Adverse and Advantageous Selection in the Medicare Supplemental Market: A Bayesian Analysis of Prescription drug Expenditure
用贝叶斯方法分析老年人补充医疗保险对处方药支出的影响，发现计划选择存在内生性，市场同时存在逆向和正向选择，补充计划使人均年支出增加41%或21%。
This paper develops an extended specification of the two-part model, which controls for unobservable self-selection and heterogeneity of health insurance, and analyzes the impact of Medicare supplemental plans on the prescription drug expenditure of the elderly, using a linked data set based on the Medicare Current Beneficiary Survey data for 2003-2004. The econometric analysis is conducted using a Bayesian econometric framework. We estimate the treatment effects for different counterfactuals and find significant evidence of endogeneity in plan choice and the presence of both adverse and advantageous selections in the supplemental insurance market. The average incentive effect is estimated to be $757 (2004 value) or 41% increase per person per year for the elderly enrolled in supplemental plans with drug coverage against the Medicare fee-for-service counterfactual and is $350 or 21% against the supplemental plans without drug coverage counterfactual. The incentive effect varies by different sources of drug coverage: highest for employer-sponsored insurance plans, followed by Medigap and managed medicare plans.