Tariff refunds ordered after the Supreme Court’s IEEPA ruling have exceeded new tariff collections every month since May, new data show, wiping out net customs revenue and complicating the administration’s fiscal claims
WASHINGTON, Aug. 23, 2026
The federal government is now paying out more in tariff refunds than it collects in tariffs, a reversal without precedent in modern customs history, according to a Tax Foundation analysis of monthly Treasury statements that drew wide attention on Friday after it was reported by Fortune.
The data show that in May, the month United States Customs and Border Protection opened its online refund portal, the Treasury refunded 21.97 billion dollars to importers while collecting 21.93 billion dollars in duties, tipping net customs revenue fractionally negative. In June the imbalance blew out: 49.18 billion dollars flowed back to importers against 23.63 billion dollars collected, a net drain of 25.56 billion dollars in a single month. The hemorrhage traces to the Supreme Court’s ruling in February that tariffs imposed under the International Emergency Economic Powers Act were unlawful, a decision that obligated the government to return roughly 166 billion dollars collected from as many as 330,000 importers between 2025 and early 2026.
As of this week, the Treasury has disbursed about 100 billion dollars in refunds since May, crossing the halfway mark of the total owed, according to Fortune. Roughly 66 billion dollars remains outstanding, and the Tax Foundation warns that the remaining claims will be slower and messier to process, even as interest accrues on every dollar the government has yet to return.
“While importers will experience some relief by receiving refunds, the economic damage from the chaotic tariff regime cannot be refunded, and the remaining tariffs means economic damage will continue to grow,” the Tax Foundation report said.
The refund drain lands at a politically awkward moment. The administration has spent 19 months promising that tariff revenue would help close the federal deficit and offset the cost of last year’s tax legislation. Instead, for at least two consecutive months and likely a third, the customs line of the federal ledger has been a net negative, even as replacement tariffs imposed under other statutes generate fresh collections.
BACKGROUND: THE BILL FOR THE IEEPA EXPERIMENT COMES DUE
The story begins with the legal architecture the administration chose in 2025. Rather than pursuing tariffs through the slower investigative procedures of Section 301 of the Trade Act of 1974 or Section 232 of the Trade Expansion Act of 1962, the White House invoked the International Emergency Economic Powers Act, a 1977 sanctions statute, to impose sweeping duties at speed, including the so-called reciprocal tariffs applied to most trading partners.
Importers sued almost immediately. The Court of International Trade ruled against the government, the Federal Circuit affirmed in August 2025, and on February 20, 2026, the Supreme Court held 6 to 3 in Learning Resources, Inc. v. Trump that IEEPA does not authorize the president to impose tariffs at all. The ruling terminated the levies and, critically, entitled the importers of record who paid them to their money back. Analyses by the Penn Wharton Budget Model at the time projected exactly the revenue reversal now visible in the Treasury data.
The administration responded on two tracks. It began rebuilding the tariff wall on firmer statutory ground, deploying Section 122 of the Trade Act for a temporary global surcharge, expanding Section 232 national-security tariffs on metals and other strategic goods, and layering new Section 301 actions, including a July action imposing duties of 10 to 12.5 percent on some 60 economies over forced-labor enforcement failures. And it began, reluctantly and then at scale, paying people back.
CBP stood up an online refund system in May, processing claims through what the agency calls CAPE declarations. The mechanics have been rocky. According to a BDO analysis of CBP filings, more than 75,000 CAPE declarations had been submitted by the end of July, with 17.69 million import entries already liquidated without the IEEPA tariffs and moved into the refund pipeline, and approximately 128.68 billion dollars in potential and certified refunds accepted for processing. But Supply Chain Dive reported that CBP has rejected nearly 30 percent of filings in the refund system, a failure rate practitioners attribute to data mismatches between entry summaries and claims, errors in identifying the importer of record, and confusion over which entries fall inside the eligible liquidation window.
Meanwhile the meter runs. Refunds not yet disbursed accrue interest at government overpayment rates, 4.5 percent on amounts of 10,000 dollars or more and 6 percent below that threshold, according to the Cato Institute. Earlier analyses estimated the interest obligation was compounding at hundreds of millions of dollars a month, a cost borne by taxpayers on top of the principal.
STAKEHOLDER REACTIONS: ‘A CHAOTIC ENVIRONMENT’
Erica York, vice president of federal tax policy at the Tax Foundation and an author of the analysis, argues that the data expose the gap between the administration’s fiscal rhetoric and the ledger. “The president himself and the administration have been talking so much about how they’re going to raise a lot of revenue with tariffs, how they’re going to supposedly fix the fiscal situation with tariffs,” York told Fortune. “And that really mismatches what we’re seeing play out in the data, which is that they have relied on really shaky legal grounds to try to impose these tariffs.”
