenue-raising charge of its tax character.

A historical anchor: when immigration charges were taxes, Congress imposed them

The taxing-power objection to the visa charge is not a modern invention pressed into service against an immigration measure. American history contains a direct and durable precedent for the proposition that a levy imposed on immigration can be a valid tax, but only because the body that imposed it was Congress. In the late nineteenth century, Congress enacted a payment on each immigrant arriving by ship, a per-passenger levy collected to defray the costs associated with immigration and to fund related government functions. When that charge was challenged, the Supreme Court upheld it, and the ground on which it upheld it is the entire point for present purposes. The Court treated the charge as an exercise of Congress’s powers, including its taxing power, and sustained it because it fell within the authority the Constitution assigns to the legislature. The charge was a tax on immigration, and a tax on immigration was permissible because Congress, the body that holds the taxing power, had enacted it.

Set that historical episode beside the visa charge and the contrast is illuminating. In both instances the government imposed a monetary charge connected to immigration, and in both instances the measure had a revenue character. The difference that decided each case was institutional. The historical charge survived because the legislature levied it through statute, exercising a power the Constitution grants Congress. The visa charge fell because the executive imposed it through proclamation, claiming a power the Constitution does not grant the president. The episodes together state the rule with a clarity that no abstract description matches: a tax on immigration-related activity is constitutionally unremarkable when Congress enacts it and constitutionally fatal when the executive imposes it alone. The classification of the charge as a tax was never the end of the analysis on its own; it became fatal only because the charge came from the wrong branch. Had the historical Congress tried to impose its immigration tax by handing the job to the executive to do by decree, that charge would have faced the same objection the visa levy did.

This history also disposes of a rhetorical move that sometimes accompanies defenses of executive immigration charges, the suggestion that treating such a charge as a tax somehow places immigration beyond the reach of fiscal policy. It does no such thing. Immigration has long been a permissible subject of federal taxation, and Congress retains full authority to tax immigration-related activity, to set substantial charges on visa categories, and to fund the immigration system or the broader government through such measures. The functional test does not immunize immigration from charges. It directs charges of a revenue character to the branch the Constitution authorizes to impose them. The visa levy was not struck down because immigration cannot be taxed. It was struck down because the president cannot tax, and the charge was a tax.

Earmarking and the general-treasury question

A natural objection to the tax classification runs through the use of the money. If a charge’s proceeds were dedicated to running the immigration program itself, the objection goes, then the charge funds a service rather than the government at large, which is the hallmark of a fee. Does dedicating revenue to a program convert a tax into a fee? The doctrine answers no, with an important qualification, and working through the answer sharpens the understanding of why the use of proceeds is only one factor among several and never the controlling one.

The use of proceeds is relevant evidence because a genuine regulatory fee typically funds the service it pays for, while a tax typically funds the government generally. But earmarking the proceeds of a charge to a particular program does not, by itself, transform a revenue measure into a fee, because the other markers of classification remain. A charge that is untethered from the cost of any service, set far above what the program costs to run, does not become a fee merely because its proceeds are routed to that program rather than to the general treasury. The reason is that a fee recovers a cost, and a charge that exceeds the cost of the service by an order of magnitude is recovering nothing; it is raising a surplus, and a surplus on a lawful activity is revenue regardless of where the revenue is deposited. If earmarking alone could convert a tax into a fee, the executive could impose any revenue charge it liked and immunize it simply by dedicating the proceeds to some program, which would defeat the taxing-power limit as thoroughly as the label problem would. The qualification is that a charge sized to the actual cost of a program, with proceeds dedicated to that program, has a strong claim to fee status, because both the proportionality and the use-of-proceeds factors line up. What matters is the combination, and the decisive factor remains the relationship between the charge and the cost. For the visa levy, no plausible accounting of immigration-program costs could justify a $100,000 charge, so the use-of-proceeds question never had a chance to rescue it. The charge was a surplus-generating measure, which is to say a tax, however its proceeds were allocated.

Who carries the burden when a charge’s classification is disputed

The classification analysis also has a procedural dimension that is easy to overlook and worth stating, because it affects how these disputes are actually litigated and how an analyst should weigh the strength of a challenge. When a party contends that a charge labeled a fee is in substance a tax, the inquiry is not a matter of deferring to the government’s characterization. Courts do not begin from a presumption that the executive’s label is correct and require the challenger to overcome it with extraordinary proof. The functional test is, by design, a judicial examination of the charge’s actual operation, conducted on the record, in which the markers, proportion to cost, exchange for a benefit, revenue function, and relationship to wrongdoing, are assessed on their merits. The government’s choice of label is entitled to no particular weight, because the entire premise of the substance-over-form rule is that the label is unreliable evidence of what the charge is.

In practice this means the strength of a classification challenge tracks the strength of the functional evidence, not the cleverness of the government’s framing. Where the gap between charge and cost is enormous and the revenue function is plain, as with the visa levy, the challenger’s task is straightforward, because the charge’s own design supplies the proof. Where a charge sits closer to the line, sized within a defensible range of program costs and exchanged for an identifiable service, the analysis becomes genuinely contestable and the outcome less certain. This is part of why a different court, weighing the same charge under a framing that emphasized the entry-condition characterization, could reason toward a different conclusion, a divergence examined in the analysis of the two rulings set head to head. But on the specific question of the cost gap, the visa charge offered no close call to exploit. The burden, however allocated, was easily carried by the simple arithmetic of a six-figure charge on a service costing a small fraction of that amount.

