I'm working on a project where we're trying to create a database model where we can (daily) update collected data in order to make RPA predictions.

We received data from Interest Rate Curves called IR-CMS(Constant Maturity Swap?) in one file and IR-OIS (Overnight Indexed Swap?) in another.

You see, as the values are exactly the same, the person who started modelling assumed both curves are and behave the same.

From what I've learned, sorry if I'm a bit lost here but I'm new in this: if my understanding is right, aren't CMS and OIS different things? Or why is it safe to assume both curves behave the same way?

  • 1
    $\begingroup$ It seems likely that the data is just bad. In any case, I don't know what RPA is, but an OIS curve would be useful for computing the NPV of future cash flows, and I don't know what a CMS curve would be used for. I do know that the latter would be much more complex to calculate. $\endgroup$
    – Drew
    Sep 18, 2019 at 12:44
  • 1
    $\begingroup$ ... I guess the CMS curve might be used to forecast CMS rates, in which case its derivation would probably utilize the OIS curve for discounting. It still seems likely that it has been done wrongly if the result is a curve identical to OIS. $\endgroup$
    – Drew
    Sep 18, 2019 at 12:54
  • $\begingroup$ Please clarify what is RPA. $\endgroup$
    – Alex C
    Sep 18, 2019 at 17:19

1 Answer 1


I'm not sure what your exact situation is, nor what an RPA is, but I can at least explain to you OIS and CMS. I think you have terminology confusion.

CMS means "constant maturity". That is an interpolation between active issues to allow for consistency. For example, when looking at treasuries you are often looking at a small set of actively trading issues. But, how can you compare the yield on the 2yr note from 2 weeks ago to the yield now? There are different numbers of days. With high issuance, it could even be a different, new, note now.

Generating a CMS curve creates a set of points for each tenor point. That would allow us to see the implied 2 year maturity on any given day. On most days you will not have an instrument (bill or note or bond) that matures at that exact point - so you are relying on interpolation of some sort.

Now OIS is different because a series of futures is used to calculate the implied yield on any given day. There are no OIS bonds or notes, so you don't have to get rid of the optical confusion created by an X-year note that is really only [X-years - y-days] on any given day.

So really an OIS curve will be the same as a CMS curve on any given day.

Does that help?

  • $\begingroup$ I'm not sure why someone downvoted this. This is correct. I trade this product every day for a living. $\endgroup$
    – JoshK
    Sep 20, 2019 at 0:32

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.