Isaac Jonas
Streetwise Economics, Abbotsford, British Columbia V2T, Canada
E-mail address: isacjonasi@gmail.com
Published on 31 August 2026
A B S T R A C T
On 22 August 2026 the United States imposed a 50 per cent tariff on approximately US$20.4 billion of Canadian goods, invoking Section 338 of the Tariff Act of 1930 for the first time since the statute was enacted. Canada replied with counter-tariffs on C$27.6 billion of American goods, tiered at 15, 25 and 50 per cent, effective 8 September 2026. Work published since has concentrated on the size of the shock. This paper asks a narrower question of one province: not how large the cost is, but who holds it.Three results follow. First, the incidence parameter in general use is an average and is probably the wrong one for British Columbia. The widely cited finding that foreign exporters absorbed about 47 per cent of the 2025 American tariff shock is an average across the fifty largest United States trading partners; the same study reports absorption near 70 per cent where the exporter is difficult to replace and near zero where it is not. Second, applying a Policy Analysis Matrix framework adapted for tariff incidence, approximately C$1.85 billion a year is borne inside British Columbia across both directions of trade, within a range of C$1.36 to C$2.56 billion. Third, that burden is concentrated by composition rather than by dependence: wood products carry 5.6 times their national weight in the provincial export basket and about half of the export-side figure, while energy, the province’s largest export category, is reached by none of the measures.The paper is a scenario decomposition under assumptions stated in the open and reproducible in a companion workbook, not a forecast. Where an underlying parameter is contested, results are reported as ranges. Section 10 states the limitations and Section 11 corrects six errors in the author’s own earlier papers.
1. Introduction
This paper is not a prediction. It does not forecast output, prices or employment, and nothing in it says what British Columbia’s economy will do next year. It asks a narrower question: given measures of known scope and rate, and given a stated assumption about who absorbs the cost, where does the money actually come from?
That question is separable from the question of magnitude, and careful work on magnitude already exists. Tombe has estimated national and provincial job effects using input–output methods [18]. The Canada West Foundation has mapped provincial exposure [17]. The Bank of Canada has measured how much of a tariff reaches consumer prices, in both the 2018 episode and the 2025 countermeasures [19,20]. This paper takes those findings as given and adds one thing: an explicit incidence assumption carried through to a decomposition across the groups that bear the cost, for one province, in both directions of trade.
The choice of British Columbia is not incidental. On the conventional measure of exposure — the share of exports going to the United States — the province is among the least dependent in the country. On the measures actually imposed in August 2026 it is the most exposed. Reconciling those two facts is the substantive contribution of Sections 3 and 4, and it turns on what the province sells rather than on how much of it goes south.
The framework is the Policy Analysis Matrix of Monke and Pearson [14], which compares what an activity earns and pays at observed prices against what it would earn and pay at undistorted prices. The gap is the transfer that policy creates. The framework is used here for the transparency of its incidence structure rather than for precedent; a reader who prefers a computable general equilibrium approach for questions of aggregate magnitude is right to prefer it, and this paper does not attempt those questions.
2. The measures in force
Commentary has generally treated the American measures as a single escalating tariff. They are three programmes under two statutes, with different products, dates and resolution paths, and a firm’s exposure depends entirely on which one reaches it. Table 1 separates them.
Table 1
The three American tariff authorities applying to Canadian goods. Sources: [3,7,9,13].
