This paper examines the extent to which the incidence of dual job holding is cyclically sensitive in the context of hours constraints on labor supply. Linear probability models of the incidence of dual job holding are estimated separately for each hours constraint regime. Selection effects are accounted for in a correlated random effects setting in which selection into overemployment, unconstrained hours, and underemployment is separately estimated each year from an ordered probit model. As measured by the local unemployment rate, transitory business cycle movements have no effect on the incidence of dual job holding. However, a sustained change in the local unemployment rate reduces the incidence of dual job holding among workers who are not hours constrained on their main jobs.