基于实际收益率分布的均值-方差-条件风险价值多目标投资优化模型  

Mean-Variance-CVaR Portfolio Optimization Model Based on Real Return Distribution

作　　者：刘燕武[1] 张忠桢[1] 

机构地区：[1]武汉理工大学管理学院,武汉430070

出　　处：《系统管理学报》2010年第4期444-450,共7页Journal of Systems & Management

摘　　要：为改善均值-方差模型不能充分反映金融资产实际收益率分布的不足,在不对金融资产收益率分布做任何假设的基础上,引入条件风险价值度量金融资产重大损失风险,建立均值-方差-条件风险价值多目标投资优化模型,提出计算模型有效前沿的理论基础和算法步骤。基于上证50指数成分股的实际数据计算了该模型的有效前沿。计算结果表明:所提出的算法具有满足投资实践所要求的可操作性;投资组合实际收益率不服从正态分布,均值-条件风险价值模型有效集并不是均值-方差模型有效集的子集;相对均值-方差模型和均值-条件风险价值模型,均值-方差-条件风险价值模型能够更好地反映金融资产的实际收益率分布,提高投资者管理投资风险的能力。The deficiency of the classical model cannot be overlooked under the background of big fluctuation of financial market.In order to improve the ability of the classical model for reflecting the real return distribution,the paper introduces conditional value at risk（CVaR） to evaluate great loss risk of financial assets,establishes mean-variance-CVaR model,and presents the theoretical fundament and algorithmic steps for calculating the efficient frontier of the model.Based on the real price data of composition stocks of SZ50 Index,the efficient frontier of the model is calculated.The computational results show that the algorithm has high tractability to satisfy the demand of investment practice,the efficient set of mean-CVaR model is not subset of the efficient set of mean-variance model because the real return distribution of the efficient portfolio is not subjected to normal distribution,and the mean-variance-CVaR model can describe the real return distribution of financial assets better and therefore improve investors＇ ability to manage financial risk.

关 键 词：均值-方差模型 均值-条件风险价值模型 均值-方差-条件风险价值模型 重大损失风险 

分 类 号：F830.9[经济管理—金融学]