This paper studies a CVaR-based approach for a portfolio optimization, and investigates its practical applications in risk management and investment decision-making. Because results of the CVaR optimization model can be obtained by solving a linear programming (LP) problem, the model can deal with optimization problems of a portfolio with hundreds of instruments.On one hand, the CVaR optimization model is discussed in detail and an empirical research on it is performed with historical data. The empirical results show that one may reduce risk and increase return of a portfolio, -then improve its performance by constructing efficient frontier on the space of Expected Return / CVaR.On the other hand, this paper concentrates on VaR-based and CVaR-based investment decision-making. A case study is implemented with historical and simulative data to demonstrate their applications in asset allocation risk management portfolio construction and performance evaluation. |