Serving the Quantitative Finance Community

 
User avatar
RiskUser
Topic Author
Posts: 12
Joined: July 19th, 2006, 4:30 pm

Creating a "Normal Swaption Volatility Cube" using multiple data sources

May 23rd, 2016, 8:52 am

Hi Forum,I am looking to build a "blended" Swaption Volatility Cube derived from prices observed from multiple data sources i.e. source A provides a term structure of Forward Swap rates and Premia / Normal Vols for differing strikes and tails, source B provides a "different" term structure of Forward Swap rates and Premia / Normal Vols for differing strikes and tails etc etc.My initial thoughts were to try and fit a SABR surface using my own term structure of Forward Swap rates to all the quotes provided and try to minimise the difference across all the quotes while applying some arbitrage checks along the way.However, I was wondering if there was a "smarter" way to do this i.e. request quotes against my term structure of Forward Swap rates or use fixed strikes and then apply a fitting routine to help reduce the degrees of freedom in the problem I am trying to solve.Any thoughts / suggestions?-RiskUser
 
User avatar
pcaspers
Posts: 30
Joined: June 6th, 2005, 9:49 am
Location: Germany

Creating a "Normal Swaption Volatility Cube" using multiple data sources

May 23rd, 2016, 6:12 pm

Not sure if I got the question, but if you have different Forwards from different providers for some reason (different timestamps, different underlying curve constructions) I would map each vol quote to ATM + spread, where ATM is the forward level from the same provider. After that you can work with "your" ATM and place the quotes via their relative-to-ATM-spreads around it. For normal vol quotes or premia this seems reasonable. Maybe also for (shifted) lognormal quotes, if the ATM differences are not big (they shouldn't be anyway ?!).