An L-Moment Approach for Portfolio Choice under Non-Expected Utility

Abstract We develop and apply a novel semi-parametric estimation method based on L-moments. Unlike conventional moments, L-moments are linear in the data and therefore robust to outliers. The estimation method provides a series expansion that quickly converges to the underlying return distribution and can be used when conventional moments do not exist. An extensive empirical analysis of portfolio choice under non-expected utility demonstrates the effectiveness of our approach. Empirical results show that our method copes well with estimation risk, yields stable portfolio returns, and reaps the information content of moment returns beyond order four.