Abstract
This article estimates an economic model of the determinants of state government spending on health care benefits provided through the Medicaid program for the old, disabled, and young populations, respectively. Spending on the mutually exclusive recipient groups, rather than the aggregate, is examined to ascertain the extent to which one population’s costs supplant spending on the alternative populations. Endogeneity bias arising from the incentive effects of health care benefit guarantees on program take-up is addressed using an identification strategy that relies on measures of time-varying state resident participation in federally administered welfare programs to control for unobservable economic and noneconomic opportunities simultaneously determining Medicaid recipiency. It finds that state demand for health care generosity for each population is interrelated with the specific costs of the alternative populations. Simulations of eligibility expansions targeting each of the recipient populations illustrate the substitution effects evident in state Medicaid spending.