Pushing the frontier : three essays on Bayesian Stochastic Frontier modelling

Abstract

This thesis presents three essays in Bayesian Stochastic Frontier models for cost and production functions and links the fields of productivity and efficiency measurement and spatial econometrics, with applications to energy economics and aggregate productivity. The thesis presents a chapter of literature review highlighting the advances and gaps in the stochastic frontier literature. Chapter 3 discusses measurement of aggregate efficiency in electricity consumption in transition economies in a cost frontier framework. The underlying model is extended to a Spatial Autoregressive model with efficiency spillovers in Chapter 4, showing good performance in simulations. The model is applied to aggregate productivity in European countries, leading to evidence of convergence between eastern and western economies over time, as in the previous chapter regarding efficiency in electricity consumption. Finally, Chapter 5 proposes a spatial model which allows for dependence in the structure of the inefficiency component while accounting for unobserved heterogeneity. This approach is applied to New Zealand electricity distribution networks, finding some evidence of efficiency spillovers between the firms. All essays explore the performance of the model using simulations and discuss the utility of the approaches in small samples. The thesis concludes with a summary of findings and future paths of research.