Determining the Number of Factors in Affine Term Structure Models

• Activity Number: 683
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Tao Zou* and Song Xi Chen
• Companies: Peking University and Peking University/Iowa State Univeristy
• Keywords: Affine Term Structure ; Bond Prices ; Penalty Criterion ; Selection Consistency

Abstract

Affine term structure models are a wide range of interest rate models and have been used to describe the dynamic of bond prices in finance. We consider determining the number of the unobservable state variables in the affine term structure models, which is an unresolved issue in the literatures of multi-factor modeling. We propose some penalty criteria and show that the number of factors can be consistently estimated. It is also found that our proposed method is very helpful to find whether the observed bond prices do or do not have pricing errors, which is required in the affine model estimation. Simulation experiments are conducted to confirm the theoretical properties of the selection consistency. We also analyze the affine term structure models employing the data of U.S. Treasury bond prices.