Robust Bayesian Portfolio Optimization

Speaker: Dr. Carlos Andres Zapata Quimbayo, ODEON, Universidad Externado de Colombia, Bogota

Abstract: We implement a robust Bayesian framework for portfolio optimization that integrates Bayesian inference with robust optimization techniques. The model considers parameter uncertainty in expected returns and covariances by combining normal-inverse-Wishart and gamma distributions through ellipsoidal uncertainty sets. We apply this methodology to a stock portfolio from the Dow Jones Industrial Average and compare its performance with traditional mean-variance and robust portfolio models.