The NAFTA Implementation Act and the Question It Answers

How did a bargain that reorganized trade among three countries become law inside the United States without passing through the treaty process? The NAFTA Implementation Act supplies the answer, and the answer takes the form of an ordinary statute. A signed international bargain, standing alone, changes no tariff rate, opens no customs house, and binds no importer or exporter in an American proceeding. Tariffs, customs procedures, agricultural programs, and procurement rules inside the United States rest on statutes enacted by Congress, and only Congress could rewrite those statutes. The problem confronting the legislators of 1993 was therefore double. A continental agreement had been negotiated and signed, yet the agreement possessed no domestic machinery of its own. Unless Congress supplied that machinery through legislation, the signatures of three heads of government would remain an international promise with no operative rules at the border, in the customs territory, or before an agency.

The formal identity of the measure is compact. The NAFTA Implementation Act is Public Law 103-182, enacted as H.R. 3450 in the 103rd Congress, signed by President Bill Clinton on December 8, 1993, and codified principally at 19 U.S.C. 3301 and following. Through that statute Congress approved the North American Free Trade Agreement, an instrument signed on December 17, 1992 and in force from January 1, 1994, and Congress enacted the changes to domestic law that the agreement required in order to operate. Not a treaty and never was. The Senate gave no advice and consent, no two thirds vote occurred in either chamber, and the House of Representatives, excluded from the treaty process by constitutional design, cast a decisive vote on the bill.

The choice of form governed every later event in the history of the agreement. A treaty would have required two thirds of the Senate and would have given the House no vote at all. A statute required a majority of the House, a majority of the Senate, and a presidential signature, the same combination that enacts an appropriations measure or a tax bill. Section 101 of the act approved the agreement and the Statement of Administrative Action that accompanied the implementing bill. Section 102 then confined the result, providing that no provision of the agreement inconsistent with United States law would have effect and that no person other than the United States would hold a cause of action under the agreement. Approval and limitation arrived in the same enactment, and the domestic life of the agreement ran through the statute rather than around it.

The instrument belongs to the category of the congressional-executive agreement, a form with deep roots in trade practice. From the Reciprocal Trade Agreements Act of 1934 onward, tariff bargaining proceeded under statutory authority, because tariff concessions require changes to schedules that only legislation can deliver. The General Agreement on Tariffs and Trade operated for the United States through statutory authority and implementing legislation, and the implementing acts for later multilateral rounds used the same form. Scholars contested the legitimacy of that practice during the North American debate. Bruce Ackerman and David Golove defended the interchangeability of treaties and congressional-executive agreements, and John Yoo defended the exclusivity of the Treaty Clause, in a debate that the Eleventh Circuit later declined to settle in favor of the treaty route when the court decided Made in the USA Foundation v. United States in 2001.

Passage tested that arithmetic under maximum political pressure. The House approved the bill on November 17, 1993 by 234 to 200, and the Senate approved the bill on November 20, 1993 by 61 to 38. A majority of House Democrats voted against a Democratic president, with 156 Democrats opposed and 102 in favor, while 132 Republicans supplied the larger share of the affirmative coalition. Majority Leader Richard Gephardt and Majority Whip David Bonior organized the opposition inside the president’s party, Minority Whip Newt Gingrich organized support inside the Republican minority, and Vice President Al Gore debated Ross Perot on November 9, 1993 in the most watched public exchange of the contest. The fast track procedure of the Trade Act of 1974 barred amendment and forced each chamber to accept or reject the negotiated package entire, a discipline that converted the roll calls into a single judgment on continental trade policy.

The agreement that the statute approved was an architecture of twenty two chapters rather than a single tariff concession. Chapters on goods established national treatment and scheduled the elimination of duties, with tariffs removed at entry into force on more than one half of Mexico’s exports to the United States and more than one third of United States exports to Mexico, most remaining duties phased out over ten years, and selected agricultural goods carried over fifteen years to 2009. Rules of origin decided which goods earned the preferential rates, through tariff classification changes, regional value content tests, tracing rules for listed components, a yarn forward rule for textiles, and an automotive threshold of 62.5 percent regional value content under the net cost method. Further chapters addressed customs procedures, energy with the Mexican state oil sector reserved, agriculture on separate bilateral tracks, government procurement, services, telecommunications, financial services, competition, temporary entry, and intellectual property.

Two chapters displaced domestic institutions in ways that outlasted the tariff schedules in public argument. Chapter 11 allowed an investor of one party to bring a claim directly against a host government before an international arbitral tribunal, under standards that included national treatment, most favored nation treatment, a minimum standard of treatment, and compensation for expropriation. Chapter 19 replaced judicial review of final antidumping and countervailing duty determinations with binational panels applying the law of the importing country, a mechanism Canada had demanded in the bilateral agreement of 1988 and defended again in the trilateral negotiation. Alongside the main text stood two side agreements on labor cooperation and environmental cooperation, negotiated in 1993 and signed in September of that year, each obliging the parties to enforce domestic law through commissions, offices, submissions, and reports, and each carrying only a narrow ladder toward an arbitral panel.

The profile that follows traces that whole structure in sequence. The opening sections examine the constitutional form and the road from the Mexican proposal of 1990 through the negotiations, the signature of December 1992, and the decision to supplement rather than reopen the signed text. The account then reconstructs the passage fight in Congress, the architecture of goods, tariffs, and origin, the two procedural chapters on investment and duty review, and the side agreements that supplied the price of enactment. The final section follows operation from January 1, 1994 to June 15, 2020, the date of the profile, and records the succession arranged by that date. The United States-Mexico-Canada Agreement had been signed on November 30, 2018, amended by a protocol of December 10, 2019, and implemented by Public Law 116-113, signed on January 29, 2020, with entry into force set for July 1, 2020. On June 15, 2020 the earlier agreement remained the operative framework, and the statute of 1993 remained the governing domestic instrument.

Not a Treaty: The Constitutional Form of the Agreement

Was NAFTA approved as a treaty?

No. The agreement was approved by an ordinary statute, the North American Free Trade Agreement Implementation Act, Public Law 103-182, passed by simple majorities of the House and Senate and signed on December 8, 1993. The Senate never voted on it as a treaty requiring two-thirds consent.

The distinction between a treaty and a statute decided everything else in the legal history of the North American Free Trade Agreement. A treaty enters domestic law through a single constitutional door. The president negotiates and signs the instrument, the Senate gives advice and consent by two thirds, and the president then ratifies it. The House of Representatives has no formal part in that process. A statute must pass both chambers by ordinary majorities and receive the signature of the president, or else pass over a veto by two thirds of both chambers. The implementing act traveled through the second door alone. Whatever legal force the agreement possessed inside the United States on January 1, 1994, the date of entry into force, it possessed because Congress enacted a law whose text defined what the agreement would and would not do in American courts and agencies.

The Treaty Clause of Article II assigns the treaty power jointly to the president and the Senate, and its supermajority requirement is deliberate friction. It forces broad consensus in one chamber before the United States binds itself in a form that Article VI ranks as supreme law alongside the Constitution and the laws of the United States. It concentrates the checking function in the Senate, the chamber designed with longer terms, and it excludes the House entirely. A commitment adopted through that route cannot be completed by the same coalition that passes an appropriations bill.

The roll calls of November 1993 showed a coalition built for a statute and not for a treaty. The House approved H.R. 3450 on November 17, 1993, by a vote of 234 to 200, with 132 Republicans and 102 Democrats in favor and 156 Democrats, 43 Republicans, and 1 independent, Bernie Sanders, against. The Senate approved the bill on November 20, 1993, by a vote of 61 to 38, with 34 Republicans and 27 Democrats in favor, 28 Democrats and 10 Republicans against, and Senator Byron Dorgan not voting. Each tally cleared the only threshold the statutory route imposed, a simple majority of each chamber. Neither tally was taken under the Treaty Clause, and the Senate tally was recorded as a vote on a bill under the ordinary rules for bills. Support and opposition both ran across party lines.

Section 101 of the implementing act performed the act of approval, and its wording defined what approval meant. Through section 101, Congress approved the agreement itself and approved the Statement of Administrative Action submitted with it. The statement described how the executive branch intended to administer the agreement and how administrative practice would change under it, and congressional approval gave that description a standing a unilateral executive document would not have possessed. Approval was a legislative act with legislative consequences. Congress was not advising the president to ratify an instrument negotiated abroad. Congress was declaring, in a public law, that the agreement negotiated among the United States, Canada, and Mexico was approved, and that declaration formed part of the statutory text that courts, agencies, importers, and exporters would later consult.

Approval alone could not make a trade agreement operate, because such an agreement reaches into tariff schedules, customs procedures, rules of origin, agricultural programs, and other subjects already governed by domestic statutes. The implementing act therefore amended domestic law so that the agreement could function inside the United States, adjusting provisions that included customs user fees and sugar and other programs, and it authorized appropriations for the administration the new obligations required. President Bill Clinton then issued Proclamation 6641 to implement the tariff schedule, exercising authority the statute supplied. That sequence, statute first and executive action under the statute second, is the signature of the congressional-executive form. The international text fixed the commitments among the three governments. The domestic statute translated those commitments into rules American officials could apply at a border crossing, in a customs ruling, or in a procurement decision.

Section 102 drew the boundary around that translation, and the boundary was drawn against the agreement itself. The section provided that no provision of the agreement inconsistent with any law of the United States would have effect, and that no person other than the United States would have a cause of action under the agreement. The first rule subordinated the agreement to United States law in the event of conflict, giving a court a statutory instruction that favored the statute. The second rule closed the courthouse door to private enforcement of the agreement as such. An importer who believed a customs decision violated the agreement could not sue on the agreement. Any claim had to rest on domestic law, on the implementing act and the statutes it amended, interpreted in light of the agreement where domestic law permitted that reading.

That structure is the meaning of the statement that the agreement was not self-executing. A self-executing provision operates as domestic law on its own force, without further legislation, and litigants invoke it directly. The North American Free Trade Agreement did not work in that fashion inside the United States. Its domestic legal effect came entirely from the statute that approved it and from the body of domestic law that the statute amended. Where the act changed a tariff provision, a rule of origin test, or a customs procedure, the change in United States law was the operative rule. Where the act left domestic law untouched, the agreement by itself supplied no rule a court could apply against the command of section 102.

The codification reflected that statutory identity. A reader tracing the law through the United States Code begins at 19 U.S.C. 3301 and following, in the title devoted to customs duties, rather than in any compilation of treaties. A reader tracing the enactment through the Statutes at Large begins at 107 Stat. 2057. A reader tracing the legislative history begins with H.R. 3450 in the 103rd Congress, with committee consideration, floor debate, and two recorded roll calls. None of those research paths runs through a Senate resolution of advice and consent, because no such resolution existed. The absence is not a gap in the record. The absence is the record, and it confirms that the constitutional event of 1993 was passage of a bill.

Why a trade agreement could take statutory form at all begins with the allocation of power over commerce. Article I gives Congress the power to regulate commerce with foreign nations and to lay and collect duties. Tariffs are legislation in their nature. They tax imports, they raise revenue, and they appear in statutes that Congress enacts and amends. An agreement promising to reduce or eliminate tariffs therefore commits subject matter at the center of congressional authority, and any reduction has to enter domestic law through the commerce and taxing powers. The congressional-executive agreement grows directly out of that allocation. The president negotiates under authority Congress has granted or with legislation in prospect, Congress approves the result and changes domestic law in the same enactment, and the finished product is a statute rather than an executive act on a subject the Constitution left to the legislature.

History supplied the working model long before 1993. From the Reciprocal Trade Agreements Act of 1934 onward, the trade program of the United States ran on statutory authority rather than on treaties. Tariff bargaining required concessions Congress could deliver only by changing its own schedules, and a negotiating partner had to receive those concessions in a form Congress would honor when the bargaining ended. Statutory authorization solved both halves at once. Congress set the negotiating authority in advance, the executive branch bargained within that authority, and implementing legislation carried the bargain into law.

The fast track procedure of the Trade Act of 1974 was built for that form, and the North American negotiations ran under it. Sections 151 to 154 of that act, codified at 19 U.S.C. 2191 to 2194, created an expedited procedure under which an implementing bill would receive specified handling in both chambers, with limited debate and a vote on passage. The procedure traded congressional control over amendment for control over timing and over the final decision. Congress could reject the finished bargain, and the roll calls of November 1993 proved rejection remained a genuine possibility until the votes were counted. Congress could not rewrite the bargain line by line on the floor, because a bargain amended after signature would cease to be the bargain the other governments had accepted. The authority had been extended in 1991 for two years. The talks had been launched in 1990 after President Carlos Salinas de Gortari of Mexico proposed an agreement between the United States and Mexico, Canada joined, the negotiations concluded on August 12, 1992, and the agreement was signed on December 17, 1992, by President George H. W. Bush, President Salinas, and Prime Minister Brian Mulroney of Canada.

Precedent beyond North America pointed the same direction. The system built around the General Agreement on Tariffs and Trade operated for the United States through statutory authority and implementing legislation rather than through a single treaty submission, and the implementing acts for the Tokyo Round and the Uruguay Round used the congressional-executive form as well. Each use strengthened the claim that major trade commitments belonged to a category of international agreement that Congress and the president could conclude together by legislation. Each use also weakened any argument that scale alone forced a trade agreement into the Treaty Clause, because the multilateral rounds reached tariffs, nontariff barriers, and institutional arrangements of great breadth.

That argument produced the central scholarly contest over the agreement, joined on both sides at full strength. The case for the statutory route had to overcome a textual difficulty. The Constitution mentions treaties expressly and prescribes a demanding procedure for them, and it does not mention congressional-executive agreements by name anywhere. If the Treaty Clause supplied the exclusive method for making significant international commitments, then a statute approving such a commitment used the wrong procedure no matter how large its majorities were, because the procedure itself was the constitutional protection. If the Treaty Clause instead supplied one available method among others, then a commitment within the legislative powers of Congress could proceed by statute, and the choice between methods was a matter of constitutional practice and political judgment. Every position in the debate reduced to a choice between exclusivity and interchangeability.

Bruce Ackerman and David Golove defended interchangeability in their 1995 Harvard Law Review article, titled Is NAFTA Constitutional?, and their defense rested on history as much as on text. In their account, the constitutional settlement of 1945 and the years that followed transformed the making of international agreements. The interwar experience, the Senate rejection of the League of Nations covenant, and the demands of a foreign policy conducted through international institutions persuaded the political branches that the Treaty Clause could not remain the sole channel for major commitments. Congressional-executive agreements emerged from that settlement as a legitimate alternative rather than an evasion, and Ackerman and Golove treated the postwar practice as a constitutional development of the kind that repeated, considered action by the political branches can produce. On that reading, the 1993 agreement stood on ground prepared by decades of practice.

Their argument also carried a democratic claim. A congressional-executive agreement passes through the House of Representatives, the chamber closest to the electorate and the chamber the Treaty Clause excludes. It requires the assent of a majority of the whole national legislature and the signature of the president. Defenders of interchangeability could ask why a procedure engaging more elected institutions, and resting on the enumerated commerce and taxing powers of Congress, should count as constitutionally inferior to a procedure engaging only the president and two thirds of one chamber. The argument recast the supermajority rule as a historical survival rather than the sole guardian of deliberation. In trade, where the substance lay inside Article I, the statutory route also matched the form to the power.

John Yoo argued the opposing view of the Treaty Clause in the same debate, beginning from the proposition that the text means what its structure implies. The founding design placed the treaty power in Article II, hedged it with a supermajority, and assigned the House no role in it, and that design would be pointless if the political branches could route any commitment of treaty scale through ordinary legislation. Interchangeability, on that reading, does not supplement the Treaty Clause. It nullifies the clause in practice, because no president who could win a simple majority would accept the burden of winning two thirds of the Senate. Yoo also pressed the point that practice cannot amend the Constitution by accumulation. A long series of statutory agreements might prove the political branches preferred the easier route, but preference, repeated, does not convert an unconstitutional method into a constitutional one. The formal amendment process exists to prevent constitutional change by drift.

