Different Trade Models, Different Trade Elasticities?

How has the development of new trade models changed our understanding of the welfare gains from trade? Answering this question depends solely on estimates of the trade elasticity obtained using techniques applicable across different models. In this paper we build on the methods of Simonovska and Waugh (2011) and we develop a common estimator for the trade elasticity that is applicable across different models that feature micro-level heterogeneity. The benefit of our approach is that, while the estimation uses the same moment conditions, it allows for different micro structures to matter. We apply the estimator to the models of Eaton and Kortum (2002), Bernard, Eaton, Jensen, and Kortum (2003), and a variant of the framework of Melitz (2003) and Chaney (2008). We find that the trade elasticity estimates differ considerably across models. The results suggest that the Bernard, Eaton, Jensen, and Kortum (2003) model yields the highest, while the Melitz (2003) model yields the lowest welfare gains from trade.