Concept
Limits on Federal Criminal Power
Constitutional Power to Exercise Criminal Jurisdiction
Limits on Federal Criminal Power
A statute that fits within a delegated power has cleared one hurdle, and the hook test from the preceding sections measures only that hurdle. The hook comes first because the Constitution's express prohibitions on government operate only where the government has authority to act in the first place, so a statute with no enumerated source fails before any prohibition is consulted.1 A statute that does have a source faces a second round, since every power the Constitution confers is limited in three ways: by its own defined scope, by structural principles such as federalism and the separation of powers, and by external guarantees such as those in the Bill of Rights.2 The hook test answers the first of those questions. This section turns to the other two, which appear here as three further checks, each aimed at a different way that delegated power can be misused. A hook is therefore necessary but not sufficient.
The first check protects the federal structure. Congress may regulate private individuals directly under a power such as the commerce power, but it may not order state executive officers to administer a federal program, because the Tenth Amendment reserves to the states the powers that the Constitution did not delegate.3 The second check protects individuals from punishment that skips the ordinary safeguards. Article I § 9 bars bills of attainder and ex post facto laws, and the prohibition on retroactive liability reaches criminal statutes that a valid hook otherwise supports.4 In Marbury v. Madison, Marshall pointed to these express prohibitions, together with the proof requirements for treason, as provisions that courts must apply in preference to a contrary statute.5 The reason for such a ban is that a legislature able to declare a named person guilty, or to declare past conduct criminal, would be doing work the Constitution assigns to the courts, and a legislature cannot sit as a judge of cases.6
The third check closes a possible back door. A treaty approved by the Senate has the same force and effect as a law enacted by Congress.7 Marshall's reading of the Supremacy Clause makes a statute supreme law only when it is made in pursuance of the Constitution, and a treaty that carries no more force than a statute gains no wider freedom from constitutional restraint.8 Without that rule, the President and Senate could reach by agreement with a foreign nation a result that Congress could not reach by passing a bill, and any of the limits above could be avoided by routing the same program through foreign affairs. The section on treaties develops this point through Reid v. Covert.
The checks operate independently, so a statute can clear the hook test and one check while failing another. That independence follows from the different purposes they serve: the federalism check protects state governments from being conscripted, the retroactivity bar protects the individual facing prosecution, and the treaty rule protects the Constitution's authority over every other source of law. The fact that carries the answer is usually a single feature of the statute, which the paired examples below isolate.
The amended statute reproduces the structure of Printz v. United States, where the Brady Act regulated handgun sales and the Court held that directing state law enforcement officers to perform background checks violated the Tenth Amendment.13 The subject of the Act, the handgun trade, was one that Congress could reach, as the preceding section explained, and that power over the trade did not extend to conscripting the officers who would carry out the regulation. Students lose this issue by finding a hook and then writing that the statute is valid, when a valid commerce hook does not save a statute that commandeers state officers, and the same is true of retroactive provisions and treaty-based claims. The subsections below take each check in turn, beginning with commandeering.
No Commandeering of States
The first of the three checks asks whom a federal statute commands. Congress may regulate private individuals directly, but it may not require a state's legislative or executive officers to enact or administer a federal regulatory program, and that limit is known as the anti-commandeering rule.15 To commandeer is to take over for one's own purposes something that belongs to another, and here the thing taken is the personnel of a state government. Printz v. United States applied the rule to the Brady Act's interim provisions, which forced local law enforcement officers to run background checks on prospective handgun buyers, and the Court held that the command violated the Tenth Amendment.16 No clause of the Constitution says in so many words that Congress may not conscript state officers, so the Court reached its result by reading three kinds of evidence together: what the early Congresses did, how the Constitution divides sovereignty between nation and states, and what the text says about state judges. This section takes history first, because the contrast it reveals between state judges and state executives is the distinction on which the rule turns.