| Section 338 | Section 232 — metals, autos | Section 232 — timber and wood | |
|---|---|---|---|
| Statute | Tariff Act of 1930, 19 U.S.C. 1338 | Trade Expansion Act 1962, s.232 | Trade Expansion Act 1962, s.232 |
| Predicate | Finding of discrimination against U.S. commerce | National security | National security |
| Scope | Canada-specific | Global | Global |
| Products | Dairy; alcohol; motor vehicles and associated goods; certain wood and paper lines | Steel and aluminium; motor vehicles and parts | Softwood timber and lumber; upholstered wooden furniture; kitchen cabinets and vanities |
| Rate | 50 per cent | Metals 50 per cent; vehicles 25 per cent | Softwood 10 per cent; furniture and cabinets 25 per cent |
| In force | 22 August 2026 | Metals 4 June 2025; vehicles 3 April 2025 | 14 October 2025 |
| Dated risk | None scheduled | None scheduled | Furniture rises to 30 and cabinets to 50 per cent on 1 January 2027 |
| Legal status | First presidential use; no case law | Settled authority | Settled authority |
Two points in Table 1 are routinely misreported. The first is the date of 1 January 2027, which belongs to the furniture and cabinet measure alone. It has circulated as a deadline attached to the Canada dispute. It is not: it is a deferred rate increase under a global national-security authority, moved from 1 January 2026 by proclamation, and unaffected by any settlement of the Section 338 measures [7].
The second is the legal standing of Section 338 itself, which cuts both ways. Subsection (d) caps additional duties at 50 per cent ad valorem, so the announced rate is at the statutory ceiling. Subsection (b) separately authorises exclusion of the offending products altogether where a country maintains its discriminations [13]. The statute has never been used before and there is no case law, so any ruling resets the analysis in either direction.
2.1. Canada’s countermeasures
Canada’s measures will take effect at 12:01 a.m. on 8 September 2026 across C$27.6 billion of American goods [1,2]. Three features matter for what follows. The rates are tiered rather than mirrored — 15, 25 and 50 per cent, matched to the American rate on the corresponding good — with steel, aluminium, furniture, clothing, dairy, pulp and paper and plastics at 50 per cent. Part of the schedule is escalation rather than new coverage: existing counter-tariffs on steel and aluminium rise from 25 to 50 per cent [24], so the incremental burden on those lines is 25 points, not 50, and the headline coverage figure overstates the incremental cost. And goods in transit on 8 September are exempt, with liability turning on origin — the measures apply only to goods eligible to be marked as a good of the United States. Both facts are actionable and Section 8 returns to them.
Alongside the tariffs, Ottawa announced C$7.5 billion in new support on top of nearly C$25 billion already provided: C$3.5 billion in rapid-response worker measures, C$2 billion for diversification capital projects and C$1.5 billion for small and medium enterprises [1].
3. Why British Columbia is different
British Columbia sent C$27.9 billion of goods to the United States in 2025, or 51.0 per cent of its total exports, against 72.5 per cent for Canada as a whole [4,11]. The provincial figure is BC Stats origin-of-export data and the national figure is customs-basis data; the two series are not identical and the comparison is indicative rather than exact.
On the measures that matter for these particular tariffs, the province is the most exposed in the country. The Canada West Foundation puts about 14 per cent of British Columbia’s American-bound exports within scope of the new measures, against roughly 5 per cent nationally — the highest concentration of any province [17].
The mechanism is composition, not intensity. British Columbia is not hit harder because it sells more to the United States. It is hit harder because of what it sells.
Fig. 1. British Columbia’s export composition against Canada’s, share of merchandise exports to the United States, 2025. Five categories carry more weight in the provincial basket than in the national one; wood products carry 5.6 times their national weight. Sources: [4,11].
Wood products are 17.8 per cent of what British Columbia sends south against 3.2 per cent nationally, a ratio of 5.6. Fish and seafood, metallic minerals and the residual commodities category also carry two to three times their national weight. Energy and machinery — the two categories that dominate the national basket and are reached by none of these measures — are underweight in the province.
That single fact organises the rest of the paper. The forestry provisions bundled into the alcohol proclamation land disproportionately here. The Section 232 softwood measure lands here. The furniture and cabinet increase due in January lands here. Agriculture compounds it: C$4.456 billion of British Columbia’s C$5.89 billion in agrifood and seafood exports go to the United States, or 75.3 per cent, far above the provincial average across all goods [12].
