Short time work (STW) policies provide subsidies for hour reductions to workers in firms experiencing temporary shocks. They are the main policy tool used to support labor hoarding during downturns, and were aggressively used during the COVID-19 pandemic. Yet, very little is known about their employment and welfare consequences. This paper leverages unique administrative social security data from Italy and quasi-experimental variation in STW policy rules to offer evidence on the effects of STW on firms’ and workers’ outcomes during the Great Recession. Our results show large and significant negative effects of STW treatment on hours, but large and positive effects on headcount employment. We then analyze whether these positive employment effects are welfare enhancing, distinguishing between temporary and more persistent shocks. We first provide evidence that liquidity constraints and rigidities in wages and hours may make labor hoarding inefficiently low without STW. Then, we show that adverse selection of low productivity firms into STW reduces the long-run insurance value of the program and creates significant negative reallocation effects when the shock is persistent.