The label does not control, and that rule protects taxpayers

It is worth dwelling on why the law refuses to take the government’s word for what a charge is, because the intuition that a charge labeled a fee is a fee is natural and widely held, and the answer to that intuition is the heart of the doctrine. The reason the label does not control is structural, not pedantic. The whole point of placing the taxing power in Congress is to subject revenue decisions to the legislative process, with its hearings, its votes, its accountability, and its requirement that revenue bills originate in the popularly elected House. If the executive could impose a charge that functions exactly like a tax, raises revenue on a lawful activity untethered from cost, and immunize it from the taxing-power limit merely by writing the word fee on the order, then the executive could tax at will. The constitutional allocation would be a formality, defeated by a thesaurus.

The functional test is the doctrine that prevents this. By classifying charges according to what they do rather than what they are called, it makes the taxing-power limit real. It tells every branch that the substance of a charge, not its branding, determines which constitutional rules apply. This is protective of the public in a direct way. It means that no matter how a revenue measure is packaged, the people retain the protection of having such measures decided by their elected representatives in the open. The rule against label control is not a technicality that let a disliked policy off the hook. It is the mechanism by which the taxing-power guarantee is enforced, and it would protect a future administration’s preferred program from a future Congress or executive just as surely as it operated here.

This is also why the entry-condition framing could not rescue the charge. The administration’s strongest characterization was that the levy was a condition on the entry of a class of foreign workers, an exercise of the president’s acknowledged immigration authority, rather than a revenue measure at all. That framing is not frivolous, and a different judge weighing the same arguments reasoned toward the opposite conclusion, which is why the existence of two conflicting rulings is itself a major feature of this dispute. But under the functional test, calling the charge an entry condition is another label, and the question remains what the charge does. A condition on entry that happens to require payment of an amount calibrated to the cost of processing entry would be a fee. A condition on entry that requires payment of an amount many times the cost of processing, designed to raise revenue, is a tax wearing the costume of an entry condition. The Massachusetts court concluded that the substance was revenue and that the entry-condition label did not change it. The functional test, applied honestly, does not care which permissible-sounding name the executive selects; it asks whether the operation is taxation, and here it found that it was.

Why can the government not avoid the tax label just by renaming it?

Because constitutional classification looks at what a charge does, not what it is called. If a label could convert a tax into a permissible fee, the executive could tax at will simply by choosing the right word, and the constitutional placement of the taxing power in Congress would become meaningless. The substance-over-form rule exists precisely to prevent that.

What it would take to make the charge a lawful fee

Understanding the test also tells you, with some precision, what a lawful version of a comparable charge would have to look like, which is more useful to an employer, an analyst, or a policymaker than any amount of commentary on the politics. There are two clean paths, and they correspond to the two ways a charge can be legitimate.

The first path is to make the charge an actual fee by tying it to cost. A charge that recovers the genuine cost of administering and regulating the program, sized to that cost rather than dwarfing it, exchanged for the service of adjudication, would be a regulatory fee, and the executive can set such fees where statute authorizes it. The defect in the $100,000 levy was the chasm between the charge and the cost. A charge sized to cost, even a generously calculated cost, would not raise the taxing-power problem, because it would not be a tax. The constraint is that a cost-based fee cannot raise the kind of revenue the proclamation sought, precisely because cost recovery and revenue raising are different objectives. You cannot have a charge that both stays within the cost of the service and produces a windfall to the treasury; the two are mutually exclusive by definition, and that mutual exclusivity is the whole content of the tax-versus-fee line.

The second path is to do it the constitutional way: have Congress legislate it. If Congress chose to impose a large charge on H-1B petitions, whether to raise revenue, to discourage reliance on the program, or both, it could do so through ordinary legislation, because Congress holds the taxing power and may tax a lawful activity for any of these reasons. A statutory charge of this kind would face its own policy debate and could be challenged on other grounds, but it would not founder on the taxing-power objection that sank the proclamation, because the body imposing it would be the body the Constitution authorizes to impose it. This is the most important practical lesson of the classification. The problem was never that a large charge on the program is inherently unconstitutional. The problem was who imposed it and how. The same charge, enacted by statute, would stand on entirely different footing.

This is also why describing the procedural defect as a mere technicality misunderstands the holding, a point developed further in the analysis of the reasoning from holding to remedy. The taxing-power ground is not a curable formality. An agency cannot cure a charge that is unconstitutional in substance by running it through more process. The only cures are to make the charge an actual cost-based fee, which abandons the revenue goal, or to obtain a statute, which abandons the unilateral executive route. Each cure gives up the very thing the proclamation was trying to achieve, which is why the classification is so consequential.

The tariff precedent and the principle that the executive cannot tax

The Massachusetts decision did not arrive in a vacuum on the question of executive taxation. It drew on a recent and closely related development: the Supreme Court’s 2026 decision in the tariff litigation, which had already addressed in a different setting whether a broad grant of executive authority carries with it a hidden power to raise revenue. In that case the Court confronted an attempt to use a sweeping economic-powers statute as the basis for imposing tariffs, and it concluded that the statute did not authorize the executive to wield what amounted to a taxing power on its own say-so. The principle the tariff decision stands for, that a general regulatory or emergency authority does not silently contain the power to tax, transferred naturally to the visa charge. In both settings the executive claimed that a broad statutory or constitutional authority over a domain, commerce in one case and immigration in the other, implicitly empowered it to impose revenue measures. In both settings the answer was that the power to raise revenue is not a default feature of broad regulatory authority; it must be granted, and it is granted to Congress.

