The Effects of State Medicaid Expansions on Medicare Beneficiaries

Prior research shows that several states’ Medicaid expansions to working-age adults led to improvements in health status and increases in healthcare utilization for new Medicaid recipients (e.g., Sommers et al., 2012; Finkelstein et al., 2012).  However, little is known about how Medicare beneficiaries in those states were impacted by the sharp increase in utilization by working-aged adults.  If healthcare providers face capacity constraints, additional utilization by patients covered by Medicaid could lead to delays in care and possible deteriorations in health status for the already-insured Medicare population.  A few studies report that physician labor supply responds to expansions in insurance coverage for children (e.g., Garthwaite, 2012), and others examine whether Massachusetts’ health reform had spillover effects on Medicare beneficiaries (Joynt et al., 2013; Bond and White, 2013). Although Joynt et al., (2013) find no effect on preventable hospitalizations, Bond and White (2013) find evidence of reduced primary care utilization among Medicare beneficiaries in areas of Massachusetts with the largest coverage increases. To the best of our knowledge, ours is the first study to examine whether past state Medicaid coverage expansions to working aged adults have spillover effects for patients enrolled in Medicare in the form of reductions in healthcare access, utilization, and outcomes. This is an important question because estimates suggest that the Affordable Care Act (ACA) will expand Medicaid coverage to 11 million Americans (Congressional Budget Office, 2012).

To test for spillover effects of Medicaid expansions on the Medicare population, we employ data from the nationally representative Medicare Current Beneficiary Survey (MCBS) Cost and Use files from 1994-2010. We focus on Medicare beneficiaries in states that had the largest Medicaid expansions for working-age adults in the 2000s, specifically, Arizona, Illinois, Maine, Michigan, New Mexico, New York, and Oregon. We estimate models of various measures of healthcare utilization and outcomes for Medicare beneficiaries as a function of Medicaid expansions for adults aged 20-64.  We test for incremental changes in these outcomes after Medicaid expansions over and above changes in these outcomes over the same period in states that do not implement Medicaid expansions using difference-in-differences. Measures of healthcare use include whether or not the individual used any office visits or inpatient care and the number of physician visits or hospitalizations, as well as receipt of specific forms of preventative care, such as flu shots, pneumonia vaccines, eye exams, and mammograms. We will also examine the 1996-97, 1998-99, 2000-01, and 2004-05 rounds of the Community Tracking Study (CTS) Physician Survey to quantify the impact of Medicaid expansions on physician willingness to accept Medicare patients.

We consider Medicaid expansions in multiple states and a range of outcomes in order to better understand the consequences of coverage expansions for Medicare beneficiaries. As the Medicare population continues to grow and coverage for non-elderly adults increases under the ACA in coming years, understanding how supply-side responses to increased non-elderly coverage affect those receiving Medicare is important for ensuring access to high quality care for the Medicare population.