Malmquist productivity index estimation with zero-value variables: The case of Greek prefectural training councils

Abstract

The paper suggests a technique for the parametric estimation of the Malmquist productivity growth index when the dataset to be analyzed contains a considerable number of observations with zero values. A dummy variable technique suggested by Battese (J. Agric. Econom. 48 (1997) 250) is extended to a translog specification of the input distance function. Moreover, technical changes (TCs) are decomposed into neutral and biased components and the sources of total productivity growth are computed via formulae explicitly accounting for the discrete nature of the data. Our approach has been applied to the Greek prefectural training councils—a state extension system for the general public. Findings indicate a considerably negative productivity growth primarily attributable to regressive TC.