Fitting And Forecasting In The Term Structure Of Interest Rates Of National Debt Based On The Nelson-Siegel Class Models

Recent years, National debt of China has increased year by year, the status of thenational debt in the financial market has gradually come out.With the development ofeconomy, the advancement of marketization of interest rate, the importance ofnational debt interest rate term structure research has revealed. As we all know thatterm structure of interest rate of national bond is the key of the design of financialproduct, the research of effectiveness and conduction mechanism of monetary policythe pricing of financial assets and the risk management of financial market, Manyscholars and policy makers are looking forward to find a suitable way of fitting andforecasting China’s national debt term structure of interest rates, most scholars believethat Nelson Siegel model and its extended model (collectively referred to as NSmodel, including NS, DNS, SV, ASV model) is more suitable for China’s nationaldebt term structure of interest rates, but during the empirical tests, maybe because ofthe method of selecting discomfort, or the sample size is too small or there is nocomparative study of model system, We has not gotten the unified conclusion. In viewof this, this paper will be based on the Nelson-Siegel model for China’s national debtterm structure of interest rates to fit and predict, we expect that we can find the mostsuitable model for fitting and predicting the term structure of interest rates of China’snational debt.This paper is based on the background that the importance of the national debtterm structure of interest rates is rising, first,we studied the traditional theory of termstructure of interest rates and the modern theory, finding that the Nelson-Siegel classmodels are more suitable for studying Chinese national debt term structure of interestrates. Then, based on the Nelson-Siegel class models, separately on the Shanghaistock exchange market and the inter-bank market, fitting national debt term structureof interest rates using the monthly data of2005-2010, Using mean square error andabsolute error to compare four dynamic model fitting effect, Then, using produced inthe process of fitting the model of dynamic parameters. Taking first-orderautoregressive (AR)(1) which is thinked as having good prediction effect by scholarsto get the parameters of prediction of2011-2013, We compare real parameter value and predictive value, Through the parameter deviation rate comparing the modelprediction effect. The empirical results show that during the four kinds of models, thebest model of fitting and forecasting our country national debt term structure ofinterest rates is Nelson Siegel model, This explains that why many countries are usingNelson Siegel model for its monetary policy. Finally, based on the empirical testresults, This article put forward the relevant policy recommendations to the Chinesenational bond market system construction from three aspects, I hope this can helpfurther improve the bond market, and laying a foundation for the study of nationaldebt term structure of interest rates in the future.