Robots and Wage Polarization: The effects of robot capital by occupation

This paper examines the distributional impacts of the increased utilization of industrial robots, emphasizing their role in specific tasks and their international trade. The study constructs a novel dataset based on tracking shocks to the cost of acquiring robots from Japan, termed the Japan Robot Shock (JRS), and analyzes these across various occupations that have adopted robots. A general equilibrium model is developed which incorporates robot automation in a large open economy, and a model-implied optimal instrumental variable (MOIV) is constructed from the JRS to address the identification challenges posed by the correlation between automation shocks and JRS. The analysis reveals that the elasticity of substitution (EoS) between robots and labor is heterogeneous across occupations, reaching up to 3 in production and material-moving jobs, which is significantly higher than the EoS between other capital goods and labor. The findings suggest that robots significantly contributed to wage polarization in the U.S. from 1990 to 2007.