Specification and estimation of stochastic multiple-output production and technical inefficiency

This paper presents a primal-based approach for specification and estimation of multiple-output production frontiers that allows simultaneous identification and estimation of determinants of technical inefficiency. The proposed model extends the technical efficiency effects model by Battese and Coelli (1995) to general multiple-output technologies. An empirical application using Swedish health care panel data from the years 1989-1994 is included. The key issue is to test for the effects on technical efficiency of an organizational reform implemented with the primary purpose to increase efficiency. The results reveal no significant effect on technical efficiency of the reform, but evidence of overall decreased technical efficiency and technical progress in the provision of health care over the studied time period are found.