One solution to uncompensated care, adopted by four states, is the formation of a state-wide program to reimburse hospitals for charity care and bad debt. The purpose of this study is to construct and test a behavioral model of the effects that these reimbursement funds have on hospital behavior, the utilization of medical care by low-income residents and the propensity to purchase private insurance. Uncompensated care reimbursement funds will be modeled as a type of insurance for hospitals and the consumer demand for insurance will be modeled as demand for a stop-loss policy. A reimbursement fund protects hospitals from losses due to uncompensated care. Like any insurance, these funds create moral hazard for hospitals which in turn may affect consumer behavior. Pooling funds over hospitals may increase moral hazard, and if hospitals are unable to accurately separate charity care from bad debt, patients who can pay as well as patients who cannot pay will receive free care. Thus, individuals may have less incentive to buy private coverage or to pay copayments and deductibles. Similarly, more generous Medicaid eligibility standards may also decrease the purchase of private insurance. These hypotheses will be tested using household data from the 1987 National Medical Expenditure Survey. A two-stage least squares model will be used to correct for the endogeneity of uncompensated care levels and the propensity to be uninsured. The effect of uncompensated care reimbursement funds and generous Medicaid eligibility standards on the propensity to purchase private insurance will be examined using Probit or Logit analysis. The impact of reimbursement funds on medical care utilization and the substitution of hospital-based services for non-hospital-based services will be examined using Tobit analysis.
