The question behind NAFTA vs USMCA is asked in classrooms, in committee staff rooms, on factory floors, and in farm offices, and the question is rarely asked in the abstract. A student assigned North American trade law wants to know which text to read first and why the second text exists at all. A staffer preparing oversight material wants to know which obligations bind the United States after July 1, 2020 and which dates on the calendar will force decisions. A producer wants to know whether a shipment, a vehicle, or a milk powder consignment faces a different rule under the successor agreement than under the agreement that governed from January 1, 1994 until termination. Each version of the question sounds like a request for a single verdict, and each version conceals a choice about what to measure.

The comparison that follows refuses the single verdict and names the measure instead. Five axes organize the measurement: market access, rules of origin, labor enforcement, investment, and duration and review. On the first axis, the price of crossing the border, the successor agreement changed little, because the earlier agreement had eliminated tariffs on substantially all goods by the end of the transition in 2008 and left a duty free baseline for the successor to inherit. On the remaining axes, the successor agreement changed the conditions attached to that baseline: higher content thresholds, a wage condition inside the origin rules, labor obligations moved into the main text with a facility specific mechanism attached, an investor remedy withdrawn or narrowed, and an expiry date placed on an arrangement that had carried no term.

Holding both answers together is the discipline of the comparison. A reader who measures only border prices will describe the replacement as modest, and the description fits the tariff evidence. A reader who measures who may complain, about what, before which body, and with what consequence will describe the replacement as large, and that description fits the enforcement evidence. The sections that follow take the axes in order, state the earlier rule and the successor rule on each axis at full strength, and reserve the combined verdict for the close, where the deciding factor, the measure chosen, is stated in terms rather than assumed.

Two Agreements, Two Statutes

A comparison between the North American Free Trade Agreement and the United States-Mexico-Canada Agreement begins with a distinction that governs every later measurement. Each agreement was an international bargain among the United States, Canada, and Mexico. Each agreement also depended, inside the United States, on a separate act of Congress. The international text stated what the three governments promised one another. The implementing statute stated what the bargain meant as domestic law, which existing statutes changed, and how federal officials were to carry the promises into practice at borders, in agencies, and in proceedings involving private parties. Confusing the agreement with the statute produces errors about authority, about enforceability, and about the means by which one framework replaced another. Keeping the two instruments distinct produces a cleaner comparison, because the change between 1994 and 2020 was a change in negotiated obligations and a change in enacted law at the same time, accomplished through the same constitutional form in both generations.

The earlier instrument was the North American Free Trade Agreement. The three governments signed the agreement on December 17, 1992. Congress approved the agreement through the North American Free Trade Agreement Implementation Act, Public Law 103-182, enacted as H.R. 3450 in the 103rd Congress and signed on December 8, 1993. The agreement entered into force on January 1, 1994. The codification of the implementing act begins at 19 U.S.C. 3301 and following, in the title of the United States Code devoted to customs duties. The House of Representatives approved the implementing bill on November 17, 1993, by a vote of 234 to 200, and the Senate approved the bill on November 20, 1993, by a vote of 61 to 38. Those votes completed the domestic approval of a bargain that had been negotiated under the fast track procedures of sections 151 to 154 of the Trade Act of 1974. A fuller account of that approval, the roll calls, and the domestic changes the statute enacted appears in the NAFTA implementation act profile, which traces the 1993 statute as a legislative event in its own right rather than as a ratification ceremony attached to a diplomatic text.

The later instrument was the United States-Mexico-Canada Agreement. The three governments signed the agreement at Buenos Aires on November 30, 2018. Congress approved the agreement through the United States-Mexico-Canada Agreement Implementation Act, Public Law 116-113, published at 134 Stat. 11, enacted as H.R. 5430 in the 116th Congress and signed on January 29, 2020. Between signature and approval, the three governments signed a Protocol of Amendment on December 10, 2019, and the statute approved the agreement as amended by that protocol rather than the text of November 2018 standing alone. The House of Representatives approved the implementing bill on December 19, 2019, by a vote of 385 to 41, and the Senate approved the bill on January 16, 2020, by a vote of 89 to 10. The agreement entered into force on July 1, 2020, and on that date the successor agreement replaced and terminated the earlier agreement as the governing framework among the three parties. A fuller account of the 2019 protocol, the passage sequence, and the machinery the 2020 statute created appears in the USMCA implementation act profile, which treats Public Law 116-113 as the product of two negotiations rather than one.

The difference in margins between the two passage sequences belongs to the identity of the instruments, even before any comparison of substance begins. The 1993 statute passed the House by thirty four votes and drew opposition from a majority of the governing party in that chamber, after a campaign in which the side agreements on labor cooperation and environmental cooperation had been negotiated during 1993 to secure support that the signed main text had not commanded on its own. The 2020 statute passed the House by three hundred forty four votes and the Senate by seventy nine votes, after a protocol negotiated during 2019 had altered the signed text on enforcement and other contested points. In the first sequence, supplementation outside the main text supplied the price of passage and the main text of December 1992 reached the statute book unchanged. In the second sequence, amendment inside the text supplied the price of passage and the main text of November 2018 reached the statute book only in revised form. The two statutes therefore record two different relationships between negotiation abroad and bargaining at home, and that difference in process left marks on the substance that a comparison has to register before reaching questions of tariffs or enforcement.

Was the successor agreement approved as a treaty?

No. Both agreements entered domestic law through ordinary statutes passed by majorities in the House and the Senate and signed by the president. Neither agreement received Senate advice and consent by two thirds, and neither agreement operated inside the United States without implementing legislation defining its domestic effect.

The shared constitutional form was the congressional-executive agreement. Under that form, the president negotiates an international bargain on subject matter that falls substantially within the legislative powers of Congress, and Congress approves the bargain and enacts the domestic changes the bargain requires in a single statutory process. The form mattered for North American trade because tariffs are legislation in character. A promise to eliminate a duty, to open a tariff rate quota, or to alter a customs procedure promises a change in rules that Congress had enacted and that only Congress could revise through law. An agreement confined to executive action could state an international intention, but the intention could operate at an American port of entry only through the statute book. The implementing acts performed that translation in both generations. Each act approved an agreement, and each act amended or supplemented domestic law so that agencies possessed authority aligned with the approved text.

Neither agreement was a treaty. The Senate gave no advice and consent to either instrument by the two thirds vote the Treaty Clause requires, and the House of Representatives, which holds no part in the treaty process, cast recorded votes that decided the fate of both implementing bills. The absence of a treaty vote was not an omission in the record of either generation. The statutory route was the chosen route in 1993, under procedures that the Trade Act of 1974 had supplied for trade agreements, and the statutory route was the chosen route again in 2020, under the trade promotion authority framework that governed the renegotiation. A reader searching for the domestic authority behind either agreement searches the Statutes at Large and the United States Code rather than a resolution of advice and consent. Public Law 103-182 and Public Law 116-113 are the operative domestic citations, and the roll calls of November 1993 and December 2019 and January 2020 are the operative domestic votes.

Neither agreement was self-executing inside the United States. A self-executing provision operates as domestic law through its own force, without further legislation, and private parties invoke such a provision directly in court. The North American agreements did not function in that manner under the implementing acts that approved them. Domestic effect ran through the statutes. Where an implementing act changed a tariff provision, a rule of origin test administered by customs officials, or a procedure governing an agency decision, the change in United States law supplied the rule that officials applied and that courts reviewed. The agreements bound the United States toward Canada and Mexico as international instruments, and the statutes governed conduct inside the United States as domestic instruments. The comparison between the two generations therefore operates on two levels throughout. One level is the bargain among governments, expressed in chapters, annexes, and tariff schedules. The other level is the enacted law of the United States, expressed in public laws that approved the bargains and adjusted the domestic code to receive them.

The architecture that the two statutes approved differed in placement as much as in content, and placement formed part of the legal identity of each instrument. The 1993 statute approved a main text accompanied by two side agreements, the North American Agreement on Labor Cooperation and the North American Agreement on Environmental Cooperation, negotiated after the main text to secure votes for passage. Labor and environmental obligations occupied instruments beside the agreement rather than chapters inside the agreement. The 2020 statute approved a main text in which labor appeared as Chapter 23 and environment appeared as Chapter 24, inside the agreement itself. The later text also contained chapters with no counterpart in the earlier text, including digital trade as Chapter 19 and macroeconomic policy and exchange rate matters as Chapter 33. State-to-state dispute settlement appeared as Chapter 20 in the earlier agreement and as Chapter 31 in the later agreement. Binational panel review of antidumping and countervailing duty determinations appeared as Chapter 19 in the earlier agreement and survived as Chapter 10 in the later agreement. A comparison that counts only chapter titles misses the significance of placement, because an obligation housed in a side agreement with a limited enforcement ladder and an obligation housed in a main text chapter subject to the general dispute settlement system occupy different legal positions even where the words describing the underlying standard resemble one another.

The renegotiation path that produced the second instrument began inside the first instrument. Administration officials elected in 2016 had campaigned on a pledge to renegotiate the earlier agreement on revised terms or to withdraw from the agreement if renegotiation failed to produce an acceptable result, and officials maintained that position after taking office. Withdrawal carried legal substance because the earlier agreement contained an exit provision, Article 2205, permitting a party to withdraw six months after giving written notice to the other parties. No consent from the other governments was required to start that interval, and no amendment to the agreement was required to make departure possible. The second negotiation was conducted under the shadow of that clause from the opening of talks to the signing of the protocol. Canada and Mexico negotiated with a partner whose stated alternative to revised terms was termination on six months notice, and the United States negotiated with partners whose exporters, investors, and supply chains depended on a framework that the exit clause placed under a defined and short horizon. The clause did not dictate any particular outcome on dairy access, automotive content, labor enforcement, or duration. The clause set the condition under which every substantive question was bargained, because failure to reach a revised bargain left withdrawal as the declared fallback rather than continuation of the earlier bargain on settled terms.

Formal negotiations opened in August 2017, under objectives and consultation requirements that Congress had enacted in advance for trade agreements of that class. The talks proceeded as a trilateral negotiation, reached a bilateral understanding between the United States and Mexico in late August 2018, and recovered trilateral form when Canada joined the arrangement on September 30, 2018. The three governments signed the resulting text at Buenos Aires on November 30, 2018. Signature completed the diplomatic stage of the renegotiation. Signature did not approve the agreement for the United States, did not change domestic law, and did not bring any provision into force. The signed text entered 2019 awaiting an implementing bill in a House of Representatives that had passed to a different party majority in the election of November 2018, and the interval between signature and introduction became the decisive legislative stage of the whole sequence.

During 2019, the House majority declined to schedule an implementing bill until enforcement concerns in the signed text had been addressed. A working group of House members negotiated with the Office of the United States Trade Representative through that year, and the administration faced a choice between a signed text that could not reach a vote and an amended text that could command a majority. Mexico enacted a labor law reform on May 1, 2019, during the same interval, and Mexico ratified the original agreement in June 2019, before the protocol existed. Neither development closed the gap in Washington, because the objection pressed by the House majority concerned the enforceability of the international text itself rather than the content of Mexican domestic law standing alone. The second negotiation, conducted between the House majority and the executive branch and then carried back to the three governments, produced the Protocol of Amendment signed on December 10, 2019. The Congressional Research Service treated the protocol as a set of targeted revisions rather than a reopening of the whole bargain, and the description fits the sequence. Tariff schedules and most chapter architecture carried over from the signed text. Enforcement machinery, the burden governing labor and environmental claims, a facility-specific rapid response mechanism that the signed text had not contained, the formation of dispute settlement panels, and a small set of other provisions changed before any American vote on approval occurred.

The contrast with the earlier path sharpens the identity of both statutes. In 1992 and 1993, a signed main text remained intact, administrations changed between signature and passage, and negotiators answered congressional objections by adding side agreements beside the text rather than by amending the text. In 2018 and 2019, a signed main text remained short of a vote for more than a year, and negotiators answered congressional objections by amending the text through a protocol before the implementing bill was introduced. The first path treated the signed bargain as fixed and built enforcement commitments around the margin of the bargain. The second path treated the signed bargain as open until the legislature that had to approve the bargain accepted the enforcement terms, and built enforcement commitments into the center of the bargain. Both paths respected the same constitutional allocation. The president negotiated, Congress approved or withheld approval, and a statute carried the result into domestic law. The leverage point differed, moving from supplementation after signature in the first generation to amendment after signature in the second generation, and that movement explains why the protocol occupies a position in the later record that no document occupies in the earlier record.

Entry into force closed the sequence and fixed the dates from which comparison proceeds. Canada completed parliamentary approval in March 2020, notifications among the three governments followed, and the successor agreement entered into force on July 1, 2020, about ten weeks before the date of comparison on September 15, 2020. On July 1, 2020, the earlier agreement terminated as the governing framework and the later agreement governed in its place, subject to transitional provisions that the approved text carried for matters arising from investments and obligations established under the earlier framework. The brevity of the interval between entry into force and comparison imposes discipline on every judgment about the successor agreement. Ten weeks of operation can establish the legal content of a text, the content thresholds a producer must meet in a given year, and the procedures a government may invoke under a chapter or an annex. Ten weeks of operation cannot establish a record of enforcement outcomes, a pattern of panel decisions, or a measured effect on trade flows attributable to the revised rules rather than to the duty free baseline the earlier agreement had already produced by the end of its transition in 2008. The comparison therefore separates questions that the texts answer on their face from questions that only administration over time could answer, and assigns each axis of comparison to the proper category before any verdict attaches to either agreement.

