Estimating and Forecasting the Term Structure of Korea Markets Using the Nelson-Siegel Model

Abstract

This paper estimates and forecasts yield curve of korea bond market using a three factor term structure model based on the Nelson-Siegel model. The Nelson-Siegel model is in-terpreted as a model of level, slope and curvature and has the flexibility required to match the changing shape of the yield curve. To estimate this model, we use the two-step estima-tion procedure as in Diebold and Li. Estimation results show our model is Quite flexible and gives a very good fit to data.

To see the forecasting ability of our model, we compare the RMSEs (root mean square error) of our model to random walk (RW) model and principal component model for out-of sample period as well as in-sample period. we find that our model has better forecasting performances over principal component model but shows slight edge over RW model especially for long run forecasting period. Considering that it is difficult for any model to show better forecasting ability over the RW model in out-of-sample period, results suggest that our model is useful for practitioners to forecast yields curve dynamics.

Citation

Lee, J.H. (2004), "Estimating and Forecasting the Term Structure of Korea Markets Using the Nelson-Siegel Model", Journal of Derivatives and Quantitative Studies: 선물연구, Vol. 12 No. 2, pp. 101-126. https://doi.org/10.1108/JDQS-02-2004-B0005