Concept
Executive Agreements
Treaty and foreign affairs powers
Executive Agreements
The Constitution gives the President one express route to a binding agreement with a foreign government, and it is a demanding one: the President may make a treaty only with the Senate’s advice and consent, and only if two thirds of the Senators present concur.1 Presidents also commit the United States by executive agreement, a device the constitutional text never mentions. The label on the document decides the procedure. A document titled a treaty requires Senate approval; a document titled an executive agreement requires no Senate ratification.2
The device carries real weight. In 1940 the “Destroyer-Bases Agreement” substantially expanded American involvement in World War II: President Roosevelt agreed to loan Great Britain 50 naval destroyers in exchange for free 99-year leases on bases at several Caribbean and Newfoundland sites.3 The Supreme Court has never held an executive agreement invalid on the ground that it exceeded the powers of the national government, and the recognized types are all treated as complete constitutional alternatives to treaties, subject to one caveat developed below.4 Critics see a loophole. In the 1940s and 1950s Senator Bricker proposed a constitutional amendment, never passed, that would have prevented the use of executive agreements; it reflects deep concern that a President can circumvent the treaty ratification process.5
Because the choice between a treaty and an executive agreement is essentially a political judgment for the executive branch, the form does not decide whether an agreement is valid. The source of the President’s authority does.6 That source falls into three groups: a prior statutory delegation or a subsequent implementing statute, the provisions of a preexisting treaty, or the independent constitutional authority of the executive branch.7 An agreement that borrows support from a statute or treaty stands on firmer ground than one resting only on the President’s own claim of power, and the answer also bears on how the agreement ranks against other law. The subsections below take the three groups in order, use Dames & Moore to show how Congress’s posture can rescue a doubtful agreement, and end by asking how a valid agreement fares against state and federal law.
In written work, a treaty is cited by its name, the date of its signing, its parties if there are only two, and references to its main sources of publication. The Bluebook generally sends a bilateral treaty to an official U.S. source if one is available, and a parallel citation to the United Nations Treaty Series may be added for a multilateral treaty.8
Agreements Implementing Statutes or Treaties
Two of the three groups of executive agreements share a feature that explains how courts and commentators treat them: the Senate or Congress has already acted before the President signs anything. One group consists of agreements authorized by a prior statutory delegation or implemented by a later statute, and the other of agreements authorized by the provisions of a preexisting treaty.9 The constitutional question in each case is whether the agreement stays within what Congress or the Senate approved. The third group, resting on the President’s own Article II power, comes in the next subsection.
Because these agreements borrow their authority, their validity is measured against the power they borrow. Suppose Congress uses its power to regulate foreign commerce to authorize the President to enter a mutual trade agreement with another state. The resulting agreement is tested against the scope of the foreign commerce power.10 If a treaty is the sole basis for the agreement, the constitutional power to make the agreement derives from the treaty’s own constitutionality.11 The analysis is a tracing exercise: find the statute or treaty, identify the power behind it, and ask whether the agreement stays inside that power. That limit rarely has teeth, since the Supreme Court has never held an executive agreement invalid for exceeding the powers of the national government.12 A second limit does. Even an agreement within the President’s powers is valid only so long as it does not violate the Constitution’s guarantees of individual rights.13
These agreements seldom raise a separation-of-powers problem, and the reason lies in the President’s own constitutional role. Article II makes the President responsible for executing the laws and treaties of the United States. When the President implements a statute or treaty through an executive agreement, the Senate’s authority is not usurped, because the Senate has already played its constitutional role by approving the statute or treaty being implemented.14 Some statutes and treaties say expressly that they may be implemented through an executive agreement, but express permission is not essential. If Congress passes a law giving the President authority to set postal rates with foreign countries, the President may legitimately use an executive agreement for the purpose whether or not the law specified how to implement it.15
The picture reverses when an agreement has nothing behind it. Executive agreements are most likely to raise separation-of-powers concerns when they neither exercise one of the President’s Article II powers nor implement a specific statute or treaty, because the President may then be using the executive-agreement route to usurp Congress’s role in lawmaking or the Senate’s role in approving treaties. An agreement of this type, especially one with a major domestic impact, may be invalid unless Congress has expressly or impliedly authorized it; with that authorization, the agreement is valid as an exercise of lawmaking authority Congress delegated.16 Dames & Moore v. Regan, taken up below, shows how much Congress’s posture can matter. The Court upheld the President’s agreement to resolve claims against Iran, and it called Congress’s express or implied approval of and acquiescence in the domestic aspects of the accord “crucial to our decision.” Without that approval, the agreement with Iran would very likely have violated the separation of powers.17
