A Advance Approval by Congress

Congress’ role in approving major international agreements has roots in a series of innovations in international trade practice. Beginning in the early nineteenth century, Congress at times provided by statute that the United States would grant favorable tariff treatment to exports of a foreign country if that country granted favorable treatment to U.S. products. This practice of tariff reciprocity did not challenge the Senate’s treaty-making power because it did not generate any formal agreements – the United States did not commit to continue favorable treatment in the future. But reciprocal trade legislation did lead to the practical innovation of Congress frequently delegating to the president the authority to determine whether other countries were offering the requisite reciprocal treatment and the power to adjust U.S. tariffs accordingly. In this system, the president’s role naturally led to informal negotiation with foreign countries regarding what foreign trade practices the president would regard as sufficiently reciprocal.Footnote ⁵⁶

From this practice it was a short step to Congress saying overtly that the president could make agreements with other countries on tariffs, which Congress did in several trade statutes beginning in the late nineteenth century.Footnote ⁵⁷ In a sense this was little more than statutory confirmation of an existing practice. Importantly, Congress, the president, and academic accounts took the position that the president’s reciprocal agreements were not binding international law, in the sense that Congress could override them at any time without violating any legal commitment made by the United States.Footnote ⁵⁸ Thus they did not infringe the Senate’s treaty-making power, which was the power to make binding international commitments.

But the new practice, and in particular its formal recognition and expansion, began to blur the traditional clear rule that treaties needed Senate supermajority approval. As one modern commentary describes, “the President … took the [tariff acts] as license to negotiate international agreements with foreign powers that looked so much like international treaties that they were frequently referred to as ‘treaties’ even though they were never submitted to the Senate.”Footnote ⁵⁹ However important the binding/nonbinding distinction was in diplomatic practice or constitutional theory, in the wider legal culture it likely had less resonance. What casual observers saw was that the president was making and Congress was approving trade agreements that looked like treaties. The impression was confirmed by the Supreme Court’s 1912 decision B. Altman & Co. v. United States, which found that, for purposes of a jurisdictional statute, a reciprocal trade agreement – even though nonbinding – was a “treaty.”Footnote ⁶⁰

As a result, when the executive branch abandoned the binding/nonbinding distinction (at least in the trade area) in the 1930s, it encountered little opposition. The legal culture at that point was accustomed to the president making trade agreements under prior statutory authority. The worldwide economic collapse – much of it blamed on high tariffs and other protectionist trade policies – contributed a sense of urgency. Franklin Roosevelt’s Administration concluded international agreements directed toward stabilizing the prices of silver and wheat in 1933, with agreements on gold following in later years. Shortly thereafter, Congress passed the 1934 Reciprocal Trade Agreements Act (RTAA), which adopted and expanded the earlier strategy of authorizing the president to make reciprocal trade agreements, and the Roosevelt Administration pursued an array of agreements lowering the ruinous tariffs of the early Depression years.Footnote ⁶¹

None of these agreements was submitted to the Senate. In one sense that did not seem remarkable, because presidents had been doing something like this for at least forty years. But critically, Roosevelt’s agreements – unlike the prior reciprocal trade agreements – were generally regarded as binding in international law.Footnote ⁶² As the Senate itself recognized in 1937 in reauthorizing the RTAA, trade agreements had now been taken out of Article II, Section 2.Footnote ⁶³ The pattern continued after World War II, albeit in a somewhat ad hoc fashion. The centerpiece of the postwar order in trade was the General Agreement on Tariffs and Trade (GATT), designed to stop trade wars and introduce gradual negotiated reductions in tariffs. But the International Trade Organization (ITO), the international body envisioned to implement GATT, could not even gain a majority in Congress. So the Executive Branch under Presidents Truman and Eisenhower claimed delegated authority from a prior statute to negotiate tariffs under GATT, although the statute did not expressly provide it, and trade agreements proceeded under GATT without further direct domestic approval.Footnote ⁶⁴ As described below, trade agreement practice eventually shifted to a different model, but advance approval by Congress remained a major (and numerically dominant) form of agreement-making as Congress increasingly passed statutes with transnational implications in other fields and increasingly adapted the trade agreement model to authorize the president to make agreements implementing those statutes.Footnote ⁶⁵

B Postsignature Approvals

The agreements made under delegated authority from Congress differed from traditional treaty approvals, not only in avoiding the supermajority requirement but also in dispensing altogether with after-the-fact legislative-branch approval. Of course that made them a powerful tool for the president, but they depended on Congress’ willingness to delegate broad authority to the president (or, at least, on the president plausibly being able to claim authority from a prior statute). A more formidable rival to the Article II treaty might arise if Congress had the ability to approve agreements after-the-fact by majority vote.

