The Treaty Power

The Treaty Power

Of all the instruments of foreign policy, the treaty is the one the Constitution describes most fully. A treaty is a formal agreement between the United States and one or more other nations that binds the countries joining it under international law. Article II, Section 2 gives the President the power to make treaties only with the Senate's advice and consent, and only if two-thirds of the senators present concur.¹ The previous section treated that clause as one of the shared powers that tie the President's foreign affairs authority to the Senate. Here the focus moves from who holds the power to what a treaty becomes once it exists. A treaty commits the nation to other governments, and it may also enter the American legal system, where courts, states, and Congress must decide how to treat it.

That second role raises most of the questions this section answers. A court asked to apply a treaty needs to know how the treaty came into being and where it ranks against state law and against acts of Congress. The court must also ask whether the treaty operates on its own or waits for Congress to implement it, how far the subjects a treaty covers can reach into matters otherwise left to the states, and what happens when the nation wishes to leave the agreement. The text answers some of these questions directly and leaves others to judicial decisions, so each has its own rule and its own reason. They are taken up one at a time below, beginning with the procedure the Constitution prescribes for making a treaty.

Making a Treaty

The Treaty Clause splits the work of making a treaty between two actors and gives each a distinct job. Article II, Section 2 gives the President the power to make treaties "by and with the Advice and Consent of the Senate," provided that "two thirds of the Senators present concur."² A treaty in the constitutional sense is a compact between the United States and a foreign nation that satisfies that requirement.³ The label carries weight, because a compact made this way becomes part of the supreme law of the land under Article VI and displaces inconsistent state law.⁴ International commitments that skip the procedure exist as well, and the sections on executive agreements below take them up. For now, the word "treaty" in the Constitution refers to the product of this process.

The Senate's place in the middle of that sequence explains a common mistake, which is to say that the Senate ratifies the treaty. It does not. The Senate gives or withholds consent, and the President then decides whether to ratify, so consent is a precondition on the President's power and not a command to use it. The difference between signing and ratifying points the same way, since the signatures of the negotiators establish that the text is authentic and definitive, yet a nation is not bound until it ratifies or otherwise approves the treaty.⁸ The Senate's option to suggest amendments fits that structure. A change to a text the other government has already accepted sends the President back to that government before any ratification can occur.

The vote threshold contains a detail that exam questions like to use. The clause measures two-thirds against the senators present, and the Senate's full membership of 100 does not set the denominator.⁹ Absent senators therefore drop out of the count, and the number of yes votes needed shrinks as attendance falls. The requirement still demands far more than a majority. With all 100 senators present, 67 yes votes are needed, and a little more than one-third of those present can defeat a treaty by voting no.

Each piece of the design answers a different institutional problem. Negotiation belongs to the executive because a single official can keep a bargaining position confidential, respond quickly to a foreign government, and speak for the nation with one voice, which a deliberating assembly of 100 cannot do. The supermajority answers the opposite risk. A treaty becomes supreme law, so a commitment approved by a bare majority would bind the country at the highest level of domestic law on the strength of a thin and possibly temporary coalition. Requiring two-thirds forces broader consensus before that happens. The clause also gives the House of Representatives no role. Because the text names only the Senate, a treaty can enter the supreme law of the land without a House vote, which is why a later section asks whether a treaty can accomplish what would otherwise require legislation passed by both houses.¹⁰

A commitment that falls short in the Senate is not a treaty in the constitutional sense, whatever the President and the foreign government may have signed. Whether the President could make the same commitment some other way depends on sources of authority that later sections examine.

Treaties as Supreme Law

A treaty made by the Article II procedure does more than commit the United States to another government. It also becomes part of American domestic law, and its rank in that law decides what happens when it meets a state statute or an act of Congress that says something different. Article VI supplies the rank: a treaty is, like a federal statute, "the supreme Law of the Land."¹² Because that label places treaties and statutes in the same tier, the conflicts they produce sort into two groups, those with state law and those with other federal law, and each group has its own rule and its own reason.

