wilmott.com

Hi All,Let's suppose I have a SV LMM, which I use to price the a product dependent on multiple forward rates. Let's also say that I am able to have an excellent calibration to the caplet vol surface. I can check this by running my MC in the appropriate measures - i.e. if I run my simulation in the Tn measure I see that my caplets with maturity Tn are correctly priced, the same happens for the caplets maturying at Tn-1 when I simulate under Tn-1 etc...However when I want to price a complex product I obviously pick one single measure to simulare all my forward rares (nothing new here). However how is this influencing my Greeks? I will see Greeks which are "measure adjusted" hence they will differ from the same Greeks I would have calculared if I was simulating each forward under its own measure. Are these the right hedges to use or am I missing something? Thank you!

I think you are missing something, although it could be that your use of "hedge" is different from what I have in mind. In a complete market, the relationship between different instrument prices are functional (it is usually possible to define the value of a security at a point in time as a function of a set of state variables) and measure independent. In that case, theoretically perfect hedges can be calculated and these will also be measure independent. If your market setting is incomplete, however, and you are implementing some kind of minimum variance hedge, then all sorts of things can be measure dependent, but then I would question the wisdom of engaging in measure changes to begin with.

Measure change is only a tool to calculate some prices. Greeks have nothing to do with measure.

or to put it yet differently, the replicating portfolio is invariant under measure changes, only the probability for the paths to materialize changes.

Thank you all! Your comments have been very helpful.