The five axes organize that separation. Market access asks what crosses the border at what price, comparing the tariff schedules and quota commitments of the two texts against the baseline each text inherited. Rules of origin ask which goods qualify for preferential treatment and what a producer must document to claim that treatment, with automotive content supplying the operative case in both generations. Labor enforcement asks what obligation a government accepted, which forum may hear a complaint about a failure to meet the obligation, and what remedy follows a finding against a government or against a facility. Investment asks what claim an investor may bring against a host government, in which forum, and under what limits each text placed on that claim. Duration and review ask how long each agreement was built to last, what schedule each text set for revisiting the bargain, and what notice a government must give to leave the bargain altogether. Each axis names a dimension along which the two instruments can be measured against one another without collapsing distinct questions into a single judgment about trade policy in general. A text can change little on one axis and change a great deal on another, and the record of the two statutes contains exactly that pattern of uneven change. Stating the instruments first, in the form in which Congress enacted them and on the dates on which the three governments bound themselves, keeps the measurement honest, because every later claim of difference must trace back to a chapter, an annex, a schedule, or a statute that a reader can locate in one of the two instruments described here.

The Five-Axis Comparison Table

The table states the comparison in compressed form, axis by axis, so that the pattern of uneven change is visible before the detailed sections argue each axis at length. Each row pairs the position under the earlier agreement with the position under the successor agreement and states the practical consequence that follows from the pairing.

Axis	Under the 1993 agreement	Under the successor agreement	Practical consequence
Market access	Tariffs on substantially all goods eliminated by the end of the transition, with the last staged categories reaching zero in 2008	Tariff schedules largely carried forward, with dairy and poultry access enlarged through tariff rate quotas and government procurement coverage narrowed	Border prices changed little, while access inside managed agricultural systems moved by stated quantities
Rules of origin	Regional value content of 62.5 percent for passenger automobiles, light trucks, engines and transmissions, and 60 percent for other vehicles and parts	Regional value content rising to 75 percent, joined by a labor value content wage test and a 70 percent North American steel and aluminum purchase requirement	Qualification became harder and heavier to document, while payment of the ordinary tariff remained the escape route
Labor enforcement	Labor content in a side agreement pledging each party to enforce domestic law, with a monetary assessment available only for a narrow band of standards and never imposed	Labor obligations in Chapter 23 of the main text, joined by a facility specific rapid response mechanism directed at named facilities in Mexico	Enforcement moved from consultation between governments toward consequences falling on the goods of a single facility
Investment	Chapter 11 gave investors direct claims against host governments before arbitral tribunals across all three relationships	Investor state claims eliminated between the United States and Canada, absent between Canada and Mexico, and narrowed between the United States and Mexico	General regulatory disputes moved toward domestic courts, with fuller protection retained only for covered government contracts
Duration and review	No term and no scheduled review, with withdrawal available on six months notice	A sixteen year term ending July 1, 2036 absent written confirmation from each party, a joint review on July 1, 2026, and withdrawal on six months notice retained	Continuation became a scheduled decision requiring affirmative confirmation rather than a default supplied by silence

Market Access: The Tariffs Barely Moved

Market access is the axis on which the comparison between the North American Free Trade Agreement and the United States-Mexico-Canada Agreement shows the least change, and the reason for that ranking sits in the calendar of the earlier agreement rather than in any modesty of ambition in the later negotiation. Under the earlier agreement, tariffs on substantially all goods traded among Canada, Mexico, and the United States had been eliminated by the end of the transition period, with the last staged categories reaching zero in 2008. A tariff schedule is a promise that unfolds across dated steps, and by the time the successor agreement entered into force on July 1, 2020, every one of those steps had long been taken. Nearly all North American goods trade was duty free by 2020, not because the successor agreement made that trade duty free, but because the earlier agreement had already done so and the intervening years had run under that settled position. The successor agreement therefore inherited a border at which the ordinary tariff, the price charged for crossing, had already been removed for the overwhelming bulk of commerce among the three parties. A negotiation that begins from zero on most lines has little left to concede on those lines, and the schedules of the successor agreement mostly restated the position the earlier agreement had produced. That restatement is the central fact of the market access axis, and every exception discussed in connection with the axis gains significance precisely because the general rule was continuity.

Those places shared a common character. The remaining restrictions did not take the form of a straightforward duty printed in a tariff line and applied to every incoming shipment without distinction. The remaining restrictions sat inside managed systems, where access depended on a quota quantity, on a pricing class, on a distinction between trade within a stated volume and trade beyond that volume, or on the coverage of a procurement chapter that named which parties owed which obligations to which suppliers. Dairy stood at the center of that category, with poultry and eggs beside dairy, and government procurement presented a different version of the same lesson, because procurement coverage under the successor agreement narrowed even as other chapters of the agreement expanded. The market access axis therefore contains two stories at once. Along the broad front of ordinary goods trade, the successor agreement changed almost nothing about the price of entry. Along the narrower fronts where access had been rationed rather than freed, the successor agreement negotiated quantities, classes, caps, and coverage, and those negotiations produced the only material movement on the axis.

If tariffs had already reached zero, where did the remaining barriers sit?

The remaining barriers sat behind the border and inside administered systems rather than in ordinary tariff lines. Canada retained managed access for dairy, poultry, and eggs through quotas and pricing classes, while government procurement coverage was defined party by party. Ordinary industrial and agricultural tariff lines offered little left to change between the two agreements in practice.

A tariff rate quota explains much of the dairy and poultry story, and the device repays a plain description. A tariff rate quota allows a stated quantity of a product to enter at a lower rate of duty, with trade beyond that quantity facing a higher rate. The border price therefore depends on volume as well as on the product itself. Control of the quantity is control of access, and an increase in the quantity is an increase in access even when neither rate in the schedule changes at all. That is the form in which the successor agreement delivered most of the market access movement on the axis. The successor agreement enlarged defined quantities for defined products from a defined exporting party, phased those quantities across a stated period, and left the underlying architecture of managed access in place around the enlarged quantities. A comparison that reads the quantities catches the movement and also catches the limit, because a quota enlarged remains a quota, and managed access enlarged remains managed access.

Canadian dairy supplied the principal market access change in the successor agreement, and the change had three distinct elements that a careful comparison keeps separate. The first element concerned domestic milk pricing classes. Canada agreed to eliminate the Class 6 and Class 7 milk pricing classes six months after entry into force. Entry into force fell on July 1, 2020, and six months after that date falls on January 1, 2021. By September 15, 2020, the elimination was a dated commitment in the agreement, not a completed act, and the comparison states the commitment in exactly that form. The second element concerned access into Canada for dairy products from the United States, delivered through tariff rate quotas. The third element concerned Canadian exports outward, in the form of caps on specified products with surcharges above the caps. Pricing, entry, and outward shipment are three different levers, and the dairy package pulled each of them. That combination explains why dairy, a single sector inside a continental agreement covering substantially all goods, carried a weight in the market access comparison that the size of the sector alone would not predict.

The Class 6 and Class 7 commitment addressed the way milk was priced inside Canada for particular uses, and the significance of the commitment lies in the connection between a domestic pricing class and the competitive position of imported product. A pricing class assigns milk to a category and attaches a price to milk in that category. When a category prices milk for a use at a level that imported ingredients cannot match on equal terms, the class itself functions as a barrier, even though no customs officer collects a duty at the border and no tariff line records the effect. Elimination of the two named classes therefore belonged on the market access axis even though the commitment operated on domestic classification rather than on a border charge. The commitment also carried a discipline of timing that the successor agreement made explicit. Canada undertook the elimination on a date fixed by reference to entry into force, six months after July 1, 2020, and that undertaking fell due on January 1, 2021. Stating the commitment with the date preserves the difference between a promise written into an agreement in force and an act already performed under that agreement. By the vantage of September 15, 2020, the classes stood under a sentence of elimination with a date attached, and nothing stronger than that statement fits the record at that vantage.

The tariff rate quotas for dairy formed the second element, and the Office of the United States Trade Representative published the schedules from which the shape of the concession can be read. Canada agreed to provide tariff rate quotas specific to the United States across dairy product lines, phased in over about six years. A country specific quota matters because access reserved for one exporting party cannot be absorbed by suppliers from other countries competing for the same quantity. The phase in matters because the full quantity arrives only at the end of the stated period, and trade in the early years of the successor agreement would run under smaller quantities than trade after the phase in completed. Summed across the product lines, the quotas amounted to roughly 3.5 to 3.6 percent of the Canadian dairy market, a figure drawn from the Office of the United States Trade Representative and university analyses of the quota schedules. That percentage states both sides of the dairy story in a single measure. Against a starting position of tightly managed access, a defined share of the market opened to the United States on stated terms represented a negotiated gain with commercial meaning for exporters able to ship within the quotas. Against the whole of the Canadian market, the same percentage shows that the greater part of that market remained behind the managed system, supplied under domestic arrangements rather than opened to import competition. Neither half of that statement cancels the other half, and the comparison holds both halves together.

The third element of the dairy package looked in the opposite direction, toward Canadian product moving outward into world markets. Canada accepted caps on exports of skim milk powder, milk protein concentrates, and infant formula, with export surcharges applying above the caps. The Office of the United States Trade Representative, in published material on the agreement, stated the first year aggregate cap for skim milk powder and milk protein concentrates at 55,000 metric tons, falling to 35,000 metric tons in the second year. A cap of that kind addresses a concern about displacement. Product priced inside a managed domestic system, if shipped outward without limit, can compete in third markets or in the market of a partner on terms that reflect the managed system rather than open competition, and a ceiling on the volume answers that concern by bounding the quantity that can move under those terms. The surcharge above the cap reinforces the ceiling with a price consequence for shipment beyond the stated quantity. Infant formula stood inside the same discipline, named alongside skim milk powder and milk protein concentrates as a product subject to a cap and a surcharge. The export element therefore completed a package that dealt with dairy as a system. Domestic classes that shaped the competitive position of imports were committed to elimination on a date, entry for the partner exporter was enlarged by quantity across about six years, and outward shipment of specified products was bounded by quantity with a charge beyond the bound.

Poultry and eggs extended the same pattern of managed access beyond dairy, on a smaller scale and with less accompanying change to domestic pricing. Canada provided tariff rate quotas for chicken from the United States, stated at 47,000 metric tons and rising to 57,000 metric tons by year six, and the successor agreement placed those quantities on top of the quota Canada already maintained under its World Trade Organization commitments. The structure of that concession deserves attention because the addition sat above an existing floor rather than replacing the floor. Access through the World Trade Organization quota continued, and the country specific quantity in the successor agreement added a further channel reserved for the United States. Egg access followed the quota pattern as well, with Canada providing tariff rate quota access for eggs and egg products from the United States inside the same managed framework that governed poultry. In both product groups, the governing fact is the persistence of the framework. The successor agreement negotiated larger quantities inside a system of rationed entry, and the comparison records enlarged access and continued management as simultaneous truths. A producer shipping within a quota gained a wider door. A producer measuring the system as a whole faced a door that remained a door, with a stated width, rather than an open border of the kind that duty free treatment on ordinary goods had created elsewhere in the agreement.

Government procurement moved the market access axis in the opposite direction, and that movement supplies a useful corrective against any assumption that the successor agreement widened access on every front. Procurement concerns the purchases governments make, and a procurement chapter states the purchases on which a party must treat suppliers from a partner party under agreed disciplines, including treatment no less favorable than the treatment given to domestic suppliers on covered purchases. Coverage is the operative question, because a procurement chapter binds only the parties between which the chapter runs and only for the entities and purchases the chapter covers. Under the successor agreement, the procurement chapter applies between the United States and Mexico. Canada stands outside that chapter, and procurement coverage between Canada and the other two parties runs through the World Trade Organization Agreement on Government Procurement instead. The result is a narrowing of coverage inside the North American instrument itself, with a separate multilateral instrument carrying the obligations for one of the three parties. On the market access axis, procurement therefore records a contraction in the reach of the agreement itself, set against continuity on tariffs and enlargement inside managed agricultural quotas.

The first reading treated the successor agreement, on market access, as a rebrand. Market participants and the trade press advanced that reading by measuring the agreement at the border, shipment by shipment, and asking what a trader would pay or encounter differently when goods crossed between the three countries after July 1, 2020. On that measure, the answer for the great bulk of trade was that nothing had changed in price, because the duty had stood at zero before the successor agreement and stood at zero under the successor agreement. A change of name, a change of instrument, and a renegotiation conducted at high political volume had produced, on the tariff lines that governed most commerce, the same number that the earlier agreement had produced by 2008. Supporters of the rebrand reading pointed to the dairy percentage as confirmation rather than refutation, since an opening summed at roughly 3.5 to 3.6 percent of one managed market in one country, phased across about six years, left the architecture of that market intact and left every other tariff line where the earlier agreement had left the line. Procurement narrowing strengthened the reading further, because an agreement presented as an expansion of North American integration had contracted the coverage of one chapter inside the instrument. Measured by what crosses the border at what price, the rebrand reading holds considerable force, and the comparison adopts that measure as the deciding test for the axis as a whole.

The second reading came from inside the sector that conceded the access, and the Dairy Farmers of Canada gave that reading its clearest voice. From the standpoint of Canadian dairy producers, the relevant measure was not the price on a single shipment or the share of the market described in a single agreement. The relevant measure was cumulative. Each successive agreement had opened a further defined share of the Canadian dairy market to imports, and the successor agreement added country specific quotas for the United States on top of access conceded in earlier instruments, inside a supply system in which producers had planned investment, production, and income around managed quantities. A share described by an exporting government as small could register, from inside the managed system, as a permanent transfer of a portion of the domestic market, because quota access once granted in an agreement does not return to domestic producers at the end of a season. The Dairy Farmers of Canada described the concession in those cumulative terms, treating the dairy provisions as a genuine loss of market rather than as a rounding adjustment inside a rebranded agreement. The phase in over about six years softened the timing of that transfer without changing the direction, and the elimination commitment for Class 6 and Class 7 added a change to domestic pricing that producers experienced as part of the same package. On that measure, the dairy provisions were the opposite of a rebrand for the sector that lived under them, and the comparison records the reading at that strength.