The best-known statutory variant is the congressional-executive agreement. The President negotiates an international commitment, and Congress approves or authorizes it by legislation, as it does for major trade agreements. A statute needs only a simple majority in each House, so the form replaces the Senate’s two-thirds vote with ordinary lawmaking and gives the House a voice in a field the Treaty Clause assigns to the Senate. Congress may delegate the negotiating authority in advance or implement the finished agreement afterward, the two statutory routes described above. The form rests on practice and a short line of decisions, not on text. In Weinberger v. Rossi, the Court read a statute’s reference to a treaty as covering executive agreements, including agreements negotiated under a congressional statute, and voiced no doubt about their constitutionality. In B. Altman & Co. v. United States, the Court called such an agreement “a compact authorized by the Congress of the United States, negotiated and proclaimed under the authority of its President.” Field v. Clark, a case about reciprocal tariff-making authority, rejected the arguments that the grant delegated legislative power and that it gave the President treaty-making power.18
The objection that the Treaty Clause names two thirds of the Senators present as the route to a binding agreement remains argued, but it has gained little traction. One reply is that validity follows from the validity of sole executive agreements: if the President alone can circumvent the Treaty Clause, the President and Congress together surely can. Courts have also avoided the merits. In Made in the USA Foundation v. United States, the Eleventh Circuit treated the question whether NAFTA required Senate ratification as a nonjusticiable political question. Practice has settled the matter more firmly than doctrine, since these agreements make up the vast majority of the international agreements the United States has concluded since World War II. One limit separates the form from a treaty. Congress’s power to carry out a congressional-executive agreement is bounded by its enumerated powers, such as the foreign commerce power, while the treaty power, as Missouri v. Holland showed, is not so bounded.19
Statute-based and sole agreements can look alike on paper, so the line between them is the source of power the President invokes. An agreement that rests on authority a statute conferred completes an arrangement Congress set in motion by delegating. An agreement that rests on inherent Article II authority, such as the recognition or Commander in Chief power, is a sole executive agreement.20 The label matters because it fixes which power the agreement is measured against.
The wholesaler’s defense rises or falls on whether the agreement is valid, because a valid executive agreement preempts conflicting state law.21 The two versions differ in one fact: whether the Senate or Congress had already acted. Article I, Section 8, Clause 3 gives Congress the power to regulate commerce with foreign nations and among the several states, which is why an agreement with no statute or treaty behind it encroaches on Congress’s sphere.22 The form the President chose does not settle the question, since the choice between a treaty and an executive agreement is essentially political.23 Students err in two directions: they declare every agreement invalid because it skips the Senate, or they declare every one valid because the Court has never struck one down. Each shortcut skips the inquiry into which statute, treaty, or power supports the agreement. That inquiry matters most for agreements with no such support. Presidents have made hundreds of agreements without Senate ratification, the Paris Agreement among them, and there is no consensus about which of them the President may make independently.24
Sole Executive Agreements
A sole executive agreement is an international commitment the President makes on a power the Constitution assigns to the President directly, with no statute or earlier treaty behind it. Article II supplies the candidate powers. Section 2 makes the President Commander in Chief of the Army and Navy, and it conditions the treaty power on Senate consent and the concurrence of two thirds of the Senators present.28 Section 3 directs that the President shall receive Ambassadors and other public Ministers, which anchors the recognition power taken up earlier in this chapter.29 A statute-based agreement is tested against the statute or treaty it borrows from. A sole agreement borrows nothing, so its validity turns on one question: whether it falls within the scope of the independent power the President is exercising.30
The leading decisions arose from American recognition of the Soviet Union in 1933. In connection with recognition, an exchange of diplomatic correspondence with Maxim Litvinov assigned to the United States all Soviet claims against Americans who held funds of Russian companies seized after the Revolution. In United States v. Belmont, the United States sued to recover funds that a Russian corporation had deposited with a private New York banker. The lower courts dismissed the suit because enforcing the Soviet confiscation would violate New York’s public policy. Justice Sutherland’s opinion for the Court treated recognition, the establishment of diplomatic relations, and the assignment as “all parts of one transaction, resulting in an international compact between the two governments,” and it held that in respect of what was done the Executive “had authority to speak as the sole organ” of the government. A compact of that kind, the Court said, is not always a treaty requiring Senate participation, and the Supremacy Clause required New York’s contrary policy to give way.
