Hello,I'd like to know how we can structure this product:the client gives 100 for a structured product with a maturity between 3 and 8 years,the first year, he touches 8.5%the second year, he touches 8.5%the third year (to the 8th), he touches 15+50%(worst performance of a 20 equities portfolio)if the worst perf is positive, the product is offif the worst perf is negative, the product continuescan someone help me with this product?

"touches" means "receives" I guess."the product is off" means that the client is paid out the cash value of the worst of 20 stocks and the structure is gone?"the product continues" means what precisely?it sounds to me like a monte carlo simulation. the simplest apprach is to directly simulate the forward of all 20 stocks given the div, vol and correlation matrix assumptions.thing is, if you want skew to go into this, you will have to assume some sort of model. people seem to be using local vol models to put skew into multi-factor exotic (but path indepependent) options such as these options nowadays.hope that helps!

It looks like a moutain range option (Everest type). Here's a link to a useful website : http://www.global-derivatives.com. Search for Everest.Regards.

That's true, but you should better use a stochastic or jump model in your monte carlo simulation.local vol models bring problems as soon as they are used in forward sensitive products, because they don't restitute the forward vol surface...

That's true, but you should better use a stochastic or jump model in your monte carlo simulation.local vol models bring problems as soon as they are used in forward sensitive products, because they don't restitute the forward vol surface...

That's true, but you should better use a stochastic or jump model in your monte carlo simulation.local vol models bring problems as soon as they are used in forward sensitive products, because they don't restitute the forward vol surface...

That's true, but you should better use a stochastic or jump model in your monte carlo simulation.local vol models bring problems as soon as they are used in forward sensitive products, because they don't restitute the forward vol surface...

That's true, but you should better use a stochastic or jump model in your monte carlo simulation.local vol models bring problems as soon as they are used in forward sensitive products, because they don't restitute the forward vol surface...

benchy - i agree with all five of your comments! but i am told that local vol, pure and simple, is used to price path-independent multifactor options such as a worst-of-20 derivative. path dependent derivatives in particular do very badly in local vol i believe.in the local vol calculation of a path independent derivative price, you only need to agree the terminal distribution (i guess it would be a lower dimensional slice of the 20 dimensional surface?) and not the entire surface, right?i mean you need to get the terminal distributions right, so that would include the skew of the T-maturity european options for every underlying, but all other info could be wrong, and you would stil get the correct path-independent derivative value.

By forward vol surface, do you refer to the Dupire's equation ?