Three Essays in Health Policy Evaluation.

Abstract

The first chapter of this dissertation examines the effect of the introduction of Medicaid between 1966 and 1970 on infant and child mortality rates. I exploit the federal requirement that Medicaid cover all cash welfare recipients, which meant that Medicaid eligibility inherited large cross-state differences in welfare receipt that had emerged decades before. I use a difference-in-differences model that compares state-level infant and child mortality rates before and after Medicaid (first difference) in states with higher and lower initial welfare-based eligibility (second difference). The results show that mortality rates in higher- and lower-eligibility states were indistinguishable prior to Medicaid, but immediately after states adopted Medicaid programs, nonwhite mortality rates fell by eight percent in high-eligibility states relative to low-eligibility states. Using newly-entered administrative data from 1963-1976, I show that children’s public insurance use increased by about six percentage points in the high-eligibility states relative to low-eligibility states. Medicaid can account for at eight percent of the aggregate decline in nonwhite child mortality from 1965 to 1979.

The second chapter examines the effect of Medicaid implementation on income-based disparities in children’s insurance coverage, health care use and medical spending. I document strong income disparities in health care use in the early 1960s and show that these disparities fell dramatically in the period after Medicaid implementation. I also use the 1963 and 1970 waves of the Survey of Health Services Utilization and Expenditure to show that after Medicaid, income disparities in insurance coverage and primary care use fell disproportionately in areas with higher pre-existing rates of welfare-based Medicaid eligibility.

The third chapter, written with Martha Bailey, estimates the effect of the Community Health Center (CHC) program on older adult mortality rates. CHCs were initially established between 1965 and 1974 and provided (rather than financed) primary care. We use data from the National Archives to construct measures of the county-level roll out of CHCs between 1965 and 1974 our estimates show that mortality for residents 50 and older fell sharply by two percent after CHC establishment, and that the effects persist for at least 15 years.