United States v. Pink returned to the Litvinov Assignment in 1942, this time over the New York assets of a Russian insurance company. New York courts had refused to enforce the agreement, but the Court ruled that states must comply with executive agreements. Justice Douglas explained that a treaty is the law of the land under the Supremacy Clause and that international compacts such as the Litvinov Assignment have “a similar dignity.” Belmont had said the same rule governs all international compacts and agreements, because complete power over international affairs lies in the national government and is not subject to any curtailment or interference by the several states.31 Together, the two decisions establish that a sole executive agreement, like a treaty, prevails over conflicting state law and policy.
The reasoning runs through recognition. The Reception Clause places recognition with the President, and the Litvinov Assignment was a term of the bargain by which recognition was extended, so the agreement shared the constitutional footing of the act it accompanied.32 That link explains why no Senate vote was needed: the transaction as a whole was an exercise of the President’s own recognition power, in which the Senate has no part. It also shows how narrowly to read the sole-organ language in this setting. Justice Sutherland wrote Curtiss-Wright as well, but in Belmont the phrase attaches to one identifiable act, the recognition of a foreign government. It supports the transaction before the Court; it does not by itself protect every agreement the President signs.
A textual objection remains, since the Supremacy Clause lists treaties and omits agreements the President makes alone. Justice Sutherland found the omission of no consequence, because the same rule results for all international compacts from the very fact that complete power over international affairs is in the national government.33 Critics have questioned his reasoning that a compact need not be a treaty, and one alternative reading of Belmont and Pink holds that the President’s exercise of the recognition power, not the agreement itself, displaced state law.34 Both readings lead to the same practical point: the force of a sole executive agreement against state law depends on the Article II power behind it. On the second reading, an agreement attached to no independent power would have no comparable claim to override state law.
The Court has never held an executive agreement invalid for exceeding the powers of the national government, and given the elastic nature of the inherent foreign-affairs power, the likelihood of its doing so is slim.35 The test nonetheless exists. When the President relies on an independent power such as recognizing foreign sovereigns or acting as Commander in Chief, the agreement’s constitutionality depends on whether it falls within the scope of the power exercised, and even a valid agreement must respect the Constitution’s guarantees of individual rights.36 The Court’s record describes outcomes without changing the test. A student who reasons from the unbroken run of upheld agreements to the conclusion that the form needs no supporting power has skipped the inquiry the cases assume.
