文献类型：期刊文献

英文题名：Stable versus fragile community structures in the correlation dynamics of Chinese industry indices

作者：Nie, Chun-Xiao[1];Song, Fu-Tie[2]

机构：[1]Zhejiang Gongshang Univ, Sch Stat & Math, Hangzhou 310018, Peoples R China;[2]East China Univ Sci & Technol, Sch Business, Dept Finance, Shanghai 200237, Peoples R China

年份：2023

卷号：167

外文期刊名：CHAOS SOLITONS & FRACTALS

收录：;EI(收录号：20230213375105);WOS:【SCI-EXPANDED(收录号:WOS:000975125300001)】；

基金：Acknowledgement This research was supported by the Shanghai Planning Office of Philosophy and Social Science (2021BJB002) .

语种：英文

外文关键词：Industry index; Crisis; Influence-strength; Network; Correlation

摘要：Correlation dynamics in the stock market is recently gained wide attention, especially its relationship to financial crises. This study examines the correlation dynamics of the Chinese market based on time series data of industry indices. We examine the dynamics of linear and nonlinear dependencies using PCC (Pearson correlation coefficient) and MIC (Maximal information coefficient), respectively. The influence-strength (IS) analysis showed that the former was more closely associated with major events. We analyse the structure of the correlation networks from both the short-and long-term perspectives. For the short-term perspective, we employ IS analysis to extract the time stamps of structural breaks in the correlation structure. The timestamps of critical events in the correlation dynamics usually correspond to major events in the real market, including the 2015 stock market crash in China. We construct a network for each correlation matrix to extract the main correlation structure and detect the community structure. We create the long-term community-based network based on community structure. The long-term community structure of the correlation dynamics has clear economic implications and is thus stable, whereas the short-term community structure is fragile. This study analyses the correlation dynamics in the Chinese market, which is valuable for asset allocation, such as identifying market states and investing in index funds.