Option Pricing Using Realized Volatility

Abstract

Estimation of the volatility process of asset returns is of paramount importance in many financial applications particularly when the ultimate objective is the pricing of derivatives. Recently the use of high-frequency intraday data for estimating daily volatilities has received extensive attention with very promising results when compared to using lower frequency interday data. In the present paper we suggest to model one such measure, the Realized Volatility, as an Inverse Gaussian distributed variable with time varying mean. We derive the appropriate model to be used for option pricing purposes, and we show that it can explain some of the mispricings often found when using traditional option pricing models using interdaily data only. We then perform an empirical analysis using options on three large American stocks, and we show that in all cases the suggested model performs significantly better than models which are estimated on return data alone. Hence the paper provides evidence on the value of using high frequency data for option pricing purposes.

Keywords: Option Pricing, Realized Volatility, Stochastic Volatility, GARCH

JEL Classification: C22, C53, G13