Measuring Changes in Disparity Gaps: An Application to Health Insurance

We propose a method for reporting how program evaluations reduce gaps between groups, such as the gender or Black-White gap. first show that reduction in disparities groups can be written difference conditional average treatment effects (CATE) each group. Then, using Kitagawa-Oaxaca-Blinder-style decomposition, we highlight these CATE decomposed into unexplained differences other observables versus composition across (e.g. “endowment”). Finally, apply this approach to study impact of Medicare on American's access health insurance.