Calculation Methods Of The Value At Risk

In recent years, the impact of the economic and financial globalization, integration,gradually the modern financial theory and financial and information technologyinnovation and so on many factors, along with the rapid development of global financialmarkets at the same time coming and market showed unprecedented volatility, therefore,the financial risk caused the global industrial and commercial enterprises, financialinstitutions and other social and close attention in academic circles. In this context, therisk management ability of homeopathy also become one of the core competencemanagement. As a risk management tool VaR (Value at Risk) is an effective financialrisk measurement, it was primarily used to measure market risk, then gradually in theworld are widely used and valued, and gradually become the mainstream party riskmanagement of financial market risk measurement method and the large investment inthe field of international standard. So, to study how to compute VaR more accurately notonly in theory, have very important sense but also in real life.The traditional calculation method of risk value often assume that satisfy somesuch as the normal distribution for a class of elliptic distribution in theory, did not fullytake into account the financial assets are often appear "fat tail" phenomenon, so theeffect in the extreme to fit the tail of the distribution is poor, so that the influence of thefinancial crisis under extreme conditions such as risk predictive effect. Based on theabove situation, this paper puts forward a new calculation model of G-H distributionbased on VaR, and using the method of stochastic simulation to important parameters inthe model, by the end of the stock market data for empirical analysis, proved that themethod of fitting effect of VaR is better than historical simulation method and normaldistribution method, and the calculation steps and method above than the extreme valuetheory method currently used is more flexible and accurate.The main research work and achievements of this dissertation:①A full study of the basic theory of the calculation of value at risk VaR and relatedmodels, a review of the existing situation of relevant research at home and abroad, haslaid the foundation for comprehensive research method to calculate the risk value VaR.②Systematically summarizes the calculation of value at risk in VaR extreme valuefitting method, learning the international advanced technology to predict the extremecase processing in financial risks, laid the theoretical foundation for the study on calculation method of value at risk VaR on "fat tail" distribution.③Aiming at the risk value to calculate the VaR model does not fully account forfinancial assets very often appear in the "fat tail" so that the influence of extremeconditions risk prediction effect problem, through the study and analysis of the relatedmathematical model, establish a fit the historical data distribution with G-H distributionalgorithm.④In view of the traditional model of G-H distribution fitting method to make thefour moments are obtained simultaneously with the target distribution fitting theconsistency problem, through the study and analysis of parameters, the traditionalthought of parameter fitting algorithm, proposed a parameter fitting method based onrandom simulation, can do four important parameters to ensure model in the operationof simple, convenient premise approximation four moments of history data at the sametime, effectively improves the goodness of fit for the model, which improves theprediction accuracy of VaR.From a securities company and the Shanghai Composite Index closed a number ofsites selected a group of sample data, fitting simulation by using a new method. Theresults show: the new VaR calculation method is not only the confidence values arehigher in the case of fitting is better than the traditional VaR method, EVT method andthe operation is currently widely used more simple and flexible.