Forecasting the U.S. Term Structure of Interest Rates using Nonparametric Functional Data Analysis
21 Pages Posted: 7 Jun 2012 Last revised: 3 Jul 2016
Date Written: April 7, 2013
In this paper we consider a novel procedure for forecasting the US yield curve by using the methodology of nonparametric kernel estimation of functional data (NP-FDA). Within this approach, each element of the sample is a monthly yield curve, evaluated at points corresponding to maturities. In this framework we attempt to capture the dynamics present in the sample of curves to forecast future values for the yield at a given maturity without imposing any parametric structure. In order to evaluated forecast performance of the proposed estimator, we consider four forecast horizons and the results are compared with widely known parametric models. Our estimates with NP-FDA present predictive performance superior to its competitors in many situations considered, especially at longer time horizons for long-term maturities. The methodol- ogy applied in this paper may be important for policy makers, fixed income portfolio managers, financial institutions and academics as it may prove useful in the construction of long-term scenarios for the yield curve.
Keywords: Term structure estimation, factor models, nonparametric method, Interest rate forecasting, Kalman filter
JEL Classification: C53, E43, G17
Suggested Citation: Suggested Citation