Robust Bayesian portfolio optimization

Abstract

We propose a robust Bayesian model using the normal-inverse-Wishart and Gamma distributions for an investment portfolio consisting of the stocks of the United States Dow Jones Industrial Index. To this end, the Bayesian approach and the robust portfolio model are integrated to determine the uncertainty of the estimated parameters in expected returns and covariances using ellipsoidal or quadratic type uncertainty sets. The results show that the proposed method exhibits better performance and diversification than the traditional mean-variance model as well as the robust portfolios.