Estimating hospital costs with a generalized Leontief function

This paper estimates a long-run hospital cost function with multiple outputs and inputs using a panel data set from Washington State hospitals during 1988-1993. We find that with our data the generalized Leontief function is more appropriate than a translog for estimating hospital cost functions. With respect to hospital costs, we find that hospitals readily adjust the use of intermediate products. Radiology, therapies and surgery, and other inpatient days, all serve as substitutes for core inpatient days. Outpatient services are found to be complementary to core inpatient services, indicating that the growth of stand-alone outpatient clinics might increase the costs of providing healthcare services. Our analysis finds that hospitals show significant economies of scale, but there is a limited amount of evidence of scope economies. Also, there is some evidence that profit-seeking hospitals achieve some of their goals by controlling costs, and that diagnostically related groups (DRG)-based Medicare services are effective in getting hospitals to control costs.