Direct and Indirect Taxes

The fourth clause of Article I, Section 9, is known as the “Direct Tax Clause.”  It provides that “[n]o Capitation, or other direct, Tax shall be laid, unless in Proportion to the Census or Enumeration herein before directed to be taken.” This Clause requires that the financial burden of any “direct” tax imposed by Congress fall equally on each state in the Union in terms of its population. For example, if two states have the same population, then the citizens of each state collectively must pay the same amount of direct tax to the U.S. Treasury. 

While it is clear how apportionment works, it was less clear at the time of the Founding which kinds of taxes qualified as “direct,” and so were subject to apportionment. According to Madison’s notes of the proceedings of the Constitutional Convention in 1787, Massachusetts delegate Rufus King at one point “asked what was the precise meaning of direct taxation?  No one answd.”  The text of Section 9 contemplates that “capitation” taxes, otherwise known as “head” or “poll” taxes, qualify as direct taxes.  Capitations are taxes on people in simple virtue of the fact that they exist. The constitutional text also seems to imply that at least one other kind of tax qualifies as direct.

The original purpose of requiring apportionment for direct federal taxes appears to have been to benefit Southern states. Specifically, the apportionment requirement was primarily designed to render impracticable federal head taxes on slaves and federal taxes on land, both major sources of wealth in eighteenth century America. Without the requirement of apportionment by state population, the burden of both kinds of federal taxes would have fallen most heavily on the South, because it possessed disproportionately more wealth in land and slaves than did the North. By contrast, apportioning any such tax by state population would have fallen most heavily on the North, because it was home to the most populous states. 

This original purpose explains why the third clause of Article I, Section 2, provides that both “Representatives and direct Taxes shall be apportioned among the several States.” The more people who live in a state, the more representation in Congress that state receives, and the more its citizens collectively pay in direct taxes. The practical consequence of requiring apportionment for head taxes on slaves and taxes on land was that the federal government did not tax slaves or land. The Thirteenth Amendment ended the possibility of head taxes on slaves, and to this day Americans do not pay property tax to the federal government. 

From the beginning, the Supreme Court has understood only very few taxes to be subject to the apportionment requirement. In its first case considering the issue, Hylton v. United States (1796), the justices who wrote opinions included only capitation and land taxes within the category of direct taxes. Throughout the nineteenth and twentieth centuries, the Court upheld federal taxes that had not been apportioned on insurance premiums, state bank notes, inheritances, trades, personal income, and corporate income. The Court reasoned that all of those taxes were excise taxes, not direct taxes.  See, e.g., Bruce Ackerman, Taxation and the Constitution, 99 Columbia Law Review 1 (1999). 

The major exception is the Court’s decision in Pollock v. Farmers’ Loan & Trust Co. (1895), which invalidated the progressive federal income tax statute then in effect. In that case, the Court remarkably and inconsistently held that federal taxation of income that was derived from real or personal property (such as rental or dividend income) was direct and so subject to apportionment, in contrast to income taxes on wages and business profits, which were not. The Pollock Court confused matters by reasoning that certain income taxes could qualify as direct taxes on the underlying property from which the income was derived. 

The Sixteenth Amendment was ratified in 1913 in order to overrule the decision.  It authorizes Congress to tax “incomes, from whatever source derived, without apportionment of the several States,” thus tracking the previous, longstanding rule on direct taxes. The Sixteenth Amendment makes clear that income, not ownership, is being taxed, so there is no requirement of apportionment and thus no constitutional problem.

Today, the Direct Tax Clause operates to render impracticable only federal capitations, federal taxes on the ownership of land, and federal taxes on personal property. Murphy v. IRS  (D.C. Cir. 2007). For example, the minimum coverage provision in the Patient Protection and Affordable Care Act (ACA) requires most individuals to either obtain health insurance or make a payment to the Internal Revenue Service. In NFIB v. Sebelius (2012), the Supreme Court, after upholding the required payment as a tax for purposes of the Taxing Clause in the first clause of Article I, Section 8, rejected the argument that it was a direct tax and so had to be apportioned. The Court’s holding is plainly correct. The required payment for going without health insurance in the ACA is a tax on those who choose to remain uninsured, not a head tax on those who simply exist, or a tax on land ownership, or a tax on personal property. It is therefore not a direct tax and need not be apportioned.

Excepting Pollock, the Court has been right all these years to define the category of direct taxes very narrowly. As discussed, a narrow definition is most consistent with the constitutional text, the original purpose of the Direct Tax Clause, the Court’s precedent, and the Sixteenth Amendment. 

A narrow definition also makes good sense from a consequentialist perspective. The Constitution confers robust federal power to tax on the theory (grounded in experience) that taxation with representation is a necessity. Requiring apportionment, however, renders federal taxation impracticable. 

Apportionment also requires the federal government to privilege regressivity over progressivity in taxation. Apportionment means that citizens of relatively wealthy states must pay at lower rates than citizens of relatively poor states in order to make the total payment for states of equal population come out the same. Such a tax regime is difficult to defend morally, particularly in an America in which the income distribution is increasingly skewed in favor of a very small number of extraordinarily wealthy Americans. 

Finally, such a tax regime is difficult to defend from a structural, federalism perspective.  States that impose progressive income taxes are at a competitive disadvantage in attracting wealthy residents relative to states that do not.  See, e.g., Akhil Reed Amar, America’s Constitution: A Biography 408 (2005).  Only progressive taxation at the federal level can overcome the collective action problem and avoid a destructive “race to the bottom.”