State law comes first, and there the treaty always wins. A valid treaty preempts, which means it displaces, every state law that is inconsistent with its terms.¹³ The reason lies in who makes the promise. When the United States agrees with a foreign government that the nationals of each country may do business in the other on equal terms, the promise comes from the nation as a whole, and the foreign government has no dealings with any single state. If a state could still enforce a contrary rule, the nation would stand in breach because of something one state did, and the national government could not cure the breach without some rule that binds the states. Supremacy is that rule. It lets one federal voice speak to other nations and lets other nations receive what was promised. A treaty guaranteeing nondiscriminatory business rights to the nationals of a partner country therefore defeats a state statute that burdens only those nationals, because the treaty creates the right and the conflicting state law is automatically preempted.¹⁴ A self-executing treaty carries the same force as a federal statute, so it overrides all state law and also overrides prior federal statutory law.¹⁵

Two ideas are packed into that statement, and they should be kept apart. A treaty is valid when it has been made by the Article II procedure and stays within the Constitution's limits, a topic that later sections of this chapter take up. Whether a court can enforce the treaty directly is a separate question. A self-executing treaty establishes enforceable domestic law without further action by Congress, while a non-self-executing treaty is a valid treaty that needs legislative implementation before its provisions have any effect as domestic law.¹⁶ Supremacy therefore describes the rank of a valid treaty, and self-execution determines whether a court may apply it. The next section explains how courts decide which kind of treaty they have before them.

The Supremacy Clause gives a treaty no rank above a federal statute. A treaty and a statute stand on equal footing, and neither is inherently superior to the other.¹⁷ The equality follows from how each is made. A statute becomes law when both houses pass it and the President signs it, while a treaty becomes law through the President's negotiation and the consent of two-thirds of the senators present, and the Constitution itself prescribes each of those routes for making supreme law.¹⁸ Neither route was built to override the other. A fixed ranking in either direction would let one lawmaking process freeze a policy against the other, since a treaty that always outranked statutes could never be changed by a statute and a statute that always outranked treaties would leave the treaty makers with no lasting effect. Equal rank avoids both results and hands the question of which instrument governs a conflict to a rule about timing.

When a treaty and a statute conflict, the one adopted most recently prevails. That rule appears in Whitney v. Robertson, decided in 1888.¹⁹ The Chinese Exclusion Case phrased the idea this way: "the last expression of the sovereign will must control."²⁰ The rule runs in both directions, since a treaty can displace an earlier statute and a statute can displace an earlier treaty.²¹ Its logic is correction. Because both instruments come from lawful processes and neither outranks the other, the nation's most recent decision governs, and the political branches can change course without being bound by an earlier choice.

Courts do not reach for the timing rule at the first sign of tension. They first try to interpret the treaty and the statute so that both can coexist, and the later-in-time rule applies only if the conflict is real and unavoidable.²² The preference follows from the reason for the rule. Each instrument is a considered act of lawmaking, and a reading that reconciles them gives effect to both choices instead of discarding one.

The order of enactment is the only fact that moves the answer when the same kind of conflict arises, and the two poles show it. A treaty increasing the quota for Canadian immigrants prevailed over an earlier act of Congress on the subject, because the treaty was the later expression of the national will.²³ Reverse the sequence and the result reverses with it. The 1954 Mutual Defense Treaty committed the United States to protect the Republic of China (Taiwan) in the event of attack. If Congress, while that treaty remained in force, approved and the President signed a bill barring any use of the armed forces to defend Taiwan, the statute would not formally terminate the treaty, yet to the extent the two conflict the statute, as the more recent enactment, would prevail.²⁴

The prevailing statute governs in American courts, and the obligation the treaty created toward the other government may remain. A treaty that is no longer valid as United States law may still be binding under international law.²⁵ A de facto repudiation of a treaty, such as the Taiwan statute, may also carry consequences for the United States on the international plane.²⁶ A later statute thus changes what American judges must apply, and what the nation owes abroad remains a matter for diplomacy and the political branches. Formal termination of a treaty is a different subject, taken up later in this chapter, and a statute that merely conflicts with a treaty does not terminate it.

