hi.. I am trying to construct the daily 3 month USD forward curve using depo rate up to 3 month, eurodollar up to 2 years and swap up to 30 years. I have some questions on this which I hope someone can help me.a) I am using local interpolation techniques for DF and intermediate swap rates for boostrapping. Unfortunately, my forward curves has many tiny humps espcially between the last eurodollar and the 2 year swap and the long end of the curve. Is there a way to obtain a smooth forward curve? or is't possible to get a smooth forward curve using local interpolation techniques?b) Which parameter has a major influence on the smoothness of the forward curve?c) Can anyone point me to some references on this? I went through the entire wilmott forum but can't seem to find any discussion on forward curve smoothing.Thanks a lot for the helpamkey04

smoothing the fwd curve is common (but still subjective) practiceit eliminates the zig-zag underspecification effect of bootstrapping using swaps (i.e. you have less market instrument prices than there are discount factor curve points) there is a good article in Oct-03 or Nov-03 Wilmott magazine called The Art and Science of Curve Building by Dr Owen WalshRgds.

Actually, local interpolation with certain kinds of splines is an OK method. The trick is to use the spline without a back-propagating second order effect.

thanks for the reply Jon and Slevin. Do you think you can PM me the article found on Wilmott magazine?

Last edited by amkey04 on April 25th, 2004, 10:00 pm, edited 1 time in total.

You could also use regression/optimization to fit the parameters of a Merton, CIR, LMM, Hull-White, etc. model. That way you actually have an implied short rate volatility and expectation, rather than just an interpolation.

Can you elaborate on that DrEvil? you mean interpolation on the swap rate?

Hi amkey...can I have the doc on the curves...txs

QuoteOriginally posted by: amkey04hi.. I am trying to construct the daily 3 month USD forward curve using depo rate up to 3 month, eurodollar up to 2 years and swap up to 30 years. I have some questions on this which I hope someone can help me.a) I am using local interpolation techniques for DF and intermediate swap rates for boostrapping. Unfortunately, my forward curves has many tiny humps espcially between the last eurodollar and the 2 year swap and the long end of the curve. Is there a way to obtain a smooth forward curve? or is't possible to get a smooth forward curve using local interpolation techniques?b) Which parameter has a major influence on the smoothness of the forward curve?c) Can anyone point me to some references on this? I went through the entire wilmott forum but can't seem to find any discussion on forward curve smoothing.Thanks a lot for the helpamkey04"Semi-Empirical Smooth Fit To The Treasury Yield Curve", Paul Diament, The Journal of Fixed Income, June 1993, pp. 55-70

Last edited by AlanB on December 1st, 2004, 11:00 pm, edited 1 time in total.

QuoteOriginally posted by: jonsmoothing the fwd curve is common (but still subjective) practiceit eliminates the zig-zag underspecification effect of bootstrapping using swaps (i.e. you have less market instrument prices than there are discount factor curve points) there is a good article in Oct-03 or Nov-03 Wilmott magazine called The Art and Science of Curve Building by Dr Owen WalshRgds.I do appreciate it if someone can possibly send me this article.Regards

Thanks Structurer, This is useful.And is this also applicable to commodity forward curves?

GZIP: On