Mean-variance portfolios constructed using the sample mean and covariance matrix of asset returns perform poorly out-of-sample due to estimation error. Recently, there are two approaches designed to reduce the effect of estimation error: robust statistics and robust optimization. Two different robust portfolios were examined by assessing the out-of-sample performance and the stability of optimal portfolio compositions. The performance of the proposed robust portfolios was compared to classical portfolios via expected return, risk, and Sharpe Ratio. The aim is to shed light on the debate concerning the importance of the estimation error and weights stability in the portfolio allocation problem, and the potential benefits coming from robust strategies in comparison to classical portfolios.
Supandi, E. D., Rosadi, D., & Abdurakhman. (2017). An empirical comparison between robust estimation and robust optimization to mean-variance portfolio. Journal of Modern Applied Statistical Methods, 16(1), 589-611. doi: 10.22237/jmasm/1493598720