The instinct to treat a treaty as a trump card is correct against state law, where supremacy settles the question, and wrong against a federal statute, where rank is equal and the dates decide. In a treaty-versus-statute problem, the first fact to find is therefore which instrument came later, and the second is whether the two can be read to coexist.

Self-Executing Treaties

The preceding section separated a treaty's rank from the question whether a court can apply it, and this section supplies the test for the second question. Treaties fall into two kinds. A self-executing treaty operates as domestic law of its own force, and a non-self-executing treaty requires implementing legislation before it can be enforced in court.²⁸ The distinction goes back to Foster v. Neilson, which treated a treaty as equivalent to an act of the legislature when it "operates of itself without the aid of any legislative provision."²⁹ Which side a treaty falls on is a question of interpretation, so a litigant who relies on a treaty in court must first show that it belongs on the self-executing side.³⁰

A non-self-executing treaty is still a valid treaty. The United States remains bound to the other governments under international law, and Congress may later enact the statute that gives the treaty domestic force.³¹ What the treaty lacks is a rule a judge can apply, so a court treats it as a promise addressed to the political branches, with domestic effect depending on legislation. That is why the question matters in litigation: a treaty that is supreme law in rank can still be useless to a private party or a state official who asks a court to enforce it.

The inquiry looks to what the treaty makers intended because of the way the Constitution divides the treaty power. The President makes the treaty and the Senate consents by a two-thirds vote. If the executive wants a treaty to operate in the courts without more, the executive can secure that result in the making, by negotiating language that plainly provides for domestic enforceability, and the Senate's supermajority then approves that choice.³² A treaty ratified without such language presents a different situation. It can become domestic law only the way other laws do, through Congress, because the power to make the necessary laws belongs to Congress and the President's role is to execute them.³³ Two further reasons support that default. A treaty that took domestic effect by implication would displace state law and bind state courts without Congress ever having decided that it should, so requiring legislation keeps that decision with the national legislature and makes it less likely that treaties will generate judicially enforceable rules that override state requirements.³⁴ Some treaty commitments also leave the political branches a choice about whether and how to comply, and automatic judicial enforcement would hand that choice to individual state and federal judges.³⁵

A pair of treaties on the same subject shows how the line falls. A treaty providing that the citizens of each country may engage in business in the other on a nondiscriminatory basis creates the right itself and needs no further action by Congress, so it is enforceable as domestic law and preempts a conflicting state statute.³⁶ Change the treaty so that it merely sets international standards for the nondiscriminatory treatment of foreign businesses and calls on each signatory to implement them through domestic legislation, and it becomes non-self-executing, because the treaty itself says that the work remains for legislators.³⁷ Treaties that require an appropriation of money are generally treated as non-self-executing as well. The Constitution lets money leave the Treasury only through appropriations made by law, and a treaty cannot satisfy that requirement by itself.³⁸

The leading illustration of the harder case is Medellin v. Texas, which asked whether a judgment of the International Court of Justice (ICJ), the principal judicial organ of the United Nations, binds American courts without an act of Congress. José Ernesto Medellín, a Mexican national, was convicted of capital murder and sentenced to death in Texas. He was not informed of his right under Article 36 of the Vienna Convention on Consular Relations to ask for help from the Mexican consulate, and his efforts to raise the claim in state and federal habeas proceedings were rejected as procedurally defaulted or on the merits.³⁹ While his federal appeal was pending, the ICJ decided the Avena case. It found that the United States had violated Article 36(1)(b) as to 51 Mexican nationals, Medellín among them, and directed review and reconsideration of their convictions and sentences without regard to state procedural default rules.⁴⁰ Medellín then filed a second state habeas application, and the Texas Court of Criminal Appeals dismissed it as an abuse of the writ, concluding that Avena was not binding federal law that could displace the state's limits on successive petitions.⁴¹ The Supreme Court framed the first question as whether Avena had automatic domestic effect under the Supremacy Clause, and it answered no: Avena creates an obligation under international law, but not every international obligation becomes binding federal law enforceable in United States courts.⁴² Whether the President could supply the missing force alone is a separate question, taken up in a later section on presidential implementation of treaties.

