Realized GARCH Option Pricing using the Filtered Historical Simulation Approach

In this thesis I combine the filtered historical simulation (FHS) method of Barone-Adesi, Engle and Mancini (2008) with the Realized GARCH model of Hansen, Huang and Shek (2011). I use the RGARCH model instead of the GJR-GARCH model, since realized data contains information on the volatility and contributes to more accurate volatility estimates. Using an empirical analysis, I apply the FHS method on in-the-money and out-of-the-money option data of the S&P 500, for the period January 2nd 2002 till December 30th 2011. Based on statistical criteria such as the root mean squared error and the mean absolute error, the FHS method in combination with the RGARCH model improves option pricing. The latter method also produces accurate Value-at-Risk forecasts.