Description

This study analyzes the relationship between machines and employment across different workers' income quintiles, utilizing firm-level and worker-level microdata for Chile from 2009 to 2023, considering a total of 80,000 firms and 2,900,000 workers. To investigate this dynamic, a panel regression is used, modeling employment as a function of machines and a set of other covariates. Additionally, a generalized propensity score is used to address the endogeneity problem. The results indicate that the relationship between machines and employment is predominantly negative, especially for workers in the middle-income quintiles and for certain economic sectors such as business services, transport, and information & communication. Furthermore, the results suggest that the relationship between machines and employment in high-income quintiles has been positive, supporting the idea that technological development requires highly qualified workers. Overall, the results suggest that automation has had heterogeneous effects on employment in the Chilean labor market, and these effects 
are smaller than those suggested by the literature.