Chief Justice Roberts, writing for the majority, treated the self-execution question as one of reading the treaties. The inquiry examines the text and related materials, such as drafting history and the parties' understanding after ratification, to learn whether the President who negotiated and the Senate that consented intended the treaty to have domestic effect.⁴³ Applied to the three instruments at issue, that method pointed the same way each time:

• The Optional Protocol gives the ICJ compulsory jurisdiction over Vienna Convention disputes, which is most naturally read as a bare grant of jurisdiction, because it says nothing about the domestic effect of ICJ decisions or how they would be enforced.⁴⁴
• Article 94(1) of the U.N. Charter provides that each member "undertakes to comply" with ICJ decisions in cases to which it is a party, and that language reflects a commitment to act through the political branches rather than a directive that courts give judgments immediate legal effect.⁴⁵
• The ICJ Statute makes a decision binding only between the parties and in respect of the particular case, and an individual such as Medellín is not a party before the ICJ.⁴⁶

The Charter's remedy provision explains why the Court read Article 94(1) as a political promise. Article 94(2) answers noncompliance with referral to the Security Council, where the United States holds a veto, and the political branches knew that when they ratified. Treating ICJ judgments as automatically enforceable would have removed the option of noncompliance and moved sensitive foreign policy choices to state and federal courts, which conflicts with the Charter's remedial scheme and with the allocation of foreign relations authority to the political branches.⁴⁷ The Executive Branch's long-held view that these treaties create no domestically enforceable law received great weight.⁴⁸ None of the many other parties to the Protocol and the Convention treats ICJ judgments as directly enforceable in its own courts, and the Court expected that parties who intended such an unusual result would have said so clearly.⁴⁹ The practical effect of this method is that a treaty silent on the question is treated as non-self-executing unless its language or later legislation provides otherwise.⁵⁰

The holding has limits. It does not call into question the ordinary enforcement of foreign judgments or arbitral awards when a treaty or implementing legislation gives them domestic effect, and Congress has implemented non-self-executing treaties by statute in many instances.⁵¹ Avena was atypical because it would have enjoined the operation of state law and compelled a state to reconsider criminal convictions, and where a treaty supplies no such remedy the Court declined to create one through judicial lawmaking.⁵²

Two other opinions show how much of the method was contested. Justice Breyer, joined by two other Justices, would have held the treaties self-executing and Avena enforceable.⁵³ He objected that many treaties the Court has found self-executing contain no express statement to that effect, and that drafters of multilateral treaties, who write for countries with different legal traditions, are unlikely to agree on language that addresses the subject.⁵⁴ In place of a demand for textual clarity he would use a practical, context-specific inquiry that weighs text, drafting history, subject matter, whether the provision confers precise individual rights that courts can enforce, and whether enforcement would require courts to create novel duties or intrude on the political branches. Applied here, the consular provision protects an individual and is judicially manageable, the review the ICJ ordered is judicial in nature and suits state courts, and the President supports enforcement while Congress has not acted to prevent it.⁵⁵ The majority replied that a multifactor approach would substitute ad hoc judicial judgments for the expectations of the political branches and make a treaty's meaning depend on later judicial decisions rather than on the text the President negotiated and the Senate approved.⁵⁶ The disagreement concerns method. The majority treats the treaty's words and the understanding of its makers as controlling, and the dissent treats them as one input into a broader judgment about whether courts can sensibly enforce the provision.

