New trade models, different distributions, same old results?

Abstract

In this paper we estimate Ricardian trade model of Eaton and Kortum (2002) using a different extreme value distribution, Weibull, for the productivity distributions of countries rather than the standard Fréchet assumption. The use of Weibull has the advantage of keeping elasticity of substitution in our estimation equation enabling us to account for demand effects. We find that incorporating this effect leads to lower comparative advantage parameter estimates. Additionally, we find that the state of technology (absolute advantage) rankings of the countries are similar in Weibull and Fréchet cases.