Concept
Tax Versus Penalty
Taxing power
Tax Versus Penalty
An exaction can clear both screens from the earlier sections, a label that does not control and some revenue, and still fall outside the taxing power. Article I, section 8 lets Congress “lay and collect Taxes, Duties, Imposts and Excises” to pay the debts and provide for the common defence and general welfare, and those words describe a power to raise money for the government.1 They do not describe a power to punish. A statute that collects money but operates mainly as a sanction has no footing in the clause, however much money reaches the Treasury.
The question is therefore functional: does the exaction work as a charge a person pays for engaging in an activity, or as punishment for an unlawful act? The Court defines a penalty in just those terms: “if the concept of penalty means anything, it means punishment for an unlawful act or omission.”2 The label cannot answer the question. As the Court said in a Sixteenth Amendment case, “Congress cannot by any definition it may adopt conclude the matter, since it cannot by legislation alter the Constitution.”3 The inquiry matters because the taxing power is a separate grant. Congress may sometimes tax what it could not regulate directly, so without a penalty inquiry a tax label would become a way around the limits on every other power.
Two features of the doctrine shape any answer. The default favors the exaction: “Every tax is in some measure regulatory,” the Court said in Sonzinsky v. United States, and it refused to inquire into Congress’s “hidden motives.”4 The current Court likewise is reluctant “to closely examine the regulatory motive or effect of revenue-raising measures.”5 Yet the default has an edge. The cases below supply the features that separate a charge from a sanction: how heavy the exaction is, what triggers it, who enforces it, and where its proceeds go.
The Child Labor Tax Case
The Child Labor Tax Case (Bailey v. Drexel Furniture Co., 1922) is the classic revenue-raising exaction held to be regulation in tax clothing. It followed Hammer v. Dagenhart (1918), which had held that the commerce power did not allow federal regulation of goods made with child labor, a reading the modern Court has thoroughly discredited.6 Congress then imposed an exaction of 10 percent of net income on those who knowingly employed children.7 The Court held that Hammer controlled the question of federal power over matters reserved to the states, perceived the tax law as an attempt to make an end run around Hammer, and found the exaction a penalty and not a tax.8
The logic of the decision follows from the separateness of the powers. Had the tax label controlled, a charge on the very conduct that Hammer put beyond the commerce power would have stood, and Hammer would have protected nothing. United States v. Butler later stated the governing principle in terms that fit taxation: “the attainment of a prohibited end may not be accomplished under the pretext of the exertion of powers which are granted.”9 The idea is familiar from McCulloch v. Maryland. The Court has since described the Child Labor Tax Case and Hill v. Wallace as involving statutes that “purported to be taxing measures” but whose purpose was to regulate conduct in the states, with “the levy of the tax a means to force compliance,” and it called that use “not a constitutional use, but an unconstitutional abuse of the power to tax.”10
NFIB later distilled three practical characteristics from the case, which it called Drexel Furniture, as the reasons the “tax” was actually a penalty:11
- The tax imposed an exceedingly heavy burden, 10 percent of a company’s net income, on those who employed children “no matter how small their infraction.”
- It applied only to those who knowingly employed underage laborers. “Such scienter requirements are typical of punitive statutes, because Congress often wishes to punish only those who intentionally break the law.”
- The Department of Labor, “an agency responsible for punishing violations of labor laws, not collecting revenue,” helped enforce it.
The case had other features that NFIB did not need to name. The child-labor tax regulated a detailed set of workplace practices, reached behavior broadly condemned as inherently immoral, and arrived close on the heels of the invalidation of a purely regulatory measure.12
Each feature points toward punishment for a reason. An ordinary excise works as a price: it attaches to an activity, and what the taxpayer owes tracks how much of the activity the taxpayer chooses to do. A charge of a tenth of net income, owed however small the infraction, bears no relation to the amount of the taxed conduct, so it operates as a fine. A knowledge requirement fits the same pattern. Enforcement by the Labor Department shows what the statute was built to accomplish. And a charge keyed to a long list of workplace standards looks like enforcement of those standards, because the employer owes it only by failing to follow the rules. No single feature decides the question. The Court found that the features together pointed one way.
