Selection-Proof Prospective Payment in Health Insurance Exchanges

Background: Despite efforts at risk adjustment, community rating regulations make the Affordable Care Act (ACA) vulnerable to selection.  This is because community rating is a prospective mechanism subject to the selection-efficiency tradeoff. While it discourages inefficiently high provision of treatment, it results in both greater losses to MCOs who attract high-risk enrollees, as well as greater rewards to MCOs who can reject or avoid them. In this way, the effectiveness of the ACA may suffer from both adverse and preferred-risk selection across health plans within the health insurance exchanges.

Objectives: This article proposes an elaborate, fully-prospective payment mechanism (BARDR) and investigates whether or not it eliminates all selection incentives, or is ``selection-proof'’, and thus not subject to the selection-efficiency trade-off.

Methods: The article models a health insurance exchange within which consumers, MCOs, and a sponsor interact. MCOs and consumers each hold private information regarding an aspect of the underlying cost of an insurance arrangement, of which the sponsor is unaware. The BARDR mechanism relies on four parts. First, consumers choose a set of health plans with which enrollment would be ``acceptable'' at a price determined within the exchange. Simultaneously, MCOs submit bids to enroll individual patients, with the winner determined by second-price sealed-bid auction from within the chosen acceptable set. Third, there is mandate that each winning MCO purchase reinsurance against the characteristics of the acceptable set chosen by the enrolled patient. These three parts determine the prospective payment. Finally, after the payment has been made and the final treatment cost realized, each MCO must report the cost of insuring each consumer enrolled under a modified Duggan-Roberts mechanism.    

Results: Under the first-best allocation of consumers to MCOs, there is no adverse or preferred risk selection. Under traditional prospective payment mechanisms, however, both forms of selection occur. This holds regardless of whether prospective payments are set centrally by the sponsor or separately by each MCO. The BARDR mechanism exhibits no evidence of either type of selection. Furthermore, numerical simulations show that the BARDR mechanism achieves the first-best outcome for over 71% of patients, while traditional prospective payment mechanisms achieve it for less than 34% of patients.

Conclusion: In an environment where traditional prospective payment mechanisms produce both adverse and preferred risk selection, the BARDR mechanism eliminates both forms of selection while remaining fully prospective, and is thus selection-proof. This advancement in health care financing allows an uninformed sponsor to achieve efficiency in treatment without the drawbacks of adverse or preferred risk selection.