The Estimation Of Term Structure In Chinese Interbank Treasury Bond Market Based On Bayesian Method And Nelson-siegel Model

Based on the outcome of maximum likelihood estimation,we propose a Bayesian estimation method for Nelson-Siegel model.We also deduce the corresponding sampling method for each parameter in the model under different prior densities.We estimate the term structure of Chinese interbank Treasury bonds between the year 2014 and 2016.With respect to the pricing efficiency of Treasury bonds,we find that the term structured estimated by Nelson-Siegel model produces smaller pricing errors than the term structure published by China Securities Depository and Clearing Co.Ltd.Compared to the maximum likelihood estimation,the Bayesian method yields more robust results when the sample is limited or distributed unevenly in term.For both maximum likelihood estimation and Bayesian estimation,the parameters in Nelson-Siegel model,especially the shape parameter,should be carefully chosen.For future application,the Bayesian method makes it possible to update term structure trade by trade and provides more information about pricing Treasury bonds.