Justice Stevens took a middle position. He concurred in the judgment while accepting that the question was close and that the text and history of the Supremacy Clause and the Court's treaty cases do not support a presumption against self-execution.⁵⁷ He nonetheless read Article 94(1)'s undertaking to comply as a promise of further steps through the political branches, a reading confirmed by the Charter's reference of noncompliance to the Security Council, and he therefore concluded that the treaties did not authorize the Court to enforce Avena.⁵⁸ He added that the international obligation remains and that the judgment does not stop Texas from acting to carry it out.⁵⁹ His opinion shares the dissent's doubt about any presumption and the majority's reading of this particular language, which shows that the result could be reached on the words of Article 94 and the Charter's remedial scheme without any presumption at all.

A court deciding whether a treaty provision can be enforced as domestic law can run the inquiry in a fixed order.

The usual mistake is to reason from supremacy directly to enforceability. A treaty's rank as supreme law settles that it beats inconsistent state law once a court may apply it, and self-execution decides whether a court may apply it at all. Step 2 and step 3 are also where a competing method would change the answer, since a court following the dissent's contextual approach would weigh individual rights and practical enforceability even where the text is silent.

Treaty Subject Matter and Federalism

A treaty needs only the President's negotiation and the consent of two-thirds of the senators present, and the House of Representatives has no part in making one. That feature raises a question about the reach of the power: which subjects may a treaty address, and may Congress then legislate on a subject it could not have regulated without the treaty? Two answers are easy to reject. If a treaty could address only matters already within Congress's enumerated powers, the nation could make no binding promise about problems that no single state and no existing federal power can solve. If a treaty could address anything at all, the President and a Senate supermajority could accomplish what Congress cannot, and the limits on Congress's powers would shrink to a formality. The Court's answer lies between those poles, and it begins with Missouri v. Holland.

The dispute concerned birds. In 1916 the United States and Great Britain concluded a treaty reciting that many species of migratory birds passed through parts of the United States and Canada, that the birds were valuable as food and for destroying insects that harm vegetation, and that they were in danger of extermination for lack of adequate protection. The treaty set closed seasons and other protections, and each side agreed to take or propose the legislation needed to carry it out.⁶² Congress then passed the Migratory Bird Treaty Act of 1918, which forbade killing, capturing, or selling the covered birds except as permitted by regulations of the Secretary of Agriculture.⁶³ Missouri sued to stop a federal game warden from enforcing the Act, arguing that it invaded powers reserved to the states by the Tenth Amendment. The state could point to earlier lower-court decisions striking down a federal statute that tried to regulate the killing of migratory birds directly, without any treaty, and from those decisions it drew a rule: what Congress cannot do unaided, a treaty cannot do either.⁶⁴

Justice Holmes's opinion for the Court rejected that rule in a sequence of steps, and the sequence is what you should be able to reproduce. The first step is textual. The Tenth Amendment reserves only powers that the Constitution has not delegated to the United States, yet Article II, Section 2 expressly delegates the power to make treaties, and Article VI declares treaties made under the authority of the United States, along with the Constitution and federal laws made in pursuance of it, the supreme law of the land.⁶⁵ The second step connects the statute to the treaty. If the treaty is valid, the statute is valid under Article I, Section 8 as a necessary and proper means of executing the powers of the government, so the case reduces to a question about the treaty alone.⁶⁶ The third step answers Missouri's test. The earlier decisions about a direct statute could not measure the treaty power, because acts of Congress are supreme only when made in pursuance of the Constitution, while treaties are supreme when made under the authority of the United States.⁶⁷ Holmes acknowledged that the treaty power has qualifications and said they must be ascertained in a different way. He added that some matters of the sharpest exigency for the national well-being lie beyond an act of Congress alone yet within reach of a treaty followed by such an act.⁶⁸ The upshot is that a treaty on the subject can support implementing legislation even if the commerce power would not reach migratory birds.⁶⁹

