Welcome Mat Effects for Dual Medicaid and Medicare Eligibles

The Affordable Care Act gives states the option to expand the Medicaid program to working-age adults with incomes up to 138% of the federal poverty level. This study investigates whether Medicaid expansions to working-age adults increased Medicaid participation by seniors eligible for dual enrollment in Medicare and Medicaid. We hypothesize that Medicaid expansions to working-age adults increase Medicaid program participation among seniors eligible for dual enrollment by raising awareness of program benefits, facilitating enrollment, and reducing the stigma of program participation. While prior studies have documented “welcome mat” effects, these studies focus on increases in participation by working-age Medicaid-eligible adults (Sommers and Epstein, 2011; Sommers, Kenney, and Epstein, 2014), and no prior study has examined welcome mat effects in the context of Medicare beneficiaries who also qualify for Medicaid. Welcome mat effects for this population have the potential to be large, since studies show that only one-third to one-half of those eligible for both programs enroll in Medicaid (Pezzin and Kasper, 2002; Ungaro and Federman, 2009).  Because Medicaid and Medicare program costs are impacted by both changes in the size and average health needs of the dual population, we also examine the general health status of those eligible duals who are more likely to participate in Medicaid following a state expansion.

To identify the effect of Medicaid expansions to working-age adults on seniors eligible for dual enrollment, we use Current Population Survey (CPS) data to examine Medicaid participation and general health among persons eligible for dual enrollment. Identification of causal effects comes through exogenous variation in coverage expansions to working-age adults across states and over time. We use data from the March supplements to the CPS from 2010 to 2015, a period of time beginning prior to ACA implementation and ending after the ACA was fully implemented. We examine the effect of early Medicaid expansions, implemented in 2010 and 2011, as well as expansions that occurred after January 1, 2014, when the ACA was fully implemented. As of November 2015, 30 states and Washington, DC had adopted the Medicaid expansion. We construct a sample of respondents aged 65 and older and eligible for Medicaid, where eligibility is based on having income below the federal poverty level, following Pezzin and Kasper (2002) and Ungaro and Federman (2009). We estimate difference-in-difference models of Medicaid program participation where the key explanatory variable will be an indicator variable that takes a value of 1 in the states that expanded Medicaid during the years that the Medicaid expansion was in effect, and 0 otherwise. We control for a rich set of individual and household traits including age, sex, race/ethnicity, marital status, education, income, urban residence, plus time-varying state-level characteristics such as Medicaid managed care penetration and unemployment rate, and state and year fixed effects. Our results have important implications for the complex and high-cost population of those eligible for dual enrollment, a large fraction of whom do not receive benefits that may improve health.