Study: Insurance and Access to Care Down, With No Boost in Work Among Tennessee Adults Losing Medicaid

September 18, 2018: We’ve updated this post to note that the National Bureau of Economic Research published Thomas DeLeire’s study as a working paper.

Tennessee’s decision of 2005 to end Medicaid for nearly 200,000 low-income adults raised the number without health insurance, reduced the likelihood that they would visit a doctor or dentist, and did not increase employment, Georgetown University’s Thomas DeLeire finds in a recent study published as a working paper by the National Bureau of Economic Research. That suggests that undoing the Affordable Care Act’s (ACA) Medicaid expansion to low-income adults, which 32 states (though not Tennessee) have adopted, “likely would reduce health insurance coverage, reduce health care access, and worsen health but would not lead to increases in employment,” DeLeire concludes.

This study adds to the preponderance of evidence showing that Medicaid enrollment among low-income adults doesn’t hurt beneficiaries’ work rates, including results from the gold-standard Oregon Health Experiment, a Kaiser Family Foundation review of studies related to the ACA’s Medicaid expansion, and another study on Tennessee’s termination of coverage for most low-income adults.

Some supporters of efforts to take Medicaid from people who don’t work or participate in work-related activities for a certain number of hours each month claim that Medicaid discourages work among beneficiaries and that work requirements are needed to offset the disincentive. These studies refute those claims.

DeLeire’s study, in particular, relates to Tennessee’s experience with Medicaid in the 1990s and early 2000s. In 1994, the state expanded its Medicaid program to cover a wider array of low-income adults. In 2005, however, the state disenrolled nearly 200,000 of them. In the aftermath of that action, the study finds:

• Tennessee adults’ work rate — higher than 70 percent — didn’t change when the state took Medicaid coverage from most low-income adults.
• The share of adults identifying a work-preventing disability rose, and more workers shifted from full-time to part-time work. That’s consistent with the expected impact of losing health coverage, particularly among those with work-limiting conditions that require regular care.
• The share of Tennessee adults enrolled in Medicaid fell by an estimated 5.4 to 5.9 percentage points.
• Adults’ uninsured rate rose by an estimated 4.7 to 5.2 percentage points.
• Adults’ private coverage rates didn’t change meaningfully, refuting claims that Medicaid coverage “crowds out” private coverage in which low-income adults otherwise would enroll.
• Adults’ self-reported health status worsened.
• Adults visited the doctor and dentist less frequently.

DeLeire’s study relied on the Census Bureau’s Survey of Income and Program Participation (SIPP), a survey that enables researchers to follow people over time. That’s the ideal data set to determine the impact of a policy change like Tennessee’s, because the analyst can see its effects on individuals over time.

This study is the third major study of the impact of Tennessee’s elimination of its 1994 Medicaid expansion on work. A 2014 Quarterly Journal of Economics study used the Census Bureau’s Current Population Survey and found that the loss of Medicaid coverage was associated with a significant rise in state employment; some proponents of rigid work requirements have relied heavily on this study to support their claims.

But more recent research examining the same policy change — DeLeire’s study and a study from 2017 — used data that was better suited to the task and found no evidence that it affected work rates. The 2017 study duplicated the Quarterly Journal of Economics results, but then conducted the same analysis using the Census Bureau’s American Community Survey (ACS), which the Census Bureau recommends for state-level analysis due to its far larger sample size. The ACS-based analysis showed that, in Tennessee, the loss of Medicaid had a very slightly negative but statistically insignificant effect on work rates.