International trade and cross-country capital composition

Most of the world's equipment is produced in a small number of rich countries. In 1996, countries in the top decile of cross-country income distribution produced 61% of world equipment and countries in the bottom decile produced only 0.2%. Rich and poor countries also differ in their dependence on imports for equipment. In 1996, poor countries imported more than half of their equipment. Structures, on the other hand, are largely domestically produced. World pattern of production and trade in equipment and structures is potentially an important determinant of composition of capital across countries.

The composition of capital differs significantly across rich and poor countries. In 1996, equipment constituted over 21% of the capital in 5 richest countries and only 8% in 5 poorest countries. While equipment capital-output ratio was a factor of more than 6 between rich and poor countries, structures capital-output ratio was less than a factor of 2. In this dissertation, I determine the quantitative relationship between international trade and cross-country capital composition. I, then, utilize the results on this relationship to examine the implications for economic development.

The starting point of my analysis is a multi-country model of trade in capital goods. There are three tradable sectors: equipment, structures and intermediate goods. Countries differ in their average level of productivity in each of the tradable sectors. International trade is subject to bilateral iceberg costs, which comprise of tariff and non-tariff barriers to trade. The theoretical model implies that the composition of capital is a function of country-specific productivity parameters and bilateral trade costs. I structurally estimate these parameters to match the pattern of bilateral trade in a sample of 76 countries.

Equipped with country specific productivity parameters and trade costs, I determine the quantitative relationship between international trade and cross-country capital composition. The calibrated model generates capital composition differences consistent with the data. Variation in log equipment to output ratio is 1.09 in data and 1.26 in the model. The model also generates cross-country differences in investment rate, income per worker and prices consistent with the data.

Through counterfactual exercises, I study the gains associated with reductions in trade costs. If all trade costs are eliminated, poor countries' welfare would increase by 39% and rich countries' welfare would increase by only 8%. If barriers only to equipment trade are eliminated, poor countries' welfare gain would be 9% and rich countries' welfare gain would be 1.4%. Reductions in barriers to flow of capital goods facilitate a more efficient allocation of the world stock of capital goods across countries and hence, are quantitatively important for economic development.