The early statutes tell a lopsided story. Congresses of the founding era placed a number of duties on state judges to enforce federal prescriptions, and the Court read those laws as showing, at most, that the Constitution was understood to permit such obligations on matters suited to the judicial power.17 Set against that record was the utter lack of statutes commanding state executives, even though that course must have been attractive to Congress, and the Court drew the inference that early lawmakers assumed no such power existed.18
The constitutional text and the nature of the two jobs both explain why judges could be treated differently. Article VI makes federal law the supreme law of the land and provides that "the Judges in every State shall be bound thereby, any Thing in the Constitution or Laws of any State to the Contrary notwithstanding."19 Article III, § 1 points the same way, because it made lower federal courts optional even though the Supreme Court alone could never hear every federal case, which made it implicit that state courts would hear federal claims.20 The functional reason is that courts apply the law of other sovereigns as a matter of routine, as when a court enforces a cause of action that arose elsewhere, so a command to hear a federal claim asks a court to do what it already does.21 An executive officer's ordinary work is carrying out the policy of the officer's own government, and an order to administer a federal program turns that officer into an instrument of a different government. That difference in the two jobs is why a federal command can fit a state trial court comfortably and still fail when it lands on a county sheriff.
The only two early statutes that did involve state executives confirm the pattern. The Extradition Act of 1793 required a state's executive authority to arrest and deliver a fugitive on the request of the executive authority of the state the fugitive had fled, but it implemented the Extradition Clause of Article IV, so it rested on a particularized constitutional command and shows nothing about a general power to conscript.22 The jail statute of September 23, 1789 points the other way. The First Congress needed state jails to hold federal prisoners, and it recommended that state legislatures pass laws making it the keepers' duty to receive them, offering fifty cents per month for each prisoner, instead of commanding the jailers.23 When Georgia refused, Congress responded only by authorizing the marshal in any noncomplying state to rent a temporary jail, a response that accepts the refusal and works around it.24 A legislature that believed it could order state executives to act would have had little reason to ask for help, offer payment, or rent a jail of its own.
History pointed against the power without settling the question, and the Court treated it that way before turning to structure.25 The structural starting point is dual sovereignty: the states surrendered many powers to the new national government but kept a "residuary and inviolable sovereignty," which the enumeration of Congress's powers in Article I § 8 implied and the Tenth Amendment made express.26 The Framers reached that design through experience. Under the Articles of Confederation, using the states as instruments of federal governance had proved ineffectual and provocative of conflict, so the Framers rejected a central government that would act upon and through the states and instead gave the nation and the states concurrent authority over the people, who are the proper objects of government.27 The Court summarized the result by saying that the Framers "explicitly chose a Constitution that confers upon Congress the power to regulate individuals, not States."28
The structural argument has two working parts, and they do different jobs. The first concerns cost and labor: if Congress could conscript state officers, it could enact a federal program and pay for none of its administration, leaving state budgets and state payrolls to carry it. The second concerns accountability and is the subtler of the two. A state official who administers a federal program remains answerable to state voters, so public anger about the program would fall on officials the voters can remove, while the Congress that designed the program escapes responsibility. The Constitution contemplates a state government that represents and remains accountable to its own citizens, and conscription would sever that link.29 Marshall's reasoning in McCulloch v. Maryland fits the same point, since the Constitution shows no intention to make the national government depend on the states for carrying out its powers, and its own means are adequate to its ends.30 For criminal law the consequence is practical, because federal prohibitions are enforced by federal prosecutors, agents, and marshals, and Congress cannot cure a shortage of federal enforcement capacity by ordering state police to fill it.
The dissent read the same materials differently, and its argument is worth knowing as a counterpoint. It reasoned that the Tenth Amendment, unlike the First Amendment, prohibits no category of laws that Article I otherwise authorizes, because its words refer only to powers not delegated to Congress. On that view the commerce power supports regulating handgun sales, the Necessary and Proper Clause supports temporarily enlisting local police, and Article VI, which requires state executive and judicial officers to swear an oath to support the Constitution and makes federal law supreme in every State, shows that federal law may place greater duties on state officials than on private citizens.31 The majority's answer starts from the words of the Supremacy Clause, which makes supreme only those federal acts that accord with the constitutional design, so an appeal to the clause raises the question whether the act is valid and does not answer it.32 The oath argument follows the same path. Officers swear to support the Constitution, and the Constitution as read through its structure withholds the power to conscript them, so the oath cannot supply a power that the structure denies.
