Research On The Quantile Dependence Model And Its Application In Finance

The financial industry,as a tool for the country to achieve optimal allocation of social resources,is the blood to promote the sound operation of the real economy and is an important guarantee for safeguarding the healthy development of the national economy.With the continuous development of the modern financial system,the relationship between the financial market and the real economy has become increasingly close,and the complexity of the economic and financial system such as non-linearity,non-stationarity,and heterogeneity has also become increasingly evident.The analysis tools based on the traditional mean-dependence theory have been difficult to satisfy the description of the complexity characteristics of economic and financial variables.Therefore,the quantile-dependence measurement model,which can better characterize the distribution characteristics of the dependence among variables,has gradually become one of the mainstream analysis tools.This paper deeply investigates four hot issure in economic and financial studies,including the measuremen of systematic risk,the behaviour of co-movement between stock markets,the quantile behaviour of stock price synchchronicity and the extreme risk spillover effect of international crude oil to Chinese stock market,in the framework of quantile theory.By constructing the corresponding extreme expectile model,quantile cointegration model,cross-quantilogram model,and quantile Granger causality model,this peper is able to characterize the heterogeneity across quantile levels for these issues,so as to better approximate and reveal the relationship between the economic and financial variables at different environments.In practice,this may provide technical support and basis for grasping the true operating rules of the financial economic system.Firstly,the extreme expectile model is constructed to measure the extreme risks of financial asset variables,called as Expectile-based VaR.Different form the quantile-based VaR,the expectile-basd VaR is sensitive to the magnitude of extreme loss.In the empirical analysis,the Expectile-based marginal expected shortfall is used to measure the systematic risks of listed commercial banks in China,and then to examine the systematic importance.The research results show that: system importance ranking of banks is related to the intensity of systemic risk destruction;the greater the destructive power,the higher the systemic importance ranking of state-owned banks is;on the contrary,when the destructive power is relatively small,the systemic importance of large-scale state-owned banks is not evident.Secondly,the concept of cointegration is introduced into the framework of quantile theroty,and qiantile cointegration model is constructed to excavate the horizontal information of non-stationary variables,wich cannot be implemented by the classical quantile regression model.In the empirical analysis,we examine the co-movement between the Shanghai Composite Index and the world's major stock market indices,such as the Nikkei 225 Index,the Hang Seng Index,the S&P 500 Index,and the Stoxx Europe 600 Index.The research results show that the long-term relationship between China's stock market and other major stock markets in the world is heterogenous at different quantile levels.When the price level of China's stock market is at a raletively low level,the rising tendency in world stock markets will pull up,to some extent,the China's stocks.However,when the price level of the Chinese stock market is at an extremely high level,the world's stock market will have minimal impact on the Chinese stock market;in other words,the Chinese stock market is almost split across the world's major stock market.Thirdly,the concept of quantile correlation and cross-quantilogram is applied to examine the heterogeneity of stock price synchronicity.The quantile-based synchronocity is able to provide information about the market condition of aggregate stock market and individual firm,which cannot be characterized by the mean-regression model.For the listed firms in China,the synchronicity has relatively higher levels at the tailed quantiles and lower levels at the center of distribution.The U-shaped character indicates that the investors are sensitive to the “panic” or “frenetic” sentiment in the stock market.Comparing the levels of synchronicity at the tailed quantiles,it can be found that the synchronocity has higher levels at the left-tailed quantiles than those at the right-tailed quantiles.This reflects the fact that the “panic” sentiment caused by extreme loss events has greater impact than the “fanatic” sentiment caused by extreme income events.Fanally,this paper investigates the extreme risk spillover of international crude oil to stock returns for 529 firms listed on the A-share market of the Shanghai stock exchange.We apply a kernel-based nonparametric method to test quantile-on-quantile Granger causality from crude oil to firm returns.From the perspective of firm-level analysis,the findings are outlined as follows: The empirical results provide strong evidence to suggest the asymmetry in the linkage of extreme movements from crude oil to firm returns,that is,the case of positive risk spillovers are more severe than the case of negative risk spillovers;Risk transmission from oil price shocks to firm returns depends on the firm's industry features.In addition to the usual focus on inter-industry difference,we also note the within-industry heterogeneity;We confirm that China's refined oil pricing reform of March 27,2013 has had intensifying effects on the negative spillovers from oil prices to firms.