Robust Portfolio Selection with Subjective Risk Aversion : A Mixed R-Vine Copula Approach with Change-Point Detection

This paper solves the robust portfolio selection problem with spectral risk measures under mixture R-vine copula uncertainty. Spectral risk measures are used to capture investors’ subjective risk aversion while R-vine copula change-point detection is employed to better construct mixture R-vine copula uncertainty. We introduce worst-case spectral risk measures (WSRM) and formulate portfolio selection problems with WSRM as risk measures. We conduct an empirical study using eight global stock indices. The results demonstrate that our robust portfolio outperforms two other portfolios that rely on a single R-vine copula or use mixture R-vine copula uncertainty constructed only through common economic crisis information, in both static and dynamic investments during the full out-of-sample period as well as crisis periods. Furthermore, the performance of our dynamic robust portfolio increases with investors’ risk aversion levels. Investors with higher risk aversion can benefit from their risk preferences by investing in our dynamic robust portfolio