The two poles differ in one variable, the branch of state government that receives the order. In Printz the Brady Act directed chief local law enforcement officers, who belong to the executive branch, to run background checks as part of a federal regulatory scheme, and the command was invalid.33 A federal statute requiring state courts of ordinary jurisdiction to hear a federal-law claim falls on the other side of the line, because Congress may require state courts of adequate and appropriate jurisdiction to enforce federal prescriptions on matters appropriate to the judicial power.34 Each statute rests on a valid federal power and each lands on state personnel, so the answer changes only with the branch commanded.
An earlier Tenth Amendment theory shows what the commandeering rule does not do. In National League of Cities v. Usery the Court held that extending federal wage and hour requirements to state employees in areas such as police protection, fire prevention, and sanitation was unconstitutional, because it displaced the states' freedom to structure integral operations in traditional governmental functions.35 That theory shielded states even from rules that bound private employers equally, and it depended on an unclear line between traditional and nontraditional functions. The Court later overruled it in Garcia, and today a law that would be valid against a private party is generally valid against a state as well. The commandeering rule survives because it asks a narrower question, which is whether Congress ordered state officers to administer a federal program.
Two features of the rule keep it from reaching further than its reasoning. States may participate voluntarily, and Congress may continue a program on a contractual basis with them, as it does when it conditions highway-safety funds on compliance with federal requirements.36 South Dakota v. Dole shows how far that route extends, because conditioning a small percentage of federal highway funds on a state's adopting a minimum drinking age of 21 was an inducement and not compulsion, although the Court recognized that a financial offer can grow coercive enough that pressure turns into compulsion.37
Some applications remain unsettled. The Court left open whether purely ministerial reporting duties, such as a requirement that state agencies report missing children to the Department of Justice, are invalid in the same way.38 A harder criminal-law application would arise if Congress required states to make their execution machinery available to carry out federal capital sentences, a challenge that courts have so far avoided by treating it as unripe.39
A familiar exam pattern tests the executive-versus-judicial line against a criminal-justice program. A federal crime-prevention bill requires state officers to take fingerprints of registering drivers for a federal data bank, and that fact pattern reproduces the Printz structure, because state officers are commanded to administer a federal program.42 The same facts also support a separate attack on the commerce hook, which means the commandeering argument is an independent issue and must be written as its own paragraph.43 Students lose points here when they treat every state officer alike and conclude that any federal duty placed on state personnel is commandeering, when a duty on state judges to enforce federal law rests on the Supremacy Clause and on early practice.
Retroactive and Punitive Laws
The commandeering rule asks whom Congress may order to act. A second group of limits asks what a criminal statute does to the person it reaches, and these limits apply whoever is commanded and whichever sovereign legislates. Article I § 9 provides that "No Bill of Attainder or ex post facto Law shall be passed."45 That sentence contains two separate prohibitions. An ex post facto law operates backward, attaching criminal consequences to conduct that was lawful when it was done. A bill of attainder is a legislative act that declares named persons guilty and punishes them, so the legislature itself decides guilt and no court does. A federal crime must first find an enumerated hook, as earlier sections showed, and these clauses then ask whether the statute survives that hook.
Each prohibition has its own logic. Criminal law works by announcing in advance what is forbidden so that people can conform their conduct, and a rule announced after the act gives the actor no chance to conform. Punishment for a completed act therefore deters nothing. An incentive problem sits underneath as well: a legislature angry at one act that has already happened could reach back and punish it, whereas a rule that looks forward binds lawmakers and their constituents alike, because nobody knows in advance whom it will catch. The attainder ban answers a different danger. A legislature that can name the guilty is doing a court's work, and the safeguards a court supplies (a charge, proof, and a neutral decision-maker) never come into play. Marshall's argument in Marbury v. Madison treated the attainder and ex post facto prohibitions as examples of specific limits on legislative acts, and he reasoned that the framers must have expected courts to follow those constitutional terms rather than a contrary statute.46 When a defendant raises the defense, a court must therefore apply it.
The same section also restricts suspension of habeas corpus. The Privilege of the Writ of Habeas Corpus may not be suspended "unless when in Cases of Rebellion or Invasion the public Safety may require it."47 The writ lets a detained person force the custodian to justify the detention before a judge, so suspending it removes the check on holding people without justification. The clause confines suspension to two emergencies and requires that public safety actually demand it. Federal law enacts the guarantee in 28 U.S.C. § 2241.48 The habeas limit appears only in § 9, but the retroactivity and attainder bans are repeated against the states, which is the contrast this section develops.
