Rough path analysis applies successfully to multidimensional stochastic processes. Ito diffusion processes on the other hand have provided models for stock market prices since F. Black and M. Scholes, Harrison and co-authors. In the following note, we intend to bring some controlled rough path considerations to the study of multidimensional Black-Scholes models with position-dependent volatilities. The Ito formula for reduced rough paths provides the syntactical analogy to bridge classical Mathematical Finance to the (Davie's type)RDE-based one. Once such connection is established, rough integration bounds are employed to estimate the errors arising from time-discretisation of the theoretical continuous-time trading strategies. We find that the portfolio trajectory resulting from the hedging strategy, when analysed pathwise, is sensitive to Greeks beyond the leading first-order Delta, and is Gamma-sensitive in particular. This provides a theoretical underpinning to reconcile traders' practice of Gamma-neutral hedging. We interpret the Gamma position in terms of (co-)variance swaps on the underlying securities, and we enlarge classical strategies in cash and stocks with positions on such swaps. Furthermore, the proposed framework addresses robustness of Delta hedging. We analyse the error that a practitioner incurs into when hedging with the wrong model. We do so without assuming an Ito's diffusion form for the actual price dynamic, whose trajectory can in particular be rougher than Brownian motion.
↧