The appeal of pay-for-performance in health care derives from the conceptual view that paying doctors and hospitals more to deliver better care will encourage them to deliver better care. What lessons can be learned from the successes and failures of pay-for-performance in health care settings that apply to pay-for-performance in population health? We argue that pay-for-performance requires conditions that are not easily met in population health settings. Pay-for-performance has focused on narrow clinical problems whose success depends on identifiable actors with the motivation and resources to change clinical processes or outcomes. In contrast, population health has broad goals, many antecedents, and no single, identifiable fiduciary (a person who holds assets in trust for a beneficiary). Nevertheless, with careful attention, conditions for successful pay-for-performance in population health might be met.