NextFin News - When the U.S. government began refunding roughly $170 billion in tariffs that the Supreme Court struck down earlier this year, the political expectation was that the money would find its way back to consumers through lower prices. A new Federal Reserve Bank of Atlanta study suggests otherwise: seven out of ten firms plan to keep at least part of their refunds as cash, according to a survey of more than 1,100 C-suite executives. The finding reframes one of the year's biggest fiscal events — a $166 billion transfer back to corporate balance sheets — as a corporate liquidity windfall rather than a consumer stimulus, with consequences for inflation, earnings quality, and the Federal Reserve's policy calculus.
The study lands in a market already pricing policy risk. On the last trading day before its release, the benchmark 10-year Treasury yield was hovering above the psychological 5 percent level and the two-year yield had climbed to 4.741 percent, its highest since July 2024, as investors anticipated further Federal Reserve rate increases. In that environment, the Atlanta Fed's conclusion that refund money is being hoarded rather than spent matters: it implies the refund will deliver little disinflationary relief and even less demand stimulus, a combination that leaves the Fed's rate path largely untouched by the tariff unwind.
The Refund That Wasn't Supposed to Sit Still
The refunds exist because of a court defeat for the administration. In February 2026, the Supreme Court ruled that the International Emergency Economic Powers Act did not authorize the tariffs imposed in 2025, and the Court of International Trade ordered the duties refunded. Federal Reserve researchers estimate the total refund will amount to approximately $166 billion — about 63 percent of all customs-duty receipts collected in 2025. By late July, roughly $100 billion had already flowed back to importers.
The Atlanta Fed's latest analysis, published September 21, draws on its Survey of Business Uncertainty, which fielded special questions to 1,156 C-suite executives between August 10 and August 21. The picture that emerges is one of caution rather than celebration. About a quarter of executives believe their firms are eligible for refunds, and among those, roughly 40 percent — or about 9 percent of the full sample when weighted by employment — have already received money. Publicly listed firms are more than twice as likely to have received refunds than non-listed peers, a gap that points to the administrative burden of the process: two-thirds of firms are handling claims with internal staff or selling their refund rights to third parties rather than outsourcing the work.
For the average eligible firm, the payout is meaningful but not transformational: refunds run about 1.7 percent of annual revenue. And when asked how they would use the money, 70 percent of firms said they anticipate retaining at least some of it as cash or cash equivalents. More than half plan to devote a portion to research and development or capital projects. Far smaller shares expect to pass the money along: 17.2 percent foresee customer rebates, 14.8 percent plan to lower prices, 12.7 percent intend worker bonuses, and 10.1 percent manager bonuses. Because respondents could select multiple uses, the categories overlap — but the ranking is unambiguous. Cash comes first.
The contrast with the political framing is stark. The refund was conceived as restitution: consumers paid the tariffs through higher prices, so consumers should get the money back. The Federal Reserve's data say the restitution stops at the corporate treasury. Once the cash lands on the balance sheet, standard corporate-finance logic takes over, and that logic points toward retention, not redistribution.
Why Cash Wins: The Textbook and the Constraint Channel
The result is exactly what standard corporate-finance theory predicts — and precisely what politicians did not want to hear. As the Atlanta Fed researchers note, a one-off windfall paid out regardless of a firm's future behavior should not, in textbook models, alter investment, hiring, or pricing decisions. The study cites Modigliani and Miller (1958) and Lintner (1956): an unconstrained firm is already operating at its profit-maximizing scale, so an unexpected cash inflow is most likely to be saved, used to repay debt, or returned to shareholders.
The mechanism turns on financial constraints. A separate Atlanta Fed analysis in July estimated that only about 34 percent of total refunds — roughly $56 billion — flows to firms classified as most financially constrained, the ones for which a cash windfall could actually relax a binding limit and fund investment, hiring, or price cuts. The remaining two-thirds goes to firms more likely to save the refund, retire debt, or distribute it, with little or no immediate effect on operations. The September survey's 70 percent cash-retention figure is the behavioral confirmation of that arithmetic.
