The effects of free trade agreements on product-level trade

We use highly disaggregated trade data to test whether "least traded products"-products either not traded previously or only traded in small amounts-experience faster trade growth after the signing of trade agreements and whether this is important for the aggregate trade creation effects of trade agreements. Because we use PPML, an estimator with unique aggregation properties, our product -level estimates are comparable to estimates based on aggregate trade data that are meant to capture aggregate trade creation effects. Using this approach, we find that virtually all of the trade creation associated with trade agreements comes from least traded products. Interestingly, for products that are very highly traded, we find that trade actually declines following an agreement. These results are robust to controlling for general mean reversion effects in the product -level trade data, which we also find evidence for, as well as for differences in observed tariff levels.