Washington Found a 1930 Law Nobody Had Ever Used — and Put 50% Tariffs on Canada With It

President Trump signed three separate Section 338 proclamations on 20 July 2026, covering alcoholic beverages, dairy and 'motor vehicles'. The statute requires a 30-day wait, so the duties were set for 12:01 a.m. ET on 19 August; the deadline was pushed three days while negotiators tried to close a deal, and the tariffs took effect early on Saturday 22 August after those talks failed. The Holland & Knight review of the proclamations found the 'motor vehicles' list to be the broadest of the three — 439 tariff lines across 18 pages — and noted that despite the title, none of them sit in the vehicles chapter of the US tariff schedule: the covered goods are agricultural products, textiles, wood, cement, furniture and machinery. The alcohol list covers beer, wine, cider and spirits; the dairy list runs to roughly 52 classifications including milk, cream, whey, lactose and casein. Two exclusions matter. Goods already carrying Section 232 national-security duties — steel, aluminium, copper, passenger and commercial vehicles, some wood, semiconductors and patented pharmaceuticals — are carved out. And, in the firm's words, 'USMCA origination does not exempt covered goods from the Section 338 duty', which is the departure: qualifying as a North American free-trade good has been the standard shield against every other tariff track, and here it is explicitly not one. CSIS put the coverage at 4.9% of total US imports from Canada in 2024. The Associated Press, reporting the day it landed, described a border that carries about $880bn of goods and services a year, roughly $2bn of goods a day and nearly 330,000 people. Canadian Prime Minister Mark Carney said 'Canada will match those tariffs dollar for dollar to protect our workers and businesses', with the countermeasures dated 8 September.

US Trade Representative Jamieson Greer's statement of 20 July made the official case plainly: Canada 'unlike other partners and allies, continues to retaliate against the United States', having 'taken U.S. alcohol products off Canadian shelves, given better market access to dairy products from the European Union, and… put a cap on U.S. vehicle exports to Canada from companies reshoring to the United States.' That is a coherent Section 338 theory — the statute is built for exactly the case where a country treats American goods worse than another country's. The timing points at a second driver. CSIS notes that the Supreme Court's rejection of IEEPA tariffs and the expiry of the Section 122 duties 'created openings for this alternative mechanism' — in other words, the authority was reached for partly because the earlier ones were gone. Section 338 is unusually convenient for that purpose: compared with Section 301 it has, in the CSIS reading, 'a less-clear investigative requirement, and a shorter timeline', and it caps out at 50%, which is where the rate landed. Holland & Knight read the whole thing as a negotiating position — a '50 percent opening bid' inside the USMCA review, with the 30-day statutory wait functioning as 'a deadline to extract concessions'. On that reading the tariffs were designed to be traded away, and the talks collapsing was the failure case, not the plan. Likely rather than confirmed: the discrimination claims are stated on the record and the legal-vacuum timing is documented, but no official has said the second thing out loud, and the internal weighting between the two is inference.

On the administration's own account the beneficiaries are named US sectors: distillers whose product came off provincial shelves, dairy producers it says are ranked below European ones, and carmakers reshoring to the US who then hit a Canadian export cap. CSIS is sceptical about at least one of those claims, noting that the dairy allegations sit awkwardly against data showing 'U.S. dairy exports to Canada remain far below the quota threshold' — if American exporters are not filling the access they already have, a tariff does not obviously create a customer. The more reliable winner is precedential. Holland & Knight expect litigation in the US Court of International Trade, and CSIS points out that no court has ever interpreted Section 338 — so until one does, the authority is live, untested, and available against anyone. CSIS reads part of the motivation as signalling to third countries, Brazil among them, what retaliating against US tariffs will cost. Uncertain, because 'who benefits' here rests on a counterfactual nobody can settle yet: if the tariffs are traded away in a USMCA deal, the beneficiary was the negotiating position; if they survive a court challenge, the beneficiary is every future use of a statute that ignores free-trade status.

The asymmetry is the whole problem. The Associated Press's figure — nearly 72% of Canadian goods exports go to the US — means Canada has no substitute market of comparable size to absorb a 50% duty, which is why Ottawa's answer is a matching tariff rather than a redirection. On the American side the cost lands on importers and the buyers behind them: the covered list is heavy in machinery and industrial inputs, not consumer luxuries, so the tariff is paid partly by US manufacturers using Canadian components. Canada's countermeasures, dated 8 September, cover steel, dairy, appliances, agricultural machinery, pulp and paper, and electronics — a list built to be felt in US export states. The structural loser is USMCA itself. Holland & Knight state flatly that USMCA origination does not exempt covered goods; CSIS calls the result 'unprecedented uncertainty regarding USMCA compliance protections' and warns about 'interconnected auto supply chains'. Formal USMCA renegotiation has begun with Mexico and has not begun with Canada. Confirmed, because these are documented figures and dated measures, not projections — the only open question is size, not direction.