China Interest Rate Term Structure Prediction Methods

In the national debt market, the Term structure of interest rates of national debt played a benchmark function in varieties of portfolio management, financial asset pricing and risk management. Recently, domestic and overseas scholars had done a mass of research on it. But the forecasting methods with regard to the Term structure of interest rates of national debt have just started in China, so, this paper attempted to summarize the results of the study abroad based on the Term structure of interest rates of national debt and conducted a accurate, detailed study on forecasting methods.In this paper, first of all, we selected the SSE national debt transaction data, estimated the Term structure of interest rates of national debt in the use of NSS model, while estimated the time-series data of the model parameters. Then established three fitting model to forecast the future development of parameters sequences, which were ARMA (1,1) model, AR (1) model and the RW model, and compared the forecast effect on the three forecasting methods in the use of recursive algorithm, we obtained: ARMA (1,1) model was suitable for short-period forecast, in the medium and long-period forecast, ARMA (1,1) model was suitable for forecasting the short-term interest rates, AR (1) model was suitable for forecasting the medium and long-term interest rates. Then, in order to find out whether it improved the forecast effect with introducing macroeconomic variables, we selected three macroeconomic variables which were closely linked to the Term structure of interest rates of national debt, and established a regression model with the parameters of NSS model, as forecasting the interest rates of different term structure in the use of stepwise regression, besides, we compared the advantages and disadvantages with the time-series forecasting methods, finally we obtained: the simple time-series forecasting methods suited to the short and medium-period forecast, regression model showed a greater advantage in the long-period forecasts. Finally, monetary policy's sudden-changes of the central bank lead to the Term structure of interest rates of national debt's sudden-changes, accordingly resulted in that the use of the above-mentioned forecasting methods can not produce a better forecast effect, this paper presented a study idea that the use of event study to amend the forecast, which provide guidance recommendations on future further research the issue.