Premium Increases and the Concentration of the Medicare Advantage Market: A Finite Mixture Modeling Approach

Prior studies have shown that enrollment in the Medicare Advantage (MA) market is highly concentrated among a few firms, and continues to become more concentrated, but hitherto no studies have examined the impact of a concentrated market on premiums faced by Medicare beneficiaries. Also, given the complexity of the MA market, research has not clearly identified the key market characteristics associated with faster premium increases. This paper studies the relationship between the concentration of enrollment in MA markets and the increases in premiums faced by MA enrollees with prescription drug coverage (MA-PDs), to examine whether, and to what extent, MA companies are using their market leverage to increase premiums for beneficiaries.  We hypothesize that plan premiums within in a county increase faster in more concentrated markets than in less concentrated markets. Our secondary hypothesis is that the increase in premiums varies by plan, company, and county characteristics, and in particular, that rural markets are likely to see faster increases in premiums than urban markets. 

We examine how changes in premiums for MA-PDs between 2008 and 2009, at the county, company and plan level, vary by the concentration of MA enrollment in companies. The time period was selected because it was a relatively stable time period with few policy changes for the MA program.  CMS MA County Enrollment Files from 2008 to 2009 were linked with information from the MA Plan Directory, Landscape Files, Benchmark Files, and the HRSA Area Resource File to compile the following information for each county:  market concentration (using the Hirschfield-Herfindahl Index), plan premiums, plan type, plan enrollment, MA penetration, number of companies offering plans, number of plans offered by companies, counties’ urban designation, tax status of companies offering plans, per capita traditional Medicare costs in the county, and standard error in county income.  We exclude plans not available for general enrollment. 

We use finite mixture modeling to identify the number of distinct markets and the county, company, and plan characteristics associated with each MA market.  Sensitivity analyses examined the extent to which outliers and the counties with monopolies influenced the results.   

We found a strongly significant and positive association between concentrated markets and change in premiums in all of our models. The observed effect was much larger in some markets than others. MA markets are not distinctly one type of a market; a combination of characteristics primes a market for larger changes in MA premiums. In particular, monopoly or oligopoly markets, many of which are in rural and lower income counties, seem most at risk for larger premium increases and faster growth rates in premiums compared to less concentrated markets.    

Our analysis indicates that Medicare beneficiaries may be adversely affected by the increasing concentration of enrollment in the MA market, and the special circumstances affecting the competitiveness of both the provider and MA markets should be considered, particularly in more rural areas.