De-Routinization of Jobs and the Distribution Of Earnings: A Cross-Country Comparison

Abstract

The Routine-Biased Technological Change hypothesis (RBTC) by Autor et al. (2023) suggests that automation processes have substituted workers operating middle-skilled routine tasks. As a result, the relative demand for complementary workers operating non-routine tasks has increased. These changes in the labor force composition imply job polarization, characterized by a growing proportion of both high- and low-skilled at the expense of middle-skilled jobs. An aspect of high socio-economic and political relevance is the distributional implications of job polarization. In this study, we use a novel dataset covering 35 countries to investigate the phenomenon of job de-routinization, job polarization, and their potential ramifications on earnings distributions. We find strong empirical evidence of job polarization in most countries, but no systematic link between job polarization and the distribution of earnings. We show that this missing link stems from the fact that occupational classes are not very strongly stratified along the earning distribution.