|Main index||Financial Markets index||About author|
Julian D. A. Wiseman
Abstract: The exponential yield curve model (1994) is obsolete. Instead please use and refer to The Magpie Yield Curve Model, 1998.
Publication history: only at www.jdawiseman.com/papers/finmkts/201304_exponential_yield_curve_model_wiseman_1994.html. Usual disclaimer and copyright terms apply.
In 1994, whilst employed at J.P. Morgan, I wrote a paper, The exponential yield curve model, describing a very smooth yield curve model. That model had some good qualities, and a set of inter-related bad qualities (as described in Cairns A. J. G., Stability of Descriptive Models for the Term Structure of Interest Rates, 1997). The old model was superseded by The Magpie Yield Curve Model, published on 24th February 1998. The Magpie model was built on the Exponential, has all of its advantages, and none of its disadvantages. It is still excellent.
The problem is that J.P. Morgan has not been thorough about keeping research. The earlier obsolete paper is still available to those with access to JPM’s website; the latter better paper is not. Hence it is the old paper is referred to by the likes of A tractable multi-factor dynamic term-structure model for risk management, Henseler, Peters, and Seydel. (Others papers also do, some even published by JPM, but Henseler et al have done so most recently.)
Please use and refer to the new excellent ‘Magpie’ model rather than the imperfect older model. Hopefully those thinking of using the earlier paper will search for it, find this brief warning, and ask for the newer paper. Hopefully.
A later version of the Magpie model, with some extra observations and a slight simplification of usage, is also described in The Magpie Yield Curve Model at SG (= Société Générale), 1st March 2012, and available on that former employer’s password-protected website.
|— Julian D. A. Wiseman|
London, 11th April 2013
August 2015: Available for sale is an implementation of the Magpie model, made to a very high standard of Excel craftsmanship. It is intended to be used of itself, and also to be used as the prototype and calculation-checker for a re-implementation in a high-speed language. If interested, contact the author.
July 2023: a formal and complete writeup has been done, being 20 pages. It should be published—if you edit a suitable journal and might be interested, please let me know.
|Main index||Top||About author|