This cross-pollination between the tariff dispute and the visa charge is one of the more analytically interesting features of the moment, and it is treated at length in the parts of this series that compare the two executive revenue measures directly. For the purpose of the tax-versus-fee question, the relevant point is narrow but firm: the principle that a president cannot convert a broad grant of authority into a license to tax was not invented for the H-1B charge. It was applied to it, drawn from a doctrine that the highest court had reaffirmed in the tariff context. The classification of the charge as a tax connected it to that principle, and the principle then supplied the conclusion.

The deeper structural idea uniting the two cases is a clear-statement instinct about the taxing power. When the government claims that a broad authority silently includes the power to raise revenue, courts have grown increasingly willing to demand that the grant say so clearly, because the power to tax is too consequential to be inferred from general language. A statute that authorizes the executive to regulate commerce, or to suspend entry in the national interest, speaks to regulation and to entry; it does not announce a power to levy charges into the treasury, and courts will not read such a power into silence. The visa charge and the tariffs both rested on exactly this kind of inference, an argument that a broad regulatory authority carried a hidden fiscal power. The rejection of that inference in both settings reflects a shared judgment that the taxing power is special, that it belongs to Congress by explicit constitutional design, and that the executive cannot acquire it through the expansive reading of a statute written for other purposes. The tax classification of the visa charge is what brought it within range of this principle, and once within range, the charge could not survive an executive imposition.

How other nations charge for skilled-worker visas without this problem

The comparative frame is where the classification stops being an American peculiarity and becomes a window onto a design choice that other skilled-immigration systems made differently, and it is the most useful lens for anyone trying to understand whether the result was an accident of American constitutional structure or a deeper feature of how a charge like this should be created. The striking thing about the major comparator systems is that they pursue the very goals the proclamation invoked, selectivity, revenue, and protection of domestic workers, without running into the constitutional wall the charge hit, and they do so for one reason: their charges are created by statute or by regulations issued under statutory authority, not by unilateral executive decree.

Consider Canada, which runs the most studied points-based skilled-immigration system in the world. Canada does not select skilled workers by attaching a flat charge to each hire. It ranks candidates in a managed pool through a points system, the Express Entry framework, scoring applicants on age, education, language ability, work experience, and other factors, and inviting the highest-ranked candidates to apply. Selectivity, the goal the proclamation cited, is achieved by ranking and choosing rather than by pricing. The fees that do exist in the Canadian system are processing and right-of-permanent-residence charges set through regulation under statutory authority, and they are calibrated as fees, not deployed as a six-figure levy to raise general revenue or to price hiring out of reach. The Canadian design pursues the policy objective through a legislated selection mechanism, which sidesteps entirely the question of whether the executive may impose a revenue charge, because the charge is not the policy instrument and the framework rests on statute.

The United Kingdom offers a different and equally instructive model, because it actually does impose a substantial charge tied to skilled-worker hiring, and yet faces no comparable problem. The British system funds itself through visa fees set by regulation, and on top of those it levies the Immigration Skills Charge, a charge employers pay when they sponsor certain skilled workers from abroad, designed in part to encourage investment in domestic training. That charge is real, it is meaningful in amount, and it is openly intended to influence employer behavior, which makes it a close functional cousin of what the proclamation attempted. The difference is constitutional architecture. The Immigration Skills Charge was established through legislation and is implemented through regulations made under statutory authority, the product of the parliamentary process rather than an executive proclamation. Britain’s unwritten constitution and parliamentary system do not separate a taxing power from an executive the way the American Constitution does, but the operative point for the comparison holds: the charge exists because the legislature created it, which is precisely the route the American Constitution requires and the proclamation skipped. A charge of similar ambition, created in the United States by statute, would have stood; created by proclamation, it fell.

Australia rounds out the picture with a points-tested skilled-migration system in the same family as Canada’s, selecting workers by ranking them against criteria rather than by imposing a flat per-hire levy, with fees set through its legislative and regulatory machinery. Across these systems a consistent pattern emerges. Comparable democracies achieve selectivity through legislated selection mechanisms and fund their systems through fees and charges grounded in statute. None of them relies on an executive imposing a large revenue charge by decree, and so none of them generates the constitutional collision the proclamation produced. The lesson the comparison teaches is not that the United States is uniquely hostile to charging for skilled-worker visas. It is that the American constitutional design insists, more explicitly than some systems, that a revenue charge of this magnitude come from the legislature, and that the proclamation’s defect was a defect of method that the comparator systems avoid by routing such charges through their lawmaking bodies.

Germany and the broader European Union add a further data point through the EU Blue Card framework, which channels skilled non-European workers into the labor market on the basis of qualifications and a salary threshold. The Blue Card scheme was created through European legislation and transposed into national law by member states, and the salary thresholds and conditions that govern it are set by statute and regulation rather than by executive fiat. Germany pairs this with active recruitment of skilled labor in shortage occupations, the same labor-market goal the proclamation invoked, pursued through legislated criteria and thresholds rather than through a punitive per-hire charge imposed by decree. Singapore, often cited as a model of selective high-skill immigration, manages its Employment Pass system through qualifying salary thresholds and a points-based assessment, with the framework and its charges established through its governmental processes. Across these systems the through-line holds without exception: where a charge or a threshold shapes skilled-worker entry, it is the product of the legislature or of regulation issued under legislative authority, not of a standalone executive revenue measure.

The breadth of the comparison matters because it forecloses the suggestion that the American result was a quirk. One could imagine arguing that the United States simply has an unusually rigid constitution that produces awkward outcomes other countries avoid through flexibility. The comparison shows the opposite. Other democracies reach the same destination by the same route. They impose selectivity and even substantial charges, but they do so through their legislatures, which is exactly what the American Constitution demands. The American structure is more explicit about separating the taxing power from the executive, but the operative norm, that a revenue charge of real magnitude comes from the lawmaking body, is shared across the systems that the proclamation’s defenders might otherwise hold up as models. The proclamation did not fail because the United States is constitutionally peculiar. It failed because it skipped the step that every comparable system treats as fundamental.

