When I was a member of the Social Security Advisory Board (SSAB) from 2006 through 2012, we were quite concerned with the rapid rise of federal disability benefit applications, awards, beneficiaries and amount of payments. The Disability Trust Fund was careening toward insolvency, large backlogs were developing in the Social Security Administration (SSA) hearings offices, and there was evidence that even the national labor market was affected significantly and noticeably by disability applicants and recipients withdrawing from work. Around this time, the hotly-debated Affordable Care Act (ACA, also called ObamaCare) was enacted and with it came large expansions of federally-funded health insurance through Medicaid. In particular, instead of narrowly covering specific groups, Medicaid would be provided to all individuals with family incomes up to 138 percent of the federal poverty level.
Federal disability and these ACA expansions were often tied together. The SSAB was told by field agents at SSA and by academic experts that one of the main motivations for disability application was access to health insurance, either through Medicaid, via Supplemental Security Income (SSI) for those with no or spotty work records, or through Medicare, via Social Security Disability Insurance (SSDI) for those with steady employment. If low-wage workers could just get access to free and sometimes better health insurance through Medicaid, they argued, then these workers would not apply for SSI or SSDI and would not withdraw from the labor force.
It turns out that this was not true.
According to a recent study, including a review of the professional literature through 2019 on this question, Williams College economists Lucie Schmidt, Lara Shore-Sheppard and Tara Watson (SSW) found that prior studies produced economically small and mixed effects of Medicaid eligibility on disability applications. But they also correctly noted that there were data and methodological problems with these studies, including possibly spurious correlations and trends.
SSW set out to correct these issues and produce a definitive answer to the question. In particular, they took advantage of the natural experiment provided by the Supreme Court decision in 2012 on the constitutionality of the ACA that made Medicaid expansion optional for states, resulting in some states not expanding Medicaid and different timing of the remaining expansions, as well as some states having more generous income eligibility limits even before the ACA. SSW compared changes from 2010 through 2016 in county-level SSI and SSDI applications and awards in about 500 contiguous county pairs across a state border with different Medicaid eligibility, thereby avoiding the strong geographic correlation at the state level in the identity of which states expanded and disability trends. Counties bordering each other are more likely to share similar labor markets and be affected similarly by macroeconomic shocks. SSW use a continuous measure of the state Medicaid income eligibility limits, including no and when expansion, compared to state-specific pre-ACA limits, as well as time-varying controls such as demographics, and county and time fixed effects, in their difference-in-difference regression equations.
SSW definitively reject that Medicaid expansions reduced either SSI or SSDI applications or awards, in their words, “a fairly precisely estimated zero” effect for the relationship between Medicaid and the receipt of disability benefits. Various robustness tests, changing specifications, and variable inclusions, do not alter these results.
It is hard to know what the intellectual and political impact would have been if we had these results in the last decade when the ACA was debated and passed, and when the SSDI program was bailed out instead of reformed. But having that knowledge now does counsel humility and prudence on big claims of cross-cutting program savings from large increases in government spending.
Reprinted with permission from AEI by Mark J. Warshawsky.