| In the modern economic environment, finance is in a very important position. Commercial banks play a crucial role on the overall financial environment. How to control the risk through the perspective of commercial banks is not only the concern of commercial banks, but also the problems which regulatory authorities and the academia focus on. In the overall risk management and control of Chinese commercial bank, how to prevent and control interest rate risk is the most important problem we should pay attention to. Because of the advance of the wave in international interest rate risk control and the accelerated pace of China’s advance in the market-oriented interest rate, how to carry out effective measurement and risk management research on commercial banks, is the important tasks that Chinese commercial banks facing. Interest rate risk management in commercial banks is a mature and innovative topic, with the gradual relaxation of controls on financial markets in the international economic environment, there has came a large scale in financial product innovation, which makes the financial market become unstable. Financial institutions reform and change from the internal management and innovative ways to market resources exploration. Financial institutions management mode gradually changed, at the same time, the prevention and control of the interest rate risk measurement model theory and practice are constantly exploring and innovation. Under this background, VaR model developed is widely recognized, which has been recognized as the market risk measurement and management tools in the Western countries. Affected by the impact of the current world economic environment and Chinese financial market development, China pays more attention to the commercial interest rate risk measurement. However, as the interest rate market fluctuations become increasingly apparent, the risk quantifiable on commercial banks in China cannot meet the gradually changed situation in market risk, which still has a distancebetween developed countries. Therefore, this paper is based on China’s interbank market, using the overnight lending rate data to do an empirical analysis, hopes that there can bo some innovation in Chinese national conditions in the theory and practice of interest rate risk analysis and make some academic contributions to the theory of interest rate risk modeling, how to modeling in interest rate risk and so on.In the process of learning the course of the interest rate risk and reading some related paper documents, we have recognized that the interest rate refers to unexpected market changes in interest rate, which will lead to the commercial bank statement and balance-sheet unstable, which will lead a direct impact on the assets of commercial banks and revenue. In the method of interest rate risk measurement, there is an emerging model method which is held in esteem by the risk of regulatory authorities and financial institutions around the world, which we call it "VaR". VaR is known as value at risk, it is a risk measurement which use statistical method. Now, VaR has become financial risk management standards in many countries, and as an important tool for analyzing tool for supervision in financial institutions risk, the dynamic monitoring and quantifying characteristics are recognized and welcomed by the financial supervisory authorities and financial institutions.In this paper, chapter three has a basic understanding and explained the definition of commercial bank interest rate risk in a basic concept, which includes the understanding the basic concept of the interest rate risk and the types of the interest rate risk in commercial banks. Under the basis definition of interest rate risk, it has a further understanding on the types of the commercial bank interest rate risk. There are four interest rate risk consisted in it, including reprising risk, basis risk, yield curve risk and optionality risk. The second part described the course of the evolution in the commercial bank interest rate risk measurement methods. The correctly understanding and choosing the interest rate risk measurement method are both the important prerequisite for the scientific management of interest rate risk, so it is necessary to understand the evolution of the commercial banks’ interest rate risk measurement tools. Under the understanding of the two traditional interest rate risk measurement, namely the interest rate sensitivity gap analysis and the duration of the model, we can sum up that the advantages of static risk measure are easy to read, easy to operate and computing, however, the disadvantage in these two methods is that transactions cannot be comprehensive respond in the account. Compared the traditional methods, VaR can be more comprehensive in the measure of complex banking interest rate risk, therefore VaR has become the most popular measurement which is promoted in the majority of the risk regulators and financial institutions. In general, the VaR method can be divided into parametric method, the non-parametric method and the semi-parametric method. This part introduces two parameter method, one is the GARCH models, the other is the mixed normal distribution based on Risk Metrics model, and also introduce two non-parametric methods called historical simulation and Monte Carlo simulation. Under the understanding of those methods, this paper will choose GARCH models and mixed normal distribution based on the Risk Metrics to do the empirical analysis, the data will choose the Shanghai inter-bank borrowing market, using the overnight lending rate to the VaR analysis.In the fourth chapter, the main job is using the Shanghai inter-bank borrowing market interest rate data to do the statistical characteristics analysis. Because the SHIBOR is more fit with the laws of market volatility than CHIBOR, and the overnight lending rate data is the most frequent, so we choose the SHIBOR for the empirical analysis, and deal the yield data with logarithm. In the fifth chapter, the main job is using the GARCH models to do the interest rate risk measurement empirical analysis based on SHIBOR yield data. Under the basis understanding of GARCH models, the first step is to ensure the autoregressive moving average model in the SHIBOR sample sequence is the AR(1). The second step is to select three types of GARCH models, including GARCH, T-GARCH, E-GARCH model to fit the conditional heteroskedasticity in model. In the choice of the residual distribution, not only use the normal distribution form, but also select the T-distribution and generalized error distribution to fit the residual distribution. The third step is to choose18kinds of GARCH models and contrast with its AIC value. Finally, choose the EGARCH(1,2)-G for the calculation of the VaR value. The last part of the fifth chapter is to use the calculated VaR value to do the backtesting test. In the sixth chapter, using the other parametric method, which is the mixed normal distribution based on Risk Metrics to fit the logarithm SHIBOR yield data. Under the assumption that the logarithm SHIBOR yield data obeys a mixed distribution, make the yield attenuation factor of0.94, then use the Matlab software for EM iteration, finally get the parameter estimates of the mixed normal distribution density function.By comparing the results of these two methods, it can conclude that both the GARCH models and the mixed normal distribution based on Risk Metrics method can be used in calculate VaR. However, there are many differences between these two methods. In general, GARCH model performs better than mixed normal distribution method in the treatment of extreme risk, while the VaR calculated by mixed normal distribution model is closer to the average level of risk. Therefore, the mixed normal distribution model is more suitable in calculating the whole interest rate risk, it is another risk measurement that worth promotion and application. |
PDF Full Text Request