Fig. 2. What the Section 338 measures reach, by commodity. Roughly 86 per cent of British Columbia’s exports to the United States are reached by none of the measures, including energy, the largest single category. Sources: [11,17]; allocation as described in Section 4.
The uncovered majority in Fig. 2 is the more important half of the chart. Energy is the province’s largest single export to the United States at C$6.2 billion and no measure touches it. Machinery and equipment at C$5.8 billion is largely untouched. Roughly 86 per cent of what British Columbia sends south is reached by none of the measures, which means the shock is narrow and deep rather than broad — a distinction with direct consequences for how relief should be designed, discussed in Section 8.
Energy’s exclusion is the most informative fact on the lists. Energy is 22.1 per cent of what the province sends south and 31.8 per cent of what Canada does, with 88.0 per cent of it going to the United States [5]. A 50 per cent tariff on Canadian crude would raise American refinery costs and pump prices directly. The exclusion is therefore leverage that has not been used, and it is the difference between a serious sectoral shock and a macroeconomic event.
4. Method
An honest statement of scope comes first. A complete Policy Analysis Matrix requires firm-level cost structures — the split between tradable inputs and domestic factors within each activity. Statistics Canada does not publish that split at the granularity these tariff lines require. What is applied here is the framework’s incidence logic rather than a fully populated matrix, and the resulting figures are transfers under stated assumptions rather than welfare estimates.
Six steps take provincial trade data to a burden estimate. Each is reproducible from the companion workbook, and each is only as good as the step above it.
Table 2
The export-side calculation. Every step is a formula on the Calculator sheet of the companion workbook.
| Step | Value | Basis | |
|---|---|---|---|
| 1 | B.C. merchandise exports to the United States | C$27,916m | BC Stats, 2025 [11] |
| 2 | Share covered by the Section 338 measures | 14 per cent | Canada West Foundation [17] |
| 3 | Statutory tariff rate | 50 per cent | 19 U.S.C. 1338(d) [13] |
| 4 | Add Section 232 softwood on B.C. softwood lumber | 10% on C$3,300m | Timber proclamation [7] |
| 5 | Share the exporter absorbs rather than passes on | 60 per cent | Assumption, argued in Section 5 |
| 6 | Split of the producer burden, margins to labour | 70 / 30 | Assumption; no published estimate |
Step 2 is another author’s number and its allocation across sectors is this paper’s. The Canada West Foundation’s 14 per cent is a share of British Columbia’s American-bound exports as a whole and is not published by commodity. Distributing it across agriculture and food, wood products, and pulp and paper — and excluding machinery — is a construction, and it drives the entire sector table.
One constraint had to be imposed, and finding it changed the answer. A straight pro-rata split gives wood products C$2,112 million of covered exports. British Columbia’s non-softwood wood exports are only C$1,683 million. The split was therefore placing softwood lumber on the Section 338 lists, where it does not belong: softwood is covered by the separate Section 232 measure. Capping wood at its non-softwood base and reallocating the residual across agriculture and pulp and paper corrects this.
The honest test of an allocation is the coverage rate it implies. The capped split assumes 33.8 per cent of the province’s wood products, 52.5 per cent of its agriculture and food and 52.5 per cent of its pulp and paper sit on the Section 338 lists. The wood figure is an upper bound by construction — it is every non-softwood wood dollar the province sends south. Whether the agrifood figure is credible is a fair question, and a reader who thinks it too high should lower it; the workbook exposes each base in a single cell and reports the uncapped variant beside it.
Two channels are never added together. A producer that holds its price loses volume; a producer that cuts its price keeps volume and loses margin. These are alternative descriptions of the same shock, not components to be summed. Converting lost volume into a dollar loss would require a profit-margin assumption that is not published at this level of detail, so no combined total is reported anywhere in this paper.
