Stochastic Dominance in Portfolio Analysis and Asset Pricing

textabstractStochastic Dominance relation is a probabilistic concept which allows random outcomes such as portfolio returns to be ranked, by utilizing the full information about the distribution of the returns, in contrast to the mean-variance rule or other mean-risk models which only use a single statistic to measure portfolio risk. This thesis focuses on algorithmic and conceptual advancements in the application of the Stochastic Dominance framework in portfolio analysis and asset pricing. In particular the thesis points out the importance of SD-efficient sets being convex, introduces and studies a concept of Almost Stochastic Dominance which is a meaningful and tractable relaxation of SD, and proposes an asset allocation strategy based on Stochastic Dominance and analyzes its performance relative to other strategies