Accounting for the non-normality of asset returns remains challenging in robust portfolio optimization. In this paper, we tackle this problem by assessing the risk of the portfolio through the "amount of randomness" conveyed by its returns. We achieve this by using an objective function that relies on R\'enyi entropy, an information-theoretic criterion that precisely quantifies the uncertainty embedded in a distribution, accounting for all moments. Compared to Shannon entropy, R\'enyi entropy features a parameter that can be tuned to play around the notion of uncertainty. It is shown to control the relative contributions of the central and tail parts of the distribution in the measure. We further propose a sample-based estimator of the exponential R\'enyi entropy by extending the robust sample-spacings estimator initially designed for Shannon entropy. The relevance of R\'enyi entropy in portfolio selection applications is illustrated with an empirical study: minimizing this cost function yields portfolios that outperform standard risk optimal portfolios along most performance indicators.
↧