The opposing argument treated the commerce power with suspicion when it was asked to carry the whole agreement. Modern trade agreements reach far past tariffs, and the North American agreement contained chapters on investment, services, intellectual property, and government procurement, alongside the labor and environmental questions addressed in the side agreements negotiated and signed in 1993. A critic of the statutory route could argue that the Commerce Clause becomes an unlimited warrant if every subject touched by international commerce becomes eligible for statutory approval on that basis alone. The Treaty Clause draws no subject matter boundary of that kind. From that side of the debate, the breadth of the agreement counted against its form. An arrangement that reorganized the economic relations of three nations and created standing international institutions resembled the classic subjects of treaty making, whatever label the submitting administration attached to the instrument.

The courts received one direct test of those arguments, and the test ended without the ruling the critics of the statutory route had sought. In Made in the USA Foundation v. United States, 242 F.3d 1300, decided by the United States Court of Appeals for the Eleventh Circuit in 2001, the plaintiffs pressed the claim that the agreement constitutionally required the treaty route and that approval by statute was therefore invalid. The court rejected the claim that NAFTA constitutionally required the treaty route. The decision left the congressional-executive form standing, and no court decision struck the agreement down for the manner of its approval. The judiciary did not supply the supermajority the political branches had not sought, and the statute remained the operative domestic instrument.

The practical consequence of the form reached further than the litigation. Because approval required only simple majorities, the agreement could pass at all. The House margin of 234 to 200 and the Senate margin of 61 to 38 were sound legislative results and inadequate treaty results, and the difference between those two judgments is the significance of the route chosen. An administration forced into the Treaty Clause would have faced a Senate in which 38 senators had recorded opposition and a House whose 234 supporters had no vote to cast. The surrounding politics ran at full intensity. Ross Perot had given opposition its most memorable phrase, the giant sucking sound of jobs moving to Mexico, during the 1992 campaign, and Vice President Al Gore debated Perot on November 9, 1993. Majority Leader Richard Gephardt and Majority Whip David Bonior opposed the agreement, and Minority Whip Newt Gingrich supported it. None of that campaigning changed the arithmetic. The statute needed half of each chamber plus one, and in the House it received little more than that.

What a simple majority enacts, a simple majority can replace, and that symmetry formed the second practical consequence. The agreement rested on a statute from its first day in force, so a successor arrangement could be approved by the same ordinary process, enacted as a new implementing act, and substituted for the earlier framework without a supermajority at any stage. The later record followed that logic. The United States-Mexico-Canada Agreement was signed on November 30, 2018 and amended by a protocol of December 10, 2019, and its Implementation Act, Public Law 116-113, was signed on January 29, 2020. Entry into force for the successor agreement was set for July 1, 2020. On June 15, 2020, the date of this account, the successor agreement had not entered into force, and the North American Free Trade Agreement remained the operative agreement. Replacement of one trade framework by another was a legislative event, completed through an implementing bill in both chambers and a presidential signature, exactly as the original approval had been.

Withdrawal belonged to the same logic. Article 2205 of the agreement, among the final provisions collected in Chapter 22, permitted withdrawal on six months notice. A party government could give that notice, run the six month interval, and leave the agreement without assembling a supermajority at home to undo what no supermajority had created. The domestic and international halves of withdrawal required separate treatment. Notice under Article 2205 addressed the obligations among the three governments. The implementing statute and the body of law built on Public Law 103-182 remained domestic law until Congress changed them, so orderly withdrawal contemplated legislative attention to the statute book even where the international act of notice belonged to the executive.

That separation between the international obligation and the domestic statute governs accurate citation. Between governments, the agreement was the controlling text. The obligations of the United States toward Canada and Mexico, the work of the Free Trade Commission of cabinet-level officials, the state-to-state dispute settlement arrangements of Chapter 20, and the binational panel review of Chapter 19 all ran on the agreement as an international instrument. Inside the United States, the controlling text was the statute. A customs broker calculating a duty, a lawyer advising on a rule of origin, or a judge reviewing an agency determination worked from the United States Code, from the implementing act, and from regulations and proclamations issued under statutory authority. Section 102 stated that limit in terms.

For a reader citing the agreement as law, the discipline follows directly. A citation to the domestic law of the agreement is a citation to Public Law 103-182, to 107 Stat. 2057, or to the codified provisions at 19 U.S.C. 3301 and following, supported where the point requires it by the Statement of Administrative Action that section 101 approved. A citation to the international obligation is a citation to the agreement and its chapters, from the objectives of Chapter 1 through the final provisions of Chapter 22. The two citations answer different questions, even when the same shorthand name covers both. The constitutional form cuts through the confusion. In the law of the United States, the agreement was what Congress enacted on the votes of November 17 and November 20, 1993, and what the president signed on December 8, 1993. It was a statute, with a statute’s authority, a statute’s limits, and a statute’s exposure to repeal and replacement by the ordinary process that created it.

The Road to the Implementing Bill

Why did the United States negotiate a trade agreement with Mexico and Canada?

The United States negotiated the agreement to secure access to the Mexican market for American exporters and investors, to lock in market reforms Mexico had begun under President Carlos Salinas de Gortari, and to extend continental trade rules beyond the Canadian arrangement already in place.

The road began on the northern border rather than the southern one. By the time Mexican officials proposed a formal agreement with Washington, the United States and Canada had already built a bilateral framework for reducing barriers between the two economies. The Canada and United States Free Trade Agreement of 1988 freed most trade between the two countries, and most trade between the United States and Canada was duty free under that agreement by the early 1990s. The 1988 agreement supplied both a precedent and a problem. The precedent was that a comprehensive trade agreement between the United States and a neighboring economy could be negotiated, approved through domestic legislation, and administered through institutions created for that purpose. The problem was that a second bilateral agreement, negotiated only between the United States and Mexico, threatened to leave Canada outside a continental arrangement that affected Canadian exporters, Canadian investors, and the rules governing trade across North America. Canadian participation in the later talks grew from that position. Canada joined negotiations already under way rather than accept exclusion from the rules that would govern the continent.

Mexico supplied the initiative that set the trilateral process in motion. For decades before the late 1980s, Mexican economic policy had relied on protection for domestic industry, restrictions on foreign participation, and a large role for the state in sectors closed to outside competition. President Carlos Salinas de Gortari reversed that orientation. Mexico joined the General Agreement on Tariffs and Trade in 1986, a step that committed the Mexican government to the disciplines of the multilateral trading system and signaled to foreign governments that the turn toward open markets had an institutional form rather than resting only on presidential preference. Salinas then carried the same logic into relations with Washington. In 1990, Salinas proposed a bilateral agreement between Mexico and the United States. The proposal reached an American administration receptive to the idea, and talks launched in 1990 after the proposal.

The reasons American officials gave for accepting the Mexican proposal combined export access, investment security, and the strategic value of binding a reform program into an international commitment. Mexico was a large neighboring market in which American goods faced barriers that Mexican goods, after the reforms of the 1980s, faced in reduced form in the United States. A negotiated agreement promised reciprocal reduction rather than unilateral opening. American exporters of manufactured goods, agricultural products, and services stood to gain scheduled access to Mexican buyers on terms written into chapters of the agreement rather than resting on Mexican administrative discretion. American investors stood to gain rules on treatment, transfers, and expropriation that a later Mexican government could not withdraw without breaching an international obligation and triggering the dispute procedures the agreement created. Members of Congress who supported the negotiation repeated that structure of argument from 1990 to 1993. Supporters described the agreement as a means of fixing in international form reforms that Salinas had carried out by Mexican domestic authority and that a successor administration in Mexico City might otherwise reverse.

The lock in argument shaped how the negotiation was understood in Washington before any text existed. Salinas had opened sectors, reduced barriers, and courted foreign investment as a matter of Mexican policy. Policy, however, could be changed by the same authority that made it. An international agreement changed the cost of reversal. If Mexico accepted obligations on tariffs, investment treatment, services, and government procurement, and if those obligations were paired with dispute settlement and the prospect of losing American market access after a breach, then the reforms acquired an external anchor. American negotiators presented that anchor as the central strategic prize of the talks. United States Trade Representative Mickey Kantor, who inherited the completed main text in 1993 and negotiated the side agreements that accompanied it, defended the agreement on the ground that it bound Mexico into a rules based system. Organized labor rejected that framing. The AFL-CIO under Lane Kirkland opposed the agreement and argued that binding Mexico closer to the American economy under the rules on offer would expose American workers to competition from a labor system that did not protect organizing and bargaining to American standards.

Canada entered the talks from a different starting point and with a narrower set of concerns. The Canadian government of Prime Minister Brian Mulroney had completed the 1988 bilateral agreement with the United States and had defended that agreement through a Canadian election fought substantially over its terms. A separate agreement between the United States and Mexico created two risks for Ottawa. The first was diversion. If Mexican goods received preferential access to the American market under rules that Canadian goods did not share, Canadian exporters could lose position in the market that mattered most to them. The second risk was rule making. A bilateral text would settle questions of origin, investment, services, and dispute settlement for North America without a Canadian voice in the drafting. Mulroney chose participation over protest. Canada joined the negotiations, and the talks that might have produced a hub and spoke pattern, with Washington at the center of two separate bilateral bargains, instead produced a single trilateral text. That choice affected the drafting in concrete ways. The duty review mechanism of Chapter 19, carried over from Chapter 19 of the 1988 agreement and described in the record as Canada’s principal demand in 1988, survived into the new agreement because Canada held a seat from which to defend it. The trilateral form also meant that concessions balanced across three economies rather than two, a feature that later made partial renegotiation so difficult.

No negotiation of that scale could proceed on executive enthusiasm alone. Congress held authority over tariffs and foreign commerce, and any agreement reducing American duties required a domestic legal basis and a congressional vote that trading partners could trust. The statutory machinery that supplied that basis was the fast track authority of the Trade Act of 1974, at sections 151 to 154 of that act, 19 United States Code sections 2191 to 2194. Fast track was a bargain between the branches struck long before the North American talks began. Congress set negotiating objectives in statute, imposed consultation duties on the executive, and received notice at defined stages of the process. In exchange, Congress committed itself to a disciplined procedure for the vote at the end. An implementing bill introduced under fast track moved through committee on a fixed schedule, reached the floor under a rule that barred amendment, and received an up or down vote in each chamber. The executive negotiated in the knowledge that the text produced abroad would not be rewritten line by line at home. Foreign governments negotiated in the same knowledge, which was the point of the device. A Mexican or Canadian concession offered in exchange for an American concession had value only if the American concession survived the legislative process in the form agreed at the table.

The credibility fast track supplied was time limited, and the North American negotiation depended on an extension granted for that purpose. Congress extended the fast track authority in 1991 for two years so that the three governments could conclude the talks. The extension debate functioned as an early congressional referendum on the negotiation itself. Members who voted for the extension did not approve any text, because no text existed in 1991. Those members accepted that the administration of President George H. W. Bush should be able to negotiate with Mexico and Canada under procedures that would make an eventual agreement capable of enactment. Members who opposed the extension recorded objections that returned in fuller form in 1993, including objections to negotiating with a country whose labor and environmental enforcement differed sharply from American practice, and objections to a procedure that denied Congress the power to amend the result. The 1991 extension therefore drew the battle lines two years before the implementing bill arrived. It also imposed deadline discipline on the negotiators. Authority that expired forced pace. The talks concluded on August 12, 1992, within the window the extension had created, and the timing of the conclusion shaped the politics of the year that followed.

What fast track promised each side can be stated precisely. To Mexico and Canada, the procedure promised that concessions accepted at the table would not be pocketed and then subjected to further demands in committee markups and floor amendments. To Congress, the procedure promised objectives stated in advance, consultation during the talks, and a final vote that retained the power of refusal. The power to reject the whole agreement remained intact, and the closeness of the eventual House vote demonstrated that the power was real rather than formal. To the executive, fast track promised a manageable legislative path for a text containing thousands of tariff lines and obligations across goods, agriculture, energy, services, investment, government procurement, intellectual property, and dispute settlement, a text that amendment by amendment consideration could have disassembled. Majority Leader Richard Gephardt and Majority Whip David Bonior, who led the opposition in the House in 1993, treated the same features as a defect. In their account, a procedure that barred amendment forced Congress to accept or reject concessions affecting American workers without the ordinary legislative tools for correcting particular provisions. Minority Whip Newt Gingrich, who supported the agreement, answered that amendment power over a concluded international bargain was limited in substance, because any change required the other governments to accept it, and a Congress that amended freely would be a Congress with which no foreign government could close a deal.

The negotiations ran from the 1990 launch to the conclusion on August 12, 1992. The talks covered the full architecture that the final agreement displayed in twenty two chapters, from objectives and definitions through market access for goods, customs procedures, energy, agriculture, emergency action, standards, government procurement, investment, cross border services, telecommunications, financial services, competition, temporary entry, intellectual property, administration, duty review, institutional arrangements and state to state dispute settlement, exceptions, and final provisions. Negotiating a text of that breadth among three countries at different levels of development required trade offs that no single chapter displayed on its own. Mexico accepted obligations in areas such as intellectual property and investment that went beyond what a developing economy had traditionally accepted in a trade agreement with the United States. Mexico protected its state oil sector under the energy chapter, a reservation recorded in the final text. The United States accepted a long phase out schedule for sensitive products, with most remaining tariffs between the United States and Mexico phased out over ten years and some agricultural goods over fifteen years to 2009. Canada defended the binational panel system for duty determinations that had been its principal achievement in 1988. Each concession was priced against concessions elsewhere in the package, and negotiators on all three sides described the result as a single undertaking in which no part stood apart from the balance of the whole.

The conclusion of the talks in August 1992 placed the finished bargain inside an American presidential campaign, and the campaign altered the conditions under which the text would reach Congress. President Bush sought reelection as the agreement author. Governor Bill Clinton of Arkansas, the Democratic nominee, supported the agreement in principle while stating that labor and environmental protections required supplementation. Ross Perot, an independent candidate for the presidency, made opposition to the agreement a centerpiece of his candidacy. During the 1992 campaign, Perot warned of a giant sucking sound of jobs moving to Mexico if the agreement took effect. The phrase entered the political vocabulary of the debate and remained attached to the agreement through the congressional fight of 1993. Perot attributed the expected loss to the wage difference between the two countries and to the incentive the agreement would create for American employers to move production south of the border in search of lower labor costs. Supporters of the agreement answered with the export and lock in arguments developed during the negotiation, and added that Mexican growth under the agreement would enlarge the market for American goods.

The text was signed on December 17, 1992. President George H. W. Bush signed for the United States, President Carlos Salinas de Gortari signed for Mexico, and Prime Minister Brian Mulroney signed for Canada. The signature took place weeks before a change of administration in Washington, since Clinton had won the November election and would take office in January 1993. Signature did not make the agreement law in the United States. Under the fast track structure, the signed text required an implementing bill, approval of the agreement and of the Statement of Administrative Action that accompanied it, and the changes to domestic law that the bill contained. The outgoing administration therefore left the incoming administration a completed international text and an unfinished domestic task. Clinton could decline to submit the agreement, could seek to reopen it, or could submit it with additions negotiated alongside the original text. The choice among those options occupied the first year of the Clinton presidency and determined the form in which the agreement finally reached Congress.

Reopening the signed text was examined and rejected, and the reasons given at the time explain much about how trade bargains hold together. A reopened negotiation would not have been confined to the provisions the Clinton administration found deficient. Once any chapter returned to the table, Mexico and Canada could return other chapters as well, including provisions the United States valued and had obtained only in exchange for concessions elsewhere. Mexican officials had defended the agreement at home as a balanced settlement, and Salinas had limited room to accept visible American revisions without demanding compensating changes. The Canadian position carried similar constraints. Administration officials concluded that reopening risked collapsing the whole bargain, leaving no agreement, no lock in of Mexican reforms, and no scheduled access to the Mexican market. Kantor pursued the alternative course. The main text remained as signed in December 1992, and the administration negotiated supplemental agreements on labor and on the environment that would enter into force alongside the main agreement. Clinton presented that course as fulfillment of the position taken during the 1992 campaign, in which support for the agreement had been conditioned on additional protections in the two areas where the main text was thinnest.

