APPLICATION OF THE GENERALISED McFADDEN COST FRONTIER AND CONDITIONAL DEMAND FUNCTIONS IN A TECHNOLOGY ANALYSIS - THE BAYESIAN APPROACH
Languages of publication
In this paper technology characteristics are compared after estimation of two competing models: a system of conditional factor demand functions and a stochastic cost frontier. Both models allow for economic efficiency, but the multiple-equation model enables decomposition of the overall efficiency to the technical and allocative components. The allocative inefficiency is based on an idea of internal (shadow) prices. The paper consists of two main parts: in the first one an application of microeconomic producer theory is presented that is later applied to construct the McFadden non-minimal cost function and the conditional factor demand functions . The second part deals with the Bayesian estimation of the conditional factor demand function and contains the comparison of technology characteristics calculated after estimation of the models. The results indicate that the system of demand functions describes the underlying technology much better than the single equation cost function. For the approximation of marginal posterior distributions of latent variables and parameters the Markov Chain Monte Carlo are used, that is the Metropolis-Hastings algorithms is implemented within the Gibbs sampler.
Publication order reference
CEJSH db identifier