Concept
Valid Legislation Distinguished
Bills of Attainder
Valid Legislation Distinguished
The Bill of Attainder Clause does not forbid a legislature to burden people it can name. Article I, section 9 provides that “No Bill of Attainder or ex post facto Law shall be passed,” and section 10 lays the same ban on the states.1 A bill of attainder is a legislative act that punishes specified persons without a judicial proceeding, and all three features must be present.2 A statute that burdens identifiable people therefore raises the question of punishment without answering it.
Legislatures govern by drawing lines, and the people on one side of a line are often known in advance. A tax, a licensing standard, or an eligibility rule may leave a small, listable group worse off. If every such burden were an attainder, the clause would ban legislating, because no legislature can try each person its statutes disadvantage. The clause aims instead at legislative punishment, and burden and classification trigger it only when they carry punishment with them.
Three situations mark the line: a legitimate class of one, the denial of a governmental benefit, and a regulatory disqualification. In each, the legislature knows whom the law will reach, so naming or listing the affected people settles nothing. What decides each case is whether the burden is punishment, which the historical, functional, and intent inquiries from the preceding sections measure.
Class of One
Neither section says how narrowly a statute may be aimed. In Nixon v. Administrator of General Services, Congress lodged a former President’s materials in the custody of the General Services Administration, pending screening by government archivists and further regulations, and the Act referred to him by name. The Court held that the Act’s specificity “does not automatically offend the Bill of Attainder Clause.”3 Reference to a single person is a warning that sends the court on to the questions of punishment and trial, and it decides nothing by itself.
The former President argued that an individual or group is attainted whenever it must bear burdens it dislikes, and he faulted the Act for taking custody of his papers but not those of other Presidents.4 The Court answered that this view proves far too much. It removes “the anchor that ties the bill of attainder guarantee to realistic conceptions of classification and punishment,” and it would cripple the process of legislating, because anyone made the subject of adverse legislation can complain that the lawmakers should have defined the affected class more generally.5
The complaint has no stopping point. Even a statute reaching all Presidents could be called overly specific, since the class might have included the whole Executive Branch, all members of the Government, or all who hold Presidential or Government papers.6 The breadth of a statute’s focus therefore cannot decide the reach of the clause.
The Court also placed the argument in its proper doctrinal home. However expansive the ban on attainders, it was not intended to serve as a variant of equal protection, invalidating every act that burdens some persons or groups but not all other plausible ones.7 The former President had first pressed an equal protection theory and then abandoned it, apparently recognizing that mere underinclusiveness is not fatal under the equal protection component of the Fifth Amendment.8 The clause remains an important “bulwark against tyranny,” in the words of United States v. Brown, but it does not limit Congress to legislating for the universe, legislating only benefits, or not legislating at all.9
Facts, not breadth, made the focus on one person legitimate. Congress had reason for concern solely with preserving his materials, because he alone had entered into a depository agreement that by its terms called for destroying some of them; the government argued that the agreement created an imminent danger that the tape recordings would be destroyed if he, who had contracted phlebitis, were to die.10 The papers of all other past Presidents were already safely housed in libraries.11 The Court therefore held that he “constituted a legitimate class of one,” which gave Congress a basis to proceed with dispatch as to his materials while accepting the status of his predecessors’ papers and ordering further consideration of general standards for his successors.12
A legitimate class of one answers only the specification question. The Court went on to say that, even if the specificity element were satisfied, the clause would not automatically apply, because forbidden legislative punishment is not involved merely because an Act imposes burdensome consequences; the further question was whether lodging the materials with the General Services Administration “inflict[ed] punishment.”13 It did not. The Act merely set a policy that historical materials be preserved, and a law is an attainder only when it both singles out a person or group for unfavorable treatment and imposes punishment.14 The unique fact did two jobs, then: it explained why Congress acted on one person, and it fit the burden Congress chose, custody for screening. It would rescue no statute whose burden is itself a penalty.
Set the pension forfeiture beside the Nixon facts, and the difference lies in what the unique fact explains. In Nixon, a depository agreement put the materials at risk, and that risk explained why custody was needed. In the pension statute, the only distinguishing fact is the legislature’s condemnation, which explains nothing but the penalty. A narrow focus explained by the problem Congress is addressing is legitimate; a narrow focus explained only by the legislature’s belief in the target’s guilt is the mark of a sentence.
Conflict-of-interest laws teach the same lesson from the side of groups. Such laws inevitably prohibit conduct by designated individuals or classes without offending the clause, as § 32 of the Banking Act of 1933 did when it disqualified officers and employees of underwriting organizations from serving as officers of Federal Reserve banks.15 Identifiability is a normal feature of lawmaking, whether the target is one person or a short list. The question is whether the burden is punishment imposed without a trial, and the section on regulatory disqualifications takes that question up for groups.
