Gone with the crisis? Job Automation during Covid-19 Pandemic

Abstract: We test the hypothesis for the COVID-19 pandemic accelerating the automation of some jobs, leading people to change industries and occupations in Peru and the US. For that, we linked four-digit job occupation codes to different measures of job automation recently developed in the literature. We also link four-digit job industry codes to government identification of essential and non-essential industries, distinguishing industries most and least affected by restrictions to operate due to COVID-19. We exploit the identification above and build a triple difference model to estimate occupation mobility responses to COVID-19. We find that workers at high-automation risk in industries most affected by lockdowns were more prompt to lose their jobs during and after the crisis. We use panel data from the National Household Survey (ENAHO) in Peru and the Panel Study of Income Dynamics (PSID) in the US.