Advocates of Medicaid Expansion Claim It Saves Lives. The Evidence Isn’t So Clear.

New research fails to detect a relationship between Medicaid expansion and mortality.

Proponents of Medicaid expansion have cast the issue as nothing less than a life-or-death struggle.

A 2019 report by the Center on Budget and Policy Priorities declared that Medicaid expansion had saved at least 19,000 lives, adding that states that declined to expand were responsible for 15,000 premature deaths. A similar article by the Center for American Progress in 2021 claimed that hold-out states could prevent 7,000 deaths each year by expanding Medicaid.

It turns out, maybe not.

A working paper released this month by the National Bureau of Economic Research (NBER) suggests that the “evidence that pre-ACA Medicaid expansions to adults saved lives is not as clear as previously suggested.” The findings challenge a flurry of studies – including Sommers et al. (2012), Sommers (2017), and Miller et al. (2021) – showing that expanding Medicaid reduces mortality, arguably the most important outcome in evaluating the policy.

This isn’t the first study to report a weak (or even positive) association between Medicaid expansion and mortality rates – see Finkelstein et al. (2012) and Blase and Balat (2020) for other examples – but it constitutes a valuable contribution to the literature. By examining more states, deploying more credible empirical tools, and analyzing more years of data than many previous studies, authors Charles Courtemanche, Jordan Jones, Antonios Koumpias and Daniela Zapata are able to perform an exceptionally thorough analysis.

The paper evaluates the effect of eight state Medicaid expansions in the 1990s and 2000s — all before the ACA was enacted — on state-level mortality rates, including all-cause, healthcare amenable, non-healthcare-amenable and HIV-related mortality.

Courtemanche et al. find no evidence that Medicaid expansion reduces any of the four mortality measures in any of the eight states or across all treated states. Not only are the estimates never statistically significant, but the pattern of signs is mixed and magnitudes are generally small.

The authors conclude: “... one should not assume that state Medicaid expansions automatically lead to reductions in adult mortality.” It’s certainly plausible, of course, that extending public health insurance coverage would improve health and ultimately cause a decline in mortality. But Courtemanche et al. speculate that the effects of state Medicaid expansions are likely a function of many factors, including the availability of services and providers for Medicaid enrollees, the demographic characteristics of the population, and other changes to the healthcare system that are implemented concurrently.

Moreover, expanding Medicaid can trigger unintended consequences. When health system capacity is strained, as it is in large swaths of the U.S., Medicaid expansion may leave fewer providers to care for those already on the program. It is worth pondering whether admitting new enrollees to Medicaid may reduce their mortality risk at the cost of heightening risk for existing enrollees, who now find themselves struggling to access care. Recent work I did with my Mercatus colleague Charles Blahous hints at this possibility, finding that Medicaid expansion shifted resources away from already-enrolled children. This is especially concerning as healthcare investments in low-income children have been found to have some of the lowest cost/benefit ratios among government policies.

In short, the null mortality result reported by Courtemanche et al. may mask sizable offsetting effects on different sub-groups – a possibility that deserves greater attention. At the very least, these new results greatly undermine the notion of Medicaid as an unmitigated boon to population health.