Why Households Cut Spending Before They Stop Buying Essentials

The subject here is the household response to the 2025 U.S. tariff shock: not only what prices did, but how people changed quantities, category participation, and product choice. The specific problem is the large gap between the modest retail-price increase and the much larger decline in spending on affected goods. The Federal Reserve evidence helps resolve that gap by separating a common contraction in purchasing from a more selective, belief-driven change in the composition of the basket.

That distinction matters because “households bought less” can describe several different behaviors. A family might buy fewer units of an essential item, abandon a discretionary category, delay a purchase, or continue buying the same number of units but choose a cheaper variety. These choices have different implications for living standards and for how the tariff’s cost is measured. The evidence points to a combination of margins rather than one uniform reaction.

The large response is quantity-led, but not one uniform quantity response

The Federal Reserve’s aggregate estimates start with three outcomes measured for the same covered retail categories: a price index, real spending, and quantity. In the benchmark exposure specification, a one-percentage-point increase in tariff exposure is associated with an approximately 0.20-percentage-point increase in prices, a 0.63-percentage-point decline in real spending, and a 0.57-percentage-point decline in quantity. At the average increase in exposure during the period, the implied price increase is about 1.2 percent while real spending falls about 3.9 percent.

The near alignment of the spending and quantity coefficients is informative. Since the paper deflates spending by the aggregate PCE goods price index, the two measures differ mainly when category-specific prices diverge from that deflator. The observed spending contraction is therefore driven largely by fewer units or less participation in the affected categories, not simply by paying more for an unchanged basket. The result is consistent with a household moving along a quantity margin, but it does not identify one single reason for every reduction.

It would be tempting to divide the spending coefficient by the price coefficient and call the result a demand elasticity. The paper explicitly warns against that reading. The coefficients come from different equations and comparisons, and the spending outcome is aggregated differently from the price index. The ratio is better understood as a diagnostic: a price-only explanation would require a very large response for ordinary groceries and household goods. That makes room for precaution, uncertainty, and category-level decisions that are not captured by a mechanical response to the price change.

The location of the contraction gives the next clue. The study classifies grocery, household, and health-and-beauty goods as essential, with other covered sectors grouped as non-essential. Under the exposure specification, the essential categories show no statistically significant spending or quantity response, while non-essential spending falls about 0.46 percent and non-essential quantity about 0.39 percent for each percentage point of exposure. The authors caution that splitting the sample reduces the cross-category variation that identifies the pooled price response. A non-significant essential coefficient is therefore not proof that no essential price changed.