Physician performance pay: Evidence from a laboratory experiment

We present causal evidence from a controlled experiment on the effect of pay for performance on physicians' behavior and patients' health benefits. At a within-subject level, we introduce performance pay to complement either fee-for-service or capitation. Performance pay is granted if a health care quality threshold is reached, and varies with the patients' severity of illness. We find that performance pay significantly reduces overprovision of medical services due to fee-for-service incentives, and underprovision due to capitation; on average, it increases the patients' health benefit. The magnitude of these effects depends, however, on the patients' characteristics. We also find evidence for a crowding-out of patient-regarding behavior due to performance pay. Health policy implications are discussed.