On the Choice of Prior in Bayesian Model Averaging

CentER Working Paper No. 2011-003

29 Pages Posted: 31 May 2011 Last revised: 26 Apr 2017

See all articles by John H. J. Einmahl

John H. J. Einmahl

Tilburg University - Department of Econometrics & Operations Research

Kamlesh Kumar

Durham University, Business School, Department of Economics and Finance, Students

J.R. Magnus

Vrije Universiteit Amsterdam, School of Business and Economics

Date Written: January 21, 2011

Abstract

Bayesian model averaging attempts to combine parameter estimation and model uncertainty in one coherent framework. The choice of prior is then critical. Within an explicit framework of ignorance we define a ‘suitable’ prior as one which leads to a continuous and suitable analog to the pretest estimator. The normal prior, used in standard Bayesian model averaging, is shown to be unsuitable. The Laplace (or lasso) prior is almost suitable. A suitable prior (the Subbotin prior) is proposed and its properties are investigated.

Keywords: Model averaging, Bayesian analysis, Subbotin prior

JEL Classification: C11, C51, C52

Suggested Citation

Einmahl, John H. J. and Kumar, Kamlesh and Magnus, Jan R., On the Choice of Prior in Bayesian Model Averaging (January 21, 2011). CentER Working Paper No. 2011-003. Available at SSRN: https://ssrn.com/abstract=1744912 or http://dx.doi.org/10.2139/ssrn.1744912

John H. J. Einmahl (Contact Author)

Tilburg University - Department of Econometrics & Operations Research ( email )

P.O. Box 90153
5000 LE Tilburg
Netherlands

Kamlesh Kumar

Durham University, Business School, Department of Economics and Finance, Students ( email )

Durham
United Kingdom

Jan R. Magnus

Vrije Universiteit Amsterdam, School of Business and Economics ( email )

De Boelelaan 1105
Amsterdam, 1081HV
Netherlands

Register to save articles to
your library

Register

Paper statistics

Downloads
90
Abstract Views
590
rank
284,083
PlumX Metrics