Canada's counter-tariffs on C$27.6B of US goods take effect Sep 8 at 15-50%

The Analysis

Canada's counter-tariffs took effect at 12:01 a.m. on 8 September, covering about C$27.6 billion of US imports at rates of 15, 25 and 50 percent. The measure is described as matching American action under Sections 338 and 232 dollar-for-dollar and rate-for-rate. Both halves of that description carry weight, and they are not the same kind of constraint.

The sectors named are steel, dairy, appliances, farm equipment, pulp and paper, and electronics, with most US steel and aluminium duties moving from 25 to 50 percent. Beside that sits a C$7.5 billion package for the firms and workers on the receiving end. That is what the announcement says. What it does not say is which tariff items, at which of the three rates, from which date, on what origin test.

Read the sector list as a press release and the schedule as the law. Nearly every error I expect to see over the next month comes from confusing the two.

Three rates, one envelope, six sector names, one commencement minute. The envelope is the part that gets skipped: C$27.6 billion is a total, not a coverage ratio. Spread evenly across six sectors it would be roughly C$4.6 billion each, and I offer that split only to show the arithmetic, since steel alone would strain it. The list therefore has to be selective inside each of those sectors, and selectivity inside a sector is expressed in tariff items. The phrase about most steel and aluminium duties rising from 25 to 50 percent describes a re-rate of things already listed. A re-rate is not a new list, and the announcement does not answer the transition question for goods ordered at 25 and entered after the change.

Dollar-for-dollar and rate-for-rate are two constraints, and the difference between them decides outcomes. Rate-for-rate is a mirror applied item by item: whatever is charged on the Canadian equivalent going south is charged on this coming north. Dollar-for-dollar is a budget on the aggregate: the whole countermeasure is sized to a number. When the two conflict the budget wins, because a rate cannot be mirrored onto an item that is absent from the schedule. That is the operative premise, and it sits in Ottawa rather than in Washington. A product facing 50 percent in the United States can face nothing here. A product facing 25 percent there can face 50 percent here. The schedule is a selection sized to a total, not a mirror of the American list, and the language of matching has obscured that.

Incidence differs by sector and then again by position in the chain. Steel and aluminium importers face a re-rate on a line that already existed, which is the harshest form, because the old number was already inside a signed quote. Appliances, farm equipment and electronics sit across a three-rate spread, so the question inside those sectors is which item you hold rather than which sector you are in. Dairy, pulp and paper carry their own rate logic and their own list. A US exporter selling into Canada pays nothing directly; its Canadian distributor is the importer of record and will be asking for the difference. A Canadian firm that buys from the United States and also ships north is exposed in both directions on overlapping product, and the two measures do not net each other out.

Size changes the binding constraint rather than the exposure. For a small importer the constraint is cash and timing: duty is paid at the border, per entry, while the support package is a program payment that arrives after an eligibility test. For a larger importer the constraint is contractual, because the duty lands on whoever the terms of sale make the importer of record, and re-quoting is a negotiation with a counterparty who has already priced the old rate into a signed order.

One minute past midnight on 8 September is an entry-time test. Anything entered before pays the old rate, anything entered after pays the new one, and the cargo on the road or on the water at that minute is the cargo nobody re-quoted. The move from 25 to 50 is the sharpest of these, because it shifts a figure that had already been sold to a customer. The C$7.5 billion does not arrive at the border. It arrives later, through programs, to firms that qualify, which leaves a gap between the cash out and the support in, and that gap is where a thin-balance importer fails even when its economics are sound.

The part not said plainly is that the unit of this measure is the tariff item, and neither the sector name nor the American rate tells you anything about it. Coverage of C$27.6 billion spread across six named sectors cannot be comprehensive inside any one of them. Two importers of steel products, facing identical American treatment, can sit at 50 percent and at zero, and the difference is a line in a schedule rather than a difference in their business. The matching language invites the opposite assumption, that exposure can be derived from the American measure by symmetry. It cannot. Symmetry is how the envelope was sized; the schedule is how the envelope was spent. Because rate-for-rate and dollar-for-dollar pull in different directions, the schedule and not the symmetry is the text to read, and the origin test in it decides which goods are inside even before the rate does.

Take a number, with the assumptions marked as mine. Suppose C$2,000,000 of US-origin steel bought in a year, delivered in four lots, re-rated from 25 to 50 percent. Duty at the old rate is C$500,000 a year; at the new rate it is C$1,000,000. The difference is C$500,000 on unchanged volume, C$125,000 per lot, paid at entry. If the goods are sold sixty days after entry, the importer finances that C$125,000 per lot in the meantime, and the support package does not touch that timing. I assume no change in volume, price or sourcing, which is precisely the assumption that will break.

The counterfactual is instructive. Had this been a pure mirror of the American schedule, an importer could look up the US rate and know its own. Because it is a budgeted selection, the lookup has to be done in the Canadian schedule, item by item, and the answer cannot be inferred from symmetry. Had the support been built as a remission at the border, the cash timing would line up with the liability. Built as a program, it lines up with an assessment cycle, and the firms that most need it are the least equipped to survive the wait.

The trap I expect is pricing on a sector average. An average computed across items sitting at 15, 25 and 50 describes no single shipment. The other trap is reading C$27.6 billion as evidence of breadth; it is evidence of selection.

Where the conclusion stops holding matters more than where it holds. If the order keys the rate to the origin of the steel rather than to the country of export, then routing the same coil through a US distributor changes nothing, while a non-US coil shipped out of a US warehouse does not become caught. If a remission or transit provision exists for goods already en route on 8 September, the entry-time reasoning narrows to a defined set of shipments and the rest of the exposure falls away. If the dispute over the underlying American measure goes Canada's way, the countermeasure does not lift by itself; it is a separate instrument with its own review. Strictly speaking, what we have is a rate range, a total and six sector names, plus one minute past midnight. The rest is a schedule we have not read, and until we have read it the only sound instruction is to obtain the order and register every line against it.

↑ Back to the key points

— By Dr. Ingrid Voss