A Hierarchical Stochastic Frontier Model for Efficiency Measurement Under Technology Heterogeneity

This article proposes an extension to the typical random-coefficients frontier model that allows the incorporation of firm management indicator(s) in the distribution of firms' technology parameters. Such a modelling approach does not only relax the homogeneous technology assumption but also empirically tests for the factors that may be responsible for variation in firms' technology parameters. The proposed approach is used to measure the technical efficiency of German dairy farms for the period 2001-2009. Estimation is performed using Bayesian techniques. The empirical findings suggest that German dairy farms achieve high levels of technical efficiency, while farms' degree of intensification indeed drives several technology parameters. Furthermore, model comparison based on Bayes factors reveals that the employed model outperforms a simple stochastic frontier model and a random-coefficients stochastic frontier model.