On regularized mean-variance-CVaR-skewness-kurtosis portfolio selection strategy

Markowitz revolutionized the concept of portfolio section in 1952, leading to what is now known as the Modern Portfolio Theory (MPT). However there exist some flaws in his proposed mean-variance model such as non-normality, use of variance as a risk measure and stability of the optimization model. The purpose of this research is to improve the dimensionality of portfolio optimization decision via Polynomial Goal Programming approach from mean-variance-skewness and mean-variance-skewness-kurtosis to a stable mean-variance-conditional-value-at-risk-skewness, thereby providing a better risk measure with the merging of variance and conditional-value-at-risk (CVaR), alleviating over-fitting or estimation risk problems via norm regularization aside considering more complete information on stock returns distribution. To provide more detailed financial outlook, we subjected our proposed model to numerical test. The empirical results show that our model is well diversified and balances the risk-return tradeoff as compared to others selected from literature.