基于动态参照点的鲁棒投资组合模型及其拓展

鲁棒投资组合模型是一种适用于收益不确定条件下寻求最优决策的方法。首先考虑投资者对底线的重视,根据当收益触及底线时,激进者和保守者在参照点上的不同变化情况,建立动态参照点模型。接着,一方面将动态参照点作为划分获益和损失的界限值,改进现有的Worst-case Omega(WOmega)模型。另一方面结合投资者对下侧风险更为厌恶的特点,以动态参照点作为下侧风险的基准,改进现有的Relative Robust Portfolio Optimization(RRPO)模型。实证研究中,对于WOmega类模型,结果表明激进行为模型在样本内表现较好,而保守行为模型在样本外表现较好。对于RRPO类模型,结果显示激进行为的收益表现良好,保守行为对标准差及最大损失值的控制较好。随着约束的放松,所有模型的收益都能得到可观提升。

Robust Portfolio Models Based on Dynamic Reference Point and Their Extensions

The robust portfolio model is a method for seeking the optimal decision under the condition of uncertainty. First, considering the investor's emphasis on the minimum requirement, we establish a dynamic reference point model based on the different attitudes of the aggressive and conservative investors when the return touches the minimum requirement. Next, on the one hand, the existing Worst-case Omega(WOmega)model is improved by using the dynamic reference point as a threshold for dividing the gain and loss. On the other hand, combined with the investor's characteristics of the downside risk aversion, the existing Relative Robust Portfolio Optimization(RRPO)model is improved by using the dynamic reference point as the benchmark for the downside risk. In the empirical study, for the class of WOmega models, the results show that the aggressive model performs best during the in-sample periods, while the conservative model performs best during the out-of-sample periods. For the class of RRPO models, the results show that the aggressive behavior yields relatively high return, while the conservative behavior controls the standard deviation and the maximum loss relatively better. With the relaxation of constraints, the return obtained by all models can be significantly improved.