Description:
This paper provides new insights regarding trends prior to temporary trade barrier (TTB) investigations and shows how they challenge the identification of trade effects. Using TTB cases initiated in the United States between 1993 and 2015, we compare static and dynamic treatment specifications with one that accounts for pre-trends. The static specification yields a considerable downward. Six years before the TTB investigation, imports from subject countries were 50 percent lower, while prices were 16 percent higher. Imports surge almost linearly until the TTB investigation. These pre-trends lead to sizeable differences between the two dynamic specifications. Comparing the dynamic specification with the one that accounts for extrapolated pre-trends, we find that import values are 43 percent lower five years after the TTB initiation, while they are 65 percent lower for the latter. This difference is critical for the net trade effects as trade destruction outweighs trade diversion only when extrapolating pre-trends