Backtesting And The Application Of The Risk Value Model

The ongoing financial crisis has reminded us of the importance of risk management in the modern financial world.Risk management should be an integral part of all financial decisions.Value at Risk(Va R)is one of the Risk measurement techniques.Scholars at home and abroad have proposed different risk measurement models to estimate the VaR,but the estimation of VaR depends on the selection of the model and the distribution of the random variables.The VaR of different internal models is often very different,and how to evaluate the accuracy of VaR estimation is particularly important.Only the model that can accurately predict the future is effective.To evaluate the accuracy of the VaR needs to use appropriate methods to test the model and the measure test(Backtesting)is a statistical test method,it can compare the actual profit and loss and the corresponding VaR estimates are consistent.In this paper,we reviews the methods of regression testing in the past few decades,and summarizes the methods for the retest of the regression test.Given the efficacy is a back inspection measurement model in evaluation of important indicators to improve the efficacy of back-test.For the purpose of this article is based on the theory of the likelihood ratio test,this paper proposes a new inspection method,the average number of failure for the first time test and the average failure rate test,and study the overall obey Pascal distribution,geometric distribution and binomial distribution situation of VaR back-test test questions,makes Kupiec [1] proposed back-test test is just a special case of the research conclusion in this paper.Then,we verify the theory by numerical calculation,and the results of numerical calculation show that the second type of error is significantly reduced,that is,the efficiency is improved.Finally,the application of high-frequency data of Shanghai 50 index in 2016 5 min to do empirical research on China's financial market risk,based on the implemented the minimum volatility(Minimum Realized volatility,MinRV)get estimates of VaR model,a new testing method was applied to VaR forecast,and compared with the traditional Kupiec test method.The result of empirical analysis is concluded that the VaR estimation model based on the likelihood ratio test method is more effective and can evaluate the risk value model more accurately.The research theory of this paper can provide useful help for us to gain an in-depth understanding of the current situation of China's financial market and effective market risk management.