Understanding the Effect of Race on Medicare Advantage Enrollment

To understand why Medicare Advantage (MA) has a relatively larger market share among racial minorities than traditional Medicare (TM).

We estimate Probit models for the choice of the MA sector versus TM by Black and Hispanic beneficiaries, as compared with White beneficiaries. We use a non-linear version of the Oaxaca-Blinder decomposition to decompose differences in the probability of MA enrollment by race into differences in explanatory variable values versus differences in the coefficients on those variables, which we identify as "preferences" for MA.

We combined 2020 Medicare Current Beneficiary Survey (MCBS) data with CMS data on MA plan payment levels aggregated to the county level, star ratings, and measures of market competition.

In the Black/White beneficiary comparison, 83% of the 17% point difference in the probability of MA enrollment was explained by differences in preferences (p < 0.001) while only 17% was explained by differences in attributes (p < 0.05). In contrast, in the Hispanic/White beneficiary comparison, 72% of the difference was explained by differences in attributes (p < 0.001) and 28% was explained by differences in preferences (p < 0.01). Attributes associated with differing rates of MA enrollment by race included both market-level characteristics (e.g., payment levels) and personal characteristics (age, level of pain, and chronic disease count). Preferences associated with differing rates of MA enrollment included coefficients of sector characteristics such as payment rates and the number of four-star+ plans available and age.

In this study, we find that the higher MA enrollment rate for Black versus White beneficiaries is largely associated with differences in preferences, while the higher enrollment rate for Hispanic beneficiaries is more associated with differences in attributes. Differences in preferences for MA sector characteristics were significant in explaining higher MA enrollment rates for both groups compared with White beneficiaries, suggesting that changes in payment rates will disproportionately impact racial minorities, particularly for Black beneficiaries. However, the reasons for different preferences for MA among racial groups remain somewhat of a puzzle, particularly given that we control for demographics, health, and market characteristics.