CFE 2015: Start Registration
View Submission - CFE
A0726
Title: Stochastic dominance criteria for normal mixture distributions Authors:  Markus Haas - University of Kiel (Germany) [presenting]
Abstract: If markets are subject to stochastic regime changes with Gaussian distributions within each regime, then the overall distribution of portfolio returns is a mixture of normals. To facilitate the application to the ranking of portfolios in markets characterized by regime-switching, recently established stochastic dominance criteria for Gaussian mixture distributions are considerably extended and simplified. E.g., the distributions of the portfolios in the comparison may have different regime probabilities, or even a different number of regimes. Moreover, both second- and fourth-order stochastic dominance criteria for normal mixture distributions are considered, where the latter is shown to fit nicely into recent definitions of higher-order risk attitudes (in particular, temperance) as preferences over particular lottery pairs. Extending the scope to fourth-order stochastic dominance allows the comparison of portfolios with the same overall variance, and, in case a risk-free asset is available, turns out to be equivalent to a comparison of the regime-specific and overall Sharpe ratios (excess return per unit of standard deviation) of the portfolios involved in the comparison.