The Receipt

Canada’s trade-war retaliatory tools were targeted, effective, and popular. Provincial alcohol bans drove an 81% drop in U.S. alcohol imports. A 25% auto surtax shifted $2.85 billion in vehicle imports away from the United States to Mexico, Japan, and Germany. Buy Canadian procurement rules cut Manitoba’s spending on American firms by 82%. On July 20, 2026, the Trump administration cited all of these measures by name in three presidential proclamations invoking Section 338 of the Smoot-Hawley Tariff Act [1] [2] [3]. It was the first time any president had imposed tariffs under the statute in its 96-year history [4]. The finding: Canada was discriminating against U.S. commerce. The evidence: the retaliatory measures themselves.

The auto and alcohol proclamations rest explicitly on measures Canada adopted after March 2025 as trade-war retaliation, and together they account for virtually all of the roughly $20 billion in imports covered by the package [4]. Section 338 contains no retaliation defence [5]. It does not ask why U.S. commerce receives worse treatment than another country’s. It only asks whether it does. Canada’s retaliatory strategy was built around singling out U.S. commerce. That is precisely the conduct the statute measures.

For eighteen months, Canadians were told the retaliatory tools were working. They were. The alcohol bans emptied shelves of American whiskey while French wine and Mexican tequila stayed put. The auto surtax hit only American-made vehicles while imports from every other country rose. Premiers staked their reputations on these measures. Voters supported them. The data confirmed they were having the intended effect.

Then the data showed up in a presidential proclamation. Not as evidence of Canadian strength, but as evidence of discrimination against the United States.

What the Proclamations Say

The motor vehicles proclamation identifies Canada’s United States Surtax Order, imposed in April 2025 [1]. It documents that the surtax applies only to U.S.-origin vehicles. It notes that U.S. auto imports to Canada fell roughly $5.6 billion while imports from Mexico rose nearly $2 billion and imports from Japan, Korea, and Germany rose between 10 and 14 percent. Other countries filled the gap the surtax created. That gap is the discrimination finding.

The alcohol proclamation identifies the provincial bans that began in March 2025 [2]. All but two provinces and territories halted purchases of U.S. alcoholic beverages without imposing similar restrictions on products from other countries. U.S. alcohol imports fell 81%, from roughly $718 million to $137 million. EU alcohol imports to Canada rose more than $100 million over the same period.

The dairy proclamation is different [3]. It does not cite trade-war retaliation. It cites a pre-existing dispute over how Canada allocates cheese quotas under CUSMA compared with CETA, arguing that U.S. cheese faces more restrictive treatment than European cheese. This is a market-access disagreement that predates the trade war. Its covered imports total roughly $97 million [4].

The dairy exception matters. It demonstrates that the administration believed it had a Section 338 theory against Canada even without the 2025 retaliation. Washington did not need the alcohol bans or the auto surtax to invoke the statute. But those two proclamations account for more than 99% of the covered trade value. The administration chose to build the overwhelming bulk of its package on the retaliatory measures because they were factually cleaner predicates [4].

Why This Law Is Different

For most of the trade war, CUSMA origin was the central compliance question. The IEEPA tariffs exempted CUSMA goods. The Section 122 surcharge that replaced them did the same. Origin certification was the shield, and it held for roughly 85% of Canada-U.S. trade [6].

That shield had already been cracked before July 20. Section 232 auto tariffs, from April 2025, applied to CUSMA-qualifying Canadian vehicles on the non-U.S. content [7]. Later metals tariffs reached CUSMA goods as well. But those breaches were confined to national-security sectors. Importers understood them as exceptions to a general rule.

Section 338 extended the breach. The proclamations state that the tariffs apply to covered goods “regardless of whether a good originates under the USMCA” [8]. The product lists reach electronics, furniture, building materials, clothing, toys, machinery, and cosmetics [9]. The general rule no longer held.