The remaining steps apply those premises to the birds. The treaty contravened no prohibitory words of the Constitution, so the only question left was whether it was forbidden by some "invisible radiation" from the general terms of the Tenth Amendment, a question the Court said must be answered in light of the nation's whole experience and of what the country has become since the founding.⁷⁰ Missouri's claim to own the birds proved a weak foundation, because wild birds belong to no one until captured, and the birds in question might be a thousand miles away within a week.⁷¹ What remained was a national interest of nearly the highest importance, one that could be protected only by national action in concert with another power, since the subject is only transitorily within any state and reliance on the states was futile.⁷² Those facts carried the result: no specific constitutional prohibition, a subject that crosses borders, and a problem the states could not solve one at a time.

The rule has two parts, and each has its own justification. The first part, that a treaty is not confined to subjects Congress could reach alone, follows from the purpose of the power. It is not lightly to be assumed that a power every civilized government needs, in matters requiring national action, is missing from the nation, and where the states individually are incompetent to act the nation must be able to.⁷³ The second part, that the treaty must still be a valid one, is policed first by the process itself. The states are represented in the Senate, and the two-thirds requirement gives them a real chance to block a commitment that threatens their interests, which is one reason courts leave the President and Senate wide room to decide what counts as a proper subject.⁷⁴

In practice the Court has left the choice of treaty subjects largely to those two actors. Holland did not say that they may pick any subject, yet nothing in the opinion invites active judicial scrutiny of whether a topic carries enough national and international concern to warrant a treaty, and the Court has never held a treaty invalid because its subject lay beyond the treaty power.⁷⁵ The same attitude answers the objection that a treaty can stand in for a statute and so shut the House out. A treaty raising the quota for Canadian immigrants would take over a subject that the Naturalization Clause gives to Congress, but the objection probably fails. Many treaties overlap Congress's Article I powers, and treaties and statutes sometimes conflict, which could not happen if they occupied separate fields.⁷⁶ A federal appeals court applied the same logic in Edwards v. Carter, upholding a treaty that transferred United States property to Panama even though Article IV gives Congress the power to dispose of such property.⁷⁷

The commerce-power cases show how much this deference can matter. In United States v. Morrison the Court struck down the federal civil remedy created by the Violence Against Women Act as beyond the commerce power, which leaves the question whether a treaty could supply what the commerce power could not.⁷⁸

Holland's second step, which lets a treaty alone justify implementing legislation, is where modern criticism concentrates. The Court had a chance to revisit Holland in Bond v. United States, a prosecution under a statute implementing the Chemical Weapons Convention, but declined to do so.⁷⁹ The petitioner argued that reading the statute to reach her conduct would force the Court either to limit or to overrule Holland, because otherwise federal power would have no limit once a treaty had been ratified, and the majority avoided the question by reading the statute and treaty narrowly.⁸⁰ Justice Scalia, joined by Justice Thomas and in part by Justice Alito, concurred only in the judgment and would have reached the constitutional question. His objection was textual. The Necessary and Proper Clause lets Congress carry into execution the government's other powers, and one of those is the President's power to make treaties, but a power to help make treaties is not a power to implement treaties already made. A treaty is complete once the President agrees and the Senate consents, so Congress must rest any implementing law on its independent Article I, Section 8 powers.⁸¹

Two further arguments sharpen the attack. Justice Scalia observed that treaties in Holland's day were typically bilateral and covered a narrow range of subjects, whereas many modern treaties are detailed multilateral instruments that regulate how states treat their own citizens and even private activity, so a Holland-style implementing power with no subject-matter limit would make the federal government's possible reach endless.⁸² A separate line of argument rests on Article V: if a treaty can enlarge the subjects on which Congress may legislate, the Constitution's amendment process stops being the only way to enlarge them, and the subject matter of legislative power ordinarily grows only by amendment.⁸³ Lower courts have also given the Necessary and Proper step some bite. In United States v. Nagarwala a federal district court held that a statute outlawing female genital mutilation, enacted under a treaty protecting women and children from discrimination, bore no rational relationship to the treaty and so exceeded Congress's Holland-based authority, stressing that states hold primary authority over criminal law.⁸⁴ That ruling is rare and debatable, yet it shows that Holland supports only legislation rationally related to carrying the treaty into effect. Holland remains the governing rule until the Court says otherwise, and an essay should state it first and flag the challenge second.

