Research On China Securities Market Based On VaR Model Via Quantile Regression

China's reform and opening up has gone through 30 years, so does China's capital market. With the development of financial market, the market risk becomes more complicated.The subprime mortgage crisis that occurred between 2007 and 2009,not only makes the US economy suffer a lot,but also makes the global economy fall into recession.After joining the WTO,the contact between countries get closer.In 2015,the total direct foreign investment of China is $126,27 billion,while the total overseas investment is $118,02 billion.Therefore,we need to focus more on risk monitoring and management.The developments of financial regulation often adapt to the development of economic development.Let us look at the past hundred years of economic development,from strict controls to financial market liberalization and the current prudential supervision,every step is accompanied by innovation and the development of financial products,but those products will bring huge financial risks too.This brings enormous challenges to the management of financial risks.With the continuous development of the financial market,the measurement requirements for financial risk management is raising too.VaR model is the most important risk measurement model and widely used.Therefore,to expand and make improvement to the VaR model is very important to strengthen and improve the management of financial market risk.VaR is a quantile in nature,but the calculation methods of VaR varies.However,many methods focus on the distribution of financial assets yield or density function estimation calculations,such as Risk Metrics model,variance-covariance method and so on.This paper introduces the quantile regression method to calculate VaR value,because quantile regression models can do directly quantile regression without considering the distribution form of financial data,and can describe extreme cases.Therefore,quantile regression can be used to China's capital market,which have a character of fat tail and volatility aggregation. At last, we make a compare between different models in order to show the advantages and limitations of VaR.This paper is divided into five parts as follows:Chapter ?:Introduction.This chapter describes the background of this article, and shows the increasing importance of financial risk management.and then illustrate the research method and ideas of this paper,as well as the innovations and shortcomings.Chapter ?: literature review.The chapter introduces the main views of current studies both domestic and aboard according to two main lines of VaR and quantile regression.Therefore to construct the model used in this article.Chapter ?:The introduction of basic theories.The chapter illustrates financial market risks,VaR and quantile regression theories respectively.Through this process, we got the key point of VaR model's measurement of financial risk and the advantages and disadvantages of each model.Based on this idea,we build a quantile regression model of this paper,this model makes a reference to a linear model proposed by Chen (2002),and simplified the model through daily held,and combining the estimate of volatility by GARCH(1,1),ARCH(1)andEGARCH(1,1) model.Chapter ?:Empirical research.First,this article make a descriptive statistical analysis of four stock market indexes,founding that China's securities market have a character of fat tail,biased and volatility clustering.Then,this paper make an stationary test and autocorrelation analysis to the time sequence data of four indexes,founding that the logarithmic rate of return is steady without autocorrelation.Later,we calculate VaR value of the four indices under at 99% and 95% confidence level.By using historical simulation,this paper make a compare between models via Copula function model,which is used to calculate the joint yield of Shanghai Composite Index and Shenzhen Composite Index.By comparing LR statistics,we find that the adaptability of the model to different securities markets is not the same.For example,in the SME composite index is more suitable for QR-ARCH(1),while the GEM index is more applicable to QR-EGARCH(1,1).Moreover,quantile regression method and historical simulation are more efficient than Gumbel Copula function.This shows that VaR measurement via quantile regression is still effective for China's securities market.Chapter V:Summary and Prospects.This chapter summarizes the empirical results of former chapter and analyze the advantages and disadvantages of the application of VaR model.The study shows that VaR model can be well adapted to Chinese market without making an assumption about the distribution of rate of return.By the way, it can accurately and effectively measure risks under extreme situations.At last, this paper discuss the future research prospects of quantile regression methods.In a word,with Comprehensive Risk Management becomes a trend,the research related to this will become more deeply.By the analyses of the value at risk of China's securities market via quantile regression,this paper found that there is a difference between the plates and the applicability of the model to different plates is not the same.So,this model should combine with other models to enhance the accuracy and effectiveness of the VaR measure.