Most health insurers in the Netherlands apply community-rating and open enrolment for supplementary health insurance, although it is offered at a free market. Theoretically, this should result in adverse selection. There are four indications that adverse selection indeed has started to occur on the Dutch supplementary insurance market. The goal of this paper is to analyze whether premium differentiation would be able to counteract adverse selection. We do this by simulating the uptake and premium development of supplementary insurance over 25 years using data on healthcare expenses and background characteristics from 110,261 insured. For the simulation of adverse selection, it is assumed that only insured for whom supplementary insurance is expected not to be beneficial will consider opting out of the insurance. Therefore, we calculate for each insured the financial profitability (by making assumptions about the consumer’s expected claims and the premium set by the insurer), the individual’s risk attitude and the probability to opt out or opt in. The simulation results show that adverse selection might result in a substantial decline in insurance uptake. Additionally, the simulations show that if insurers were to differentiate their premium to 28 age and gender groups, adverse selection could be modestly counteracted. Finally, this paper shows that if insurers would apply highly refined risk-rating, adverse selection for this type of supplementary insurance could be counteracted completely.

Additional Metadata
Keywords Adverse selection, Adverse selection spiral, Death spiral, Health insurance, Premium, Premium differentiation, Supplementary insurance
Persistent URL dx.doi.org/10.1007/s10198-017-0918-2, hdl.handle.net/1765/101061
Journal The European Journal of Health Economics
Citation
Winssen, K.P.M, van, van Kleef, R.C, & van de Ven, W.P.M.M. (2017). Can premium differentiation counteract adverse selection in the Dutch supplementary health insurance? A simulation study. The European Journal of Health Economics, 1–12. doi:10.1007/s10198-017-0918-2