Calculation Of VaR Of Garch Models And Its Empirical Study In Financial Markets

With the rapid development of national economy, more and more financial derivatives andfinancial tools show up followed by the higher requirements of risk research in financial market.Risk measurement is the key link in the risk management with its method being constantlyimproved and innovated. VaR (Value at Risk) method, one of the effective methods to measurethe financial risk, is widely used in domestic and foreign financial institutions. VaR method tosome extent compensates for the shortage of many risk measurement methods.Begin with the theoretical framework of financial market risk, this paper introduces the VaRrisk measurement methods in detail, summarizes the basic principle, characteristic andcalculation steps of VaR method, describes three main calculation methods, and analyzes theevaluation of VaR method. In the latter part, this paper introduces the GARCH model, togetherwith its development model and the method of processing back peak sequence in the model.In the empirical analysis, the benchmark Shanghai composite index, logarithm yieldsequence is chosen as the research object. According to the sequence of yield under The normaldistribution assumption, t-distribution assumption and GED distribution assumption, itrespectively establishes a GARCH (1,1), EGARCH (1,1) and PARCH (1,1) model. Aftercomparison among the models, ideal results are obtained. Putting the GARCH (1,1), EGARCH(1,1) and PARCH (1,1) model under the hypothesis of GED distribution, calculating of VaRvalue under different confidence level respectively and using the Kupiec failure test to test modelcome to a conclusion, that is GARCH (1,1) model under the hypothesis of GED distribution isthe optimal model of the Shanghai composite index return risk.