Time Series Analysis Of Industry Index Stocks

The gradual development and improvement of the financial m arket in recent years have further promoted the development of the real economy,im-proved the efficiency of re source al location an d ut ilization,an d sa tisfied the capital needs of enterprises and individuals.A sound and complete financial system plays an indispensable role in the market economy system.The most prosperous and vibrant market in the financial market is the stock m arket.Af-ter a long time of development,the theory and method of measuring the risk of stock market are gradually enriched.The risk of the stock market is mainly manifested by irregular and violent fluctuations in the stock price,and when the stock price of one country or region fluctuates significantly,the stock price of other countries and regions will also be affected.The chain reaction of this fluctuation i s t he s pecific pe rformance of fin ancial max ket correlation.When measuring the risk of financial markets,thec orrelation b etween assets is an important research object.Pearson’s correlation coefficient mainly studies the linear correlation between variables,and stock return series often have the char-acteristics of peaks and thick tails.The correlation between different assets is usually dynamic and asymmetric correlation may occur.For this complex situation,the copula function needs to be used for correlation measurement.There are few studies on the correlation between industries in the existing literature,and the data dimension is relatively low.Therefore,based on the Shenwan first-level i ndustry i ndex,t his p aper u ses t he i ndustry i ndex price data from 2017 to 2019 to study the correlation between multi-dimensional industries.The GARCH(1,1)model and GARCH(1,1)-t model are used to fit the logarithmic return series of the industry index,and the standard residu-al of the model is transformed by the probability integral to be used as the marginal distribution of the copula model.The Kendall rank correlation co-efficient of the t-copula model shows that there is no super strong correlation among industries;automotive,mechanical equipment,light industry manufac-turing,commercial trade,electrical equipment,chemical industry,textile and clothing have strong correlation with many industries.The upper and lower tail correlation coefficients indicate that except for the upper tail correlation coefficient of the banking industry and 16 industries,and the food and bev-erage and seven industries,which are greater than the lower tail correlation coefficient,the upper tail correlation coefficients between the other industries are less than the lower tail correlation coefficients,indicating that the possi-bility of these industry indexes plummeting at the same time is greater than the possibility of all industry indexes rising at the same time.When studying dynamic correlations between industries the GAS structure is applied to the factor copula model to build a dynamic single factor copula model.In the fac-tor copula model,it is assumed that the common factor is subject to partial t distribution and the specific factor is subject to t distribution,and then the maximum likelihood estimation method is used to estimate the parameters of the copula model.The change trend of factor loading in factor copula model re-flects the dynamic change of correlation between industries.Study shows that there is a strong correlation between different industries,and the factor load-ing of each classification has changed significantly since the end of September 2018.From 2017 to the end of the third quarter of 2018,although the overall fluctuation of factor loading is large,the degree of impact by common factors is almost unchanged,and the common factors has the strongest impact on industrial manufacturing and the weakest impact on financial business.With the further aggravation of China-US trade friction,the trend of factor loading has changed significantly from October 2018 to the end of 2019.Among them,the factor loading value of the chemical minerals category exceeds that of the industrial manufacturing category,the factor loading value is the largest,and the chemical mineral category is most affected by the common factor;the fac-tor loading value of the daily consumption category is smaller than that of the financial commerce and trade category and becomes smallest,which is least affected by the factor.