The distance between the two readings is a distance between measures, and naming the measures resolves much of the apparent contradiction. The rebrand reading measures the whole of North American goods trade and finds the dairy opening small against that whole, the tariff schedules unchanged, and procurement coverage reduced. The dairy sector reading measures one managed market from inside and finds the opening large against the portion of that market that producers regarded as theirs, enlarged by accumulation across agreements and paired with the end of two pricing classes on a fixed date. Both readings can be accurate at once because each reading answers a different question. The verdict for the axis follows the measure the comparison announced for the agreement as a whole. Judged by what crosses the border at what price, market access changed modestly, and dairy stands as a bounded exception inside a general continuity rather than as evidence of a general opening. Judged by the experience of the sector that administered the concession, the change carried weight that a continental average conceals. Questions of employment belong to a different measurement again, and the employment evidence assembled on the earlier agreement addresses a record that the price based comparison of the two schedules does not attempt to measure a second time.

One further discipline governs every statement on the axis, and that discipline is the vantage of September 15, 2020. The successor agreement had been in force since July 1, 2020, a span of about ten weeks by the middle of September. The dairy quotas had years of phase in ahead under the schedules, with the full quantities set to arrive across about six years rather than on the day of entry into force. The Class 6 and Class 7 elimination falls due on January 1, 2021, six months after entry into force, and enters this comparison as a dated commitment rather than a completed act. The chicken quota quantity of 47,000 metric tons described an early stage of a schedule rising to 57,000 metric tons by year six, not the final quantity in operation at the vantage. The tariffs barely moved because the movement on tariffs had happened under the earlier agreement and had finished in 2008. The movement that remained for the successor agreement sat in quantities, classes, caps, and coverage, and the successor agreement moved each of those by a stated amount, on a stated timetable, inside systems that the movement left standing.

Rules of Origin: From Where It Is Made to How Workers Are Paid

A rule of origin performs a single decisive task inside a preferential trade agreement. The agreement promises that goods traded among the parties enter at a preferential tariff, in most North American trade at a rate of zero, and the rule of origin decides which goods have earned that promise. A good produced entirely outside the region receives no preference, however it is routed, and a good produced inside the region receives no preference either unless it satisfies the conditions the text attaches to regional production. The rule therefore operates as a gate between two tariff treatments for the identical physical object. The same automobile, the same transmission, and the same component cross the same border at the preferential rate if the origin conditions are met and at the ordinary rate applied to goods from any other source if the origin conditions are not met. Everything on the rules of origin axis follows from that gatekeeping function, because a change to the conditions changes the practical meaning of the preference itself.

By September 15, 2020, the successor agreement had been in force since July 1, 2020, and the tariff schedules carried over from the earlier agreement had changed little, since duties on substantially all goods among the three countries had already been eliminated under the earlier agreement. Measured on the rules of origin axis, however, the successor text was the more restrictive instrument, and the restrictiveness was deliberate. A higher origin threshold does not raise a tariff in form, since the preferential rate itself remains zero, but it narrows the class of goods able to claim that rate without altering production. A producer whose supply chain satisfied the earlier threshold with room to spare faced a recalculation under the successor text, and a producer whose supply chain sat close to the earlier threshold faced a choice among reorganizing sourcing, reorganizing production, absorbing additional documentation, or surrendering the preference on some output. The comparison on this axis is therefore not a comparison of rates. It is a comparison of conditions, and the conditions moved in one direction across every element of the automotive rule, toward more regional content, more regionally purchased metal, and, in the labor value content requirement, a condition of a kind the earlier agreement had never imposed at all.

The earlier agreement measured automotive origin principally through regional value content, and it prescribed the net cost method for that measurement in the vehicle sector. The net cost method begins with the total cost of the good and subtracts the value of materials that did not originate in the region, so that the resulting percentage expresses the share of the vehicle’s cost attributable to regional production rather than to imported content. The method matters because it ties qualification to the economics of production instead of to a simple change in tariff classification. A vehicle assembled in the region from a kit of imported components could change classification on assembly and yet contribute little regional value, and the net cost calculation was designed to catch exactly that case. Conversely, a vehicle built in the region with deep regional sourcing accumulated qualifying value at each stage, since the cost of regional labor, regional parts, and regional overhead entered the calculation on the qualifying side. Administration followed the calculation. A claim for preference rested on a certification that the good met the stated content threshold, and customs authorities in the importing country retained the power to verify the claim against production and purchasing records.

Under the earlier agreement, the threshold for passenger automobiles and light trucks stood at 62.5 percent regional value content under the net cost method, and the same 62.5 percent figure applied to engines and transmissions, the components whose production anchored the vehicle supply chain. Other vehicles and parts faced a lower threshold of 60 percent. Those figures had governed North American automotive production since the earlier agreement entered into force on January 1, 1994, and a generation of sourcing decisions had been built around them. Plants were located, supplier contracts were written, and component designs were costed on the assumption that a vehicle meeting the 62.5 percent mark qualified for duty free movement across the three countries. A producer planning a model cycle of several years could design to the threshold with confidence that the figure would hold for the life of the model, and the resulting supply chain was calibrated to a known and durable test.

The successor agreement raised the regional value content requirement for passenger vehicles, light trucks, and core parts to 75 percent, and it staged the increase across four years so that the full figure did not apply on the first day of the agreement. The schedule published by the Office of the United States Trade Representative set the requirement at 66 percent in 2020, at 69 percent in 2021, at 72 percent in 2022, and at 75 percent from 2023. A producer received time to renegotiate supplier contracts, to shift component sourcing toward regional suppliers, or to relocate stages of production, but the schedule denied the option of waiting for the requirement to recede. The structure of the rule also reached below the finished vehicle. Under the schedules described by the Chicago Fed, principal parts were set to reach 70 percent regional value content and complementary parts 65 percent by 2023, so the tightening applied at the component level as well as at the level of the vehicle as a whole. A vehicle could therefore fail qualification through a part even where the finished vehicle calculation appeared comfortable, and compliance planning had to follow the rule down through tiers of suppliers rather than resting at the assembly plant.

Why does a wage test belong inside a rule about origin?

A wage test belongs inside the rule because the drafters treated pay as part of origin itself. A vehicle assembled inside the region from regional parts could fail qualification if too much of its value came from facilities paying less than sixteen dollars an hour on average.

The genuinely original instrument in the successor text was labor value content, a requirement with no counterpart in the earlier agreement. Labor value content provides that a stated percentage of a vehicle’s production value must come from facilities paying an average base hourly wage of at least 16 dollars. The test operates on value rather than on the physical composition of the vehicle. A part does not pass or fail the test because of where its materials were mined or where its subcomponents were fabricated. The test asks where the value embodied in the vehicle was created and what the workers who created that value were paid. Production value generated at a qualifying facility counts toward the required percentage, and production value generated at a facility paying less than the specified average does not count, even where the facility sits squarely inside the region and even where every material in the product originated in North America. The wage figure functions as a threshold in the same formal sense as the content percentage, a line with compliance on one side and disqualification on the other, but the line runs through payroll records rather than through bills of materials.

For passenger vehicles, the labor value content schedule set the required share at 30 percent in 2020, at 33 percent in 2021, at 36 percent in 2022, and at 40 percent from 2023. Light trucks followed a steeper path to a higher destination, moving through 30 percent, 35 percent, and 40 percent to reach 45 percent in 2023. The staging mirrored the staging of regional value content, and the parallel schedules compounded the planning problem for producers, since a sourcing plan that satisfied the content test in a given year also had to satisfy the wage test in the same year, and both tests tightened on the same calendar. The wage at the center of the test was defined as an average base hourly wage, which meant that the calculation turned on the composition of a facility’s workforce and pay structure rather than on the wage of any individual worker. A facility employing a mix of higher paid and lower paid production workers could qualify on the average, and a facility could move toward qualification by changing that mix, by raising base pay, or by shifting which stages of production occurred under its roof. The average also meant that verification required payroll evidence aggregated across a workforce, a form of proof foreign to the origin rules of the earlier agreement.

The text softened the wage test with credits for high wage expenditures outside direct production labor. A producer could earn credit toward the labor value content requirement through high wage technology expenditures and through high wage assembly expenditures, categories that recognized research, engineering, information technology, and capital intensive assembly operations as contributions to the high wage share of a vehicle’s value. The credits acknowledged a mechanical problem. A highly automated plant might generate substantial production value with a small production workforce, and a test confined to production wages alone could penalize exactly the investment in technology that raised productivity and supported higher pay for the workers employed. By crediting qualifying technology and assembly spending, the drafters allowed a producer to reach the required percentage through a combination of direct high wage production and high wage supporting activity. The credits also widened the range of compliance strategies. A producer short of the threshold could raise wages, relocate production, expand high wage technical operations attached to vehicle programs, or combine those responses, and each response carried a different cost and a different consequence for the organization of production.

The deeper significance of labor value content lies in the conversion it performs. A traditional rule of origin is a rule about geography. It asks where a good was made, where its materials came from, and how much value the region added, and it is indifferent to the conditions under which regional production occurred. Two vehicles with identical regional content receive identical treatment under such a rule even if one was built by workers paid well above prevailing wages and the other by workers paid at the lowest wage the regional market offered. Labor value content breaks that indifference. Under the successor text, geography remained necessary but ceased to be sufficient, since regional production counted toward the wage share of the test only when the production occurred at facilities meeting the pay threshold. A rule drafted in the vocabulary of origin thus acquired the function of a wage floor for a defined share of production. The conversion was partial by design, since the test governed the share of value that had to be produced at qualifying pay rather than the pay of every worker in the supply chain, and production below the wage threshold remained lawful, marketable, and eligible to count toward regional value content even where it could not count toward labor value content. The distinction preserved a space for lower wage production inside the region while withdrawing part of the preferential reward that such production had received under the earlier agreement.

A second addition tied origin to the metals supply chain. The successor agreement required that 70 percent of an automaker’s steel and aluminum purchases, measured by value, originate in North America, a purchase based test applied at the level of the producer rather than at the level of the individual vehicle. The shift in unit of analysis mattered for administration. A vehicle level test is answered from the records of a single product, while a purchase level test is answered from the procurement records of the firm across a period, and a producer had to demonstrate that the aggregate flow of metal into its operations met the regional share. The aluminum requirement applied the same regional logic as the steel requirement, treating the two metals as twin pillars of vehicle construction whose sourcing the agreement sought to anchor in the region. Layered on the purchase test was a further condition about how steel was made. The requirement that steel be melted and poured in North America addressed the distinction between steel finished in the region from imported substrate and steel whose production from molten metal occurred in the region, and the agreement placed that requirement on a seven year runway, so the full condition was a matter of enacted schedule rather than a test applied in full from the first day. By September 15, 2020, the purchase share obligation stood in force while the melted and poured condition moved toward application along the stated runway, and producers planning furnace and mill relationships had to plan against both the obligation in force and the condition scheduled to follow.

Compliance under the successor text demanded more than recalculation. Three new certifications accompanied the regional value content certification, one for labor value content, one for steel, and one for aluminum, so each claim for preferential treatment on a vehicle rested on a stack of attestations rather than on a single calculation. Each certification carried its own evidentiary trail. The labor value content certification depended on wage and payroll records capable of showing an average base hourly wage at each facility whose production was counted, organized in a form a customs authority could audit. The steel and aluminum certifications depended on purchase records capable of showing the origin of metal by value across the producer’s operations. The regional value content certification depended on the costed bills of materials and supplier information that the earlier system had already required, deepened by the higher thresholds and the tiered treatment of parts. Suppliers faced corresponding demands, since an assembler’s certification was only as strong as the declarations and records supplied by the firms whose parts and materials entered the calculation. Documentation therefore traveled up the supply chain as a condition of trade, and firms that had treated origin as a periodic accounting exercise faced a heavier and more continuous administrative discipline under the successor text.

Against that burden stood a permanent feature of every origin regime, the option to decline the preference. A producer unable or unwilling to certify compliance could ship the good under the ordinary most favored nation tariff, the nondiscriminatory rate applied to goods from any source, and pay the duty instead. The option places a ceiling on how binding any origin rule can be. A rule binds a producer only up to the point where the cost of compliance, counting reorganization, higher cost sourcing, wage adjustments, and administration, exceeds the duty saved by claiming the preference. Beyond that point the rational producer pays the tariff, accepts treatment no better than a producer outside the region, and treats the agreement as irrelevant to that line of production. The ceiling operates unevenly across products, since the value of the preference depends on the ordinary tariff attached to the good in question, and a rule that binds tightly where the ordinary tariff is substantial may bind hardly at all where the ordinary tariff is low. The successor text tightened the conditions of preference considerably while leaving the escape route untouched, and any assessment of the tightening on this axis has to hold both facts together. The gate became harder to pass, and the option of walking around the gate by paying the ordinary rate remained exactly where the earlier agreement had left it.

The readings of the tightened rules divided along predictable lines, and each reading deserves statement at full strength. Automakers warned that the combined effect of higher content thresholds, the wage test, and the metals requirements would raise the cost of building vehicles in North America and complicate supply chains that had been optimized over decades around the earlier thresholds. The United States Chamber of Commerce pressed the compliance side of the same case, arguing that layered certifications and record keeping obligations imposed administrative costs that fell on firms regardless of whether the underlying production changed, and that complexity itself operated as a tax on regional trade. On that reading, some production would be pushed toward paying the ordinary tariff, some sourcing decisions would be distorted toward compliance rather than efficiency, and regional competitiveness against vehicles built outside North America would suffer.

