Option pricing when correlations are stochastic: an analytical framework (Q2425554)

This paper presents an analytically tractable model for capturing the joint behavior of prices of an underlying set of assets, options on the individual underlying assets, and derivatives on baskets of the underlying assets. This model allows for non-trivial stochastic volatility of asset returns and stochastic correlation of cross-sectional asset returns is fully consistent with the sorts of smile and skew effects that are apparent in typical market pricing of plain vanilla option prices. In this model, prices (conditional to volatilities) evolve according to a lognormal diffusion while the stochastic variance-covariance matrix follows a Wishart process. This paper improves previous literature on correlation risk: first the authors characterize all linear correlation structures between underlying asset returns and their volatilities, which are fully consistent with the smile and skew effects on vanilla options. Second, it is quantified the effect of stochastic correlations on option prices by introducing the notion of implied correlation together with its smile and skew effects. It is provided a complete treatment in the case of a best-of basket option, where the whole procedure can be carried out explicitly. In this case the appearance of an implied correlation skew can be explained in terms of the asymmetric reaction of correlations with respect to positive or negative shocks on returns.