Interest rates factor model

Abstract

Interdependence of the interest rates of the US, the UK, and Japan is analyzed in this work by means of spectral analysis and network methods. A predominant effective factor in the interest rate market is which country floats a bond issue, and a minor effective factor is time to maturity of bonds. Power-law cross-correlation among different countries is analyzed by the detrended cross-correlation analysis method. Long-range cross-correlation is found between the first factors of interest rate, while there is no cross-correlation between some of the second factors. The tail dependency is indicated by tail indices from Archimedean copulas, including an empirical copula. In contrast to other pairs, the US-UK first factor pair has tail dependencies in both the upper-tail and lower-tail. Dynamic properties of interest rate are modeled by a stochastic volatility model. The properties of mean reverting and volatility clustering are observed and reflected in this model. The proposed simulation method combines the dependence structures and the factor dynamics model; it simultaneously describes the interest rates of different countries. (C) 2011 Elsevier B.V. All rights reserved.