Working Paper: NBER ID: w10204
Authors: Gautam Gowrisankaran; Robert J. Town
Abstract: We seek to investigate whether managed health care can affect mortality, and if so, through which mechanisms. We estimate the impact of Medicare+Choice (M+C), Medicare's managed care program, on elderly mortality, using a county-level panel from 1993 to 2000. We control for endogenous M+C penetration rates with county fixed effects and instrumental variables. We construct instruments using the identification created by the fact that M+C payment rates are based on 3 to 8 year lagged fee-for-service (FFS) costs in the county. We find that enrollment in managed care without prescription drug coverage significantly increases mortality while enrollment in managed care with drug coverage has no significant impact, both relative to FFS. The impact of managed care penetration on mortality from heart disease appears to follow a similar pattern. The estimates suggest that a 10-percentage point increase in M+C non-drug coverage would cause 51,000 additional deaths among the aged population in 2000.
Keywords: managed care; mortality; elderly; Medicare; drug coverage
JEL Codes: I11; I12; I18; C33
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Managed care (MC) penetration without drug coverage (I13) | Elderly mortality (J26) |
| Managed care (MC) penetration with drug coverage (I13) | Elderly mortality (J26) |
| Managed care (MC) penetration (I11) | Elderly mortality from heart disease (I12) |