Article I § 10 forbids a state to pass a bill of attainder or an ex post facto law.49 The repetition matters because one act can violate both federal and state criminal law, and a person prosecuted by either sovereign can raise the same defense. Justice Story made a related point in Martin v. Hunter's Lessee when he answered Virginia's claim of sovereignty by noting that the people had limited state sovereignty through restrictions such as those in Article I.50 The protection belongs to individuals. In South Carolina v. Katzenbach the Court refused to entertain a State's argument that the Voting Rights Act was a bill of attainder, because courts have treated the Bill of Attainder Clause as a protection for individual persons and private groups who are vulnerable to nonjudicial determinations of guilt, and a State cannot invoke it against the federal government.51
These prohibitions work as a filter separate from the power question. Earlier sections asked whether a sovereign has authority over the subject, through an enumerated hook for Congress or the reserved general power for the states, and the commandeering rule asked whom Congress may command. A statute can pass all of those tests and still fall, because §§ 9 and 10 restrict how a legislature may punish, whatever the subject matter. Anchoring the analysis on timing keeps it simple: the question is whether the law reaches conduct completed before it took effect.
On an exam, a stem that pairs a clear commerce hook with an effective date later than the conduct charged is testing this limit, and the answer turns on the date and not on the hook. Students lose points when they stop the analysis after finding a valid enumerated power and call the statute constitutional. The same statute made effective only from enactment raises no ex post facto problem, and a statute that names individuals for punishment without a trial raises the attainder problem whatever power supports it.
Treaties Cannot Override
The first two limits in this part applied to statutes that rest on an enumerated hook in Art. I § 8. A third limit arises when the government points to an international agreement as its source of power. The argument has surface appeal, because Article VI makes the Constitution, the federal laws made "in Pursuance thereof," and all treaties made "under the Authority of the United States" the supreme law of the land, and it binds state judges notwithstanding any contrary state law.53 Read quickly, the word supreme suggests that a treaty outranks every other constitutional rule, including the guarantees that protect criminal defendants. Reid v. Covert rejects that reading and holds that no agreement with a foreign nation can confer power on Congress, or on any other branch, free of the Constitution's restraints.54
The Court began with the text. Nothing in Article VI's language suggests that treaties, or laws enacted to carry them out, escape the Constitution's provisions, and nothing in the debates over drafting and ratification suggests it either.55 The wording does differ between the two kinds of law, because federal statutes are supreme only when made in pursuance of the Constitution while treaties are supreme when made under the authority of the United States.56 The Court traced that difference to a backward-looking purpose. Treaties were not tied to the Constitution so that agreements made under the Articles of Confederation, including the peace treaties that ended the Revolutionary War, would remain in effect.57 A drafting choice that protected old agreements gives no basis for reading new ones as exempt from constitutional prohibitions.
The structure of the arrangement explains why the limit exists. A treaty is negotiated by the President and approved by two-thirds of the Senate, and an executive agreement needs only the President's authority, so neither route runs through the amendment procedure the Constitution prescribes in Art. V.58 If an agreement could confer power free of constitutional prohibitions, a small group of officials would hold a shortcut around that procedure. They would also have a motive to use it whenever a prohibition stood in the way of a policy that a foreign partner wanted too. The Court drew the same conclusion: such a reading would permit amendment in a manner Art. V does not sanction, and the prohibitions were designed to bind every branch, so the Executive, or the Executive and the Senate together, cannot nullify them.59
A second structural point reinforces the first. Statutes and treaties stand on equal footing as federal law, so a later statute that conflicts with a treaty displaces it to the extent of the conflict, yet a statute must conform to the Constitution. It would be anomalous to hold that a treaty may ignore constitutional limits when a statute that must obey them can override it.60
The facts of Reid show the principle at work. Two servicemen's wives, one living in Japan and one in England, were tried by court-martial under Art. 2(11) of the Uniform Code of Military Justice on charges of killing their husbands, and neither trial gave a jury or the other Bill of Rights protections.61 Congress's power over the armed forces is express, and courts-martial of service members rest on that grant and on a Fifth Amendment exception for cases arising in the land or naval forces, but these defendants were civilians.62 The government instead defended the statute as necessary and proper to carry out executive agreements with the host nations, under which American military courts held exclusive jurisdiction over offenses by servicemen and their dependents.63 The Court answered that the Constitution in its entirety applied to the trials and that the courts-martial satisfied neither Art. III § 2 nor the Fifth and Sixth Amendments.64 Art. III § 2 requires that the trial of all crimes, except impeachments, be by jury, and a court-martial supplies no jury.65