In the most detailed telling of this part of the story, by Bruce Ackerman and David Golove, the dramatic point came at the end of World War II, as the United States contemplated a series of agreements to shape the postwar world order.Footnote ⁶⁶ The roots of this form of agreement-making appear to extend a little earlier: a 1923 statute authorized the president to renegotiate Allied war debt arising from World War IFootnote ⁶⁷ and Congress approved U.S. membership in the International Labor Organization (ILO) in 1934.Footnote ⁶⁸ These were new ways to approve agreement-making, because in both cases Congress gave assent to existing agreements. Joining the ILO involved accepting the ILO constitution, which had been developed in the aftermath of World War I. The statute authorizing renegotiation of the war debts expressly required the president to bring the completed deals back to Congress for after-the-fact approval. As with the reciprocal trade legislation, these early developments did not directly challenge the Senate’s general power of supermajority consent to treaties. The debt agreements were justified under Congress’ power to dispose of U.S. property, and the ILO membership was defended (perhaps wrongly) on the ground that it imposed no binding obligations on the United States.Footnote ⁶⁹ But, again like the reciprocal trade legislation, they began a blurring of categories that ultimately contributed to the collapse of Senate exclusivity.

As Professors Ackerman and Golove recount, by 1945, rightly or wrongly, conventional wisdom held that the Senate’s rejection of the League and the consequent U.S. withdrawal from the affairs of Europe after World War I had set the world on course for the World War II. The Executive Branch under Presidents Roosevelt and Truman was determined to move forward with the new postwar world order with or without the Senate. The two presidents thus presented a series of key agreements to Congress for postsignature majority approval, backed up by a proposed constitutional amendment in the House that would have eliminated the required Senate supermajority altogether.Footnote ⁷⁰ The result was a potential constitutional crisis, resolved by a majority of the Senate approving the agreements as submitted to both houses without insisting on any special supermajority approval by the Senate.Footnote ⁷¹

This outcome inaugurated a new model of agreement-making: postsignature approval by Congress. In the Ackerman/Golove account, it quickly became a usual procedure regarded as interchangeable with supermajority Senate consent as a means of constitutionally approving international agreements. As Ackerman and Golove quote President Truman in 1947, asking for Congressional consent to an agreement establishing U.S. trusteeship in the Pacific Islands territories:

In 1954, apparently confirming the interchangeability of alternate approaches, Congress approved an agreement regarding construction of the St. Lawrence Seaway, which the Senate had earlier rejected when it had been presented for approval as a treaty.Footnote ⁷³ Congress also began expressly authorizing the president to negotiate international agreements in particular areas subject to after-the-fact approval by Congress – notably in the Atomic Energy Act of 1954 and the Arms Control and Disarmament Act of 1961.Footnote ⁷⁴ Later, Congress approved the 1972 Strategic Arms Limitation Treaty (SALT I) via the latter procedure,Footnote ⁷⁵ confirming Congress’ authority over agreements in a core area of national defense.

A more sophisticated version of this sort of authorization became the foundation for modern U.S. trade law in the Trade Act of 1974, which shifted trade agreements from their traditional procedure of advance authorization into a regime of after-the-fact approvals.Footnote ⁷⁶ As discussed, the postwar innovation of after-the-fact congressional approval had not taken immediate hold in trade law because Congress was unwilling to approve the ITO, and the Executive Branch instead negotiated the initial GATT rounds under a claim of prior statutory approval. That pattern continued until the 1974 Act, which required after-the-fact approval, but also made the process more attractive by developing the “fast-track” procedure of mandating a vote within a specified time and without amendments.Footnote ⁷⁷ Agreements approved under the Act’s procedures included the Tokyo Round of GATT (the first GATT round subject to after-the-fact approval) and free trade agreements with Israel (1985) and Canada (1988).Footnote ⁷⁸

Thus, by the time the Third Restatement was published in 1987, the new form – increasingly called the “Congressional-Executive agreement” – appeared to be the emerging dominant approach. To be sure, the traditional Senate approval had not been wholly superseded, and Ackerman and Golove’s account may overstate the extent to which the treaty-making clause was eclipsed.Footnote ⁷⁹ Even in the immediate postwar period, some important international agreements had been approved under Article II, Section 2: in particular, the UN Charter, the North Atlantic Treaty Organization (NATO) Treaty, and other linchpins in the network of early Cold War security pacts. But momentum appeared to be on the side of the Congressional-Executive agreements, which not only moved into areas more traditionally associated with treaties (such as arms control), but also laid claim to areas typically handled through more open-ended prenegotiation approvals of executive action, such as trade. Moreover, no convincing theory seemed able to explain why some agreements were approved as Article II, Section 2 treaties and some as Congressional-Executive agreements, lending credence to the idea that either could be used for any purpose. Not surprisingly, the Third Restatement affirmed the idea of interchangeability in Section 303: “The prevailing view is that the Congressional-Executive agreement can be used as an alternative to the treaty method in every instance.”Footnote ⁸⁰ “Interchangeability,” Ackerman and Golove concluded, “had become part of the living Constitution.”Footnote ⁸¹

But this too was not the whole picture. First, by the 1980s the Senate was already beginning to fight back, particularly in the arms control area. Despite the precedent of SALT I, senators insisted that the controversial SALT II agreement be submitted under the treaty-making clause.Footnote ⁸² In approving the agreement on Intermediate-Range Nuclear Forces (INF) in 1987, the Senate added a declaration that seemed (despite some ambiguity) to call for future arms control agreements to be submitted under the treaty-making clause.Footnote ⁸³ Second, there was another contender we have yet to consider: the sole Executive agreement, which had enjoyed a parallel rise to prominence – a part of the story taken up in the next subsection.