The side agreements were negotiated during 1993 and signed in September 1993. The North American Agreement on Labor Cooperation created a Commission for Labor Cooperation with a ministerial Council and a Secretariat based in Dallas, required each country to maintain a National Administrative Office to receive submissions, and set out eleven labor principles in its annex. The labor agreement built an escalation ladder of consultations, ministerial consultations, and evaluation by experts, with an arbitral panel available only for a persistent pattern of failure to enforce occupational safety and health, child labor, or minimum wage technical standards. Monetary enforcement assessments under the labor agreement were capped at 0.007 percent of trade in goods, and trade sanctions followed only if an assessment went unpaid. The North American Agreement on Environmental Cooperation created a Commission for Environmental Cooperation with a Council, a Secretariat in Montreal, and a Joint Public Advisory Committee. The environmental agreement allowed citizen submissions under Articles 14 and 15 leading to factual records, and provided party to party consultation and, for a persistent pattern of failure to enforce environmental law, an arbitral panel and possible suspension of benefits.

The Mexican context that American negotiators and members of Congress described during this period requires care, because American descriptions of Mexico were arguments in a domestic debate as much as reports on a foreign country. In the account given by administration officials and by congressional supporters, Mexico before Salinas had been a closed economy whose state dominated major industries, whose tariffs and licensing kept American goods at the border, and whose treatment of foreign investors left American firms exposed to shifting administrative practice. Salinas reforms, in that account, had begun a transformation that the agreement would complete and secure. Approval by Congress would tell Mexican reformers that the United States would meet Mexican openness with American openness, and rejection would weaken the officials who had staked a program on integration with the American market. Opponents in Congress offered a different description of the same country. In the opponents account, Mexican labor law on paper concealed enforcement practice that left independent organizing exposed, Mexican environmental enforcement lacked the resources and the political backing to restrain industrial pollution along the border, and Mexican wages reflected those conditions in ways no side agreement confined to consultations could correct.

The package that arrived in Congress in the autumn of 1993 had three distinct layers, and members considering the implementing bill had to hold all three in view at once. The first layer was the signed main text of December 17, 1992, with its twenty two chapters, its tariff schedules, and its phase out periods running ten and fifteen years for trade between the United States and Mexico, while most trade between the United States and Canada stood on the duty free footing the 1988 agreement had already established. The second layer consisted of the two side agreements negotiated in 1993 and signed in September of that year, each with its own institutions, its own procedures, and its own limited enforcement ladder, neither altering the main text and both scheduled to enter into force with it on January 1, 1994. The third layer was domestic. The implementing bill, H.R. 3450 in the House, translated the international package into American law. Section 101 of the bill as enacted approved the agreement and the Statement of Administrative Action, the executive document that set out how the administration understood the agreement and how domestic agencies would administer it. Other provisions of the bill amended domestic law, addressed customs user fees and particular commodity programs, and authorized appropriations. The President would implement the tariff schedule by proclamation, issued as Proclamation 6641, after enactment.

The Statement of Administrative Action occupied a position between the international text and the statute. The statement was submitted with the implementing bill, was approved by section 101 alongside the agreement itself, and served as the authoritative American explanation of how the agreement would be carried out. For members of Congress, approval of the statement meant that the administration interpretation placed before the chambers during the debate formed part of what the vote endorsed. The fast track procedure bound the three layers together procedurally as well. The bill moved without amendment, so a member could not vote for the main text while striking the domestic changes, and could not add conditions beyond those the side agreements already contained. The vote in November 1993 was in that sense a vote on the entire road traveled since 1990, the Salinas proposal, the trilateral expansion, the 1991 extension, the August 1992 conclusion, the December 1992 signature, the change of administration, and the decision to supplement rather than reopen. Each stage had narrowed the choices remaining at the next, until the choice before the House and Senate stood in the binary form fast track had promised from the start.

Passage: A President Against His Own Party’s Majority

How did the implementing bill move through Congress?

The implementing bill moved under fast track rules that barred amendments, passed through House Ways and Means and Senate Finance after informal mock markups, and reached the floor after an autumn whip count in which Republican support offset Democratic opposition in both chambers.

The procedure that governed passage came from the Trade Act of 1974. Sections 151 to 154 of that act, codified at 19 U.S.C. 2191 to 2194, created an expedited route for legislation implementing a trade agreement negotiated under authority granted in advance by Congress. The executive branch negotiated within parameters Congress had set, consulted committees during the negotiation, and then submitted an implementing bill together with the agreement and a Statement of Administrative Action. Once the bill was formally introduced, the statute imposed a compressed timetable for committee action and floor consideration, limited debate, and prohibited amendments. The North American Free Trade Agreement had been negotiated under that framework after the authority was extended in 1991 for two years. Negotiations concluded on August 12, 1992, and President George H. W. Bush signed the agreement on December 17, 1992, together with President Carlos Salinas de Gortari of Mexico and Prime Minister Brian Mulroney of Canada.

The prohibition on amendments carried the central strategic consequence of the entire passage fight. In ordinary legislative process, a member who supported most of a bill but objected to one title could offer or support an amendment, build a record of attempted repair, and then vote for a revised product. Fast track removed that middle position. The text presented to the House and the Senate was the product of a three country negotiation, and a unilateral change by one chamber would have reopened terms that Canada and Mexico had accepted on the understanding that Congress would vote the bargain up or down as written. Every member therefore faced a single choice on the whole package: tariff reductions phased over years, rules of origin, investment protections, intellectual property provisions, agricultural schedules, customs changes, and the labor and environmental side agreements negotiated in 1993. That structure converted the roll calls into a referendum on trade policy in general, a framing both sides understood and used, because no member could claim to have fixed the objectionable part and no leader could assemble a majority by trading away sections of the text.

Committee consideration followed the same constrained pattern in both chambers. The bill was referred in the House to the Committee on Ways and Means, which holds jurisdiction over revenue, tariffs, and trade agreements, and in the Senate to the Committee on Finance, which holds parallel jurisdiction over customs, tariffs, and reciprocal trade. Under the expedited timetable, the committees had a limited period to report the bill, and failure to act within the statutory window did not kill the measure, since the procedure provided for discharge and floor placement. The practical work of the committees therefore centered on examination rather than revision. Hearings and staff review tested how the agreement would interact with existing customs law, sugar and other commodity programs, rules of origin administration, and the appropriations and fee provisions that the implementing bill itself contained.

Before formal introduction, the process ran through informal channels that the trade statutes and committee practice had developed over successive agreements. Administration officials and committee staff conducted mock markups, sessions in which the draft implementing language was reviewed section by section as if in a formal markup, with committee members and staff signaling objections, requesting changes in the domestic implementing provisions, and identifying points that required clarification in the Statement of Administrative Action. The mock markup device mattered precisely because formal amendment was barred after introduction. Influence had to be exercised before the text froze. A concern raised early enough could be addressed through the drafting of the implementing bill, through administrative commitments recorded in the statement, or through the side agreements that the Clinton administration negotiated during 1993 after the main text had been signed. A concern raised after introduction could only be answered with assurances, report language, or a decision by the member to vote no.

United States Trade Representative Mickey Kantor directed the administration trade effort within that structure. The Office of the United States Trade Representative had negotiated the side agreements on labor cooperation and environmental cooperation, signed in September 1993 and scheduled to enter into force on January 1, 1994 with the main agreement. Kantor and White House legislative staff treated the side agreements as part of the answer to Democratic objections concerning labor standards and environmental enforcement in Mexico, and administration spokesmen presented them as evidence that the agreement submitted in 1993 differed from the text signed in December 1992. The committee process aired both positions without resolving them, because the procedure offered no amendment through which a stronger enforcement title could be substituted and tested on the floor.

The alignment of congressional leaders departed from the normal pattern of a president relying on the majority party leadership to deliver a signature measure. President Bill Clinton was a Democrat, and Democrats held the majority in both chambers of the 103rd Congress. Majority Leader Richard Gephardt, however, organized against the bill, and Majority Whip David Bonior, the official ordinarily charged with counting and persuading Democratic votes for the leadership program, also organized against it. Bonior in particular became a central figure in the opposition whip operation, assembling vote counts for the opponents, distributing material critical of the agreement to Democratic offices, and presenting the case that tariff liberalization with Mexico would disadvantage American workers. Gephardt stated opposition in terms of the effect of the agreement on wages and employment in the United States and on the bargaining position of workers facing employers able to relocate production.

Support inside the House came instead from an unusual combination of the White House, a portion of the Democratic membership, and the organized leadership of the Republican minority. Minority Whip Newt Gingrich supported the bill and helped organize Republican votes for passage. Gingrich and other Republican supporters presented the agreement as an extension of market opening policy associated with Republican administrations, noted that a Republican president had negotiated and signed the agreement, and argued that rejection would damage American credibility with trading partners and forfeit export opportunities in Mexico for American firms and farms. The Republican whip organization therefore supplied structure that the Democratic whip organization, under Bonior, was using in the opposite direction. Vote counting in the autumn proceeded on parallel tracks: administration officials counted a coalition that crossed party lines, Republican whips counted a minority conference that was expected to provide a large share of the affirmative votes, and Democratic opponents counted a majority of the Democratic caucus that was expected to vote no. Each count depended on a bloc of undecided members whose offices received delegations from both coalitions and whose local economic circumstances received close attention from both sides.

The White House gave the coalition a public form on September 14, 1993, when Clinton appeared at the White House with former Presidents Ford, Carter, and Bush. The event placed a Democratic president alongside two Republican former presidents and a Democratic former president, each associated with support for the agreement, and the staging carried a deliberate institutional message. Bush had signed the agreement less than a year earlier, so his presence linked the bill to the administration that negotiated it and denied opponents the argument that the measure belonged to a single party or a single presidency. Carter and Ford extended that continuity further back through earlier trade and foreign policy. Administration officials used the event to argue that passage concerned national commitments and relations with Mexico and Canada rather than ordinary partisan advantage.

Through the autumn, Kantor and the White House legislative operation counted votes member by member. The count distinguished firm supporters, firm opponents, members leaning in either direction, and members who stated that a decision had not been reached. Persuasion followed the classification. Firm supporters received requests to speak to undecided colleagues and to state a position publicly early enough to create momentum. Undecided members received sustained attention: briefings on tariff schedules and rules of origin, meetings with administration officials, information concerning the effect of the agreement on exports from a district or state, and discussion of the labor and environmental commissions created by the side agreements. Because amendment was impossible, persuasion could not take the form of a textual concession offered on the floor. It took the form of explanation, assurance, and commitments concerning administration of the agreement after enactment, together with sector specific understandings. The absence of an amendment route thus concentrated the entire bargaining process in private and committee settings, leaving the formal floor debate to ratify alignments that the whip operations had largely formed in advance.

Organized labor conducted the most visible opposition campaign outside Congress. The AFL-CIO, under its president Lane Kirkland, opposed the agreement and mobilized affiliated unions against passage. Kirkland and federation officials argued that lowering tariffs with a lower wage economy would encourage employers to shift production to Mexico or to use the threat of relocation in bargaining with American workers, and that the labor side agreement lacked enforcement strong enough to protect organizing rights and workplace standards. Local unions carried that argument into congressional districts through meetings, calls, and visits to member offices, directing pressure at undecided Democrats in particular. Labor opposition also carried an electoral dimension stated openly during the campaign. Union officials warned that support from organized labor in future elections could be withdrawn from members who voted for the agreement, a warning that converted a policy vote into a question of a member’s political base.

Business organizations organized with comparable intensity on the other side through USA*NAFTA, a coalition formed to support passage. The coalition represented companies and associations that expected gains from reduced Mexican tariffs, clearer investment rules, and protection for intellectual property, and its members argued that Mexico represented a growing export market that American producers would lose to European and Asian competitors if Congress rejected the agreement. Coalition activity supplied the administration with district level material that the White House used in persuasion, including information about firms exporting to Mexico, suppliers dependent on North American production chains, and agricultural producers seeking access under the tariff schedules. Business supporters also contacted undecided members directly and placed the case for passage in terms of jobs associated with exports, answering the labor argument about jobs associated with import competition and relocation.

Ross Perot supplied the opposition with its most recognizable public spokesman outside organized labor. Perot had campaigned for president in 1992 as an independent critic of the agreement, and his phrase from that campaign, the giant sucking sound, described his prediction of jobs moving south to Mexico if tariffs fell. Perot repeated the argument during the 1993 passage fight in television appearances and in print, directing his case at members who had not committed and at voters whose calls and letters reached congressional offices. On November 9, 1993, Vice President Al Gore debated Perot in a televised forum devoted to the agreement. Gore defended the administration position, argued that exports to Mexico supported American employment, challenged Perot on the accuracy of his claims about the agreement, and presented rejection as a retreat from engagement with a neighboring economy. Perot restated the case against tariff reduction with Mexico and attacked the side agreements as inadequate. Contemporary accounts and participants in the whip operations attributed a movement among undecided members and in public opinion to the debate performance, and administration officials treated November 9 as a turning point after which the count of leaning supporters grew firmer.

Stated reasons for an affirmative vote clustered around market access, foreign policy, and institutional credibility, and named supporters returned to those points in floor debate and public statements. Clinton argued that the agreement would lower barriers facing American exports, create a framework for North American production, and strengthen relations with Mexico at a moment of political and economic change in that country. Kantor argued that the negotiated schedules and rules, including the automotive regional value content requirements and the yarn forward rule for textiles, protected the value of tariff preferences for production within North America rather than production imported from outside the region. Gingrich and Republican supporters argued that Congress should honor a bargain negotiated under authority Congress had granted, that rejection would invite other countries to doubt American commitments, and that expanded trade with Mexico and Canada served Republican principles concerning open markets. Business supporters in USA*NAFTA made a narrower version of the same case, concentrating on tariff lines, procurement access, financial services, telecommunications, and other chapters of direct concern to member firms.

Stated reasons for a negative vote clustered around employment, wages, standards, and the adequacy of enforcement, and named opponents presented those points with equal consistency. Gephardt argued that the agreement failed to protect American workers against competition based on lower wages and weaker enforcement, and that the side agreements did not cure the defect. Bonior argued in the whip operation and in public that members should weigh the experience of workers in districts where manufacturing employment had already contracted, and that assurances about export growth did not answer constituents who expected plant relocation. Kirkland and the AFL-CIO argued that labor rights in the side agreement were subject mainly to consultation and review, with arbitral panels available only for a persistent pattern of failure to enforce occupational safety and health, child labor, or minimum wage technical standards, and that monetary enforcement assessments under that agreement were capped at 0.007 percent of trade in goods. Perot argued that wage differences between the United States and Mexico made job movement a predictable result of tariff elimination and that the agreement served large corporations able to organize production across borders at the expense of workers without that mobility. Republican opponents, a smaller bloc in both chambers, included members who objected on grounds of national sovereignty, constituency level import competition, or distrust of dispute settlement mechanisms, including the Chapter 19 binational panels that replaced judicial review of final antidumping and countervailing duty determinations.

The final votes were assembled in part through concessions that operated outside the statutory text, because the text itself could not be changed. Administration officials gave district level and sector level assurances to members whose support depended on treatment of a particular commodity or industry, and those assurances concerned administration of the agreement, attention to import surges under safeguard provisions, and support for protections affecting citrus, sugar, and winter vegetables, together with other named sectoral concerns raised by members from producing regions. Sugar illustrated the pattern, since the implementing act itself amended sugar and other programs and domestic treatment of sugar imports had direct consequences for producers in a limited number of states and districts whose senators and representatives held pivotal votes. Citrus and winter vegetable producers faced direct seasonal competition from Mexican production, and members representing those districts pressed for commitments concerning monitoring, safeguards, and the handling of perishable and seasonal products under the agreement schedules. Each assurance was narrow, directed at a defined constituency, and intelligible only within the arithmetic of the floor count. The mechanics were those of majority assembly under a no amendment rule: where a textual change was unavailable, the administration substituted commitments about execution, and a member accepting such a commitment could state to constituents that a specific local exposure had been addressed even though the printed bill remained identical for every district.

The House of Representatives passed the bill on November 17, 1993, by a vote of 234 to 200. In favor were 132 Republicans and 102 Democrats. Against were 156 Democrats, 43 Republicans, and 1 independent, Bernie Sanders. The figures record the central fact of the passage story in numerical form. A majority of the president’s own party voted no, since 156 Democrats opposed the bill and only 102 supported it, while Republican votes supplied the margin, since the 132 Republican affirmative votes exceeded the entire Democratic affirmative total and exceeded the overall margin of 34 votes many times over. Without Republican support organized in substantial part through the minority leadership, the bill would have failed on the votes of the majority party alone. Without the 102 Democratic votes that the White House retained despite the opposition of Gephardt and Bonior, Republican support alone would also have fallen short of a majority. Passage therefore rested on a coalition that neither party could have produced internally, assembled under a procedure that denied every member an amended alternative and required a judgment on the agreement as a whole.

