The Measurement Of Commercial Bank Interest Rate Risk Based On The Interest Rate Market

China’s interest rate market reform began in the mid-1990s, a recent breakthrough reform measure is on July 20.2013, all of the financial institutions lending rates were let of, it marks that China’s interest rate market reform has entered a substantive stage. After the marketization of interest rate, it will intensify the competition between commercial banks and increase the influence on operating and management affected by interest rate fluctuations. Then, the interest rate risk will become one of the main market risk faced by commercial banks, therefore, commercial banks should pay attention to identify and prevent the interest rate risk. However, identify and guard against interest rate risk, the key is to effectively measure the interest rate risk, this paper is to explore the interest rate risk measurement based on the interest rate market background, makes the commercial bank can be based on the method to predict the interest rate risk, and take appropriate measures to prevent risks.So this paper builds two different types of interest rate risk measure models, through the comparative, choose the better one as this paper’s interest rate risk measure model.The first model is APGARCH-GED-MC nonparametric VaR model which is based on building different GARCH class parameter and nonparametric VaR models. By comparing the random variables respectively to obey normal, t and GED distribution GARCH(1,1), EGARCH(1,1) and APGARCH(1,1) 9 models, select the most optimal APGARCH-GED model to predict the conditional variance which was used as a standard in Monte Carlo simulation. By Kupiec failure rate test, APGARCH-GED-MC VaR model calculate the interest rate risk of VaR value, compared with the actual logarithm yield, the difference is small, failure rate is low,while compare to the parametric and nonparametric methods.The second model is integrated consider the VaR definition (VaR is a quantile under a certain confidence level) and in building the first kind of model to find that which distribution form the yields obeys has a larger impact on the calculation of VaR, then, from the perspective of quantile to build a model of VaR to measure interest rate risk. This paper builds two kinds of quantile regression models:non-recursive quantile regression and quantile regression of recursion. The non-recursive quantile regression VaR model is QR.APGARCH-GED VaR, this model’s interpretation variables are made up by the APGARCH (1,1)-GED model’s forecast. The recursive quantile regression VaR model is the asymmetry absolute and slope of conditional quantile auto regression model (AAVS-CAViaR), this model considers the VaR’s relevance and yields of asymmetry. Also use the Kupiec failure test to compare the two VaR quantile regression models, it is concluded that the quantile regression QR-APGARCH-GED VaR model, not only coverage the real yields better, but also reduce the interest rate risk VaR values predicted failure rate.To sum up, this paper builds the interest rate risk VaR models from two different aspects. At first, through the Kupiec inspection finding that based on the APGARCH-GED-MC VaR model to forecast the interest rate risk VaR, the failure rates is lower than parametric and nonparametric VaR models, so choose the APGARCH-GED-MC VaR model as the first model in this paper. Second, in the procedure of building the quantile regression VaR model, based on the QR. APGARCH-GED VaR model to forcast the interest rate risk VaR is better covering the real yields and the failure rates is lower than AAVS-CAViaR model, so choose the QR. APGARCH-GED VaR model as the second model of this paper. Finally, by comparing the failure days of APGARCH-GED-MC VaR model and QR. APGARCH-GED VaR model, found that the quantile regression method greatly reduces the failure days when forecast the VaR value, also using the QR. APGARCH-GED VaR model to calculate VaR has good range coverage, it can also be volatile with the actual yield curve swings in the strange time. Based on this, this paper finally chooses QR. APGARCH-GED model as the commercial bank interest rate risk measurement model under current level of interest rate liberalization.This model not only provide a theoretical basis for the commercial banks’measurement and management, but also provide a reference for the higher lever of interest rate marketization.