By Si Thu Naing

August 20, 2026

Late Tuesday night, less than two hours before a midnight deadline that would have triggered 50% tariffs on roughly $20 billion of Canadian goods, U.S. President Donald Trump announced a three-day reprieve — posting that Canada and the United States “have a DEAL,” subject to “the finalization of documents” — with the pause holding only until midnight on Friday.

Prime Minister Mark Carney’s own statement was considerably more cautious: “substantial progress has been made, although there is important work still to be done.”

That rhetorical gap is not diplomatic hedging. It is a sign that Canada’s trade problem is not only about Washington.

One of Donald Trump’s most prominent demands has been that Canadian provinces restock American wine, beer and spirits, pulled from shelves in 2025 in protest against Trump’s tariff war. On Wednesday, Carney personally asked the premiers, on a call, to restock American alcohol, telling them he would absorb any public backlash himself.

Nova Scotia’s Tim Houston agreed outright, calling it an ask that “came directly out of the negotiating table.” British Columbia’s David Eby, who had tied restocking to softwood lumber relief, offered only warmer language about “substantial progress across strategic sectors” without committing to act, while Quebec’s Christine Fréchette said she needed more information before deciding, citing supply management and Quebec’s “cultural exception.” Ontario’s Doug Ford has said he is “more than happy” to restock, but only alongside firm sector protection, not a partial rollback.

On Thursday, Manitoba Premier Wab Kinew came out more forcefully than any other premier on the hazards of not only making concessions pre-emptively to Trump, but of negotiating with this president at all. “You can’t make a good deal with a bad person,” Kinew told reporters. “He’s erratic, he’s irresponsible, and he’s not to be trusted.”

Commentators have mostly described this as ordinary Canadian federalism: the usual friction of a 14-government federation trying to speak with one voice. That is true as far as it goes, but political scientists have a more precise name for what is actually happening. In 1988, Harvard’s Robert Putnam described international negotiations as “two-level games”: a chief negotiator bargains simultaneously at an international table and a domestic one, and any deal reached abroad must still fall within the domestic “win-set,” the range of outcomes the negotiator’s own constituents will actually accept.

Putnam’s “ratification” is a term of art for that domestic acceptance, not the formal legal act of treaty ratification, which in Canada rests with the federal executive alone. Ottawa’s negotiators are not just facing Jamieson Greer in Washington; they are also facing a second table at home, and that table is not one constituency but several: Ontario, British Columbia, and Quebec each hold their own win-set, and satisfying one does not obligate the others.

A second bargaining table, and a familiar way to lose it

For anyone who has sat across an actual bargaining table in the labour movement, that second table looks familiar. Ottawa is behaving like a national union executive trying to sign a master agreement while several of its components (union-within-a-union bodies representing a specific bargaining unit or sector) refuse to lift their own job action until local grievances are addressed.

A national executive cannot force a component to abandon a position touching its own members’ conditions, and a component does not ordinarily surrender its primary leverage against a promise to act later; it holds that leverage until the concession is in hand. Push a settlement through regardless, and the result is a ratification crisis, not unity: components refuse to endorse the deal, and the executive is sent back empty-handed.

Ford and Eby’s refusal to restock ahead of firm sector commitments has the same strategic function, whatever their private reasoning: it keeps leverage in hand until a concession is secured.

Ottawa is closer to that position than it may like to admit.

Liquor distribution in Canada falls under provincial jurisdiction, administered through provincially owned boards such as the LCBO in Ontario, the SAQ in Quebec and the BC Liquor Distribution Branch. Ottawa has told provinces to be ready to restock “as soon as possible” and to drop Buy Canadian procurement preferences, but it cannot deliver provincial compliance on its own signature.

This is settled law, not just convention, and it turns on a distinction separate from Putnam’s: Ottawa alone signs a trade agreement, but implementing any obligation touching provincial jurisdiction is a different legal step it cannot take on the provinces’ behalf.

In the 1937 Labour Conventions case, the Judicial Committee of the Privy Council ruled Ottawa cannot compel provinces to implement treaty obligations touching their jurisdiction, even though only Ottawa can sign the treaty. It can negotiate, but not implement on the provinces’ behalf, and can only persuade, province by province.

The comparison to a union has a real limit. A component’s authority to bargain locally flows from the national body’s own constitution. A province’s authority does not flow from Ottawa at all; it is independent and constitutionally entrenched. Provinces are not components awaiting instructions from head office; they are constitutionally autonomous co-negotiators whose cooperation Ottawa needs to make an international bargain work at home.

