Comparisons of Model Averaging Techniques: Assessing Growth Determinants

Shahram Amini, Virginia Polytechnic Institute and State University
christopher parmeter, University of Miami

Abstract

This paper replicates three important studies on growth theory uncertainty that employed Bayesian model averaging tools. We compare these results with estimates obtained using recently developed frequentist and alternative Bayesian model averaging techniques. Overall, we successfully replicate all three studies using freely available software in the statistical environment R, provide an easily implementable algorithm to operationalize the frequentist model averaging methods and find that the sign and magnitude of these new estimates are reasonably close to those produced via traditional Bayesian methods.