Analysis of Correlation based Threshold Networks of Dow Jones stocks of USA: An Econophysics Approach

ABSTRACT

We have investigated the time series of constituents of the Dow Jones Industrial Average (DJIA) for a period of 18 years (2000-2018). DJIA is a dominant stock market index comprising of thirty US based companies. We have applied the Random Matrix Theory (RMT), complex network analysis and hierarchical clustering techniques to extract out the information from the time series of DJIA stocks. The impact of sub-prime crisis of 2008(FC08) on structure and dynamics of network of DJIA stocks is studied by diving the periods under consideration into three distinct periods;pre crisis (PRC), during crisis (DUC) and post crisis (POC) on the basis of volatility. The RMT analysis shows that data analyzed contain important information. Network analysis reveals high correlation among the stocks in the DUC period. The MST and hierarchical clustering techniques support the results of RMT analysis. Degree centralities, closeness centralities and clustering coefficients of DJIA networks increases in DUC period. High correlation and closeness among stocks in DUC period is depicted in various analyses. The dynamic analysis is also carried out which detect various extreme events such as Covid-19. In conclusion, investigation shows that during the period of crisis, there are significant changes in the structure and dynamics of DJIA network. The findings of investigation can be utilized as risk indicator and detection of such crises in future. © 2022. School of Science, IHU. All rights reserved.