Labor organizations read the same provisions from the opposite direction and treated the wage linkage as the point of the exercise. The AFL-CIO and allied labor organizations had long argued that the earlier agreement rewarded the movement of production to the lowest wage location inside the region, since origin rules counted low wage regional production on equal terms with high wage regional production, and that the equal treatment built a downward pressure on wages into the structure of North American trade. On that reading, labor value content corrected a defect rather than adding a burden, because a preference financed by the three governments through forgone tariff revenue ought not to subsidize a race toward the cheapest labor available within the region. The 16 dollar average wage threshold, the rising share of value required to meet it, and the credits for high wage technology and assembly spending all served, in the labor reading, to tilt the incentive structure toward production models that sustained higher pay. Supporters in the labor movement also treated the staging as evidence of seriousness, since the schedule did not stop at a symbolic first step but rose to 40 percent for passenger vehicles and 45 percent for light trucks by 2023, levels high enough to change behavior at firms intending to claim the preference across a full model range.

By September 15, 2020, roughly ten weeks after entry into force, only the first step of each schedule had taken hold. The 66 percent content figure governed claims in 2020, the first labor value content percentages applied, and the 70 percent steel and aluminum purchase share stood in force, while the steps to 69 percent and beyond, the final content levels of 2023, and the melted and poured condition on its seven year runway remained matters of enacted schedule. The comparison on the rules of origin axis therefore describes a trajectory as much as a state of affairs. The earlier agreement had asked where a vehicle was made and had set that test at 62.5 percent for vehicles, engines, and transmissions and at 60 percent for other vehicles and parts. The successor agreement asked the same geographic question at 75 percent, asked a further question about the origin of the metal a producer purchased, and asked a question the earlier text had never posed, how the workers who produced a stated share of the vehicle’s value were paid. Whether the tightened gate would redirect production, raise wages, raise costs, or push trade toward the ordinary tariff in significant volume depended on choices producers had only begun to make under the first step of the schedules and on the balance between compliance cost and preference value as the schedules rose toward 2023.

Labor Enforcement: From Consultation to the Facility Door

On the labor enforcement axis, the distance between the two agreements is wider than on any other axis of comparison, and the distance is structural rather than rhetorical. The earlier agreement placed its labor content outside the agreement itself, in a side agreement negotiated after the main text had been completed to secure votes for passage. The successor agreement placed its labor content inside the main text, attached consequences to defined obligations, and built a procedure that ends at the loading dock of a single named facility rather than at a communique. The comparison between the two agreements has been summarized in a single line, a rebrand on tariffs and a rewrite on enforcement, and the labor enforcement axis is where the second half of that line earns its wording. Labor enforcement changed in kind, because the question the procedure asks, the party who bears the cost of an adverse finding, and the speed at which a finding can arrive all changed at once.

What could the earlier system compel a government to do?

The earlier system could compel discussion, review, and public reporting, but it could compel payment only in a narrow band of cases involving workplace safety, child labor, or minimum wage enforcement, and that payment stage was never reached in practice during twenty six years of operation under the side agreement.

The earlier design began from a deliberate restraint. Under the North American Agreement on Labor Cooperation, the side agreement that carried the labor content of the earlier bargain, each party pledged to enforce its own labor law. No common labor code was imposed across the three countries, and no party accepted an obligation to raise its standards to a shared floor set by the agreement. The commitment ran to enforcement of domestic law as each party had written that law, and the agreement organized its subject matter as eleven labor principles arranged in three tiers. The tier into which a principle fell determined how far a complaint about that principle could travel, and the available distance differed sharply by tier. The architecture therefore answered two questions before any facts were examined. The first question was whether a government had enforced the law on its own books. The second question was whether the principle at stake had been assigned a procedural route long enough to reach a consequence, and for most principles the assigned route ended well short of one.

The ladder itself had four recognizable rungs, and a matter could stop permanently at any of them. The first rung sat in the National Administrative Office that each party maintained for the purpose. That office received submissions raising questions about labor law enforcement in another party, and it could hold reviews and hearings on the questions presented. The product of that rung was examination and publicity, a record assembled in the open, rather than an order directed at the government whose enforcement had been questioned. The second rung was consultation between labor ministers, where the governments themselves took up the matter as a subject of discussion between sovereigns. Consultation could produce agreement to cooperate, to study, or to report, and it could also produce nothing beyond the exchange itself, because the rung carried no power to decide the dispute against a government that declined to move. The third rung, an Evaluation Committee of Experts, stood beyond consultation but behind a gate. A committee of that kind could be convened only for matters in the second and third tiers of principles, and its function was analysis of patterns of enforcement practice, delivered as expert assessment rather than as judgment. The fourth rung, an arbitral panel, stood at the top of the ladder and behind the narrowest gate in the entire design. A panel, the only stage in the process that could produce a monetary enforcement assessment, was available only for a persistent pattern of failure to enforce a defined subset of standards, occupational safety and health, child labor, and minimum wage requirements. A single lapse could not reach that stage, and a principle outside the named subset could not reach that stage at any level of persistence.

Even at the top of the ladder, the financial consequence had a ceiling fixed in advance. Annex 39 of the side agreement capped a monetary enforcement assessment at 20 million dollars in the first year, and thereafter at no more than 0.0075 percent of total trade in goods between the parties concerned. If an assessment went unpaid, the complaining party could suspend benefits in that amount, which converted the remedy into a withdrawal of trade advantage equal to the sum assessed. The cap mattered to the design in two directions at once. A government weighing defiance could calculate the maximum exposure before the process began, and a government weighing compliance could compare the cost of changing enforcement practice against a bounded figure unrelated to the scale of the failure alleged. The Congressional Research Service, in its analyses of the side agreement, and the Department of Labor, in its presentation of the agreement text, describe a mechanism whose stages were used for review and consultation across the life of the earlier agreement, and whose final stage remained theoretical from beginning to end. In twenty six years of operation, from entry into force on January 1, 1994 to termination on July 1, 2020, no monetary enforcement assessment was ever imposed under the side agreement, and no case reached the panel stage at all. The record of the earlier design is therefore a record of process without penalty, of findings that could embarrass and consultations that could exhort, attached to a sanction that existed on paper for the whole period and in practice for none of it.

The deepest feature of the earlier design was not the length of the ladder but the placement of the rights that would later sit at the center of the successor mechanism. Freedom of association and the right to bargain collectively, the rights that concern whether workers may organize a union of their own choosing and negotiate through it, sat in the first tier of the eleven principles. First tier placement carried a precise procedural meaning. A matter involving those rights was eligible for review by a National Administrative Office and for consultation between ministers, and it was eligible for nothing beyond those two stages. The expert committee track was closed to it by the tier gate, and the arbitral panel track was closed to it twice over, once by the tier gate and once by the subject matter limit that confined panels to occupational safety and health, child labor, and minimum wage enforcement. A government could therefore face sustained questioning about the suppression of organizing activity inside its territory, could be asked to consult about that questioning at ministerial level, and could then decline every proposed change while remaining entirely inside its obligations under the side agreement, because the procedure assigned to the rights in question had no stage at which refusal acquired a price. The rights most directly connected to bargaining power inside a workplace were the rights the earlier procedure could pursue the least distance, while the rights it could pursue to a monetary assessment were technical standards, important to the workers they protected, that did not determine who held power in the employment relationship. Enforcement reach and organizing rights moved in opposite directions under the side agreement, and that inversion, more than the absence of any imposed assessment, is the baseline against which the labor enforcement axis measures change.

The successor agreement answered the placement problem by relocation before it answered any question of remedy. Labor content moved from a side agreement into the main text of the agreement itself, where Chapter 23 carries labor obligations as obligations of the agreement, subject to the standing of the agreement as a whole rather than to the lesser standing of an accompanying instrument negotiated to close a domestic vote count. The chapter obliges each party to adopt and maintain, in its law and practice, the rights stated in the 1998 Declaration of the International Labour Organization, the declaration in which that organization set out the fundamental rights its members accept by virtue of membership, and freedom of association and the effective recognition of the right to bargain collectively stand among those declared rights. Relocation changed the legal character of the commitment in a way that tier placement never could. A failure to maintain the declared rights became a failure under the trade agreement itself, arguable through the agreement’s own state to state dispute settlement in Chapter 31, instead of a subject confined to a consultation process housed in a separate instrument. The chapter also reached past general obligation into the domestic law of one party by name. Annex 23-A, addressed to Mexico, required the enactment of collective bargaining and union democracy reforms, specifying changes to the framework under which unions gain recognition, contracts gain legitimacy, and workers vote on the agreements that govern them. Mexico enacted its labor law reform on May 1, 2019, before the successor agreement entered into force, so the domestic precondition was performed in advance of the obligations it supports rather than promised for performance afterward. The sequence carries weight on the labor enforcement axis. Under the earlier design, domestic law was the fixed point and the agreement pledged only that each party would enforce whatever law it had. Under the successor design, one party’s domestic law was itself an object of the bargain, amended on a schedule tied to the agreement, so that the enforcement machinery of the Protocol of Amendment would operate on a legal framework the agreement had already reshaped.

The hinge between the signed text and the enforceable design was the Protocol of Amendment, signed on December 10, 2019, a little more than a year after the agreement itself had been signed at Buenos Aires on November 30, 2018. The Protocol reworked enforcement on two planes at once, the burden a complaining party carries in an ordinary labor case under the chapter, and the creation of a separate mechanism that bypasses the ordinary case structure altogether. On burden, the signed text had framed key labor obligations so that a complaining party carried the requirement of showing that a failure to adopt or maintain the protected rights had occurred in a manner affecting trade and investment between the parties, a framing that placed proof of trade effect at the center of every case and invited dispute about markets before any dispute about rights could be decided. The Protocol eased that burden for the key obligations, establishing a presumption that a failure affects trade unless the responding party shows otherwise. The shift reallocates the work of the case. Argument begins from effect presumed and moves to rebuttal, instead of beginning from effect unproven and moving through market evidence toward the workplace conduct that prompted the complaint. On the second plane, the Protocol created the facility specific rapid response mechanism, the element of the labor enforcement axis with no ancestor anywhere in the earlier agreement, and the element that gives this comparison its sharpest contrast in kind.

The mechanism, created by Annex 31-A for relations between the United States and Mexico and by Annex 31-B for relations between Canada and Mexico, runs in both forms against covered facilities in Mexico in priority sectors, and its unit of action is the individual facility rather than the national enforcement system. A party that believes workers at a named facility in Mexico are being denied the rights of free association and collective bargaining may ask the host government to review the situation at that facility. The request identifies a place, an employer, and a workforce, which distinguishes the procedure from every stage of the side agreement ladder, where the subject was always a pattern of enforcement across a national system and never the conduct surrounding one organizing effort at one address. If the matter is not resolved through that review, an independent rapid response labor panel may determine whether a denial of rights exists at the facility. The panel’s question is factual and local, whether the named workers hold the named rights in practice, and its determination carries remedies that fall on the goods of the facility concerned. Remedies include suspension of preferential tariff treatment for those goods and penalties, and on repetition the ladder of consequences climbs from the facility itself to sister facilities of the same entity, and then to denial of entry for the goods of the entity. Escalation therefore follows corporate structure rather than national territory, widening from a plant to its siblings and then to the border itself, while the government whose territory houses the facility is a party to the process rather than the object of the remedy. Timelines run in weeks and months, counted against deadlines built for a dispute whose subject, an organizing campaign or a contested contract vote, can be won or lost inside a single season, and not in the years the side agreement ladder consumed in consultations whose conclusion no stage compelled. Read mechanically, the mechanism is a sequence of named facility, governmental review, independent determination, and remedy on goods, with each step defined in advance and each consequence priced in trade terms from the first step onward.

Against that architecture, the early record at the date of comparison must be stated expressly and without decoration, because decoration in either direction would misstate the evidence available on September 15, 2020. The successor agreement entered into force on July 1, 2020, and had been in force about ten weeks by September 15, 2020. In that interval the rapid response mechanism had never been invoked. No petition had been filed under it, no panel had convened under it, and no remedy had been imposed under it, by September 15, 2020. The mechanism existed as text, as annex, and as procedure awaiting a case, and no case had been presented under it. The labor side event closest to entry into force was preparatory rather than operational. On June 30, 2020, one day before entry into force, the United States published interim procedural guidelines for petitions through its Interagency Labor Committee, setting out how a petition could be presented for consideration. The guidelines marked an administrative doorway built in advance, and supplied no evidence about what would pass through it. Any assessment of the mechanism on September 15, 2020 is therefore an assessment of design alone, made on the documents both sets of parties signed, because operation had not begun. The twenty six year record of consultation without assessment on one side of the comparison faces ten weeks of untested architecture on the other.

The meaning of that design divided organized labor and organized business along lines that each side stated at full strength during the ratification debate, and the division survives statement without endorsement in either direction. The AFL-CIO, the federation whose support for the successor agreement marked a break from the position labor organizations had held toward the earlier agreement, argued that facility level remedies change the incentives of the employer at the point where organizing succeeds or fails. Under the side agreement, on that reading, an employer that resisted a union faced a cost structure in which resistance was cheap, because the available international process ran against governments rather than against firms, moved through consultation measured in years, and ended, in the whole twenty six year record, without a single monetary assessment. A plant level decision to dismiss organizers, to sponsor a compliant union, or to hold a contract vote under conditions management controlled could therefore be taken with confidence that no trade consequence would reach the plant or the goods the plant shipped. The rapid response mechanism, on the AFL-CIO reading, reprices that decision. A denial of rights at one facility exposes the goods of that facility to loss of preferential treatment and penalties, exposes sister facilities on repetition, and exposes the entity to denial of entry at the end of the escalation, so the expected cost of suppression enters the calculation of the firm that controls the workplace rather than dissipating into a diplomatic exchange between capitals. Speed compounds the repricing, because a remedy that can arrive in weeks or months can arrive while the organizing effort it concerns remains alive, restoring a vote or a negotiation to conditions closer to the moment of denial, whereas a consultation concluded years later addresses a workforce that has dispersed and a contract that has hardened into place. On that reading, the mechanism deters less by the remedies it imposes than by the remedies a rational employer declines to risk, and its facility door focus is the entire point, since bargaining power is exercised or extinguished at the facility door and nowhere else.

