Adam Slater
The USMCA is a globally significant trade grouping. Mutual trade among its three members – the US, Mexico, and Canada – was worth around $1.5 trillion in 2023, or about 8% of world imports. As a result, the recent threat by US President-elect Donald Trump to put sweeping 25% tariffs on imports from Canada and Mexico, could have potentially huge effects both regionally and globally. We consider the implementation of blanket tariffs, particularly on a permanent basis, as a very low probability scenario, compared to the more targeted tariffs we assume in our baseline. But such a scenario is nevertheless worthy of analysis given the scale of the possible effects and the general uncertainty regarding this issue.
We have already illustrated some of the possible regional impacts in recent work, but with a focus on the short-term effects as we assumed tariffs would be lifted by mid-2026. In this report, we supplement our previous research with an analysis that looks at the effects of permanent 25% tariffs and also covers some of the global spillovers, using the Global Trade Analysis Project trade model of Purdue University.
An important initial observation is that regional trade is far more important for Canada and Mexico than it is for the US. Trade with the US is 15%-20% of Canadian and Mexican GDP, but US trade with its USMCA partners is only around 3% of US GDP.
Similarly, 12%-13% of the value-added in Mexican and Canadian exports originates in other USMCA partners, but less than 2% of the value-added in US exports comes from Mexico and Canada. Regional supply chain disruption is a much bigger deal for Mexico and Canada than for the US.
This implies that the costs of regional trade disruption via large-scale tariffs would fall disproportionately on Canada and Mexico. Both countries would need to reorient their trade heavily to prevent steep declines in total exports and output – something that would be very challenging and likely take a lengthy period.
…and would be massively disrupted by 25% tariffs
Our main scenario using the GTAP model assumes that the US imposes permanent 25% tariffs on all goods imports from Canada and Mexico from early 2025 plus an additional 10% tariff on imports from China (in line with President-elect Trump's recent threats).
In this scenario, intra-USMCA trade flows drop sharply. US imports from Mexico and Canada fall by around 50% and US exports to Mexico and Canada drop by around 60%. This implies a price elasticity of demand for regional trade in the GTAP model of around -2, meaning that every 1ppt rise in tariffs cuts mutual trade by around 2%.
This is broadly in line with historical evidence. In the early 1930s, we estimate Canadian export volumes to the US fell 20% in response to a tariff rise of 7ppts-10ppts, and the evidence from the US-China trade war from 2018-2023 shows that a 1ppt rise in US tariffs on China has cut imports from China by around 2.5%. There is nevertheless considerable uncertainty about the likely elasticity; academic estimates vary and elasticities between sectors will also likely be very different.
The impact on trade flows could be significantly mitigated if some key industries were exempt from tariffs. In an alternative ‘Canadian exemptions’ scenario, we assume that the Canadian extraction, chemicals, and auto industries are spared US tariffs and Canada spares the same US industries. These Canadian industries are heavily integrated with their US counterparts so imposing large tariffs on them might be seen as self-defeating by the US. Similar exemptions for Canada occurred when US President Nixon imposed a 10% import surcharge in 1971.
This would exempt around half of US imports from Canada from tariffs and greatly reduce the drop in US-Canadian trade. US imports from Canada would nevertheless still fall 18% and US exports to Canada would drop by around 30%.
The imposition of 25% tariffs on all regional trade flows in North America would lead to a large-scale reshuffling of the geographical pattern of the USMCA members' trade and considerable displacement of imports by domestic production.
Our GTAP results show that the share of Mexico’s and Canada’s exports that go to the US would roughly halve, as would the combined share of Mexico and Canada in US imports. The main gainers in terms of US market share would be Europe, Asia (excluding China), other emerging economies, and the US’s other FTA partners. Mexican exports would reorient towards Canada, Europe, China, other Asian economies, and especially other emerging economies. Meanwhile, Canadian exports would shift towards Europe, China, other Asian economies, and other emerging economies. The reorientation of Canadian trade would be even more dramatic than that seen during the 1930-1933 period, when the combination of big changes in tariff levels and collapsing US GDP resulted in a sharp shift from US to Commonwealth trade.
