Stability analysis of optimal mean-variance portfolio due to covariance estimation

The objective of this paper is to study the stability of the mean-variance portfolio optimization. The results of the mean-variance optimal selection problem are very sensitive to the model parameters (portfolio calibration window and frequency of portfolio rebalancing). There are presented three methods of stabilization of covariance matrix estimation and further analysis are focused on the influence of estimation of covariance matrix to robustness of optimal selection. For practical verification, the proposed approaches were tested the suitability of these methods to the performance of the investment portfolio. This were done within the framework of 2000 to 2016 using daily data of 100 companies from the New York Stock Exchange.