Three countries, one treaty, three different names for it. Canada calls it CUSMA. The United States calls it USMCA. Mexico calls it the T-MEC. Nobody actually disagrees about what’s in it—the text is identical, negotiated by the same three delegations at the same table—they just couldn’t agree on what to call it, so each government went home and named it in whatever order flattered itself most. It’s a small, almost comic detail, and also a pretty accurate preview of how the whole negotiation went: substantial agreement on the mechanics, real friction over who got to claim the win.

The deal itself very nearly didn’t happen on schedule. By the last week of September 2018, the U.S. and Mexico had a text. Canada didn’t, and Washington had made it clear it would publish the U.S.–Mexico deal alone and dare Ottawa to sign on later from a weaker position. Chrystia Freeland, then Canada’s foreign minister, spent the final days in more or less continuous shuttle diplomacy, and the three sides landed a deal right at the September 30 midnight deadline the Americans had set: the kind of down-to-the-wire finish that makes for a better story than most trade negotiations, which typically end in a scheduled press conference rather than a deadline sprint.

Background

What got replaced was NAFTA, which had run continental trade since 1994. The Trump administration, campaigning in 2016 on NAFTA as one of the worst deals America ever signed, forced a renegotiation rather than a renewal. A new name was required, since selling a rebranded product is easier than admitting you kept most of the old one. CUSMA was signed in late 2018, amended in December 2019 after the U.S. Congress balked at the labour and enforcement provisions, and took effect on July 1, 2020, in the middle of a pandemic that briefly made trade rules feel like the least of anyone’s problems.

The substance is closer to continuity than rupture. Tariff-free trade across the three countries survived mostly intact. What changed was mostly at the margins: tighter in some places, modernized in others, with a handful of genuinely new chapters for things that barely existed as trade issues in 1994, digital commerce chief among them.

Key actors

Three governments, unevenly matched. In Canada, trade policy runs through the ministries responsible for international trade and finance, with Global Affairs Canada doing the legwork and the Prime Minister’s Office weighing in whenever the stakes get large enough, which, with CUSMA, is often. Provinces matter more than the org chart suggests, because dairy, autos, and forestry (the sectors that actually fight over this agreement) are regulated provincially or clustered in a handful of them. Ontario cares about autos in a way Prince Edward Island does not; that asymmetry shapes the politics as much as anything in the text.

Industry groups and unions, especially in the auto sector, punch well above their formal role, because the agreement’s most contested rules are about manufacturing supply chains that cross all three borders multiple times before a finished vehicle rolls off the line. A single car’s parts can cross into and out of the U.S. four or five times during assembly. That’s not a rhetorical flourish. It’s why automotive rules of origin get fought over harder than almost anything else in the text.

What changed from NAFTA

The headline change is automotive: a vehicle now needs 75 percent North American content to qualify for tariff-free treatment, up from 62.5 percent under NAFTA, plus a new requirement that a meaningful share of the vehicle be built by workers earning at least US$16 an hour, a rule aimed squarely at pulling production out of lower-wage Mexican plants and back toward the U.S. and Canada, with mixed results depending on who you ask.

Labour and environmental enforcement got sharper teeth, including a rapid-response mechanism that lets a facility-specific labour complaint trigger an actual investigation rather than disappearing into a diplomatic note. There’s a digital trade chapter (genuinely new, since digital trade barely existed as a category in the NAFTA era) that bars customs duties on electronic products and limits how much governments can force companies to store data locally. State-to-state dispute settlement survived; the investor-state provisions that let companies sue governments directly, which had been one of NAFTA’s more controversial features, did not, not in the form they took before.

For Canada, the sharpest political cost was dairy: expanded American access to a supply-managed market that Canadian farmers had spent decades protecting, which is the kind of concession that shows up in farm country long after the news cycle has moved on. In exchange, Canada held onto its cultural-industries exemption, the carve-out that lets Ottawa protect broadcasting and publishing from being treated as just another traded good, a fight Canada has been having with Washington since long before CUSMA existed.

Numbers

Roughly three-quarters of what Canada sells abroad goes to the United States. That single fact does more to explain Canadian trade policy than almost any other number. It’s why CUSMA’s fine print matters more to Canada’s economy than most trade agreements matter to most countries. Two-way trade across the border runs into the high hundreds of billions annually, and millions of Canadian jobs trace back to it directly or indirectly. Mexico is a smaller, growing piece of the picture, but the U.S. relationship is the one that decides whether the Canadian economy has a good year or a bad one.

Policy stakes

The mechanism that makes 2026 matter is a sunset clause, and it’s a genuinely unusual piece of design: CUSMA runs sixteen years, but six years in, this year, the three governments have to formally affirm they still want it, or the agreement starts sliding toward expiry through a series of annual reviews. NAFTA never had anything like this; it just ran indefinitely until someone decided to blow it up. CUSMA was built to force a re-examination on a fixed schedule instead, on the theory that a periodic gut-check beats letting the rules ossify for a generation. Whether that theory was wise is now being tested for the first time, and the answer isn’t automatically yes.

Beyond the review itself, the recurring irritants haven’t gone anywhere: tariff threats aimed at specific sectors, the long-running softwood lumber dispute that predates CUSMA and will probably outlive it, dairy grievances that never fully go away, and pressure on Canada to fall in line with U.S. trade and investment screening toward third countries. For an economy this dependent on one customer, the question never really changes: how do you keep the access without keeping all of the exposure that comes with it.

Reading list

  • The full text of CUSMA and the official summaries published by Global Affairs Canada
  • The 2019 protocol of amendment and the labour and enforcement provisions
  • Canadian government backgrounders on automotive rules of origin
  • Parliamentary Budget Officer and committee analyses of CUSMA’s economic effects
  • Trade-policy research on the 2026 joint review and sunset clause