Tail Correlation Measurement And Empirical Analysis Based On TailCoR Model

With the rapid development of financial market people are increasingly insep-arable from the financial market.Investors hope to get returns through the financial market.Enterprises want to expand production from the financial market.The govern-ment promotes the steady development of the economy through the financial market.However,financial crises continue to occur.As regions and industries become more and more closely linked,the impact of financial crises will become greater and greater.The financial crisis in 2008 and the debt crisis in Europe from 2010 to 2012 have caused tremendous damage to financial markets.More and more attention has been paid to the measurement of tail correlation.Tail correlation may be caused by linear correlation,Pearson correlation coefficient or non-linear correlation.In the stage of model building,this paper draws on the TailCoR model proposed by Ricci et al.(2015),a new tail-related measurement model.The original model has the advantages of small sample robustness,independent of specific distribution and can study asymmetric tail correlation.However,the TailCoR obtained from the original model is a number,which has many limitations in application.In this paper,TailCoR model is improved.In the process of calculating TailCoR,the quantile regression is used to replace the calculation of quantile deviation,and a dynamic TailCoR model is obtained.Under the assumption of elliptic distribution,TailCoR can be decomposed into linear and non-linear parts.In this paper,dynamic linear correlation coefficients are calculated by DCC-GARCH model.Finally,dynamic non-linear correlation coef-ficients are obtained by dynamic TailCoR and dynamic linear correlation coefficients.In the empirical research stage,we selected the daily return data of ICBC,Bank of China,Construction Bank and China Merchants Bank from January 2008 to June 29,2018.Through the full sample TailCoR model,piecewise TailCoR model and dynamic TailCoR model,we analyzed the nature and change rule of tail correlation of the four banks.There are the following conclusions:First,TailCoR will rise rapidly when the bull market starts and the market starts to collapse,and remain stable for the rest of the time;second,tail correlation is mainly contributed by the non-linear part;third,three large state-owned commercial banks,namely ICBC,Bank of China and Construction Bank,have high consistency of tail correlation,and in extreme market,tail correlation is very high.The correlation coefficient of ministries increased greatly,while the TailCoR between China Merchants Bank and three large state-owned commercial banks was relatively stable.