C Agreement-Making by Presidential Authority

As discussed, presidential agreement-making without formal congressional or Senate approval has roots going almost to the founding and in limited form may be defensible under the Constitution’s text. But it has evolved from a minor diplomatic convenience principally directed to one-time claims settlements into a major element of U.S. foreign policy. This development has occurred in at least three ways: (a) the rise of sole Executive agreements; (b) the innovation of “nonbinding” agreements; and (c) increased claims of implicit congressional approval or acquiescence.

Sole Executive agreements’ move to prominence roughly coincides with the historical pattern described above. In the post–Civil War era, presidents began making more significant agreements, especially in the military area. Several agreements with Mexico in the 1880s authorized each country’s military forces to cross the border in pursuit of outlaws (a very significant concession, had it been expected that the Mexican Army would actually act upon it).Footnote ⁸⁴ Executive agreements made up a substantial part of the interim military and political arrangements after Spain’s capitulation in the Spanish-American War.Footnote ⁸⁵ By the early twentieth century it was probably no longer possible to say (as it had been earlier) that Executive agreements never covered matters of diplomatic consequence.

As with congressionally approved agreements, the biggest steps are associated with President Franklin Roosevelt. In 1933, Roosevelt controversially extended diplomatic recognition to the Soviet Union and negotiated a settlement of substantial outstanding claims between the two nations.Footnote ⁸⁶ He avoided seeking any sort of congressional approval, likely the result of popular mistrust of the Communist regime at the time. Like other steps in our story, this was not in itself a big one: presidents had settled claims unilaterally before; diplomatic recognition was thought to be a core presidential power; and presidents were taking increasingly prominent roles in foreign affairs, including through the tariff negotiations discussed above and the nonbinding agreements discussed below.

But Roosevelt took the unusual step of trying to enforce his agreement as a matter of U.S. domestic law. Under the agreement, the Soviet Union assigned its claims against private parties in the United States to the U.S. government, and the United States filed suits to collect against private citizens. New York state law refused to recognize the validity of the Soviet claims to property located in New York, and Roosevelt sought to override the state law. Doing so meant insisting that the Executive agreement was a binding commitment under international law of equivalent status to a treaty under Article VI of the Constitution.Footnote ⁸⁷

In United States v. Belmont,Footnote ⁸⁸ the U.S. Supreme Court upheld Roosevelt’s position, ruling that he had power to enter into the recognition agreement and that the agreement overrode state law in the same manner as a treaty: “[W]hile this rule [supremacy over preexisting state law] in respect of treaties is established by the express language of cl. 2, Art. VI, of the Constitution, the same rule would result in the case of all international compacts and agreements.”Footnote ⁸⁹ Five years later, the Court reaffirmed its position in United States v. Pink.Footnote ⁹⁰ If read broadly, these decisions might underwrite a dramatic shift of agreement-making power to the president. At minimum, they contributed, along with parallel developments in congressionally approved agreements, to a blurring of categories surrounding agreement-making. Just as the treaty-making clause was losing its exclusivity as the way to make “treaties,” Article VI appeared to be losing its exclusivity as the way agreements became “supreme Law of the Land.”

Roosevelt followed up in 1940 with the important destroyers-for-bases agreement with Britain, also done on sole executive authority and a major step toward U.S. involvement in World War II.Footnote ⁹¹ Attorney General and future Supreme Court Justice Robert Jackson defended the agreement under the president’s Commander-in-Chief power: while conceding that the president could not violate existing statutes, Jackson did not rest on any congressional authorization of the agreement.Footnote ⁹² Thus another avenue for agreement-making – at least in core executive areas – was becoming entrenched.Footnote ⁹³

Even before this time, a further potentially powerful executive form of agreement-making had emerged, this one associated with Franklin Roosevelt’s cousin Theodore. An aggressive proponent of executive power, Theodore Roosevelt as president developed the idea of the “Gentlemen’s Agreement” – that is, as he envisioned it, a diplomatic arrangement based on the good faith of particular leaders but not a formal binding agreement under international law.Footnote ⁹⁴ Because they were not binding, these “agreements” were not “treaties” in the constitutional sense – the sine qua non of treaties being their bindingness – and so no Senate approval was needed. So far this argument paralleled the contemporaneous view of reciprocal trade agreements discussed above. But Roosevelt went further in denying the need for congressional approval as well. The president’s authority arose from the president’s executive diplomatic power, which Roosevelt read expansively, and since (unlike the reciprocal trade agreements) he did not expect his promises to become part of U.S. law, he thought Congress had no role to play.Footnote ⁹⁵

