The Labor Market Incidence of New Technologies

Abstract

This paper develops a general framework for evaluating the incidence of labor market shocks, with a particular focus on automation technologies and artificial intelligence. Labor market shocks are shared among workers across occupations depending on their substitutability. Central to our theory is the concept of distance-dependent elasticity of substitution (DIDES), where substitutability between occupations declines with their distance in skill space. Consequently, shocks concentrated in skill-adjacent occupations lead to limited employment shifts but substantial wage disparities, leading to an uneven distribution of the economic incidence. We develop our model in both static and dynamic settings, embedding a Roy framework with a correlated productivity distribution to capture realistic labor substitution patterns. Empirically, we measure technological exposures using LLM-based task evaluations and estimate the substitution structure with employment and wage effects of automation technologies. Our findings indicate that automation and AI both cluster strongly in skill-adjacent occupations, restricting worker mobility and exacerbating labor market inequality. The dynamic model further demonstrates that limited mobility persists both during the transition and in the long run, constraining the labor market's ability to absorb a rapid AI adoption.

Keywords: Automation, Artificial Intelligence, Labor Market Incidence, Distance Dependent Substitutability, Skill Space