Modeling Term Structure Dynamics of Interest Rates

2006 Fiscal Year Final Research Report Summary

• Project/Area Number: 17510116
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Social systems engineering/Safety system
• Research Institution: The University of Tokyo
• Principal Investigator: FUJII Mariko The University of Tokyo, Research Center for Advanced Science and Technology, Professor, 先端科学技術研究センター, 教授 (90323550)
• Co-Investigator(Kenkyū-buntansha): OHTSUKA Kazumichi The University of Tokyo, Research Center for Advanced Science and Technology, Specially Appointed Research Associate, 先端科学技術研究センター, 特任教員(特任助手) (30361869)
• Project Period (FY): 2005 – 2006
• Keywords: term structure modeling / regime shift / Nelson-Siegel model / state-space model

Research Abstract

In this research project, we constructed a model of Japanese term structure interested rates, explaining from short to long in a consolidated manner. Based on this model, we analyze the relation between the macroeconomic variables and three dimensional changes of the term structure for the period of 1993 to 2003.
First we test the hypothesis of regime shift in describing the dynamic behavior of the term structure. By applying the models of both Markov switching and random level-shift of regime to the Japanese time series data, we find the statistical significance of regime changes in both specifications. These changes appears to correspond to changes in macroeconomic situations and monetary policy.
Based on these findings, our next step is to construct a model incorporating the macroeconomic variables explicitly. We fit the term structure with the exponential function proposed by Nelson and Siegel (1987). In this setting, yield curve is characterized by three factors ; the slope, and the curvature and dynamic behavior of the term structure are described as changes of these three factors. A set of macroeconomic variables such as inflation rate, industrial production, and stock prices was chosen to explain these changes of factors.
In order to construct a three-dimensional term structure model, we incorporate the macroeconomic variables into the framework of state-space time series model of Nelson-Siegel three factors. Parameters are estimated by the method of the extended Kalman filter. Obtained model fits relatively well, however, the choice of macroeconomic variables may be improved and the tests by extensive simulation using the constructed model are in our future agenda.