Formally, Roosevelt may have been right,Footnote ⁹⁶ and indeed the informal modus vivendi had long been an aspect of international diplomacy. But practically, his enhancement of the nonbinding agreement was substantial compared to earlier practice, as a sampling of these agreements suggests. For example, the 1907 “Gentlemen’s Agreement” with Japan limited Japanese emigration to the United States;Footnote ⁹⁷ the 1905 agreement with Santo Domingo – done in place of a treaty the Senate refused to approve – involved the United States taking over much of Santo Domingo’s financial affairs.Footnote ⁹⁸ President Woodrow Wilson followed Roosevelt’s practice in the Lansing-Ishii Agreement regarding policy in the Far East.Footnote ⁹⁹ In the World War II era, Franklin Roosevelt continued the practice, making far-reaching arrangements with the Allied Powers on his own authority without the use of binding agreements, as with the Atlantic Charter and the Yalta Agreement.Footnote ¹⁰⁰

Thus, by the end of World War II two forms of pure executive agreement-making had emerged (although the separation between them was not always clear), along with Congressional approval through either ex ante authorization or ex post consent. The distinctions became even less clear in the decades after the war, driven by a decline in formalism generally and by the pervasive influence of Justice Jackson’s concurring opinion in the U.S. Supreme Court’s seminal 1952 case Youngstown Sheet & Tube Co. v. Sawyer.Footnote ¹⁰¹

Youngstown on its face had nothing to do with agreement-making, and Justice Hugo Black’s formalist majority opinion did not offer much accommodation for executive actions not expressly authorized by statute or the Constitution.Footnote ¹⁰² But Jackson, who had defended the destroyers-for-bases deal as attorney general, outlined a more flexible approach, while agreeing with Black on the case’s outcome. Separation of powers issues, Jackson wrote, could usefully be analyzed in a three-category framework: (1) where Congress had approved of a presidential action; (2) where Congress was silent; and (3) where Congress disapproved. In the first category, Jackson argued, presidential action would almost always be valid. Jackson made clear that congressional approval or disapproval could be informal or inferred from circumstances; in Youngstown itself, he found Congress’s implied disapproval to be decisive.Footnote ¹⁰³

Though Jackson joined Black to rule against executive action in Youngstown, his opinion pointed the way for further expansion of executive power, including with respect to international agreements. If the executive branch could find an implicit congressional approval of its agreement-making, it would be in Jackson’s category one, where its constitutionality seemed almost assured. GATT was an obvious example. Congress had voted down the ITO, but it had (sort of) authorized the president to negotiate a tariff regime apart from the ITO. As presidents pursued additional rounds of negotiations under GATT, Congress did not object, and indeed passed legislation assuming the validity of the GATT regime. And Jackson’s structure could be pushed much further, depending on the extent to which generalized and informal congressional actions might be construed as consent.

This model of presidential agreement-making received a potentially broad endorsement from the Supreme Court in 1981 in Dames & Moore v. Regan. President Carter negotiated a deal with Iran to end the embassy hostage crisis; among other things, the Algiers Accords required that suits pending against Iran in U.S. courts be terminated and transferred to an international arbitral body. When President Reagan, succeeding Carter, sought to enforce the Accords, Dames & Moore (a claimant against Iran) objected. The Court, expressly embracing Jackson’s Youngstown framework, found that Congress had “acquiesced” in presidential settlement of claims, a practice the Court traced back to the Wilmington Packet in 1799.Footnote ¹⁰⁴

Dames & Moore, if read broadly, was more radical than it sounded: the agreement in question was much more than the one-time settlement agreements of the nineteenth century, and in any event those agreements had never been seen as making domestic law.Footnote ¹⁰⁵ By incorporating Jackson’s flexible and informal requirements for approval, and arguably expanding them, Dames & Moore made it much easier for the Executive Branch to argue that it was making agreements with congressional approval. The need to seek formal approval, either through Article II, Section 2 or through Congress, correspondingly receded. On the other hand, Dames & Moore seemed to signal a retreat from sole Executive agreements of the Pink/Belmont variety. An aggressively pro-executive opinion could have cited Pink and Belmont prominently and exclusively, dispensing with the need to squeeze Congressional approval out of Congressional inaction. The Court’s more prominent citation of Jackson’s Youngstown concurrence pushed Pink and Belmont, and sole Executive agreements generally, to the background, and built up the need for some nominal tie to Congress for authority to make binding agreements. Whether for that reason or others, nonbinding agreements (which in practice did not purport to require any sort of congressional acceptance) continued to play a major role in U.S. diplomacy. Examples included President Kennedy’s settlement of the Cuban Missile Crisis, President Nixon’s diplomacy with China and his resolution of U.S. involvement in the Vietnam War, the Helsinki and Sinai Accords under President Ford, and various agreements under President Carter.Footnote ¹⁰⁶

These threefold developments left Executive agreements in a state of considerable uncertainty. As to any particular agreement, it might be very difficult to identify the claimed source of authority. For example, how could one tell if an Executive agreement was a “political” (nonbinding) commitment or a legal commitment? Some agreements, of course, might say so expressly: the 1975 Helsinki Final ActFootnote ¹⁰⁷ expressly declared that it was nonbinding, while others recited that the parties intended to be legally bound. But others might be subject to substantial dispute as to their bindingness. Similarly, the idea of implicit Congressional approval or acquiescence made it difficult to tell, absent specific direction from the Executive Branch, whether an agreement should be categorized as done on presidential authority or done on authority provided by Congress.