There is also a scale effect. A refund equal to 1.7 percent of revenue is large enough to matter for a balance sheet but small enough that it rarely changes a company's strategic calculus. For a retailer with thin margins, it can repair a quarter's gross profit. For a manufacturer sitting on unused capacity, it does not create new demand. The refund compensates for a cost already incurred; it does not fund a project that would not otherwise clear the company's hurdle rate. A firm that was not going to build a new factory at a 15 percent required return is not going to build it because a tariff it never wanted was briefly imposed and then refunded.
The constraint channel does matter at the margin, and it is the one place where the refund could still do real economic work. A financially constrained importer that was rationing working capital can now restock inventory, pay suppliers on time, or fund an expansion it had deferred. But that channel is capped at the $56 billion flowing to constrained firms — a third of the total — and even there, the first use of cash is often balance-sheet repair, not new commitments.
The Retail Exception: Price Cuts as Marketing, Not Economics
The conspicuous exceptions are customer-facing retailers, where price perception is itself a strategic asset. Home Depot reported $730 million in tariff refunds during its fiscal second quarter and applied roughly $685 million of that to cost of goods sold, lifting gross margin by 0.3 percentage points. Walmart said it is eligible for about $2.9 billion and plans to use the funds to lower prices. TJX Companies applied its $331 million refund against second-quarter cost of sales. Target credited $752 million — $1.65 per share — to net earnings, with a $994 million pretax benefit to gross margin and operating income. Kohl's put $100 million into second-quarter gross margin and said it would invest the rest in deeper inventory.
Not everyone is following that script. Lowe's received roughly $80 million in repayments, an 11-cent earnings boost, and explicitly declined to cut prices.
"We feel strongly that we want to deliver strong profitability for our shareholders and make sure that we don't follow any aggressive pricing action," said Marvin Ellison, chief executive of Lowe's.
The divergence is not random: value-positioned chains have a strategic reason to proclaim price relief, while others treat the refund as margin repair.
Even the price-cutting is partly theater. Bryan Eshelman, a managing director at consulting firm AlixPartners, said untangling the rebate from other cost pressures is nearly impossible for shoppers.
"To me, a lot of this is marketing," Eshelman said.
The refund becomes a public-relations instrument — a way to claim credit for price cuts that competitive pressure might have demanded anyway.
The retail case also reveals a deeper accounting problem. Refunds have muddied earnings reports because companies have booked them inconsistently — some against cost of sales, some as other income, some as margin boosts. Over 40 S&P 500 companies have reported some $9.6 billion in refunds in the past quarter, including at least $2.1 billion in cash already received. That dispersion makes cross-company comparison difficult and gives management teams wide latitude in how much of their quarter's strength they attribute to operations versus a one-time government payment.
The Second-Order Effects: What Hoarded Cash Does to the Macro Picture
The aggregate consequence is that the refund's fiscal multiplier is close to zero. Money that sits on balance sheets as cash does not circulate; it does not lift demand, and it does not accelerate disinflation. The refund impulse, in other words, fades almost as soon as it is booked. For policymakers who counted on the unwinding of tariffs to deliver a double benefit — lower import costs plus refund-fueled price cuts — the Atlanta Fed's findings trim that expectation sharply. Less pass-through to prices means less disinflationary help from the refund channel, but also less demand stimulus to overheat the economy. The net effect is roughly neutral for growth with a modestly hawkish tilt for inflation.
This is the second-order point the market has not fully absorbed. The consensus narrative treats the tariff refund as inherently disinflationary: tariffs pushed prices up, so refunds should push them down. But if 70 percent of firms retain the cash and only 14.8 percent plan explicit price cuts, the refund does not reverse the inflation impulse — it merely stops it from growing further. The price level stays elevated; only the rate of increase changes. That distinction matters for a Federal Reserve watching core inflation prints and for bond investors positioning around the rate path.
There is a subtler distortion building in the earnings data. The one-time items have turbocharged second- and third-quarter profits, but they also set a higher bar for the following year. As Eshelman put it, the windfall creates an unfair positive comparison to the prior-year quarter and an unfair negative comparison to the year ahead. Investors reading headline earnings growth into 2027 will need to strip out the refund effect or mistake a lapping headwind for an operational slowdown. The companies that booked the largest refunds — retailers chief among them — face the steepest comparables problem.