The Senate passed the bill on November 20, 1993, by a vote of 61 to 38. In favor were 34 Republicans and 27 Democrats. Against were 28 Democrats and 10 Republicans. Senator Byron Dorgan did not vote. A majority of Senate Democrats also voted no, since 28 Democrats opposed the bill and 27 supported it, a division that mirrored the House pattern within the president’s party even though the overall result differed. The Senate margin was wider in both absolute and proportional terms, and the description of Senate approval as comfortable applies to that margin relative to the House, where 34 votes separated the sides after months of uncertainty concerning whether a majority existed at all. In the Senate, 23 votes separated the sides in a chamber of 100 members, the outcome was not in comparable doubt during the final count, and support crossed party lines in the same configuration as the House, with Republicans providing the larger share of affirmative votes and a minority of Democrats joining them to form the majority. The Senate result confirmed rather than reversed the judgment the House had made three days earlier, and it completed congressional action on H.R. 3450, the bill that became Public Law 103-182.

President Clinton signed the North American Free Trade Agreement Implementation Act on December 8, 1993. The act approved the agreement and the Statement of Administrative Action, enacted the domestic changes required for implementation, and set the framework under which the President issued Proclamation 6641 to implement the tariff schedule. The agreement entered into force on January 1, 1994, together with the labor and environmental side agreements. Passage closed with those two dated facts after a legislative contest in which a Democratic president obtained a central measure over the opposition of a majority of Democratic members in both chambers, on the strength of Republican votes, under a procedure that allowed Congress no third option between approval and rejection of the bargain as negotiated.

The Architecture of the Agreement: Goods, Tariffs, and Origin

The goods provisions formed the commercial core of the North American Free Trade Agreement, and the opening chapters arranged that core in a deliberate sequence. The text first stated its purposes, then supplied definitions for reading those purposes, then imposed national treatment and a program of tariff elimination on trade in goods, then defined the origin rules that decided which goods qualified for the program, and then surrounded the bargain with customs procedures, special treatment for agriculture and energy, and an emergency mechanism for import surges.

Chapter 1 set out the objectives of the agreement, and the implementing legislation treated those objectives as more than a preamble. The objectives included the elimination of barriers to trade in goods and services, the promotion of conditions of fair competition in the free trade area, an increase in investment opportunities, protection for intellectual property rights, the creation of procedures for administration and dispute resolution, and a framework for further cooperation among the three countries. For the goods chapters, the operative objective was the first in that sequence. Barriers fell through scheduled tariff elimination rather than through a single act of removal on January 1, 1994, and the chapters that followed translated the broad commitment into product coverage, staging periods, and conditions of qualification.

The interpretive rule in Chapter 1 required the agreement to be read in light of its objectives and in accordance with applicable rules of international law. An interpretation that defeated an express objective carried a burden that an interpretation advancing the objective did not carry, although the rule never licensed a reader to override clear operative text in the name of purpose. Operative chapters controlled specific questions of duty treatment, origin calculation, and procedure, while the objectives supplied context for gaps and borderline cases. Chapter 2 then supplied the general definitions that made the operative chapters administrable across three legal systems. Definitions fixed the meaning of territory, customs administration, duty, goods, enterprise, and person, and the chapter incorporated understandings associated with the General Agreement on Tariffs and Trade, so that a term with an established meaning in trade practice did not acquire an unrelated meaning merely through placement in a North American text.

What did the agreement do to tariffs?

At entry into force on January 1, 1994, the agreement eliminated tariffs immediately on more than one half of Mexico’s exports to the United States and more than one third of United States exports to Mexico, phased most remaining duties out over ten years, and extended selected agricultural transitions to 2009.

Chapter 3 carried the central goods obligation under the heading of national treatment and market access for goods. National treatment required each party to treat goods of another party no less favorably than like domestic goods once the imported goods had entered the domestic market, in matters covered by the chapter such as internal taxation and regulation tied to sale and distribution. The market access component addressed the border itself. The chapter prohibited the adoption or maintenance of customs duties on originating goods beyond the scheduled phaseout, restricted new duties, and placed limits on other border charges and quantitative restrictions that could have reproduced the effect of a tariff under another name.

The tariff elimination program operated through the schedules annexed to Chapter 3, and each party maintained a schedule that stated the base rate for a product and the path by which the rate fell to zero. At entry into force on January 1, 1994, elimination was immediate for a large share of trade in both directions across the United States and Mexico border. More than one half of Mexico’s exports to the United States entered free of duty from the first day, and more than one third of United States exports to Mexico received the same treatment. Most remaining tariffs between the United States and Mexico then fell in stages across ten years, with reductions spread through the decade rather than concentrated at its end. A smaller set of agricultural goods received fifteen years, with elimination completed in 2009.

The political weight of the program came from an asymmetry that the language of reciprocal liberalization tended to conceal. At the opening of the negotiations, Mexican tariffs on American goods stood substantially higher than American tariffs on Mexican goods. Because Mexican protection started higher, equal movement to zero required larger reductions on the Mexican side and opened a larger share of the Mexican market in proportional terms. The immediate elimination figures reflected the starting positions of the two schedules. A larger fraction of Mexican exports already faced low or zero American duties before 1994, so more than one half could move to duty free treatment at entry into force. Fewer United States exports enjoyed comparable access to Mexico at the start, so the immediate share on that side was more than one third, with the balance carried into the staged reductions.

The asymmetry also explained the distribution of support and opposition during the passage debate in the United States. Exporting industries that faced high Mexican tariffs stood to gain access that had no equivalent in prior bilateral practice, and business coalitions organized in support of approval emphasized that gain. Import competing producers and organized labor focused on the other direction of the same schedule, in which low American tariffs fell to zero and Mexican goods entered under rules that labor organizations regarded as insufficient on labor standards. Trade between the United States and Canada occupied a different starting position. Most trade between the two countries was already duty free under the Canada and United States Free Trade Agreement of 1988, so the 1993 text largely carried forward an accomplished liberalization rather than opening a new one on that border. The significance of Chapter 3 for the northern border lay in consolidation under a single trilateral framework, even though the tariff history of each bilateral relationship differed.

Chapter 4 answered the question that every free trade area must answer and that a customs union answers by a different route. A customs union adopts a common external tariff, so goods admitted into one member circulate without a further inquiry into origin, because the duty cost at the external border is the same at every point of entry. A free trade area keeps separate external tariffs. Under the North American agreement, Canada, Mexico, and the United States each retained a distinct tariff toward the rest of the world, and the difference among those tariffs created an arbitrage opportunity. A third country producer could ship goods into the partner with the lowest external duty and then move the goods across an internal border at the preferential rate, obtaining the benefit of the agreement without producing in North America. Rules of origin existed to close that route.

An originating good under Chapter 4 was a good that satisfied the conditions the chapter prescribed, and qualification depended on production within the territory of the parties rather than on the nationality of the firm that owned the plant or shipped the product. A good wholly obtained or produced entirely in North America qualified on that ground, a category that covered products grown, harvested, or extracted in the region without outside content. Most manufactured goods combined materials from inside and outside the region, and for those goods the chapter supplied tests that measured whether North American production was substantial enough to earn preferential treatment. The tests operated as a gate to the Chapter 3 schedule. A good that failed the origin tests entered the importing country at the ordinary rate applied to goods from the rest of the world rather than at the preferential rate, and the producer lost the preference rather than access to the market itself.

The first principal test was a change in tariff classification, often described as a tariff shift. The Harmonized System classified every traded product and every material under headings and subheadings, and the product specific rules in Chapter 4 stated the change that non originating materials had to undergo for the finished good to qualify. Where a rule required a change of heading, imported materials classified under one heading had to emerge from North American production as a good classified under a different heading. The logic of the test treated a classification change as evidence of substantial transformation, on the ground that the nomenclature itself reflected differences in character and use among products. The level of change varied by product. Some rules required a shift at the chapter level of the nomenclature, others at the heading or subheading level, and some rules combined a classification change with a value content requirement or offered value content as an alternative.

The second principal test measured regional value content, and Chapter 4 supplied two methods of calculation. The transaction value method compared the value of non originating materials with the transaction value of the finished good, the price based measure drawn from the sale of the good, and expressed North American content as a share of that value. The net cost method compared non originating material value with net cost, a measure that started from total cost and removed specified items such as sales promotion, marketing, royalties, shipping and packing, and certain interest costs, so that the calculation concentrated on production cost rather than on price formation after production. For most goods, the general thresholds were 55 percent regional value content under the transaction value method and 50 percent under the net cost method. The five point difference between the thresholds reflected the different denominators of the two calculations rather than a lesser commitment under one method, because transaction value included elements that net cost excluded and a producer measuring against price therefore faced a higher nominal bar to reach an equivalent showing of regional content.

Tracing sharpened that discipline for the components that negotiators regarded as central to a finished good. Under the tracing rules, the value of listed non originating components counted against regional content even when those components passed through intermediate production stages or entered the finished good as part of a larger subassembly produced in North America. Without tracing, a producer could dilute a critical imported component inside a regional subassembly and present the subassembly as originating material, so that the imported content at the heart of the product escaped the calculation. Tracing followed the listed component through the chain and kept its value in the non originating column. The device mattered most in industries built from high value components with concentrated foreign supply, and the automotive rules gave tracing its fullest application.

Textiles and apparel received a rule of distinctive strictness in the yarn forward principle. Under that rule, a textile or apparel good qualified as originating only when production from the yarn stage forward occurred in the territory of the parties, so that spinning or extruding the yarn, weaving or knitting the fabric, and cutting and sewing the finished article each took place within North America. The rule was stricter than a simple change in tariff classification for apparel, because cutting and sewing imported fabric readily produced a classification change without creating demand for regional yarn or regional cloth. Yarn forward closed that gap by moving the origin inquiry upstream to the stage at which the essential material of the industry came into existence. The design protected regional textile mills and yarn producers as the intended beneficiaries of apparel preferences, and apparel makers accepted tighter sourcing limits as the price of preferential access for finished garments.

The automotive rules stood at the top of the origin system in strictness, and the statute stated the thresholds with exactness. Under Chapter 4 and 19 U.S.C. 3332, passenger automobiles, light trucks, and their engines and transmissions had to reach 62.5 percent regional value content under the net cost method, while other vehicles and parts had to reach 60 percent under the same method, figures recorded in CRS Report R44907. The phase in ran from 56 percent for the first category and 55 percent for the second in the years from 1998, and the full levels of 62.5 percent and 60 percent applied from 2002. No transaction value alternative softened the automotive test. Vehicles and core components qualified through net cost alone, with tracing applied to listed components, so the calculation reached into engines, transmissions, and other specified parts rather than stopping at the value of the assembled vehicle.

The height of the automotive threshold reflected the structure of the industry that negotiators confronted and the fear that a loose rule would convert the agreement into an assembly platform for outside producers. Vehicle production in North America was already integrated across the three borders before 1994, with engines, transmissions, stampings, wiring, and finished vehicles crossing between plants in Canada, Mexico, and the United States during a single production sequence. The recorded reasoning of the negotiators also ran toward exclusion. A low threshold would have allowed a producer to import a large share of Asian components, perform final assembly in Mexico at lower wage cost, and sell the finished vehicle in the United States at the preferential rate, capturing the preference for a production pattern that created limited North American content beyond the assembly line. The 62.5 percent figure answered that scenario by requiring the substance of the vehicle, measured in production cost, to be North American. Supporters of the rule in the passage debate presented the threshold as proof that the agreement favored continental production over transshipment and light assembly, while labor organizations that opposed approval argued that content rules alone could not equalize the labor standards under which the North American share was produced.

Chapter 5 gave the origin and tariff system an administrative machinery in the customs procedures provisions. The chapter provided for certificates of origin, records retention, verification, and determinations of origin, and it distributed responsibility among exporters, producers, importers, and the customs administrations of the three parties. A certificate of origin was a declaration, completed by the exporter or producer on the basis of knowledge or records of the production process, that stated a good qualified as originating. The importer claimed preferential treatment on the strength of that certificate and on possession of the supporting information the chapter required. Customs administrations retained authority to verify a claim after entry, through questionnaires and through visits and inquiries conducted under the procedures the agreement prescribed, and an administration that found a claim unsupported could deny preference for the shipment in the circumstances the chapter described.

Chapter 7 took agriculture out of the general goods track and placed farm trade on separate bilateral tracks, so that the commitments between the United States and Mexico, between Canada and Mexico, and between the United States and Canada each followed the schedule and instruments suited to that relationship. Between the United States and Mexico, the chapter and its annexes provided for the phaseout of tariffs and the conversion or elimination of certain import licensing and quantitative measures over the transition periods, with the longest periods reserved for the most sensitive staples. Corn and beans occupied that position for Mexico. Both crops were produced by large numbers of small Mexican farmers for whom the crops carried income and subsistence weight beyond market value, and both faced competition from United States production supported by scale, mechanization, and established export infrastructure. The extended transition for corn and beans gave Mexican agriculture the longest lawful interval in the agreement to adjust to open competition, and the choice of those two products as the outer limit of gradualism stated where the negotiators judged the social cost of rapid liberalization to be highest.

Chapter 8 supplied the emergency mechanism under the heading of emergency action, and the chapter contained two distinct instruments. The bilateral safeguard allowed a party, during the transition period for a good, to suspend further tariff reduction or to restore a duty up to a level tied to the rate applied to goods from the rest of the world, when imports of an originating good had increased in such quantities and under such conditions as to cause or threaten serious injury to a domestic industry producing a like or directly competitive good. The remedy was temporary, bounded by the transition framework, conditioned on an investigation and findings under domestic procedures, and accompanied by compensation or the possible suspension of equivalent concessions by the exporting party. The second instrument reserved rights under the global safeguard law of the wider trading system. The agreement preserved the ability of a party to take emergency action against imports from all sources under the rules governing safeguards in the General Agreement on Tariffs and Trade, and the chapter regulated the treatment of partner imports within such a global action, including the conditions for excluding the other parties when imports from those parties did not account for a substantial share of total imports or did not contribute importantly to the injury found.

Chapter 6 dealt with energy and basic petrochemicals in a shorter compass, and its central feature for the goods architecture was a reservation rather than a liberalization. Mexico protected its state oil sector in the agreement, retaining state ownership and control over the exploration, production, and refining activities that Mexican law reserved to the state and to its state oil enterprise, and the chapter recorded that reservation against the market access and investment disciplines that applied elsewhere. The goods chapters therefore liberalized trade in energy products within stated limits while leaving the ownership structure of Mexican oil untouched, a compromise that allowed the wider bargain to proceed.

The question of whom the goods chapters bound, and what those chapters left untouched, defined the outer limits of the architecture. The chapters bound the three governments in the exercise of customs, tariff, and regulatory authority over goods, and through the implementing legislation the obligations entered domestic law in the form that the 1993 act prescribed, with section 102 of the act providing that no provision of the agreement inconsistent with United States law had effect and that no person other than the United States held a cause of action under the agreement. What the chapters left untouched was equally structural. The agreement created no common external tariff, so each country kept a separate tariff toward the rest of the world. Each country also kept its own trade remedy system of antidumping and countervailing duty law, subject to the binational panel review of Chapter 19 rather than to abolition or harmonization, and each country retained the domestic institutions that investigated injury and calculated duties. The free trade area was an agreement among sovereign tariff systems to admit originating goods on preferential terms, not a merger of those systems into a single customs territory.

Read as a whole, the goods architecture rested on a small number of load bearing choices. Objectives and definitions in Chapters 1 and 2 supplied a common language and a stated direction for interpretation. National treatment and scheduled elimination in Chapter 3 removed border duties and guarded the concessions against internal erosion. Origin rules in Chapter 4 confined the preference to goods with substantial North American production, through classification change, value content, tracing, a yarn forward rule for textiles, and an automotive threshold of 62.5 percent that stood as the strictest bargain in the system. Customs procedures in Chapter 5 made the preference claim provable and auditable. Agriculture in Chapter 7 proceeded on bilateral tracks with the longest transitions reserved for Mexican corn and beans, energy in Chapter 6 proceeded around a Mexican reservation of the state oil sector, and emergency action in Chapter 8 supplied a temporary and bounded response to surges without reopening the destination of zero duties.