The rule does not make a treaty yield to every constitutional objection. In Missouri v. Holland, Missouri sued to stop enforcement of the Migratory Bird Treaty Act, which implemented a treaty with Great Britain protecting migratory birds, on the ground that the statute invaded rights the Tenth Amendment reserves to the states.66 The Court rejected the claim. The Tenth Amendment alone did not answer the question, because the Constitution expressly delegates the treaty power in Art. II § 2 and makes treaties supreme under Art. VI, and the treaty contravened no prohibitory words in the Constitution.67 Reid describes Holland in the same terms: the treaty there was not inconsistent with any specific provision of the Constitution, and to the extent the United States can validly make treaties, the people and the states have delegated that power to the National Government, so the Tenth Amendment is no barrier.68 Holland did not claim that the treaty power is unlimited. It said that the qualifications on that power must be ascertained by a different method than asking whether Congress could act alone.69 The two decisions therefore divide on the kind of objection raised. A specific prohibition on government action, like the jury guarantee and the Fifth and Sixth Amendments in Reid, limits what a treaty can authorize. The Tenth Amendment's reservation of undelegated powers does not limit a power the Constitution has delegated.
The Necessary and Proper Clause fits into this structure as a derivative power. It authorizes laws necessary and proper for carrying into execution the powers listed in Art. I § 8 and all other powers vested in the government of the United States, so it extends a power the government already validly holds and creates none of its own.70 Holland shows the clause working from a valid treaty: once the treaty was valid, no dispute remained about the implementing statute under Art. I § 8 as a necessary and proper means of executing the government's powers.71 Reid shows the clause failing from the other end. The government's necessary-and-proper argument depended on the agreements as a source of power, and because an agreement cannot supply power free of constitutional restraints, the clause could not stretch the agreements into authority to try these defendants without the protections the Constitution guarantees.72
An earlier opinion in the same litigation had relied on the Insular Cases, which allowed certain constitutional safeguards to be withheld in recently acquired territories that Congress governed under Art. IV § 3.73 The Court distinguished those cases and refused to extend them. They concerned the temporary governance of territories with wholly dissimilar traditions and institutions, whereas in Reid the basis for governmental power was American citizenship, and the Court said that neither the cases nor their reasoning should receive any further expansion.74 The Court added that if foreign commitments ever reach a point where the government cannot operate within the Constitution's bounds, the Constitution can be amended by the method it prescribes, and the Court had no authority to read exceptions into it.75 Amendment is therefore the route when existing limits prove too confining, and the later discussion of states as parties returns to the Eleventh Amendment as an example of that route in operation.
Treaties also face a separate question, whether they have domestic force at all. In Medellín v. Texas, a Mexican national sentenced to death in Texas for capital murder relied on an International Court of Justice judgment, Avena, which directed the United States to provide review and reconsideration of the convictions of affected Mexican nationals because consular notice under the Vienna Convention had not been given.76 The Texas court dismissed his second habeas application, concluding that neither the judgment nor the President's memorandum was binding federal law that displaced state limits on successive filings. The Court framed the issue as whether the judgment had automatic domestic legal effect, and it applied the distinction between self-executing treaties, which operate of their own force as domestic law, and non-self-executing treaties, which can be enforced only through legislation carrying them into effect.77 The two inquiries differ. Reid asks how much power an agreement can confer, and the answer is no more than the Constitution allows. Medellín asks whether the agreement operates as domestic law without implementing legislation. A treaty that clears the second filter remains subordinate to the Constitution under the first.
On an exam, a stem in which Congress or the President invokes a treaty or executive agreement to justify a procedure that skips a constitutional guarantee tests the Reid principle, and the guarantee controls. Students lose points when they read supreme law of the land as placing a treaty above the Constitution, since Article VI gives treaties the same status as federal statutes, and both sit beneath the Constitution's prohibitions.82 A stem in which a state objects only that a treaty-implementing statute invades its reserved powers tests Holland instead, and the answer favors the federal government unless a specific prohibition is also in play.