Stochastic Volatility Model with Time‐dependent Skew.pdfVIP

Applied Mathematical Finance, Vol. 12, No. 2, 147–185, June 2005 Stochastic Volatility Model with Time- dependent Skew VLADIMIR V. PITERBARG Bank of America, 5 Canada Square, London, E14 5AQ, UK ABSTRACT A formula is derived for the ‘effective’ skew in a stochastic volatility model with a time-dependent local volatility function. The formula relates the total amount of skew generated by the model over a given time period to the time-dependent slope of the instantaneous local volatility function. A new ‘effective’ volatility approximation is also derived. The utility of the formulas is demonstrated by building a forward Libor model that can be calibrated to swaption smiles that vary across the swaption grid. KEY WORDS: Stochastic volatility, volatility smile, time-dependent local volatility, effective volatility, effective skew, average skew, homogenization, averaging principle, effective media, forward Libor model, Libor market model, LMM, BGM, volatility calibration, skew calibration Introduction Stochastic volatility models are a popular choice in volatility smile modelling (the body of literature on stochastic volatility models is immense and constantly grow- ing, so we only mention the classic book; Lewis, 2000). To be fully useful, a stochastic volatility model must possess efficient numerical methods for pricing European options. As a rule, such methods are only available for stochastic volatility models with time-homogeneous parameters (Andersen and Brotherton-Ratcliffe, 2001; Hagan et al., 2002; Zhou, 2003; Andersen and Andreasen, 2002). Empirical evidence from many markets suggests however, that the parameters of the volatility smile (such as its level, slope and curvature) are different for different option expiries. While for vanilla options it is us