The Two Chapters That Outgrew the Agreement: Investment and Duty Review

What was the investment chapter in NAFTA?

Chapter 11 was NAFTA’s investment chapter, Articles 1101 to 1139, guaranteeing foreign investors national treatment, most favored nation treatment, a minimum standard drawn from customary international law, and compensation for expropriation, and allowing an investor to sue a host government directly before an international arbitral tribunal.

Most of the agreement concerned measures at a border. Tariffs fell on a schedule, rules of origin determined whether a good qualified for preferential treatment, and customs procedures governed crossings. Chapters 11 and 19 operated differently. Neither set a tariff rate, and neither defined a qualifying good. Each moved decision making away from a domestic institution that would ordinarily have controlled the outcome. Chapter 11 moved part of the protection of foreign investment away from diplomacy and from host state courts into an arbitral tribunal convened at the election of a private investor. Chapter 19 moved review of final antidumping and countervailing duty determinations away from the domestic court that would ordinarily have reviewed an agency into a binational panel drawn from the two countries concerned. That shared feature explains the attention the two chapters attracted in 1993, and why argument about each returned repeatedly to sovereignty.

The substantive promises of the investment chapter

Chapter 11 applied to investors of one party and to investments of such investors in the territory of another party, in each reciprocal combination among Canada, Mexico, and the United States. Section A set the standards for treatment of such an investor and such an investment, doing work a bilateral investment treaty ordinarily did inside a far broader trade instrument.

National treatment formed the first promise. A host government undertook to accord investors of another party treatment no less favorable than treatment accorded, in like circumstances, to its own investors, with respect to establishment, acquisition, expansion, management, conduct, operation, and sale or other disposition of investments, and the undertaking extended to the investments themselves. The comparison was internal. The question was whether a foreign investor had been treated worse than a domestic investor in like circumstances, not whether treatment matched an abstract benchmark. Treatment could be accorded through practice as well as legislation, so a measure neutral in text could nonetheless be examined for the treatment delivered in operation.

Most favored nation treatment formed the second promise, and the comparison ran outward. A host government undertook to accord investors of another party treatment no less favorable than treatment accorded, in like circumstances, to investors of any other party or of a non party. A preference granted to investors of one foreign state could not be withheld from a NAFTA investor entitled to such treatment. National treatment guarded against preference for domestic investors, and most favored nation treatment guarded against preference for investors of a third country, subject to exceptions and reservations the parties scheduled against the chapter.

The minimum standard of treatment formed the third promise, and it depended on no comparator. Each party undertook to accord covered investments treatment in accordance with international law, including fair and equitable treatment and full protection and security. Interpretation tied that undertaking to the customary international law minimum standard of treatment of aliens. The tie mattered because that standard set a floor independent of how a government treated its own nationals or investors of other states. A government could satisfy both nondiscrimination promises and yet breach the floor, through conduct such as a denial of basic procedural fairness, and the standard framed disputes about arbitrariness that did not reduce to discrimination. Because the floor drew content from law outside the agreement text, its precise height stayed contested among tribunals, governments, and commentators.

Compensation for expropriation formed the fourth promise. No party was to nationalize or expropriate an investment directly or indirectly, or take a measure equivalent to nationalization or expropriation, except for a public purpose, on a nondiscriminatory basis, in accordance with due process of law, and on payment of compensation. Indirect expropriation extended the promise beyond formal seizure of title. A government could leave title untouched and nonetheless face a claim that a measure had destroyed the value of an investment to a degree equivalent to a taking. Compensation was to reflect fair market value at the relevant time, be paid without delay, and be effectively realizable. The boundary between a compensable indirect expropriation and a noncompensable exercise of regulatory authority became the most contested line in the chapter, because significant regulation often reduced asset values, and the chapter could not convert every reduction into a taking without displacing ordinary government.

The performance requirement prohibitions completed the substantive core. A party was barred from imposing specified requirements in connection with establishment, acquisition, expansion, management, conduct, or operation of a covered investment, and from conditioning an advantage on specified commitments. The prohibited categories addressed requirements to achieve a given level of domestic content, to purchase or prefer goods produced in the host territory, to relate imports to exports or foreign exchange earnings, to restrict sales in the host territory by reference to exports, and to transfer technology or a production process to a person in the host territory. The logic connected investment discipline to trade discipline. A government that had surrendered a tariff against a covered good retained an incentive to reach a similar protective result through conditions on the investor producing that good, and the prohibitions closed part of that route.

Section B and the procedural invention

Section B, Articles 1115 to 1138, supplied enforcement, and the mechanism was the invention that distinguished the chapter. An investor of one party could submit a claim to arbitration on its own behalf, or on behalf of an enterprise of the host party that the investor owned or controlled, alleging a breach of Section A and resulting loss or damage. The investor did not petition its own government to espouse the claim as diplomacy. The investor filed directly against the sovereign host state.

Arbitration could proceed under the Convention on the Settlement of Investment Disputes between States and Nationals of Other States, administered through the International Centre for Settlement of Investment Disputes (ICSID), where the conditions for that route were satisfied, under the Additional Facility rules of that centre where only one state concerned belonged to the convention system, or under the arbitration rules of the United Nations Commission on International Trade Law (UNCITRAL). The tribunal applied the agreement and applicable rules of international law. An award could include monetary damages and interest, or restitution of property subject to the government’s option to pay damages instead. A tribunal could not order a government to repeal a statute. Opponents discounted the limit, because a damages award priced a regulatory choice that a government then had to count.

The reasoning attributed to the negotiators who designed Section B ran on two tracks. The first concerned host state courts. An investor committing capital to a plant, a mine, a waste facility, or a distribution network exposed that capital to the host legal system for decades, and a dispute with a municipality, a province or state, or a national ministry would ordinarily be heard in the courts of the very government whose measure was challenged. Negotiators reasoned that a foreign investor could doubt the neutrality of that forum, face procedures written for domestic litigants, and suffer delay destroying the value of the investment before judgment. A neutral tribunal constituted for the single dispute under international rules answered that doubt without requiring proof of actual bias in a particular court. The second track concerned politics. Under the older diplomatic model, an injured investor asked its own government to take up the claim, and that government weighed the claim against every other interest in the bilateral relationship. Direct filing, the negotiators reasoned, depoliticized the dispute by removing the home government’s discretion to espouse or abandon a claim. Among Canada, Mexico, and the United States the mechanism drew sharper attention, because Canada and the United States each possessed developed courts and experienced judiciaries, and arbitration across the northern border treated each system as an insufficient guarantee for the other’s investors.

The litigation record investors built

Ethyl Corporation v. Canada arose from a Canadian ban on the fuel additive MMT. Ethyl, a United States corporation, challenged the measure under the investment chapter, and the dispute settled in 1998. Canada repealed the MMT ban and paid about 13 million dollars. The settlement carried no tribunal holding that the ban had breached the agreement, and that absence did not soften the political effect. Defenders treated the settlement as a commercial resolution of a contested measure. Opponents treated the sequence, a ban withdrawn after a foreign claim and a payment made, as evidence that the chapter reached ordinary public regulation.

Metalclad v. Mexico supplied a large contested award of the early record. Metalclad, a United States corporation, sought to operate a hazardous waste landfill in Mexico and encountered a block at the municipal level after other levels of the Mexican state had dealt with the project. The award issued in 2000 required Mexico to pay about 16.7 million dollars. The case turned on relations between levels of government in a federal state. The measure defeating the project came from a municipality, but the obligation bound Mexico as a party, and Mexico answered internationally for conduct the tribunal attributed to the Mexican state. For investors, the case showed the chapter reaching beyond a discriminatory national statute to permitting decisions across levels of government.

S.D. Myers v. Canada belonged to the same early Canadian docket and extended argument from expropriation toward the nondiscrimination and minimum standard promises, because a measure defended in environmental terms could likewise be tested against domestic competitors and against the floor set by international law. Methanex v. United States tested the opposite boundary. Methanex, a methanol producer, claimed against the United States over a California ban on the gasoline additive MTBE, a substance produced using methanol, asking a tribunal to treat a state environmental measure as a breach of the chapter. The tribunal dismissed the claim in 2005. Loewen v. United States, arising from litigation in the courts of Mississippi and the treatment the claimant alleged in that judicial process, was also dismissed. The Loewen result mattered because the claim had asked an international tribunal to sit in judgment on the fairness of a domestic trial, and dismissal preserved a distance between the investment chapter and appellate review of domestic court proceedings.

Measured across decided claims, the record by the start of 2020 matched neither side’s most sweeping prediction at passage. Investors had prevailed in only a minority of decided claims, and the United States never lost a decided case. Supporters of the mechanism, including the business coalitions gathered in USA*NAFTA that had supported the agreement at passage, read that record as discipline on arbitrary regulation. On that reading, the chapter deterred measures singling out a foreign investor, measures defeating a specific investment through an unfair process, and takings without compensation, while leaving general regulation intact. Methanex stood as proof that a bona fide environmental measure survived scrutiny, and Loewen as proof that a domestic trial did not become an international appeal because a foreign litigant disliked the result.

Critics of the mechanism, including the AFL-CIO under Lane Kirkland, which had opposed the agreement at passage, read the same record from settlements and filings rather than final awards. On that reading, a government did not need to lose a decided case to feel the constraint. Ethyl showed a ban repealed in a settlement, and Metalclad showed a municipal land use decision converted into national liability in millions of dollars. Defense costs, uncertainty from a floor drawn from customary international law, and the risk that a tribunal might call a regulation an indirect expropriation entered a government’s calculation before a measure was adopted. A regulator that softened, delayed, or abandoned a measure for that reason left no award to count, so a minority success rate understated the effect on regulation never brought to a tribunal. That reading treated Chapter 11 as a chilling constraint on legitimate regulation, strongest where public purpose was contested and the affected investment was large.

Binational panels and the review of duty determinations

Antidumping and countervailing duty law allowed an importing country to impose additional duties on goods found to have been sold at less than fair value, or to have benefited from a countervailable subsidy, where dumped or subsidized imports caused material injury to a domestic industry. Such determinations came from domestic agencies and ordinarily faced judicial review in the importing country’s own courts, where a judge tested the agency record against domestic law and the domestic standard of review. Article 1904 of Chapter 19 displaced that route between the parties. At the request of an interested party entitled to seek judicial review, a binational panel reviewed the final determination instead of a domestic court.

Panel composition and governing law defined the compromise. Each panel drew members from the two countries concerned, selected from rosters of persons with expertise in trade law and its administration, and sat as experts rather than as judges commissioned by either state. The panel applied the law of the importing country whose determination was under review, so a panel reviewing a United States determination applied United States law, a panel reviewing a Canadian determination applied Canadian law, and a panel reviewing a Mexican determination applied Mexican law. The standard of review was the standard the importing country’s own court would have applied. The panel examined the agency record, heard argument under the procedures the chapter supplied, and could affirm the determination or remand it to the agency for action not inconsistent with the panel decision.

Extraordinary challenge committees guarded the integrity of that substitute process. Where a party alleged that a panel member had been guilty of gross misconduct, bias, or a serious conflict of interest, that a panel had departed from a fundamental rule of procedure, or that a panel had manifestly exceeded its powers, authority, or jurisdiction, and that such conduct had materially affected the decision and threatened the integrity of the binational panel review process, a committee drawn under the chapter could vacate or remand the panel decision. The threshold was deliberately demanding. An extraordinary challenge was not an appeal on the merits, and disagreement with a panel’s reading of domestic law did not satisfy it.

Binational panel review had been Canada’s principal demand in the negotiation of the 1988 Canada and United States Free Trade Agreement, and Chapter 19 of that bilateral agreement first installed the mechanism between the two countries. Canada sought the mechanism because Canadian exporters faced repeated exposure to United States antidumping and countervailing duty proceedings, and because the Canadian government regarded review by United States courts as an insufficient safeguard against political pressure surrounding trade remedy cases. The United States resisted the demand in 1988 as a surrender of judicial sovereignty. Review of a federal agency by a federal court formed part of the ordinary structure of United States administrative law, and transfer of that review to a treaty panel staffed in part by nationals of the exporting country struck United States negotiators and members of Congress as a transfer of a judicial function outside the constitutional courts. The United States accepted the mechanism provisionally in 1988, preserving the substance of United States trade remedy law while yielding the forum of review for the life of the bilateral agreement. In 1993, when the trilateral negotiation added Mexico, the parties retained binational panel review in Chapter 19 of the North American Free Trade Agreement rather than allowing the bilateral arrangement to lapse.

The softwood lumber disputes supplied the recurring worked example. Lumber crossed the northern border in large volume, Canadian production rested on provincial forest arrangements that United States producers challenged as subsidized, and United States proceedings against Canadian softwood lumber recurred across successive rounds of investigation, determination, and review. Each round placed the same structural question before a panel. A United States agency had determined that Canadian lumber benefited from a countervailable subsidy or had been dumped and had assessed duties on that basis, and Canadian interests sought review. A panel applied United States law and the United States standard of review to the agency record and decided whether the determination survived the test a United States court would have applied. Remands were read in the United States as foreign interference with a domestic remedy and read in Canada as the mechanism performing the function Canada had demanded in 1988, namely application of United States law by a body including Canadian members and standing outside the political atmosphere of the importing country.

The sovereignty objection, pressed by the United States in 1988 and repeated by opponents of the mechanism in Congress, held that judicial review belonged to domestic courts, that an appointed binational panel lacked the constitutional footing and democratic accountability of a court, and that a determination made under domestic law ought to be tested only in the forum the domestic legal system provided. Constitutional questions about the panels were debated throughout the life of the chapter, and no court decision striking the chapter down ever resolved those questions. The answer, pressed by Canada as the demanding party in 1988 and by defenders of the compromise in the United States, held that panels imported no foreign law. A panel applied the importing country’s statute, regulations, and administrative practice as domestic law defined them, and the importing country’s standard of review. The concession concerned forum rather than substance. A government accepting panel review surrendered the exclusive right of its own courts to perform the review and retained the right to write the law under which review proceeded.

Why the two chapters endured as the center of controversy

Of the twenty two chapters of the agreement, Chapters 11 and 19 generated the most durable controversy because each converted an abstract argument about sovereignty into a concrete institution that opponents could name and defenders had to justify. Chapter 11 allowed a private corporation to summon a sovereign government before an international tribunal and obtain a money award for a measure adopted through that government’s own political process. Chapter 19 removed a class of cases from the courts of the importing country and assigned review to panels including nationals of the exporting country. Ethyl could be retold as a fuel ban withdrawn after a foreign claim, Metalclad as a municipal permit decision priced by an international tribunal, and a softwood lumber remand as a domestic agency corrected by a panel on which the exporting country sat. The decided record, in which investors prevailed in a minority of claims and the United States never lost, did not displace those stories, because the controversy had never rested on win rates alone.

The map of the controversy followed the sovereignty argument each side made at passage. Defenders in 1993, including the Clinton administration that signed the implementing act on December 8, 1993 and the business coalition organized as USA*NAFTA, argued that a government bound itself to obtain binding commitments from other governments, that the surrender in each chapter was reciprocal, and that the substance of domestic law survived. A United States investor gained a tribunal against Canada or Mexico because a Canadian or Mexican investor gained the same tribunal against the United States, and a Canadian exporter gained a panel against a United States determination because a United States exporter gained the same panel against a Canadian determination. Opponents in 1993, including Majority Leader Richard Gephardt and Majority Whip David Bonior in the House and the AFL-CIO under Lane Kirkland, argued that reciprocal form concealed asymmetric effect, that institutions answering to no electorate had been created, and that a damages award or a panel remand reached inside the domestic process in a way a tariff concession at the border did not. The House approved the implementing act on November 17, 1993 by 234 to 200, and the Senate approved the act on November 20, 1993 by 61 to 38.

The Side Agreements: The Price of Enactment

What were the NAFTA side agreements?

The North American Agreement on Labor Cooperation and the North American Agreement on Environmental Cooperation were parallel accords negotiated in 1993, signed in September 1993, and placed in force on January 1, 1994, requiring each country to enforce its own labor and environmental law through new trilateral institutions.

