Research Of The Tail Risk On The Rate Of Stock Index Return In Chinese Security Markets

For insurance companies, banks and other financial institutions, risk management is an important subject, risk measure model is very important. in order to accurately measure the risk,It is necessary to ensure that fitting distribution hypothesis selection has better priority degree, this article attempts to modify the financial rate of return distribution hypothesis by improving risk measurement precision. In the past academic research history, traditional financial development is based on normal distribution theory to study the return on assets, many experts and scholars discovery that the actual return on assets does not meet the normal distribution, the development, based on the normal distribution heory,proposed in view of the actual assets yield stable distribution the peak, thick tail characteristics. However, according to the actual data proof, stable distribution has thicker tail. Then, scholars put forward by the normal mean--variance and generalized inverse Gauss distribution for the theoretical foundation of the fitting generalized hyperbolic distribution real return on assets, the distribution of the tail than the normal distribution to be thick, than the stable distribution to be thin, can better fit the assets yield peak, thick tail and skewness features, in recent years in foreign financial field has obtained the rapid development, but in the domestic research about the financial sector is still rare, under this background, this paper bases on the domestic and foreign advanced achievements, the generalized hyperbolic distribution is applied to the empirical research on the field of risk measurement China stock market, with the Shanghai Composite Index, Shenzhen as a sample, the generalized hyperbolic distributions in fitting the distribution of asset returns the goodness of fit of empirical study.The first chapter introduces the meaning of choosing this topic, including the theory of generalized hyperbolic and risk measurement and the practical significance, the starting point of this article and the value. Empirical results and previous research, the domestic and foreign research scholar of generalized hyperbolic distribution of literature summary and the return distribution research summary. Then put forward the desired depth, the difficulties and we hope to highlight the innovation and the article structure arrangement.The development process in the second chapter, distribution fitting to start, respectively introduces the distribution fitting mainstream yields, including the assumption of normal distribution and stable distribution hypothesis, to sort out the development process. Then focus on the introduction of generalized hyperbolic distributions of the history, background, theoretical basis. On this basis, the five parameters generalized hyperbolic distribution common representation methods; parameter representing the meaning, academic circles to parameter most commonly used representation, distribution and the limit distribution of the generalized hyperbolic distribution. Then, we need to know that how to estimate the parameters of the generalized hyperbolic distribution. The parameter estimation process is complex, the function more, the estimation process is attached to the appendix.The third chapter mainly analysis’s the data character of Shanghai stock index and Shenzhen index. In this paper, firstly, the logarithmic transformation of the data, using the logarithmic rate of return, then the test of normality of data (kurtosis, skewness, and QQ plot), stationary test, and autocorrelation test, reflect the characteristics of returns with the actual data. Then, using the R software, calculate five kinds of distributions, including generalized hyperbolic distribution, parameters of Shanghai index and Shenzhen index value, at the same time, the calculated number of iteration, the likelihood function value, AIC value, according to the numerical size, judge the fitting goodness of fit comparison.The fourth chapter mainly carries on the tail risk measurement, the sample data of Shenzhen and Shanghai stock index is extended to2014April, calculated in the generalized hyperbolic distribution risk value of group VaR sub distribution and the ultimate distribution assumption. And the method of inspection, the failure frequency calculation results of VaR test. Then calculate the tail expect ES, the calculation results are compared with Value at Risk, evaluation of the two risk measure index.The fifth chapter, summay and evaluation, summarizes the logical process, combing of the text of all conclusions in the empirical process, suggestions.