The direct/indirect tax dichotomy remains important because it affects the types of tax the federal government can impose. Classification between the two categories, as well as application of the apportionment and uniformity tests can determine the validity of modern statutes.

Very difficult to satisfy, the apportionment requirement can be met only with a capitation—a direct tax on humans simply because they are humans. To satisfy apportionment, the tax would necessarily be the same per person nationwide. Congress has never enacted such a tax and arguably is unlikely to do so in the foreseeable future. That equality requirement is a powerful restriction on the taxing power: remember, the income tax is progressive and applies much more heavily on high-income persons than on others. Because the income tax is not subject to apportionment—largely because of the Sixteenth Amendment—progressivity is possible. In addition, Congress cannot impose a property tax on land. Apportioning such a tax would be impossible because the amount of land per person is not the same in every state. Land taxes—known as ad valorum and “property taxes”—however, are very important to the states, with many states raising a substantial portion of their revenue from them. Restricting such taxes to the states is another very significant restriction on the federal taxing power.

Although Congress cannot impose a property tax directly on personal property—such as cars, furniture, stocks or bonds—as opposed to land—, it has two fairly easy work-arounds for such taxes. NFIB v. Sebelius (2012). Since 1796, the Supreme Court has viewed a tax on the use of personal property as an indirect tax subject to uniformity rather than to apportionment. Hylton v. United States (1796) (Chase, J.). As a result, it could easily style a tax on automobiles as a tax on the use of the item and thus avoid apportionment. Indeed, because the number of cars (or any other personal item) is unlikely ever to be the same per capita in every state, without the Hylton decision, a federal personal property tax would be impossible because it could never satisfy apportionment. Although Congress does not often impose direct taxes on personal property, Congress routinely imposes similar taxes on the purchase of personal property such as tires and gasoline, styling them as excises subject merely to uniformity. Significantly, it could impose a nationwide automobile or telephone usage tax.

Until 1913, a tax on either personal or real property income was effectively forbidden because such taxes were considered direct and not easily apportioned. Pollock v. Farmers’ Loan & Trust Co. (1895). The Sixteenth Amendment resolved this by replacing the income tax apportionment requirement with a new requirement that a tax on income need not be apportioned so long as the tax is imposed on income “derived from a source”—a serious, but different restriction.

In 2012, the Supreme Court narrowly read the direct tax definition in relation to the Affordable Care Act (ACA). The Court held the ACA penalty on persons without minimum health insurance to be a tax; however, the Court also held it not to be “any kind of direct tax.” NFIB v. Sebelius (2012). In so doing, the Court made two critical points. First, by citing Pollock favorably, the Court removed any argument that the “penalty” was supportable as an income tax subject to the “derived” test. That was important because the “tax” is, in part, a function of a person’s income. Without the Sixteenth Amendment as a possible foundation, the tax/penalty must satisfy either apportionment or uniformity. Second, the Court relied on the often-quoted 1796 Hylton language: a direct tax is one imposed “without regard to property, profession, or any other circumstance.” Finding the lack of health insurance an “other circumstance,” the Court found that the mandate to purchase insurance was not a direct tax, and it rejected apportionment as applying to the ACA.

Interestingly, the Court quoted Justice Chase out of context. Chase actually said: “I am inclined to think, but of this I do not give a judicial opinion, that the direct taxes contemplated by the Constitution are only two, to wit, a capitation or poll tax simply, without regard to property, profession, or any other circumstances, and the tax on land.” Hylton v. United States (1796) (Chase, J.). The NFIB Court thus omitted Chase’s independent clause and quoted only the dependent clause, which he described as not his “judicial opinion”; however, it labeled the quotation as “opinion of Chase, J.” which is best described as misleading. Nevertheless, the NFIB decision appears settled: the ACA tax is neither a direct tax nor an income tax.

But discussion of a tax which is neither direct nor an income tax necessarily raises the issue of uniformity. Uniformity is a much easier-to-satisfy requirement than apportionment.  It is all about geography: the tax must apply the same in every state.

Few Supreme Court decisions have applied uniformity and none has invalidated a tax because of it. Essentially, the geographic element of uniformity applies in a laxer manner if the natural geographic variations among states justify a lack of uniformity or if a particular state’s behavior causes it. For example, the Court upheld a charge on immigration through sea ports though it had no impact on land-locked states, finding that there was substantial uniformity. Edye v. Robertson (Head Money Cases) (1884). Similarly, in 1983 the Court upheld a difference between taxes on income from oil pumped above the arctic circle and taxes on income from oil pumped elsewhere, noting that there were important natural differences in the oil and the extraction process. United States v. Ptasynski (1983). In 1927, the Court upheld the estate tax although it ultimately applied differently to Florida. Florida v. Mellon (1927). The difference, however, resulted from a particular Florida statute which created the lack of uniformity. Thus a state cannot self-impose non-uniformity and then complain about it.

Significantly, the NFIB decision did not discuss “uniformity.” The Court likely avoided addressing this issue because it was premature: because the Act was not yet operable, no facts indicating uniformity or the lack thereof were available. Since 2014, however, the tax on persons without health insurance has applied differently in the various states.  It varies as a function of the cost of health insurance in each state, as well as the federal “poverty level,” which itself varies between Alaska, Hawaii, and other states. As a result, the ACA tax can vary from one person to another even though the two persons may reside less than a mile apart. The only important distinction may be the man-made political boundary between them: the state line.  Whether the Supreme Court will hear a case challenging the ACA for lack of uniformity is unclear. If it does, that may be the first instance in which the constitutional restriction has a real impact.