Viewing the uninsured rate through foggy lenses

Subscribe to xpostfactoid

Snapshots of the health insurance status of the U.S. population are blurry.

When you look at the CMS tally of Medicaid enrollment increase since the pandemic struck, it seems, simply, that the increase swamps most estimates of the number of people that lost employer-sponsored insurance. Other factors are at work, of course. But the one large number is considerably larger than the other large number.

But official Medicaid enrollment totals may not be an entirely reliable measure of how many people are actually covered by Medicaid, and know themselves as such -- particularly during this pandemic, when disenrollments have been paused since March 2020. State Medicaid agencies are, to varying extents, blind beasts.

I know a young man who, during a year of transition, lived in two states and worked at three jobs, with a period of unemployment. At different points in the year he applied for Medicaid in two (blue) states and received rejection notices. From both of those states, months later (and months apart), while insured through a new employer, he was sent managed Medicaid membership cards and informed that he'd be enrolled since shortly after his application was completed.  In both states, it took some doing and some time to get himself disenrolled.

That's a one-person, two-state sample. But there is a ton of churn in Medicaid, and the bureaucracy churns slowly.  In a typical month, there are 2-3 million applications, including, in about a dozen states, re-enrollments and redeterminations. Net enrollment totals normally change by a few tens of thousands (in the pandemic, by hundreds of thousands). 

Assessments of the insured population are mostly dependent on surveys, which in some ways may be more reliable than hard enrollment data (see above; are you really insured by a program you don't know you're enrolled in? -- and how promptly does a state agency determine your status?)  -- and in some ways less reliable (a lot of people in managed Medicaid programs branded by insurer don't know they're in "Medicaid").

So far, analyses of the changes in the uninsured rate during the pandemic find either downticks that don't reach statistical significance (1, 2) or slight upticks. Statistical rigor is good, and we do have some idea of what happened (public programs newly covered roughly as many people as lost employer-sponsored insurance). In a creaky, as-good-as-it-gets-in-America way, the safety net, bolstered by the ACA and having survived Republican assaults, worked. The single strongest bulwark may have been the pause in Medicaid disenrollments effected by the Families First Act in March 2020.  The enhancements in ACA marketplace subsidies enacted in the American Rescue Plan Act in March 2021 have added a significant secondary boost this year to date.  

But the overall picture of how the health insurance landscape changed in 2020 and 2021 is still fuzzy and  will remain so for some time (e.g., until Census bureau data for 2021 comes out this time next year), and, to a degree, forever.

Subscribe to xpostfactoid