The Effect of Extended Medicaid on the Emergency Department Visits: An analysis of the 2008 Oregon Health Program

Abstract

I analyze the effect of the extended Medicaid program on the emergency department use behavior by using the data from the Oregon Health Program conducted in 2008. Although previous literature report that the Medicaid increases the probability of ED use, they fail to properly account for the endogeneity due to an underidentification. Our contribution is twofold. First, by a copula regression, we achieved a flexibility by using different joint distributions other than multivariate normal. Our result shows that the model performs better under the nonconventional likelihood function. Second, more importantly, we achieved identification of the parameters of interest by using the copula decomposition under the conditional independence assumption. Although our result agrees with the previous research that the Medicaid indeed increases the chance of ED use in Oregon, the average partial effect is estimated to be lower than the earlier estimates.