Do Tariffs Lower Inflation?

Tariffs lower inflation: 150 years of evidence

Régis Barnichon and Aayush Singh employ a long-term historical approach to identify the causal macroeconomic effects of tariff shocks by exploiting 150 years, from 1870 to 2025, of US tariff policy variation, along with comparable data from France and the UK.

The authors rather than using standard econometric techniques on potentially endogenous policy decisions, they conduct a meticulous narrative historical analysis of major US tariff changes since 1870 to establish exogeneity. They find that because Republicans historically favored high tariffs to protect Northern industrialists, while Democrats favored low tariffs, representing Southern agricultural interests, tariff policy changes were not systematically responding to economic conditions.

Estimation approaches:

1. OLS with recursive identification: Treats tariff innovations as exogenous, ordering them first in Vector Autoregressions (VAR) with two lags of tariff changes, unemployment, and CPI inflation.
2. Instrumental variables: Uses narratively identified tariff shocks from nine dates explicitly motivated by long-run political considerations rather than cyclical conditions (1872, 1890, 1909, 1922, 1947, 1967, 1979, 1994, 2018).

Both approaches yield remarkably consistent results across full historical sample (1869-2020), pre-WWII period (1869-1939), and post-WWII period (1946-2020). The authors also examine responses of stock prices and volatility to explore potential transmission mechanisms.

Findings

• There is an inverse relationship between tariffs and inflation: Contrary to standard economic theory predicting that tariffs raise costs and inflation, the empirical evidence shows that a tariff hike of 4 percentage points raises unemployment by approximately 1 percentage point and lowers CPI inflation by roughly 2 percentage points over the short run. 
• Historical consistency: The negative inflation response is robust across dramatically different historical periods. Pre-WWII estimates, with large tariff variations of 10-15 percentage points annually, show the same directional effects as post-WWII estimates, despite the post-war period having much smaller tariff variation of -5 to 5 percentage points.
• International evidence: Using France and the UK as independent sources of variation yields similar conclusions: tariff increases lower inflation and contract economic activity, establishing generalizability beyond the US experience.
• Mechanism: The authors find evidence consistent with aggregate demand channels, rather than cost-push effects, that move unemployment and inflation in the same direction. Tariff shocks coincide with declining stock market valuations and increased stock price volatility, suggesting that uncertainty and wealth effects depress economic activity, leading to lower inflation despite higher import costs.

What Is a Tariff Shock? Insights from 150 years of Tariff Policy
Authors: Régis Barnichon, Aayush Singh
From: Federal Reserve Bank of San Francisco

Tariffs do not lower inflation: Measuring tariff pass-through with daily price data

Alberto Cavallo, Paola Llamas, and Franco Vazquez employ a novel high-frequency approach combining daily retail price microdata with detailed country-of-origin and tariff classification information to track the short-run transmission of 2025 US tariffs to consumer prices. The authors capture immediate price-level effects, i.e. 0-6 months, from tariffs raising import costs, which mechanically increase retail prices.

Data sources: Daily prices from five major US multichannel retailers (October 2024–September 2025) covering approximately 359,000 products representing 28.5% of official CPI weights. Country-of-origin information obtained through two methods: UPC matching to external databases (308,000 products) and generative AI predictions with 85-88% accuracy (covering food categories).

Tariff assignment: AI-based hierarchical classification assigns each product a 10-digit Harmonized System code, linked to both statutory tariff rates from HTS and Federal Register and applied/effective tariff rates, calculated as duties paid divided by customs value from Census Bureau import data.

Price index construction: Chained matched-model price indices following BLS methodology, computed as geometric means of daily price relatives within product groups (by country, domestic/imported status, tariff exposure). Pre-tariff trends estimated from October 2024–March 2025 serve as counterfactual benchmarks.

Empirical specifications:

• Event study around March 4 Liberation Day announcement with daily data, comparing exposed (imported and domestic-affected) to unexposed goods within categories
• Distributed-lag pass-through regressions at product level relating monthly price changes to applied tariff changes over six-month horizon
• CPI impact estimation aggregating sectoral indices using official expenditure weights scaled to sample coverage

Findings

1. Pass-through magnitude and timing: Retail prices responded immediately but adjusted gradually. Imported goods prices rose within days of March 4 announcement and continued increasing, reaching 5.4% above pre-tariff trends by September, versus 3% for domestic goods. Estimated retail pass-through is 14-20% after six months, meaning a 20 percentage point applied tariff raised retail prices by approximately 4%.
2. Domestic goods spillovers: Domestic goods in categories competing with imports exhibited price increases nearly matching imported goods, while unaffected domestic goods remained stable—highlighting competitive pricing mechanisms.
3. Cross-country heterogeneity: Chinese goods rose 4.6% relative to trend (facing 38% applied tariffs by July), Canadian goods 3.6% (3% applied rates), Mexican goods only 0.1% (5% applied rates with extensive USMCA exemptions). Turkey and Poland showed largest increases due to product composition (carpets, steel products).
4. Sectoral variation: Largest impacts in Furnishings & Household Equipment (6.7% above trend) and Miscellaneous Goods (5.2%), both heavily reliant on Chinese imports. Food showed smaller effects (1.6%) due to lower import penetration.
5. Distributional effects (“cheapflation”): Within categories, cheapest quartile prices rose 5% versus 2.5% for most expensive quartile, suggesting firms with lower markups passed through larger cost shares—implying regressive distributional consequences.
6. Aggregate inflation impact: Tariffs contributed approximately 0.7 percentage points to all-items CPI by September 2025. Without tariffs, August 2025 annual inflation would have been 2.2% rather than observed 2.9%—much closer to Federal Reserve’s target.

Tracking the Short-Run Price Impact of U.S. Tariffs
Authors: Alberto Cavallo, Paola Llamas, Franco Vazquez
From: Harvard University, Northwestern University, Universidad de San André