5. Who absorbs a tariff
When the United States puts a 50 per cent tariff on a Canadian good, somebody pays it. If the Canadian exporter holds its price, the American importer pays and passes on what it can. If the exporter cuts its price to keep the business, the Canadian producer pays. In practice the burden splits, and where it splits is an empirical question.
The figure in general circulation comes from a CEPR analysis of the 2025 American tariff shock, which estimates that foreign exporters absorbed roughly 47 per cent of it — United States incidence of about 53 per cent — within a reported range of 40 to 50 per cent [15]. It is a careful study: unit-value regressions on ten-digit monthly data covering the fifty largest American trading partners and more than 95 per cent of United States goods imports.
The headline is an average, and the variation underneath it is what matters here. Absorption rises with the exporter’s share of the American market. Dominant suppliers — those an American buyer cannot easily replace — absorbed about 70 per cent of the tariff. Exporters holding three or four per cent of the market passed it through entirely, absorbing essentially none.
Fig. 3. Absorption and the exporter’s share of the American market. The 47 and 70 per cent values are published [15]; the 60 per cent central case adopted here is this paper’s assumption, argued in Section 5 and changeable in the workbook. The line is illustrative, not fitted.
The reason is straightforward. A seller’s ability to pass on a tax depends on how easily the buyer can go elsewhere. Where there are ten interchangeable suppliers, the buyer switches and the tariffed seller either matches the others’ delivered price, eating the duty, or loses the order. Where there is effectively one supplier, the buyer has to pay. The residual demand curve facing a dominant supplier is flatter, and a flatter curve means more of the tax lands on the seller.
British Columbia appears to sit toward the concentrated end of that spectrum. Canada is a large supplier to the American market in several of the categories the province is weighted toward, and wood products are the clearest case: alternative sources at comparable volume are limited, which is part of why American homebuilding interests have argued against these duties rather than for them.
On that reasoning this paper adopts 60 per cent as the central case, between CEPR’s 47 per cent all-partner average and its 70 per cent dominant-supplier figure. This is an interpolation, not an estimate. Nothing has been estimated here; a number has been chosen and a reason given for choosing it. A reader who finds the reasoning unpersuasive should use 47 per cent, which is the published figure, and every table in this paper reports that column. The workbook recalculates on any value.
The assumption is testable. Absorption is observable in export unit values: if British Columbia exporters are cutting prices to hold volume, unit values in the covered categories should fall after 22 August; if they are holding price and losing sales, volumes should fall and unit values should not. That data will exist by November.
6. Results
Applying 50 per cent to the covered provincial exports gives a gross Section 338 liability of about C$1.95 billion a year. The Section 232 softwood measure adds C$330 million on the province’s C$3.3 billion of softwood lumber exports. Total gross liability on the export side is about C$2.28 billion before any quantity response. That is the bill; who pays it depends on absorption. Fig. 4 and Table 3 report the range across three scenarios rather than defending a single point.
Fig. 4. The decomposition across three scenarios. Export side: covered B.C. exports at 50 per cent, plus the Section 232 softwood measure, at exporter absorption of 47, 60 and 70 per cent [15]. Import side: B.C.’s share of the C$27.6 billion countermeasure schedule [2] at 15, 25 and 50 per cent, split 75/25 between firms and households on Cavallo et al. [20].
Table 3
Annual incidence of all measures on British Columbia. Author’s calculation. The firm–worker split within the producer burden is a stated 70/30 assumption. Import-side rows use countermeasure rates of 15, 25 and 50 per cent respectively.
