November 29th, 2002, 11:38 am
Paul, I've been thinking a bit more on the subject. And, I see what you mean, looking at i.e. a up&out: its indeed very sensitive too what volatility one quotes. And especially, it also depens un where the barrier is in relation too the current spot price. Far away you want the vol to bee high, but of course the value decreases witgh higher vol. near the barrier. So, you could of course price in a stoch vol environment. But, how usable are these models, I mean isnt predicting the parameters "a real pain in the as" (considering the uncertainty, I mean, is there such a thing as stable correlations or smile for the vol)? The voll often seems to be clustering, soo I think it shoud be possible too come up with a model which is sort of "state dependent"? But, purhapse its even more simple too have an idea of a "volatility band", i.e. like " the vol stays in the range x-y with some probability". But know, that woudnt help, since no one would trade with you on that basis. So, the practioners has to have some kind of optimal replication strategy (like the ideas of static hedging) ?!??