The United States Chamber of Commerce, speaking for business organizations that had supported the earlier agreement and that assessed the successor agreement with greater reserve on enforcement grounds, stated the opposing reading at equal strength, centered on compliance exposure and on the character of the complaint process itself. On that reading, a mechanism that attaches trade consequences to a single facility converts every contested organizing campaign at a covered plant in Mexico into a potential trade action, and the exposure falls not only on the Mexican operator of the plant but on the supply chains built around its output, including purchasers in the United States and Canada that planned production around preferential treatment for goods that a panel determination could strip of preference, penalize, or exclude. Exposure of that kind, the Chamber argument runs, is difficult to price in advance, because the trigger is a belief about denial of rights presented by a party to the agreement, the determination is entrusted to a panel convened for the occasion rather than to the courts of the country whose labor law governs the plant, and the timeline that labor organizations praise for speed compresses the time available to assemble a factual record about workplace events that are themselves contested. The second strand of the reading concerns motive and selection. A complaint channel that begins with a petition invites complaints, and complaints can be filed for reasons mixed with competitive interest, by rivals that gain when a facility loses preference, by parties to a local bargaining dispute seeking leverage beyond the merits, or by organizations whose objective is the disruption of a supply relationship rather than the restoration of a right. On that reading, the risk is a mechanism used as a competitive instrument, in which the threat of a petition becomes a bargaining asset detached from conditions at the named facility, and in which firms respond by adding legal review, documentation, and caution to every labor decision at covered facilities, costs that fall on compliant and noncompliant operators alike. Business organizations holding that view did not dispute that the side agreement record contained no assessments. The dispute concerned whether zero penalties proved a system too weak to matter or a system appropriately confined to consultation between sovereigns, and whether the Protocol exchanged a known and bounded weakness for an unbounded and untested exposure.

The two readings agree on the facts of design and divide on the weight those facts should carry, and by September 15, 2020 nothing in the operation of the mechanism had arbitrated between them. Both readings describe the same relocation into Chapter 23, the same Mexican reform of May 1, 2019 as domestic precondition, the same eased burden and presumption under the Protocol of Amendment, and the same escalation from facility to sister facilities to denial of entry. The AFL-CIO reading treats consequence at the facility as the missing condition that makes declared rights real inside a workplace. The Chamber reading treats consequence at the facility as a trade remedy attached to the most fact dependent and contestable events in industrial relations, with exposure radiating outward to firms and purchasers far from the original dispute. On the labor enforcement axis, the comparison between the two agreements therefore ends in a judgment about architecture that the evidence available by September 15, 2020 fully supports, joined to a judgment about performance that the same evidence cannot supply in either direction. The earlier system ran for twenty six years and never imposed its sanction. The successor system had run for about ten weeks and had never been invoked. Between a sanction that existed and was never used and a mechanism that exists without a single invocation on record, the difference in design is categorical and documented, and the difference in result remained, at the date of comparison, entirely unmeasured.

Investment: The Most Litigated Feature, Reversed

Investment was the axis on which the distance between the earlier agreement and the successor agreement was widest, and the reason lay in what the earlier agreement had created. Chapter 11 of the North American Free Trade Agreement gave an investor of one party a direct right of action against the government of another party in which the investor had placed capital. An American company with a plant or a concession in Mexico or Canada, a Canadian company with an investment in the United States or Mexico, and a Mexican company with an investment in the United States or Canada could each bring a claim in its own name, for its own loss, before an international arbitral tribunal. The claim did not depend on the investor persuading its home government to take up the grievance, and it did not require the investor to sue in the courts of the host government whose measure was under challenge. That structure departed from the older pattern of diplomatic protection, under which an injured investor relied on its own government to decide whether a dispute deserved to be pressed at the state to state level, and from the ordinary pattern of domestic litigation, under which a foreign investor challenged a measure in the host government courts on the same footing as a domestic litigant.

The substantive protections were written in broad terms. Chapter 11 required each party to accord investors of another party treatment no less favorable than the treatment accorded to its own investors and treatment no less favorable than the treatment accorded to investors of any other country, and it protected covered investments against expropriation and against measures tantamount to expropriation without compensation. It also carried a minimum standard of treatment drawn from international law. The breadth of that language meant that a wide range of government action could be recast as a treaty claim. A denial of a permit, a change in a regulatory condition, a phaseout of a substance, and a court judgment that altered the value of an investment could each become the basis for an allegation that the host government had discriminated, had failed to meet the minimum standard, or had taken value without payment. The tribunal sat outside the domestic judicial hierarchy of any of the three countries, its award was not subject to appeal on the merits through the ordinary courts of the respondent, and the investor was not required to exhaust domestic remedies before filing. Those features explain why investment became the most litigated feature of the earlier agreement across its twenty six years in force, with dozens of claims filed and a body of arbitral reasoning that had no counterpart in the tariff chapters, where change had been scheduled in advance and had largely run its course by 2008.

The litigation record reached well beyond ordinary commercial disputes. Claims under Chapter 11 challenged environmental measures, including measures that restricted activities on environmental grounds and measures that conditioned industrial operations on environmental assessment. Claims challenged public health measures, including measures directed at substances and products that governments regulated on health grounds. Claims also challenged judicial measures, in which the conduct under review was not a statute or an executive decision but the handling and outcome of proceedings in domestic courts. That reach was central to the controversy. A tariff commitment either held or it did not. An investment claim, by contrast, could ask a tribunal to evaluate the purpose, the process, and the effect of a regulatory program that a legislature or an agency had adopted for reasons unrelated to trade in goods. The prospect that a government might have to pay compensation for a measure that remained in force added to the political weight of the mechanism, even though the record for this comparison is kept qualitative and no claim count or award total settles the argument.

The sovereignty objection was stated most directly by the governments that had defended claims and by the organizations that campaigned against the mechanism during the renegotiation. The government of Canada and the government of Mexico, in their official statements on the modernization of the agreement, and labor and environmental organizations, including the AFL-CIO in its trade policy positions, argued that investor state dispute settlement placed a private arbitral process above domestic democratic and judicial processes. In that account, a foreign investor held a remedy that a domestic investor did not hold, because a domestic investor challenging the same measure had access to domestic courts and to protections under domestic law but had no access to an international tribunal applying treaty standards. The objection extended to regulatory autonomy. Governments and organizations pressing the objection argued that the cost and risk of arbitration could deter a government from adopting or maintaining an environmental or public health measure, even when the measure applied without discrimination, because officials had to weigh the public purpose against the exposure created by a treaty claim. For those who held that position, claims reaching environmental, public health, and judicial measures were the predictable result of combining broad substantive language with a private right of enforcement.

The defense was stated with equal clarity by the business organizations that had supported Chapter 11 through the debate over its successor. The United States Chamber of Commerce and other business organizations representing companies with cross border investments argued that direct investor protection was a necessary discipline where an investor commits capital that cannot be moved once a plant is built or a concession is developed. In that account, the value of the mechanism lay in its independence from the host government courts and from the diplomatic calculations of the investor home government. A government that wished to attract long term capital, the business organizations argued, gained credibility by accepting in advance that a foreign investor could obtain a neutral hearing if property was taken or if treatment turned discriminatory after the investment had been made. On regulation, those organizations argued that Chapter 11 did not prohibit regulation and that tribunals distinguished between a compensable taking and a nondiscriminatory measure adopted for a public purpose, so that the proper remedy for an excess in a particular award was tighter drafting rather than abolition. That defense carried particular weight for sectors in which investments are large, fixed, and dependent on government contracts over long periods, a point that later shaped the narrower protection preserved between the United States and Mexico.

What did the successor agreement do to investor claims?

The successor agreement eliminated investor state claims between the United States and Canada, left Canada and Mexico to another agreement, and narrowed United States and Mexico claims to direct expropriation and nondiscrimination after domestic litigation, with fuller protection only for covered government contracts in specified sectors.

Chapter 14 of the successor agreement did not produce a single investment rule for North America. It produced three different answers for the three bilateral relationships, and the differences were deliberate. Between the United States and Canada, investor state dispute settlement was eliminated. An American investor in Canada and a Canadian investor in the United States held no direct claim under the successor agreement against the host government before an arbitral tribunal, and the text provided no replacement annex, no narrowed list of claims, and no sector specific carve out for that pair. Between Canada and Mexico, investor state dispute settlement was absent under this agreement. The text did not carry a bilateral investor state mechanism for that pair, and investors in that relationship relied on another agreement between Canada and Mexico rather than on Chapter 14. Between the United States and Mexico, the mechanism survived in a narrowed form, through two annexes that divided claimants by the character of the claim and by the character of the investment.

Annex 14-D governed the general run of claims between the United States and Mexico, and its confinement was substantial when measured against Chapter 11. Most claims under Annex 14-D were confined to direct expropriation and to breaches of national treatment and most favored nation treatment. Indirect expropriation claims and claims founded on the broader minimum standard of treatment, which had supplied much of the contested terrain under the earlier agreement, were placed outside the general route. The procedural confinement matched the substantive confinement. A claimant under Annex 14-D was required to pursue domestic litigation for thirty months before an arbitral claim could be brought. That requirement reversed the logic of Chapter 11. Under the earlier agreement, the investor could bypass domestic courts and proceed directly to arbitration. Under Annex 14-D, domestic courts received the dispute first, held it for a defined period measured in months, and only after that period had run could the investor move the surviving categories of claim to arbitration. The design gave the host government an opportunity to correct a measure through its own judicial process, gave domestic courts a record building role that Chapter 11 had denied them, and filtered claims that could be resolved or abandoned during the domestic phase. It also changed the timing and the cost of enforcement, because a remedy that had once been immediate in procedural terms became dependent on a preliminary period of litigation in the courts of the respondent government.

Annex 14-E preserved a fuller measure of protection for a defined class of investments between the United States and Mexico. The annex applied to covered government contracts in specified sectors, identified in the record for this comparison as oil and natural gas, power generation, telecommunications, transportation services, and certain infrastructure, together with one further listed category. For investments held under such contracts, the protections available were broader than the confined list in Annex 14-D, reflecting the argument, pressed by business organizations during the negotiation, that an investor who enters a long term contract with a government in a capital intensive sector faces a distinct risk that the contracting government will use sovereign authority to alter the bargain after capital has been committed. Annex 14-E answered that concern without restoring Chapter 11 in general form. The fuller protection attached to the contract and to the listed sectors, not to every investment in the Mexican or American economy, and an investor outside the covered contracts remained subject to the confinement and to the domestic litigation requirement of Annex 14-D. The line between the two annexes was a line between general regulatory disputes, which the successor agreement pushed toward domestic courts, and contract based disputes in sectors where the government acts as counterparty as well as regulator, which retained a stronger arbitral protection.

Superimposed on the three bilateral answers was a transitional rule for investments established under the earlier agreement. Annex 14-C permitted legacy claims arising from investments established while the North American Free Trade Agreement was in force, under the conditions that annex prescribed, for three years after the termination of the earlier agreement. The earlier agreement terminated on July 1, 2020, when the successor agreement entered into force, and the legacy window ran as a matter of enacted schedule to July 1, 2023. The legacy provision prevented an abrupt extinguishment of rights that had accrued around investments made in reliance on Chapter 11, while also preventing those rights from continuing without limit under an agreement that had ceased to exist. An investor whose investment predated the termination held a time limited route that an investor who entered after July 1, 2020 did not hold. After the legacy period closed under the schedule, the Chapter 11 basis for a claim closed with it, and the bilateral answers in Chapter 14 governed without the transitional overlay.

The practical consequence can be stated by identifying which claimants lost a remedy, which kept a remedy in altered form, and which forums absorbed the disputes that Chapter 11 had drawn into arbitration. Canadian investors in the United States and American investors in Canada lost the direct arbitral remedy altogether. A grievance that either investor might once have framed as a Chapter 11 claim had, after July 1, 2020, to be pursued in the domestic courts of the host country, under domestic constitutional and administrative law, or to be raised through diplomatic channels in the hope that the home government would take up the matter at the state to state level under Chapter 31 of the successor agreement.

Mexican and American investors kept a remedy, but a remedy divided and delayed by design. A claimant with a direct expropriation claim or a claim of discriminatory treatment between the United States and Mexico retained access to arbitration under Annex 14-D, subject to the thirty month domestic litigation requirement that stood between the measure and the tribunal. A claimant whose complaint rested on an indirect taking theory or on the broader minimum standard of treatment found that the general route no longer carried those theories. A claimant holding a covered government contract in a specified sector under Annex 14-E retained fuller protection, because the successor agreement treated the contract based investor as the case in which the argument for an external remedy retained its force. The consequence was a sorting of disputes by subject matter. Claims about the fairness of general regulation moved toward domestic courts, claims about direct takings and discrimination survived in narrowed arbitral form after a domestic phase, and claims rooted in government contracts in the listed sectors retained the closest analogue to the earlier protection.