| Borne by | Low (47%) | Central (60%) | High (70%) |
|---|---|---|---|
| B.C. exporting firms — margin | C$751m | C$959m | C$1,119m |
| B.C. workers — hours and shifts | C$322m | C$411m | C$480m |
| B.C. producers, total | C$1,074m | C$1,370m | C$1,599m |
| U.S. importers and final buyers | C$1,210m | C$914m | C$685m |
| B.C. firms buying U.S. inputs | C$216m | C$361m | C$721m |
| B.C. households, through prices | C$72m | C$120m | C$240m |
| Total borne inside British Columbia | C$1,362m | C$1,851m | C$2,560m |
The firm–worker split is the least supported number in this paper. No published Canadian estimate decomposes exporter absorption into margin compression and labour adjustment. The 70/30 allocation is informed by observed dependence ratios — 76.4 per cent of jobs in automobile and light-duty vehicle manufacturing, 77.6 per cent in aluminium production and 67.0 per cent in iron and steel mills exist because of American demand [8] — which imply that labour adjustment is a material rather than residual channel. It remains an assumption, and a reader who prefers 60/40 or 80/20 can substitute it directly; the workbook reports the sensitivity.
This paper does not estimate provincial job losses, and the omission is deliberate. Tombe has published those figures using input–output methods — 87,200 jobs at risk nationally against about C$28 billion of covered goods, of which roughly 11,000 in British Columbia [18]. That work is done properly and duplicating it with a weaker method would add nothing.
Fig. 5. The burden on British Columbia producers by sector, central case, export side only. Wood bears both a Section 338 allocation and the Section 232 softwood measure, which is why it leads. Trade-remedy duties on softwood are separate and are not included — see Section 6.
Three sectors carry the export-side burden. Wood products bear about C$703 million, or 51 per cent of the total, because they are reached by the Section 338 wood and paper lines and the Section 232 softwood measure at the same time. Agriculture and food bear about C$527 million and pulp and paper about C$141 million. Energy, the province’s largest export category, does not appear at all, and neither does machinery.
A further burden is not counted anywhere above, and a British Columbia mill pays it every day. Canadian softwood lumber also faces antidumping and countervailing duties. The rate currently being collected is the sixth administrative review’s final result of 8 August 2025: 35.19 per cent combined for non-selected respondents. The seventh review’s post-preliminary result of 30 June 2026 is 25.18 per cent and is not yet final [16]. With the 10 per cent Section 232 measure stacked on top, the combined burden on softwood is therefore roughly 45 per cent at the rate now in force, or about 35 per cent if the pending review is confirmed. Either figure predates the August measures entirely and would survive any settlement of them. Neither is included in the estimates above, which cover only the 2026 measures.
7. The other direction: the 8 September countermeasures
Canada’s countermeasures are analytically different from the American ones, and the difference is routinely reversed in public discussion. A tariff levied on imported American steel, appliances, agricultural equipment, pulp and paper and electronics is collected from Canadian importers of record and passed to Canadian buyers. The first-round burden is domestic. That is not an argument against imposing it — a deterrent that costs the other side nothing does not deter — but the deterrent is bought with domestic money, and the composition of the list determines the price.
British Columbia imported about C$25.2 billion from the United States in 2024, or 34.4 per cent of its C$73.2 billion of total merchandise imports. The 8 September schedule covers C$27.6 billion of the C$361.7 billion Canada buys from the United States, or 7.63 per cent. Applying that national coverage share to the provincial import figure and a blended 25 per cent rate gives a gross cost of about C$481 million a year, split 75/25 between firms and households on the lower of the two Bank of Canada pass-through estimates.
Four warnings apply to that number, and all of them point the same way. The tier mix is unknown: the schedule publishes tariff items, not covered value by tier, so the blended rate is an assumption bounded at 15 and 50 per cent. Part of the C$27.6 billion is a rate increase rather than new coverage, so the incremental cost is smaller than the headline. Province-of-clearance accounting inflates British Columbia’s recorded imports, since goods clearing customs in the province may be consumed elsewhere. And provincial imports are not published by commodity. The import-side estimate is an upper bound and is weaker than the export-side estimate.
