This article develops a new method of decomposing the cost difference between HMO and non-HMO plans into observed risk selection, unobserved risk selection, utilization differences, and differences in provider reimbursement rates. We implement this method using a large national sample of employer-sponsored health insurance enrollees from the Community Tracking Study Household Survey. We find no evidence that HMO plans attract a disproportionate share of low-risk enrollees; the US$188 difference between HMO and non-HMO medical expenditures per enrollee can be explained by the relatively low provider reimbursement rates paid by HMO plans. This indicates there may be little need for employers to risk adjust insurance premiums or otherwise restrict employee choice of plan types.
View full article...