Bayesian analysis in econometrics (Q1262067)

Five basic propositions on Bayesian analysis in econometrics are put forward, discussed and illustrated through applications in Bayesian estimation, prediction, control and decision processes. The five propositions deal with the unity of science principle, the Jeffreys- Wrinch simplicity postulate, the prediction principle, a subjective concept of probability and the learning model embedded in Bayes' theorem on prior and posterior probabilities. For applied econometric problems, the following applications are of particular importance: (a) estimation of Stein's n-means, (b) estimation using an asymmetric loss function and (c) control problems viewed as regression processes.