Nonlinear Kalman Filtering in Affine Term Structure Models

Peter Christoffersen, Christian Dorion, Kris Jacobs, Lofti Karoui

Research output: Contribution to journalJournal articleResearchpeer-review

Abstract

The extended Kalman filter, which linearizes the relationship between security prices and state variables, is widely used in fixed-income applications. We investigate whether the unscented Kalman filter should be used to capture nonlinearities and compare the performance of the Kalman filter with that of the particle filter. We analyze the cross section of swap rates, which are mildly nonlinear in the states, and cap prices, which are highly nonlinear. When caps are used to filter the states, the unscented Kalman filter significantly outperforms its extended counterpart. The unscented Kalman filter also performs well when compared with the much more computationally intensive particle filter. These findings suggest that the unscented Kalman filter may be a good approach for a variety of problems in fixed-income pricing.
Original languageEnglish
JournalManagement Science
Volume60
Issue number9
Pages (from-to)2248–2268
ISSN0025-1909
DOIs
Publication statusPublished - 2014

Cite this

Christoffersen, P., Dorion, C., Jacobs, K., & Karoui, L. (2014). Nonlinear Kalman Filtering in Affine Term Structure Models. Management Science, 60(9), 2248–2268. https://doi.org/10.1287/mnsc.2013.1870