D Conclusion: The Third Restatement and U.S. Agreement-Making in 1987

As the American Law Institute prepared the Third Restatement in the mid-1980s, the constitutional authority for international agreement-making in the United States was highly confused. At least five distinct alternatives to Article II, Section 2 treaty-making existed in theory and practice: express advance Congressional authorization; express postsignature Congressional approval; implicit Congressional approval; sole Executive agreements; and nonbinding agreements. No theory stood as a comprehensive explanation of them all, and practice appeared to adopt one or the other according to individual circumstances.

Perhaps understandably, the Restatement finessed the entire mess, making only two points. First, Congressional approval could substitute for supermajority Senate approval, and second, the president could make agreements independently in areas of the president’s constitutional authority.Footnote ¹⁰⁸ For the most part these assessments were not incorrect, just incomplete. The comments’ assertionFootnote ¹⁰⁹ of complete interchangeability between treaties and Congressional-Executive agreements, it is true, lacked full theoretical and practical support. A better answer would have been that no one could explain why some agreements took one form and some took the other, nor predict which form would prevail in an ongoing struggle that, at least in the arms control area, seemed to be reemerging. More broadly, the history of agreement-making in the Constitution’s second hundred years reveals the shifting fortunes of various approaches, with none dominant in theory or practice. That pattern, as we will see, continued in the years after the Restatement.

A The Persistence of Treaties

Professor Edwin Borchard, in the midst of the World War II era debate over treaty-making power, asked “Shall the Congressional-Executive Agreement Replace the Treaty?”Footnote ¹¹⁰ By the time of the Third Restatement, the consensus legal answer seemed to be that it could – that is, a Congressional-Executive agreement could do anything a treaty could.Footnote ¹¹¹ And by the numbers, it may appear that the treaty has indeed been largely displaced: only a small fraction of U.S. international agreements are in modern practice approved through the Article II, Section 2 process.Footnote ¹¹² Professor Oona Hathaway’s 2008 academic study, fittingly titled Treaties’ End, concluded that “Congressional-Executive Agreements have gradually eclipsed [the Treaty Clause] as the central method of international lawmaking in the United States.”Footnote ¹¹³

Yet the Article II, Section 2 treaty persists. The raw numbers hide a peculiar feature of modern practice. Despite the Third Restatement’s declaration of “interchangeability,” since its publication the treaty form has dominated the most important agreements in the most important areas of U.S. international relations. The only prominent exception is trade law, where the non–Article II approvals of NAFTA, the World Trade Organization (WTO), and various bilateral or regional free trade agreements stand not so much as a signal of treaties’ decline, but rather a puzzling exception to an otherwise fairly consistent practice.

To begin, consider the range of subjects where treaties have dominated, and indeed in some cases reasserted their importance. In arms control, for example, the 1961 Arms Control Act expressly contemplated approval of arms control treaties outside Article II, and the path-breaking SALT I agreement with the Soviet Union was approved by Congress as a whole (albeit with the gesture of calling it an “interim” agreement).Footnote ¹¹⁴ But thereafter the focus shifted back to Article II, Section 2. As discussed, the pattern began even before the Third Restatement went to press, with SALT II in 1979 and the INF treaty in 1987.Footnote ¹¹⁵ The next two decades entrenched the treaty path: among others, the Strategic Arms Reduction Talks (START I) Treaty and the Treaty on Conventional Armed Forces in Europe (CFE) in 1991, START II in 1993, the Chemical Weapons Convention in 1997, the Comprehensive Nuclear Test Ban Treaty in 1999, the Moscow Treaty in 2003, and the New START Treaty in 2010 were all submitted to the Senate under the treaty-making clause.Footnote ¹¹⁶ No major arms control treaty was submitted or approved under Article I during this period. Moreover, after 1991 the Senate expressly used arms control to contest interchangeability, routinely attaching a declaration to its consent stating that arms control agreements should use the Article II supermajority process.Footnote ¹¹⁷ As a Senate report explained in 1999, “some in the executive branch persist in the mistaken belief that it is constitutionally acceptable to undertake militarily significant international accords by Executive agreement, approved by a simple majority vote of both Houses.”Footnote ¹¹⁸

The Article II, Section 2 approach has also dominated the U.S. approval of human rights agreements, despite the Senate’s reluctance to give consent. During the Cold War era, the Senate declined to act on a series of important multilateral human rights agreements such as the 1948 Convention against Genocide (ultimately approved in 1986) and the 1966 International Covenant on Civil and Political Rights (ICCPR). The Senate approved the latter only in 1992, after the end of the Cold War, and made it subject to numerous reservations, understandings, and declarations – including a declaration that the Convention was not self-executing – that minimized its effect in the United States.Footnote ¹¹⁹ Other major human rights treaties submitted through the Article II, Section 2 process include the Convention Against Torture and the International Convention on the Elimination of All Forms of Racial Discrimination (both approved with reservations after long delays in 1994),Footnote ¹²⁰ and conventions on the rights of women, children, and the disabled (which remain unapproved).Footnote ¹²¹ Despite difficulties in obtaining Senate approval, no major human rights agreement has been submitted or approved solely through Article I, even though Congress also implemented several of these treaties because they required enactment of criminal laws.Footnote ¹²²

