Robust Portfolio Selection With Higher Moments And S-shaped Utility

We develop a model for robust portfolio selection using arctan estimator that addresses two major shortcomings of the Markowitz approach:the ability to handle higher moments and estimation error.Our model has several attractive features:(1) it can be computed by solving a single nonlinear program,where robust estimation and portfolio optimization are performed in a single step;(2) it is based on S-shaped utility,which provides the decision-theoretic foundation;and(3) it can incorporate higher moments and does not require to explicitly compute any estimate of the coskewness and cokurtosis matrix.We also show analytically that the resulting portfolio weights are stable. Moreover,our numerical results on simulated and empirical data suggest that the proposed portfolios have good out-of-sample performance.