The chronology controls the interpretation. Negotiations concluded on August 12, 1992, and President George H. W. Bush, President Carlos Salinas de Gortari, and Prime Minister Brian Mulroney signed the main text on December 17, 1992. By early 1993 the text existed as a closed bargain among three governments. Reopening a tariff line or an investment rule in one chapter would have invited a corresponding demand in another chapter. The agreement therefore arrived in Washington as a document that a new President could accept, reject, or supplement, but could not readily rewrite.

President Bill Clinton took office after a campaign in which support for the agreement had been conditioned on additional protections for labor and the environment. Majority Leader Richard Gephardt opposed the agreement, Majority Whip David Bonior opposed the agreement, and organized labor under Lane Kirkland at the AFL-CIO campaigned against passage. The main text contained chapters on goods, investment, services, and intellectual property, but no discipline requiring any party to enforce domestic labor or environmental statutes. For members whose districts contained manufacturing workers, that absence supplied a concrete reason to withhold a vote.

The administration chose a parallel instrument rather than an amendment. United States Trade Representative Mickey Kantor pursued two separate agreements during 1993, designed to enter into force on the same date as the main agreement. A parallel text left the December 1992 bargain intact and gave Mexico and Canada a bounded set of new obligations to price. A parallel text also gave undecided members a specific commitment to cite and a set of institutions to describe to constituents. Secretary of Labor Robert Reich formed part of an administration that presented labor enforcement as a necessary complement to market access. The agreements were signed in September 1993, two months before the House and Senate votes. A document demanded as a condition of support and signed before a close vote functioned as part of the enactment bargain.

The labor agreement stated its scope in an annex of 11 labor principles. The principles addressed freedom of association and protection of the right to organize, the right to bargain collectively, and the right to strike. The principles addressed the prohibition of forced labor, protections for children and young persons, and minimum employment standards covering matters such as minimum wages and overtime pay. The principles addressed elimination of employment discrimination, equal pay for women and men, prevention of occupational injuries and illnesses, compensation in cases of occupational injuries and illnesses, and protection for migrant workers. Breadth formed part of the legislative purpose, because a member concerned about child labor, workplace safety, or freedom of association could each locate a stated principle inside the text. The force of the annex turned on a separate question, concerning the obligation attached to each principle and the procedure available when a government failed to act.

The core obligation was procedural rather than substantive. Each party undertook to promote compliance with, and effectively enforce, its own labor law. The agreement enacted no continental minimum wage, no continental code on hours or dismissals, and no power to strike down a national statute, order reinstatement, or certify a bargaining representative. Domestic law remained the standard, and the international obligation concerned enforcement of that standard. A submission could challenge a sustained failure to inspect, prosecute, or adjudicate under law already on the books. A submission could not challenge the adequacy of the law itself merely because another country maintained a higher standard.

Institutions gave the obligation a permanent address. The labor agreement created a Commission for Labor Cooperation, governed by a ministerial Council of labor ministers from the three countries, with a Secretariat seated in Dallas. Each country established a National Administrative Office inside its own labor ministry structure. A union, a worker organization, or another interested person could file a submission with an office alleging that another party had failed to enforce labor law in a specific matter. The receiving office reviewed conduct in one of the other two countries, so accountability ran across a border. Hearings, requests for information, and a public report formed the ordinary output of review.

The escalation ladder beyond a report was graduated and narrow at the top. Consultations between the parties came first, and ministerial consultations could produce an agreed course of action or no agreement at all. The Council could then convene an evaluation committee of experts, charged with examining patterns of practice in enforcement and reporting to the Council. Expert evaluation added analysis and publicity but imposed no remedy. Only after those stages could a party request an arbitral panel, and only for a single class of dispute. The dispute had to allege a persistent pattern of failure to effectively enforce occupational safety and health standards, child labor standards, or minimum wage technical labor standards. A single inspection failure or a single contested election could not satisfy that formulation.

The narrowness of the panel gate determined the meaning of the annex. Freedom of association, collective bargaining, and the right to strike generated the most sensitive disputes, because organizing campaigns involved direct contests over representation. None of those principles could carry a dispute to a panel. Discrimination, equal pay, forced labor, and protection for migrant workers also stopped short of the panel stage. For 8 of the 11 principles, the maximum formal outcome was consultation, expert examination, and public reporting. For the remaining three categories, a panel could support a monetary enforcement assessment, capped at 0.007 percent of trade in goods between the parties concerned. Trade sanctions entered only at a further remove, if an assessment went unpaid. The first sanction stage in that sequence was never reached.

The first test arrived in 1994. Mexican unions filed submissions at the Mexican National Administrative Office concerning operations of Honeywell and General Electric in the United States. The submissions alleged failures connected with the rights of workers at facilities in the United States, placing the United States as respondent in a process often described during the 1993 debate as a mechanism directed at Mexican enforcement. The Mexican office accepted the submissions for review, conducted hearings, and produced a report addressing the matters raised. A Mexican labor organization had invoked a trilateral procedure and obtained governmental examination in Mexico of enforcement conduct in the United States. Public examination of organizing disputes through an international filing was a new channel.

The same submissions established the limit in concrete form. Freedom of association and collective bargaining lay near the center of the allegations, and disputes over those principles had no path to a panel and no path to an assessment. The review could clarify facts, convene testimony, and recommend further ministerial consultation. The review could not change an election outcome, compel recognition of a union, or impose a financial consequence. Defenders of the design could point to a functioning office, a hearing record, and a published report within the first year as evidence that the institutions were real. Opponents could point to the same episode and observe that a complaint about the most important principle in the annex terminated in documentation, because the text had assigned documentation as the final remedy for that class of complaint.

The environmental agreement followed a parallel architecture with a more independent public face. The agreement created the Commission for Environmental Cooperation, governed by a Council of cabinet level environmental officials, with a Secretariat seated in Montreal. A Joint Public Advisory Committee, drawn from the public in the three countries, advised the Council. The controlling obligation required each party to ensure that domestic environmental law was effectively enforced. The agreement set no continental emissions standard and empowered no tribunal to rewrite a national statute. Sovereignty over the content of environmental law remained national, and international discipline concerned enforcement, transparency, and cooperation.

The citizen submission process under Articles 14 and 15 supplied the most distinctive procedure in either agreement. A resident or organization in any of the three countries could file a submission with the Secretariat asserting that a party was failing to effectively enforce environmental law. The Secretariat assessed threshold criteria, sought a response from the party whose enforcement was challenged, and advised the Council whether preparation of a factual record was warranted. A factual record gathered information from the submitter, the party, and independent experts, and the Council decided whether to make the record public. A factual record made no binding finding, imposed no penalty, and ordered no cleanup. The record nevertheless created a documented, trilateral account of an enforcement controversy that a domestic agency might otherwise have left inside an administrative file.

A second track addressed disputes between governments. A party that considered another party engaged in a persistent pattern of failure to effectively enforce environmental law could request consultations, with Council involvement in an effort at resolution. If consultations failed, the complaining party could request an arbitral panel. A panel could determine whether a persistent pattern existed, and the parties could agree on an action plan. Failure to implement an action plan could support a monetary enforcement assessment, and suspension of benefits remained available if an assessment went unpaid. Sanctions were therefore theoretically available through a defined sequence. No party carried a dispute through that sequence to an assessment or a suspension of benefits through June 15, 2020.

Comparison between the two agreements clarifies the common bargain. Both agreements converted a political demand, that trade liberalization should not reward a government for neglecting domestic protective law, into procedures for filing, review, consultation, and reporting. Neither agreement converted that demand into a power to harmonize standards upward or to penalize a government for maintaining a lower standard than a trading partner. The labor agreement distributed review through offices housed in national ministries, keeping the first stage close to governments. The environmental agreement placed intake in an international Secretariat and added a standing advisory committee, giving that process a more independent public face. In each case the design privileged a public record over a coercive remedy, and in each case that privilege was a choice made in the text rather than an accident of implementation.

The enforcement critique, stated at full strength, began from that design and treated the Honeywell and General Electric episode as confirmation. The AFL-CIO under Lane Kirkland contended that a labor regime unable to reach freedom of association through a panel could not protect the right that made other standards enforceable, because workers without an independent union lacked a durable means to report safety, child labor, or wage violations without fear of discharge. Environmental organizations that had criticized the 1993 settlement contended that a factual record without a binding finding left an agency free to accept publicity as the cost of inaction. Richard Gephardt and David Bonior pressed a related point in the passage debate. Consultation rewards delay, expert committees reward study, and a respondent government controls much of the information on which a pattern must be proved. The 0.007 percent ceiling compounded the problem, because a capped assessment, available only after an action plan process and collectible through trade measures only after nonpayment, signaled that maximum exposure for a persistent failure was bounded and distant.

Those opponents also challenged the premise that enforcement of existing law was the right unit of discipline. A country could enforce a weak statute faithfully and remain in full compliance, even if the statute permitted conduct that trading partners prohibited. The obligation froze the distribution of standards at enactment and offered no mechanism to raise a floor. Competition for mobile investment operated through the content of law, through rules on elections, dismissal, inspections, penalties, and permitting. A regime that examined only enforcement left that channel outside the discipline the agreements purported to create. The complaint was not that the offices failed to function. The complaint was that the offices functioned as designed, and that the design answered a different question from the question posed when a plant closed in one country and production expanded in another under lower labor costs.

The defense, stated at full strength, answered that the critique measured the agreements against a purpose the negotiators had rejected. Mickey Kantor and Robert Reich defended the agreements as a first establishment of a principle, namely that a trade bargain properly includes labor and environmental commitments alongside tariff commitments. Before 1993, a continental agreement could have been enacted without any labor annex, environmental commission, or public submission procedure, and the absence would have created no legal defect in the trade text. After September 1993, a negotiator proposing silence on labor and environment would have to explain a retreat from a structure that already existed. The defense treated precedent as a form of power, slower than a sanction but capable of altering the starting point of every later negotiation. The agreements were never designed as supranational enforcement, and the defense presented that limit as a stated condition of the bargain rather than a defect discovered afterward.

The defense also rested on institutional permanence and reciprocity. The commissions, the National Administrative Offices, the Montreal Secretariat, and the Joint Public Advisory Committee were standing bodies with staff, dockets, and publication practices. A factual record, an office report, and an expert committee finding each placed official information into public circulation in a form that legislators, unions, and community organizations could cite. A government that faced no fine could face a documented record and ministerial questions all the same from two trading partners. A supranational inspectorate empowered to order remedies inside workplaces in any of the three countries would have been rejected in at least one capital, leaving the main agreement without any complement at all. The United States also accepted reciprocal exposure, as the 1994 submissions demonstrated, and that reciprocity gave the regime a legitimacy that a one directional system aimed only at Mexico would have lacked.

Measured against the vote the administration needed, the side agreements performed the function assigned in 1993. The implementing bill passed the House of Representatives by 234 to 200 on November 17, 1993, with 102 Democrats joining 132 Republicans in favor, and passed the Senate by 61 to 38 on November 20, 1993. President Clinton signed the North American Free Trade Agreement Implementation Act, Public Law 103-182, on December 8, 1993, and the main agreement and both side agreements entered into force together on January 1, 1994. No accounting can assign a specific number of votes to the side agreements, because members also weighed the tariff schedule, the investment chapter, and district level calculations. The role of the side agreements in the contest is nevertheless identifiable. The agreements removed the argument that the package contained nothing on labor or the environment and allowed a member who favored liberalization but feared that objection to vote in favor with a stated basis. A provision can be weak as enforcement and decisive as enactment politics at the same time.

The subsequent record vindicated opponents on sanctions and defenders on institutional survival. Submissions continued to be filed, offices continued to hold proceedings, and factual records and reports continued to appear under both agreements between 1994 and the replacement negotiation that began a quarter century after enactment. No monetary enforcement assessment was imposed under either agreement, and no suspension of benefits was imposed under either agreement. Weak enforcement moved from a prediction offered by opponents in 1993 to the central criticism of the whole settlement in later debate, repeated in congressional hearings, in campaign argument, and in the stated rationale for renegotiation. The target was a model in which labor and environmental obligations lived in separate instruments, rested on enforcement of domestic law, and terminated for most subjects in consultation and publication.

The successor negotiation translated that criticism into a different architecture. The United States-Mexico-Canada Agreement was signed on November 30, 2018 and amended by a protocol signed on December 10, 2019. Its Implementation Act, Public Law 116-113, was signed on January 29, 2020, and entry into force was set for July 1, 2020, after the June 15, 2020 date of the history recorded here. In the form enacted and scheduled at that date, labor and environment moved into chapters of the main text, subject to the dispute settlement system of the principal agreement. The successor text also created a facility specific rapid response labor mechanism, directed at denial of freedom of association and collective bargaining rights at a covered facility, with a timetable and a remedial path aimed at a single workplace rather than at a national pattern proved over years. Freedom of association, placed first in the 1993 annex and last in remedial reach, became the principle around which the successor mechanism was built. The rapid response mechanism answered the Honeywell and General Electric problem by constructing a procedure in which an identified facility, an identified right, and a defined period for action replaced the open ended sequence of consultations.

The side agreements therefore require a double judgment that follows from the texts. As instruments of coercion, the agreements were weak by construction, with panel access confined to occupational safety and health, child labor, and minimum wage technical standards, with assessments capped at 0.007 percent of trade in goods, with trade measures available only after nonpayment, and with environmental sanctions theoretical through June 15, 2020. As instruments of enactment, the agreements were central to the sequence that produced Public Law 103-182, because the agreements answered the condition on which the President had campaigned and the objection around which House opposition had organized. As instruments of precedent, the agreements established commissions, offices, submission rights, and public records that made labor and environment permanent subjects of North American trade law and supplied the acknowledged limits from which the successor design was drawn. The price of enactment was a pair of agreements strong enough to pass and weak enough to sign, and the quarter century between January 1, 1994 and the scheduled succession on July 1, 2020 measured the distance between those two properties.

The Chapter Map

The chapter map condenses the architecture described in the preceding sections into a single reference. Each row states the subject an instrument governed, the constituency that pressed for the instrument during negotiation, and the treatment the instrument received in the successor agreement enacted in January 2020 with entry into force set for July 1, 2020. The map records disposition rather than evaluation, and the prose sections supply the evidence behind each row.

Chapter or instrument	What it governed	Who wanted it	Fate in the successor agreement
Chapters 1 to 3: objectives, definitions, goods	Tariff elimination and national treatment for goods	All three governments	Carried forward in revised form
Chapter 4: rules of origin	Which goods qualify for preferential tariffs	All three governments, with automotive producers pressing for strict content rules	Tightened; automotive content threshold raised to 75 percent
Chapter 6: energy	Energy trade, with Mexico’s state oil sector reserved	Mexico	Carried forward in revised form
Chapter 7: agriculture	Farm trade on separate bilateral tracks	All three governments	Carried forward in revised form
Chapter 10: government procurement	Access to government purchasing	All three governments	Narrowed to the United States and Mexico
Chapter 11: investment	Investor protections and investor-state arbitration	The United States and Canada, extending the 1988 bilateral model	Eliminated between the United States and Canada; narrowed for Mexico
Chapters 12 to 15: services and regulated sectors	Cross-border services, telecommunications, financial services, competition	All three governments	Carried forward and expanded, with digital trade added
Chapter 16: temporary entry	Short-term entry for business persons and professionals	All three governments	Carried forward substantially unchanged
Chapter 17: intellectual property	Copyright, patent, and trademark protection	The United States	Expanded and revised
Chapter 19: duty review	Binational panels replacing court review of duty determinations	Canada, its principal demand since 1988	Retained as Chapter 10 of the successor
Chapter 20: institutions and disputes	Administration and government-to-government dispute settlement	All three governments	Carried forward in revised form
Labor side agreement	Enforcement of each country’s own labor law	The Clinton administration, to assemble the congressional majority	Moved into the main text with a facility-specific rapid response mechanism
Environmental side agreement	Enforcement of each country’s own environmental law	The Clinton administration, to assemble the congressional majority	Moved into the main text

Operation, Amendment, and the Road to a Successor

Is NAFTA still in force?

Yes. On June 15, 2020, the North American Free Trade Agreement remained in force as the operative agreement among the United States, Canada, and Mexico. Its successor had been enacted into United States law in January 2020, with entry into force set for July 1, 2020.

