We develop a multi-sectoral matching model to predict the impact of the lockdown on the US unemployment, considering the heterogeneity of workers to account for the contrasted impacts across various types of jobs. We show that separations and business closures that hit the workers with the first level of education explains the abruptness of the unemployment rise. The existence of significant congestion externalities in the hiring process suggests that a comeback to the pre-crisis unemployment level could be reached in 2024 in a scenario with a double wave. In the same scenario, a calibration on French data leads to more pessimistic forecasts with a comeback to the pre-crisis unemployment level expected until 2027.
COVID-19 unemployment dynamics search and matching worker heterogeneity