Trade Specialization in the Gravity Model of International Trade

Abstract

One criticism of the gravity model of international trade is that it takes no account of comparative advantage. This critique is particularly important when the gravity model is considered for policy applications such as identifying priority markets for trade promotion programs. For example, the potential for trade expansion might be greater with countries with complementary patterns of comparative advantage than those with similar patterns. We introduce a trade specialization variable into a gravity model to capture the degree of complementarity of trading partners' comparative advantage. This short paper describes the test of this thesis. Our general findings are as follows: The trade specialization index clearly distinguishes countries that are generally believed to be "most similar" from those that are believed to "most different." Its explanatory power in the gravity equation is good, comparing well with other established variables and it improves the overall goodness of fit of the gravity equation. It thus appears to be a useful addition to the gravity model toolkit. At the same time, it still leaves a large residual unexplained variation that weakens the gravity model's utility for some policy applications.

Keywords: Gravity model, comparative advantage, trade specialization index

JEL Classification: F14