A Bayesian approach to analyze regional elasticities

This paper presents a bayesian approach to analyze regional elasticity distributions with a regular translog cost function. It is known that a proper statistical analysis concerning elasticities can be performed only with the bayesian approach. Morover we can take advantage of this methodology to form reasonable priors using national data. This way we can produce sounder inferences without much elicition by the analyst. To compare results, this approach is applied to a cost function for the main regions in Italy with a diffuse prior too. Price and substitution elasticities are derived conditional on factor shares or covariates. The low posterior probability than inequality constraints hold with an noninformative prior shows how bayesian methods can be fruitfully employed to assess regional elasticities with a proper prior obtained from national data.