Research On The Linkage Between My Country's Real Estate And Building Materials Industry

The related structure study of a financial asset or financial markets is an important content of modern financial analysis,seek a financial asset or financial markets between the related structures of reliable estimates have already became a hot issue of many scholars attention,however,with the traditional linear model analysis the limitation to the economic and financial issues,both from the explanatory and from the perspective of predictability,linear model can fully describe the structure of the actual financial data related to the generation process,also did not consider the related structure present different features in different historical periods.This paper focuses on the characteristics of the non-linear correlation structure of financial assets or financial markets in different historical periods such as bull market and bear market,and how to describe them?In fact,for this topic,we need to answer at least how to divide the bull market and the bear market,and what kind of function should be adopted to describe the non-linear correlation structure of financial assets.Based on this,this paper believes that first of all,the hybrid Copula function should be adopted to describe the nonlinear correlation structure of financial assets,in order to avoid the shortcomings of a single Copula function in describing the nonlinear correlation structure;Secondly,the hybrid Copula that describes the non-linear correlation structure between financial assets or financial markets should not be fixed throughout the sample cycle,but different in different historical periods such as bull market and bear market,that is,the hybrid Copula function should have different weights in different periods.In this context,the hybrid variable structure Copula model is proposed to answer the above questions.The specific research thinking of this paper is as follows:the first step is to take the Shanghai composite index as the basis for dividing the bull market and the bear market into sub-samples.This step will adopt the revised BB law to divide the bull market and the bear market.The second step is to establish the hybrid variable structure Copula model for the research object.For the hybrid variable structure Copula model,a five-step estimation algorithm is proposed.2)determine the initial value of hybrid Copula parameters;3)weight parameter estimation in hybrid variable structure Copula;4)parameter estimation of Copula in hybrid variable structure Copula;5)repeat step 3 and step 4 until all Copula function parameters and weight parameters converge,and the final convergent parameter value is the parameter estimation result of mixed variable structure Copula.During the process,the organic combination of full samples and sub-samples of each sub-interval,empirical distribution function,maximum likelihood estimation,EM algorithm and iterative method in numerical calculation were adopted to complete the parameter estimation of hybrid variable structure Copula model.Third,based on the research method proposed in this project,a hybrid variable structure Copula model was established for empirical analysis of real estate and building materials plate data to characterize the nonlinear correlation structures of the two in different periods.The fourth step is to use the estimation results of the hybrid variable structure Copula model to calculate the value of VaR through monte carlo simulation method,and to conduct Kupiec test on multiple Copula models to further illustrate the accuracy and stability of the hybrid variable structure Copula model in predicting the risk measurement of investment portfolio.The empirical results show that,from the Angle of weighting parameters,in a bear market,Clayton copulas connect the weight parameter is greater than 0.6279 Gumbel copulas connect the weight parameter of 0.3721,in a bull market period,Gumbel copulas connect 0.5089 slightly larger than the weight of Clayton copulas connect the weight of 0.4911,in bear markets Clayton has dominated absolute weights of copulas connect,on the one hand,the financial markets than chase after go up an exodus copycat more likely;From the results of Copula parameter estimation,in the hybrid variable structure Copula model,the parameter of Clayton Copula in the bull market is 0.4911 less than that in the bear market is 0.6279.Similarly,the parameter of Gumbel Copula in the bull market is 0.5089 more than that in the bear market is 0.3721,indicating that Gumbel Copula is dominant in the bull market and Clayton Copula is dominant in the bear market.There was an asymmetric tail correlation between the real estate industry and the building materials industry in the whole sample,with the lower tail correlation coefficient of 0.8257 greater than the upper tail correlation coefficient of 0.6492.It seems to illustrate a fact that the impact of bad news on the market is greater than the impact of good news on the market.Investors are more motivated to kill the fall than to chase the rise,with greater probability and stronger financial risk contagion.