Benefit Denials
A legislature can disadvantage identifiable people through the programs it funds as readily as through the sanctions it writes, and the clause reaches a withdrawal of benefits only when the withdrawal is punishment.17 The ban binds Congress through section 9 and the states through section 10, so the whole question is whether ending or conditioning a benefit inflicts punishment.18 Two decisions answer it. In Flemming v. Nestor, Congress terminated a deportee’s Social Security benefits under § 202(n), and the Court rejected the claim that the termination was a bill of attainder.19 In the draft-registration case, Selective Service System v. Minnesota Public Interest Research Group, the Court reached the same result for a statute denying federal student aid to students who had not registered, and together the two decisions show where a benefit rule ends and punishment begins.20
Flemming first treated the cutoff as an ordinary classification. A deportee’s presumably permanent residence abroad bears obviously on eligibility, because one benefit the economy draws from Social Security is the national purchasing power created by paying retirees and disabled people who generally spend a comparatively large percentage of their benefits, and that advantage is lost on payments made abroad. The Court added that it is constitutionally irrelevant whether this reasoning actually underlay the legislative decision, and irrelevant that the section does not extend to everyone to whom the rationale might apply.21 A rational classification does not dispose of an attainder claim, however, because a rational law can still be punishment. The Court therefore asked whether terminating the benefits was punishment in the constitutional sense, the characterization on which the attainder, trial, and ex post facto claims all depended.22
The answer turned on the object of the legislature’s concern. Where the source of concern can be thought to be the activity or status from which the individual is barred, the disqualification is not punishment even though it may bear harshly on the person affected; the contrary is true where the statute is evidently aimed at the person or class disqualified.23 Cummings v. Missouri showed the contrary case: the oath there was required “to reach the person, not the calling,” because the listed acts were thought to deserve punishment and not because they showed unfitness. The plurality in De Veau v. Braisted put the test as a question: whether the legislative aim was to punish the individual for past activity, or whether the restriction comes about as a relevant incident to the regulation of a present situation, such as the proper qualifications for a profession.24 Measured that way, the termination of the deportee’s benefits was not punishment. The draft-registration Court later summarized Flemming: the sanction was “the mere denial of a noncontractual governmental benefit,” no affirmative disability or restraint was imposed, and Congress had inflicted “nothing approaching the ‘infamous punishment’ of imprisonment.”25
Severity cannot supply the distinction. A retiree who loses a monthly payment may lose a great deal, yet the severity of a sanction is not determinative of its character as punishment.26 Courts look instead for the legislature’s reason, in the record and in the explanations the law can bear. In Lovett, which invalidated a statute forbidding further payment of the salaries of three named government employees, the finding of punishment rested in large measure on the specific congressional history, and the plurality in Trop v. Dulles rested its characterization, at least in part, on its inability to discern any alternative purpose the statute could serve.27 A benefit cutoff with a plausible program-based explanation and a record that contains no verdict on the person stays on the regulatory side.
The challenger’s objection deserves a fair statement. He argued that the termination punished him without a judicial trial, imposed punishment by legislative act, and rested on conduct that was not unlawful when engaged in.28 The objection has force, because Lovett also involved nothing more than the stoppage of payments. The Court answered by accepting that the loss is real and resting on two features: the benefit was noncontractual, and the concern behind the cutoff could be thought to be a status relevant to eligibility. On that view, the loss of an expected payment, however painful, does not turn a program rule into a verdict.
The draft-registration case applies the same inquiry to a condition instead of a withdrawal. Section 12(f) denied federal student aid to students who had not registered with the Selective Service. The district court read the statute to reach students who registered late, and on that reading it found that the statute singled out an ascertainable group based on past conduct, legislatively determined that group’s guilt, and punished it by depriving students of the practical means to pursue an education.29 The Supreme Court did not accept that reading. Because § 12(f) allows late registration, it is “clearly distinguishable” from the provisions struck down in Cummings and Garland, which were absolute barriers to professions for those who could not file the required loyalty oaths; no one who had served the Confederacy could comply, for his status was irreversible. Section 12(f)’s requirements, by contrast, can be met readily by timely or late filing.30
The Court then ran the three inquiries, framed by two cautions. Severity does not decide whether a sanction is punishment, and a legislative aim to encourage compliance does not establish that a statute is mere regulation, since punishment is not limited to retribution for past events and may include deprivations inflicted to deter future misconduct.31 The second caution keeps a legislature from escaping the clause by announcing a forward-looking goal. Under the historical inquiry, section 12(f) imposed none of the burdens associated with punishment: the sanction was the denial of a noncontractual benefit, not imprisonment, banishment, punitive confiscation of property, or a bar from a profession.32 Under the functional and intent inquiries, the law reasonably furthered nonpunitive goals by limiting federal aid to those who meet their responsibilities by registering, and the legislative history gave convincing support for the conclusion that Congress sought to promote compliance with the registration requirement and fairness in allocating scarce federal resources.33
The two decisions fit together once curability is placed correctly. The deportee in Flemming could not undo the status that ended his payments, and the cutoff survived because the concern could be thought to be a status relevant to the program. The students could end the denial by complying, and that feature helped the government without being the whole explanation, because the Court still ran the punishment inquiries. An analysis that stops at the observation that a group defined by past conduct lost something has located only the specification element, and the work continues through the punishment inquiries.