For Canada and Mexico, the consequence was relocation rather than simple loss or retention. Because investor state dispute settlement was absent between that pair under this agreement, a dispute that might once have been brought under Chapter 11 had to be located in another agreement between Canada and Mexico if a treaty remedy was to be found at all, or pursued in domestic courts, or pressed through state to state channels. Canada and the United States accepted domestic courts for each other investors. Canada and Mexico looked outside the agreement. The United States and Mexico retained arbitration in confined and contract specific forms. Uniformity, a defining characteristic of Chapter 11, was the characteristic the successor agreement abandoned most completely on this axis.

Disputes that no longer reached an arbitral tribunal under Chapter 14 did not disappear. They routed elsewhere, and the routing changed the character of the dispute. A claim taken to domestic courts became a case about domestic law, argued before judges of the host country, subject to domestic appellate correction, and limited to the remedies domestic law provides. A grievance taken up by a home government became a state to state matter under Chapter 31, in which governments weighed trade and diplomatic interests alongside the investor loss, controlled the framing of the complaint, and decided whether settlement served a broader purpose. The removal of those filters had been, depending on attribution, either the achievement of the earlier agreement, as the business organizations that defended the mechanism described it, or its central defect, as the governments and organizations that opposed the mechanism described it. The successor agreement restored the filters for two of the three relationships in full or in relocated form, and restored the domestic court filter in part for the third, while preserving an arbitral exit for the categories the negotiators chose to protect.

Measured on the investment axis alone, the comparison supports the larger judgment without expanding it. On tariffs, the successor agreement largely carried forward a duty free structure that the earlier agreement had completed more than a decade before the successor entered into force. On investment, the successor agreement withdrew a private right of action that had operated for twenty six years, replaced it with no right for one pair, a right located in another instrument for a second pair, and a narrowed and delayed right for the third pair, subject to a legacy schedule that ran to July 1, 2023. The subject matter that had generated the most contested litigation, including claims that tested environmental measures, public health measures, and the acts of domestic courts, was redirected toward the institutions whose authority the sovereignty objection had invoked, namely domestic courts and governments acting in their own names. The sectors in which governments act as contracting parties retained a measure of the older discipline, because the negotiation accepted, in that confined setting, the business argument that fixed capital committed under a government contract requires a remedy outside the contracting government own courts. That division, between general regulation returned to domestic forums and contract based protection retained in specified sectors, was the settlement Chapter 14 enacted, and it is the reason investment stands apart from market access in any honest comparison of the two agreements.

Duration and Review: Renegotiation Put on the Calendar

On the fifth axis of comparison, duration and review, the difference between the two agreements was not a matter of degree in the staging of a tariff or the height of a threshold. The earlier agreement was written as if North American trade integration, once established, would continue without a date on which the parties had to decide again whether it should continue at all. The successor agreement was written on the opposite premise. Its text placed a term on the arrangement, placed reviews on a calendar, and made continuation depend on an affirmative act by each party rather than on the absence of an act by any party. That inversion, from permanence assumed to permanence requiring renewal, changed what the agreement was as a legal instrument even where the goods crossing the border and the duties applied to them changed little. Duration was itself a substantive provision, because a scheduled reconsideration alters the calculations of governments that administer the agreement and firms that build under it.

The earlier agreement contained no term and no scheduled review. The North American Free Trade Agreement entered into force on January 1, 1994, and its text set no anniversary on which the parties were required to meet and assess its operation, no date on which its obligations lapsed, and no procedure by which continuation had to be affirmed. A party that wished to leave could do so, and the provision governing that choice, Article 2205, allowed withdrawal on six months notice given to the other parties, with the agreement remaining in force for the parties that stayed. Withdrawal was therefore the only temporal mechanism the text supplied. It was unilateral, available at any time, and it required no cause, no review, and no joint finding that the agreement had succeeded or failed. Between entry into force and withdrawal, the text contemplated nothing. The silence was the design.

That design rested on an assumption embedded in the structure itself. An agreement without a term treats the bargain struck at negotiation as a settlement rather than as a provisional arrangement subject to a later verdict. Tariff elimination schedules in the earlier agreement ran forward through staged reductions to a completed state in which duties had been removed, and once that state had been reached the schedules had no further work to do. The structure pointed toward integration maintained by inertia. A government dissatisfied with the agreement had remedies for particular disputes and a right of exit for the whole arrangement, but it had no regular occasion, supplied by the agreement itself, on which dissatisfaction could be converted into renegotiation without the drama and the legal consequence of threatening to leave. Permanence was the default, and the burden of action rested entirely on the party that wanted change. No scheduled occasion existed for correcting drift, for revisiting sectors transformed by technology, or for requiring the arrangement to justify itself again.

The successor agreement reversed the default. Article 34.7 of the United States-Mexico-Canada Agreement, as set out in the text approved by the United States-Mexico-Canada Agreement Implementation Act, Public Law 116-113, signed on January 29, 2020, provided that the agreement terminates sixteen years after its entry into force unless each party confirms in writing that it wishes the agreement to continue. Entry into force took place on July 1, 2020, on which date the successor replaced and terminated the earlier agreement, so the sixteen year term stated in Article 34.7 runs to July 1, 2036. Termination on that date follows under Article 34.7 unless each party has confirmed continuation in writing. Confirmation by two parties does not satisfy the provision, and silence by one party does not satisfy the provision. Continuation beyond the term requires an affirmative written confirmation from Canada, from Mexico, and from the United States alike.

Does a fixed term change how a trade agreement works even before any review takes place?

Yes. A fixed term changes behavior from the date of entry into force because every government, producer, and investor can read the review dates on the calendar and must plan for a decision point at which continuation requires an affirmative written confirmation from each of the three parties.

The review machinery that leads toward that decision point was stated with the same precision. Article 34.7 provided for a joint review by the parties on the sixth anniversary of entry into force, which falls on July 1, 2026, and for joint reviews at least every six years after that review. A joint review under the provision is not a ceremonial stocktaking detached from the term, because the review is the occasion at which each party states, in writing, whether it wishes the agreement to continue. If a party withholds confirmation at a review, Article 34.7 does not produce immediate termination. Instead, the parties review the agreement annually after that review, and the agreement terminates ten years after the review at which confirmation was withheld unless confirmation is given during that interval. Every element of that sequence is stated here as the enacted schedule of Article 34.7 and as nothing else. By September 15, 2020, the successor had been in force since July 1, 2020, for a period measured in weeks, no joint review under Article 34.7 had taken place, and no party had confirmed or withheld confirmation under the provision.

Alongside the term and the reviews, the older escape route survived in its own provision. Article 34.6 preserved a right of withdrawal on six months notice, separate from Article 34.7 and independent of its calendar. The term did not replace sovereign exit with collective review. The term added a collective review, with legal consequences for continuation, to an arrangement that retained unilateral exit, so a government possessed both a scheduled forum in which continuation itself was the question and an unscheduled right to depart on six months notice whether or not any review was pending. In the longer sequence of American trade statutes traced in the postwar record of trade legislation and its successive grants of negotiating authority, an expiry date of this kind had no close precedent, and in that longer record the successor agreement stands as the first major American trade agreement with an expiry date built into it.

What a scheduled review does to bargaining incentives follows from the mechanism itself, without any need to predict how a particular government will use a review that, by September 15, 2020, lay almost six years in the future. Under the earlier agreement, a party that wanted a provision changed had to persuade the other parties that reopening a settled text was worthwhile, and persuasion faced the inertia of a permanent arrangement in which doing nothing left the agreement intact. Under the successor, doing nothing at a review had a different meaning. A party that wanted change could wait for an occasion that the text itself supplied, arrive with demands, and withhold the written confirmation on which extension depended until the other parties addressed those demands or accepted the consequences of annual reviews and a ten year run toward termination. The confirmation requirement converts each party into a necessary party to continuation, and a necessary party holds leverage that a merely dissatisfied party under a permanent agreement does not hold. That leverage exists whether or not it is exercised, because the other parties must calculate, in advance of each review, what price in concessions might be demanded for a confirmation that cannot be dispensed with.

Investment horizons respond to the same calendar through a different channel. A plant, a mine, a transmission investment, or a supplier network built for continental production recovers its cost across a period that can exceed a single review cycle, and a firm committing capital after July 1, 2020 could read in Article 34.7 the dates on which the preferential terms governing that investment would be reconsidered. The provision did not make investment impossible, and the sixteen year initial term, with a further ten year run-off after a withheld confirmation, is longer than the payback period of many projects. The provision did introduce a form of political risk that the earlier agreement had excluded, the risk that preferential access itself, rather than a tariff rate or a regulation within the agreement, would be the subject of a scheduled decision.

Against that cost, the mechanism offered a benefit that its advocates described in terms of discipline. The Office of the United States Trade Representative, in presenting the successor agreement, treated the review and sunset structure as a device that would prevent a trade agreement from drifting onward after the conditions that justified it had changed, because a future government would not need to manufacture a crisis or threaten withdrawal in order to force a reckoning. On that reading, the calendar disciplines all three parties continuously, because a government that tolerates noncompliance in a sector covered by an enforceable chapter knows that a review approaches at which the other parties can connect that record to the question of continuation.

The contrary reading received its fullest statement from the trading partners whose confirmation the provision made necessary and from business organizations whose members planned across the review dates. The governments of Canada and Mexico, in their official statements during the negotiation and approval of the successor, resisted an expiry structure of the kind proposed during the talks and accepted a review mechanism whose long term and long run-off softened the original proposal. The United States Chamber of Commerce and other business organizations argued, in submissions and statements on the agreement, that a recurring expiry injects uncertainty into decisions that depend on stable preferential access, because a firm cannot know at the moment of investment whether the terms governing the later years of that investment will survive a review at which any one of three governments can withhold confirmation for reasons that may have little to do with the sector in which the firm operates.

The companion change on this axis concerned the machinery for deciding disputes between the parties themselves, and here the successor addressed a failure that the earlier agreement had exhibited in operation rather than in design alone. State-to-state dispute settlement under the earlier agreement sat in its Chapter 20, which provided for consultations between the parties, referral to the commission established by the agreement, and, where a matter remained unresolved, the formation of a panel to examine the matter and report. On paper the ladder resembled the dispute procedures of other trade agreements of its period. In practice the ladder depended on steps that a reluctant respondent could obstruct. Panel formation required a roster of panelists from which panel members could be drawn, and the roster under Chapter 20 was never maintained in the manner the text contemplated. Without a maintained roster, the composition of a panel became a negotiation in itself, and a respondent that preferred delay, or that preferred the existing state of affairs to any ruling at all, could stall the process by refusing to cooperate in the appointments on which a panel depended. A procedure that can be prevented from starting offers no discipline to a party that expects to lose, and it offers no remedy to a party that expects to win, because the expected ruling never arrives.

The record matched the vulnerability. The last panel formed under Chapter 20 sat in 2001, and after that date the mechanism fell into disuse, not because disputes between the three parties disappeared, and not because the substantive obligations of the agreement ceased to matter, but because the procedural path to a panel had been shown to be blockable. The parties conducted their disagreements through other channels, through diplomacy, through domestic measures, and through procedures under other agreements where those procedures were available. Disuse of that kind carries a consequence beyond the individual cases that were never brought. An obligation in a trade agreement derives part of its force from the credible prospect that a breach can be taken to a panel and decided against the breaching party. Where the prospect is not credible, because the respondent can prevent the panel from forming, the obligation rests on voluntary compliance and on retaliation threatened outside the procedure, and both substitutes are weaker and less predictable than a ruling delivered under the text. The Chapter 20 failure was distinct and purely procedural. The text promised adjudication between governments and delivered adjudication only where the respondent permitted adjudication to begin.

The repair was made not in the agreement as signed at Buenos Aires on November 30, 2018, but in the Protocol of Amendment signed on December 10, 2019, which altered several parts of the successor before any party brought it into force. For state-to-state dispute settlement, the protocol repaired the panel formation rules in Chapter 31 of the successor so that a respondent cannot block a panel by refusing roster appointments. The amended design supplied the appointments machinery that Chapter 20 had lacked in operation, with rosters established and maintained and with default steps that allow a panel to be composed where a party declines to participate in composing it. A complainant under Chapter 31 possesses a path to a constituted panel that does not run through the consent of the government complained against. Consultations and commission referral remain part of the sequence, and a party remains free to settle a matter at any stage, but settlement is no longer the only alternative to a stalled proceeding. The respondent retains every defense on the merits that the text allows and the benefit of any exception the text provides. What the respondent no longer retains, under the repaired formation rules, is a veto over the existence of the tribunal that would decide those arguments. For an agreement that had also moved labor and environment obligations into its main text, where they became subject to the state-to-state procedure, the repair strengthened chapters that the dispute chapter never mentions, because enforceable obligations in Chapters 23 and 24 depend for their credibility on the same panel machinery.

One further element of the earlier dispute architecture survived the transition intact in function, though under a different chapter number. Binational panel review of antidumping and countervailing duty determinations, established in Chapter 19 of the earlier agreement, allowed a determination by a domestic investigating authority to be reviewed by a binational panel instead of by the domestic courts of the importing country, at the request of an interested party in the cases the chapter covered. That review was carried into the successor as Chapter 10, retaining binational panel review of duty determinations in substantially the role it had occupied before. The retention matters for the comparison on this axis because it separates the part of the dispute system that had failed from the part that had functioned. Chapter 20 state-to-state settlement had fallen into disuse after panel blocking and roster neglect, and the protocol rebuilt its formation rules. Chapter 19 binational review had operated as a working channel for challenges to duty determinations, and the successor preserved it, renumbered, rather than rebuilding it. The successor therefore treated dispute settlement not as a single inheritance to be kept or discarded but as two mechanisms with two different records, one repaired because it had stalled and one retained because it had worked.

