Research On The Impact Of Stock Market Disaster On The Stock Network Of A-share Listed Companies

In 2015,China's capital market experienced abnormal fluctuations.After two rounds of cliff-type declines,China's stock market fell by more than 45%.The stock market disaster has many adverse effects on China's social stability and stable economic operation,and has also profoundly affected the structure of the A-share market.Therefore,research on the impact of stock market crashes on the A-share market is of great value.The correlation between stocks is the relationship between stock prices or yields among different stocks.It is an important indicator for studying the systemic risk of stock market and the effectiveness of portfolios.The partial correlation coefficient between stocks is the effect of eliminating other stocks in the whole market.Under the premise,considering the relationship between two stocks,it can better reflect the "real relevance" between stocks,so it is necessary to measure the partial correlation coefficient.Suppose the variable follows a multivariate normal distribution,the inverse of the covariance matrix,also known as the precision matrix,where each element represents the partial correlation coefficient between the variables,so the precision matrix needs to be estimated.This paper selects 200 stocks before and after the 2015 stock market crash,and uses the Graphical lasso method to estimate the precision matrix.Based on this,this paper studies the structure of the stock network characterized by the precision matrix before and after the stock market disaster,and explores whether the network structure before the stock market disaster,the stock market disaster and the stock market disaster has changed significantly.It turns out that the three stages of the stock network are significantly different,and the stock market disaster makes it different.The correlation between individual stocks within the industry and the industry was closer;the financial industry was selected for further research,and then two representative stocks of Ping An and China Merchants Bank in the samefinancial industry were selected to calculate the partial correlation coefficient.The confidence interval is used to study the variation of the partial correlation coefficient of the three stages.It is also found that the correlation is enhanced and has a certain persistence.The above hypothesis test and confidence interval estimation are all estimated by the self-sampling method.In the case of high-dimensional,this method is superior to the traditional asymptotic estimation.Finally,the research results of this paper have certain reference value for the allocation of investment portfolio and the policy makers' policy formulation,which is beneficial to improve the effectiveness of the investment portfolio,and put forward the policy development should fully consider the impact on various industries,and must have A certain continuity and other suggestions.