An application of Constrained M-Estimator in construction of robust portfolio

The mean-variance portfolio model assumes that the returns follow a multivariate normal distribution. Unfortunately in actual financial markets, the empirical distribution of asset returns may in fact be asymmetric or multivariate elliptically symmetric with heavier tails. Therefore, the resulting optimal portfolio by this model will be heavily biased. For this reason, in this paper, we construct a robust mean-variance portfolio that has better stability performances. The robust portfolio is constructed using certain robust estimator, i.e. Constrained M-Estimator. Based on simulation and empirical results, we can conclude that our proposed robust portfolios are better than classical portfolios in all cases investigated.