Taken together, the two halves of this axis describe the successor’s approach to time and to obligation more completely than either half does alone. On duration, the successor abandoned the assumption that a continental trade agreement, once made, continues until a party chooses the drastic remedy of withdrawal, and substituted a term of sixteen years to July 1, 2036, a joint review on July 1, 2026 and at least every six years after, annual reviews with a ten year run toward termination after a withheld confirmation, and a separate six month withdrawal right in Article 34.6 that survives alongside the calendar. On dispute settlement, the successor abandoned the assumption that a panel procedure stated in the text would operate because it had been stated, and substituted formation machinery in Chapter 31, repaired by the Protocol of Amendment of December 10, 2019, under which a respondent cannot prevent a panel from being formed by refusing to cooperate in appointments, while binational review of duty determinations passed from Chapter 19 of the earlier agreement into Chapter 10 of the successor. On the measure of what crosses the border at what price, that record supports a judgment of modest change. On the measure of how long the arrangement is promised to last, who must affirm that it continues, and whether a government accused of breaching it can be brought before a panel at all, the change was among the largest that the successor made.

The Verdict: Name the Measure Before Judging the Change

The comparison yields two answers because the comparison measures two different things, and a verdict that refuses to name its measure conceals that difference instead of resolving it. On market access, the later agreement left the tariff structure largely where the earlier agreement had left it, since staged elimination under the North American Free Trade Agreement had run to completion for the last categories in 2008. On rules of origin, the later agreement raised the regional value content threshold for automobiles and added a wage condition and metals sourcing requirements that had no counterpart in the earlier text. On labor enforcement, obligations moved from a side agreement pledging each party to enforce domestic law into the main text, joined by a facility specific mechanism aimed at denials of free association and collective bargaining at named workplaces in Mexico. On investment, a direct investor remedy supplied for twenty-six years was withdrawn from two of the three bilateral relationships and narrowed sharply in the third. On duration and review, the later agreement acquired an expiry date, a scheduled joint review, and a repaired method for forming state to state panels.

Measured by what crosses the border at what price, the change was modest, because tariffs were already at or near zero and the market access gains are concentrated in a handful of agricultural quotas. The earlier agreement had phased out tariffs on substantially all goods among the three countries, and by the time negotiators addressed a replacement, remaining protection sat in particular sectors and instruments rather than spread across the schedule. The Office of the United States Trade Representative presented the agricultural openings as the market access centerpiece, and their structure confirms the concentration: gains arrived as tariff rate quotas in named product lines, phased in across about six years, inside a Canadian supply management system that the agreement left standing. Canada agreed to eliminate its Class 6 and Class 7 milk pricing classes six months after entry into force, a commitment falling due on January 1, 2021 that stood as a dated obligation at the vantage of this comparison rather than a completed act, and to provide United States specific tariff rate quotas across dairy lines summing to roughly 3.5 to 3.6 percent of the Canadian dairy market, on analyses by the Office of the United States Trade Representative and university researchers. Canada also accepted export caps on skim milk powder, milk protein concentrates, and infant formula, with surcharges above the caps and an aggregate cap for skim milk powder and milk protein concentrates of 55,000 metric tons in the first year falling to 35,000 metric tons in the second year, and added chicken quotas rising to 57,000 metric tons by year six on top of the quota maintained under the World Trade Organization. Government procurement moved the other way, since the procurement chapter applies between the United States and Mexico while Canada is covered through the World Trade Organization Agreement on Government Procurement, so one channel of guaranteed access narrowed even as several agricultural channels widened.

Measured by what a government or a firm can be made to do and by whom, the change was large, and the deciding factor is the measure chosen. The origin rules carry the first part of that verdict because a wage condition entered the origin rules: the earlier agreement required regional value content of 62.5 percent under the net cost method for passenger automobiles, light trucks, engines, and transmissions, and 60 percent for other vehicles and parts, while the later agreement raises the threshold for passenger vehicles, light trucks, and core parts to 75 percent, phased in at 66 percent in 2020, 69 percent in 2021, 72 percent in 2022, and 75 percent from 2023, with principal parts reaching 70 percent and complementary parts reaching 65 percent by 2023, on schedules published by the Office of the United States Trade Representative and summarized by the Chicago Fed. Labor value content, which has no counterpart in the earlier agreement, requires a share of production value from facilities paying an average base hourly wage of at least 16 dollars, set for passenger vehicles at 30 percent in 2020, 33 percent in 2021, 36 percent in 2022, and 40 percent from 2023, with light trucks reaching 45 percent in 2023. A further requirement sets 70 percent of the purchase value of the steel and aluminum of an automaker to originate in North America from 2020, with steel melted and poured in North America applying across a seven year runway, and three added certifications covering labor value content, steel, and aluminum accompany the regional value content certification.

Labor enforcement carries the deepest part of the large verdict because the earlier architecture placed the rights at the center of the later mechanism beyond the reach of its own strongest remedy. Under the North American Agreement on Labor Cooperation, each party pledged to enforce domestic labor law, and eleven labor principles sat in three tiers: National Administrative Offices could receive submissions and hold reviews and hearings, ministers could consult, and an Evaluation Committee of Experts could be convened for second and third tier matters, but an arbitral panel, the only stage capable of producing a monetary enforcement assessment, was available only for a persistent pattern of failure to enforce occupational safety and health, child labor, or minimum wage standards. Freedom of association and the right to bargain collectively sat in the first tier, eligible for review and consultation and structurally outside the expert committee and panel track, and the capped assessment at the top of that ladder was never imposed: no monetary enforcement assessment was ever imposed under the side agreement in twenty-six years, and no case reached the panel stage, on analyses of the Congressional Research Service. Chapter 23 of the later agreement places enforceable labor obligations in the agreement itself, including adoption and maintenance of the rights stated in the 1998 Declaration of the International Labour Organization, and Annex 23-A required Mexico to enact collective bargaining and union democracy reforms, carried into domestic law in the reform of May 1, 2019, before entry into force. The Protocol of Amendment of December 10, 2019 removed the requirement to prove that a violation occurred in a manner affecting trade and investment as the signed text had framed that burden for key obligations, substituting a presumption that a failure affects trade unless the respondent shows otherwise, and created the facility specific rapid response mechanism in Annex 31-A for the United States and Mexico and Annex 31-B for Canada and Mexico, running against covered facilities in Mexico in priority sectors. A party believing that workers at a named facility face a denial of rights may ask the host government to review, and an independent rapid response labor panel may then determine whether a denial exists, with remedies falling on the goods of the facility, escalating on repetition from the facility to sister facilities of the same entity and then to the entity, across timelines of weeks and months. A facility in one country became answerable to a panel convened under a trade agreement, a form of answerability the earlier settlement, with labor content in a side agreement negotiated after the main text, had withheld.

Investment extends the large verdict by subtraction, and duration completes the verdict by reversing a default. Chapter 11 of the earlier agreement gave investors of one party direct claims against host governments of another before arbitral tribunals, and that investor state dispute settlement became the most contested and most litigated feature of the agreement across twenty-six years and dozens of claims, including claims directed at environmental and public health measures. Chapter 14 of the later agreement eliminates the mechanism between the United States and Canada, leaves the mechanism absent between Canada and Mexico under this agreement, with investors in that relationship relying on another agreement between those two countries, and narrows the mechanism between the United States and Mexico, where Annex 14-D confines most claims to direct expropriation, national treatment, and most favored nation treatment after thirty months of domestic litigation, while Annex 14-E preserves fuller protections for covered government contracts in specified sectors including oil and natural gas, power generation, telecommunications, transportation services, and certain infrastructure. Annex 14-C permits claims arising from investments established while the earlier agreement was in force for three years after termination, a schedule running to July 1, 2023. On duration, the earlier agreement had no term and no scheduled review, with withdrawal on six months notice under Article 2205, whereas Article 34.7 of the later agreement terminates the agreement sixteen years after entry into force, on July 1, 2036, unless each party confirms in writing a wish to continue, with a joint review on the sixth anniversary, July 1, 2026, and at least every six years afterward, annual reviews and termination ten years after a review if confirmation is withheld at that review, and withdrawal on six months notice surviving separately under Article 34.6. The later agreement is the first major American trade agreement with an expiry date built into the text, and the dispute settlement repair reinforces the same axis, since the Chapter 20 process of the earlier agreement could be stalled because rosters were never maintained and a respondent could block panel formation, leaving the mechanism in disuse after the last panel in 2001, while the Protocol of Amendment repaired panel formation in Chapter 31 so that a respondent cannot block a panel by refusing roster appointments.

The verdict compresses into one sentence that names both measures without blending them: “A rebrand on tariffs and a rewrite on enforcement.” The first half earns its place because the tariff schedules changed least where trade was largest, and a different name on an agreement whose border prices a predecessor had already set invites the skepticism that the word rebrand expresses. The second half earns its place because enforcement changed most where the predecessor was weakest, in labor obligations, investment exposure, panel formation, and the lifespan of the agreement itself. Neither half should be quoted without the other, since each disciplines the other. Depth is a property of the texts under a named measure, and under the two measures adopted here the texts support a modest verdict and a large verdict simultaneously.

Does stronger enforcement language guarantee stronger enforcement in practice?

No. By September 15, 2020, the rapid response mechanism had produced no petition, no panel, and no remedy. Its text created legal capacity to challenge a denial of rights at a named facility, but capacity describes what a party may attempt, not what any proceeding has established or changed.

The counter reading deserves full weight, because stronger enforcement text is not the same as stronger enforcement outcomes, and the distance between those propositions is the distance between a statute book and a workplace. The successor agreement had been in force since July 1, 2020, about ten weeks at the vantage of September 15, 2020, and the facility mechanism is untested architecture: no petition had been filed by September 15, 2020, no panel had been convened, and no remedy had been imposed. The Interagency Labor Committee published interim procedural guidelines for petitions on June 30, 2020, a step describing preparation to receive complaints rather than experience resolving complaints. The enforcement verdict is therefore a verdict about legal capacity, about what the text empowers a party to attempt and what consequences the text authorizes if an attempt succeeds, and not a verdict about results in any factory, warehouse, or mine. The origin rules carry a parallel caution, because the effects of the origin rules depend on producers choosing certification over the ordinary tariff, a choice the text cannot compel: a producer facing the higher threshold, the wage condition, the metals requirements, and three added certifications may elect to pay the ordinary tariff rather than certify, and that election is the practical ceiling on any origin rule. Where the ordinary tariff is low, the incentive to rebuild a supply chain around a wage condition weakens accordingly, the dairy quotas create access that exporters may or may not fill in a given quota year, and the investment withdrawal gains practical weight only through disputes that governments and investors choose to bring or to avoid. Across every axis where the change was large, the change consists of altered capacities, incentives, and defaults, and outcomes follow from choices made inside that altered structure.

For the student encountering North American trade law as a subject of study, the earlier agreement repays study first, because the later agreement is legible only as an answer to the earlier text: the side agreement architecture, the tiering that placed freedom of association beyond the reach of a panel, the investor remedy in full breadth, and the absence of any term or review each explain a corresponding feature of the successor by contrast, and a grounded study plan for working through trade statutes in sequence points in the same direction as the texts, toward reading the agreement approved in 1993 before turning to the agreement approved in 2020 and asking, provision by provision, what changed and what record in the earlier design justified the change.

For the staffer advising a member of Congress or a committee on trade oversight, the later agreement repays study first, because the obligations binding the United States, the deadlines structuring executive branch action, and the dates on which decisions fall due live in the successor text and the implementation act signed on January 29, 2020: Chapter 23 and Annex 31-A before the side agreement, Article 34.7 and the joint review scheduled for July 1, 2026 before the withdrawal provision of the earlier agreement, and the narrowed investment annexes governing exposure after July 1, 2020 before the broader chapter that governed exposure during the preceding twenty-six years, with the earlier text entering the work as a baseline for measuring departure rather than as the operative document.

For the researcher measuring how trade agreements affect conduct inside firms and governments, the later agreement repays study first as a research design and the earlier agreement repays study as the comparison case that gives the design meaning, because the successor varies several treatments at once in ways that invite separation axis by axis: a wage condition entered the origin rules for one sector, a facility specific remedy attached to two named rights in one country, an investor remedy disappeared from two relationships and narrowed in a third, and an expiry date entered an arrangement that previously had none, so hypotheses about certification choices, sourcing decisions, complaint behavior, and regulatory caution can be stated from the successor text and tested against the long baseline that twenty-six years under the earlier agreement supply, under the discipline that the vantage of 2020 permits hypotheses about capacity and incentive rather than findings about effect.

For the journalist explaining the replacement of one agreement by another to an audience that encounters trade through prices, plants, and paychecks, the later agreement repays study first and the earlier agreement repays quotation first, because reporting on the successor concerns provisions taking effect on July 1, 2020 and commitments falling due on stated dates such as January 1, 2021 for the dairy pricing classes and 2023 for the final content thresholds, while explanation of why those provisions exist requires the earlier record in compact form: tariffs already eliminated, a labor side agreement that never produced an assessment, an investor remedy that drew claims against environmental and public health measures, and a state to state process in disuse after 2001; command of the successor text supports accuracy about what changed on paper, and command of the earlier record supports explanation, without exaggeration in either direction, of how a change that left border prices largely untouched occupied Congress, the three governments, organized labor, and industry across three years of negotiation.

The closing assessment returns to the measure, because the measure is where the comparison began and where responsible use of the comparison ends. A reader who asks only about prices will find the replacement a modest event, a set of agricultural quotas and export caps laid over a tariff structure completed years before, and that finding is correct within its limits and worthless beyond them. A reader who asks about power will find the same replacement a large event, a reallocation of who may complain, about what, before which body, under what deadline, and with what consequence, extending from a single facility in Mexico to the lifespan of the agreement itself, and that finding is equally correct within its limits and equally worthless beyond them. The agreements differ most where the earlier settlement was most deliberately permissive, in the enforcement of labor rights, in the exposure of governments to investor claims, and in the assumption of indefinite continuation without a scheduled reckoning, and the agreements differ least where the earlier settlement had already done its work, in the elimination of tariffs among three economies trading on a continental scale.

