Money Market Interest Rate Forecasting Model

Purchase-back agreement rate of Chinese Treasury bonds on the inter-bank market is the most typical short-term interest rate in Chinese monetary market. Utilizing the data of the 7-day repo rates on the inter-bank treasury bond market, this paper analyses some statistical features of the rates. Based on the characters of non-negativity, autocorrelation and volatility clustering, ARIMA model and GARCH model are estimated respectively, and their predictive abilities are compared. The economic mechanism of these models shows that, the 7-day repo rate on the Chinese inter-bank treasury bond market has an obvious mean reverting effect, and with GARCH effect being introduced, it improves the fitness of the short-term rate volatility. The empirical result indicates that neither model is an exact dynamic model for the short-term rate in the Chinese bond market. However, ARIMA model performs better in predicting the trends of the short-term rates, on the other side, GARCH model will be a better utility to capture the interest rates volatility. The result of this study may not only help the financial institutions properly pricing their financial products and hedging interest rate risk, but also has implications in conducting the monetary policy destination setup by the central bank.