When the US imposed sweeping tariff increases in 2025, most economic forecasters predicted a sharp rise in consumer prices and significant import disruption. Instead, retail prices rose modestly and import-dependent sectors held up. This column uses data for the 50 largest US trading partners to show that foreign exporters absorbed roughly 40–50% of the 2025 US tariff increases through lower export prices – much more than earlier studies suggested. The absorption was highly unequal: dominant suppliers bore the bulk, while smaller ones passed the tariff on almost fully. The aggregate result vindicates optimal tariff theory. It does not mean the tariffs made America better off.

When the United States imposed sweeping tariff increases in 2025 – averaging 10–50% across trading partners – most economic forecasters predicted a sharp rise in consumer prices and significant import disruption. It did not happen. Retail prices rose modestly. Import-dependent sectors held up. 

Trade theory offers an explanation. When a large importer like the US imposes a tariff, foreign exporters can respond by lowering their prices to limit the loss in sales. Part of the tariff burden then falls on the exporter, not the consumer – the classic ‘terms-of-trade effect’ that has been central to trade theory. 

The pattern shows up clearly in individual cases. Chinese exporters of toys and dolls saw tariffs rise by 28 percentage points while cutting export prices by 22%, absorbing roughly 80% of the shock. South Korean automakers faced a 22 percentage point tariff increase and lowered export prices by 12%, absorbing about half. Indian cotton bedsheet producers saw a 28 percentage point tariff increase and reduced export prices by 16%, also absorbing roughly half. In each case, exporters with large market shares and limited US substitutes had both the incentive and the room to absorb rather than pass through the tariff.

Yet the recent empirical literature on both the 2018–19 and 2025 tariff shocks finds near-complete pass-through to importers, with Amiti et al. (2026), Hinz et al. (2026), and Fajgelbaum and Khandelwal (2026) all reporting US incidence above 90% for the 2025 shock.

In this column, I argue those estimates do not tell the full story because they treat all product-country trade relationships symmetrically, giving equal weight to a tiny textile code with $300 in annual imports as to passenger vehicles accounting for $82 billion. They also don’t account for differences across countries, where the top exporter of a specific good typically represents half of product imports. Once the country-product trade flows are weighted by economic importance – or restricted to the dominant exporters of the most significant products – a very different picture emerges.

What the price data show

Figure 1 tells the central story in a single picture, using monthly bilateral product-level import data for the 50 largest US trading partners – covering over 95% of total US goods imports. In 2023 and 2024, export prices, tariff rates, and landed prices (the all-in cost to US importers) move together near their 2024 average baseline. Then from April 2025, tariffs begin rising sharply, reaching over 10% above its 2024 average by mid-year. But export unit values move in the opposite direction: they fall approximately 7–9% below their 2024 average. The result is that landed prices rise by only 1–3% – a fraction of what full pass-through would imply

Notes: All three series are normalized so that the 2024 calendar-year average equals one, allowing the pre-tariff baseline and post-tariff adjustment to be read directly from the chart. The unit value series (dashed red) is the value-weighted geometric mean of the ratio of each product-country pair’s monthly unit value to its average 2024 unit value. The duty multiplier series (solid blue, shaded) is the value-weighted average of (1 + applied duty rate) across matched pairs, normalized by its 2024 average. The landed price series (solid purple) is the product of the two — the price a US importer pays inclusive of the duty, normalized to the same base. Author’s calculation using 50-country sample at the HTS 10-digit level from September 2023 through January 2026.

Why other studies miss it

Previous work estimates the average price response treating all product-country pairs symetrically. The problem is that US imports are extraordinarily concentrated. The top 1% of HTS-10 products (186 out of 18,643) account for over half of total import value. The bottom half of all product codes combined accounts for less than 1% of trade. The top exporting country of a product typically holds a roughly 50% market share.

An unweighted analysis gives equal weight to every product-country observation regardless of economic significance. The tens of thousands of small, marginal exporters – genuine price-takers who cannot absorb tariffs – dominate the sample and drive the coefficient toward full pass-through. The result is statistically precise but economically misleading: it accurately describes the typical trade relationship, but the typical trade relationship is economically negligible.

