A Fractional Differencing Analysis of Yield Curves by Means of Wavelet Analysis

Esben Hoeg
Department of Information Science,The Aarhus School of Business
eh@hdc.hha.dk

Abstract

It is well known in finance that theoretical models of the term structure of interest rates predict shapes of the yield curve.

An important question is whether yields at the various maturities are I(d) processes with d < 1, or d=1. With d less than unity the corresponding process is mean-reverting, whereas this is not the case when d=1 . On the other hand as long as the process is non-stationary.

We use wavelet methods to draw inference on the differencing parameter d. Furthermore we compare with other non-wavelet methods.

The empirical validity is illustrated on a large set of weekly Danish Government Bond prices quoted on the Copenhagen Stock Exchange.