Does Combining Models Help With Pricing of European Options?

In this study I combine six individual models to price European call options on the S&P500 index. Each model relaxes one or more assumptions of the Black-Scholes (1973) framework. The combining methods considered include three parametric and three non-parametric approaches. Combining outperforms most individual models in terms of the Root Mean Squared Error during the overall period (January 2006-December 2011) and three sub-periods. A simple and easy-implementable way of combining based on the historical performance of the individual models yields the best results among all individual and combining models considered. This method attaches most weight to the Merton (1976) jump diffusion model.