Modern trade theory suggests several determinants of trade and specialization. They include technology, factor endowments, trade costs, tastes, and returns to scale. Various theories exist that explain the effects of these determinants and a number of empirical studies have been performed to check if these theories provide good explanations of the data.
There is evidence now that all of the above determinants of trade have a significant effect on the pattern of trade and specialization in production. It would be interesting then to somehow compare the importance of these determinants. Unfortunately, there is little theoretical guidance that can help us make this comparison.
To make this comparison empirically requires a model that incorporates all of these determinants of trade, and a methodology for measuring and comparing the importance of these determinants. This paper sets out to accomplish these tasks, with the exception of studying the effects of the returns to scale. The results presented in this paper will help us sort through the various determinants of trade and specialization and contribute to our understanding of how these determinants create the currently observed patterns of specialization and trade. (1)
The model used in this paper includes constant returns to scale, fixed factor endowments that are different across countries, industry-level productivity differences, and taste differences across countries. It also incorporates trade costs that are different for each industry and each pair of countries. To explain the within-industry two-way trade between countries, this paper allows for producer heterogeneity within industries using the framework of Eaton and Kortum. (2) The paper presents the evidence, both on the micro and macro levels, that supports this model.
The effects of various determinants of trade are studied by performing counterfactual simulations. The importance of the determinants is measured by the magnitudes of their effects on specialization and pattern of trade. The paper also checks if the importance of the determinants varies with country income and shows the effects of various determinants of trade on welfare.
Simulations produce several interesting results. Productivity differences across countries and industries are found to be the most influential determinants of trade and specialization. Preferences are found to be second most influential, signifying that the lack of attention paid to them in the trade literature is unwarranted. At the same time, factor endowments are found to be the least influential in the average country of the data set.
Trade costs are found to be very important determinants of trade and specialization. While previous literature has demonstrated their importance for the volume of trade, this paper shows how significant they are for specialization and the pattern of trade. The trade costs' direct effects on trade, which occur because trade costs differ across countries and industries, is found to be moderate. However, trade costs' indirect effect, whereby they interact with other determinants of trade, is very significant. For example, trade costs strongly link preferences with specialization and net exports. They also influence trade by affecting access to intermediate goods. In addition, trade costs limit the geographical range of comparative advantage, forcing it to be determined within a neighborhood of a country instead of the whole world.
This paper also finds that most of the above determinants have a greater effect on trade and specialization in poorer than richer countries. Interestingly, factor endowments stand out in this respect, because their influence is many times larger in poorer countries than in richer ones. Correspondingly, the effects of factor endowments on trade and specialization in richer countries can be called economically insignificant. This result gives support to the notion that the Heckscher-Ohlin explanation of trade is more relevant to poorer than to richer countries. …