The power-based test follows from where the danger lies. Where an agreement exercises an Article II power or implements a statute or treaty, separation-of-powers concerns are minimal. The danger arises when it does neither, because the President may then be using the executive-agreement route to usurp Congress’s role in lawmaking or the Senate’s role in approving treaties, particularly when the agreement has a major domestic impact.37
Two agreements show the safe pole. An executive agreement with France to conduct joint naval exercises in the Pacific rests on the Commander in Chief power, so the President may make it without a Senate vote. An agreement designating the location of a United States embassy or consulate abroad rests on the power to receive ambassadors, which includes recognizing foreign governments and establishing diplomatic relations with them, so it too needs no Senate consent.38 Recognition itself can carry an exchange. A President who recognizes a newly formed nation in return for its agreement to transfer title to its bank assets in the United States, so that American creditors of that nation can be reimbursed, acts within the recognition power just as in Belmont and Pink. The agreement stands so long as it violates no constitutional limit, such as those in the Bill of Rights.39
The funding agreement fails where the naval and embassy agreements succeed because the Constitution assigns those subjects to the President and assigns the spending of public money to Congress. Money may leave the Treasury only in consequence of appropriations made by law, and a treaty that sought to appropriate funds without a congressional appropriation would likely violate that limit.40 If a treaty cannot do that, an executive agreement, which skips the Senate entirely, cannot either. Calling a document an executive agreement avoids a Senate vote, but it supplies no power the Constitution withholds from the President. Agreements with severe domestic effects add a further difficulty, because Dames & Moore upheld the Iranian claims settlement with Congress’s express or implied approval as well as the President’s own authority, as the next subsection explains.41
Dames & Moore and Claims Settlement
Dames & Moore v. Regan holds that the President, acting with at least implicit congressional approval, may settle American citizens’ claims against a foreign government by executive agreement, and may suspend those claims in United States courts while an arbitral tribunal hears them. The case joins the last two subsections. The settlement of Soviet claims in Pink was bound up with recognition, so no delegation from Congress was needed there. The Iran agreement had no comparable power behind it, and it had a severe domestic impact: it totally displaced the contract, tort, and other state law claims that private parties held against Iran. Congressional approval was therefore critical.44
The Court framed the dispute as one more episode in the continuing tension between a President who must meet each day’s new challenge and a Constitution that embodies checks and balances. It set Curtiss-Wright’s description of the President as sole organ beside Justice Jackson’s Youngstown response to claims of virtually unlimited executive power, and then looked to what Congress had done.
The crisis began on November 4, 1979, when the American Embassy in Tehran was seized and its diplomatic personnel were taken hostage. President Carter responded under the International Emergency Economic Powers Act (IEEPA) by declaring a national emergency on November 14 and blocking the removal or transfer of Iranian government property subject to United States jurisdiction. Treasury regulations then declared null and void any attachment or other judicial process against Iranian property unless licensed, and they reserved the power to amend or revoke any license at any time. An American engineering firm with a contract with the Atomic Energy Organization of Iran sued Iran, the Organization, and several Iranian banks that December for more than $3.4 million for services performed, and the district court attached property of the banks to secure any judgment.45
The crisis ended through an executive agreement. The hostages were released on January 20, 1981, under an agreement entered into the day before, in which the United States and Iran undertook to terminate all litigation between each government and the nationals of the other and to settle the claims through binding arbitration before an Iran-United States Claims Tribunal.46 President Carter implemented the agreement through executive orders that revoked licenses to exercise rights in Iranian assets, nullified non-Iranian interests acquired after the blocking order, and required banks to transfer Iranian assets to the Federal Reserve Bank of New York. President Reagan then ratified those orders and suspended every claim that could be presented to the Tribunal, directing that such claims have no legal effect in pending court actions. The firm, which had by then won summary judgment against Iran and the Atomic Energy Organization, sued the United States and the Treasury Secretary to block enforcement, arguing that the orders exceeded the President’s statutory and constitutional powers.47 The Court took the case before the court of appeals ruled, on an expedited schedule, because the Solicitor General reported that Iran could treat the United States as in breach of the agreement unless the Government acted by July 19, 1981.48
The Court divided the President’s orders into two pieces, because each rested on different support. The easier piece was the nullification of the attachments and the transfer of the blocked assets out of the country. IEEPA delegates broad authority to the President to act in a national emergency with respect to the property of a foreign country, and the Court held that this authority reached both acts.49 It added that the power to nullify attachments does not depend on the President’s providing a forum in which claimants can recover.50 Because a statute supplied the authority, this piece falls in Justice Jackson’s first category, and nothing further about congressional attitudes was required.