Three structural features made the statute uniquely threatening. It requires no investigation, no national-security finding, and no hearing. The president acts by proclamation alone [5]. The tariffs have no time limit. And if the “discrimination” persists, the statute authorizes escalation to outright exclusion of the offending country’s products [10]. The 50% rate is the tariff ceiling. Exclusion is the step above it.

The timing is important. The Supreme Court struck down the administration’s IEEPA tariffs in February 2026. Section 301 requires a USTR investigation. Section 232 requires a national-security finding. Section 338 requires none of that. After the court closed one door, the administration reached for the statute that had the fewest procedural requirements [5] [11].

The Paradox

Section 338 authorizes tariffs when a country “discriminates in fact” against U.S. commerce [10]. It asks a single factual question: does U.S. commerce receive worse treatment than another country’s? There is no exception for retaliation, no enquiry into justification, no consideration of who started the dispute.

Retaliation, by definition, singles out one country. You hit back at the country that hit you. Canada’s alcohol bans pulled American products and left everyone else’s on the shelves. The auto surtax applied only to American vehicles. Each measure was U.S.-specific because that is what retaliation is. Under Section 338, that specificity is the trigger [5].

The data proving the tools worked is the data proving discrimination. An 81% drop in U.S. alcohol imports, a $2.85 billion shift in vehicle sourcing away from the United States. The proclamations cite those numbers directly. The measures were effective, and the evidence of effectiveness is what fills the finding.

The obvious alternative was to tax everyone equally rather than targeting the United States alone. But that would have required Canada to impose costs on countries that had done nothing to it and to risk breaching its own WTO tariff bindings [5]. Normal trade retaliation works in the opposite direction: obligations are suspended against the offending country, not against everyone. There was no practical version of effective counter-pressure that avoided producing the differential Section 338 measures.

An important qualification: Canada’s retaliation did not cause the administration’s desire to escalate. The administration was already searching for replacement tariff authorities after the courts struck down IEEPA [11]. Canada’s U.S.-specific counter-measures gave Section 338 an unusually direct factual hook. The retaliation supplied the legal vehicle, not the motivation.

The strongest counter-interpretation deserves its full weight. The fact that Washington devoted a new legal mechanism to neutralizing Canada’s retaliatory tools can itself be read as evidence that the tools had leverage. You do not revive a 96-year-old statute to kill something that is not working. If the retaliatory measures eventually brought Washington to the table, Section 338 is not proof that retaliation failed. It is proof that retaliation became costly enough for the other side to build a legal architecture to force its removal. As of August 22, Canada has received nothing in exchange. Negotiations were suspended, and the 50% Section 338 tariffs took effect. The retaliatory tools that were spent outside the deal — the DST rescission and the bulk tariff removal — cannot be re-spent. The leverage was consumed. The return has not materialized.

The Mexico Question

Mexico faced the same initial tariffs from the same president under the same trade agreement. Mexico threatened retaliation in March 2025 and then did not follow through with sustained U.S.-specific counter-measures after Trump exempted USMCA goods [12]. Mexico maintained dialogue, aligned its trade policy with U.S. priorities on China, and did not impose provincial-style alcohol bans or procurement exclusions [13].

Mexico has not faced a Section 338 action [14].

This is a suggestive comparator, not causal proof. Mexico’s bilateral negotiations were further advanced. Mexico had made concessions on migration and security. The manufacturing integration between the two countries is deeper. The Sheinbaum government maintained a lower-temperature political relationship throughout. No authoritative U.S. statement has been located saying Mexico was spared Section 338 because it did not retaliate.

What the record does show: Mexico did not create a documented record of U.S.-specific discrimination. No 81% import drop. No $2.85 billion import diversion. The country that produced those numbers was targeted under a statute that measures exactly those numbers. The country that did not produce them was not.

The Cost of the Clock

In May 2025, the United Kingdom signed the first deal of the trade war. Steel and aluminum tariffs: 25%, locked in under a tariff-rate quota [15]. One month later, Trump doubled the rate to 50% for every other country. The UK was explicitly exempted.