How do other countries charge for skilled-worker visas without this problem?

They route charges through legislation. Canada and Australia select skilled workers by points-ranked pools rather than flat per-hire levies, and the United Kingdom’s Immigration Skills Charge, though substantial, was created by statute and regulation. Because the legislature sets the charge, no question arises of an executive imposing a tax by decree, which is exactly what the American Constitution forbids.

The domestic user-fee doctrine as a sanity check

It is worth grounding the comparison back home, because American law has a well-developed body of doctrine on exactly when a charge is a permissible fee and when it crosses into taxation, and the H-1B levy fails that domestic test for the same reasons it fails the broader constitutional one. In the domestic user-fee context, courts have long distinguished charges that survive as fees from those that operate as taxes, and the line they draw matches the functional test precisely. A charge tends to be treated as a valid regulatory fee when it is assessed against those who use a service or trigger the regulation, when it is reasonably proportioned to the cost of the service or regulatory activity, and when the proceeds are used to fund that service or activity rather than poured into the general treasury. A charge tends to be treated as a tax when it is imposed on a broad class to raise general revenue, when it bears no reasonable relationship to any service the payer receives, and when the money funds government broadly.

Run the $100,000 charge through that domestic framework and the result is the same. It was not proportioned to the cost of any service; it exceeded the cost of adjudication many times over. It was not exchanged for a calibrated benefit; the benefit, adjudication of a petition, costs a fraction of the charge. Its function was to raise revenue on a lawful activity. By the settled markers of the user-fee doctrine, it lands in the tax column. This domestic doctrine matters because it shows the classification was not a novel or strained reading invented to reach a result. It is the ordinary application of a test American courts have used for generations to keep charges honest about what they are, and the visa levy failed it on every prong that distinguishes a fee from a tax.

The domestic user-fee cases also supply a useful vocabulary of degree that the visa charge makes almost irrelevant by its extremity. In the harder cases that have shaped this doctrine, courts have wrestled with charges that sit in a genuine gray zone: a regulatory charge that recovers somewhat more than the bare cost of a service, a charge that funds a program broadly related to the payers but not strictly limited to the service they receive, a charge calibrated to a class of users rather than to each individual transaction. These cases are difficult because the markers pull in different directions, and reasonable judges can weigh them differently. They are the reason the tax-versus-fee line generates litigation at all. What makes the visa charge an easy case rather than a hard one is that it does not sit in any gray zone. It is not a charge that recovers somewhat more than cost; it is a charge that bears no relationship to cost. It is not a charge calibrated to a class of users in a defensible way; it is a uniform six-figure levy on a lawful activity. The doctrine’s hard cases test the boundaries of the fee category, and the visa charge is nowhere near those boundaries. It is deep in tax territory, which is why the classification, contested though the broader case is on other framings, was on the merits a comfortable conclusion.

A fee-versus-tax test, scored against the charge

The functional test is easiest to hold in mind as a set of features, and it is most useful when applied directly to the charge in question. The table below lays out the markers that distinguish a regulatory fee from a tax and scores the $100,000 H-1B levy against each. This is the one place in this analysis where a table earns its place, because the comparison is genuinely a side-by-side scoring and reads more clearly in columns than in prose.

Feature	Marks a regulatory fee	Marks a tax	How the $100,000 charge scored
Relationship to service cost	Reasonably proportioned to the cost of the service or regulation	Untethered from service cost; far exceeds it	Exceeded the few-thousand-dollar cost of adjudication by more than an order of magnitude, scoring as a tax
Exchange for a benefit	Paid for a discrete service or benefit to the payer	Not the price of a service; imposed on a transaction to raise money	The adjudication benefit costs a fraction of the charge, so the charge was not a price for the service, scoring as a tax
Primary function	Recovering the cost of providing a service or regulating an activity	Raising revenue for government support	Designed to raise revenue on a lawful program, scoring as a tax
Use of proceeds	Defrays the cost of the service or regulatory program	Funds government generally	Operated to extract revenue into the treasury, scoring as a tax
Relationship to wrongdoing	Not applicable; a fee is not a punishment	Not a penalty; falls on lawful activity rather than punishing unlawful conduct	Imposed on lawful hiring with no finding of illegality, confirming it was a tax and not a penalty
Who may impose it	The executive may set fees where statute authorizes	Only Congress may levy a tax	Imposed by presidential proclamation, which the Constitution does not authorize for a tax

Read down the right-hand column and the verdict is not a close call. On every marker that separates a fee from a tax, the charge scored as a tax, and the final row supplies the consequence: a tax imposed by the executive, rather than by Congress, cannot stand. The table is not a substitute for the reasoning above; it is a compression of it, the functional test made legible at a glance.

The strongest version of the opposing reading, and why it still fails

Intellectual honesty, and the requirement of a fair account, demand engaging the best version of the argument that the charge was not a tax, because a different judge found that version persuasive and an appeal will press it hard. The strongest opposing reading does not deny the functional test. It accepts that substance controls and then argues that the substance here was not revenue but regulation of entry. On this view, the charge is a condition the president attached to the admission of a class of foreign workers, an exercise of the broad authority over entry that statute and the executive’s foreign-affairs role confer. The amount, the argument continues, reflects a judgment about how strongly to discourage reliance on the program, not a revenue target, and a high entry condition is still an entry condition rather than a tax. The point of comparison is not the cost of adjudication, which is irrelevant to an entry condition, but the policy weight the executive assigned to limiting the category. Read this way, the charge is a regulatory restriction that happens to take monetary form, and regulatory restrictions on entry are squarely within executive power.

