Asymptotic Analysis of Volatility推荐.pdfVIP

CHAPTER 55 asymptotic analysis of volatility In this Chapter. . . • asymptotic analysis and small or large parameters • a series solution for implied volatility under arbitrary stochastic volatility 55.1 INTRODUCTION Asymptotic analysis is a systematic method for exploiting the largeness or smallness of a parameter in some equation. In our context the equation is that for an option under stochastic volatility and the parameters measure the speed of mean reversion of volatility and the volatility of volatility. The mean reversion is fast (large parameter) and the volatility of volatility is large (another large parameter). Since we’ll be looking for asymptotic solutions we’ll see how the precise specification of the model is irrelevant as far as finding closed-form solutions is concerned. In other words, we don’t have to sacrifice accuracy for tractability anymore. This chapter is heavily based on a paper by myself and Henrik Rasmussen, Rasmussen Wilmott (2002). 55.2 FAST MEAN REVERSION AND HIGH VOLATILITY OF VOLATILITY We consider the pricing of options when the underlying asset value S and its volatility σ are described by the stochastic differential equations, dS/S = r dt + σ dX dσ = A dt + B dY dX · dY = ρ dt where X and Y are Brownian motions, r is the short rate, and the coefficients A and B are functions of only σ . When calibrating such a stochastic volatility model to market prices, one 902 Part Five advanced topics usually finds that the volatility of volatility B/σ is greater than the volatility σ of the underlying. For instance (Wiggins, 1987), σ ∝ 0.2 B ∝ 0.2, in which case the ratio between volatility and the volatilit