VaR Model In Semiparametric And Parameter Methods

Since the beginning of the finance market, it is character is yields are kept both of risks. Under the ordinary condition, the bigger the risk taking, the bigger the yields requested by the investors. After entering 90's in 20th century, along with the turning worse of variation and system risk of finance markets, people give much more attention to the financial risk. The supreme mortgage crisis that is exploded to the United States in 2008 brings substantial concussions to the whole world's finance industry, this makes people start control and supervise the financial risk while using developing financial tool at in great quantities. In order to consume the yields as soon as to control the investment risks, the first thing we need to do is to valuate the financial risk correctly, considering there are so many kinds of financial risk, my paper begins with market risk and securities market that acquaint with most from people commence, the risk of securities market has very close relation with the wave motion of its price. Here are so many ways to value the financial market risk, but among them, the Value at Risk model is the most common estimate method. Value at Risk model is a new risk management tool which is developed at last century of 90's, and as a tool to measurement and control financial risk, Value at Risk model is easy to operate and wide application. So Value at Risk model is more worthy to use than traditional financial management patterns. The security market in China is developed more than ten years, we have obtained some certain results, but still existed a lot of antis specification compared with the mature securities market of developed countries, so the risk faced by investors is more strong, this needs more to select by examinations a compliant financial management to our country.This text chose representative stock index in the Shanghai and Shenzhen security market as a research objects to analyze our country's market risk in the securities market. At the first, there are some introductions about Value at Risk model, then give the highlighting information about VaR which is the main model I use in this paper. The contents are including its computing principle and computing method including parametric method, semi-parametric method and non- parametric method, and then compare these main computing methods to each other, in order to give the reasons to why to choose the parametric method and semi-parametric method to compute the Value at Risk in this text. On the bases of the speculative knowledge before, now to study the Shanghai stock index and Shenzhen stock index, then the profitability of them is tested through normality test, stability test, and GARCH-LM test. The outcomes of the tests indicate that GARCH models are suitable to be applied to calculating Value at Risk of our country. So we choose the GARCH,GARCH-M,TARCH and EGARCH models under the distributions of student's t and GED. Both the Semi-parametric method and parametric method's GARCH-VaR models need taking advantage of conditional standard deviations in the garch models. The statistics results to compute Value at Risk in quasi parameter law. At last Kupiec test is adopted to test the accuracy of the model to find suitable model for each securities market. From the comparisons, we can conclude that firstly the average value at risk is bigger in the Shenzhen stock market than Shanghai, so the investors in the Shenzhen stock market face more risk. The Garch-VaR models can reflect the time series'wave motion characteristic. Under the parametric, TARCH-t model is the best model. The best predictive abilities of the models are not equal to the best precise of the results. Under the Semi-parametric method, the accuracy that the four ranks is obviously better than the parameter law on the garch pattern. The quasi parameter law that four ranks just give the value zone that the risk is worth of, have never given an accurate value, it gains everyday a possible fall to describe the dimension of comparatively risk in the market; the parameter law computing of GARCH pattern of value with securities market rate of return future per diem probably the biggest point present of fall absolute risk dimension in the absolute risk. So we'd better to take both of the models into consideration during our investment.