This is a serious argument, and it is why the dispute is genuinely contested rather than one-sided. But it does not survive the functional test as the Massachusetts court applied it, for a reason that goes to the difference between a condition and a charge. A genuine condition on entry regulates who may come and on what terms in a way that operates as a restriction: a health requirement, a documentation requirement, a security screening, even a numerical cap. These conditions shape entry by governing eligibility and conduct. A demand for a large sum of money, untethered from any service and payable into the treasury, does not govern eligibility or conduct in that sense. It governs nothing except the transfer of money, and a requirement whose entire operation is the transfer of money to the government is the definition of a fiscal exaction. The entry-condition framing tries to characterize a payment obligation as a regulation, but the payment obligation does not regulate; it raises revenue. The court’s response to this framing was to ask what the charge actually does, and what it does is collect money on a lawful activity in an amount unrelated to cost. That is taxation, and dressing it as an entry condition does not change the operation, only the description.

The opposing reading also proves too much, which is often a sign that an argument has slipped its constitutional moorings. If a monetary demand of any size could be recharacterized as a regulatory entry condition immune from the taxing-power limit, then there would be no executive revenue measure in the immigration field that the limit could ever reach, because every such measure could be framed as a condition on entry. The president could impose a charge of any magnitude on any immigration benefit and call it a condition, and the taxing power would cease to constrain the executive in the one domain where the executive’s authority is broadest. A reading that erases the constitutional limit it is supposed to be navigating cannot be the right reading. The functional test exists precisely to prevent characterization from defeating substance, and the entry-condition framing, however sophisticated, is a characterization. The Massachusetts court declined to let it override the substance, and the structural logic of the taxing power supports that choice even as a different court, weighing the framing differently, came out the other way.

The namable claim: the revenue-versus-cost test

If this analysis reduces to a single portable rule, it is this: a charge crosses the line from fee to tax when it is untethered from the cost of the service it supposedly pays for and functions to raise revenue, whatever it is called. Call it the revenue-versus-cost test. It is the through-line of every part of the analysis above. It explains why the cost gap mattered so much, because the gap is the evidence that the charge was not recovering a cost. It explains why the label could not save the charge, because the test looks at function and not at branding. It explains why the no-net-revenue argument failed, because each payment raised revenue regardless of aggregate effects. It explains why a cost-based fee or a statute would have been lawful, because the first stays on the cost side of the line and the second comes from the body that may tax. And it explains why the comparator nations avoid the problem, because they create such charges through legislation rather than executive decree.

The value of stating the rule this plainly is that it travels. An attorney can deploy it against the next executive charge that claims to be a fee. A student can use it to classify any charge on an exam or in a brief. A policymaker can use it as a design constraint, knowing in advance that a charge meant to raise serious revenue will be a tax and must come from the legislature. An employer can use it to read the durability of any future charge: a cost-based fee set by an agency is on firm ground, a revenue charge imposed by proclamation is not. The revenue-versus-cost test is the doctrine reduced to something a reader can actually use, which is the entire point of working through it.

Verdict on the doctrine’s bearing

The tax-versus-fee classification was not one issue among many in the H-1B litigation. It was the issue, the hinge on which the constitutional outcome turned, and the functional test was the tool that decided it. The charge fit the tax category on every feature that matters: it dwarfed the cost of the service it claimed to price, it was designed to raise revenue on a lawful activity, it was not a penalty because the activity was not unlawful, and it was imposed by an executive who holds no power to tax. None of the administration’s reframings changed that, because the functional test does not yield to labels, to entry-condition characterizations, or to arguments built on a charge’s deterrent effect. The substance was taxation, and taxation by the executive is something the Constitution does not permit.

The durability of this classification is high, precisely because it rests on a settled and structural feature of American constitutional law rather than on a contestable reading of immigration policy. The functional test is old, well-anchored in precedent, and reaffirmed in the tariff context. The placement of the taxing power in Congress is textual and foundational. An appeal will test other things, including the entry-condition framing that persuaded a different court and the procedural questions, but the tax characterization sits on the most stable ground in the case.

There is a final reason the classification deserves the weight this analysis has given it, which is that it reframes what the whole dispute is actually about. Read superficially, the case looks like a fight over immigration: how open the skilled-worker program should be, how much employers should pay, whether the executive may tighten access. Read through the tax-versus-fee lens, it is not really about immigration at all. It is about the allocation of a fundamental power, and the immigration setting is incidental. The same charge on a different lawful activity would have raised the same question and met the same answer, because the constitutional defect was structural rather than substantive. That is why the classification has reach beyond this program and beyond this administration. It states a rule about who may impose revenue charges in a constitutional system that deliberately separated that power from the executive, and it enforces that rule through a test that no label, no policy rationale, and no clever recharacterization can evade. A reader who understands why the charge was a tax understands not just this case but the principle that will govern the next executive attempt to raise revenue by decree, whatever form it takes and whoever attempts it.

The single sentence that captures the holding is the revenue-versus-cost test itself: a charge that raises revenue far beyond the cost of the service it claims to fund is a tax, and the executive cannot impose a tax, however the charge is named. Everything else in the decision flows from that, and any reader who carries away that one rule has the key to the entire dispute.

Frequently Asked Questions

Q: Why is the H-1B $100,000 charge considered a tax and not a fee?