At that point, Canada held four cards the United States had publicly identified as irritants: the Digital Services Tax, the alcohol bans, roughly $60 billion in retaliatory tariffs, and Buy Canadian procurement policies. The ceiling Canada would have been negotiating down from was 25% on steel and aluminum.

The DST was rescinded in June 2025 to restart trade talks that subsequently collapsed. The White House called it a “cave” [16]. Canada received no tariff concession in exchange. The bulk retaliatory tariffs were dropped unilaterally in August 2025 [17]. Roughly $44.7 billion in counter-tariffs removed, again without a reciprocal U.S. reduction. The alcohol bans and procurement exclusions were on the table when talks collapsed. They have not been formally surrendered, but the cards played to reach that table are already spent.

While those cards were being spent, the structural environment worsened. Steel and aluminum tariffs doubled to 50%. The assessment base widened to full customs value. CUSMA was not renewed. Section 338 was created. Each month of delay added either a U.S. escalation or a Canadian concession given away outside the deal.

The anchor shift is documented in the Canadian government’s own public statements. In April 2025, Carney said the system of global trade Canada had relied on since the second world war “is over” [17]. By August 2025, he was framing Canada as having “the best trade deal with the United States” and “better than that of any other country” [17]. The reference point had shifted. Success was no longer measured against where Canada was in January 2025. It was measured against where other countries were now.

Before February 2025, Canadian steel and aluminum entered the United States tariff-free under a full CUSMA exemption. No tariff, no volume quota. The only constraint was a surge-monitoring mechanism that required consultations before any duty could be reimposed [18]. In nearly six years, it was invoked once. The deal that was on the table before talks collapsed would have replaced that with a tariff-rate quota: a per-unit tariff on every tonne within quota, and a higher rate above it. Both the cost and the volume ceiling would be new. Whether those terms return to the table, and at what price, is now an open question.

Manitoba Premier Wab Kinew captured the tension before talks collapsed. “I think we should fight,” he said on August 20 [14]. He called Trump “erratic” and “not to be trusted.” He questioned whether accepting permanent tariffs was worth giving up the leverage. But he also acknowledged he was not at the negotiating table and said he would go along as part of a “Team Canada” approach. On Carney’s ask to restore American alcohol to Manitoba shelves: “I wouldn’t say that he was begging us, but what is a step before begging?” [14]

Kinew’s position rests on a reading of the tariff landscape that treats Section 338 as more of the same. It is not. The 50% that was already on metals was a Section 232 action, global in scope, with CUSMA carve-outs. The 50% under Section 338 was Canada-specific, CUSMA-piercing, backed by a statute with no expiry, and one step short of the statutory authority to exclude Canadian goods entirely. Same number. Different architecture.

Section 338 may not survive in court. Legal scholars have identified serious vulnerabilities: the tariff lists cover products far removed from the alleged discrimination, the 50% rate may exceed the statutory “offset” requirement, and later trade statutes may have implicitly displaced Section 338 [5]. The Supreme Court’s IEEPA ruling demonstrated willingness to constrain novel executive tariff claims. If Section 338 is narrowed or invalidated, the paradox dissolves. The vulnerability exists as an exercise of executive power unless and until courts rule otherwise.

What Would Change This Assessment

This analysis rests on specific, testable claims. It would be weakened or overturned if:

  • The auto and alcohol proclamations are shown to rest on factual findings independent of Canada’s post-March 2025 retaliatory measures. The spine requires that those two proclamations cite trade-war counter-measures as their basis.
  • Mexico is subsequently hit with a Section 338 action despite not sustaining U.S.-specific retaliation. That would break the comparator.
  • A court rules that Section 338 cannot override CUSMA, or that the statute contains an implied retaliation defence. That would mean the structural vulnerability was never enforceable.
  • Canada could have retaliated in a non-discriminatory way that was comparably effective without triggering Section 338. That would mean a practical off-ramp existed.