Who Pays When Everyone Adjusts Differently
The subject in this chapter is the distribution of the 2025 U.S. tariff burden across households. The focus is not simply which income group cut spending the most, but how price incidence, budget flexibility, category choices, and the definition of welfare interact. The reader should be able to distinguish a larger dollar adjustment from a larger hardship, and to read the available welfare estimates as answers to different questions rather than as competing versions of one number.
That distinction is necessary because the household data show a superficially surprising pattern: low-, middle-, and high-income households reduce spending on exposed goods by broadly similar amounts, yet low-income households experience the largest price pass-through and the greatest loss relative to income. The same total change in spending can therefore be produced by different constraints. One household may pay more for necessities; another may remove discretionary purchases; a third may make a larger dollar cut that represents a smaller sacrifice.
Similar spending changes conceal different burdens
The Federal Reserve study divides households into three broad annual-gross-income groups: below $40,000, $40,000–$125,000, and above $125,000. It estimates price, spending, and quantity responses separately for each group using the benchmark tariff-exposure measure. The price coefficients decline with income. A one-percentage-point increase in exposure is associated with about a 0.28-percentage-point price increase for low-income households, 0.20 percentage points for middle-income households, and 0.15 percentage points for high-income households.
This ordering is the first distributional result. The tariff is not passed through uniformly to the baskets observed in the panel. Low-income households face the largest price response, while high-income households face the smallest. The paper relates that pattern to differences in shopping behavior, retailer pricing, and the availability of close substitutes. Higher-income households may be better positioned to change product quality or retailer, while lower-income households may have fewer practical alternatives and must absorb more of the price movement in continuing purchases.
The spending coefficients, however, are much closer: about −0.54 for low-income households, −0.56 for middle-income households, and −0.57 for high-income households per percentage point of exposure. High-income households show the largest absolute quantity adjustment. It would be a mistake to read this as evidence that the groups bear the same cost. The quantity or spending response records an adjustment in the covered basket. It does not reveal how much income that adjustment consumes, whether it represents a forgone necessity or a forgone luxury, or how much room remains after the change.
The similar spending responses arise from opposing forces. High-income households have more discretionary spending and can defer or eliminate purchases without immediately threatening baseline consumption. Low-income households have less slack, but they face higher pass-through and have fewer options for reorganizing the basket. A low-income household can therefore make a smaller physical adjustment while bearing a larger price burden. A high-income household can make a larger dollar reduction while avoiding a proportionately large loss of resources.
The study’s income-specific price indexes also show that the basket purchased by low-income households experienced higher cumulative price growth over the sample period than the baskets of middle- and high-income households. This matters because welfare depends on the prices of what a household actually buys, not on a single population-wide average. A tariff applied to a commonly purchased essential can matter more to a constrained household than a larger price movement in a product that household can simply stop buying.