Because it raised revenue far beyond the cost of the service it claimed to price. Adjudicating a petition costs a few thousand dollars at most, yet the levy was set at $100,000, untethered from that cost and designed to extract money on a lawful activity. A charge that raises revenue rather than recovering a service cost is a tax in function, regardless of the label the proclamation attached to it. The Massachusetts court applied this functional analysis, looked past the word fee, and concluded the charge operated as a tax. Because the Constitution places the taxing power in Congress, an executive-imposed tax cannot stand, and that classification controlled the outcome.

Q: What is the difference between a tax and a regulatory fee?

A regulatory fee is the price of a specific government service or the cost of regulating a particular activity. It is paid in exchange for a discrete benefit to the payer, is reasonably proportioned to the cost of providing that benefit, and funds the service rather than the government at large. A tax is fundamentally different: it raises revenue for general government support and is not tied to the cost of any service the payer receives. The clearest practical marker is proportion to cost. A fee tracks the cost of the service, while a tax is untethered from it. The H-1B charge dwarfed processing cost and functioned to raise revenue, which placed it firmly in the tax category.

Q: What legal test decides whether a charge counts as a tax?

Courts apply the functional test, which classifies a charge by what it actually does rather than by what the government calls it. The inquiry asks whether the charge is reasonably proportioned to the cost of a service, whether it is exchanged for a discrete benefit, and whether its function is to raise revenue or to recover a cost. A charge that is untethered from service cost and operates to produce revenue is a tax, even if every official document labels it a fee. The substance controls the form. This same functional approach has been used for generations in the domestic user-fee context and was reaffirmed at the Supreme Court level in modern decisions, which is why the Massachusetts court could apply it with confidence.

Q: Does the Constitution let only Congress impose taxes?

In substance, yes. Article I vests legislative power in Congress, lists the taxing power first among its enumerated authorities, and requires that revenue-raising bills originate in the House of Representatives. The framers placed the power to extract money from the public in the branch closest to the electorate, a direct response to the revolutionary grievance over taxation without representation. The president holds significant authority over immigration and the enforcement of laws, but nothing in the Constitution grants the executive an independent power to tax. The executive may administer taxes Congress enacts and exercise delegated authority, but it cannot create a tax by proclamation. That structural limit is why classifying the H-1B charge as a tax was fatal to it.

Q: How did the Affordable Care Act decision shape the test used here?

The 2012 Supreme Court decision on the Affordable Care Act’s individual mandate is the leading modern statement of the functional test. There, Congress had called a payment a penalty, and the Court held that the label did not control the constitutional question. Examining how the payment actually operated, collected through the tax system, producing revenue, applying broadly, with no finding of wrongdoing, the Court classified it as a tax and upheld it under the taxing power. The Massachusetts court used the same substance-over-label reasoning, but it cut the other way. The functional test converted a charge the executive called a fee into a tax, and because the executive may not tax, that classification doomed the charge rather than saving it.

Q: Why did a tobacco comparison appear in the reasoning?

The tobacco analogy answers the argument that a charge cannot be a tax if it discourages the taxed activity. Congress taxes cigarettes heavily, and a recognized purpose of those taxes is to reduce smoking. Yet a cigarette tax is unquestionably a tax, even though it is designed to shrink consumption and may reduce total receipts as it succeeds. The comparison isolates a principle: a tax remains a tax even when it suppresses the activity it falls on. Applied to the visa charge, it defeats the claim that the levy’s deterrent effect on hiring stripped it of its tax character. A charge can aim to discourage, can in fact discourage, and still be revenue-raising in operation and therefore a tax in nature.

Q: Can a charge be a tax even if it reduces overall revenue?

Yes. The character of a charge as a tax is assessed by how it operates when paid, not by its aggregate effect on government collections. Every payment of the H-1B levy that was made put money into the treasury, which is revenue, and that does not change because the charge might deter other employers from filing and paying. Taxes routinely suppress the activity they fall on, the way tobacco taxes reduce smoking, and they remain taxes. Accepting the contrary theory would mean a tax stops being a tax the moment it succeeds at discouraging conduct, which is untenable. The court declined to let the levy’s chilling effect convert it out of the tax category.

Q: Why can the government not avoid the tax label just by renaming it?

Because constitutional classification looks at what a charge does, not at what it is called. If a label could convert a tax into a permissible fee, the executive could tax at will simply by writing the right word on an order, and the constitutional placement of the taxing power in Congress would be a formality defeated by drafting. The substance-over-form rule exists precisely to prevent that. It makes the taxing-power limit real by tying the applicable rules to the operation of the charge rather than to its branding. This protects the public, because it guarantees that revenue measures, however packaged, must come from elected representatives through the legislative process.

Q: What features mark a charge as a regulatory fee rather than a tax?

A regulatory fee typically carries three features. It is assessed against those who use a service or trigger the regulation, so it falls on the people who receive the benefit. It is reasonably proportioned to the cost of providing that service or regulating that activity, so the amount tracks the cost rather than dwarfing it. And its proceeds are used to fund that service or regulatory program rather than the general treasury. When a charge has all three features, it functions as a fee. The H-1B levy had none of them in the way that matters. It exceeded the cost of adjudication many times over, was not a calibrated price for the service, and operated to raise revenue, which is why it failed the fee test on every prong.

Q: Is a user fee the same thing as a tax?

No, and the distinction is the heart of the doctrine. A user fee is the price someone pays for a particular government service they use, such as a passport application charge or an inspection fee, and it is sized to the cost of providing that service. A tax raises revenue for the support of government generally and is not the price of any specific service rendered to the payer. American courts have long distinguished the two by asking whether the charge is proportioned to service cost, exchanged for a discrete benefit, and used to fund that service. A user fee answers yes to those questions; a tax does not. The H-1B charge, sized far above service cost and aimed at revenue, was a tax rather than a user fee.

