The Calculation Of VaR And Its Applications In Risk Management

VaR technique as a new risk measurement method, compared to the traditional riskmeasurement methods, such as scene simulation method, pressure test method, the method ofsensitivity, because of its method of intuitive, quantify, and easy to understand characteristics,get the favour of risk managers, also get many scholars’ research.This paper first introduced the basic theory of the VaR and some traditional calculationmethods.Then, choose the SSE180Index, Shenzhen Component Index, Hong Kong HangSeng Index as the research object. We go on some basic statistical analysis about this indexreturn series, indicate that the index return series of these markets are Leptokurtic and fattailed,with the effect of ARCH. So that it is appropriate to use the GARCH model on thesethree markets.Next, using GARCH model and APARCH model with the normal innovations,Student-t innovations, GED innovations, skewness t innovations and skewness GEDinnovations to calculate the value of the VaR of each index. Also we made preliminaryanalysis of the model and results. At the same time, we use MCMC methods to estimate theGARCH model parameters, then compare it with the traditional maximum likelihoodestimation methods.At the end of this paper, we try to combine multiple back testing methods together whichcan evaluate a VaR model from three aspects-accuracy, conservative and effectiveness. Themain conclusions are:(1) Because through the three security markets data of return thecalculating APARCH model parameter1are non-zero, and it’s positive, indicate that thereare marked "leverage effect" in Shanghai, Shenzhen and Hong Kong security markets.(2) UseMCMC methods to estimate GARCH model parameters, calculate the VaR of SSE180Index and the Hang Seng Index under different confidence levels, show that the effectivenessof the GARCH-N-MCMC model is more than the other two models, In the model accuracyfrom that premise, the use of this method, can let investors has a minimum reserve ofopportunity cost.(3) the comparison of the APARCH and GARCH model: When theconfidence level is high, on the one hand, at the side of accuracy and effectiveness, APARCHmodel has increased than GARCH model, but at the same time APARCH model is also relatively conservative. On the other hand, Based on the generalized error of skewnessdistribution APARCH model can capture more of the financial market of all kinds ofcharacteristics, such as volatility clustering, leptokurtic and fat–tailed, etc. And APARCHmodel is more effectiveness than GARCH model. At this time we can use APARCH modelwith skewness GED innovations to calculate the value of the VaR of these three markets, theeffect is better. In this paper, the calculating method of the VaR made some attempt to lookforward to provide a reference to risk managers.