When a tariff changes, how long does trade take to respond — and does the usual estimate capture it?
The Long and Short (Run) of Trade Elasticities
Christoph Boehm, Andrei Levchenko, Nitya Pandalai-Nayar (2023) · American Economic Review 113(4)
Article ↗ · Replication data ↗ (openICPSR)
The archive requires an account we do not hold, so the files were not opened. What follows rests on the published abstract and on this platform's own observations.
What the authors found
Boehm, Levchenko and Pandalai-Nayar exploit an overlooked source of clean variation: when a country changes its Most Favoured Nation tariff, partners trading on MFN terms get a plausibly exogenous shock while partners with preferential agreements do not. Comparing the two groups across product-level tariffs and trade flows from 1995 to 2018, they estimate the trade elasticity at successive horizons and find it is far smaller in the short run than at long horizons — the response builds over years, so a short-run estimate substantially understates the eventual effect.
What our data adds
The slow adjustment they document is what the trade-war series in this database looks like from the outside. US goods imports from China fell 16 per cent in the first year after the 2018 tariffs, from 563 billion dollars to 472 billion — but seven years later the figure is 379 billion, a further 20 per cent below the 2019 level. If one had estimated the response from 2019 alone, one would have concluded the tariffs mattered roughly half as much as the eventual path suggests. That is the paper's point arriving in a single bilateral series.
| Figure | Value | Indicator |
|---|---|---|
| US imports from China, 2018 | 563bn USD | IMTS.MG_CIF_USD.A |
| One year later, 2019 | 472bn USD | IMTS.MG_CIF_USD.A |
| Seven years later, 2025 | 379bn USD | IMTS.MG_CIF_USD.A |
Every figure here is computed from observations in this database and can be charted, downloaded or queried.
What this cannot show
This is illustration, not estimation. Seven years of one country pair contains a pandemic, a shipping crisis and several further rounds of policy, none of which the paper's design has to contend with because it compares MFN partners against a control group. Reading the later decline as tariff response is exactly the confounding their identification exists to avoid.
The data behind it
Related notes
- When a country exports, how much of the value was actually made there — and how far does the usual export statistic mislead us?
- Who actually paid for the 2018 US tariffs, and what did they do to the pattern of trade?
- When two countries trade with each other and neither is the United States, whose currency is the price in?