Q: Why does the gap between the charge and the cost of processing a petition matter?

The gap is the central piece of evidence that the charge was not recovering a cost. A regulatory fee approximates the cost of the government work it pays for, so a fee and the cost of the service stay in the same neighborhood. Processing an H-1B petition costs a few thousand dollars, while the levy was set at $100,000. A charge that exceeds the cost of the underlying service by more than an order of magnitude cannot be explained as cost recovery, because there is no plausible cost it could be recovering at that level. The only remaining explanation for the charge is that it was raising revenue, which is the defining function of a tax. The size of the gap is what made the classification clear.

Q: Is the charge a penalty rather than a tax, and does the difference matter?

It is not a penalty, and the difference matters a great deal. A penalty punishes unlawful conduct and usually presupposes culpability, attaching to behavior the law condemns. Hiring through the H-1B program is lawful, and the charge made no attempt to declare the activity forbidden. It simply attached a price to a lawful choice and raised money. That makes it a tax rather than a penalty. The distinction matters because the penalty characterization was one of the few ways the charge might have escaped the tax category. The court closed that route by observing that the levy did not seek to establish that the activity was illegal, which meant the only remaining classification, given that it was plainly not a cost-based fee, was a tax.

Q: What does it mean for a court to look at substance over form?

It means the court classifies a charge by its actual operation rather than by the name the government gives it. Form is the label and the packaging; substance is what the charge does in practice, who pays it, what they get, how it relates to cost, and whether it raises revenue. Under a substance-over-form analysis, a charge that operates as a tax is a tax even if labeled a fee, and a charge that operates as a fee is a fee even if labeled something else. This approach prevents the government from defeating a constitutional limit through careful word choice. For the H-1B charge, substance over form meant the proclamation’s use of the word fee was irrelevant once the charge’s revenue function became clear.

Q: Does calling the charge a national-interest measure stop it from being a tax?

No. A stated policy purpose, including protecting the domestic labor market, does not change a charge’s classification under the functional test. A charge can pursue a regulatory goal and operate as a tax at the same time, exactly as a cigarette tax both discourages smoking and raises revenue. The presence of a national-interest rationale does not subtract the revenue function from a charge that raises money on a lawful activity untethered from service cost. If a policy justification could remove a charge from the tax category, the executive could escape the taxing-power limit on any revenue measure by announcing a reason for it, which would swallow the constitutional rule. The court declined to let purpose launder the charge out of its category.

Q: Does the sheer size of the charge alone make it a tax?

Size matters as evidence, but it is not the formal test by itself. What makes a charge a tax is that it functions to raise revenue rather than to recover the cost of a service, and that it is untethered from that cost. A very large charge is strong evidence of tax character precisely because it cannot plausibly be explained as cost recovery, since no service in the program costs anything close to $100,000 to provide. But a charge could in theory be modest and still be a tax if it raised revenue without any relationship to service cost, and a charge could be sizable yet remain a fee if a genuinely expensive service justified it. For the H-1B levy, the magnitude made the untethering from cost undeniable.

Q: Could the administration redesign the charge so it qualifies as a lawful fee?

Only by abandoning its revenue purpose. To be a lawful regulatory fee, a charge must be tied to the cost of the service or regulation, sized to that cost rather than exceeding it. The executive can set such cost-based fees where statute authorizes. But a cost-based fee cannot raise the kind of revenue the proclamation sought, because cost recovery and revenue raising are mutually exclusive by definition. You cannot have a charge that both stays within the cost of adjudication and produces a large windfall to the treasury. So a redesigned charge that genuinely qualified as a fee would have to be a fraction of the $100,000 figure, which would defeat the point. The other path, a statute, is discussed separately.

Q: Why does it matter constitutionally whether a charge is classified as a tax?

Because the classification determines who may impose the charge and through what process. If a charge is a tax, only Congress may levy it, through legislation that originates in the House and runs the full legislative gauntlet of debate, votes, and accountability. If a charge is a fee, the executive may set it where a statute authorizes. The classification therefore decides whether an executive action is permissible at all. For the H-1B charge, finding it to be a tax meant it could not be imposed by proclamation, because the executive holds no taxing power. The classification was not academic. It was the difference between a valid exercise of authority and an action the Constitution forbids.

Q: How does the tax-versus-fee distinction apply outside immigration?

The distinction is general and applies to government charges across every field. Courts use the same functional markers, proportion to service cost, exchange for a discrete benefit, use of proceeds, and revenue function, to classify charges on activities ranging from regulated industries to public services. The principle that an executive or an agency may set cost-based fees but may not impose a revenue tax without legislative authority operates everywhere, not just in the visa context. This is why the H-1B ruling connects to broader disputes about executive revenue measures, including the tariff litigation, where the same question arose in the context of trade. The doctrine is a unified one, and the visa charge was simply its latest and most vivid application.

Q: What role did the purpose of raising revenue play in the classification?

Revenue purpose was the decisive feature. The functional test asks whether a charge’s function is to raise money or to recover a cost, and the court found the levy was designed to raise revenue from a lawful program. That finding, combined with the charge’s untethering from service cost, placed it in the tax category. Revenue function is what separates a tax from a fee at the most basic level, because a fee exists to pay for a service while a tax exists to fund government. The administration’s attempt to recharacterize the revenue function as a regulatory or deterrent purpose did not change the analysis, because a charge can have a policy motive and still operate to raise revenue, which keeps it a tax.

