On Theories Explaining the Success of the Gravity Equation

We examine whether two important theories of trade, the Heckscher- Ohlin theory and the increasing returns theory, can account for the empirical success of the so-called gravity equation. Since versions of both theories can predict this equation, we tackle the model identi- fication problem by conditioning bilateral trade relations on factor endowment differences and on the share of intraindustry trade. Only for large differences in factor endowments does the Heckscher-Ohlin model predict perfect production specialization in different countries as well as the gravity equation, and trade is purely in goods produced with different factor intensities. Our empirical analysis yields three findings. First, the predictions of the perfect specialization versions of both theories are rejected by the data and so are unlikely explanations for the empirical success of the gravity equation. Second, a model of imperfect specialization that includes both increasing re-turns and factor endowments as sources of trade has a mixed performance: it correctly predicts production of more differentiated goods when the level of intraindustry trade is greater; however, the predicted link to factor proportions is tenuous. Third, the predictions of a model with imperfect specialization that relies solely on differences in factor endowments find support in the data. These results suggest that factor endowments and increasing returns explain different components of the international variation of production patterns and trade volumes.