Parameter Estimation Of Double Factor Interest Rate Models

Interest rate is a important content in economic research.It plays a certain role in the pricing of financial assets,financial risk management and currency appreciation.Therefore,the study of interest rate is important.With the opening of domestic financial market especially after joining WTO,interest rate has played a more and more significant role in the national macroeconomic regulation and control.The parameter estimation of the interest rate model has always been a basic and important problem in the financial field,which is of great research value.The term structure model of interest rate is divided into static model and dynamic model.Because interest rate in the financial market is closer to the dynamic model,the main purpose of this paper is to study the dynamic interest rate term structure model.This paper first introduces the single factor interest rate model,then introduces the related contents of the double factor Vasicek model.In this paper,the method of estimating dynamic interest rate model is Kalman filter.We need to use spot interest rate data in Kalman filtering method,which is generated in the static estimation model.Before introducing the Kalman filtering method,we introduce the Nelson-Siegel model and Nelson-Siegel-Svensson model for the estimation method of static interest rate.Because the Nelson-Siegel-Svensson model is suitable for most of the bond markets,this paper uses the Nelson-Siegel-Svensson model to get the spot interest rate.After introducing the static estimation method of the term structure model of the interest rate,the relevant contents of the Kalman filtering method are introduced.By adopting the national debt data of Shanghai Stock Exchange,this paper uses Nelson-Siegel-Svensson model to generate spot interest rate data,and uses Kalman filtering method to get the estimated parameters of double factor Vasicek model.Comparing the estimated data with the actual data,it is concluded that the data of the double factor Vasicek model generated by the Kalman filtering method is close to the actual data as a whole.