稳健的双重均值方差管理在投资组合构建中的应用  

Research on Robust Mean-Variance Managed Portfolios in Chinese Stock Markets

作　　者：熊巍 沈童 唐晓彬 XIONG Wei;SHEN Tong;TANG Xiao-bin(School of Statistics,University of International Business and Economics,Beijing 100029,China)

机构地区：[1]对外经济贸易大学统计学院,北京100029

出　　处：《数理统计与管理》2025年第1期55-69,共15页Journal of Applied Statistics and Management

基　　金：国家社会科学基金重大项目(22&ZD164);国家自然科学基金青年项目(12001101);对外经济贸易大学中央高校基本科研业务费专项资金(CXTD14-05)。

摘　　要：本文通过利用适应性变系数因子模型(FACE)估计资产收益率的协方差,构建了一种对投资组合内部头寸及总体头寸进行双重均值方差管理的方法,并在我国股票市场数据上进行了检验。研究发现,组合内运用动态稳健协方差估计的均值方差策略在收益上大幅优于其他策略,卖空限制对策略收益有较大负面影响。FACE模型能够通过因子系数的变动较好的拟合股票市场变化,从而提供动态准确的估计。在组合总体头寸方面,均值方差管理能够在市场剧烈波动时期及时降低头寸,从而大幅降低损失并提升收益。进一步对收益率预测准确性的影响进行研究,发现收益率精准度对组合内均值方差管理影响较大而对总体头寸管理影响有限。本文的研究弥补了以往均值方差方法实证文献中对组合协方差估计不够精确的不足;同时,本文提出的方法可视为一种新的择时方法,对于投资者在市场高风险时期降低损失、保护收益具有较高的现实意义。This paper aims to construct double mean variance managed portfolios by employing the adaptive varying coefficient factor model(FACE)to estimate the covariance of the return on assets.We test the performance of these portfolios using data from the Chinese stock market.The results demonstrate that mean variance managed portfolios,with covariance estimated by FACE,significantly outperform other strategies in terms of returns.Additionally,we find that short selling restrictions have a significant impact on reducing returns.The FACE model has a capability of accurately capturing changes in the stock market by adjusting the factor coeficients,providing dynamic and accurate estimates.In terms of portfolio positioning,mean variance management allows for timely reduction of position during periods of market volatility,leading to significant loss reduction and improved returns.We also examine the impact of return prediction accuracy and find that it has a greater effect on mean variance management within the portfolio,but a limited effect on overall position management.This paper also addresses the accuracy of covariance estimation,an aspect that has often been overlooked in previous literature.Our proposed method can be considered as a new timing strategy with practical significance for investors seeking to minimize losses and protect gains during periods of high market risk.

关 键 词：均值方差模型 投资组合 动态结构 因子模型 协方差矩阵估计 

分 类 号：O212[理学—概率论与数理统计]