Hi again all,

The title of the subject was a recent question during an interview as part of a longer discussion. The interviewer wanted to know how would I go about discriminating models for exotic derivative pricing and explain my findings to a trader. In other words, how do I explain a trader which model is best to use depending on what I wanna price.

I found the question quite broad and so I thought a broad answer would be fitting. I explained that a model should be chosen based on how easy it allows us to find appropriate and understandable hedge ratios, but I could not come up with anything more specific than that. The interviewer was not overly pleased with that I think, I imagine he wanted some specific examples.

I would like to know practitioners' take on the subject and know what they would have replied instead? I suppose the gist of the question is describing a trader the behaviour of different complex models using plain and easy terms.

Thanks in advance guys.

Hedging is one thing, if need be you calculate the hedge ratios by brute force. More importantly, does the model capture the risks of the product sufficiently.

For example, a forward start option is very easy to price and hedge using Black-Scholes. Would you use Black-Scholes? No, as the risk is forward volatility risk which is not captured at all by Black-Scholes.

Depends on your asset class.

Depends on your asset class.

Indeed, also asset class - good point.

You could talk about the number of factors that govern the contract's value, and thus the number of factors needed by a model.

Maybe a bit dated comment with advances in computational power but back in day the trade off for longer-dated fx/ir hybrid exotics was could I get my EOD portfolio greeks by start of trading the next day (so use less theoretically elegant yet quicker calculation-wise model during the week) & then on the weekend run the more theoretically elegant but slower model for COB Friday & do a sanity check comparison of the greeks from both models on Monday morning.