Fitting And Forecasting The Term Structure Of Government Bonds In Shanghai Stock Exchange

The term structure of interest rate is the basics of investments and pricings of financial tools. Despite powerful advances in the term structure of interest rate modeling in the last 20 years, the key practical problem of fitting and forecasting the term structure of interest rate has just started. Following the present researches, this thesis tries to do some systematic researches on fitting and forecasting the term structure of interest rate of government bonds in China, based on the Nelson-Siegel model.Firstly, we modify the Nelson-Siegel model by adding the government bond supply factor. About the value we take for the exponential decay rateλt, we write it as a time-varying step function. The empirical results show that we can use the estimated At value by Kalman filter as the exponential decay rate in the modified Nelson-Siegel model.Secondly, we use the modified Nelson-Siegel model to fit and forecast the term structure of interest rate of government bonds in China. The linear regression results show that the fitting error is small and stable. When forecasting the term structure of interest rate, we consider AR(1) model, ARMA(1,1) model and Random Walk model to forecast the four state variables of the modified Nelson-Siegel model and the results show that, the Random Walk model is inferior to AR(1) model and ARMA(1,1) model on both forecasting errors and stability. ARMA(1,1) model performs better in short-term forecast, while AR(1) model is superior in long-term forecast and is more stable.Finally, we point out the problems we find in our research and propose guidance recommendations in the term structure of interest rate of government bonds in China.