Objective: The Monroe-Livingston demonstration project's capitation payment system (CPS) was evaluated to determine whether capitated funding of mental health care, compared with fee-for-service funding, could reduce hospitalization rates and improve functioning and symptoms for severely and persistently mentally ill adults without increasing the total cost of care. Methods: The experiment was a communitywide prerandomized clinical trial involving 422 patients. Patients who were randomized into the experimental group were eligible for enrollment in a capitated funding program administered by one of five community mental health centers. Those randomized into the control group received standard fee-based services. Follow-up interviews with patients one and two years after enrollment in the study assessed changes in symptoms and functioning. Data files of the membership corporation that coordinated community mental health services for the CPS provided measures of study patients' use of inpatient mental health services. Results: During the two-year follow-up period, patients in the experimental group had significantly fewer hospital inpatient days than patients in the control group, but the two groups bad no significant differences in functioning or level of symptoms. Conclusions: The CPS successfully maintained severely ill patients in the community but did not improve their functioning or level of symptoms.