Congress’s label does matter for some purposes, so keep the constitutional question separate from statutory ones. When the question is how two statutes relate, the best evidence of Congress’s intent is the statutory text, and the Court has therefore applied the Anti-Injunction Act to statutorily described “taxes” even where the label was inaccurate, including the “Child Labor Tax” it had struck down as exceeding the taxing power.13 Statutory text decides whether the Act applies. Substance decides whether an exaction is a tax in the Constitution’s sense.
Sonzinsky shows a genuine excise. A firearms dealer challenged the National Firearms Act of 1934, which required every dealer to register with the Commissioner of Internal Revenue and to pay a special excise tax of $200 per year.14 The Court sustained the levy, observing that a tax is no less a tax for having a regulatory effect, and it distinguished the Child Labor Tax Case and Constantine.15 The firearms levy lacked the detailed regulation of the child-labor statute and the criminal trigger of the Revenue Act at issue in Constantine.16 A regulatory aim therefore cannot separate the cases. What separates them is how the exaction is built and triggered.
Constantine and Criminal Triggers
United States v. Constantine (1935) isolated a cleaner signal in a simpler statute. The Revenue Act of 1926 imposed a $25 annual tax on every retail liquor dealer and added a $1,000 annual tax on dealers who operated in violation of state law.18 The Court held the added $1,000 a penalty and not a true tax.19 NFIB later cited the case for two considerations: the amount of the exaction, and the fact that it was imposed for violation of a separate criminal law.20
Each consideration has its own logic. The trigger matters because a tax ordinarily works as a price: a dealer pays it and carries on lawfully, so payment buys the freedom to act. A charge that falls only on dealers already breaking the law cannot work that way, because paying it does not make the business lawful. It sits on top of the violation like a fine, and the taxing clause supplies no power to fine. The amount supplies evidence of purpose. The lawful dealer paid $25, which marks what Congress valued the privilege of selling liquor at for revenue, and the violator paid forty times that. Revenue goals do not readily explain the extra sum, so deterrence and punishment remain. The comparison also gives a court a yardstick drawn from the statute itself, which spares judges from deciding how much revenue Congress needed.
These signals track the general definition of a penalty. United States v. La Franca called a penalty “an exaction imposed by statute as punishment for an unlawful act,” and NFIB repeated the idea.21 Unlawfulness is therefore the core of the definition, and the amount is evidence of what Congress meant the exaction to do. Justice Cardozo, joined by Justices Brandeis and Stone, read that evidence differently: “Not repression, but payment commensurate with the gains is . . . the animating motive.”22 The dissent shows that treating a large amount as punishment is an inference about purpose, and reasonable judges can resist it.
Sonzinsky marks the other side of the line. The firearms dealer there owed a flat $200 for doing business, with no added sum that depended on any violation, and the Court distinguished Constantine on the ground that the Revenue Act carried a criminal trigger.23 Regulatory aim cannot separate the two, since every tax is in some measure regulatory. The firearms dealer owed the charge for doing business, and the liquor dealer owed the extra $1,000 only for doing business unlawfully.
