A Review of Term-Structure Models and Their Applications: [forms Part Of: Report of the Fixed-Interest Working Group, B.A.J. 4, II Pg.213–383]

1998 ◽  
Vol 4 (2) ◽  
pp. 323-383 ◽  
Author(s):  
G.B. Chaplin

ABSTRACTThe literature on ‘Term-Structure Models’ is extensive with many contributions from financial economists over the last twenty years. This paper reviews examples of term-structure models from different categories (‘equilibrium’, ‘evolutionary’ and ‘descriptive’) with particular emphasis on their intended application. The Vasicek (one-factor equilibrium), Richard (two-factor), and Hull & White (evolutionary) models are discussed in some detail.The paper reviews a particular class of descriptive polynomial models which is flexible and in widespread use both in the academic and the practitioner community. The model is cast in terms of forward rates, applied to the gilt market, and techniques are used to determine how many terms in the polynomial expansion are statistically required in order to describe the market accurately. The model is a linear model of forward and spot rates and is stable; this allows the calculation of risk measures for each bond which give a superior approach, in principle, to portfolio hedging.Selection of model should be driven by its application. If the objective is a reasonably accurate description of the curve and, by implication, an accurate indication of yields which can be obtained in the market, then a model which fits the market accurately is preferable. The ‘descriptive’ approach is therefore most appropriate in this context.

2007 ◽  
Vol 42 (1) ◽  
pp. 41-80 ◽  
Author(s):  
Huarong Tang ◽  
Yihong Xia

AbstractWe examine the yield curve behavior and the relative performance of affine term structure models (ATSMs) using government bond yield data from Canada, Germany, Japan, the U.K., and the U.S. We find strong predictability of forward rates for excess bond returns and reject the expectations hypothesis in all five countries. A three-factor model is sufficient to capture movements in the yield curve of Canada, Japan, the U.K., and the U.S., but may not be enough for Germany. An exhaustive comparison among ATSMs with no more than three factors reveals that the three-factor essential affine model (A1(3)E), with only one factor affecting the volatility of the short rate but with all three factors affecting the price of risk, performs best in all five countries. Simulations provide inconclusive evidence on whether this best affine model can successfully generate the rich yield curve behavior observed in the data.


Author(s):  
Carl Chiarella ◽  
Chih-Ying Hsiao ◽  
Thuy Duong To

Sign in / Sign up

Export Citation Format

Share Document