Pre‐trends and trade effects of temporary trade barriers

Abstract The identification of the trade effects of temporary trade barriers (TTBs) hinges critically on the parallel pre‐trends assumption and the counterfactual continuation of pre‐trends. Using US TTB investigations between 1993 and 2015, we document large and gradual pre‐trends: relative to imports of the same sector and country, import quantities of targeted goods increase by 50% and prices drop by 16% over the six years before the investigation. Trade effects are significantly smaller under a static approach than under an event study design. Under the latter, the long‐run effects differ significantly depending on the assumed continuation of pre‐tends. Under the parallel pre‐trends assumption, the trade destruction after 5 years is 43%; however, if the 6‐year linear pre‐trend is extrapolated, it is 64%. The treatment of pre‐trends is critical to determining the net per good trade effects. By examining pre‐trend heterogeneity, we also provide new insights into the “de facto” application of TTBs.