The GATT vs Inflation: Tokyo Drift

This paper demonstrates that ignoring the source of variation in tariff levels yields misleading conclusions regarding the effects of tariff changes. Specifically, we show that changes in the ad valorem equivalent (AVE) tariff are insufficient to quantify the magnitude or sign of the effect of tariff changes on economic outcomes. To illustrate this point, we construct the first database of annual US statutory tariffs from 1972 to 1988 and use it to explore the consequences of liberalization in the years spanning the Tokyo Round of the General Agreement on Tariffs and Trade (GATT). Decomposing the aggregate AVE into its statutory and endogenous components, we document two distinct liberalizations in our sample. Prior to 1979, inflation combines with specific tariffs to create an ``accidental liberalization'', despite the absence of widespread statutory tariff changes. After 1979, endogenous changes in the composition of imports mask the large Tokyo Round statutory tariff liberalization. Generalizing an exact hat framework to accommodate specific tariffs, we show that inflation-driven reductions in AVEs account for roughly half of the tariff liberalization in our sample, but these changes ultimately reduce imports and welfare. Finally, we show that specific tariffs remain an important determinant of cross-sectional variation in AVEs: between 2000 and 2017, one-third to one-half of the annual cross-sectional variance in good-level AVEs is driven by variation in the AVE of specific tariffs.