Abstract
This paper studies how constraining physicians' working hours affects hospitals' healthcare provision and patients' health. We evaluate a 1982 reform in Denmark that strongly restricted the scheduling of overtime, weekend, and nighttime shifts. Consistent with an hours-to-workers substitution, hospitals' labor demand increased by 35% one-year after implementation. Drawing on administrative data covering the universe of hospitalizations in Denmark, we combine this labor demand shock with variation in exposure to the policy across hospital wards in a dose-response difference-in-differences strategy. As exposure, we use the pre-reform share of emergency cases in each hospital ward, that we show to be highly predictive of post-policy labor demand change. We find that a one-standard-deviation difference in exposure to overtime restrictions led to 2.4 fewer deaths in hospital for every thousand admissions (-5.8\%), three years after the onset of the reform. The overall mortality decline is driven by greater gains among patients with urgent medical needs, with no changes on post-discharge readmission rates or average length-of-stay. Our findings highlight how working hour regulations can help correct for life-costing inefficiencies in the healthcare sector.