Dynamic Risk Adjustment in Markets with Persistent Risk and Manipulable Signals: Market Design for Health Insurance
Abstract
Risk adjustment is a payment mechanism, not only a prediction problem. I extend optimal risk adjustment to dynamic insurance markets in which plans capture future residuals from persistent risk. Under an efficiency criterion, payments should reflect expected profits and losses over the enrollee relationship on margins plans control, not only one-year spending predictions. A finite-cell model separates selection, health production, and manipulable measurement. The framework implies that annual recalculation can tax prevention, high-R² prediction can reduce welfare, and lagged claims anchors are useful only with gaming safeguards.