The purpose of this paper is to study hospital costs in the event of introduction of a Prospective Payment System in France. We use a nested three dimensional database (stays- hospitals-years) in order to identify hospital unobservable heterogeneity and a transitory moral hazard component of cost variability. Econometric estimate are performed on a sample of 7,314 stays for acute myo-cardial infarction observed in 36 French public hospitals over the period 1994 to 1997. Transitory moral hazard is far from negligible: its estimated standard error is about 50 % of the standard error we estimate for cost variability due to permanent unobervable heterogeneity between hospitals. Simulations show that a cost reduction of about 16 % can be expecte...