Incentives for risk selection and omitted variables in the risk adjustment formula

Risk adjustment in health insurance raises the question of how to treat variables which influence health care expenditures but do not capture acceptable costs differences. We argue that these variables should be included in the explanatory model and neutralized afterwards for the computation of the premium subsidies. This explicit approach is better than the conventional approach in removing the incentives for cream-skimming. We illustrate the empirical relevancy of the problem with data for Belgium.