The comparison with the ongoing trade negotiations captures the ratification logic and the leverage problem, not the source of provincial authority.

None of this is unprecedented. Ottawa’s “Team Canada” coalition during the first NAFTA renegotiation was, according to a 2026 paper by political scientist Noah Fry in the Canadian Foreign Policy Journal, built as what he calls a “trap” strategy, a spanning coalition of industry, labour and civil society that functioned as both a defence against internal rifts Washington could exploit and an offensive tool in its own right.

The current standoff suggests that such discipline has not been fully reproduced this round, with the qualifier that the good cop/bad cop routine is also a longstanding feature of negotiations.

The comparison also understates the difficulty: it points to an internal coalition failure, not an external one. Ontario’s priorities run through steel and autos. British Columbia’s run through softwood lumber. Quebec, under Premier Christine Fréchette, has held its own line on dairy supply management. A national executive bargaining across components with genuinely different interests faces the same problem: resolving one grievance does not move another any closer to ratifying, the same coalition-management failure Fry’s “trap” strategy was meant to prevent, only a level down.

What labour relations gets right that trade diplomacy has missed

Collective bargaining has spent decades refining ways to manage exactly this: how a central table and local tables move together without either side losing leverage or trust. Three tools translate directly.

First, sub-tables that run alongside the main table. Sector-specific issues are typically negotiated in parallel with a master agreement, not deferred until after it is signed. Canada already has a version of this machinery: the federal-provincial Committee on Trade, or C-Trade, meets quarterly and is, according to political scientist Christopher Kukucha’s research on provincial trade policy, one of the main channels through which Ottawa and the provinces coordinate on trade.

Kukucha’s broader finding is that provinces have shaped trade outcomes for decades not through formal treaty power, which they lack, but through sustained participation in structures such as this.

What is missing under deadline pressure is those standing tables running alongside the main talks, with an understanding that relief gets finalized only once the sector tables affecting holdout provinces are resolved. Right now, that sequencing is inverted: Ottawa is negotiating the umbrella first and treating provincial buy-in as an afterthought.

Second, compensating an unresolved concession rather than simply demanding it. When a national executive cannot get an employer to move on a component’s issue, it negotiates something else internally to offset the loss, a side letter, a targeted fund, a commitment on a related issue. Ottawa has such tools, from softwood-community backstops to industrial adjustment funding, but has not treated that compensation as the price of asking a province to give up leverage early.

Third, treating provincial buy-in as a precondition for signing, not a hoped-for consequence of it. An executive that signs an agreement its components will not ratify has not reached a deal; it has created a second negotiation with its own membership. Ottawa risks the same: a federal-level agreement Washington considers final, while provinces keep treating their liquor boards as leverage because their conditions were never met. That does not end the trade war; it relocates it to intergovernmental relations, when Ottawa most needs the provinces onside.

What the two situations share is narrower than a full equivalence: a negotiator who needs the buy-in of semi-autonomous units, and a temptation to treat their grievances as bargaining chips. Putnam explains why that temptation is dangerous; labour relations shows what it looks like when it goes wrong; and the Labour Conventions case explains why Ottawa has no legal shortcut around it.

The choice in front of Ottawa

Jamieson Greer has already signalled how Washington will read any wobble: in the days before the pause, he told reporters the U.S. would welcome a “more conciliatory approach” from Canada, but that America would do “what’s best for America” regardless. That is what an employer says when it senses a union’s internal coalition is not fully behind its own negotiators: an invitation to fold, aimed at the gap between the table and the membership.

Ottawa’s task before the documents are finalized is not simply to extract a better tariff number. It is to give Ford, Eby and Fréchette a reason to accept and implement whatever gets signed, rather than a reason to keep de-stocking as leverage indefinitely.

A three-day pause is not a settlement; it is a deadline Ottawa has, for now, talked its way past. Labour relations has a name for the result when this goes wrong: a settlement negotiated at the main bargaining table but rejected by the membership. Trade policy is about to learn whether its equivalent failure bears a similar cost.

Si Thu Naing is a public servant in the federal government and an MBA candidate in Sustainable Innovation at the University of Victoria’s Gustavson School of Business. He has served as a shop steward for the Union of National Defence Employees. The views expressed are his own.