Methods & data

What the tool measures

This page sets out how we estimate the share of Canadian jobs exposed to U.S. tariffs, by industry, province and territory, and Census Division. We report two measures:

Each measure is reported as a share of the employment in a place and as a count in job-equivalents. Here, “jobs” refers to the 2021 Census employed workforce assigned to the industry of their job (counted by place of residence). Job-equivalents are fractional allocations of that population. The number of job-equivalents at-risk from U.S. tariffs is not a forecast of layoffs: the method assumes exposure is proportional to the predicted export-volume loss and does not estimate firms’ employment responses.

We compute the second measure for the tariffs in force on five dates between June 2025 and September 2026, each compared with January 2025. It is reported as a range. The high end and the low end differ in how tariffs are measured, how U.S. buyers respond, and which import bans count. The range shows how much the result depends on these assumptions; it is not a confidence interval, and actual outcomes are not guaranteed to fall within it. Trade redirection and subsequent labour-market adjustment are outside the scope of the estimates.

What it doesn't measure

• Layoffs or firm closures. Firms may absorb part of a tariff in margins, renegotiate contracts, adjust hours, or weather temporary duties without cutting headcount.
• Whether a particular person will lose their job, or layoffs that have already happened.
• Direct services exports, Canada's counter-tariffs, and exchange-rate shifts. Direct exports of services face no U.S. customs tariff and are outside scope (though Canadian service jobs that supply domestic goods exporters are captured under Direct + suppliers). Retaliatory tariffs and currency movements are held fixed.
• Trade redirection and subsequent labour-market adjustment. Sales lost in the U.S. are not offset by new buyers in other countries or within Canada.

How the method works

The method runs in two parts. The first turns tariffs into loss rates:

1. Tariff ratios for each 10-digit U.S. Harmonized Tariff Schedule (HTS-10) tariff line, split into multilateral tariffs (applying to all or most exporters, e.g. §232, §122, and §301, where U.S. buyers switch between imports and U.S. domestic production) and bilateral tariffs (applying to Canada only, e.g. §338 and anti-dumping and countervailing duties, where U.S. buyers can switch from Canada to other foreign suppliers), compared against the line’s January 2025 tariff.
2. Blended elasticities that combine substitution elasticities with U.S. total industry demand elasticity, weighted by the tariffed suppliers’ share of the U.S. market (all imports under a multilateral tariff; Canada alone under a bilateral tariff, following Hallren & Riker 2017).
3. Product loss rates: the share of each product’s exports to the U.S. predicted to be lost, computed for each HTS-10 tariff line and averaged to 6-digit Harmonized System (HS-6) products using 2024 U.S. imports from Canada.
4. Industry loss rates by province: product loss rates aggregated to 163 industry groups separately for each province, using 2024 provincial export weights and Statistics Canada Supply and Use Table product-to-industry mappings.

The second part turns employment into the two measures, each building on the one before:

1. Job-equivalents dependent on U.S. goods exports: multiply an industry’s employment by its U.S. export intensity (from Statistics Canada Table 12-10-0100, Value Added in Exports, 2022).
2. Job-equivalents at-risk from U.S. tariffs: multiply those job-equivalents dependent on U.S. goods exports by the industry’s loss rate from stage 4.

Only the second measure differs between the high end and the low end of the range.

Direct + supplier jobs and Census Division allocation

Indirect (supplier) jobs at the national and provincial levels. Statistics Canada Table 12-10-0100 also reports, for each province and industry, the jobs due to U.S. demand: jobs in the industry (across both goods and service sectors) that depend on exports to the U.S., either its own or those of the industries it supplies. Subtracting the industry’s direct export jobs gives its indirect jobs across all U.S. exports, so each job is counted once, in the industry where it is held. Because U.S. customs tariffs apply to goods, we retain the share of each supplier industry’s indirect jobs that serves goods-exporting industries and link them to the goods-exporting industries they supply using a provincial domestic Leontief input-output model built from the 2022 Supply and Use Tables at basic prices (scaled to Table 12-10-0100’s published within-province indirect totals per exporter); each supplier industry’s indirect jobs then take the weighted loss rate of those exporting industries. Indirect jobs are reported at the national and provincial/territorial levels and are not allocated to Census Divisions.

