This paper investigates whether short‐time work (STW) programmes achieve their stated goal of being devices intended to preserve jobs and keep workers employed in times of crisis. Our identification strategy exploits a change in the financial incentives provided to employers and employees for the temporary suspension of work contracts or the reduction of working time. We use longitudinal administrative data and estimate difference‐in‐differences regressions and instrumental variable bivariate probit models with endogenous covariates, which try to take account of the potential endogeneity of participation in STW. Our results suggest that discretionary policy changes in the incentives of STW schemes can be effective in the short run but they lose their ability when the decline in demand and the lack of work are more permanent.