The suspension of the claims was the harder piece, because the claims exist apart from the attachments that accompanied them. The President invoked both IEEPA and the Hostage Act, an 1868 statute, and the Court concluded that neither specifically authorizes suspending claims pending in American courts. It found both statutes relevant nonetheless, in the looser sense that they show congressional acceptance of a broad scope for executive action. IEEPA gives the President wide authority over foreign property in an emergency, and the Hostage Act signals congressional willingness to leave the President broad discretion when responding to the hostile acts of foreign sovereigns.51 Congress cannot anticipate every action a President may need to take, so a failure to delegate specifically does not imply disapproval, especially in foreign policy and national security. Legislation closely related to the subject that shows an intent to give the President broad discretion may be considered to invite measures on independent presidential responsibility, at least where nothing shows a contrary legislative intent and a history of congressional acquiescence exists.52 This is the logic of Jackson’s middle category, in which congressional behavior short of a specific statute does the work.
The history the Court examined was a long practice of settling claims by agreement. Claims by nationals of one country against the government of another are a source of friction between the two sovereigns, and nations have often resolved the friction by agreeing to settle those claims. The United States has done so repeatedly, sometimes by treaty and also, by longstanding practice, by executive agreement without the advice and consent of the Senate. In those agreements the President has renounced or extinguished the claims of American nationals against foreign governments in return for lump-sum payments or arbitration procedures.53 Since 1952 the President has concluded at least ten binding settlements, including one with the People’s Republic of China for $80 million.
Congress, the Court said, had implicitly approved the practice, and that approval was crucial to the decision. The International Claims Settlement Act of 1949 created a commission with jurisdiction to make final and binding decisions on claims against settlement funds. By creating a procedure to implement future settlement agreements, Congress placed its stamp of approval on such agreements, and it has since amended the Act repeatedly for particular settlements without questioning them or the President’s power to make them.54 The legislative history of IEEPA likewise stated that nothing in the Act was meant to impede the settlement of claims of American citizens.55 The reasoning is a chain. Congress had repeated occasions to object to settlements and used them to build machinery that presupposed them. A systematic, unbroken executive practice, long pursued with Congress’s knowledge and never questioned, may be treated as a gloss on the executive power Article II vests, and long-continued practice known to and acquiesced in by Congress raises a presumption that the President acted with its consent.56 Practice alone creates no power, which is why the presumption depends on Congress’s knowledge and acceptance.
The firm raised a statutory and a structural objection, and the Court answered both. It argued that the Foreign Sovereign Immunities Act had moved suits against foreign states out of foreign affairs and into the courts, so that suspending its claim stripped the federal courts of jurisdiction in violation of Article III. The Court answered that the order only suspends claims; those outside the Tribunal’s jurisdiction revive and become enforceable again in United States courts, so the President changed the substantive law governing the lawsuits without taking jurisdiction from the courts. The Act aimed at the barrier of sovereign immunity and cannot fairly be read to bar the President from settling claims. Its history cut the other way from Youngstown’s. The Congress that enacted it had considered and rejected proposals to limit the President’s power to make executive agreements, including claims settlement agreements, and a year later it enacted IEEPA with a statement that it did not mean to impede claims settlements.57 In Youngstown, Congress had rejected the very power the President claimed, so its silence could not be read as consent. Here the rejected proposals would have restrained executive agreements, so reading silence as consent comes far more naturally.
Two further facts buttressed the result. The President had provided an alternative forum, so claimants received something in return for the suspension, namely access to an international tribunal before which they might well recover something on their claims.58 Congress also held hearings on the agreements without enacting legislation or passing a resolution of displeasure, and the relevant Senate committee called the Tribunal of vital importance to the United States.59 Congress has shown that it can object when it wishes. After the President agreed in 1973 to settle more than $105 million in claims against Czechoslovakia for $20.5 million, Congress promptly enacted legislation requiring that the agreement be renegotiated, a reaction the Court contrasted with its silence about the Iranian agreements.60 The practice of settling claims was the same in both episodes, but Congress’s posture reversed, and a congressional enactment contradicting a settlement would put the action in Jackson’s third category, where acquiescence cannot be claimed.
