Dynamic relative Mean-RCVaR portfolio model based on stochastic benchmark

A dynamic portfolio selection problem with stochastic benchmark is investigated. The expected relative terminal wealth is maximized under a new risk constraint, RCVaR, which is defined by a relative wealth process. In a Black-Scholes setting, stochastic analysis method and nonlinear programming theory are used to obtain explicit solutions of the optimal strategies and the efficient frontiers, which include the riskless asset, revised market portfolio and benchmark portfolio. The results exhibit three-fund separation theorem. Numerical examples are presented.