Estimates of moral hazard in health insurance markets can be confounded by adverse selection. This paper considers a plausibly exogenous source of variation in insurance coverage for childbirth in California. We find that additional health insurance coverage induces substantial extensions in length of hospital stay for mother and newborn. However, remaining in the hospital longer has no effect on readmissions or mortality, and the estimates are precise. Our results suggest that for uncomplicated births, minimum insurance mandates incur substantial costs without detectable health benefits. * Josh Angrist, Janet Currie, David Cutler, Carlos Dobkin, Lena Edlund, Randall Ellis, Michael Greenstone, Hilary Hoynes, Rick Hornbeck, Ellen Meara, Doug Miller, Roberto Rigobon, Jon Skinner, Tom Stoker, Tavneet Suri, and seminar participants at Harvard University, Boston University, BYU, and UC Irvine provided helpful comments and discussions. We also thank Jan Morgan of the California Healthcare Information Resource Center for helpful advice and discussions, Nicole Radmore for help with the National Hospital Discharge Survey data, and Sammy Burfeind, whose birth inspired our identification strategy.† Columbia University and NBER: da2152@columbia.edu ‡ MIT and NBER: jjdoyle@mit.edu 1 VOL. VOLUME NO. ISSUE
LENGTH OF POSTPARTUM HOSPITAL STAYS 1
AbstractEstimates of moral hazard in health insurance markets can be confounded by adverse selection. This paper considers a plausibly exogenous source of variation in insurance coverage for childbirth in California. We find that additional health insurance coverage induces substantial extensions in length of hospital stay for mother and newborn. However, remaining in the hospital longer has no effect on readmissions or mortality, and the estimates are precise. Our results suggest that for uncomplicated births, minimum insurance mandates incur substantial costs without detectable health benefits.