中国股市行业板块的动态相关性建模——基于DCC—MVGARCH模型

相关性的讨论是现代金融分析的重要内容，行业板块的相关分析是组合投资的关键步骤。基于能刻画动态相关的DCC-MVGARCH模型对我国波动剧烈的六个股市行业板块进行了相关性研究，结果表明：十个板块相关性序列可看成常量，Pearson相关系数仍能刻画其相对大小，这为机构投资者按Pearson相关系数进行组合构建提供了实证依据；五个动态相关性序列是宽平稳而非严平稳的，适合采用随机过程建模以实现预测，另外动态相关性与时变波动率存在一定的关系，当波动率增强时，相关性有随之增大的趋势。

Dynamic Correlations Modeling of Industry Segments in Chinese Stock Market-Based on DCC-MV-GARCH Model

Correlation discussion is very important in modern financial analysis, and correlation analysis of industry segments is the key step in the portfolio. In this paper, based on DCC - MVGARCH model, which can represent the dynamic correlations, correlations of six fluctuant industry segments in Chinese stock market are investigated, and the results show that: ten correlations between segments can be thought as constants, and Pearson correlation can still depict their relative size, which provides empirical evidence for institutional investors to construct portfolio by employing Pearson correlation. Five dynamic correlation series are stationary but not strict stationary, so it is suitable to forecast by stochastic process modeling. Furthermore, dynamic correlation and time-varying volatility has certain relationship that the correlations have the tendency of increase with the more fluctuant heteroskedasticities.