The Hidden Cost of Tariffs: When Consumers Pay More Than the Tax Itself

Evidence from European wine tariffs shows how supply-chain markups quietly magnify trade taxes into higher household costs.

In the economic debate surrounding tariffs, many people look at the “percent pass-through,” the percentage increase in consumer prices compared to the percentage increase in tariffs. But even when the percent pass-through looks low at the cash register, consumers can still end up paying more than the government collects. That’s the central, and underappreciated, lesson from a new NBER working paper by Aaron Flaaen, Ali Hortaçsu, Felix Tintelnot, Nicolás Urdaneta and Daniel Xu—“Who Pays for Tariffs Along the Supply Chain? Evidence from European Wine Tariffs.”

Using the 2019 Airbus dispute tariffs on European still wines (the ≤14% alcohol by volume [ABV] category) as a clean natural experiment, the authors trace prices bottle by bottle from the foreign producer through the U.S. importer and distributor to the final retail sale. The picture that emerges demonstrates how markups and multitiered distribution systems transform an upstream tax into downstream inflation and shifting margins.

Wine is perfect to illustrate this point because U.S. alcohol markets legally separate the three tiers—importers, distributors/wholesalers, retailers—so the data reflect arm’s-length pricing at each stage. The tariff shock was targeted (25% on French, Spanish, German and U.K. still wines at ≤14% ABV in bottles ≤2 liters), with sparkling wines and >14% ABV still wines spared. So the authors can compare “treated” and “untreated” products that are otherwise similar, including within the same producers.

Percent vs. Dollar Pass-Through: The Hidden Amplifier

At the border, pass-through is incomplete: Foreign suppliers cut their prices by about 5.2% in response to the 25% tariff. Roughly one-quarter of the tariff revenue is borne abroad, and the rest lands in the U.S. economy. The importer then raises its sell price to U.S. distributors by 5.4%. Finally, retail prices paid by consumers rise by 6.9%. If you stop there and think in percentages, you might be tempted to say, “See? A 25% tariff, but only a 6.9% price increase. Tariff pass-through to consumers is modest.”

That would be the wrong conclusion because percentages hide the role of markups. The paper by Flaaen et al. draws a crucial distinction between percent pass-through and dollar pass-through. When a product moves through multiple stages, each with its own markup, small percentage changes early in the chain get magnified in dollars by the time the good hits the shelf. In the authors’ worked example, a bottle that cost $5 at the border before the tariff faces $1.19 in tariff liability. The estimated retail response translates to $1.59 more paid by the consumer. In other words, consumer dollar costs exceed tariff revenue. The tax wedge is amplified downstream.

Markups and Lagged Effects

That wedge shows up in margins, too. The importer’s gross markup contracts after the tariff, absorbing part of the shock. By contrast, the combined distributor/retailer markup expands. This is not “greedflation” hand-waving; it’s the mechanical result of compounding margins on a higher cost base with bargaining and participation constraints. In plain terms: When it becomes more expensive to put the bottle on the shelf, the stage that controls shelf space has leverage to keep its dollar margins whole, or better.

Timing matters as well. The importer’s prices adjust in roughly three months; consumer prices take almost a year to fully reflect the tariff, and they remain elevated well after the tariff is suspended. If you’re a policymaker or a central banker trying to map tariffs to inflation, that lag is not a rounding error. Inventory cycles and the number of rungs in the ladder determine when the pressure shows up in Consumer Price Index (CPI)-type measures.

Tariff Engineering

There’s another important piece of the paper that should temper how we read customs data: tariff engineering. Because the 2019 policy taxed ≤14% ABV wines but not >14% ABV, producers and importers pivoted. The authors use “certificate of label approval” files to show a sudden jump in new products registered above the 14% cutoff and even in “threshold switchers”—bottles of the same brand/appellation reappearing at 14.5% after the tariff, then drifting back once the tariff is gone. Aggregate unit values therefore become a muddled instrument for border pass-through: Product composition is shifting precisely in response to the policy. If you only look at average prices in customs data and conclude that the tariff’s cost has been fully absorbed by the foreign country, you’re likely misled.

Policy Lessons: The Real Incidence of Protectionism

There are three implications for the broader tariff moment we’re living through:

1) The incidence question is not “foreigners or us?” It’s “which us?” The paper documents that within the U.S., the importer takes a hit while downstream distribution improves its dollar margins. Policymakers who talk as if tariff revenue is a free lunch financed by foreign producers miss the fact that the domestic reshuffling, lower importer profits and higher distributor/retailer take can still leave households footing a bigger bill than Treasury collects.

2) Beware the comforting arithmetic of percent pass-through. In economies where Bureau of Economic Analysis-style gross markups between producer and purchaser prices are large (on average a little over 100% across personal consumption expenditure goods), the same logic that turns a 6.9% retail increase into $1.59 on a $23 bottle generalizes. With multistage markups, low percent pass-through can still mean high dollar pass-through. For inflation, disposable income and welfare, dollars matter.

3) The lag structure is policy-relevant. If we introduce or expand tariffs in Q2 and then ask in Q3 why the CPI hasn’t “fully reflected” the policy (or, worse, declare victory when CPI cools as inventories clear), we are kidding ourselves. This study’s approximately 12-month consumer price lag is a warning: Tariffs are a slow fuse that keeps burning even after they’re no longer in effect.

Two final virtues of the paper are methodological and conceptual. Methodologically, stitching together confidential importer invoices, distributor postings and retail prices at the product-vintage level gives us rare visibility into the same item through the entire chain. Conceptually, the authors show how easy it is for a standard “full pass-through in levels but not in percentages” result to emerge once we remember that a modern economy is layers of intermediation, not a single competitive market with a single markup. Their simple model with a retailer participation constraint squares the circle: incomplete border pass-through, a squeezed importer and a downstream sector that moves retail prices enough in dollars to more than cover the tax.

Tariffs are not paid in journal-article percentages; they are paid in household dollars. If you ignore markups and the supply chain, you’ll systematically understate how much tariffs really cost consumers—and you’ll miss who, within our own borders, quietly benefits.