That distribution of sameness and difference is the finding the five axes jointly support, and the finding carries a discipline for every later comparison between an agreement and a successor. Claims of transformation should be required to name the axis on which transformation occurred and the evidence that the axis moved. Judged under that discipline, the successor agreement stands as a conservative instrument on prices and an ambitious instrument on compulsion, drafted by parties who left the border largely as a predecessor had arranged the border and who rewrote, in the main text and under a deadline measured in years rather than decades, the rules governing what the parties to the agreement could be made to answer for.

The comparison ends where the measurement began. Border prices moved least, because a predecessor had finished the tariff work years before the successor entered into force, and compulsion moved most, because obligations, remedies, and the lifespan of the arrangement itself were rewritten in the main text. A careful reader keeps the two findings separate, names the axis behind every claim of change, and treats dated commitments as commitments until the schedule date arrives and performance can be examined on evidence. Readers assembling a study file for that longer examination can keep your statute notes, citations, and case chronologies together free on VaultBook, and readers building the underlying civics foundation alongside the trade material can practice and revise US government and civics material on ReportMedic. Either habit serves the same end: a record organized well enough that a later comparison can test text against outcome rather than against recollection.

Frequently Asked Questions

Q: What is the difference between USMCA and the 1993 agreement?

The difference runs along five axes rather than along a single line. On market access, the successor agreement largely carried forward a tariff structure that the earlier agreement had completed by 2008. On rules of origin, the successor raised automotive content thresholds and added a wage condition and metals sourcing requirements with no earlier counterpart. On labor enforcement, obligations moved from a side agreement into Chapter 23 of the main text, joined by a facility specific mechanism. On investment, a direct investor remedy was withdrawn or narrowed. On duration, an agreement with no term gave way to an agreement with an expiry date and a scheduled joint review.

Q: Did USMCA change tariffs much compared with the earlier deal?

Measured at the border, the change was small. Tariffs on substantially all goods among the three countries had been eliminated under the earlier agreement by the end of the transition, with the last staged categories reaching zero in 2008, so nearly all North American goods trade was duty free before the successor agreement entered into force on July 1, 2020. The successor schedules mostly restated that position. Movement on the axis sat inside managed systems instead: dairy and poultry access enlarged through tariff rate quotas, export caps placed on specified Canadian dairy products, and government procurement coverage narrowed between the parties. Continuity on tariff lines and change inside quotas describe the same axis at once.

Q: How did USMCA change labor enforcement from the earlier deal?

The earlier deal housed labor content in a side agreement under which each party pledged to enforce domestic law, with review and consultation as the working stages and a monetary assessment available only for a narrow band of standards. The successor agreement placed enforceable labor obligations in Chapter 23 of the main text, required Mexican collective bargaining and union democracy reforms under Annex 23-A, eased the burden on a complaining party through a presumption on trade effect, and created a facility specific rapid response mechanism directed at named facilities in Mexico. By September 15, 2020 that mechanism remained untested: no petition had been filed, no panel had convened, and no remedy had been imposed.

Q: Did USMCA remove investor state dispute settlement?

Removal describes two of the three relationships and narrowing describes the third. Between the United States and Canada, investor state dispute settlement was eliminated, leaving investors to domestic courts or to state to state channels. Between Canada and Mexico, the mechanism was absent under the successor agreement, and investors in that relationship relied on another agreement between those two countries. Between the United States and Mexico, the mechanism survived in confined form: most claims under Annex 14-D were limited to direct expropriation and nondiscrimination after thirty months of domestic litigation, while Annex 14-E preserved fuller protections for covered government contracts in specified sectors. Legacy claims under Annex 14-C ran on a schedule to July 1, 2023.

Q: Why does USMCA expire when the earlier deal did not?

The expiry date was a deliberate design choice written into Article 34.7. The earlier agreement carried no term and no scheduled review, so continuation was the default and change required a party to threaten or use withdrawal on six months notice. The successor agreement reversed the default: the agreement terminates sixteen years after entry into force, on July 1, 2036, unless each party confirms in writing a wish to continue. A joint review on the sixth anniversary, July 1, 2026, supplies the occasion for that confirmation. Supporters presented the structure as discipline against drift, while critics presented the same structure as scheduled uncertainty for investment planned across review dates.

Q: Which is better for autos, USMCA or the earlier rules?

The answer depends on the interest being measured, and the texts support each reading. Under the earlier rules, a vehicle qualified at 62.5 percent regional value content, a threshold around which supply chains had been designed for a generation. Under the successor rules, the threshold rises to 75 percent, principal and complementary parts face tiered thresholds, a labor value content test ties part of production value to facilities paying an average base hourly wage of at least 16 dollars, and 70 percent of steel and aluminum purchases must originate in North America. Automakers stressed cost and complexity. Labor organizations stressed that the wage linkage stopped rewarding movement toward the lowest wage location inside the region.

Q: Did USMCA fix the blocked dispute panel problem?

On formation, the successor text addressed the failure directly. Under Chapter 20 of the earlier agreement, panel rosters were never maintained as the text contemplated, a respondent could stall panel formation by refusing cooperation on appointments, and the last panel sat in 2001 before the mechanism fell into disuse. The Protocol of Amendment of December 10, 2019 repaired panel formation in Chapter 31 of the successor agreement so that a respondent cannot block a panel by refusing roster appointments, with default steps allowing a panel to be composed. The repair concerns the existence of the tribunal. Defenses on the merits and exceptions under the text remain available to a respondent in any case that reaches a panel.

Q: Which should a student study first, USMCA or the earlier agreement?

The earlier agreement repays study first, because the successor agreement reads as an answer to the earlier text and the answer is legible only against the question. The side agreement architecture explains why labor obligations moved into the main text. The tiering that placed freedom of association beyond the reach of a panel explains the facility specific mechanism. The breadth of the investor remedy explains the narrowing in Chapter 14. The absence of any term or review explains the expiry date and the scheduled joint review. A student who begins with the successor text encounters provisions whose purpose sits in the record of the earlier design, and provision by provision comparison supplies that purpose.

Q: What changed for dairy farmers under USMCA compared with NAFTA?

The dairy package pulled three levers at once. Canada agreed to eliminate the Class 6 and Class 7 milk pricing classes six months after entry into force, a commitment falling due on January 1, 2021 and stated at the comparison date as a dated obligation rather than a completed act. Canada also provided tariff rate quotas specific to the United States across dairy product lines, phased in over about six years and summing to roughly 3.5 to 3.6 percent of the Canadian dairy market. Canada additionally accepted caps on exports of skim milk powder, milk protein concentrates, and infant formula, with surcharges above the caps. The managed supply system itself remained standing around the enlarged quantities.

Q: Did USMCA add a digital trade chapter that NAFTA did not have?

Yes. The successor agreement contained chapters with no counterpart in the earlier agreement, and digital trade, carried as Chapter 19, stood among the additions, alongside macroeconomic policy and exchange rate matters as Chapter 33. The earlier agreement had been negotiated before electronic commerce occupied a central place in North American trade, and the text carried no digital trade chapter. Placement matters alongside existence: labor and environment content moved from side agreements into Chapters 23 and 24 of the main text during the same transition, so the successor agreement differed from the earlier agreement both in subjects covered and in the legal position occupied by subjects carried across from the earlier framework.

Q: What happened to NAFTA’s environmental side agreement under USMCA?

Environmental content changed position in the same manner as labor content. Under the earlier bargain, environmental obligations sat in the North American Agreement on Environmental Cooperation, a side agreement negotiated after the main text to secure votes for passage, beside the agreement rather than inside the agreement. Under the successor agreement, environment appeared as Chapter 24 of the main text itself. An obligation housed in a main text chapter occupies a different legal position from an obligation housed in an accompanying instrument, because the chapter stands subject to the general dispute settlement system of the agreement, as repaired by the Protocol of Amendment for panel formation in Chapter 31. The relocation formed part of the broader architectural change between the two instruments.

Q: Can a country still withdraw from USMCA the way it could from NAFTA?

Withdrawal survived the transition as a separate right. Under the earlier agreement, Article 2205 permitted a party to withdraw six months after giving written notice to the other parties, with no consent from the other governments required and no cause stated. Under the successor agreement, Article 34.6 preserved withdrawal on six months notice, independent of the term and review calendar in Article 34.7. A government therefore holds two distinct temporal instruments under the successor text: a scheduled collective review at which continuation requires affirmative written confirmation from each party, and an unscheduled unilateral right to depart on six months notice whether or not a review is pending.

Q: What happens at the scheduled joint review of USMCA?

The joint review is the occasion on which each party states in writing whether the agreement should continue. Article 34.7 sets the first joint review on the sixth anniversary of entry into force, July 1, 2026, with further reviews at least every six years after that review. Confirmation by each party carries the agreement beyond the sixteen year term that otherwise ends on July 1, 2036. If a party withholds confirmation at a review, termination does not follow at once: the parties review the agreement annually, and the agreement terminates ten years after that review unless confirmation is given during the interval. Every step of that sequence is enacted schedule, stated in advance of any review taking place.

Q: Are claims under NAFTA’s investment chapter still possible after it ended?

A transitional route survived termination for a defined period. Annex 14-C of the successor agreement permits legacy claims arising from investments established while the earlier agreement was in force, under the conditions that annex prescribes, for three years after termination of the earlier agreement. The earlier agreement terminated on July 1, 2020, so the legacy window runs as enacted schedule to July 1, 2023. The provision prevented abrupt extinguishment of rights accrued around investments made in reliance on Chapter 11, while preventing those rights from continuing without limit. An investor who entered after July 1, 2020 holds no legacy route, and Chapter 14 of the successor agreement governs that investment without the transitional overlay.

Q: How does the steel and aluminum rule change what automakers must buy?

The rule operates on purchases rather than on the composition of a single vehicle. The successor agreement requires that 70 percent of an automaker’s steel and aluminum purchase value originate in North America, a test answered from procurement records across a period rather than from the bill of materials of one car or truck. A further condition, that steel be melted and poured in North America, addresses the difference between steel finished in the region from imported substrate and steel produced from molten metal in the region, and that condition applies on a seven year runway as enacted schedule. Two certifications, one for steel and one for aluminum, accompany the regional value content certification for each claim of preferential treatment.

Q: Did USMCA keep NAFTA’s binational panels for duty disputes?

Binational panel review survived the transition, renumbered. Under Chapter 19 of the earlier agreement, a determination by a domestic investigating authority in an antidumping or countervailing duty matter could be reviewed by a binational panel instead of by the domestic courts of the importing country, at the request of an interested party in covered cases. The successor agreement carried that review into Chapter 10 in substantially the same role. The retention separates the part of the dispute architecture that had functioned from the part that had stalled: state to state settlement under Chapter 20 fell into disuse after panel blocking, while binational review of duty determinations operated as a working channel and was preserved rather than rebuilt.

Q: Why was the signed USMCA text amended before Congress approved it?

The signed text could not command a House majority until enforcement concerns had been addressed, and amendment supplied the price of passage. After signature at Buenos Aires on November 30, 2018, the House majority declined to schedule an implementing bill during 2019 while a working group negotiated with the Office of the United States Trade Representative. The resulting Protocol of Amendment, signed December 10, 2019, changed enforcement machinery, eased the burden governing key labor and environmental obligations, created the facility specific rapid response mechanism that the signed text had not contained, and repaired dispute settlement panel formation. Congress then approved the agreement as amended, and the implementing statute, Public Law 116-113, was signed on January 29, 2020.

Q: What is the labor value content rule and how is it measured?

Labor value content requires that a stated share of a vehicle’s production value come from facilities paying an average base hourly wage of at least 16 dollars. For passenger vehicles the required share stands at 30 percent in 2020 and rises through 33 percent in 2021 and 36 percent in 2022 to 40 percent from 2023, while light trucks reach 45 percent in 2023. The test operates on value and payroll rather than on physical composition: production value generated at a facility meeting the average wage threshold counts toward the share, and production value generated below the threshold does not count, even where the facility sits inside the region. Credits for high wage technology and assembly expenditures may also count toward the requirement.

Q: Which agreement gives governments more power over a single factory?

The successor agreement gives a government a route that reaches a single factory, and the earlier agreement supplied no comparable route. Under the side agreement ladder, the subject was always a pattern of enforcement across a national system, pursued through administrative office review, ministerial consultation, expert analysis, and, for a narrow band of standards only, a panel. Under Annex 31-A and Annex 31-B, a party that believes workers at a named facility in Mexico face a denial of free association and collective bargaining rights may request a host government review, and an independent panel may determine whether a denial exists, with remedies falling on the goods of that facility. By September 15, 2020 the mechanism remained untested, with no petition filed and no panel convened.

Q: Did USMCA narrow government procurement coverage compared with NAFTA?

Coverage inside the North American instrument narrowed. A procurement chapter binds only the parties between which the chapter runs and only for covered entities and purchases, so coverage is the operative question. Under the successor agreement, the procurement chapter applies between the United States and Mexico. Canada stands outside that chapter, and procurement coverage between Canada and the other two parties runs through the World Trade Organization Agreement on Government Procurement instead. The result records a contraction in the reach of the agreement itself on the market access axis, set against continuity on ordinary tariffs and enlargement inside managed agricultural quotas. Procurement therefore supplies the corrective case against any claim that the successor agreement widened access on every front.