Plan Selection in Medicare Part D

Featured in print Digest

Fewer than 10 percent of individuals enroll in what for them would be the most cost-effective plans.

In Plan Selection in Medicare Part D: Evidence from Administrative Data (NBER Working Paper No. 18166), co-authors Florian Heiss, Adam Leive, Daniel McFadden, and Joachim Winter analyze data on medical claims in Medicare Part D drug insurance programs. They find that fewer than 10 percent of individuals enroll in what for them would be the most cost-effective plans. This is apparently because seniors pay more attention to their out-of-pocket premiums than to the overall benefits of the dozens of drug plans available to them. Equally significant, the researchers believe that how seniors decide whether to enroll in Medicare Part D, and what plans they select, is important not only for management of the Part D program, but also is indicative of how consumers behave in real-world decision situations with a complex, ambiguous structure and high stakes. The researchers add that their findings may yield predictions for how seniors will handle plan choices in the new general health insurance exchanges that will implement the Patient Protection and Affordable Care Act of 2010.

This is one of the first papers to provide a comprehensive analysis of plan choice in the Part D market using a large random sample from the entire Medicare-eligible population. The data is derived from Plan D claims records for 2006-8, combined with Parts A and B claims records for 2002-8. The information includes plan choice, drug use, health conditions, out-of-pocket costs, and premiums. The authors then simulate the relevant attributes of alternative plans available to each consumer, using the administrative data on drug spending to characterize Part D enrollment decisions. This simulation predicts the beneficiaries' out-of-pocket spending among each available stand-alone Part D plan in their regions.

The analysis of enrollment and choice among different levels of plan generosity suggests that the share of eligible consumers without drug insurance is in the range one would expect if risk reduction and the option value of avoiding late enrollment penalties in the future are ignored, and the only criterion is whether enrollment is first-year actuarially favorable. In choosing between Silver (standard) and Gold (generic gap coverage) plans, it appears that consumers undersubscribe to the Gold plans. This result is consistent with earlier findings that consumers pay more attention to premiums than to benefit generosity, with the result that they tend to favor low-premium standard or equivalent plans.

Relative to the benchmark of a static decision rule, similar to the Plan Finder provided by the Medicare and Medicaid administrations, which makes next year's plan choice conditional on the drugs consumed in the current year, enrollees lost an average of about $300 per year. While these losses are modest compared to the losses associated with not enrolling at all, they are difficult to reconcile with decision costs alone. It appears that a sizeable fraction of consumers either value plan features that are not reflected in total cost, or else do not optimize effectively.

--Matt Nesvisky