York’s sharper point is about uncertainty as an independent economic cost. By her count, tariff policy has changed more than 50 times since January 2025, a churn she says has been as disruptive as the duties themselves, leaving companies scrambling to adapt supply chains and deferring hiring and wage decisions. “It hasn’t just been, ‘Here’s a new tariff done in a very transparent way,’ and then businesses can plan around it,” she said. “It has been a chaotic environment.” Even with the emergency-powers saga concluded, she expects no calm: “Even though we’re past the IEEPA saga, we’re not past the chaotic tariff environment saga. I think we are stuck in that for at least the next couple of years.”
The administration’s defenders counter that the refund wave is a transitional artifact, not a verdict on tariff policy. Net customs revenue, they note, will swing back positive once the refund backlog clears, because the replacement tariffs under Sections 122, 232 and 301 continue to collect on a broad import base. Treasury officials have consistently characterized the refunds as the orderly execution of a court mandate.
The importer community, for its part, is focused less on the fiscal optics than on getting paid. Trade counsel report that the first phase of refunds, covering entries liquidated in the normal course, moved relatively smoothly for large importers with clean records, while smaller importers and those with complex entry histories are disproportionately represented in the 30 percent rejection pile. The Liberty Justice Center, which represented the winning plaintiffs before the Supreme Court, has continued publishing procedural updates urging importers to perfect claims before liquidation deadlines pass.
Litigation pressure on the remaining tariff structure is building as well. In late July, spice importer Burlap and Barrel filed suit challenging Section 301 tariffs, part of a broader legal campaign that seeks to extend the Supreme Court’s skepticism of executive tariff power to the statutes the administration now relies on. Few practitioners expect those challenges to succeed as completely as the IEEPA cases, since Section 301 rests on decades of practice, but the filings guarantee that legal uncertainty will shadow the replacement regime too.
ECONOMIC IMPACT: THE DAMAGE THAT CANNOT BE REFUNDED
The refund program is, on one level, the largest business relief event of the year, a 166 billion dollar transfer back to companies that financed the duties out of working capital. Research from the Federal Reserve Bank of New York found that importers, overwhelmingly American firms, bore the direct cost of the 2025 tariffs, so the refunds land on the balance sheets of the companies that paid, from multinationals to family import businesses, many of which took on debt to cover duty bills at the peak.
But economists broadly agree with the Tax Foundation’s framing that much of the damage is unrecoverable. The tariffs raised consumer prices while they were in force: Federal Reserve Bank of St. Louis research found the levies lifted prices of pharmaceuticals and household utensils by more than 4 percent over the year studied, among other categories. Those price increases were paid by millions of households that will receive nothing from the refund process. Supply chains rerouted at real cost; contracts were renegotiated; investments were deferred; and some import-dependent small businesses failed outright before the Supreme Court ruled. None of that is refundable.
The fiscal arithmetic is also less favorable than the gross collection figures once suggested. Tariff revenue was always a small fraction of total federal receipts, and the Penn Wharton Budget Model projected after the February ruling that refunds would claw back a substantial share of everything the IEEPA tariffs had raised. The Treasury data now confirm it: measured from May, the tariff program’s net contribution to the fisc has been negative, and interest owed on pending refunds deepens the hole each month. York expects net customs revenue to turn positive again within months as the backlog clears, but the episode has permanently discounted the argument that tariffs are a reliable deficit tool, because any tariff imposed on contested authority carries a contingent refund liability of its own.
There is a subtler cost that trade economists emphasize: credibility. Foreign negotiating partners watched the United States collect 166 billion dollars on a legal theory its own Supreme Court rejected, and they now negotiate with an administration whose remaining tariff instruments are themselves under challenge. That dynamic cuts both ways, giving counterparties reasons to slow-walk concessions while litigation runs, and giving the administration incentives to lock in deals before courts move again.
IMPLICATIONS FOR IMPORTERS, EXPORTERS AND US BUSINESSES
For importers still owed money, the practical guidance from the trade bar is unambiguous: file accurately or wait longer. With nearly a third of refund filings bouncing, the difference between a clean claim and a defective one is measured in months of foregone cash and interest complications. Importers should reconcile their own entry data against CBP records before filing, confirm importer-of-record designations, particularly where customs brokers or freight forwarders acted as intermediaries, and calendar the liquidation-related deadlines that govern the harder second phase of claims. The remaining 66 billion dollars will not be distributed evenly or automatically; it will go to the companies with the cleanest paperwork first.
For finance teams, the refunds are a windfall with accounting consequences: duty recoveries flow through cost of goods sold or other income depending on treatment elected when the duties were paid, and the interest component is taxable. Companies that priced tariff costs into customer contracts face commercial questions about whether recoveries must be shared, an issue already generating disputes in supply agreements with tariff pass-through clauses.
