Gini's Mean Difference and Portfolio Selection: An Empirical Evaluation

Abstract

Yitzhaki [19] recently developed two portfolio selection criteria (EG and EI“) based on the mean and Gini's mean difference. Similar to mean-variance(EV), the EG criterion uses two summary statistics to describe the probability distribution of a risky prospect, the mean and one-half Gini's mean difference. Gini's mean difference is defined as the average of the absolute differences between all possible pairs of observations of a random variable. Yitzhaki's development concentrated on the theoretical aspects of EG and EI“ and the theoretical relationships among EG, EI“, EV, and stochastic dominance (SD) selection criteria. He did not address either the empirical properties of EG and EI“ or the relationship between the empirical efficient sets of EG and EI“ and other portfolio selection criteria. Yitzhaki suggested that the next step in the development and application of his proposed selection criteria should be an empirical investigation of how the EG and EI“ criteria compare with other selection criteria.