wilmott.com

Do you use a 3-months period or a 6-months period? and why?The calibration instruments (caplets and swaptions) depend on both periods (caplets: typically 3mths at the short end, 6mths beyond 2 years or so, swaptions: 3mths for floating leg, 6mths for fixed leg)If you want to price CMS(20yr)-CMS(6mths) you might prefer to have a 3months period model to deal with the CMS(6mths)...Also, funding side of the deal you want to price could be 3mths Libor + spd.I am interested in hearing people's views on that matter!Do you make approximations (and make life easier) or do you deal with all these things "exactly"?ps: the above refers to JPY, USD, GBP (not Euro)

As i know, in US market caplets are based on 3m rates and swaptions are based on 6m rates. So u can extract vols of 3m rates from cap vols and convert them into vols of 6m rates and then by which correlation of 6m rates can be extracted from swaption vols. but i still got a trouble for pricing notes based on 3m rates like 3m libor range accrual. another way is to simulate 3m rates and get correlation of 3m rates from swaption vols, nevertheless i think swaptions depend more basically on 6m rates than 3m rates.any advice?

these things are a pain.For calibration:as it really is JPY I m into, apparently it is 6m for everything, so calibration is easy.For pricing:- if a deal has 3m libor, it is ok to assume it is 6m.- if a deal has frequency 3m for the structured side then I calibrate as if calibration instruments were 3m. I checked against another guy who was modeling frequency exactly and our prices were close.And we don t have many of those deals.But yes, these things are really a pain and I wonder how people deal with them (do they do it exactly od they make assumptions to make life easie etc).

Hert, did u have two periods in ur calibration algorithm?

hope any experts in libor markt models can help me.form the perspective of modeling, if choose 3months/6months periods as underlyings tenor in ur libor market models, u are in fact modeling the joint distribution of 3months/6months forward rates.from the perspective of US market, caps depend on the marginal distribution of 3months forward rates and swaptions basically on the joint distribution of 6months forward rates. my Qs are (1) will be over-modeling for swaption if choosing 3months as underlyings in libor market models (2) cannt deal with the structures based on 3months libor if choosing 6months as underlyings in libor market modelsbtw, brigo suggests, in his excellent book " libor market model in theory and practice", to model 6months rates and provides a tool converting vols of 3months rates into vols of 6months rates.

no one know that???