In parametric efficiency studies, two alternative approaches exist to provide an estimate of the long-run efficiency of firms: the dynamic stochastic frontier model and the generalised true random-effects model. We extend the former in order to allow for heterogeneity in the long-run technical efficiency of firms. This model is based on potential differences in firm-specific characteristics and in firms’ inefficiency persistence. The model is applied to an unbalanced micro-panel of German dairy farms over the period 1999 to 2009. Estimation of long-run technical efficiency and inefficiency persistence is based on an output distance function representation of the production technology and estimated in a Bayesian framework. The results suggest that heterogeneity in long-run technical efficiency of farms is mostly attributed to discrepancies in farm-specific factors rather than differences in farms’ inefficiency persistence. Farm size is positively related to long-run technical efficiency while subsidies exert a negative effect on the long-run technical efficiency of farms. Inefficiency persistence is found to be very high, but heterogeneity in this persistence is low.
- Dairy farms
- dynamic stochastic frontier
- inefficiency persistence
- long-run technical efficiency