In a large multi-specialty group practice treating approximately equal numbers of health maintenance organization (HMO) and fee-for-service (FFS) patients, we analyzed a natural experiment by the administration to introduce a dual incentive system for physicians. We examine the impact on care when they announced that each physician would be remunerated for HMO care based on a per capita budget, but for FFS care based on billable services. Data were 86,230 episodes for treating patients under age 65 with seven common illnesses. There was no evidence that the intended impact (reducing HMO care) occurred; instead, there were undesired and unintended effects (reduced care for FFS and upset physicians and threats to their corporate culture).