宏观因子、投资者行为与中国股债收益相关性——基于动态条件相关系数的实证研究

Research on Macro Finance Variables, Investor Behavior and Stockbond Return Correlation in China——A Dynamic Conditional Correlation Approach

本文基于DCC-MGARCH模型和分位数回归模型,计算了中国股债市场收益率的动态条件相关系数,进而研究了中国股债收益相关性的时变机制。研究发现,由于特殊的债券投资者结构,传统"避险行为"理论对中国市场的解释力不足;基准利率、经济增速和自相关性是中国股债收益相关系数条件均值的主要影响因素;并且,中国股债收益相关系数的条件分布也具有时变性,其尾部风险受基准利率和自相关性的显著影响。这些研究结果解释了中国股债收益相关性独特的波动路径,并为管理股债资产组合的相关性风险提供了分析工具。

This paper employs DCC-MGARCH model and quantile regressions to investigate the dynamics of China＇s stock-bond return correlation. It is found that the interest rate, economic growth and auto-correlation drive the conditional mean of China＇s stock-bond return correlation and that the ＂flight-to-safety＂ theory is invalid due to the structure of China＇s bond investors. Moreover, the conditional distribution of China＇s stock-bond correlation is time variant, and its tail risk is de- pendent on interest rate and auto-correlation. These findings explain the special variation pattern of China＇s stock-bond cor- relation and provide tools for analyzing the related risk of asset portfolio.