Figure 2 presents estimates of the US tariff incidence from different specifications, weighted versus unweighted, and after stripping out product price trends and country pricing tendencies. 100% implies the US importer pays the full tariff, 0% means the foreign exporter pays the full tariff.

The contrast is stark. In the preferred weighted specification (Product×Time + Country fixed effects), the estimated US incidence is 53% – meaning foreign exporters absorbed roughly 47 cents of every dollar of tariff through lower export prices. In the unweighted version of the same specification, the estimate shows more than full pass-through, an implausible result driven by the many small price-taking pairs in the sample. The finding holds up across a range of different ways of slicing the data.

Dominant exporters absorb more 

A potential explanation is that dominant exporters are likely to be the lowest-cost, highest-markup producers. Standard trade models with heterogeneous firms predict that such exporters earn rents that give them room to compress margins in response to a tariff rather than pass the cost through to US buyers. Smaller exporters, operating near the competitive margin, have no such buffer and must pass the tariff through in full or exit.

Indeed, Figure 3 reveals a striking pattern about which exporters absorb tariffs. Countries are ranked within each product by their pre-tariff market share. Then, considering only country-product pairs where the tariff increased by more than 5 percentage points, the absorption ratio is calculated by rank, where 100% means full exporter absorption. 

The dominant supplier in each product – the country with the largest share of US imports – absorbs as much as 70% of the tariff on average, meaning US buyers pay only 30 cents of every dollar of tariff. That absorption fades quickly as market share falls: only the top three or four exporters, those supplying more than 10% of US imports in a product, show meaningful absorption. The smaller exporters, with market shares of just 3–4% or less, pass the tariff through entirely.

This makes intuitive sense. A dominant supplier can afford to lower its price to protect its position in the US market. A small supplier has no such leverage and no cushion to absorb costs.

It also explains why other studies find little evidence of exporter absorption. When every exporter is treated equally regardless of size – giving a 3% supplier (as well as the multitude of much smaller exporters) the same weight as a 59% supplier – the signal from the dominant exporters is buried under thousands of small price-takers. The headline result looks like full pass-through, because most exporters, by count, are indeed passing the tariff through in full. But those exporters account for a tiny fraction of what Americans actually buy.

Trade policy implications

The finding that foreign exporters absorbed roughly half the 2025 tariff increase has three important implications.

• First, the results help explain why consumer prices have risen so much less than the headline tariff numbers implied. When dominant foreign exporters absorb half the tariff through lower export prices, the landed cost to US importers rises by only half the tariff rate. Absorption by retailers, as highlighted by Cavallo et al. (2021), compresses this further. A 10-20% tariff translating into a small retail price increase is entirely consistent with these estimates.
• Second, it vindicates the terms-of-trade motive for trade agreements. Bagwell and Staiger (1999, 2002) argue that the GATT and WTO exist precisely because large countries can improve their terms of trade through tariffs, and that reciprocal liberalisation allows them to escape the prisoner’s dilemma this creates. If tariffs were fully passed through, the terms-of-trade motive would be empirically weak. The 2025 results show it is not: the US did extract a significant price concession from foreign exporters, consistent with a large country successfully exploiting its market power.
• Third, these terms-of-trade gains reinforce the value of the rules-based trading system they help undermine. Precisely because large countries can extract such gains, reciprocal agreements that constrain tariff-setting are economically valuable. The fact that the US could do this to its trading partners means its trading partners had strong incentives to retaliate and to withdraw from the agreements that previously constrained them. Retaliation and the erosion of the multilateral system may prove to be the most consequential legacy of the episode.

What the results do not imply is that the tariffs were successful on welfare grounds. The static terms-of-trade gains must be weighed against the costs of distorted trade, reduced import variety, and policy uncertainty. A fuller treatment of the welfare implications of heterogeneous tariff incidence is a priority for future work.

Source : VOXeu