Research On Measurement Of Interest Rate Risk Of Chinese Commercial Banks Based On VaR Model

Due to the deepening of the liberalization of China's market,the frequency of interest rate fluctuations has been strengthened.In particular,the volatility changes during the period from 2010 to 2013 are particularly fierce.At the same time,commercial banks also need to respond to some increasingly severe challenges to interest rate fluctuations.For the selection of risk measurement methods,more traditional static research methods are used,which do not apply to the analysis of the timeliness of interest rate risk measurement.The VaR method can dynamically measure the interest rate risk in a concise and intuitive way.It is a common international dynamic risk measurement method.However,the domestic research results on the VaR method are not yet mature,especially the related research on the characteristics of China's interest rate risk fluctuation is relatively lacking.Through the comprehensive comparison and judgment of duration analysis method,sensitivity analysis method and VaR method,this paper finally selects VaR method as the methodological guidance of this article.Taking into account the Shibor interest rate of higher degree of marketization,so choose Shibor O/N yield data from April 2007 to April 2017 as the basic data of this empirical study,establish 36 kinds of GARCH family model,select the model effect Good AR(1)-GARCH(2,1)-N model,AR(1)-GARCH(1,1)-G model,AR(1)-TGARCH(1,1)-N model,AR(1))-TGARCH(1,2)-G model,AR(1)-EGARCH(1,2)-N model establishes the conditional variance equation of the Shibor yield series { } calculates the VaR predicted value and performs an accuracy check.Since the { } sequence has a periodic fluctuation feature,the dummy variable Dummy^' is introduced and the segment AR(1)-GARCH(1,1)-G model is selected,and the segment AR(1)-TGARCH(1,2)-G model and the segmented AR(1)-EGARCH(1,2)-N model establish a conditional variance equation for the Shibor yield series{ }to verify the accuracy of the VaR prediction value.The test results show that for the sample data with phased fluctuation characteristics,the segmented GARCH family model can be given priority after the introduction of dummy variables.The segmented GARCH family model can effectively compensate for the sample sensitivity shortcomings of the traditional GARCH family model.Improve the accuracy of VaR forecasting risk.Based on the characteristics of interest rate fluctuations in China,commercial banks can: on the one hand,learn more advanced VaR models by learning advanced measurement methods;on the other hand,they adopt measures to innovate interest rate derivatives,choose the level of confidence that suits their circumstances,and the duration of holding periods.It will not only help improve the accuracy of risk measurement,but also help strengthen the ability to respond to extreme risks.