What have independent analyses concluded about tariffs' effect on U.S. consumer prices since 2018?
Executive summary
Independent academic and central‑bank analyses conclude that U.S. tariffs since 2018 raised import and retail prices in a measurable way — in many studies the costs were largely borne by U.S. importers and consumers — but the macroeconomic inflationary effect is modest and sensitive to tariff size, scope, and behavioral responses [1] [2] [3]. Estimates of the uplift to headline consumer price measures vary: some place the immediate effect in the tenths of a percentage point range while scenario and model‑based work shows possible short‑run rises around 1% or more under stronger assumptions [4] [5] [3].
1. The 2018–19 trade war: near‑complete pass‑through into import prices and big distributional effects
Close empirical work on the 2018–19 tariffs — notably Amiti, Redding and Weinstein and follow‑up NBER work — finds that tariffs were essentially “passed through” into domestic import prices and that the incidence fell on U.S. consumers and importers rather than foreign producers, producing sharp price increases for intermediates and final goods and a measurable hit to real U.S. income (including an estimated $1.4 billion per month decline in real income by late 2018) [1] [2].
2. How large were the headline inflation impacts?
Independent institutions differ but cluster on modest aggregate effects: Richmond Fed calculations translating empirical pass‑through into headline measures put the 2018–19 episode at roughly a 0.3 percentage‑point rise in the CPI attributable to the change in effective tariff rates [4], while banking and policy analysts produced higher scenario estimates — J.P. Morgan projected PCE prices might rise 1–1.5% under some measures of tariff increases [3] and Yale’s Budget Lab modeling estimated a roughly 1.3% short‑run rise in consumer prices for a particular 2025 tariff package assuming full pass‑through [5].
3. Why estimates vary: mechanisms, scope and assumptions matter
Differences in quantitative conclusions stem from methods and assumptions: some studies use direct import‑price pass‑through and input‑output accounting, others use daily retail barcode data or macro empirical techniques, and modelers often assume full (100%) passthrough for convenience or as a central case — an assumption that materially raises projected consumer price effects [6] [7]. Empirical studies sometimes find retailers absorb part of the cost in margins rather than immediately hiking retail prices, while input‑price channels and reduced product variety create welfare losses that aren’t fully captured by headline CPI movements [8] [9] [10].
4. Time profile and behavioral responses: slow, partial pass‑through and supply‑chain shifts
Several analyses emphasize that price effects can be gradual: researchers tracking daily prices after the 2018–19 measures and newer episodes find pass‑through can unfold over months or years rather than instantaneously, and firms and consumers respond by reshoring, switching suppliers, or dropping products — responses that mute headline inflation but reduce variety and real welfare over time [10] [11]. Central‑bank work also notes that initial tariff episodes can resemble negative demand shocks (pulling back spending and dampening inflation) before pass‑through and second‑round effects lift prices later, making short‑run measurement tricky [12].
5. Net judgment and open uncertainties
The independent literature converges on two core conclusions: tariffs raise import prices and much of the burden tends to fall on U.S. buyers, and the macro effect on headline inflation is measurable but typically modest unless tariffs are large, prolonged, and accompanied by limited exchange‑rate offset or rapid supply‑chain substitution [1] [4] [7]. Important caveats remain: point estimates vary by methodology and by which tariffs and time windows are studied; model‑based scenarios that assume 100% passthrough or no behavior change generate larger price effects [6] [5], while more recent Fed work finds pass‑through can be weaker today than in 2018–19 because of changing trade shares and delayed implementation [11].