Mean-Variance Efficiency Versus Positive Skewness Seeking In Portfolio Selection

This paper compares the performance of efficient portfolios based on the Markowitz (1952) mean-variance model with portfolios with high skewness. The assumption that the return of assets follows the normal distribution is the basis of the mean-variance model. However, the return of financial assets often deviates from the normal distribution, namely due to positive skewness, which may offer some advantages to investors. Previous literature reports several obstacles that make it difficult to include skewness in portfolio optimization, and that there is a trade-off between return and skewness maximization. Mean-variance optimization versus the search for positive skewness is addressed in this paper by estimating comparative performance ratios between mean-variance efficient portfolios and portfolios with high skewness. The paper also estimates probit models which highlight the probability of obtaining higher return from portfolios with high skewness than from mean-variance optimized portfolios. The probability given by our estimations is, in general, relatively low, which suggests that mean-variance optimization must be preferred to the search for positive skewness as method of portfolio choice.