Abstract: The effects of higher-order risk aversion on optimal cross-sectional portfolio choice are investigated using portfolio optimization with Stochastic Dominance constraints. Tractable sufficient conditions for higher-degree dominance are introduced that take the form of a system of linear inequalities. Existing studies of active equity industry rotation are extended from lower degrees to higher degrees of dominance. Fourth-degree dominance assumes that investors are ‘prudent’ and ‘temperate’ and therefore like skewness and dislike kurtosis. Using this dominance criterion leads to superior out-of-sample investment performance, by allowing for more concentration in recent winner industries which tend to show persistent positive abnormal returns and a favorable higher-order risk profile due to the industry-level price momentum effect.

Keywords: Portfolio choice; Higher-order risk; Portfolio optimization; Linear programming; Active portfolio management