Trade & Sourcing
The $50.33 Billion Duty Increase Was Larger Than the Import Bill Could Explain
Across all 57 tracked families, a symmetric accounting decomposition assigns +$52.02 billion of the 2024–2025 calculated-duty change to observed-rate changes and −$1.69 billion to changes in customs import value.
Editorial evidence cutoff: September 9, 2026. Published September 29, 2026. Observation periods are stated throughout; older figures are retrospective evidence.
For September 2026 cost planning, the completed 2024 and 2025 records offer a useful test of an easy explanation for a growing duty bill: perhaps import spending simply grew. This retrospective uses the observations already available by September 9. It separates the arithmetic factors before asking what any of them can say about the next purchasing decision.
The tracked manufacturing import bill shrank in 2025. The calculated-duty bill grew by $50.33 billion. Those two facts sit beside one another in the customs record, but they do not explain how much of the second movement is associated with the first.
An accounting decomposition of all 57 product families supplies the missing bridge. Changes in their observed duty rates contribute $52.02 billion to the increase. Changes in their customs import values contribute negative $1.69 billion. The smaller value base offsets part of the increase associated with rates; it does not explain the rising bill.
The distinction matters because a dollar total is the product of two moving quantities. Without separating them, a report can mistake a bigger buying program for a higher burden, or mistake shrinking purchases for evidence that the burden has eased. Here the recorded dollars support neither shortcut.
THE IMPORT BASE SHRANK BY ALMOST $17 BILLION
Across the same 57 families, customs value declined from $576.27 billion in 2024 to $559.30 billion in 2025. Calculated duties rose from $18.80 billion to $69.13 billion. These are annual observations from the saved USITC DataWeb snapshot, rather than an extrapolation from one unusually expensive shipment or month.
The panel includes manufacturing materials, components and equipment. It is not every US import. The families are held constant across the two years, so adding new categories cannot manufacture the result. Every annual duty denominator equals the reported customs value used in the calculation.
There is another distinction that must stay attached to the headline. Calculated duties are statistical amounts reported through the trade system. They are not an audited cash-collections ledger. The Census statistical definitions explicitly caution against equating the measure with amounts of duty paid. The investigation concerns the recorded burden, with that limitation intact.
TWO MOVING FACTORS, ONE EXACT RECONCILIATION
For each family, calculated duty equals customs value multiplied by its observed effective rate. The decomposition asks what happens when those two factors move between years. One ordering changes the rate first and import value second. The other changes value first and the rate second. Averaging the two assigns their interaction equally.
In practical terms, the rate component multiplies the change in rate by the average of the two years’ customs values. The value component multiplies the change in customs value by the average of the two rates. Adding the components reproduces the observed duty-dollar change for each family and for the entire panel.
This is an established symmetric decomposition, not a new economic identification strategy. Its usefulness is that it makes the accounting explicit and leaves no unexplained residual. A reader can reproduce the result from the two annual value-and-duty records for each family, without accepting a narrative about why buyers acted.
BATTERIES AND STEEL SHOW THE OFFSET AT WORK
For batteries, the rate component adds $2.91 billion while the customs-value component subtracts $408.12 million. The total increase is therefore smaller than the rate component considered on its own. Describing only the falling import value would miss the larger duty burden recorded against those imports.
Flat-rolled steel provides a second substantial example. Its rate component is positive $1.66 billion; its value component is negative $478.79 million. That negative contribution describes what the changed customs-value base does within the identity. It is not a measured benefit of a particular purchasing decision.
Other families have a different combination. Wire and cable has positive contributions from both rate and customs value. The full 57-family calculation shows why a shrinking aggregate import base does not imply that customs value fell in every purchasing category. Positive and negative value components coexist inside the panel, and their dollar weights determine the combined result.
THE RESULT SURVIVES BOTH WEIGHTING ORDERS
An accounting result can look decisive because of the way its interaction term was assigned. To test that concern, the rate contribution was recalculated using only 2024 customs values, then using only 2025 values.
The baseline-value calculation gives positive $53.14 billion. The current-value calculation gives positive $50.90 billion. The symmetric result, $52.02 billion, lies between them. Changing the ordering affects the attribution by a meaningful amount, but it does not reverse the conclusion: rate changes account for a large positive component while the value component offsets part of it.
The numerical reconciliation is also direct. The maximum discrepancy between the two components and a family’s recorded duty change is less than one-millionth of a dollar, a floating-point rounding residue. More important than that computational detail is the substantive check: no family or missing denominator was silently dropped to make the aggregate add up.
AN OBSERVED RATE IS NOT A PURE POLICY VARIABLE
The rate term should not be renamed the tariff-policy effect. A family’s observed rate can change because its products, origins, preference claims and other entry characteristics change. A broad family can contain goods with very different treatments, and their relative weights can move even if the published treatment of each individual item were unchanged.
Customs value is also not physical volume. Prices, quantities and product composition can all change its dollar total. The negative $1.69 billion component therefore cannot tell us how many units were not imported, why a firm changed an order, or how much any policy reduced trade.
The useful conclusion is narrower and stronger than those claims. In this matched panel, the recorded duty increase persists after separating the shrinking customs-value base. Investigating its causes requires the next layer of evidence: matched products, origins, entry treatment and quantities. The decomposition identifies where that investigation should begin; it does not substitute for it.
PUT THE BRIDGE BESIDE THE BILL
A manufacturing purchasing review can use the same structure on its own records. Start with verified duty and customs values, preserve the product scope, and show the rate and value components separately. Keep a third column for changes in specification or classification that make one year’s goods unlike the other’s.
That produces a clearer question for a supplier or broker than asking why the total bill went up. Which part reflects a different buying base, which part reflects a different recorded rate, and what transaction facts explain the rate movement? The national family data cannot answer those company questions, but it demonstrates why they need separate answers.
Source window: calendar 2024–2025; 57 matched families. Annual observations were carried forward in the saved snapshot fetched September 6, 2026. Calculations retain full dollar precision; displayed amounts are rounded.
Sources and evidence
Evidence period: Calendar 2024 versus 2025. The frozen evidence record lists the source files and verified hashes available September 9, 2026. Source revision: 5e4fb7726c3d40060c0151c086c903baae856cab. Later live-data updates do not alter the historical evidence in this article.
census.gov/foreign-trade/guide/sec2.html
link.springer.com/article/10.1007/s10888-011-9214-z
Published 2026-09-29.