The case that surfaced the debate also produced the doctrine with the most practical force, a clear-statement rule. Congress had enacted a statute to implement the 1999 Chemical Weapons Convention, and the government invoked it against a person who, in a feud born of romantic jealousy, spread chemicals on a neighbor's doorknob and mailbox and caused a minor burn to a thumb.⁸⁵ The Court held that the statute lacked the clear statement needed to reach such conduct. Treaties and their implementing legislation must be read consistently with principles of federalism, and reading this statute to cover the dispute would turn a law whose core concerns were acts of war, assassination, and terrorism into a vast federal regime reaching the simplest assaults.⁸⁶ In the Court's words, "Absent a clear statement of that purpose, we will not presume Congress to have authorized such a stark intrusion into traditional state authority."⁸⁷

The reason for the rule is structural. A treaty and its implementing statute can reach deep into fields such as ordinary criminal law that the states normally control, and the states' main protection is the Senate ratification process and the federal lawmaking process. Those protections work only if the intended domestic reach is stated clearly enough that senators and legislators know what they are approving, and without that clarity the states' ability to defend themselves through those processes would be severely impaired.⁸⁸ In Holland the treaty and its implementing statute were both highly specific about their domestic reach, so no clarity problem arose.⁸⁹ The rule operates at the level of interpretation, so it narrows what an implementing statute covers without deciding whether Congress had the power to go further, which is why it left Holland untouched.

Broad deference on subject matter does not turn a treaty into an all-purpose substitute for Congress. Some functions that the Constitution assigns to Congress, by specific text or by evident purpose, probably cannot be performed by treaty alone, at least without congressional authorization.⁹⁰ Spending is the clearest case, because Article I, Section 9 lets money leave the Treasury only through appropriations made by law, and a treaty that tried to appropriate funds by its own force would likely violate that textual limit. The same reasoning explains why treaties that require an appropriation are generally treated as non-self-executing, as the preceding section noted.⁹¹ Defining crimes is a second case. The Fifth Amendment's due process guarantee traces to the old rule that the Crown could take life, liberty, or property only by the common law or an act of the legislature, so criminal liability resting on a treaty alone, with no statute behind it, would likely fail that guarantee.⁹² Declaring war is a third. It has been argued that a treaty cannot replace a congressional declaration, because the power was given to Congress to keep the President from taking the nation into war without the people's consent, and letting the President and Senate alone declare war would circumvent that check.⁹³ The Canadian quota differs for exactly this reason: nothing in the naturalization power exists to stop unilateral executive action, whereas the war power was placed with Congress for that purpose.

The specific limits above rest on a general one that Reid v. Covert later stated. There the government defended a statute subjecting civilians abroad to military trial as a necessary and proper means of carrying out executive agreements with other nations, and a plurality answered that no agreement with a foreign nation can confer power on Congress, or on any other branch, free from the restraints of the Constitution.⁹⁴ The plurality also explained that Holland is consistent with that view, because the treaty there was inconsistent with no specific constitutional provision, and the Tenth Amendment is no barrier to the extent the United States can validly make treaties.⁹⁵ Holland therefore defeats a general federalism objection and leaves every specific prohibition in place, a point a later section develops in full.

