Title

THE ECONOMICS OF THE CONNECTICUT HOSPITAL INDUSTRY: COMPETITION, STRUCTURE, AND PERFORMANCE

Date of Completion

January 1982

Keywords

Economics, General

Degree

Ph.D.

Abstract

This study uses the structure - conduct - performance paradigm of Industrial Organization as a framework for analyzing the Connecticut hospital industry. Efforts are made to quantify the relationship between structural and performance factors while accounting for several unusual features of the industry.^ An initial phase of the research involves an evaluation of a variety of market definitions needed for the calculation of standard measures of market concentration. This assessment of actual market patterns reveals that Health Systems Agency areas best meet valid market boundary criteria.^ The principal part of the study involves the estimation of demand and cost functions for the industry. The multiple regression results suggest that hospital utilization rates are very much affected by rivalrous behavioral interchanges among hospitals. As conditions for rivalrous behavior decrease, utilization rates increase. Other determinants of hospital demand that are tested and found to be important include education, alternate kinds of care, insurance, and the value of time. Only weak support for supplier - induced demand hypotheses is generated.^ In performing the cost analysis, a multiple output - multiple input cost function is specified. This approach represents a significant methodological advancement over much previous research. Parametric tests provide statistical evidence for rejection of the use of conventional functional forms which are contrained to exhibit separability of outputs and constant returns to scale. Several production statistics are calculated from the estimated parameters including input elasticities of substitution, own input price elasticities, economies of scale, and marginal costs. A final effort is directed to the identification of determinants of deviations from average costs and reveals that competitive factors, case mix, insurance type, and physician influence are important. ^

COinS