The Relationship Between Industries Indices Return

The stock market is a typical complex network system,and there is a universal connection between different stocks.On one hand,the correlation of stocks in the same industry will be greater than the correlation of stocks in different industries;on the other hand,the closer the two companies are in the industrial chain,the more relevant the correlation is.This article mainly discusses the relationship of the latter,that is,the possible correlation characteristics between different industries.To comprehensively and meticulously study this problem,this article not only uses typical linear analysis methods,such as correlation coefficient analysis and linear Granger causality analysis but also uses linear analysis methods.There are also common nonlinear methods,such as analysis based on Copula functions.By constructing the minimum spanning tree network,Granger causality network,and the vine model of the Shenwan Industry Index,the author examines the characteristics of the industry index association from three levels.First,based on the minimum matrix model of the distance matrix and R-vine model,the representative industry,and the marginal industry in the index system can be determined.Studies have shown that chemical and mining are representative industries of A-shares.The finance and food-beverage industries are marginal.Secondly,through the construction of Granger causal network,it can be seen that the leading effect of the industry index during the stock market turmoil is stronger,and Granger causality is more obvious.Finally,by constructing the Copula quantile model,it is possible to quantitatively describe the strength of association between industry indicators.To examine the industry risk spillover effect under extreme risk conditions in the market,this paper calculates the %Co Va R of different industries and constructs the %Co Va R matrix to reflect the risk spillover effect at the tail of the index of each industry.Also,it uses D vine Copula quantile regression to measure the impact of banks,non-bank finance,and entire real estate on extreme market risks.