Terminating a Treaty

Article II, Section 2 gives the President the power to make treaties with the Senate's advice and consent, provided two thirds of the senators present concur.⁹⁸ Nothing in the Constitution says how a treaty ends, meaning how a party may withdraw from it or cancel it, a step usually called abrogation. The question of who may end a treaty therefore has to be argued from structure and history, and both sides of the argument have real force.⁹⁹

One argument holds that ending a treaty should mirror making one. A commitment created by the President and two-thirds of the Senate should not be undone by the President alone, because a single official could then erase what a supermajority approved. The competing argument observes that the text requires no Senate or congressional approval for ending a treaty, which suggests that the authority rests with the President as part of the executive power.¹⁰⁰ Each argument reads the same silence in the opposite direction, and the text alone does not choose between them.

Historical practice does not choose either, because the United States has ended treaties in a variety of ways:

• by a statute directing the President to send notice of termination;
• by a statute that ends the treaty without separate presidential notice;
• by the President with the approval of both Houses of Congress;
• by the President with the consent of the Senate;
• by the President alone.¹⁰¹

Because precedents exist for both unilateral and shared action, each side of the dispute can point to examples, and no settled practice has hardened into a rule that a court could treat as the constitutional answer.

The one time the Court confronted the question directly shows how little it settled. In Goldwater v. Carter the Court dismissed a challenge by members of Congress to President Carter's unilateral termination of the 1954 Mutual Defense Treaty with Taiwan.¹⁰² In 1978 the President had recognized the People's Republic of China as the sole government of China and announced that the treaty with the Republic of China was ended. The plaintiffs claimed that a President may abrogate a treaty only with the consent of two thirds of the Senate or a majority of both Houses. Six Justices voted to dismiss, but they did not agree on why: four concluded that how a treaty may be abrogated is a political question, one found the challenge unripe, and the sixth did not disclose a reason.¹⁰³

With no majority rationale, the decision stands for the dismissal and nothing more. It does not hold that a President may terminate a treaty alone, and it does not hold that Senate or congressional approval is required, so the underlying question stays open until the political branches resolve it or the Court chooses to take it up.¹⁰⁴ The plurality's reluctance has a structural basis. A dispute between the President and Congress is a contest between coequal branches, and each has constitutional powers to protect its own interests outside a courtroom, tools a private litigant lacks.¹⁰⁵ A separate opinion rejected the political question label and gave reasons that show how much was left undecided. It found no constitutional text that commits termination to the President alone, since the Constitution says nothing explicit about it and its provisions on making treaties and on their supremacy point the other way. It also found that ordinary interpretation could resolve the case, and it added that if Congress, by appropriate formal action, had challenged the President's authority, the Court would have a duty to decide the issue.¹⁰⁶

Jackson's framework from earlier in this chapter offers one way to organize the open question, though no court has adopted it for terminations. On this view, where a treaty and Congress are silent about withdrawal, the President acting alone would be in Category Two and could terminate. If Congress, by statute or through the treaty's own terms, prescribed a withdrawal procedure, a unilateral termination would move into Category Three. How far Congress or the Senate may prescribe such a rule remains largely unaddressed.¹⁰⁷

A different problem arises when Congress does not purport to end a treaty at all but passes a statute that conflicts with it. While the 1954 treaty with Taiwan was in effect, a bill barring any use of the armed forces to defend Taiwan, approved by Congress and signed by the President, would not formally terminate the treaty. To the extent the two conflict, though, the statute, as the more recent enactment, would prevail.¹⁰⁸ The reason is the equal domestic rank of treaties and statutes covered earlier, so Congress can in effect end a treaty's domestic force by enacting a contrary statute, as Whitney v. Robertson held.¹⁰⁹ The statute operates only to the extent of the conflict, and the treaty remains formally in existence. The nation's obligation to the other government may also survive, because a de facto repudiation of this kind may carry consequences for the United States under international law.¹¹⁰

The two situations differ in what is unsettled. President Carter's announcement ended the treaty as a whole and raised the question of who holds the termination power, which no majority of the Court has answered. The statute leaves the treaty alive and resolves a conflict in domestic courts by the later-in-time rule, which is settled. On an exam, a stem describing an announced withdrawal calls for the open question and the dismissal in Goldwater, while a stem describing a later statute with contrary terms calls for the timing rule.