By Census Division. Provincial values of direct job-equivalents are distributed to Census Divisions (CDs) in proportion to each CD’s share of the province’s jobs in the industry. By construction, CD totals add up to provincial totals. Jobs come from the 2021 Census of Population: 293 CDs across 13 provinces and territories, by 4-digit NAICS. The Census does not publish cells under 10; where an industry has no published CD counts in a province, its CD shares come from its subsector, then its sector, then all jobs.

Full equations are set out in the Technical Note (PDF); an interactive step-by-step calculation walkthrough (Trace) is coming soon.

Why a range

The share of job-equivalents at-risk from U.S. tariffs is reported as a range with a high end and a low end (the first measure — job-equivalents dependent on U.S. goods exports — is evaluated before tariffs and is the same at both ends). The high end uses empirical substitution elasticities from trade research and the statutory tariff rates written in U.S. law, derived by adapting and extending the Budget Lab at Yale’s tariff-rate-tracker for Canada (incorporating Canadian CUSMA claiming shares, stacking rules, AD/CVD orders, and subsequent policy updates). The low end uses the duties collected at the border as estimated from U.S. Census import records (calculated duty over customs value for each HTS-10 line, in the months each set of tariffs was in force), with anti-dumping and countervailing duties (AD/CVD) added at their legal rates because U.S. Census duty data exclude them. The low end sets both substitution elasticities to 1. Lines without U.S. Census data, and dates U.S. Census data do not yet cover (September 2026), use the rates in law.

Dated snapshots of U.S. tariffs in force

Each date is compared with the January 2025 baseline; thus the baseline rate is 0% for most Canadian products (CUSMA duty-free) except for products with pre-existing duties (such as softwood lumber, drywall, and utility-scale wind towers). Only tariffs added or increased since January 2025 count toward job-equivalents at-risk.

Note: “Latest in tool (September 2026)” reflects the most recent U.S. tariff schedule logged in the model (as of September 29, 2026). Subsequent U.S. tariff changes after that date are not yet reflected here.

Assumptions & limitations

• Full pass-through (ρ = 1) at both the high end and the low end.
• Total demand elasticity (θ = 1) is set to 1 for all industries, in line with the average own-price elasticities of demand for food and agricultural products compiled by the USDA Economic Research Service (USDA ERS 2006).
• Suppliers keep their prices unchanged (partial equilibrium). We do not model how any supplier’s price responds to the tariff.
• Other trade costs are unchanged: the tariff is the only change in the cost of selling to the U.S. Transport, insurance, exchange rates, and other trade costs stay at their January 2025 level.
• Tariff-line weights: within an HS-6 product, HTS-10 tariff lines are weighted by 2024 U.S. imports from Canada (or the simple average of lines when none had 2024 imports).
• Market-share fallbacks: 98 HS-6 products (C$7.7 billion of exports, mostly electricity, which carries no tariff) have no U.S. market-share data and take m = 0.058 and p = 0.435 (the median product’s import penetration).
• Low-end substitution elasticities: at the low end, both substitution elasticities are set to 1 for every product.
• Provincial export intensities, product mixes, and supplier linkages: jobs are counted only in industries for which their province exported to the U.S. in 2024 (with NL, NS, and PE excluded from softwood lumber AD/CVD orders). Within a province, Census Divisions share the province's industry export intensity and product mix and differ through their local 2021 Census industry employment mix.

Data sources & references

References

• Ahmad, S. and D. Riker (2020). Updated Estimates of the Trade Elasticity of Substitution. USITC Economics Working Paper 2020-05-A.
• Fontagné, L., H. Guimbard and G. Orefice (2022). “Tariff-Based Product-Level Trade Elasticities.” Journal of International Economics.
• Hallren, R. and D. Riker (2017). An Introduction to Partial Equilibrium Modeling of Trade Policy. USITC Economics Working Paper 2017-07-B.
• Pierce, J. and P. Schott (2012). “A Concordance Between Ten-Digit U.S. Harmonized System Codes and SIC/NAICS Product Classes and Industries.” Journal of Economic and Social Measurement.
• Riker, D. (2025). USITC Economics Working Paper 2025-06-B.
• Statistics Canada. Supply and Use Tables (Table 36-10-0478-01); Value Added in Exports (Table 12-10-0100); Census of Population 2021 (Table 98-10-0592-01); Canadian International Merchandise Trade.

About, citing & downloads

• Technical note (PDF, hosted by IRPP) to come
• Downloads: results by place, industry and snapshot (Parquet / CSV) to come
• What changed in this version changelog: to come

Every release pins its input datasets, precomputed views, and engine code so that any number cited from the tool can be reproduced.