OBJECTIVE: To determine whether hospital utilization and expenditures have declined more rapidly in metropolitan statistical areas (MSAs) with high health maintenance organization (HMO) penetration compared with MSAs with low HMO penetration. STUDY DESIGN: Levels and rates of change in hospital expenditures and hospital utilization in MSAs with varying levels of HMO penetration (1982 to 1996) were compared in a natural experiment. METHODS: MSAs were grouped into 4 categories based on HMO penetration rates in 1996. Levels and rates of change in hospital admission rates, hospital inpatient days, emergency room visits, total expenditures per capita, and expenditures per adjusted inpatient day from 1982 to 1996 were compared. A first-difference multivariate model was evaluated for 1993 to 1996. RESULTS: At the MSA level, the rates of change in hospital utilization and hospital expenditures varied only modestly with the level of HMO penetration. Changes in hospital admission rates did not vary systematically with HMO penetration rates except in the 1993 to 1996 period, when MSAs with the highest HMO penetration had the largest decline. Reductions in hospital days per capita and expenditures per day did not vary systematically by level of HMO penetration. Emergency room days declined most rapidly in the MSAs with the highest HMO penetration in the 1982 to 1993 period and were similar in the 1993 to 1996 period. Hospital expenditures per capita showed the greatest association with managed care penetration. They averaged 1.6% slower annual growth in MSAs with high versus low HMO penetration in the 1982 to 1996 period. CONCLUSIONS: This national study using data from 1982 to 1996 suggests that the effects of HMO penetration on hospital expenditures and hospital utilization at the MSA level are small (generally less than 1% per year).