2

I am often told that options priced under SLV models, the Greeks cannot be exactly replicated by finite differences, but are computed at the level of the grid used to solve the PDE.

Can someone please explain how these Greeks (delta and gamma) are computed and to what extent they can be replicated by finite differences?

Thank you in advance.

AIEA
  • 21
  • 2
  • Apart from small numerical errors, there should be almost no differences between getting Greeks from the PDE vs bump and reprice (FD). Why do you not ask the people who keep telling you it doesn't work directly? – AKdemy Mar 11 '24 at 21:21
  • Thank you @AKdemy for you comment, I'll rephrase my question: when you bump and reprice you option you'll be taking another grid for you second option, this can lead to a noise in your delta especialy for small enough bumps. My question is if this noise is quantifiable and how to limit it ? (would taking smaler grid steps be a good idea ?) my question is for people that are used to work with these type of models and can give me more insights on how they validate it with finite diffrences (I'm at 4% error for now but can it be better ?) – AIEA Mar 12 '24 at 20:54

0 Answers0