Analysis of Dynamic Correlation of Japanese Stock Returns with Network Clustering
Analysis of Dynamic Correlation of Japanese Stock Returns with Network Clustering
复制标题
日本股票收益动态相关性的网络聚类分析
DOI：
10.1007/s10690-017-9230-5
复制
发表时间：
2017
影响因子：
1.7
通讯作者：
Isogai Takashi
中科院分区：
文献类型：
--
作者：
土屋垣内 晶;平野 好幸;竹林 由武;清水 栄司;中川 彰子;Isogai Takashi
本文利用动态条件相关模型（DCC-GARCH）估计了日本股票收益率的动态相关性，并实证研究了它们之间的动态相关性。该模型很难同时适用于整个股票市场，因此，采用基于网络的聚类方法对样本数据进行降维。两种类型的相关性结构估计：同质组的股票在一个平衡的大小，通过聚类观察组内的相关性，而一个单一的投资组合，包括组投资组合收益率也被创建观察组间的相关性。估计结果揭示了由估计的相关矩阵的最大特征值表示的相关强度的动态变化。在危机期间，即雷曼破产和东日本大地震之后，观察到更高水平的相关强度和波动性，对于组间和组内相关性。还证实了相关性变化的模式在组之间是显著不同的。该方法可以有效地监控大规模投资组合中资产收益的动态相关性。
In this paper, the dynamic correlation of Japanese stock returns is estimated by using the dynamic conditional correlation (DCC–GARCH) model to study their correlation dynamics empirically. It is difficult to fit the model to the whole stock market jointly at the same time; therefore, a network-based clustering is applied for the dimensionality reduction of the sample data. Two types correlation structures are estimated: homogeneous groups of stocks in a balanced size are created by clustering to observe within-group correlation, while a single portfolio that comprises group portfolio returns is also created to observe between-group correlation. The estimation result reveals dynamic changes in correlation intensity represented by the largest eigenvalue of the estimated correlation matrix. A higher level of correlation intensity and volatility are observed during the crisis periods, namely after both the Lehman collapse and the Great East Japan Earthquake, for the between- and within-group correlations. It is also confirmed that the pattern of correlation change is significantly different between the groups. The proposed method is useful for monitoring dynamic correlation of asset returns efficiently in a large scale of portfolio.
影响因子：
2.1
作者：
Isogai, Takashi
通讯作者：
Isogai, Takashi
测试多元 GARCH 模型中的条件相关结构：广义跨谱方法
DOI：
10.1111/j.1468-2354.2011.00657.x
发表时间：
2011
期刊：
Wiley-Blackwell: International Economic Review
影响因子：
--
作者：
N. McCloud;Yongmiao Hong
通讯作者：
Yongmiao Hong
通过高斯混合模型对 GARCH 过程进行快速聚类
DOI：
10.2139/ssrn.2071716
发表时间：
2012
期刊：
ERN: Estimation (Topic)
影响因子：
--
作者：
Gian Piero Aielli;M. Caporin
通讯作者：
M. Caporin