For all businesses exposed to trade, the strategic message of the negative-revenue months is that tariff policy in this cycle is not a stable tax but a contingent one, imposed quickly, litigated constantly, and sometimes reversed with interest. That argues for three standing disciplines: model duties under multiple legal scenarios rather than current rates alone; preserve refund optionality by keeping entry records litigation-ready; and treat every new tariff, including the Section 232 and 301 measures now carrying the load, as potentially refundable years later, with all the cash-flow planning that implies.
INSIDE THE 30 PERCENT REJECTION PILE
The rejection statistics deserve a closer look, because they are becoming the defining operational story of the refund program’s second act. A near 30 percent failure rate on filings, in a system where the government concedes it owes the money, reflects the collision between the improvised speed of the original tariffs and the exacting formality of customs law.
Practitioners describe several recurring failure modes. The most common is the identity problem: refunds run to the importer of record, but during 2025 many companies imported through customs brokers, freight forwarders acting as nominal importers, or foreign suppliers selling on delivered-duty-paid terms. In those arrangements, the party that economically bore the tariff is often not the party legally entitled to the refund, and CBP’s system rejects claims where the filer and the importer of record do not match. Unwinding this requires assignments, powers of attorney, or commercial renegotiation, and in some delivered-duty-paid cases the refund may flow to a foreign seller with no obligation to pass it on to the American buyer who actually absorbed the cost in pricing.
The second failure mode is data reconciliation. A refund claim must tie to specific entry numbers, and companies with high-volume, multi-port operations frequently discover discrepancies between their internal records, their brokers’ records and CBP’s database, any one of which will bounce a filing. The third is timing: entries fall into different procedural categories depending on liquidation status, and claims filed under the wrong pathway for their category fail even when the underlying entitlement is undisputed. The Tax Foundation’s observation that the remaining 66 billion dollars involves claims filed after the established liquidation period, raising unresolved procedural questions, suggests the rejection rate may worsen before it improves.
The interest accrual gives these frictions a fiscal edge. Every month a valid claim sits unpaid, the government’s obligation grows at 4.5 to 6 percent annualized. For the Treasury, faster processing is cheaper processing. For claimants, the calculus is reversed in one narrow sense, since interest accrues in their favor, but no treasurer prefers accrued interest to cash, and the time value rarely compensates for the working-capital strain of waiting.
THE REVENUE QUESTION HANGING OVER FISCAL POLICY
The negative-revenue months also feed a broader fiscal debate that will intensify as budget season approaches. Tariff receipts were written into the administration’s deficit narrative, and outside scorekeepers now have to model not only how much the replacement tariffs will collect but how durable those collections are against litigation. The Section 122 surcharge is expressly temporary by statute. The Section 232 programs are legally robust but commodity-specific. The new Section 301 forced-labor architecture is broad but novel, and the Burlap and Barrel litigation is the first of what practitioners expect to be several challenges probing its boundaries.
The Supreme Court’s IEEPA ruling supplies the template that worries revenue forecasters: a tariff can be collected for a year and repaid with interest in the next, converting an apparent revenue stream into a liability. Penn Wharton’s post-ruling analysis made the mechanical point that gross collections overstate the fisc’s true position whenever legal risk is unresolved. Prudent scorekeeping now applies a haircut to tariff revenue projections, and congressional staff on both sides of the aisle have begun asking the Congressional Budget Office how contingent refund liabilities should be represented in baseline estimates. However those methodological questions resolve, the political fact is simpler: through the spring and summer of 2026, the tariff line on the federal ledger has been bleeding, and the administration’s opponents have the Treasury’s own monthly statements to cite.
For exporters abroad and their American customers, the refund saga also carries a forward-looking pricing lesson. During 2025, many supply contracts allocated tariff costs to the American buyer on the theory that duties were a fixed feature of the landscape. The refunds have now revealed those allocations to be bets on litigation outcomes, and sophisticated counterparties are responding by writing contracts that specify not only who pays a tariff but who owns any subsequent refund, how interest is split, and who controls refund claims and litigation participation. Trade counsel report that refund-allocation clauses have moved from exotic to standard in large supply agreements over the course of the summer, one of several quiet ways the IEEPA episode is rewriting commercial practice even in industries that never sued anyone.
There is a final irony in the timing. In the same week the refund data circulated, the administration paused new tariffs on Canada for negotiations, waived beef tariffs to fight food inflation, and fielded a call from Brazil’s president seeking tariff relief. Each move, in its own way, prices tariffs as a cost to be traded away rather than a revenue source to be maximized. The ledger and the diplomacy are telling the same story from different directions.
The refund wave will eventually recede. The 330,000 importers will be made whole, more or less, and net customs revenue will climb back above zero. What will remain is the precedent: the largest tariff program in modern American history, unwound by the courts, its proceeds returned, and its costs, in the Tax Foundation’s phrase, unrefundable.