The Court stressed how little it decided. It did not hold that the President has plenary power to settle claims, even against foreign governmental entities, and it observed that the magnitude of such a power, set against the complexity of modern international trade, counsels against a broader construction than necessary. The holding reaches a settlement that is a necessary incident of resolving a major foreign-policy dispute with another country and that Congress has acquiesced in.61 The takings question received the same restraint. Both sides conceded that whether suspending the claims amounted to a taking without just compensation was not ripe, so the Court left it open, noting that the firm could bring a Tucker Act action in the Court of Claims, where the Government conceded that the treaty exception to that court’s jurisdiction was no bar.62 Justice Powell, concurring in part, would have left both the claim about the suspension and the claim about the nullified attachments for case-by-case resolution in the Court of Claims.63
The usual exam error is to cite the case for a general power to settle claims, or for the idea that congressional silence authorizes executive action in foreign affairs. The Court relied instead on statutes showing broad delegation, a long practice known to Congress, the absence of disapproval, and a substitute forum, all in the setting of a major foreign-policy crisis. The treaty case Medellin v. Texas, treated earlier in this chapter, shows the same attention to what the political branches intended, though it asked the question of a treaty’s language and drafting history and found no intent to give International Court of Justice decisions domestic effect.64
Run on the Iran facts, the nullification of the attachments is resolved at the second step, and the suspension of claims carries through the third to the sixth. Substitute the Czechoslovakia reaction at the fifth step and the result changes, because the presumption of consent depends on Congress’s having left the practice alone.
Rank Against Other Law
Whether an executive agreement counts as supreme law depends on what it is compared with. The Supremacy Clause lists treaties among the supreme law of the land and mentions no other international accord, so it is fair to ask whether an instrument that skips the Senate can carry equal force.65 Against state law, a valid executive agreement ranks like a treaty. Against earlier federal statutes and treaties, its rank depends on whether Congress or a treaty stood behind it, and one important case is unsettled. This subsection takes the two comparisons in that order.
The rule against state law comes from the Litvinov Assignment cases. New York courts had refused to enforce the agreement, but in United States v. Pink and United States v. Belmont the Court ruled that states must comply with executive agreements, which, like treaties, prevail over state law and policy. Justice Douglas wrote in Pink that international compacts such as the Litvinov Assignment have “a similar dignity” to a treaty, and Belmont rested the result on the principle that “complete power over international affairs is in the national government and is not and cannot be subject to any curtailment or interference on the part of the several states.”66
The Supremacy Clause’s silence has a possible explanation. The Constitution contemplates other accords with foreign powers, such as the “Agreement or Compact” a state may make if Congress consents, and the federal government is presumably free to enter any of them.67 The Framers may not have expected executive agreements to displace state law. They may have thought such agreements would be limited to diplomatic and overseas matters, with any accord of significant domestic impact taking the form of a treaty. If so, the reluctance in Dames & Moore to recognize a blanket power to use executive agreements in place of treaties fits their intent. Except where an agreement can be defended as an exercise of one of the President’s Article II powers, an international accord that conflicts with state law may be valid only if Congress delegated the authority to make it.68
American Insurance Association v. Garamendi shows the rule at work. During the Nazi era Germany confiscated the assets of Jews across Europe, including the value of insurance policies, and many of those assets never reached the policyholders or their heirs. When German courts lifted a postwar moratorium on such claims in the mid-1990s, class actions followed in American courts against companies that had done business in Germany.69 Germany and its insurers protested, and President Clinton signed an executive agreement under which Germany would create a foundation, funded by the German government and by voluntary contributions from German companies, to compensate people who suffered at the hands of German companies. The United States agreed to tell any court hearing such a suit that American policy favored dismissal on any valid ground, and to use its best efforts to have state and local governments treat the foundation as the exclusive mechanism for resolving these disputes. California, meanwhile, had enacted its Holocaust Victim Insurance Relief Act, which required every insurer doing business in the state to disclose details of policies issued in Europe between 1920 and 1945 and threatened license suspension for noncompliance and criminal prosecution for false statements.70
The agreement contained no preemption clause, so the question was one of implied preemption. The Court found a sufficiently clear conflict anyway, based on the intent of the executive agreements and statements of high-level Executive Branch officials, and it held that the California statute compromised “the very capacity of the President to speak for the Nation with one voice in dealing with other governments.”71 The conflict was one of obstacle, not impossibility. Preemption analysis of that kind asks what the federal objective is and then how far state law interferes with it.72 Here the two schemes shared a goal, compensation for Holocaust victims, but they differed in means. The President had deliberately chosen to encourage voluntary action by European governments and companies, a choice that served the nation’s interest in amicable relations with its allies. California’s disclosure regime, in an area that was not a traditional state concern and that reached conduct outside the state, would have undercut that diplomatic discretion. The Court summed it up this way: “California seeks to use an iron fist where the President has consistently chosen kid gloves.”
