December 17th, 2012, 8:18 am
Here is a scenario:You have an equity underlier (lets call it X) which you are trading based on a volatility surface. You also have a list of other equity underliers with their own, unique, vol surfaces. Your task is search through that list of other underliers to find the vol surface which is the best match (i.e. the most similar) for X's vol surface. You can do this is one of two ways:1) Run a 4 year, weekly, historical correlation of X against all other underliers and see which of the list are at least 75% correlated or more. From these, pick one which matches X's country of origin or business type. (i.e. try and pick within the list of those which are highly correlated one which is the most similar to X).2) Test for those which are co-integrated with X and pick one of those.I realise that both methods are not great, but this is the situation. Which is the lesser of two evils? Will searching for those which are co-integrated help at all in finding underliers which are a good match for X's vol surface? Thanks.