Panel Paper: ACA Medicaid Enrollment Among Previously-Eligible Adults: How Powerful Was the Welcome Mat?

Friday, November 4, 2016 : 1:50 PM
Columbia 10 (Washington Hilton)

*Names in bold indicate Presenter

Kosali Simon1, Benjamin Sommers2,3 and Coady Wing1, (1)Indiana University, (2)Harvard University, (3)Brigham and Women's Hospital


The Medicaid program has expanded its coverage of low-income adults dramatically since 2010. In addition to insuring the newly eligible, this expansion has potential to increase Medicaid take-up rates among those who were already eligible for the program. Prior literature such as Aizer and Grogger (2003) finds evidence of such spillover effects, which are often called “welcome-mat” effects. Because the Affordable Care Act (ACA)’s Marketplace insurance expansions and other provisions occurred even in states with no Medicaid expansion, these welcome-mat effects may be felt in non-expansion states as well. Streamlined application processes, simplified income counting rules, and elimination of asset testing may have decreased the transactions costs of applying for Medicaid substantially. Increased awareness of health insurance programs, and the individual shared responsibility provision can also increase Medicaid take-up of those already eligible. However, expectations regarding the welcome-mat effect and its budgetary implications may be a factor that influences state decisions regarding Medicaid expansion, thus motivating the need for timely research.

We ask the following two questions: (1) Is there evidence that welcome-mat effects occur nationally? (2) How do the effects compare in expansion vs non-expansion states? We answer the first question by using an interrupted time series analysis of Medicaid take-up rates in 2014 among those who were eligible by pre 2013 eligibility standard; we compare to year-to-year changes prior to 2014, to account for the influence of measurement error in income and other reasons why there may be increases in take-up of Medicaid among the always-eligible, even absent Medicaid expansion. We answer the second question by conducting a difference-in-difference analysis, which compares the changes in expansion states before and after 2014 with the changes in non-expansion states. Here too, we compare year-to-year changes that occurred prior to 2014 in states without prior changes in Medicaid eligibility during our study period. 

We use 2008-2014 data from the American Community Survey (ACS).  It’s large sample size allows us to conduct an interrupted time series analysis, and to control for detailed individual and family characteristics in both of our analyses. We merge the ACS data with the information on state Medicaid eligibility rules compiled from various sources.

Our preliminary analysis of the 2013-2014 ACS data shows the mean rate of Medicaid coverage among non-disabled parents whose income is below pre-ACA Medicaid eligibility thresholds increased by 5.6 percentage points in expansion states and by 4.6 percentage points in non-expansion states. The mean rate of any insurance coverage among the same population increased by 7.0 percentage points and 6.4 percentage points in expansion and non-expansion states, respectively. These preliminary results indicate that welcome-mat effects occurred nationally and the effects were similar between the expansion and non-expansion states. 

Our study provides insight into the mechanism of how welcome-mat effects work. Our current results suggest that these effects occurred not so much because of the Medicaid expansion itself but rather because of other components of the ACA, including the decreased transaction costs associated with applying for Medicaid, greater outreach, and the individual shared responsibility requirements.