Essays on Technology and Trade

Abstract

This thesis consists of essays on technology and trade. In Chapter 1, I study how technology in the 21st century has changed the pattern of trade. I document that skill-abundant countries no longer have a comparative advantage in skill-intensive sectors. While this empirical relationship was strong in the 1980s, it weakened in the 1990s and disappeared by the 2000s. The decline is more pronounced in countries and sectors with higher automation. I find no such heterogeneous effects among countries and sectors more exposed to offshoring. Using a quantitative trade model incorporating automation and offshoring, I confirm that the observed changes in automation can account for the evolution of comparative advantage while observed changes in offshoring cannot. I conclude by revisiting the relationships between globalization, technology, and inequality through this model. Automation increases skill premia in developed countries with high automation and also raises welfare globally, whereas offshoring leads to smaller, more evenly distributed welfare gains.

In Chapter 2 (joint with Daniel G. O'Connor), we turn to the geographic consequences of technology and trade by analyzing the role of granularity—the dominance of a few large firms in local labor markets. We propose a new economic geography model featuring granular firms subject to idiosyncratic shocks. We show that average wages increase in the size of the local labor market due to that granularity, and provide a sufficient statistic for the contribution of our mechanism. We further prove that too few firms enter in equilibrium. Using Japanese administrative data on manufacturing, we provide evidence consistent with our mechanism and quantify it. Our mechanism implies that markets with around 2 firms per sector have an elasticity of wages to population of 0.05 and firms capture only 85% of their contribution to production in profits. In large markets like Tokyo, the elasticity is around 0.001, and firm entry is approximately efficient. Enacting optimal place-based industrial policy would increase the number of firms in modest-sized cities by more than 30% and actually decrease the number of firms and people in Tokyo.

In Chapter 3 (joint with Sagiri Kitao), we study the distributional consequences of technological and trade-induced polarization—wage and employment losses of middle-class workers relative to low- and high-skill groups. We build a model of overlapping generations who choose consumption, savings, labor supply, and occupations over their life-cycles, and accumulate human capital. We simulate a wage shift observed since the early 1980s and investigate individuals' responses. Polarization improves welfare of young individuals that are high-skilled, while it hurts low-skilled individuals across all ages and especially younger ones. The gain of the high-skilled is larger for generations entering in later periods, who can fully exploit the rising skill premium.