On the firm–household split the two Bank of Canada studies disagree and the lower one is used here. Cavallo and co-authors, examining daily posted prices at seven major retailers during the 2025 countermeasures, put consumer pass-through at about 25 per cent [20]. Lam, using synthetic control on the 2018 episode, puts the share of the tariff reaching Canadian consumer prices nearer 60 per cent [19]. There is a further mismatch that should be stated: Cavallo’s coefficient is estimated on consumer goods at retail, whereas the 8 September list is weighted toward capital goods and intermediate inputs, which mostly do not pass through a retailer at all. Applying a retail coefficient to a schedule of capital goods probably understates the eventual burden.
A class of firms pays on both sides of the ledger at once. A Fraser Valley fabricator shipping into the American market while buying American plate steel faces the Section 338 or Section 232 measures on its output and the countermeasures on its inputs simultaneously. No estimate here or elsewhere counts these firms separately, because Statistics Canada does not publish the overlap between exporters and importers, although its business survey frame appears to permit identifying it.
Because the list is weighted toward capital goods and intermediates, monitoring the consumer price index will not reveal this. A farmer replacing a combine, a builder buying appliances and a fabricator buying steel each pay more, and the cost propagates into farm, construction and manufacturing cost structures with a lag.
8. Implications, by group
Advice in this area often stops at “diversify” or “find new markets”, which is sound in the long run and hard to act on in a quarter. What follows is organised by group, with the economics that makes each item work, or the reason it may not.
8.1. Exporting firms
Exposure is a tariff-line fact, not a category one. The 439-subheading proclamation is titled for motor vehicles but spans a wide range of agricultural and manufactured goods, and the 63-subheading proclamation bundles alcohol with wood, paper and hockey equipment. Press summaries have varied. The only reliable check is a firm’s own HS codes against the published lists.
Whether a sale is priced delivered or at the border determines who is legally liable for the duty, which is a different question from who bears it economically. The two can be separated contractually, and a duty-sharing clause is negotiable.
Differentiation is what buys pricing power, and Section 5’s logic runs in both directions. A supplier the buyer cannot replace absorbs more of the tariff but keeps the order; a supplier who is one of ten loses the order outright. Diversification is real but should be costed as a multi-year investment: market entry carries fixed costs — certification, distribution, standards, relationships — that are sunk before the first sale, which is why “sell elsewhere” is not a quarterly response.
8.2. Firms exposed in both directions
Finance Canada operates a remission process for tariffs on certain American goods, designed for the case where an input has no available domestic or third-country substitute. Firms that can document the absence of an alternative source have a route to relief, and the documentation is better started before the first invoice than after it.
Origin, not supplier, triggers the countermeasure. Re-sourcing where an input is made moves it outside the schedule entirely; changing distributor does not. The in-transit exemption on 8 September is a one-time timing benefit, and duty drawback and duties-relief provisions can recover tariffs paid on inputs that are subsequently re-exported.
8.3. Workers and the support programmes
Where a large share of a sector’s jobs exist because of American demand, a lost order does not shave margins evenly across a workforce. It removes shifts. That is why the labour channel in Table 3 is a material share rather than a rounding error, and why the design of the C$3.5 billion in rapid-response measures matters.
Retention beats replacement when a shock is temporary, and the reverse when it is not. A worker and an employer who have invested in each other hold match-specific capital that a layoff destroys and that is expensive to rebuild. Subsidising the employment relationship through reduced hours preserves that value when the shock is expected to pass; when it is permanent, the same policy holds workers in jobs that are not returning, and retraining and mobility support does more. The programmes as announced do not make that judgement in public, and the judgement is the design question.
8.4. Government
Relief should follow the shape of the shock. Where roughly 86 per cent of provincial exports to the United States are untouched and half the export burden sits in one sector, support delivered as general liquidity spreads money across an economy that is largely not being tariffed. Diversification funding does more per dollar aimed at the specific fixed costs that block market entry than as general credit.
Softwood is a separate file and needs separate advocacy. The combined trade-remedy and Section 232 burden predates this dispute and would survive its settlement. A province whose forest sector supports close to 100,000 jobs and contributes about C$13 billion a year [23] has a strong claim to representation on that file with its own timetable.
