Serving the Quantitative Finance Community

 
User avatar
pcg
Topic Author
Posts: 0
Joined: September 13th, 2004, 11:11 am

cap vols: flat vs spot

April 12th, 2006, 5:30 am

can someone say how one converts spot cap vols into flat vols and the other way round ?
 
User avatar
mutley
Posts: 20
Joined: February 9th, 2005, 3:51 pm

cap vols: flat vs spot

April 12th, 2006, 8:12 am

I guess you could call it 'vol bootstrapping'.Flat vols: this is the constant vol number you give to all caplets underlying a certain cap such that the sum of their PVs equals the market cap price. Clearly, some will be over priced and others will be underpriced. You can get an idea for the spot volatility [the per-caplet vol that prices all caplets correctly so the running sum equals the PV of the market caps as we extend the cap tenor] if you take a functional form for the underlying spot vol and optimise the parameters governing its behavior to minimise the fit between the market and model cap prices.Basically spot vols are defined instantaneously, sigma_1(t), and flat vols are defined over a period/tenor, sigma_2(0,t).
 
User avatar
PedroRaquel
Posts: 0
Joined: August 7th, 2006, 2:12 pm

cap vols: flat vs spot

October 12th, 2006, 6:45 am

QuoteOriginally posted by: mutleyI guess you could call it 'vol bootstrapping'.Flat vols: this is the constant vol number you give to all caplets underlying a certain cap such that the sum of their PVs equals the market cap price. Clearly, some will be over priced and others will be underpriced. You can get an idea for the spot volatility [the per-caplet vol that prices all caplets correctly so the running sum equals the PV of the market caps as we extend the cap tenor] if you take a functional form for the underlying spot vol and optimise the parameters governing its behavior to minimise the fit between the market and model cap prices.Basically spot vols are defined instantaneously, sigma_1(t), and flat vols are defined over a period/tenor, sigma_2(0,t).Mutley, what kind of functional form would you recomend? Would this be a decent one: σ(t) = (a+bt)exp(-ct)+d?Thanks