The distribution of access matters, too. Publicly listed firms are more than twice as likely to have already received refunds, and two-thirds of firms are managing claims internally or selling the rights to third parties. A secondary market has emerged in which companies trade away future refund claims for immediate cash, often at a steep discount to face value. Smaller importers without dedicated tax and legal staffs face a choice: wait months for the full amount or sell today for a fraction of it. The refund process itself, in effect, transfers value from the least-resourced firms to the liquidity providers who buy their claims — a redistribution no court order contemplated.
The Counter-Thesis: Investment Could Still Surprise
The bear case against the "cash hoarding" reading is not trivial. Seventy percent of firms retaining at least some cash still leaves more than half planning to allocate a portion to R&D or capital projects — and the $56 billion flowing to financially constrained firms could fund investment that would not otherwise happen. If those constrained firms deploy the money quickly, the refund could still show up as a measurable lift in equipment spending and hiring by late 2026 or early 2027. The price and rebate actions, while a minority, also mean a nontrivial share of refunds reaches customers and employees directly.
That case is plausible but likely overstated. The investment intentions are spread over time and compete with existing capital budgets; the Atlanta Fed's July analysis explicitly warned that the headline $166 billion figure could overstate the stimulus effect because the payments compensate for past duties regardless of what firms do next. And the refund's small size relative to revenue — 1.7 percent — limits how much incremental spending it can finance at firms that were not already planning to spend. The dominant use remains defensive: repairing balance sheets bruised by a policy shock that the courts later deemed unauthorized.
There is also a timing mismatch. Capital projects approved in August do not appear in investment data until the spending occurs, often quarters later, and they remain vulnerable to cancellation if demand softens. The cash retention, by contrast, is immediate and visible in the next quarter's balance sheet. When the evidence is weighed by immediacy and magnitude, the hoarding story dominates the investment story.
The falsifying signal is specific. If the Atlanta Fed's next Survey of Business Uncertainty wave shows the cash-retention share falling below roughly 50 percent while reported capex and price-cutting actions rise materially, the "hoarding" thesis weakens. Equally, if core goods inflation excluding tariff effects turns decisively negative in the months after refunds peak — evidence that refunds are being competed away through prices rather than retained — the consumer-pass-through story would regain credibility. Until then, the survey says the opposite.
What to Watch: Three Horizons
Before laying out the horizons, the central judgment needs a clear classification. The refund episode is a cyclical liquidity event, not a structural shift in corporate behavior. It is mean-reverting by construction: the cash arrives once, is booked once, and then lapses against the following year's comparables. Nothing about it permanently raises the propensity to hold cash or permanently lowers the propensity to consume. The structural story lives elsewhere — in the supply-chain rewiring the tariffs triggered — while the refund itself is a one-time balance-sheet transfer whose macro footprint fades as it is lapped.
In the short term, expect more earnings beats driven by one-time refund items and a rise in corporate cash ratios. The second quarter was only the beginning; as the remaining $60 billion to $70 billion of refunds clears the Treasury, more companies will book windfalls that have nothing to do with operating performance. Analysts will need to adjust their models, and companies that lean on refund-fueled margins in their guidance are storing up a credibility problem for when the items lapse.
Over the medium term, the distortion cuts the other way. Refund-fueled margins lap against cleaner comparables, and the refund impulse fades from growth rates. The constrained-firm investment channel may begin to show up in capital-expenditure data with a lag, but it represents at most a third of the total — and only if those firms were genuinely binding on cash. The earnings narrative flips from "tariff refunds boosted profits" to "profits grew despite the absence of refunds," and that is a harder story to sell.
The long-term question is separate from the refunds themselves: whether the tariff episode, even reversed, has permanently altered supply-chain decisions. Refunds make firms whole on duties paid; they do not restore confidence that the rules will not change again. That uncertainty premium is the part of the tariff shock that no court order can refund, and it is the part most likely to show up in durable form — in diversified sourcing, in nearshoring commitments, and in the inventory buffers companies now consider essential.
The refund was sold to the public as a return to consumers. The balance sheets say it is a return to capital. Seven out of ten firms are keeping the cash — and the economy will feel the difference between those two stories for quarters to come.
Explore more exclusive insights at nextfin.ai.