NFIB shows the same definition working in favor of the taxing power. The shared responsibility payment aimed to induce people to buy insurance, but the mandate need not be read to declare that going without coverage is unlawful, because neither the Act nor any other law attaches negative legal consequences to that choice beyond the payment to the IRS.24 The Government agreed that a person who pays rather than obtains insurance has fully complied with the law.25 That concession separates the mandate from the liquor levy. The dealer who paid the added $1,000 was still breaking state law, while the individual who paid the IRS had done what the statute allowed. An estimated four million people a year were expected to choose payment, and Congress’s tolerance of that number suggests that it did not think it was creating four million outlaws.26
Earmarked and Coercive Exactions
Where the money goes supplies a third signal. In Butler the Court wrote that a tax, “in the general understanding of the term, and as used in the Constitution, signifies an exaction for the support of the Government,” and that the word “has never been thought to connote the expropriation of money from one group for the benefit of another.”28 It concluded that the processing tax before it was “a mere incident” of an act regulating agricultural production.29 Charles C. Steward Machine Co. v. Davis later summarized the point: the so-called tax was “not a true one,” because its proceeds were earmarked for the benefit of farmers who complied with the prescribed conditions.30
Failing as a tax ends the taxing-power defense but not necessarily the measure. Butler conceded that an imposition of that kind is constitutional when imposed to effectuate regulation of a matter in which both groups are interested and over which Congress has power of legislative regulation.31 A measure that fails as a tax must therefore find support in a regulatory power.
Read Butler with care. Its holding on the scope of Congress’s taxing and spending power remains good law, but its ruling on the Tenth Amendment is no longer followed.32 This section uses the case only to show how the Court once distinguished a true tax from an exaction that serves as an instrument of regulation.
Justice Stone’s dissent shows that the classification can be contested. For him the processing tax differed from the penalties in the Child Labor Tax Case, Hill v. Wallace, Linder v. United States, and Constantine, which “were themselves the instruments of regulation by virtue of their coercive effect on matters left to the control of the states.” Here, he wrote, “regulation, if any there be, is accomplished not by the tax but by the method by which its proceeds are expended.”33 The majority and the dissent thus share the principle that an exaction which itself coerces is not a tax. They differ over whether the processing tax coerced or merely funded a program whose spending did the regulating.
Steward Machine (1937) upheld an excise on employers that allowed a credit for contributions to a state unemployment fund. The challenger relied on Butler, and the Court summarized that decision as resting on three objections: the so-called tax was not a true one because its proceeds were earmarked for complying farmers, the plan attempted to regulate production without the consent of the state in which production was affected, and the payments to farmers were coupled with coercive contracts.34 The Court answered the first objection by finding that the proceeds of the employer excise were not earmarked for a special group.35 The other two objections concern state authority and coercive conditions, and they fall outside this section.
The Court then marked the line between two kinds of tax. It is one thing, it said, to impose “a tax dependent upon the conduct of the taxpayers” where the conduct to be stimulated or discouraged is “unrelated to the fiscal need subserved by the tax in its normal operation,” and the Child Labor Tax Case and Hill v. Wallace were decided in the belief that those statutes were exposed to that reproach. It is “quite another thing to say that a tax will be abated upon the doing of an act that will satisfy the fiscal need, the tax and the alternative being approximate equivalents,” and in that case “inducement or persuasion does not go beyond the bounds of power.”36 The credit for state contributions fit the second description.
Carter v. Carter Coal Co., decided the same year as Butler, applied the reasoning to a coal “tax.” The Government rested the exaction’s validity on the commerce power and not the taxing power, and it agreed that the “tax” must fall if the act’s labor and price-fixing provisions were not upheld. The Court agreed.37 Once validity rested on the commerce power, the taxing power dropped out of the case, and the whole scheme stood or fell with its regulatory provisions.
Butler and Steward Machine both involved measures that hoped to advance a collateral policy, so a collateral aim cannot separate them. The difference lies in where the money went and in whether a levy or an exemption forced compliance. In Butler the proceeds flowed to the farmers who complied with the plan. In Steward Machine the excise singled out no group to receive its proceeds.
Modern Deference and Its Limits
NFIB describes two eras. In the first, “[a] few of our cases policed these limits aggressively, invalidating punitive exactions obviously designed to regulate behavior otherwise regarded at the time as beyond federal authority,” and the Court cites Butler and the Child Labor Tax Case as examples. “More often and more recently,” the Court has “declined to closely examine the regulatory motive or effect of revenue-raising measures.”39 The later cases sustained excises on firearms, marijuana, and wagering.40 The penalty argument therefore remains available on an exam, but the default result is that the tax stands.