One of the three American grievances names a provincial decision: the alcohol proclamation’s stated predicate is that provincial liquor boards halted purchases of American product without applying equivalent restrictions to other suppliers. That was a sub-national decision now carrying a national cost, which is a fact about federalism worth recording whatever one concludes about the decision itself.
Finally, Statistics Canada should publish the exporter–importer overlap. Doing so would allow direct measurement of two-way tariff exposure rather than the qualitative description this paper is limited to, and it would use data already collected.
9. What to watch
The estimates here rest on assumptions that observable data will shortly test. Table 4 sets out when.
Table 4
Dates on which the assumptions in this paper become testable. Sources: as cited throughout.
| Date | Event | Why it matters |
|---|---|---|
| 2 September 2026 | Bank of Canada policy rate decision | First decision after the tariffs took effect, against a second quarter that grew 3.3 per cent annualised [21] |
| 4 September 2026 | Labour Force Survey, August | First labour data covering the post-22 August period — about two weeks of exposure. An early signal only |
| 8 September 2026 | Canadian countermeasures take effect | The date the domestic cost begins |
| October–November 2026 | Export unit values in the covered categories | Settles the absorption question in Section 5 empirically rather than by assumption |
| Autumn 2026 | Softwood administrative review, final results | The post-preliminary rate of 25.18 per cent for non-selected respondents is not yet final [16] |
| 1 January 2027 | Section 232 furniture and cabinet increase | Rates rise to 30 and 50 per cent unless deferred again. Unaffected by the Section 338 dispute |
| Ongoing | Litigation over Section 338 | A statute never before invoked, now the basis for US$20 billion of tariffs. Any ruling resets the analysis |
10. Limitations
These are stated in order of how much they should concern a reader.
The sector allocation is a judgement. The Canada West Foundation’s coverage figure is published for the province as a whole, not by commodity; distributing it across three categories and excluding machinery is this paper’s construction and drives the entire sector table.
The 60 per cent absorption central case is an interpolation. CEPR’s 47 per cent average and 70 per cent dominant-supplier figure are published; the value between them is not. The argument for sitting above the average is defensible; the specific number is a choice.
The import side is an upper bound and is weaker than the export side, for the four compounding reasons given in Section 7.
The firm–worker split is an assumption, not an estimate, and no published Canadian evidence decomposes exporter absorption into margin and labour channels.
The base is pre-tariff. Coverage values rest on 2024 and 2025 flows. Canadian merchandise exports to the United States fell 5.3 per cent in 2025 [4], and trade in the covered categories is contracting faster than the aggregate, so gross liabilities are ceilings on an already-reduced base.
Two measures reaching British Columbia are not estimated at all. The Section 232 metals measure and the January 2027 furniture and cabinet increase both apply here, but provincial exports of those specific lines are not published separately. Both are omitted from the dollar figures, which therefore understate total exposure.
The elasticity is conventional, not estimated: the volume illustration uses a value of −1.0, and sector-specific elasticities would be preferable and are not published at this granularity.
Finally, the Policy Analysis Matrix is applied outside its customary domain, and its use here rests on the transparency of its incidence structure rather than on precedent.
11. Conclusion
The size of the 2026 tariff shock has been estimated carefully by several authors. This paper asks who holds the cost inside one province, and reaches approximately C$1.85 billion a year across both directions of trade, within a range of C$1.36 to C$2.56 billion, concentrated in wood products, agriculture and food, and pulp and paper.
Two findings travel beyond British Columbia. The first is that provincial composition rather than provincial dependence drives exposure to measures of this kind, which means the conventional export-share metric ranks provinces incorrectly for this purpose. The second is that the all-partner absorption average now in general use conceals the variation that matters at provincial level, and applying it uniformly produces a distribution the underlying study does not support.
The estimates rest on assumptions that are stated, bounded and testable, and export unit-value data will begin to settle the central one by November. The author intends to publish that answer whichever way it falls.
