鲁棒均值-CVaR投资组合模型及实证:基于安全准则的视角

Empirical Study on Robust Mean-CVaR Portfolio Model:A Safety Criterion Perspective

考虑证券市场的不确定性,将资产的收益率看成区间随机变量。利用鲁棒优化方法,构建鲁棒均值-CVaR投资组合模型。采用对偶理论,将鲁棒均值-CVaR投资组合模型转换为线性规划问题,降低了模型的求解难度,有助于计算大规模的资产组合。进一步地,考虑投资者的安全性需求,在模型中引入最大违反概率,控制模型的保守程度,并直观反映投资者的安全性要求。采用实证的方法,研究模型的有效性。结果表明:鲁棒均值-CVaR投资组合模型具有较好的稳健性,且满足投资者的安全性要求,在实际的投资决策中具有可行性。

Considering the uncertainty in the real stock market, the paper regards the security return as an interval random variable and develops a robust mean-CVaR portfolio model based on the robust theory. Following the duality theory, the proposed model can be transformed as a linear programming problem, which reduces the computational complexity and contributes to solve a large-scale portfolio model. To consider investors＇ safety requirement, a concept of the most violated probability is introduced, which can be used to adjust the conservatism of the proposed model and reflect investors＇ safety requirement intuitively. The performance of the proposed model is empirically studied. The result shows that the proposed model can be used to construct portfolios that exhibit robustness against the expected return, meet investors＇ safety requirement and are viable in real investment.