Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

One of the fundamental questions in health economics is how health insurance affects the demand for medical care. In general, health insurance causes ex post moral hazard (that is, an increase in the demand for medical care as a result of the decreased net price of care). Moreover, health insurance that is specific to just one type of medical care - prescription drugs, long-term care, or mental health care - could influence consumption of other types of medical care. This change in medical care consumption stems both from the differential cost-sharing features of insurance for different types of care as well as the relative effectiveness of each type of care in producing or maintaining health. The resulting changes in morbidity and mortality affect all future medical care expenditures. We use data from the longitudinal Medicare Current Beneficiary Survey Data (MCBS) from 1992 to 2001 to jointly estimate a system of dynamic empirical equations representing supplemental insurance coverage decisions, drug and other medical care demand, and health production. Specifically, our findings quantify the effect of prescription drug coverage (through Medicaid, employer and private insurance plans, or Medicare's managed care option) on the demand for drugs as well as hospital and physician services among Medicare beneficiaries. We also examine the effect of each medical care input on chronic condition status, functional status, and mortality, and the effect of health on subsequent medical care consumption over time. We evaluate the long-run (five-year) effect of drug coverage by simulating behavior under different drug coverage scenarios and updating endogenous explanatory variables year by year. Universal prescription drug coverage would increase prescription drug expenditures in our sample by 7 to 27 percent over five years (depending on the type of drug coverage provided). The associated changes in hospital and physician service expenditures differ depending of the source of drug coverage and the subpopulation of interest, but some offsets in expenditures are realized. While some of the increase in total expenditures is directly attributable to changes in insurance, the increase results from changes in health as well. Long-run survival probabilities increase, leading to larger proportions of elderly survivors with functional limitations. Our projections of changes in both expenditures and health, however, are smaller than those produced by extrapolating static models that fail to incorporate the dynamic consequences of increased prescription drug use on health and consumption of other Medicare-covered services.

This paper extends the literature on moral hazard induced by health insurance in several ways. We are the first to model the dynamic effects of insurance and drug coverage on health and Medicare-covered expenditures over time. A few papers have tried to estimate the static effect of prescription drug coverage on other forms of medical care expenditures, but never before in a dynamic framework. Static models miss much of the total effect of prescription drug coverage, because prescription drug use affects future morbidity, mortality, and medical care expenditures, not just current ones. Furthermore, because our model allows for both permanent and time-varying heterogeneity, we show that medical care behavior of the elderly is highly correlated over time. Our policy simulations not only show modest cost offsets over five years, they break down the changes into morbidity and mortality effects.