Black-Litterman Model for Portfolio Performance Enhancement - An Out-Of-Sample Evaluation of the Black-Litterman Model on a U.S. Stock-Dominated Portfolio

Abstract

In this thesis, the Black-Litterman model is evaluated out-of-sample and compared to mean-variance and naïve allocation. Two references are implemented in the Black-Litterman framework, the minimum-variance and naive portfolios. The study complements previ-ous work by considering a stock-dominated portfolio, where all assets are from the U.S. market. Although performance enhancements are observed, there is no significant dif-ference between the Black-Litterman and mean-variance portfolios for three different in-vestor types. Furthermore, the initial parameters are varied to control for different market environments and portfolios compositions. Finally, transaction costs are added and varied to investigate different trading conditions. This test concludes that the BL portfolios require more precise return and volatility estimates to compensate for high portfolio turnover.