Abstract

This paper asks whether the comovement of economic activity across different sectors in the United States was different during the Covid recession than during preceding recessions. A dynamic factor model fit to data for the pre-Covid period is used to characterize historical patterns of comovement. When extrapolated over the Covid period, the fitted values from this model show how the various sectors were predicted to evolve based on their pre-Covid correlations. The paper documents large deviations from these predicted patterns. Including an additional factor for the Covid period improves the fit substantially. This new ‘Covid factor’ takes on large values from March to June 2020, but was less important after June 2020. This Covid-factor augmented model provides a parsimonious representation of comovement during the Covid recession and its aftermath.