Serving the Quantitative Finance Community

 
ferdelo
Topic Author
Posts: 9
Joined: September 11th, 2016, 8:31 am

SABR: One parametrization to rule the world .. ?

December 5th, 2017, 7:13 am

I have two questions that prevent me from sleeping well... If someone could help me .. :( 

1. I was wondering whether traders use the same smile model (parametrization)  to price vanillas (EUR cash swaptions) and CMS. When they contribute to Totem at end of month, do they use the same parametrization to contribute the central smile (with the ZWC), wings and CMS .. ?

2. What is the standard to price vanillas in EUR currency (Cash Swations .. ?). Do people rely on a mapping and integrate the pay-off .. ? Or do people price them "a la Black" , for example, by using a cash annuity measure (and using somehow the corresponding forward cash swap)

Thank you very much in advance ... !!
 
ferdelo
Topic Author
Posts: 9
Joined: September 11th, 2016, 8:31 am

Re: SABR: One parametrization to rule the world .. ?

December 6th, 2017, 9:30 am

i am afraid i will keep sleeping not so well... 

thank you anyway...
 
User avatar
kfp22
Posts: 3
Joined: October 8th, 2014, 1:34 pm

Re: SABR: One parametrization to rule the world .. ?

February 5th, 2019, 12:51 am

I would very much like to know too..
 
User avatar
Finatos
Posts: 11
Joined: July 28th, 2012, 8:18 pm

Re: SABR: One parametrization to rule the world .. ?

March 2nd, 2019, 4:58 am

This is what Bloomberg does:

For vanilla European option there is no point of using complicated models. Black/Normal model definitely is good enough. Of course there is no blocker of using more complicated models such as Hull-White or LMM models to price these vanilla instruments but since these models are usually calibrated to these vanilla instruments, they should be able to backout the vanilla instrument's price otherwise the calibration error is huge which means the model is not implemented well.

For CMS swaps you need covexity adjustments. Usually Black and Normal model or use replication method is good enough. No need to use sophisticated stochastic models. Advanced models are used to price complex or exotic rates derivatives. Just name a few, Bermudan swaptions, CMS spread options, range accruals, etc. These are path dependent deals don't have analytic solutions. You need to use MC simulations.