Estimating The Term Structure Of Interest Rate

Firstly, this paper reviews various methodologies used to estimate the term structure of interest rates from data of government securities, including polynomial splines model (McCulloch), exponential splines model (Vasicek), B-splines model (Shea, Steeley) and exponential function model (Nelson, Siegel). By comparing these models theoretically, the paper discusses the distinctions in coping with transaction cost problems and robustness problems among all the four models.Because the domestic bond market has properties of small scale, inactive transactions and various tax-rates according to different traders, it is easily disturbed by the inaccurately priced bonds while estimating the term structure of interest rate of domestic market. This paper introduces the methodology of quantile regression into a well-chosen model and gets a robust estimation. Besides, the paper utilizes the advantages of quantile regression and finds out some probably inaccurately priced bonds.Although empirical research of the term structure has came into maturity, new techniques are springing continually. Upon the comparison, the paper suggests that, while doing similar researches on the domestic market, the differences between the domestic market and the mature markets overseas should be taken into account, and simply tracing new techniques cannot solve the characteristic problems of domestic market.