Simulations for American option pricing under a jump-diffusion model: Comparison study between kernel-based and regression-based methods

Abstract

There is no exact analytic formula for valuing American option even in the diffusion model because of its early exercise feature. Recently, Monte Carlo simulation (MCS) methods are successfully applied to American option pricing, especially under diffusion models. They include regression-based methods and kernel-based methods. In this paper, we conduct a performance comparison study between the kernel-based MCS methods and the regression-based MCS methods under a jump-diffusion model. © 2008 Springer-Verlag Berlin Heidelberg.