Using Behavioral Economics to Improve Plan Choice in Medicare: A Randomized Experiment

A variety of research finds that consumers make systematic mistakes when selecting health insurance plans, and exhibit several behavioral economic biases that can lead them to choose or remain enrolled in plans with higher expected costs than they might otherwise prefer. In two new randomized experiments, we consider whether certain behavioral tendencies—specifically, salience and choice overload—can be used to redesign plan information materials in a manner that improves plan choice in Medicare, specifically by shifting consumer preferences away from defaulting and inertia, and towards lower expected cost plans.

A sample of 487 soon-to-be Medicare enrollees and 210 current beneficiaries were recruited from Middlesex County, Massachusetts, and randomized to receive county-specific plan information for the 2015 enrollment period, either presented as in the state’s Medicare And Youpublication (Control) or one of several treatment formats informed by behavioral insights on salience, choice overload, or combinations of the two. Respondents were asked to identify the Medicare Advantage plan they believed would be best for them, and whether they would select that plan over either Original Medicare (soon-to-be enrollees) or their current plan (current beneficiaries). To increase the likelihood that hypothetical answers would reflect actual behavior, surveys were conducted over the three weeks before Medicare’s open enrollment began. 

The new enrollee analysis indicated that reformatting Medicare Advantage plan information tables to make each plan’s expected costs highly salient yielded a statistically significant 10 percentage point decrease in preferences for defaulting, while a choice overload intervention drawing attention to a smaller choice set reduced this preference by a statistically insignificant 6 percentage points. Considering the expected costs of each respondent’s preferred plan, analyses found statistically significant savings of $12 and $14 per month from the salience and choice overload treatments, respectively. These expected cost effects were not driven by a shift in responsiveness to plan premiums.

Focusing on current beneficiaries, the second intervention found that, relative to the control group, a combined salience-choice overload treatment yielded a 13 percentage point increase in preferences for switching plans. The corresponding change in expected costs was a statistically insignificant reduction of about $15 per month, again, not due to a shift in responsiveness to plan premiums.

Overall, these results show that behavioral economics can be leveraged to shift consumer preferences in Medicare towards lower expected cost plans and away from defaulting and inertia, simply by redesigning plan information materials in accordance with common behavioral tendencies, and without limiting respondents’ choice sets.