The shift tracks the commerce power. A regulatory tax threatens the enumeration of powers only when it reaches something Congress could not regulate directly. The Child Labor Tax Case and Constantine were decided when the Court, under the enclave theory, read Congress’s regulatory powers narrowly and treated matters reserved to the states as beyond federal reach.41 The commerce power grew, and the growth began early. In Swift & Co. v. United States (1905), the Court treated local stockyard bidding by meat dealers as a temporary stop in the interstate sale of cattle.42 The 1937 decisions upholding Social Security Act taxes, Steward Machine and Helvering v. Davis, reflect the same turn. Today Hammer’s reading is thoroughly discredited, federal regulation of goods made with child labor is valid, and a federal tax on the same subject would likely be sustained.43 As the commerce power broadened, fewer subjects remained that Congress could reach only through a tax,44 and calling a regulatory tax a penalty gained little. The most likely result of that label is that the measure is sustained under the commerce power instead.45
A second reason concerns who decides. The alternative to restraint is a court ranking the competing motives of a legislature and striking down a statute on the strength of the ranking. The older answer to a complaint that a tax is oppressive points elsewhere. The power to tax may be exercised oppressively, but “the responsibility of the legislature is not to the courts, but to the people by whom its members are elected,” and a tax that bears heavily upon a class “cannot, for that reason only, be pronounced contrary to the Constitution.”46
The deference has an edge, and NFIB marks it. The Court maintained, quoting Kurth Ranch and the Child Labor Tax Case, that “there comes a time in the extension of the penalizing features of the so-called tax when it loses its character as such and becomes a mere penalty with the characteristics of regulation and punishment.”47 The Court did not say where that time arrives. The shared responsibility payment lay well inside the line, so the Court left the precise point undecided, and it added that the “power to tax is not the power to destroy while this Court sits.”48 The two statements fit together. The principle that taxation can destroy is pertinent only where there is no power to tax the subject at all, and the weight of a genuine tax on a proper subject is a matter for the voters.49 Heaviness alone therefore does not invalidate a levy, though it remains evidence in the penalty inquiry.
A gun-free school zones tax shows where deference and the residual limit meet.
The older cases point toward penalty. Congress acted right after the Court invalidated a purely regulatory measure aimed at the same behavior, the same pattern as in the Child Labor Tax Case. The school-zone limit can be read as the “detailed and specified course of conduct” that the earlier decision condemned, and the $10,000 charge looks so exotic as a matter of taxing tradition that it differs in kind from the orthodox firearms, marijuana, and wagering excises.50 Under those cases the odds are high that the Court would call the “tax” a penalty, because its regulatory effect and purpose are palpable.51
Several features point the other way. The charge has none of the detailed regulations of the Child Labor Tax Law, and its trigger is a single act, possession in a school zone, which resembles the general description of the taxable activity in Sonzinsky.52 Whether the lack of detail is enough to distinguish the child-labor statute depends on the deference due Congress, and the current Court has reaffirmed its reluctance to examine regulatory motive or effect.53 Constantine supplies a second comparison. The charge is not tied by its terms to a separate crime, so the criminal-trigger signal is weaker than it was there, though the size of the charge and its role as a substitute for the invalidated ban give a reasonable case for applying that decision.
The consequence of the label is also harder to predict than it once was. If the charge is a penalty, the taxing power cannot support it, and the commerce power cannot either, because Congress may not use that power to penalize noneconomic conduct of this kind.54 The usual assumption that a penalty finding only moves the measure to the commerce power weakens after Lopez, Morrison, and NFIB narrowed that power, so the label could decide the outcome and not merely relabel the measure.55 A strong exam answer sets out the features pulling each way, the timing and size against the single-act trigger and the Court’s reluctance to examine motive, and then commits to a conclusion while acknowledging that the question is open.