Operation under the North American Free Trade Agreement Implementation Act, Public Law 103-182, began from a distinction that shaped administration and replacement alike. Section 101 approved the agreement and the Statement of Administrative Action, but the act was not self-executing. Section 102 provided that no provision of the agreement inconsistent with United States law had effect, and that no person other than the United States had a cause of action under the agreement. Between the three governments, the agreement created obligations accepted through each party’s domestic process. Inside the United States, enforceable duties ran through the implementing act and the older statutes it amended, rather than through the agreement text standing alone. A customs officer, a court, or an agency applied United States law. The agreement supplied the international commitment behind that law, and the Statement of Administrative Action supplied the executive explanation of administration, but neither displaced a conflicting statute.

The central administrative body was the Free Trade Commission, composed of cabinet-level officials from the three parties. Chapter 18 addressed publication and administration, and Chapter 20 placed institutional arrangements and state-to-state dispute settlement within the structure the Commission supervised. The Commission interpreted the agreement, considered questions about operation, and provided the political level at which disagreements that committees and working groups could not resolve could reach officials with authority to settle them. Routine work was distributed among committees and working groups on goods, agriculture, standards, and customs. Inside the United States, administration remained with departments and agencies that already handled customs, trade negotiation, agriculture programs, and regulatory enforcement, coordinated through the United States Trade Representative and the other cabinet officials representing the United States on the Commission.

Tariff administration showed the pattern in concrete form. The President issued Proclamation 6641 to implement the tariff schedule, exercising the proclamation authority through which staged reductions approved by Congress took effect. At entry into force on January 1, 1994, tariffs were eliminated immediately on more than one half of Mexico exports to the United States and on more than one third of United States exports to Mexico. Most remaining tariffs between the United States and Mexico phased out over ten years, and some agricultural goods phased out over fifteen years, to 2009. Most trade between the United States and Canada was already duty free under the 1988 Canada and United States Free Trade Agreement. Customs administration required origin determinations under Chapter 4, tariff shift and regional value content tests, and tracing rules for listed components. Automotive regional value content reached 62.5 percent under the net cost method for passenger automobiles, light trucks, and engines and transmissions, phased from lower levels between 1998 and 2002.

The implementing act also carried domestic housekeeping that determined how the agreement fit inside the United States Code. The act amended domestic law, including customs user fees and sugar and other programs, and authorized appropriations for administration. Congressional approval was therefore not a single vote of assent followed by unbounded executive discretion. Congress changed fees and program rules in the same instrument that approved the agreement, and administrators operated within the amended text. Chapter 17 on intellectual property was paired with a provision restoring copyright in certain motion pictures, a change requiring statutory language because private rights under copyright law could not be altered by proclamation. Procurement under Chapter 10, temporary entry under Chapter 16, financial services under Chapter 14, and telecommunications under Chapter 13 each depended on existing regulatory statutes and agency practice described in the Statement of Administrative Action. Where a conflict existed, section 102 gave priority to United States law.

Dispute settlement operated on three tracks, and the balance among them defined much of the operational record. Chapter 20 supplied the state-to-state procedure. A party that considered another party in breach could request consultations, refer the matter to the Free Trade Commission, and, if disagreement continued, seek an arbitral panel. Formal Chapter 20 proceedings remained rare in comparison with the volume of trade covered and with activity on the other two tracks. The design assumed that most disagreements would end through consultation or ministerial intervention, because the parties had continuing relationships across hundreds of products and had reasons to avoid a formal loss setting an unwelcome precedent. Rarity did not mean inactivity. Consultations handled questions without a panel report, and the possibility of a panel gave legal shape to negotiations outside public view. Governments chose, case by case, how far to press a claim toward judgment.

Chapter 19 produced a different pattern. Under Article 1904, binational panels replaced judicial review of final antidumping and countervailing duty determinations. Each side nominated panellists, and extraordinary challenge committees stood behind the panels for gross misconduct or comparable breakdown in the process. The mechanism had been carried over from Chapter 19 of the 1988 agreement, where panel review had been Canada’s principal demand, and was retained in 1993. A panel applied the domestic law of the importing country to the administrative record and asked whether the agency determination conformed to that law, preserving each party’s substantive trade remedy statutes while changing the reviewing forum. Trade remedy determinations arrived on a regular cycle, each creating a reviewable record with defined deadlines and stakes, which helps explain heavier use of Chapter 19. The softwood lumber disputes recurred in that environment, showing that panels could remand determinations but could not by themselves produce a durable political settlement of that commodity conflict.

Chapter 11 created the third track, operating without a government sponsor. Section B, Articles 1115 to 1138, allowed an investor of one party to bring a claim against another party for breach of obligations including national treatment, most-favored-nation treatment, the minimum standard of treatment, and compensation for expropriation. Arbitration proceeded under the rules of the International Centre for Settlement of Investment Disputes, its Additional Facility, or the United Nations Commission on International Trade Law. Ethyl Corporation v. Canada settled in 1998, after which Canada repealed the MMT ban and paid about 13 million dollars. Metalclad v. Mexico produced an award in 2000 of about 16.7 million dollars. Methanex v. United States ended with dismissal in 2005, and Loewen v. United States was also dismissed. By the start of 2020, investors had prevailed in only a minority of decided claims, and the United States had not lost a decided case. A covered investor could file directly, and a host government could not block a claim by declining to espouse it, as a government could under Chapter 20.

The side agreements operated through machinery deliberately weaker than the tracks inside the main agreement. The North American Agreement on Labor Cooperation created a Commission for Labor Cooperation with a ministerial Council and a Secretariat in Dallas, and each country created a National Administrative Office to receive submissions. The labor agreement listed 11 labor principles in its annex, but the enforceable obligation centered on effective enforcement of each country’s own labor law rather than a common code. Escalation ran from consultations to ministerial consultations and evaluation by experts. An arbitral panel was available only for a persistent pattern of failure to enforce occupational safety and health, child labor, or minimum wage technical standards. Monetary enforcement assessments were capped at 0.007 percent of trade in goods, trade sanctions were available only if an assessment went unpaid, and the first sanction stage was never reached. The first submission, filed in 1994 by Mexican unions, placed Honeywell and General Electric complaints at the Mexican National Administrative Office against United States companies. The North American Agreement on Environmental Cooperation created the Commission for Environmental Cooperation, with a Council, a Secretariat in Montreal, and a Joint Public Advisory Committee. Citizen submissions under Articles 14 and 15 could lead to factual records, and sanctions under that agreement were never used.

Amendment followed from the constitutional form chosen in 1993. The agreement was a congressional-executive agreement, approved by majorities of both houses and signed by the President, rather than a treaty approved through Senate advice and consent by a two-thirds vote. The Eleventh Circuit addressed the objection in Made in the USA Foundation v. United States, 242 F.3d 1300, decided in 2001, and rejected the claim that the agreement constitutionally required the treaty route. Bruce Ackerman and David Golove defended interchangeability of treaties and congressional-executive agreements in Is NAFTA Constitutional?, published in the Harvard Law Review in 1995, and John Yoo argued the opposing view of the Treaty Clause in the same debate. Change therefore arrived through legislation enacted by Congress and signed by the President, and through protocols among the three parties requiring domestic implementation where United States law had to change. No separate Senate consent procedure governed the path, and the threshold was the ordinary legislative threshold of majority support in both houses and a presidential signature, or a veto override.

Measured against that available procedure, the striking feature of the record between January 1, 1994 and June 15, 2020 was the absence of comprehensive amendment. The main agreement was not rewritten through a general renegotiation during that interval, and the implementing act was not replaced until enactment of Public Law 116-113 in January 2020. Adjustment accumulated at the edges through protocols among the parties, committee decisions, interpretations through the Free Trade Commission, side understandings, and changes in domestic regulations. Tariff staging ran on the schedule fixed at entry into force, and rules of origin administered the categories written into Chapter 4. Stability came from the value each government placed on predictable access terms and from the cost of reopening a bargain in which concessions in agriculture, energy, investment, and procurement had been traded against one another. A party seeking change in one chapter faced demands for offsetting change in another, which left the architecture recognizable across the entire period.

The Uruguay Round Agreements Act of 1994 formed the largest interaction between the implementing act and a later trade statute in the first years of operation. That act implemented the Uruguay Round agreements, which created the World Trade Organization, in United States law one year after the NAFTA implementing act. The two statutes had to be read together by administrators and by firms whose trade crossed both regional and multilateral systems. Rules on safeguards, subsidies, customs valuation, and intellectual property operated under both frameworks, and a measure satisfying one framework could raise a question under the other. Copyright illustrated the continuity. The NAFTA implementing act had restored copyright in certain motion pictures, and Uruguay Round legislation continued the project of aligning United States intellectual property law with negotiated commitments that included restoration for categories of foreign works. Each new agreement arrived as an implementing bill amending the code and adjusting earlier provisions, leaving administrators to reconcile overlapping obligations through regulation and practice.

The limits of the implementing act define the statute as sharply as its operative provisions. The act did not itself create the trade flows that followed entry into force, and it did not itself produce the employment gains or losses attributed to the agreement in the passage debate and the renegotiation debate. Vice President Al Gore and Ross Perot presented opposing expectations in the televised debate of November 9, 1993, with Perot building on a phrase from the 1992 campaign about a giant sucking sound of jobs moving to Mexico. Organized labor under the AFL-CIO, led by Lane Kirkland, opposed passage on grounds including expected pressure on United States workers, while the business coalition USA*NAFTA supported passage on grounds including expected gains from access to the Mexican market. Those statements record positions held by named participants. Neither set of statements answers the empirical question of what portion of later trade or employment change was caused by the agreement rather than by exchange rates, macroeconomic conditions, technology, domestic policy, or firm decisions. That inquiry belongs to the companion impact article, and effect claims in this profile remain attributed positions, not findings.

Other limits were written into the design. The act did not harmonize labor law, and the labor side agreement obliged each party to enforce its own law against the background of the 11 principles in the annex. A wage rule, a union recognition procedure, or a safety standard differing across the border on December 31, 1993 could continue to differ on January 1, 1994 unless a legislature changed it for domestic reasons. The environmental side agreement followed the same structure, creating factual records and consultations instead of a continental environmental code. The agreement also did not create a customs union. A customs union requires a common external tariff, and the parties retained separate external tariffs, separate trade remedy systems subject to panel review, and separate authority to negotiate with outside governments. Preferential origin rules existed precisely because external tariffs differed, and the yarn-forward rule for textiles and the automotive content rules were burdens following from a free trade area that stopped short of a customs union. Finally, the act did not settle the softwood lumber conflict, which predated the agreement and remained a recurring subject across the operational period.

The renegotiation that produced a successor followed a dated sequence. The United States-Mexico-Canada Agreement was signed on November 30, 2018. A Protocol of Amendment followed on December 10, 2019, revising the signed text before implementation legislation completed passage. The House passed the implementation bill on December 19, 2019, by 385 to 41. The Senate passed the bill on January 16, 2020, by 89 to 10. The United States-Mexico-Canada Agreement Implementation Act, Public Law 116-113, was signed on January 29, 2020, with entry into force set for July 1, 2020. On June 15, 2020, that date remained in the future. The North American Free Trade Agreement remained in force, its implementing act remained the governing statute for the regional arrangement, and the successor statute stood enacted with a scheduled commencement that had not yet arrived. Article 2205 of the earlier agreement allowed withdrawal on six months notice, giving each party a formal alternative to negotiated replacement, but replacement arrived through agreement and legislation rather than through withdrawal followed by an interval without a regional framework.

Automotive rules supplied the most detailed successor comparison. The 1993 rules had set regional value content at 62.5 percent under the net cost method for passenger automobiles, light trucks, and engines and transmissions. The successor raised automotive regional value content to 75 percent and added a labor value content rule requiring that 40 to 45 percent of content be produced by workers earning at least 16 dollars an hour. A higher threshold narrowed the share of a vehicle that could originate outside North America while qualifying for preferential treatment, and the wage condition tied part of the calculation to pay rather than geography alone. In the renegotiation debate, supporters of the labor value content rule presented it as a response to wage competition inside the regional production system, and opponents of tighter rules presented higher thresholds as a source of cost and supply chain disruption. The underlying method survived, because preferential treatment depended on regional content accounting and customs verification, but a wage variable entered a field the 1993 text had governed through origin and value alone.

Investment arbitration supplied the clearest reversal. The successor eliminated investor-state arbitration between the United States and Canada and narrowed it between the United States and Mexico. Elimination for Canada removed the direct filing right Chapter 11 had created for that relationship, returning disputes to domestic courts and state-to-state channels. Narrowing for Mexico preserved a reduced form of the mechanism rather than ending it across the agreement. The United States had not lost a decided Chapter 11 case by the start of 2020, and investors had prevailed in only a minority of decided claims, yet the successor coalition chose to curtail the forum. In the renegotiation debate, defenders of the original chapter, including business advocates supporting investment protection, presented arbitration as a safeguard against discriminatory treatment. Critics, including labor and public interest organizations, presented the forum as an avenue for challenging regulation. The enacted restriction indicates that objections concerned the structure of a private claim against a sovereign and distribution of access to the forum, rather than a count of losses alone.

Other parts of the 1993 design survived. Binational panel review was retained in Chapter 10 of the successor, carrying forward review of final antidumping and countervailing duty determinations by binational panels rather than returning review entirely to domestic courts. Retention after more than two decades of softwood lumber litigation indicated that governments and affected industries valued a forum applying domestic law while removing review from the courts of the importing country alone. Labor and environment moved in the opposite direction, from the periphery toward the center. The successor placed labor and environment in the main text, replacing side agreement placement of 1993, and added a facility-specific rapid response labor mechanism. Main text placement changed the relationship of those subjects to dispute procedures, and the rapid response mechanism addressed enforcement at an individual facility rather than only through the country-level pattern analysis of the labor side agreement. In the renegotiation debate, labor organizations that had opposed the 1993 settlement treated main text placement and facility-level enforcement as tests of whether the successor corrected a structural weakness, while business participants pressed for predictable enforcement language for firms operating plants in Mexico. Read together, the changes form a selective audit. Panel review, duty free trade, and origin rules survived, investor-state arbitration contracted, labor and environment advanced into the main text, and automotive rules tightened, marking the points where the original settlement had weakened under experience and political argument.

The closing observation concerns form rather than any single chapter. Replacement was available through the ordinary legislative process because the original commitment had been made through that process. A treaty rests on Senate advice and consent by a supermajority, and that origin gives a treaty a particular entrenchment, since the coalition needed to create the commitment is broader than the coalition needed to pass an ordinary statute and the expectations of a treaty partner form around that higher threshold. A congressional-executive agreement rests on majorities of both houses and presidential approval, the same combination that enacts any other federal statute. Majorities enacted Public Law 103-182, signed December 8, 1993, after House passage by 234 to 200 on November 17, 1993 and Senate passage by 61 to 38 on November 20, 1993. Later majorities enacted Public Law 116-113, signed January 29, 2020, after House passage by 385 to 41 and Senate passage by 89 to 10. No constitutional amendment, supermajority consent proceeding, or judicial invalidation intervened. The Eleventh Circuit decision of 2001 had left the congressional-executive form intact, and the political branches used the flexibility that form provides. Ease in the constitutional sense means the door stood open to a successor statute whenever a president and congressional majorities could agree on terms with Canada and Mexico. On June 15, 2020, the earlier agreement remained in force, with entry into force of the successor set for July 1, 2020, as the final operational fact of the 1993 act.

Closing Assessment: Why the Form Decided the Politics

The record assembled in the profile supports a single organizing judgment. The constitutional form of the North American bargain explains the politics of the bargain, at enactment and at replacement alike. Because the agreement traveled as legislation, the House of Representatives voted, the president’s party could split without defeating the measure, and a coalition of 132 House Republicans and 102 House Democrats could carry a bill that a treaty proceeding would have placed beyond reach. The procedure that made that coalition sufficient had a statutory source, and a reader tracing the device to its origin can consult the Trade Act of 1974 pillar guide for the statutory source of the fast track procedure. The mechanics of the device itself, the bargain under which Congress surrendered amendment in exchange for a guaranteed vote, are set out in how trade promotion authority works for the procedure that carried the bill.