In terms of import displacement by domestic production, our modelling with GTAP shows that US imports of manufactured goods from countries subject to tariffs would fall around 45%, with the gap in the US market filled by higher domestic sales by US manufacturers and higher imports from other economies in a roughly 2:1 ratio. There is some uncertainty about this ratio though. Our analysis shows that for the electronics sector, around one-third of displaced Chinese imports were replaced by US domestic production and about two-thirds by imports from other countries from 2018-2022.
The disruptions to trade and output in our main GTAP scenario would be particularly large in industries which are strongly regionally integrated. Chief among these would be the auto industry, where exports by the USMCA countries would slump and output would drop by almost 30% in Canada. Meanwhile, Mexico would experience big drops in machinery and electronics output, due to these sectors being heavily focused on exports to the US. Output in the US chemicals industry would also fall. By contrast, output would rise in some sectors, helped by import substitution. In the US and Canada, this would include electronics and machinery, and in Mexico it would include chemicals and other manufacturing.
Output losses would likely be larger in the near term
Our main GTAP scenario shows long-term/structural output losses of around 0.2% for the US and 1.3%-1.5% for Canada and Mexico. These output losses result from lower exports and higher prices (and lower real incomes) brought about by tariffs. The much larger losses in Mexico and Canada reflect their much greater exposure to intra-USMCA trade noted earlier. In our alternative scenario, where there are sizeable exemptions to the tariffs between the US and Canada, output losses are modestly smaller for the US and considerably smaller for Canada.
Short-term output losses would likely be significantly larger than this. The modelled long-term losses are constrained by very substantial trade reorientation which would probably take some years to achieve, especially since much production is largely geared to the specifics of the regional North American market. Indeed, to limit long-term export losses to 11%-12%, Canada and Mexico would need to increase exports to non-US destinations by a huge 70%-100%. We would note that for firms to consider making the big changes needed to allow such reorientation, there would need to be a belief that the tariffs are permanent.
While all three USMCA economies struggled to adjust to new tariff levels and reorient towards other markets, short-term declines in investment, employment, and consumer spending would be deeper and price rises larger than the long-term impacts generated by our GTAP scenarios. Near-term financial market disruption would also probably be larger. In particular, to cushion the impact of tariffs on exports to the US and help reorientation of exports to other markets, there is a high probability that there would be a large structural depreciation of the Mexican and Canadian exchange rates, which would partially help neutralise the impact of higher US tariffs. The UK experience in 2016 following the Brexit vote is a possible model here.
While the GTAP model is good at looking at the long-term, structural effects of permanent tariff changes, we think our Global Economic Model (GEM) is better suited to capturing some of the shorter-term dynamics. We recently used the GEM in this way, constructing a scenario where 25% tariffs are imposed at the start of 2025 but then lifted in 2026. In this scenario, US GDP declines by as much as 2% below our baseline at the worst point in early 2026 while Mexico and Canada suffer peak GDP losses of 3.5%-4%. These results with the GEM illustrate that the adjustment to radically different long-term trade patterns implied by our GTAP results could be very painful.
The GTAP model is made up of a global database describing bilateral trade patterns, production, consumption, and intermediate uses of commodities and services and a multi-region, multi-sector computable general equilibrium model. This allows the effects of changes in trade policy such as tariffs to be studied, including shifts between domestic and foreign production and the reorientation of exports and imports to new destinations/suppliers/products due to changes in relative trade costs and resource shifts.
A limitation of the GTAP model used in this Research Briefing is that it provides 'before and after' static impacts of trade policies, rather than time series results. This abstracts from the dynamic impacts of trade policy changes and may miss some of the adjustment costs to new trade patterns. As such, the impacts shown by the GTAP should be seen as the long-run effects of a permanent increase in tariffs.
Our Global Economic Model has a different structure and focuses more on short-term output and price dynamics and incorporates additional channels such as financial market effects.
Adam Slater is the lead economist at Oxford Economics.
BUSINESS REPORT