Interest Rate Risk Measurement Study Under The Condition Of Interest Rate Liberalization

Interest rate risk is one of the most important risk among the financial market risks. Since 1996. the real start of China's liberalization reform of interest rate, it has gone through fourteen years of the reform process. With financial liberalization. especially the development of the liberalization of interest rate, the interest rate reform in China has entered the most critical aspect and overall interest rate liberalization is just around the corner. At this point, the most accurate measure of interest rate risk is urgent. So to study the measure of interest rate risk, to urge financial institutions to establish a sound interest rate risk management system has important theoretical significance and practical significance.The main method of interest rate risk measurement used in this article is a combination of Copula function and variance-covariance method for step forward VaR calculation.Early interest rate risk measurement methods are:sensitivity gap analysis, duration gap analysis, option-adjusted spread (OAS) and so on. With the deepening of the marketization of interest rate. the traditional measurement methods starting from the balance sheet items have appeared to be inadequate:and risk-oriented measurement method "VaR" has significant advantages. Three traditional measures methods of VaR are:variance-covariance method, historical simulation, Monte Carlo simulation. The core of the variance-covariance method is the calculation of variance-covariance matrix. The traditional variance-covariance matrix is calculated based on the linear relationship between series. In addition, there is an assumption of this method is:return series follow a normal distribution. Large empirical studies have shown that these do not match the real of financial variables. Because the Copula function without any assumptions to the distribution can describe the non-linear and asymmetric relationships well, so this paper do empirical research by Copula functions and variance-covariance method. First I estimated the correlation coefficient with Copula functions. Secondly with the variance predicted by the mean equation of each sequence and you can get variance-covariance matrix. Thirdly calculate the step forward VaR.The main conclusions of this paper is:when you forecast the step forward VaR of the repo rate market, you should combine variance-covariance method and Copula function, this method is more efficient and more accurate than the traditional variance-covariance method, and posteriori test results support this conclusion too. and thus it has more practical significance to commercial banks at measurement and management of interest rate risk.