Regulatory Disqualifications
Legislatures routinely decide who may hold a particular position, and because the holders of a position are usually few and known, a disqualification often falls on an identifiable group. United States v. Brown shows both outcomes. The Court there invalidated a law declaring it unlawful for a member of the Communist Party to hold employment in a labor union, reasoning that the law impermissibly designated a class of persons for punishment.39 The same decision made clear that conflict-of-interest laws, which inevitably prohibit conduct by designated individuals or classes of individuals, do not contravene the guarantee, and it noted § 32 of the Banking Act of 1933, which disqualified officers and employees of underwriting organizations from serving as officers of Federal Reserve banks.40
The doctrine tolerates such disqualifications because conflict-of-interest rules work by excluding the very people who occupy the conflicting roles. A clause that condemned every disqualification of identifiable persons would leave legislatures unable to enforce any such rule or to set standards for a licensed calling. A disqualification that passed whenever the legislature labeled it regulation, on the other hand, would let a majority write a sentence in the form of an eligibility rule. The Court reconciles the two concerns by asking whether the designation of persons based on past conduct simply names individuals for punishment or promotes a nonpunitive goal based on reasonable criteria over which the individual has some control.41
That formulation has two working parts, and each guards against a different abuse. The first is a reasonable criterion, one that tracks a trait or situation bearing on the position. A person who is an officer of an underwriting organization and also an officer of a Federal Reserve bank serves two interests that may diverge, so the conflict explains the Banking Act provision whoever the officers happen to be. When a legitimate goal accounts for the burden in that way, the inference that the legislature meant to punish loses its footing, which is the logic of the functional test from earlier in this chapter. The second part is control. Someone who wants the Reserve bank post can end the disqualification by leaving the underwriting organization, so the rule operates as a standard that present choice can satisfy. The early oath cases supply the contrast: a condition counts as a qualification only if it is attainable, in theory at least, by everyone, and what a person’s past condition or conduct puts beyond reach is a disqualification that operates as a perpetual bar.42
The De Veau question from the benefit cases carries over directly: was the legislative aim to punish past activity, or is the restriction a relevant incident to regulating a present situation, such as the proper qualifications for a profession?43 A disqualification aimed at the position is regulation even when it bears harshly on the person affected. One evidently aimed at the person is punishment, as in Cummings, where the oath was exacted “to reach the person, not the calling.”44
Two illustrative statutes in the medical field show the sorting at work; neither is a decided case. A law declaring that every doctor who ever worked at one named hospital may no longer practice medicine could be struck down as an attainder, because the past conduct that defines the group serves only to identify persons for punishment and measures no trait that makes anyone unfit to practice.45 A law prohibiting doctors previously convicted of medical malpractice from continuing to practice might be upheld as the establishment of nonpunitive licensing criteria.46 Both laws look backward, burden an identifiable class, and end a medical career, so none of those features separates them. What separates them is the fit between the criterion and the legitimate objective of protecting patients. A malpractice conviction records the doctor’s own conduct in the very work the license covers, and it follows a judicial finding of culpability. Employment at a particular hospital rests on no finding about any doctor, so the legislature’s own condemnation is the only explanation left for the bar.
Two decisions upholding disqualifications show that the form of the measure settles nothing and the legislature’s focus does. The Court refused to invalidate a municipal ordinance requiring employees to swear that they had never belonged to the Communist Party or a similar organization advocating violent overthrow of the government, because it believed the ordinance only established eligibility for employment. It also sustained a state law keeping convicted felons from holding positions with longshoremen’s unions, observing that the prohibition rested on a judicial determination of culpability and created no further implication of guilt.47
A conflict-of-interest label does not end the analysis, because a legislature can call a measure a qualification and still impose a sentence, as the oath cases show. The question remains whether the criterion fits a real concern about the position and whether the person can meet it. A disqualification that clears both, aimed at a present situation, is regulation even though the legislature knew exactly whom it would reach.