Incentives in health care payment systems

This dissertation studies how incentives operate in health care payment systems. Relevant theory is developed and public policy implications are drawn. The first chapter models the important health policy dilemmas of risk selection and moral hazard. When providers can increase revenues by selecting favorable risks, capitation or purely prospective payment is unlikely to be optimal. A "second best" payment system may involve mixed levels of both demand- and supply-side cost sharing: consumers may prefer to pay deductibles and co-payments rather than have health care providers receive large financial rewards for skimping on care or discriminating against expensive-to-treat patients. Provider agency and risk adjustment can improve the terms of the social trade-off between inefficient utilization and inequitable coverage.; The second chapter links the optimal payment literature with multi-task principal-agent theory by focusing on design of health insurance-payment systems when providers engage in multiple "tasks". The model formalizes the insight of Newhouse (1996) that when supply-side cost sharing rewards providers not only for legitimate cost control but also for risk selection, there is a trade-off between efficiency in production and selection that may be best addressed by mixed payment. The incentive to shun unprofitable patients and to attract profitable patients is exacerbated by competition without risk adjustment. Given the current limitations of risk adjustment, caution is warranted regarding "high-powered" incentives in the health sector, such as combining competition and capitation payment.; The third chapter, joint with Winnie Yip, focuses on payment system reforms in urban China. We empirically examine the impact of introducing prospective hospital payment in the southern province of Hainan, using a pre-post study design with a control group to analyze claims data from 31 hospitals between 1995 and 1997. We find that prospective payment as opposed to fee-for-service is associated with both a slower rate of growth of overall expenditures and of program spending per inpatient admission, and a slower increase in spending on expensive drugs. Although the broader social welfare impact of reform remains unclear, these findings suggest cautious optimism regarding the effectiveness of prospective payment for controlling costs.