Tariffs Can Lose in Court and Still Win in the Market

The cleanest myth in trade policy is that a bad tariff can be unwound. A court strikes it down, the government cuts refund checks, companies reset prices, and the market returns to whatever it was before the experiment began. I do not buy it.

On February 20, the Supreme Court ruled that President Donald Trump had exceeded his authority by using the International Emergency Economic Powers Act, a 1977 sanctions law known as IEEPA, to impose sweeping tariffs on imports from much of the world; the levies had raised roughly $166 billion before they were shut down, according to Associated Press reporting on the replacement tariff push². That ruling matters. It limits a dangerous claim of presidential power. But as an economic event, it is less like erasing a line in a legal brief and more like trying to pull dye out of water.

The refund fight makes the point. General Motors now expects a $500 million refund from the overturned IEEPA tariffs, but the company told the AP it had not yet received the money and did not have a specific estimate for when it would arrive; Customs and Border Protection has said more than 330,000 importers paid about $166 billion across more than 53 million shipments, and approved claims will take 60 to 90 days while the system rolls out in phases, according to the AP’s report on GM’s expected refund¹. That is not an administrative footnote. It is the shape of the problem.

Tariffs do not land where the law says they land. Customs law can identify the importer of record. It can reliquidate an entry, correct paperwork, or refund a duty. The economy is messier. A retailer may have raised prices. A manufacturer may have paid more for inputs. A supplier may have lost a contract because a buyer shifted countries. A consumer may have paid the higher price months ago and will never be located, much less reimbursed. The legal payor and the economic bearer are often different people.

The best evidence from the first Trump trade war points in that direction. Mary Amiti, Stephen Redding and David Weinstein found that the 2018 tariffs produced “complete passthrough” into domestic prices of imported goods, altered supply-chain networks, reduced imported varieties and cut U.S. real income by about $1.4 billion per month by the end of 2018, according to their NBER paper³. The U.S. International Trade Commission later found a similar mechanism in the Section 232 and Section 301 tariffs: U.S. importers bore nearly the full cost because import prices rose at about the same rate as the tariffs, according to the USITC’s 2023 report summary⁴.

That does not mean every tariff-induced move is waste. The strongest defense of tariffs is that they can force useful diversification. If a company left a China-only sourcing model and discovered a reliable supplier in Vietnam, Mexico, India or the United States, that may make the firm less exposed to coercion, war risk, pandemic closures or Beijing’s industrial policy. The Biden administration’s own 2024 review of China Section 301 tariffs made this case, saying the tariffs reduced U.S. imports from China and increased imports from alternate sources, including allies and partners, potentially supporting supply-chain resilience, according to USTR’s four-year review release⁵. I take that seriously.

But it does not rescue the broader argument. A useful supply-chain discovery is not proof that the policy was clean, fair or reversible. It proves the opposite: tariffs change behavior before the legal system finishes asking whether they were lawful. Once a buyer qualifies a new supplier, rewrites contracts, changes logistics, hires customs counsel, books warehouse space and adjusts prices, the old market is gone. Some of the new structure may be efficient. Some of it may be a workaround. Some of it may be pure deadweight cost. The court cannot separate those threads with a refund check.

Trade rerouting is a good example. A Harvard Business School working paper on the U.S.-China trade war found that Vietnam’s exports to the United States that were effectively rerouted from China rose by 1.74 percentage points for the average tariff hike, with the increase driven by new establishments and Chinese-owned firms; the authors estimated that 8.8 percent of Vietnam’s export growth to the United States from 2018 to 2021 reflected rerouting, according to Harvard Business School’s summary of the paper⁶. That is not a simple story of “bring production home.” It is a story of global firms learning how to route around U.S. policy, sometimes by moving real production, sometimes by shifting enough activity to change origin labels, and sometimes by creating a new compliance fog.

The other reason I think protectionism wins even after losing in court is that relief itself becomes a market asset. The Government Accountability Office found that USTR received about 53,000 exclusion requests for China-tariff products from 2018 to 2020 and did not fully document its internal procedures, a flaw GAO said mattered for consistent review, according to GAO’s 2021 report⁷. A later study by Veljko Fotak, Hye Seung Lee, William Megginson and Jesus Salas found that campaign contributions and lobbying increased the likelihood of exemptions during the first Trump administration; their data covered 1,022 approved exemptions and 5,993 rejected applications, and they estimated that an exemption was worth about $51 million for the median firm in their sample, as summarized by the Cato Institute⁸ and detailed in the Journal of Financial and Quantitative Analysis paper⁹.

The narrow rebuttal is fair: discretionary exemptions under Section 301 are not the same as court-driven refunds after an IEEPA ruling. Refund eligibility should be more rule-bound. It should turn on entries, liquidation status, deadlines and documentation, not whether a company has friends in Washington. But “more rule-bound” is not the same as equal. Firms with customs departments, trade lawyers, clean records and cash cushions will move faster. Smaller importers that passed costs downstream, missed procedural deadlines, or lack documentation may recover less or later. Consumers will not recover at all.

That distributional point is where the legal victory starts to look thin. The Supreme Court can say the president used the wrong statute. It cannot make every buyer whole. It cannot restore a small importer’s lost customer. It cannot make a downstream manufacturer un-pay the higher input costs embedded in its own contracts. It cannot stop the administration from trying to recreate the same protectionist effect under other authorities, which it is already doing: two days after the Supreme Court decision, the administration imposed 10 percent temporary tariffs under Section 122 of the Trade Act of 1974, and officials have pursued Section 301 investigations that could generate new import taxes, according to AP’s account of the post-ruling tariff strategy².

The global audience can see this. On May 1, China expanded tariff-free treatment to 53 of Africa’s 54 countries, excluding Eswatini because it recognizes Taiwan, while covering large economies such as South Africa, Egypt, Nigeria, Algeria and Kenya, according to the AP¹⁰. Beijing is not doing charity. China-Africa trade reached a record $348 billion in 2025, with China exporting $225 billion to Africa and importing $123 billion, widening Africa’s deficit, according to the same AP report¹⁰. Brookings’ Yun Sun has also warned that Africa’s exports to China remain concentrated in minerals, raw materials and low-value products, and that duty-free access alone may deepen Africa’s role as a raw-material supplier unless it supports industrial upgrading, according to Brookings¹¹.

Still, China does not need to be a better economic partner in some moral sense to win the comparison. It only needs to look more predictable at the margin. Washington is offering emergency tariffs, court reversals, temporary replacement tariffs, new investigations and refund queues. Beijing is offering a simple sentence: tariff-free access for countries that recognize China. African governments know the catch. They also know the value of predictability.

My view is that the Supreme Court ruling stopped an abuse of legal authority, but it did not stop the protectionist machine. The machine has already worked through prices, supply chains, exemptions, legal bills and diplomatic perceptions. The next indicator to watch is not just how many refund checks CBP sends by midsummer. Watch whether tariff-exposed product prices fall back toward comparable non-tariff goods by early 2027, whether smaller importers recover at the same rate as large firms, and whether U.S. import shares simply migrate from IEEPA-hit countries into new Section 122 or Section 301 categories. If those gaps persist 12 months from now, the verdict will be clear: the tariffs lost in court, but the market still lives with them.