Practical illustrations of the two recent contributions to stochastic frontier models

Abstract

In a recent year, the conditional expectation for technical inefficiency (i.e., the predictor) in stochastic frontier production models and its confidence interval are proved to be increasing in the individual firm's inefficiency effect, which is defined as the mean of the normal distribution that is truncated at zero. This paper illustrates how the two recent contributions work in practice and how these findings are significant, by using the Battese and Coelli (1995) type specification and giving an empirical study on the Japanese pharmaceutical industry.

Acknowledgements

We thank Sango Kim, Hajime Hori, Osamu Kamoike, Yoshiro Tsutsui for their helpful comments. This research by the second author is supported by a grant-in-aid from the Zengin Foundation for Studies on Economics and Finance in 1999 and by a grant-in-aid 13630104 from the Ministry of Education, Science and Culture of Japan.