As some of your may know from my other posts, I am working on a Dynamic Nelson Siegel (DNS) based relative value trading model. On simulated data (which satisfies all the assumptions) of the DNS it worked well (unsurprisingly).

However I now need to get real data/results.

Part of the logic of the model is to look at the relative value of the most liquid points say 2,5,10y relative to all the remaining points.

So it's important that the zero coupon data represents ideally traded or at least traceable rates for all tenors rather than stale rates or even interpolated rates (as in this case I am just comparing different interpolation schemes!).

Does anyone know of a Bloomberg page that offers such rates? Or at least makes a distinction between which rates are real and which are interpolated?

Thanks Baz


For the US Treasury market, zero coupon bonds are traded and they are called STRIPS. You can access them through "S GOVT" (coupon Strips) or "SP GOVT" (principal strips) on BBG.

With regard to relative value trading, it's actually pretty rare that we fit models to zeros, because a lot of them are not liquid and trade differently from their coupon counterparts.

Instead, what you should do is to fit your model to coupon bonds or par swaps directly. Like you said, you could potentially fit the 2y, 10y, and 30y bonds, then look at relative value elsewhere. Once you have fit a model to coupon Treasuries, you can then assess STRIPS against this same model.

For some real life examples, Tuckman's Fixed Income Securities (3rd Edition) Chapter 11 is quite good.

  • $\begingroup$ Thanks for this! I just want a zero coupon curve (derived from the most liquid instrumnents) not a curve of zero coupon isntruments sorry for any confusion. Problem seems to be that a lot of contributors to bloomberg only give market rates for a small subset of the curve with the remaining rates being interpolated. Bloomberg sets a minimum of four instruments for the zero curve construction (so would hope these at least are liquid)! But even if I know or could hazard a guess as to which instruments are being used it still leaves the problem that the remainig points are interpolated and so I $\endgroup$ – Bazman Jun 23 '14 at 15:12
  • $\begingroup$ would just be comparing interpolation schemes rather than true relative value? Even if I fit to bonds and par swap directly it's still not possible to know which rates are market observed and which are interpolated? Plus when stripping you have to start making assumptions about the interpolation method i could very well use the Nelson Siegel framework but then my stripped "market" points would simply match the model and there would be no relative value. $\endgroup$ – Bazman Jun 23 '14 at 15:23
  • $\begingroup$ @Bazman I might not be understanding you completely, but all bond prices on BBG should be quoted; i.e., the prices you see for all 300is Treasuries are tradable market prices, not model prices. As far as I know, the par swap rates are composite quotes too. At a minimum, 1y, 2y, 3y, 4y, 5y, 7y, 10y, 12y, 15y, 20y, 25y, and 30y par swap rates are quoted and not interpolated. $\endgroup$ – Helin Jun 23 '14 at 15:59
  • $\begingroup$ Thanks that helps but from talking to bloomberg it seems their only requirement for contributed curves is that a minimum of four rates be market observed (still trying to ascertain if four market rates is the norm of the exception). So you have a handle on how liquid rates like 11y are? Do they trade every day? $\endgroup$ – Bazman Jun 23 '14 at 16:35
  • $\begingroup$ Hi @Bazman, I assume you're referring to BBG's curve builder when you mentioned "a minimum of four rates"? I was actually talking about retrieving quoted rates directly, using tickers such as USSW2, USSW10, etc. These should get you the latest quoted rates. 11yr is not actively traded... $\endgroup$ – Helin Jun 23 '14 at 17:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.