Study Of Dynamic Value At Risk By EVT

In recent years, more and more people recognize the importance of the risk offinancial assets, by the financial crisis caused by subprime lending crisis that leadingto wide fluctuations of every country. In research of financial crisis, researcher andinvestor of the focus of attention is that estimating and forecasting risk of loss underextreme case. In1994, JP Morgan bank propose a risk measurement method by value atrisk that is widely used in modern times’ financial risk measurement.At present, there are two mainly aspects about the research of VaR: traditional riskmeasurement ways often suppose the series of profit and loss as normal distribution,however, the actual financial data are not normal distribution fat-tailed but leptokurtosisand fattaill, on the one hand,making the accurate and appropriate model by actualfinancial data on the other one hand. Considering above two aspects, the paper describethe series of loss’ tail by using the POT model of EVT and describe the volatility ofseries of loss’ tail using FIAPARCH model. The paper measure VaR by combining thePOT model with the FIAPARCH model and do empirical analysis with SHI. The resultshow that it is accurate and more conserved of measuring VaR based onPOT-FIAPARCH model. At the same time, considering the volatility of between assertsand spreading risk, the paper put forward based on dynamic condition correlationmultivariant FIGARCH and FIAPARCH models which are used for measuringportfolio’s VaR. At last the paper make use of SHI and SZI for empirical analysis, theresults show that DCC-FIAPARCH model is much more accurate and effective.