An untested 1930 tariff clause now taxes $20bn of the $451bn Canada ships south

Fifty per cent of $20bn is $10bn, and that is the ceiling on what the Section 338 order can collect in a year if every covered shipment keeps moving at the old volume, which it will not ^[1]. Set against the $451bn of goods Canada sent into the US in 2025 ^[5], that maximum take is about 2.2% ^[2], and the covered slice itself works out to 4.4%, which the Fortune author rounds to roughly 4% ^{[4] [6]}. Canada's answering schedule, which the South China Morning Post reported as running from 15% to 50% ^[13], therefore implies duty of somewhere between $3bn and $10bn on its own $20bn ^[5]. Canada's measures took effect on Aug. 25, 2026 ^[1].

The more interesting term is not the rate but the concessions that stopped existing once the rate held. Ottawa says it had a preliminary deal to bring US steel and aluminium tariffs down from 50% to 25% ^[9]; when that vanished, Canada pulled the reopening of the Keystone oil pipeline deal that Joe Biden cancelled in 2021, which it had put on the table ^[11], and it dropped any willingness to reverse the provinces' removal of US liquor from store shelves, where American spirits sales have already plunged ^[12]. Neither of those two withdrawn items carries a dollar figure in the source material, and both are worth more attention than the duty arithmetic, because a pipeline and a liquor listing are structural while a tariff line is a signature. Note also that Canada is raising its own steel tariff to 50% from 25% ^[13], abandoning on the way up the exact number it wanted Washington to come down to ^[9].

Washington's stated posture, per the SCMP account, is that retaliation will not be tolerated, with officials observing that only two countries, Canada and China, have used counter-tariffs ^[16], and no US official will say how or when Trump responds ^[17]. The day after Canada moved, Trump called it one of the worst countries in the world to deal with, and Mark Carney said Trump was trying to destroy Canada's auto industry ^[20]. The Fortune piece, written by an economist who says he has studied trade policy for 49 years ^[21], makes the Smoot-Hawley comparison explicit: economists generally agree the act's escalating retaliation extended the Great Depression, and US trade fell by two-thirds from 1929 to 1932 ^[19]. That mechanism works through list growth, not through any single 50% rate, which is why the variable to price here is how fast $20bn becomes $40bn.

My read is that at $20bn a side this is price discovery on political pain rather than macroeconomic damage, and the thing that would break that read is a plant-level response landing before November: a Louisville appliance line or a Wisconsin dairy cutting output because Canadian orders stopped ^[8]. The material supplies two $20bn headline figures and no company's Canadian revenue, so the defensible position is that the cost is real, capped by the size of the lists, and not yet measurable at the firm level ^{[4] [7]}.