Patching SABR

surya2cents · February 21st, 2014, 4:58 am

The classical SABR runs into trouble at low strikes. Various patches have been proposed. Is there any consensus on what could be considered the best approach?

Pat · February 21st, 2014, 11:19 am

At the moment, I'd suggest Arb Free SABR ... it solves the problem at the expense of numerially solving an effective 1 dimensional PDE. Article should appear in Wilmott shortly. May be a better way soon.

cemil · February 21st, 2014, 12:56 pm

I can suggest a paper here and hereand here

Alan · February 21st, 2014, 2:32 pm

Also, this thread has some comments that may be relevant.

surya2cents · February 21st, 2014, 2:49 pm

Thanks Pat, look forward to reading that upcoming article. Thank you - cemll and Alan, my weekend is going to be busy!

surya2cents · February 22nd, 2014, 2:17 pm

I am also interested in pricing mid curve swaptions. Here a market model such as BGM(LMM) seems to be the way to go since we want to back out a volatility term structure. However, mid curve swaptions also exhibit a skew. So again we end up needing to do a LMM + skew model. Is there a nice unified model, that is also parsimonious ? Or is it too much to ask for!