Industry Systemic Risk In Stock Market Based On Correlation Networks

The transformation of economic development and the optimization of industrial structure have promoted the deepening of interrelationship between industries in our country.At the same time,extensive and close ties have also emerged between the financial industry and the real economy industries through lending and investment relationships,which has greatly increased the probability of systemic risk contagion across industries.Therefore,analyzing the complex relationships among industries by constructing correlation networks and studying the influence of network structure on the systemic risk spillover effect of industries can help prevent the occurrence and spread of systemic risks.This paper selects 28 first-class industry indexes as the research object,builds risk correlation networks and conducts research.Firstly,this paper measures the transfer entropy of Va R series among different industries based on information theory to portray the risk information transfer between industries.With the threshold method to filter redundant information,industry risk correlation networks are conducted and analyzed.Secondly,this paper uses Co Va R model to measure the systematic risk spillover effect of different industries,the DCC-GARCH method is chosen to calculate the ΔCo Va R series of industries and the dynamic correlation coefficients of industries and stock market system.The averageΔCo Va R of industries in the sample period is ranked and the dynamic evolution characteristics of the systematic risk spillover effect of different industries are analyzed.Finally,a panel model is constructed based on the industry risk correlation networks to explore its influence on the industry systemic risk spillover effect in terms of the structural characteristics of the network nodes.The findings are as follows:(1)The risk correlation networks of industry in China are both consistent and time-varying: there are central nodes of the network throughout and the network structure changes over time.The household appliances,food and beverage,and banking sectors are at the center of the network for most of the time windows during the sample period,other sectors show different positions in the network during bull and bear markets,with the network centrality of the non-banking financial sector increasing during bear markets.The small-world nature of the network strengthens during stock market volatility,and network density increases significantly during bear markets.(2)The systematic risk spillover effects varies significantly across industries,but the overall fluctuation trend is consistent,reflecting that the systematic risk spillover effect of industries is influenced by both industry-specific factors and macro factors.In terms of industry differences,the magnitude of systematic risk spillover effect varies across industries,as well as the specific timing and duration of fluctuations.In terms of volatility trends,the systematic risk spillover effect of industries increases during periods of stock market volatility,especially during the 2015 stock market disaster.(3)The risk correlation network structure has an impact on the systematic risk spillover effect of the industries.The node out-degree,in-degree,out-intensity and in-intensity of the lagged period are inversely related to the magnitude of the systematic risk spillover effect of the industry,and the risk information transmission between industries plays more of a risk-sharing and risk-prediction role.The findings of this paper will provide some reference value for regulators to understand inter-industry linkages and prevent cross-industry contagion of systemic risk.