Minneapolis, MN, USA: University of Minnesota (doctoral dissertation).
University of Minnesota, Minneapolis, MN, USA
Employer-provided health insurance is a major component of the health care financing system in the United States. “Managed competition” pricing (Enthoven and Kronick, 1989) became a buzzword in employment-based health insurance in the 1990s, but this fixed-dollar strategy for employer contributions was implemented primarily without the recommended health risk adjustment to the employer subsidy. This work compares managed competition pricing, with and without health risk adjustment, to traditional percent-of-premium pricing. Using a model that predicts employees’ individual health claims and plan selection patterns, this work demonstrates that managed competition pricing provides the greatest average employee welfare among the three contribution methods, while holding employer costs fixed.
The development of the employee cost and choice model is based on three years of enrollment and claims data for a large Twin Cities employer. The detailed nature of the claims data allows the use of sophisticated risk measures, unique in the health plan choice literature, using Bayesian inference to estimate a multiperiod, multinomial probit model. This model allows correlation in the unobservable preferences across time and across plan choices, for the first time integrating these correlations in a logically consistent way in health plan choice modeling.
The resulting model suggests that the commonly used multinomial logit model may not be appropriate in health plan choice modeling, even when modified by the nesting of choices. This may explain some of the literature’s inconsistent results regarding the influence of health risk on plan selection. In addition, the model identifies high correlations in the unobservable preferences over time, leading to great inertia in health plan choice. The estimated model also suggests risk aversion on the part of employees in their health plan selection.
Please log in/register to access.