Why Did Medicare Advantage Enrollment Grow as Payment Fell?

Objective: To explore the growth of Medicare Advantage (MA) enrollment after the Affordable Care Act’s (ACA) payment cuts were implemented.

Design: We conducted descriptive analyses of MA enrollment between 2009 and 2017 by plan type nationally, by region, and by state. We then estimated county-level first-differences models describing the association between changes in MA penetration and the availability of high-quality plans, the availability of zero premium plans, Part D premiums, and MA benchmarks (for 2,833 counties). We also estimated an alternate first-differences model including changes in demographic factors like the share of the population over age 65, the share of the elderly aged 65-69, the share of the elderly below poverty, and the share of the elderly with a disability. These demographic data were only available in 790 large counties

Results: Between 2009 and 2017, MA enrollment grew from 10.4 million to 19.1 million and varied widely across the country and across plan types, with 21.7% of the growth in MA enrollment attributable to growth in employer plans and 11.9% was attributable to growth in SNPs, both of which are restricted-enrollment plans. Regionally, 23.9% of all MA enrollment growth occurred in the South Atlantic region and 15.3% occurred in the East North Central region.

Our first-differences models show that gaining access to a zero-premium plan or having one in both years was associated with an increase in MA penetration relative to never having a zero-premium plan (1.5 and 3.0 percentage points, respectively). In addition, gaining access to a five-star plan was associated with a 7.3 percentage point increase in MA penetration. Gaining or maintaining access to a four-star plan was also associated with an increase in MA penetration (2.3 and 1.9 percentage points, respectively). Finally, increases in Part D premiums were associated with lower MA penetration growth (-0.2 percentage points per dollar increase in monthly premiums).

Adding demographic characteristics to our model showed that only the change in the share of the elderly below poverty was significantly associated with MA penetration and the coefficient was negative, suggesting that MA did not grow as quickly in high-poverty counties, but did not change other results for MA market characteristics.

Discussion: The ACA was expected to reduce MA enrollment, but instead MA grew substantially. This growth was unevenly distributed by region and plan type, however, and senior choice was only part of the story, as over one-third of enrollment growth was attributable to restricted-access plans like employer-sponsored plans or SNPs. County-level MA penetration growth was related to access to zero premium plans and four or five star plans, suggesting that beneficiaries are choosing MA more frequently where MA plans are generous or high quality. MA penetration growth was not related to the age distribution of beneficiaries in a county, suggesting that growth is not driven by younger beneficiaries enrolling because more comfortable with managed care. We could not measure the effects of other factors, including benefit design simplicity and advertising, that could have contributed to MA enrollment growth.