The Effect of Medicaid Expansion on Non-Elderly Adult Insurance Rates: Estimating the Treatment Effect on Non-Expansion States

We estimate the effect of Medicaid expansion on non-expansion states in 2014. We make three primary contributions to the existing literature. First, in contrast to the existing literature, we directly target treatment effects on non-expansion states (ETU). This effect is relevant if we wish to understand the magnitude of the foregone coverage gains associated with these states’ decision not to expand Medicaid. Second, we extend the synthetic controls approach to estimate this effect, which we call a “synthetic treatment.” Abadie et al. (2010) proposed synthetic controls to estimate counterfactuals absent an intervention; in this paper we adapt the methodology to predict the unobserved counterfactual given exposure to an intervention. Using this approach, we estimate that that the ETU is -2.08 (-1.34, -2.82). This estimate is lower in magnitude than comparable estimates of the treatment effect on the treated (ETT) in the existing literature. We provide evidence that factors associated with Republican governance drives this differential. This result highlights the mediating role of implementation and related factors in determining Medicaid enrollment. If the ultimate goal of the 2014 Medicaid expansion was to maximize insurance coverage, our results suggest that non-expansion states would have seen lower coverage gains than expansion states in 2014. The federal government may therefore wish to adopt policies to encourage enrollment among eligible individuals.