Why does employment in all major sectors move together over the business cycle?

Abstract

In recessions, employment falls in all major sectors. Positive correlation of employment across sectors is a puzzle, because a standard two-sector business-cycle model driven by aggregate productivity shocks predicts negative correlation of total hours of work in the consumption-goods sector and the investment-goods sector. I start from the observation that most of the variability of total hours worked takes the form of variations in the number of workers. Hours per employed worker is only a secondary source of variation. The extensive margin is therefore critical in understanding the positive correlation of sectoral labor market variables, yet neglected by existing studies. This paper advances the literature on cross-sectoral correlation of employment by making unemployment an explicit feature of the model. I construct a novel two sector model with search and matching friction, capital adjustment costs, and partial wage stickiness. The model explains the positive cross-sectoral correlation through movements of workers in both sectors into and out of unemployment.