基于尾部风险差异性态度的多目标投资组合策略  

Multi-Objective Optimal Portfolio with Different Tail-Risk Attitudes

作　　者：赵霞 时雨 欧阳资生 ZHAO Xia;SHI Yu;OUYANG Zisheng(School of Statistics and Information,Shanghai University of International Business and Economics,Shanghai 201620;School of Business,Hunan Normal University,Changsha 410081)

机构地区：[1]上海对外经贸大学统计与信息学院,上海201620 [2]湖南师范大学商学院,长沙410081

出　　处：《系统科学与数学》2022年第5期1129-1144,共16页Journal of Systems Science and Mathematical Sciences

基　　金：国家自然科学基金(71671104,11971301);湖南省自然科学基金(2021JJ30196)资助课题。

摘　　要：近年来金融市场危机频发,极端风险在资产管理中受到较大关注.考虑到投资者的尾部风险态度对投资策略的影响,文章利用改进的峰度指标和K-means方法对资产进行分组,对不同类别资产组的尾部极端风险赋予不同的厌恶态度,基于均值-方差-CVaR准则提出了两类优化模型,进行投资组合策略的模拟分析和实证研究.研究结果表明:在给定目标收益的条件下,文章提出的优化模型通过减小总投资比例,减小了投资组合的风险暴露;与不考虑尾部风险态度差异的模型相比,新模型能够更好地控制风险,并具有更稳定的外样本表现;特别是文章提出的第二种模型,可以通过调整投资者对不同资产组的尾部风险厌恶水平,得到个性化的投资策略.Financial crises have occurred frequently in recent years,and hence,the extreme risk has attracted a great attention in asset management.Considering the effect of investor’s aversion toward tail-risk on investment strategy,this paper uses the adjusted kurtosis coefficient and K-means method to cluster assets.Subsequently,considering that investors hold different attitudes toward the tail-risks of different asset groups,two multi-objective optimization models are constructed based on meanvariance-CVaR criterion.From the simulation and empirical study,we find that the proposed models control the risk of portfolio by reducing the total investment ratio for a given target return.Compared with the model without the consideration of investor’s different attitudes,the proposed models can better control the risk and have a more robust out-of-sample performance.Especially,when the investor’s aversion toward tail-risk changes,the second model proposed in this paper can obtain a more adaptive investment strategy by adjusting the aversion coefficients for different asset groups.

关 键 词：均值-方差-CVaR准则 投资组合 尾部风险态度 K-MEANS聚类算法 

分 类 号：O212.1[理学—概率论与数理统计] F830.9[理学—数学]