A common generalization, in the face of the above practice, is that Congressional-Executive agreements dominate in economic matters while treaties are used elsewhere.Footnote ¹²³ This is also not accurate. For example, in the early 1980s the U.S. State Department inaugurated a new program to provide legal protection for U.S. foreign investment, superseding the prior Friendship, Commerce, and Navigation (FCN) Treaties, which had been used until the 1960s. The new form, the Bilateral Investment Treaty (BIT), followed the old in its approval process under Article II, Section 2.Footnote ¹²⁴ The United States is now party to over forty BITs, most negotiated since the 1980s and all approved under Article II, Section 2.Footnote ¹²⁵ The only substantial modern investment agreement not approved through Article II, Section 2 is the Agreement on Trade-Related Investment Measures (TRIMs), approved by Congress as part of the WTO agreements in 1995 (although congressionally approved trade agreements such as NAFTA often also include investor protection provisions similar to the BITs).Footnote ¹²⁶

Similarly, in the area of intellectual property, almost all major recent agreements have been approved by the Senate under the treaty-making clause, including all of the agreements administered by the World Intellectual Property Organization (WIPO).Footnote ¹²⁷ Again the principal exception is a WTO-related agreement, the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), which was approved by Congress.Footnote ¹²⁸ Finally, tax treaties historically took the Article II approval route, and modern practice has showed no change in this pattern.Footnote ¹²⁹

In sum, the core pattern appears to be that, despite what Restatement Section 303 might seem to foreshadow, major international agreements (other than trade agreements) continue to be approved by the Senate through the Article II process.Footnote ¹³⁰

A second key point is that the Senate supermajority approval process did not become an automatic rubber stamp. If it had, one might argue that it persisted only because it lacked practical significance. But to the contrary, Senate opposition in a number of areas complicated presidential diplomacy. Some treaties were blocked or rejected. In 1999 the Senate voted down the Comprehensive Nuclear Test Ban Treaty despite heavy lobbying by President Clinton.Footnote ¹³¹ For many years the Law of the Sea Convention remained unapproved because it lacked the votes in the Senate.Footnote ¹³² President Obama made approval of that Convention a priority after 2008 but it again failed to obtain the needed supermajority, and in 2012, a prominent senator pronounced the Convention “dead” in light of thirty-four senators’ commitments to vote against it.Footnote ¹³³ Yet most of these agreements had the support of a majority of senators, and so might have been approved as Congressional-Executive agreements.

In other areas, the difficulty of obtaining Senate consent forced major changes or caused the agreements to languish unapproved. In 1997 the Senate made clear that it would not consent to the Kyoto Protocol on climate change.Footnote ¹³⁴ As noted, approval of several key human rights treaties could only be obtained by the president accepting non-self-execution declarations (despised by the human rights community). Yet even with non-self-execution declarations attached, the Senate voted down the Convention on the Rights of Persons with Disabilities in 2012 (although a majority supported it) and refused even to vote on the Convention on the Elimination of all Forms of Discrimination against Women (CEDAW) despite the requests of multiple presidents.Footnote ¹³⁵ The U.N. Arms Trade Treaty, which went into effect in 2014 without U.S. ratification, was so strongly opposed in the Senate that the President Obama did not even submit it for approval, although he signed it and signaled his support.Footnote ¹³⁶ And even where the Senate has given consent, it sometimes came only after a difficult campaign. Most prominently, the New START Treaty, a centerpiece of the Obama Administration’s arms control policy, gained the requisite supermajority consent in 2010 only by a narrow margin after considerable suspense and debate.Footnote ¹³⁷

At the same time, the academic view of the Article II treaty underwent some adjustment. Whether or not the Third Restatement was entirely correct in saying that the conventional academic view accepted interchangeability in 1987, the next two decades witnessed substantial cracks in this supposed consensus. First, in connection with the approval of the WTO in 1994, several prominent academics, notably Laurence Tribe and Anne-Marie Slaughter, argued that supermajority Senate approval should be sought.Footnote ¹³⁸ That provoked an extensive debate between Professors Ackerman and Golove on the one hand and Professor Tribe on the other in the Harvard Law Review, with Tribe making an extended textual case for the constitutional exclusivity of the Article II, Section 2 process.Footnote ¹³⁹ Though the WTO was approved as a Congressional-Executive agreement (albeit with two-thirds of senators consenting), subsequent academic studies were not kind to interchangeability. Separate pieces in 2001 by John Yoo and Peter Spiro conceded the use of Congressional-Executive agreements in trade law but argued for Article II, Section 2 approval of major agreements elsewhere.Footnote ¹⁴⁰ Yoo followed up with a New York Times op-ed in 2009 (with John Bolton) expressly insisting on the exclusiveness of the treaty form for arms control agreementsFootnote ¹⁴¹ – a prescription followed by the Obama Administration despite the near-defeat of the New START Treaty a year later. Even Professor Hathaway’s 2008 antitreaty manifesto Treaties’ End conceded that shifting the approval process entirely to Congress might require a constitutional amendment.Footnote ¹⁴²