The same form explains the replacement. A commitment created by simple majorities could be retired by simple majorities, and the succession required no supermajority at any stage. Later majorities enacted Public Law 116-113 in January 2020 after House passage by 385 to 41 and Senate passage by 89 to 10, and entry into force for the successor agreement had been set for July 1, 2020, with the agreement of 1993 operative through June 15, 2020. A reader comparing the two settlements chapter by chapter can consult the comparison of NAFTA and its successor for what the successor changed. The longer arc within which both statutes sit, running from the postwar trade rounds to the trilateral succession, is traced in the sequence of United States trade legislation since 1950 for the wider statutory sequence.

The side agreements deserve assessment as price rather than afterthought. The labor and environmental accords of September 1993 altered no tariff line and rewrote no chapter, and the enforcement ladders inside each accord stopped, for most subjects, at consultation, expert examination, and public reporting. Panel access under the labor agreement reached only occupational safety and health, child labor, and minimum wage technical standards, and monetary enforcement assessments were capped at 0.007 percent of trade in goods. Judged as coercion, the instruments were weak by construction. Judged as enactment politics, the instruments answered the condition on which President Clinton had campaigned in 1992 and the objection around which House opposition had organized, and the agreements allowed members who favored liberalization to vote in favor with a stated basis on the two subjects the main text had left unaddressed. The succession ratified the criticism and the precedent at once, moving labor and environment into the main text and building a facility-specific rapid response mechanism around freedom of association, the principle that the 1993 annex had placed first and reached last.

The durable institutional legacy sits in the two procedural chapters. Chapter 11 detached investment protection from diplomacy and from host state courts, and the early record, from the Ethyl settlement of 1998 and the Metalclad award of 2000 to the dismissals in Methanex and Loewen, gave defenders and critics of the mechanism concrete cases rather than abstractions. Chapter 19 detached review of duty determinations from domestic courts, and the softwood lumber disputes gave that transfer a recurring test. The successor settlement audited the two chapters and reached opposite verdicts, curtailing investor-state arbitration while retaining binational panel review, and the divergence confirms that the controversy had always concerned forum and sovereignty rather than tariff levels. Tariff elimination, the commercial core of 1993, generated the least durable argument, because scheduled reductions, once completed, left no institution behind to defend or attack.

One further consequence of the form deserves emphasis. Because section 102 subordinated the agreement to United States law and denied private parties a cause of action under the agreement itself, the operative questions for importers, investors, agencies, and courts remained questions of domestic law throughout the period of operation. A customs determination turned on the Code, a procurement dispute turned on domestic regulation, and a challenge to the constitutional route itself ended in the Eleventh Circuit with the statutory form intact. The international text supplied the commitments among governments. The statute supplied the rules that officials applied, and that division of labor, settled in 1993, governed administration for the entire life of the agreement through June 15, 2020.

The profile has deliberately confined effect claims to attributed positions. Organized labor under Lane Kirkland predicted displacement and warned of relocation pressure, the business coalition USA*NAFTA predicted export gains, and Ross Perot gave the opposition a phrase that outlived the 1992 campaign. Attribution records who held each position and why each position mattered to the vote. Attribution does not measure what the agreement caused in employment, wages, or trade flows across more than two decades of operation, against the background of exchange rates, technology, and macroeconomic change. A reader seeking measurement rather than legislative history can consult the evidence assessment of NAFTA’s economic effects for measured effects, which this profile deliberately does not litigate.

The statute rewards study as a statute. The operative materials are a public law, a codification at 19 U.S.C. 3301 and following, a Statement of Administrative Action approved by section 101, two recorded roll calls, twenty two chapters, two side agreements, and a successor enactment with a scheduled commencement date. Each element carries a citation, a date, and a procedural history that a careful reader can verify against the record. A researcher assembling those materials for sustained work can keep your statute notes, citations, and case chronologies together free on VaultBook as the closing reader step.

Frequently Asked Questions

Q: Is NAFTA a treaty?

No. The North American Free Trade Agreement was approved by an ordinary statute, the North American Free Trade Agreement Implementation Act, Public Law 103-182, passed by simple majorities of the House and the Senate and signed on December 8, 1993. The Senate never voted on the agreement as a treaty, and no two thirds consent proceeding took place. Bruce Ackerman and David Golove defended the interchangeability of treaties and congressional-executive agreements, John Yoo argued the opposing view of the Treaty Clause, and the Eleventh Circuit rejected the claim that the agreement constitutionally required the treaty route in Made in the USA Foundation v. United States, decided in 2001.

Q: How did NAFTA pass Congress?

The agreement passed as an implementing bill under the fast track procedure of the Trade Act of 1974. The bill, H.R. 3450 in the 103rd Congress, moved through the House Committee on Ways and Means and the Senate Committee on Finance after informal mock markups, reached each floor without amendment, and received an up or down vote in each chamber. The House passed the bill on November 17, 1993 by 234 to 200, the Senate passed the bill on November 20, 1993 by 61 to 38, and President Bill Clinton signed the North American Free Trade Agreement Implementation Act on December 8, 1993. The agreement entered into force on January 1, 1994.

Q: Did most Democrats vote for NAFTA?

No. A majority of Democrats in each chamber voted against the implementing bill. In the House, 156 Democrats voted against the bill and 102 Democrats voted in favor, while 132 Republicans voted in favor and 43 Republicans voted against, with 1 independent, Bernie Sanders, also against. In the Senate, 28 Democrats voted against the bill and 27 Democrats voted in favor, while 34 Republicans voted in favor and 10 Republicans voted against, and Senator Byron Dorgan did not vote. Majority Leader Richard Gephardt and Majority Whip David Bonior opposed the agreement, Minority Whip Newt Gingrich supported the agreement, and Republican votes supplied the larger share of the winning coalition in both chambers.

Q: What were the NAFTA side agreements?

The side agreements were the North American Agreement on Labor Cooperation and the North American Agreement on Environmental Cooperation, negotiated in 1993 after the main text had been signed and signed in September 1993. Each agreement entered into force on January 1, 1994 with the main agreement. Each agreement obliged the parties to enforce domestic labor or environmental law rather than adopt a common code. The labor agreement created a Commission for Labor Cooperation with a Secretariat in Dallas and a National Administrative Office in each country. The environmental agreement created a Commission for Environmental Cooperation with a Secretariat in Montreal and a citizen submission process under Articles 14 and 15 leading to factual records.

Q: What was Chapter 11 in NAFTA?

Chapter 11 was the investment chapter of the agreement, Articles 1101 to 1139. Section A guaranteed investors of one party national treatment, most favored nation treatment, a minimum standard of treatment drawn from international law, and compensation for expropriation. Section B, Articles 1115 to 1138, allowed an investor to submit a claim directly against a host government to arbitration under the rules of the International Centre for Settlement of Investment Disputes, the Additional Facility of that centre, or the United Nations Commission on International Trade Law. Ethyl Corporation v. Canada settled in 1998, Metalclad v. Mexico produced an award of about 16.7 million dollars in 2000, and Methanex v. United States was dismissed in 2005. By the start of 2020, investors had prevailed in only a minority of decided claims, and the United States had never lost a decided case.

Q: What was Chapter 19 in NAFTA?

Chapter 19 governed review of antidumping and countervailing duty determinations. Under Article 1904, a binational panel replaced judicial review of a final determination by the courts of the importing country. A panel applied the law and the standard of review of the importing country to the agency record, and a panel could affirm the determination or remand the determination to the agency. Extraordinary challenge committees stood behind the panels for gross misconduct, bias, or a manifest excess of authority. Binational panel review had been Canada’s principal demand in the negotiation of the 1988 Canada and United States Free Trade Agreement, the mechanism was retained in the trilateral agreement of 1993, and the softwood lumber disputes supplied the recurring example of panel review in operation.

Q: When did NAFTA take effect?

The agreement entered into force on January 1, 1994. The chronology that produced that date ran through several distinct steps. Negotiations concluded on August 12, 1992, and the agreement was signed on December 17, 1992 by President George H. W. Bush, President Carlos Salinas de Gortari, and Prime Minister Brian Mulroney. Congress passed the implementing bill in November 1993, and President Bill Clinton signed the implementing act on December 8, 1993. The labor and environmental side agreements, signed in September 1993, entered into force on the same date as the main agreement. At entry into force, tariffs were eliminated immediately on more than one half of Mexico’s exports to the United States and more than one third of United States exports to Mexico.

Q: What were NAFTA rules of origin?

The rules of origin in Chapter 4 decided which goods qualified for preferential tariffs in a free trade area where each party kept a separate external tariff. A good qualified through a change in tariff classification, called a tariff shift, or through regional value content. For most goods, the regional value content threshold was 55 percent under the transaction value method and 50 percent under the net cost method. Passenger automobiles, light trucks, and engines and transmissions had to reach 62.5 percent regional value content under the net cost method, and other vehicles and parts had to reach 60 percent, with the full levels applying from 2002. Tracing rules counted listed non originating components against regional content, and textiles followed a yarn forward rule.

Q: What is the public law number of the NAFTA Implementation Act?

The public law number is Public Law 103-182. The measure was enacted by the 103rd Congress as H.R. 3450, appears in the Statutes at Large at 107 Stat. 2057, and was signed by President Bill Clinton on December 8, 1993. The principal codification of the act sits at 19 U.S.C. 3301 and following, in the title of the United States Code devoted to customs duties. A citation to the domestic law of the agreement therefore runs to the public law, to the Statutes at Large, or to the codified provisions, rather than to a treaty series or to a Senate resolution of advice and consent, because approval took statutory form from the beginning.

Q: Who signed the NAFTA Implementation Act?

President Bill Clinton signed the North American Free Trade Agreement Implementation Act on December 8, 1993, after the House passed the bill on November 17, 1993 and the Senate passed the bill on November 20, 1993. A separate signature belongs to the agreement itself. President George H. W. Bush signed the agreement on December 17, 1992, together with President Carlos Salinas de Gortari of Mexico and Prime Minister Brian Mulroney of Canada. The distinction matters for citation. The signature of December 1992 completed the international instrument among the three governments, and the signature of December 1993 completed the domestic statute that gave the agreement operative effect inside the United States from January 1, 1994.

Q: What did the NAFTA Implementation Act do?

Section 101 of the act approved the agreement and the Statement of Administrative Action submitted with the implementing bill. Section 102 provided that no provision of the agreement inconsistent with United States law would have effect and that no person other than the United States would have a cause of action under the agreement, provisions that made the act the operative domestic instrument rather than the agreement standing alone. Other provisions amended domestic law, including customs user fees and sugar and other programs, authorized appropriations for administration, and restored copyright in certain motion pictures under the intellectual property chapter. President Clinton then issued Proclamation 6641 to implement the tariff schedule under authority the statute supplied.

Q: Why was NAFTA not submitted as a treaty?

The agreement was not submitted as a treaty because trade commitments rest on powers the Constitution assigns to Congress and because trade practice had run on statutory authority for decades. Article I gives Congress the power to regulate commerce with foreign nations and to lay and collect duties, and a tariff concession requires a change to schedules that only legislation can deliver. From the Reciprocal Trade Agreements Act of 1934 onward, tariff bargaining proceeded under statutory authority, the General Agreement on Tariffs and Trade operated for the United States through implementing legislation, and the fast track procedure of the Trade Act of 1974 was built for the statutory form. Critics answered that scale and breadth belonged to the Treaty Clause, and that scholarly contest accompanied the statutory choice.

Q: How did fast track shape the NAFTA vote?

Fast track converted the congressional decision into a single up or down judgment on a negotiated package. The procedure, at sections 151 to 154 of the Trade Act of 1974 and codified at 19 U.S.C. 2191 to 2194, imposed a compressed timetable for committee action and floor consideration, limited debate, and barred amendment after the implementing bill was introduced. Congress had extended the authority in 1991 for two years so the three governments could conclude the talks. Because amendment was barred, influence operated before introduction through consultation and informal mock markups, and persuasion during the autumn of 1993 took the form of explanation and administrative assurances rather than textual concessions offered on the floor of either chamber.

Q: What is a congressional-executive agreement?

A congressional-executive agreement is an international agreement negotiated by the president and approved by a statute passed by majorities of both houses of Congress and signed by the president. The form differs from a treaty, which requires advice and consent by two thirds of the Senate and gives the House of Representatives no vote. The form also differs from a sole executive agreement, which rests on presidential authority alone. Trade agreements became the leading modern use of the congressional-executive form because tariffs and foreign commerce fall within the legislative powers of Congress. Bruce Ackerman and David Golove defended the interchangeability of the two forms for major commitments, and John Yoo argued that the Treaty Clause supplied the exclusive route for commitments of treaty scale.

Q: Which committees handled the NAFTA implementing bill?

The implementing bill was handled in the House by the Committee on Ways and Means and in the Senate by the Committee on Finance. Ways and Means holds jurisdiction over revenue, tariffs, and trade agreements in the House, and Finance holds parallel jurisdiction over customs, tariffs, and reciprocal trade in the Senate. Under the fast track timetable, each committee had a limited period to report the bill, and the procedure provided for discharge and floor placement if a committee failed to act. Before formal introduction, administration officials and committee staff conducted mock markups that reviewed the draft implementing language section by section, a device that mattered because formal amendment was barred after the bill was introduced in the autumn of 1993.

Q: What did the Gore and Perot debate change in the NAFTA fight?

Vice President Al Gore debated Ross Perot on November 9, 1993 in a televised forum devoted to the agreement, eight days before the House vote. Perot had campaigned against the agreement in 1992 and had warned of a giant sucking sound of jobs moving to Mexico, and Perot restated the case against tariff reduction and attacked the side agreements as inadequate. Gore defended the administration position, argued that exports to Mexico supported American employment, and challenged the accuracy of the claims Perot presented. Contemporary accounts and participants in the whip operations attributed a movement among undecided members and in public opinion to the debate, and administration officials treated November 9, 1993 as a turning point after which the count of leaning supporters grew firmer.

Q: Did NAFTA need Senate ratification?

No. Ratification is the step that follows Senate advice and consent in the treaty process, and the agreement never entered that process. The Senate acted on the agreement as a bill, passing H.R. 3450 on November 20, 1993 by 61 to 38 under the ordinary rules for legislation, after the House had passed the bill on November 17, 1993 by 234 to 200. President Bill Clinton signed the bill into law on December 8, 1993, and the agreement entered into force on January 1, 1994 on the strength of the statute and the proclamation that implemented the tariff schedule. A treaty route would have required two thirds of the Senate and would have excluded the House from the decision entirely.

Q: What happened to investor-state arbitration in the successor agreement?

The successor agreement curtailed the mechanism that Chapter 11 had created. The United States-Mexico-Canada Agreement was signed on November 30, 2018 and amended by a protocol of December 10, 2019, and the Implementation Act for the successor, Public Law 116-113, was signed on January 29, 2020. In the enacted successor text, investor-state arbitration was eliminated between the United States and Canada and narrowed between the United States and Mexico, removing the direct filing right for the Canadian relationship and preserving a reduced form of the mechanism for the Mexican relationship. Entry into force for the successor agreement was set for July 1, 2020, and the agreement of 1993 remained the operative agreement on June 15, 2020.

Q: Where is the NAFTA Implementation Act in the United States Code?

The act is codified principally at 19 U.S.C. 3301 and following, in Title 19 of the United States Code, the title devoted to customs duties. The enactment itself appears at Public Law 103-182 and 107 Stat. 2057, and the legislative history runs through H.R. 3450 in the 103rd Congress. Related provisions sit elsewhere in the same title, including the automotive and general rules of origin stated at 19 U.S.C. 3332. A researcher tracing the domestic law of the agreement therefore works from the Code and the Statutes at Large, supported where a point requires the explanation by the Statement of Administrative Action that section 101 of the act approved in 1993.

Q: How could NAFTA be replaced without a new treaty?

Replacement required no new treaty because the original commitment had never been a treaty. A statute enacted by simple majorities could be succeeded by another statute enacted by simple majorities, and the succession followed that path. The United States-Mexico-Canada Agreement was signed on November 30, 2018 and amended by a protocol of December 10, 2019. The House passed the successor implementation bill on December 19, 2019 by 385 to 41, the Senate passed the bill on January 16, 2020 by 89 to 10, and the Implementation Act, Public Law 116-113, was signed on January 29, 2020, with entry into force set for July 1, 2020. Article 2205 of the earlier agreement also permitted withdrawal on six months notice, but replacement arrived through agreement and legislation.