|
For 2) and 3), you cannot use real (or historical) drift for option pricing, as it will probably cause arbitrage opportunities. For 1), I am very agree with Chemist_MZ's explanation.
Recommended readings: Paul Wilmott on Quantitative Finance, Ch. 15.
I rarely use risk neutral concept to price derivatives. From my experiences, it's very easy to introduce logical errors which are hard to discover, especially for interest rate products.
My suggestion is to derive the PDE with strict mathematics, and then applying Feyman-Kac formula to work back the so called risk neutral drift. Yes, it's boring but it minimize the chance of getting logical error.
|