Nonetheless, the Article II, Section 2 treaty cannot claim victory over the Restatement’s “interchangeability” view, only survival. Its rivals also persist, and, as discussed below, to some extent have grown stronger. It remains true that at the level of routine diplomacy, the treaty has been largely “eclipsed” (Professor Hathaway’s word)Footnote ¹⁴³ by Executive agreements, which may or may not claim some sort of prior approval from Congress, and may or may not be formally binding in international law. Further, to the extent the claims for the Article II, Section 2 process rest upon a constitutional imperative, difficulty persists in reconciling text and practice. As Ackerman and Golove argued, Professor Tribe’s uncompromising defense of Article II, Section 2 seemed to imply that almost the entirety of U.S. diplomatic practice was unconstitutional.Footnote ¹⁴⁴ The more practical position, associated with Professor Slaughter, that major agreements require Senate supermajority consent, is difficult to reconcile with practice in the trade area and hampered by the extreme difficulty of defining the line between “major” and “minor.” Neither Professor Yoo nor Professor Spiro, in their prominent attacks on interchangeability, managed successfully to tie their policy prescriptions to the Constitution’s text (and indeed Spiro directly affirmed that difficulty).Footnote ¹⁴⁵

The result has been theoretical incoherence, but a reasonable degree of practical accommodation. Despite the Third Restatement’s claim of interchangeability, almost all subsequent major international agreements have been submitted to the Senate through the Article II, Section 2 process, other than in the area of trade law, where postsignature approval by a majority of Congress has been entrenched as the standard approach. In contrast, minor agreements are rarely submitted to the Senate (or to Congress) for after-the-fact approval. There may be debates as to whether particular agreements conform to this pattern, but there seems little doubt that the pattern exists, just as there seems little theoretically satisfactory explanation for it.

B The Modern Scope of After-the-Fact Congressional Approval

The Congressional-Executive agreement, approved after signature by a majority of Congress, is the Article II treaty’s greatest theoretical rival. As Professors Ackerman and Golove argued in 1995, and Professor Hathaway developed further in 2008, approval through Congress is most responsive to democratic and separation-of-powers values.Footnote ¹⁴⁶ It retains a substantial check on executive power, vindicating the Framers’ desire that treaty-making not be entrusted to the president alone. Yet it avoids the antidemocratic result that a small number of senators – perhaps, if they come from small states, representing only a tiny percentage of the population – might block agreements desired by broad majorities.

But despite its theoretical attractions, its academic supporters, and its validation by the Third Restatement, the formal ex post Congressional-Executive agreement has not prevailed in the last twenty-five years of practice. To be sure, it has become entrenched, far beyond where it was at the time of the Restatement, in the trade area. In 1993 NAFTA was approved as a Congressional-Executive agreement without serious constitutional objection, even though it was highly controversial as a policy matter and likely would not have carried two-thirds of the Senate.Footnote ¹⁴⁷ In 1994 NAFTA’s approval was overshadowed by the even more significant approval of the WTO, which itself involved an array of separate agreements covering almost all aspects of trade policy. As noted, unlike NAFTA the WTO did provoke some constitutional objections, but they failed to gain traction; the WTO package was approved by a wide margin. Then, as subsequent rounds of the WTO failed to produce further multilateral agreements, the United States devoted more effort to regional and bilateral free trade agreements – all of which were submitted to Congress as Congressional-Executive agreements. In 2002, Congress reauthorized the “fast-track” procedure for trade agreements until 2007,Footnote ¹⁴⁸ and a new series of agreements were approved under this process, including the Dominican Republic-Central America Free Trade Agreement (CAFTA) and bilateral free trade agreements with Chile, Singapore, Australia, Morocco, Oman, Bahrain, and Peru.Footnote ¹⁴⁹ Constitutional objections were negligible; the Ackerman-Golove/Tribe debate seemed to have run its course with Ackerman and Golove triumphant. The second Bush Administration’s trade agreements were simply assumed to be destined for the Article I approval process, a pattern that continued in the Obama Administration with free trade agreements with Colombia, South Korea, and Panama.Footnote ¹⁵⁰

But Article I’s victory in trade was not reflected elsewhere. As noted above, Congressional-Executive agreements lost out to treaties even in other economic areas, such as investment and intellectual property (except when considered in conjunction with trade), and tax. They did not become a realistic alternative elsewhere, even in areas where a minority of the Senate appeared especially hostile to agreement-making. For example, environmental organizations speculated that climate change agreements might be switched to the Article I process to ease approval, but no effort has been made in that direction. Though the Article I process for arms control remains on the books, no effort was made to use it to circumvent the Senate, even when the Senate blocked the Comprehensive Nuclear Test Ban treaty or threatened to block the New START Treaty. Indeed, in the modern era there has been no repetition of the St. Lawrence Seaway affair, the high-water mark of interchangeability, when the President submitted under Article I an agreement previously defeated under Article II.Footnote ¹⁵¹ Even as the Senate threatened (and ultimately blocked) at least four Obama Administration treaty priorities (the Law of the Sea Convention, the Arms Trade Treaty, CEDAW, and the Disability Rights Convention), and even when the President’s party had majorities in both Houses, there was no material suggestion that these might be withdrawn and resubmitted through the Article I process.

