The Studies On Topological Structure And Stability Of Stock Markets Based On The Directed Partial Correlation Networks

Although Chinese stock market has gradually recovered from its abnormal fluctuation in 2015,it has always been a hot topic for researchers and regulators to understand the topological structure and stability of stock market complex network and to prevent possible risk of stock market spreading and then deteriorating into a crash in a timely manner.This paper constructs the directed network of constituent stocks of CSI800,and discusses the characteristics of the network topology by using Partial Correlation Planar maximally filtered Graph(PCPG).The findings suggest that the out-degree network of CSI800 is a scale-free network,however,there is no scale-free characteristic for the in-degree networks of CSI800.The extent of each stock influenced by other stock's price fluctuations is about the same.Specially,we attack the stock market network using both random and deliberate way in order to measure the stability of the stock market network.The consequence indicates that owing to the closer correlation between stocks during the abnormal fluctuation period,the network global efficiency decreases more rapidly during the abnormal fluctuation time than other times,and the network has a more unstable structure once it is attacked.In the part of the sectorial network analysis,the sectors of information technology and energy have a great impact on the network before the stock market abnormal fluctuation.During the market abnormal fluctuation period,the industrial sector became the largest out-degree node in the stock market,while the consumer industry became the largest node after the abnormal fluctuation.In addition,the impact coefficient of financial sector has negative values before and during the stock market abnormal fluctuation,but it becomes the largest one among all the sectors with positive value of impact coefficients after the abnormal fluctuation.In the part of the inter-industry network,the intensity of industrial network shows the tendency from small to big first,and then from big to small in the three periods,while that in the sector of information technology and energy shows the opposite tendency.Especially,the density of the financial inter-industry network increases during the period because of the intention of value investing proposed by China Securities Regulatory Commission,and the implementation of Shenzhen-Hong Kong Stock Connect accelerates the communication and cooperation among financial institutions.Then,this paper analyses the risk contagion of sectorial directed network by using SIR model,and the result suggests that once few sectors that has a high out-degree experience a dramatically fluctuation,will most likely cause a massive infection to the stock market and destroy the stock market.This paper helps to reveal the change of Chinese stock market,to provide theoretical basis for regulators to better develop comprehensive regulatory policies,to give new enlightenment for listed companies to maintain stability of their stock price,and to propose investors instructional suggestions for making scientific investment decision.