Suboptimal Employee Health Plan Choices: Longitudinal Evidence from a Natural Experiment

Wednesday, June 13, 2018: 8:40 AM
Salon IV - Garden Level (Emory Conference Center Hotel)

Presenter: Mary Hamman

Co-Author: Taggert Brooks;

Discussant: Brian E. McGarry

As of 2017, 58 percent of people with employer sponsored health insurance coverage had more than one plan option, 57 percent had a high deductible health plan (HDHP) option, and 94 percent of firms that offered HDHPs offered other plan types too (Kaiser Family Foundation 2017 Employer Health Benefits Survey). HDHP premiums can be substantially lower than other plan offerings and employer contributions to Health Savings Accounts (HSAs) can partially offset or even eliminate cost sharing making the total premium and out-of-pocket costs cheaper than other offerings regardless of healthcare consumed. In this paper, we examine plan elections among public employees across four years of benefit elections before and an increase in employer HSA contributions made the HDHP plan offerings financially optimal for all employees. Using two waves of survey data linked to administrative plan choice data, we identify predictors of optimal choice in Wave 1 and test the explanatory power of these predictors in Wave 2 among new participants. Additionally, we exploit two natural experiments: the exit of one insurance carrier which required employees covered under those plans to actively elect benefits, and the introduction of a new interactive plan decision aid. We find only 10 percent of employees that had to actively enroll switched to the dominant plan but this was double the rate of those who did not need to actively enroll. Employees that chose HDHP plans were more knowledgeable about plan attributes than those who selected other options. Use of the decision tool did not predict enrollment in the dominant plan and only 27 percent of employees used the tool. This analysis makes a unique contribution to the existing literature on employee plan choice because identical provider networks, gatekeeping, and coverage were available under financially dominant and dominated plans allowing us to isolate the impact of changes in financial dominance and objectively determine optimal plans despite heterogeneous preferences. These findings suggest default enrollment in financially optimal plans or elimination of dominated choices may be the best policy options for improving choice behavior.