Q: How do other countries charge for skilled-worker visas without creating this problem?

They route such charges through legislation rather than executive decree. Canada and Australia pursue selectivity through points-ranked pools that choose skilled workers by criteria rather than by attaching a flat per-hire levy, and the fees in their systems are set through statute and regulation. The United Kingdom does impose a substantial charge tied to skilled-worker sponsorship, the Immigration Skills Charge, but it was created by legislation and implemented through regulations under statutory authority. Because the legislature establishes the charge in each system, no question arises of an executive imposing a revenue measure on its own. The comparison shows the proclamation’s defect was one of method. A charge of similar ambition, enacted by Congress, would not have faced the taxing-power objection.

Q: Is the United Kingdom’s Immigration Skills Charge a tax or a fee?

The British charge functions much like the kind of charge the proclamation attempted, a substantial levy on employers who sponsor skilled workers from abroad, designed in part to influence behavior and encourage domestic training. In the American constitutional vocabulary it has features of both a regulatory charge and a revenue measure. The decisive difference for the comparison is not how it would be labeled but how it was created. The Immigration Skills Charge was established through legislation and is implemented by regulations made under statutory authority, the product of the parliamentary process. Whatever its precise classification under British law, it came from the legislature, which is exactly the route the American Constitution requires for a charge of this nature and the route the proclamation bypassed.

Q: Does the functional-tax test give judges too much discretion?

The test asks judges to assess operation rather than apply a bright-line rule, which inevitably involves judgment, but its markers are concrete enough to constrain that judgment substantially. Whether a charge is proportioned to service cost, exchanged for a benefit, and aimed at revenue are answerable questions grounded in the charge’s design and the record. In close cases the test can produce disagreement, which is part of why two courts reached different conclusions on the same charge under related framings. But the H-1B levy was not a close case on the cost question. A charge exceeding service cost by more than an order of magnitude leaves little room for discretion to reach a different result. The test is principled, and its application here was constrained by the size of the gap between charge and cost.

Q: If the charge is a tax, does that make every government charge suspect?

No. Most government charges are ordinary fees that comfortably satisfy the functional test, because they are sized to the cost of a service and exchanged for a discrete benefit. Filing fees, permit charges, inspection fees, and the routine fees in the immigration system are valid fees and raise no taxing-power problem. The functional test does not threaten them; it confirms them. What the test flags is the unusual charge that abandons the relationship to cost and operates to raise revenue while claiming to be a fee. The H-1B levy was flagged precisely because it was an outlier, a charge many times the cost of the service it claimed to price. The ordinary, cost-based charges that fund government services are exactly what the test protects.

Q: What is the single clearest sign that a charge is a tax and not a fee?

The clearest sign is a large and unexplained gap between the charge and the cost of the service it claims to pay for. A fee tracks the cost of the service, so a charge that vastly exceeds that cost cannot be cost recovery and must be doing something else, which is raising revenue. That gap is what made the H-1B classification straightforward. The charge was set at $100,000 while the service cost a few thousand, a gap of more than an order of magnitude that no cost-recovery explanation could bridge. When you see a charge that dwarfs the cost of the underlying government work, the revenue-versus-cost test points to a tax, and the burden shifts to anyone claiming otherwise to explain a cost that simply is not there.

Q: Does dedicating the revenue to a specific program change whether the charge is a tax?

Not by itself. The use of proceeds is one factor in classification, because a genuine fee tends to fund the service it pays for while a tax tends to fund the government generally. But earmarking the proceeds to a program does not convert a tax into a fee when the other markers still point to a tax. A charge set far above the cost of the program is recovering nothing and raising a surplus, and a surplus on a lawful activity is revenue no matter where it is deposited. If earmarking alone could rescue a charge, the executive could immunize any revenue measure by routing the money to a program. The decisive factor remains the relationship between the charge and the cost, and the visa levy failed that regardless of where its proceeds went.

Q: Who has the burden of proving a charge is a tax or a fee?

The functional test is a judicial examination of the charge’s actual operation, not a matter of deferring to the government’s label. Courts do not start from a presumption that the executive’s characterization is correct and demand that challengers overcome it; the entire premise of the substance-over-form rule is that the label is unreliable. In practice the strength of a classification challenge tracks the strength of the functional evidence. Where the gap between charge and cost is enormous and the revenue function is plain, the challenger’s task is straightforward because the charge’s own design supplies the proof. Where a charge sits nearer the line, the analysis becomes genuinely contestable. The visa levy, with its six-figure charge on a service costing a fraction of that, offered no close call to exploit.

Q: Could a charge be partly a fee and partly a tax?

In principle a charge can have a fee component and a tax component, and courts sometimes treat the portion of a charge that recovers genuine service cost as a valid fee while treating the excess as a tax. Under that approach, the part of a charge proportioned to the cost of adjudication would stand as a fee, and only the surplus above cost would be the tax. Applied to the visa levy, this split does not save it, because the offending portion is the overwhelming majority of the charge. The few thousand dollars of genuine processing cost is a rounding error against a $100,000 charge, so the charge is a tax in all but a negligible fraction. The possibility of a mixed charge matters for closer cases, but here the revenue portion so dominates that the levy is a tax in substance and effect.

---

Working through doctrine like the tax-versus-fee distinction rewards careful note-taking and a place to keep the test, the precedent, and the comparisons in one organized set. You can save and annotate this analysis and build your own issue tracker free on VaultBook, and if you are anchoring a memo, a paper, or a syllabus on the classification, you can build a study guide and reference set on ReportMedic to keep the functional test and the supporting authority at hand.