References
[1] Department of Finance Canada, Canada announces targeted countermeasures and substantive support for workers and businesses in response to U.S. tariffs, 22 August 2026. canada.ca
[2] Department of Finance Canada, List of products from the United States subject to counter-tariffs effective September 8, 2026, list updated 26 August 2026. canada.ca
[3] White & Case LLP, Trump administration imposes 50% tariffs on certain Canadian products in first use of Section 338, 2026. Proclamation texts at federalregister.gov.
[4] Global Affairs Canada, Office of the Chief Economist, Highlights of Canada’s merchandise trade performance — 2025 update. Customs basis, from Statistics Canada Table 12-10-0171-01.
[5] Statistics Canada, Focus on Canada and the United States: trade, 2025. Reference year 2024, customs basis, domestic exports.
[6] Statistics Canada, Canadian international merchandise trade, December 2025, The Daily, 19 February 2026. Balance-of-payments basis.
[7] Wiley Rein LLP, White House imposes Section 232 tariffs on imports of timber, lumber and their derivative products, 2025. Effective 14 October 2025; increase for furniture and cabinets deferred to 1 January 2027 by proclamation of 31 December 2025.
[8] Statistics Canada, Value added and job creation associated with Canadian manufacturing exports to the United States, Catalogue 13-605-X, reference year 2024.
[9] Blake, Cassels & Graydon LLP, U.S.–Canada tariffs: timeline of key dates and documents, 2026.
[10] Bank of Canada, Monetary Policy Report, April 2026 — in focus: assessing the impact of US trade restrictions.
[11] BC Stats / Province of British Columbia, Annual B.C. exports, released 4 August 2026; British Columbia and U.S. tariffs; trade data.
[12] B.C. Ministry of Agriculture, Sector snapshot 2024: international trade, 2025.
[13] 19 U.S. Code § 1338 — Discrimination by foreign countries, subsections (b), (d), (f). Cornell Legal Information Institute.
[14] E.A. Monke, S.R. Pearson, The Policy Analysis Matrix for Agricultural Development, Cornell University Press, Ithaca, 1989.
[15] CEPR / VoxEU, Foreign exporters absorbed nearly half of the 2025 US tariff shock, 2026. Unit-value regressions, HTS 10-digit monthly data, September 2023 – January 2026.
[16] Global Affairs Canada, Softwood lumber — recent developments, 2026. Seventh administrative review, post-preliminary results, 30 June 2026.
[17] Canada West Foundation, New U.S. tariffs hit B.C. hard, largely spare the Prairies, Trade and Economic Policy Team, 30 July 2026. National comparison drawn from RBC Economics, 24 July 2026.
[18] T. Tombe, New U.S. tariffs on $28 billion in Canadian goods put 87,200 jobs at risk, The Hub, 26 August 2026. Supersedes the 90,000 estimate of 20 August 2026.
[19] A. Lam, Assessing tariff pass-through to consumer prices in Canada: lessons from 2018, Bank of Canada Staff Analytical Note 2025-18, June 2025.
[20] A. Cavallo, O. Kostyshyna, O. Kryvtsov, M. Vieyra, The price impact of Canadian retaliatory tariffs, Bank of Canada Staff Working Paper 2026-22, June 2026.
[21] Statistics Canada, Gross domestic product, income and expenditure, second quarter 2026, The Daily, 28 August 2026.
[22] BC Stats, Quarterly population highlights, 2026. Population as at 1 April 2026.
[23] Province of British Columbia, Supporting B.C.’s softwood lumber industry, 2026.
[24] Willson International, Canadian counter-tariffs on U.S. products effective September 8, 2026.
Streetwise Economics Working Paper 2026-06. Prepared 31 August 2026; every source opened and checked on that date. A companion workbook reproduces every figure as a live formula. This paper is analytical and educational; it is not investment advice. Corrections are welcome at isacjonasi@gmail.com.
Leave A Comment