Occupational Exposure to Chinese Import Competition: Evidence from U.S. Occupation-Industry Level Data

Abstract

This paper analyzes how increased Chinese import competition affects U.S. labor markets at the occupation level. Prior research has addressed similar forms of analysis, but not on the occupation level with U.S. data. This paper constructs approximately 40,000 occupation-industry-period cells that match the Bureau of Labor Statistics’ Occupational Employment and Wage Statistics with bilateral trade data from the United Nations Comtrade database from 2002 to 2024. Building from Autor, Dorn, and Hanson (2013) and Basco, Liégey, Mestieri, and Smagghue (2025), this paper utilizes an occupation-level China exposure index that weights industry level import shocks by wage-bill intensity shares. Additionally, this paper instruments U.S. exposure to Chinese imports with the import growth of eight other high-income countries. The estimates from this analysis indicate that a one unit increase in occupational exposure reduces employment by approximately 21 percent but does not estimate any significant effect on wages. When the sample data is split between different sectors, the employment effect is concentrated entirely in non-manufacturing cells (coefficient: -0.283, p = 0.001), while manufacturing cells show no significant effect (+0.067, p = 0.170). This result does not mean that manufacturing is unharmed (mean manufacturing employment declines by 12.2 percent) but rather implies that the China Shock contracts all manufacturing occupations uniformly, leaving no cross-occupation variation for the regression to detect. This significant non-manufacturing result effectively captures supply-chain spillovers. This means that occupations with greater exposure to manufacturing experience worse employment outcomes even in their non-manufacturing sector cells. Because the regression identifies only differential effects across occupations, any uniform aggregate decline that affects all occupations equally is absorbed by the period fixed effects and goes undetected. The true employment losses attributable to Chinese import competition are therefore likely larger in magnitude than the estimates reported here.