Three Essays in International Trade and Macroeconomics.

Abstract

The first chapter studies the impact of globalization on the income gaps between the rich and the poor. This paper presents a new piece of empirical evidence showing that executive-to-worker pay ratio is higher among exporting firms than non-exporting firms. It then builds a model with heterogeneous firms, occupational choice, and executive compensation to model analytically and assess quantitatively the impact of globalization on the income gaps between the rich and the poor. The key insight of the model is that the "gains from trade" are not distributed evenly within the same firm. The compensation of an executive is positively linked to the size of the firm, while the wage paid to the workers is determined in a country-wide labor market. Any extra profit earned in the foreign markets benefits the executives more than an average worker. The model is then calibrated to create a counterfactual world where the only source of change is the access to the global markets. Model simulations show that around one-third of the surge in top income shares in the U.S. can be attributed to globalization between 1988 and 2008. The second chapter, joint with Ruediger Bachmann, studies the relationship between nonconvex capital adjustment costs at the firm level and aggregate investment dynamics. We study this question quantitatively with a two-sector lumpy investment model with inventories. We find that with inventories, nonconvex capital adjustment costs dampen and propagate the reaction of investment to shocks: the initial response of fixed capital investment to productivity shocks is 50% higher with frictionless adjustment than with the calibrated capital adjustment frictions, once inventories are introduced. The last chapter, joint with Ruediger Bachmann and Andrei Levchenko, presents a set of novel empirical facts that the aggregate U.S. imports, exports, and real exchange rate show conditional heteroscedasticity. We estimate two ARCH family time series models and show that conditional heteroscedasticity is statistically significant for imports and exports between 1970 and 2012, and for the real exchange rate between 1973 and 2012.