Garamendi applies the same logic as Crosby v. National Foreign Trade Council, where a Massachusetts law barring state purchases from companies doing business with Burma gave way to a federal sanctions statute that let the President modify the sanctions, even though the statute had no express preemption clause.73 The cases differ in the source of the federal policy: a statute giving the President discretion in Crosby, the President’s own agreement in Garamendi. Neither makes foreign affairs a field that automatically displaces the states, since the fact that a state law touches foreign affairs does not by itself require preemption, though the fact is hardly irrelevant.74
Four Justices were uneasy about how the majority reached its result. Justices Ginsburg, Stevens, Scalia, and Thomas, dissenting, would limit preemption in the executive-agreement setting to cases in which the President has spoken clearly to the issue, meaning cases in which the express terms of the agreement call for preemption.75 Their objection concerns who makes the decision. An executive agreement is made without the Senate and without either House, so when a court infers from general policy and officials’ statements that the agreement displaces a state law, neither a legislative body nor a clear presidential statement has chosen that result. That position did not carry the Court, so the majority’s implied-preemption approach is the law. The dissent still marks the doctrine’s weak point: an agreement that says nothing about state law.
Rank against federal law is more qualified, and the qualification that matters is authorization. When Congress or the terms of a valid treaty authorized the agreement, the later-in-time rule that governs conflicts between statutes and treaties also governs conflicts between the agreement and an earlier statute or treaty. A congressionally authorized agreement adopted after a treaty therefore prevails over the treaty’s conflicting terms.76 The logic follows from where the agreement’s force comes from. As the earlier subsection showed, an authorized agreement borrows authority from Congress or from the Senate’s consent, so it counts as a later act of that authority, and the date decides which expression of the lawmaking power controls.
The harder case is the agreement that relies solely on the President’s own authority. Such an agreement clearly preempts contrary state law when it is self-executing, but its power to alter earlier federal law is contested, and the Supreme Court has not addressed the controversy.77 One position, associated with United States v. Guy W. Capps, Inc., a Fourth Circuit decision affirmed on other grounds, holds that a purely presidential agreement cannot change preexisting federal law without congressional implementation. The reason is structural. A statute became law through passage by both Houses and presentment to the President, and a treaty through the Senate’s two-thirds vote, while the agreement passed through neither, so letting it prevail would allow the President alone to undo what other actors enacted. The opposing position starts from the Constitution’s allocation of power. If the President makes the agreement under an independent power such as the power to recognize foreign nations or the Commander in Chief power, the agreement is valid federal law with the same dignity as a statute or treaty, and the only question is whether it falls within the scope of the executive power. If it does, the later-in-time rule applies to it as to any other federal law. The dispute tracks Youngstown’s categories: an agreement contradicting an earlier statute sits in the third category, where the President prevails only by showing exclusive control over the subject, and the second position amounts to the claim that powers like recognition supply that control.
The first version is the unsettled one, and the second is the ordinary application of later-in-time, so what separates them is whether Congress stood behind the agreement. Two shortcuts cost points on this subject. One is to carry the Garamendi result across to federal law and announce that executive agreements outrank everything, which confuses the state-law answer with the federal-law answer. The other is to reason from the Supremacy Clause’s silence that an agreement cannot preempt state law at all, which ignores Pink and Belmont. A complete answer gives both halves: an agreement defeats conflicting state law, and it yields to a conflicting federal statute or treaty unless Congress or a treaty authorized it and it came later in time. For a purely presidential agreement set against earlier federal law, the answer is that the question is open.