Further, Congressional-Executive agreements, despite their democratic credentials, have been put on the constitutional defensive. The twenty-five years following the Third Restatement witnessed a revival of interest in formalism, textualism, original meaning, and Framers’ values as modes of constitutional interpretation.Footnote ¹⁵² That development has been especially notable (although not uncontroversial) in foreign affairs law, where practicality and the needs of the Cold War had almost entirely displaced ordinary constitutional interpretation in prior years.Footnote ¹⁵³ Among other things, it posed theoretical challenges for Congressional-Executive agreements because they are, despite their attractiveness, hard to reconcile with the formal Constitution.

This challenge was especially sharp in one area. Pure interchangeability, as pressed by the Restatement, proclaimed no limits on the subject matter of Congressional-Executive agreements.Footnote ¹⁵⁴ Such a view might have been plausible at a time when conventional constitutional theory presumed essentially no limits on Congress’ domestic legislative powers. But developments in constitutional theory in the late 1980s and 1990s attacked that presumption by returning to the Framers’ idea of limited enumerated legislative powers. In United States v. Lopez in 1995 and United States v. Morrison in 2000, the Supreme Court established that Congress’ legislative powers were not unlimited. Pure interchangeability could not accommodate that development, because it rested on Article I’s grant of legislative power, and if that grant was not plenary, then Congressional-Executive agreements could not be plenary. So long as Missouri v. Holland remained good law, treaties appeared to have a wider scope.Footnote ¹⁵⁵

Even beyond the issue of enumerated powers, Congressional-Executive agreements lacked firm constitutional defenses, and the modern era did not produce any. Professors Ackerman and Golove may even have undermined Congressional-Executive agreements’ standing by showing them to be largely a post–World War II innovation; their approach depended on a theory of informal constitutional amendment many people did not accept,Footnote ¹⁵⁶ plus (as Professor Spiro argued in 2001)Footnote ¹⁵⁷ a somewhat overblown account of post–World War II events. Subsequent scholarship did little to find a textual or historical basis for a broad Article I agreement power, although much of it accepted Article I approvals in particular areas.Footnote ¹⁵⁸ Thus by the second decade of the twenty-first century, the NAFTA/WTO-type Congressional-Executive agreement in the trade area appeared (both practically and theoretically) as an anomaly, not as a model.

C Executive Agreements: The Dominant Form or a Footnote?

While much academic debate has focused on treaties and Congressional-Executive agreements, numerically speaking the vast majority of U.S. international agreements are made by the president without specific approval of either the Senate or Congress.Footnote ¹⁵⁹ As noted, within this category are at least three distinct subcategories: areas where Congress has given express advance approval of presidential agreement-making; areas where the president claims some implicit or nonspecific advance approval from Congress or a prior Article II treaty; and agreements (either binding or nonbinding) made under the president’s constitutional authority.Footnote ¹⁶⁰

At the outset, it should be noted that, as in prior times, it is often difficult to determine in which subcategory a particular agreement belongs. For example, the Anti-Counterfeiting Trade Agreement (ACTA), mentioned at the outset of this chapter, was signed by President Obama, and its particular terms were not directly approved by Congress or the Senate.Footnote ¹⁶¹ Thus, in the terminology used here, it was an Executive agreement – but what sort? Initially, the Obama Administration claimed that it was made on the president’s sole authority. At the same time, the administration suggested it was nonbinding, at least in the sense that it did not require any changes to U.S. law. After significant objections were raised in Congress and in commentary, the administration shifted ground and explained that the agreement was made under prior Congressional approval.Footnote ¹⁶² However, the new position did not rest on any specific delegation of agreement-making authority to the president, but rather upon statutory language expressing Congress’ view that some (unspecified) actions should be taken to protect intellectual property abroad.Footnote ¹⁶³ Given the various justifications, it seems difficult to say with confidence whether ACTA should be regarded as a sole Executive agreement or a congressionally approved Executive agreement. Indeed, given Dames & Moore’s emphasis on Congressional acquiescence to support a sole Executive agreement,Footnote ¹⁶⁴ the categories may be too blurred to allow reliable categorization.Footnote ¹⁶⁵

Nonetheless, some generalizations may be suggested. First, nonbinding agreements remain an important if underexamined tool in U.S. agreement-making. They have begun to encounter some constitutional resistance, though only a little.Footnote ¹⁶⁶ Second, sole Executive agreements received substantial judicial and academic attention during the modern period, with the ultimate result that they may appear weaker and more constitutionally suspect. Third, as foreshadowed by Dames & Moore, the congressionally authorized Executive agreement is the most consequential subcategory, at least in numerical terms, although it remains unclear how consequential this form is in terms of the overall direction of U.S. foreign relations.