The refund that went to companies, not shoppers
In February 2026 the US Supreme Court ruled that the emergency powers law behind the White House's broadest 2025 tariffs does not allow the president to impose tariffs. That covered the tariffs on imports from most countries and the ones imposed on Canada, Mexico and China over drug trafficking. Tariffs imposed under other laws, such as those on steel, aluminum and cars, were not affected. The ruling itself did not order refunds. That came in a separate order from the US Court of International Trade, which directed customs to settle the affected imports without those duties. The government appealed that order. Customs filings put the duties collected under the emergency law at about $166 billion, paid by more than 330,000 companies. That is the amount collected, not the amount already paid back.
None of the money the government returns goes to shoppers. US customs says it can refund these duties only to the importer of record, meaning the company named on the customs paperwork, or to a party that company designates. Shoppers who paid more at the till have no claim on the government. Some have sued retailers instead, Walmart among them, arguing under state law that the companies should pass part of their refunds on. No federal customs law requires it. Whether state law does is for the courts to decide.
That gap, between who handed the money to the government and who ended up poorer, is the whole question of who pays a tariff. A company that raised its prices to cover a tariff and now gets the duty back is being repaid for a cost its customers may have carried. The refund shows who paid customs. It does not show who bore the cost.
The question is back in the news. In late September the US and China set out procedures for considering lower tariffs on roughly $30 billion of each other's goods. The document does not itself cut any tariff. Separately, the US has charged 50 percent duties on specified Canadian goods since August 22, and since September 29 it has banned imports of specified Canadian dairy products, packaged beer, wine and motorcycles with engines over 800 cc.
- CBP : IEEPA duty refunds
- Supreme Court of the United States : Learning Resources, Inc. v. Trump, slip opinion
- US Court of International Trade : Slip Op. 26-94
- Consumer Notice : Tariff refund lawsuit
- Open Class Actions : Walmart tariff class action
- The White House : US-China Board of Trade working procedures
- Troutman Pepper Locke : From duties to import bans, Section 338 actions against Canadian goods
Step one : the bill is paid at your own border
A tariff is a tax collected by the country doing the buying. When a container of toys arrives from China, the company bringing it in pays the duty to its own customs office before the goods are released. China's government pays nothing. The Chinese factory usually pays nothing directly either. The exception is a foreign seller that registers as the importer itself, as some do when they sell with the duty already included in the price. Then the seller pays customs directly, but a foreign government still never does.
So the question politicians argue about, whether the foreign country pays, has a plain legal answer : it does not. The refund shows this. The money came from importers, so it goes back to importers. The real question is a different one. Does the cost stay with the importer, slip back to the foreign seller, or move forward to the shopper?
Step two : foreign sellers barely lowered their prices
A foreign seller could, in principle, cut its price to keep its customers, quietly taking part of the tariff on itself. Economists can check whether this happened by looking at the price at the border before the tariff is added. For the US tariffs of 2018 and 2019, two separate teams found that it mostly did not happen. Prices of the targeted imports did not fall, so the tariff was added almost in full to what US buyers paid.
The 2025 tariffs look the same. In the May 2026 version of their study, economists Gita Gopinath and Brent Neiman estimate that 92 percent of the tariff showed up in US import prices. Economists at the Federal Reserve Bank of New York, part of the US central bank, put it at around 90 percent. In other words, of every dollar of tariff, roughly 90 cents stayed in the United States.
There is one notable exception, and it runs the other way. When China hit back at US farm goods in 2018, American exporters did cut their prices, so part of China's tariff fell on US farmers. One shipment of soybeans is much like another, so buyers can easily walk away to a different supplier. Gopinath and Neiman also looked at the seven exporters facing the largest tariff increases in 2025. Only shipments from China saw their average pre-tariff price fall, by 3.1 percent over the twelve months to December 2025. That is an average across shipments, not a cut every Chinese seller made. These cases show that the split is not a law of nature. It depends on how easily buyers can go elsewhere.
- Amiti, Redding and Weinstein, Journal of Economic Perspectives
- Fajgelbaum, Goldberg, Kennedy and Khandelwal, The Return to Protectionism
- Cavallo, Gopinath, Neiman and Tang, Tariff pass-through at the border and at the store
- Gopinath and Neiman, The incidence of tariffs : rates and reality
- Federal Reserve Bank of New York, Staff Report 1201
Step three : companies and shoppers split the rest
Once the cost is inside the country, it is shared between the companies selling the goods and the people buying them. Shoppers took the smaller share at first. A Harvard team led by Alberto Cavallo tracked daily prices at large US retailers and found that seven months after the 2025 tariffs began, store prices reflected up to 24 percent of the tariff.
That figure understates the shopper's share, because the imported item is only part of what a store charges. Rent, wages and transport make up the rest. The team assumed that the whole tariff was added at the border and that imports make up half of the store price. On those assumptions it estimates that US consumers had paid up to 43 percent of the tariff's cost by then, and that US companies absorbed the rest, mostly in thinner margins.
Shoppers also paid in a way that is easy to miss. Prices of American-made goods rose too, by 3.6 percent above their earlier trend, against 6.2 percent for imports. Domestic makers pay more for imported parts, and with foreign rivals more expensive they can charge more themselves. The New York Fed study finds that about a third of the tariff's effect on store prices came through goods made at home.
What moves the split between company and shopper
Time is the first thing. Imported goods get more expensive almost at once, but the rise in American-made goods builds over nine to twelve months, as higher costs work through supply chains. Over a full year, the New York Fed economists estimate that about 26 percent of a tariff reaches consumer prices, measured against goods less exposed to the tariff. That is a comparison between goods, not a forecast of overall inflation.
Expectations come second. A store that thinks a tariff will vanish next month has little reason to reprint its price tags. Canada offers a clean test. When it put counter-tariffs on US goods in 2025, a Bank of Canada study found that store prices of those goods peaked at about 6 percent, roughly a quarter of the 25 percent tariff. They moved with trade news and fell back quickly once the tariffs were lifted.
The third is surprisingly simple : whether the shelf says so. In the same Canadian study, prices rose more for products that stores labeled as tariffed. The study records the pattern, not its cause. One reading is that shoppers accept a rise more readily when they are told why, but the labeling result alone does not prove it.
The sentence to keep
A tariff is charged at your own border, and the open question is how the bill is split at home. In the recent US tariffs, almost none of it went back to the foreign seller. Companies swallowed most of it at first, and shoppers took a larger share as the months passed, including on goods made at home.
These numbers describe particular tariffs, products and years. They are not a rule for every case, and they say nothing about whether a given tariff is worth its cost. But they give a way to read the next tariff headline. Don't ask which country will pay. Ask how easily buyers can switch suppliers, how long the tariff is expected to last, and whether stores will say it on the label.
