Minimizing Conditional Value-at-Risk under Constraint on Expected Value

Conditional Value-at-Risk (CVaR) measures the expected loss amount beyond VaR. It has vast advantage over VaR because of its property of coherence. This paper gives an analytical solution in a complete market setting to the risk reward problem faced by a portfolio manager whose portfolio needs to be continuously rebalanced to minimize risk taken (measured by CVaR) while meeting the reward goal (measured by expected return). The optimal portfolio is identified whenever it exists, and the associated minimal risk is calculated. An example in the Black-Scholes framework is cited where dynamic hedging strategy is calculated and the efficient frontier is plotted.