August 22nd, 2010, 5:01 pm
Agree with Piterbarg. However, one also has to take into account things like convexity, skew/smiles and distributions. Swaption market would essentially give you everything about 2y swap rate distributions at every time point. For 10-2 spread, there is indicative broker market and totem survey, which might give you some ideas about 10-2 distributions. What's the relationship between the two? That's something you have to think about. Or from 10s distribution and 2s distributions, plus assumption between the two, you could derive the 10-2 distributions so that it matches 10-2 broker market. You would have to do numerical integration here.Piterbarg,Any comment? What's your thought on which model to take? Maybe you can share a little thought before your book is published. Do you discuss the cms spread option market in your book?