Published online by Cambridge University Press: 06 April 2009
In a series of recent articles ([2], [3], [4], [5]) R. B. Porter and his associates have conducted empirical comparisons of the Mean-Variance (EV) and Stochastic Dominance portfolio choice criteria. The basic methodology of all these studies was first to compute the set of EV-efficient portfolios by an optimizing algorithm, then to find through heuristic methods “stochastically dominant” portfolios, and finally to compare the two. A major finding of these studies was that most EV-efficient portfolios survived the second-degree stochastic dominance (SSD) test against the randomly generated portfolios. The purpose of this note is to show that, for all cases of practical interest, a portion of the EV frontier is a subset of the SSD-efficient set. In other words, we offer here an exact theoretical justification of some empirical results of the aforementioned studies.