Multivariate asset models using Levy processes and applications

In this paper, we propose a multivariate asset model based on Levy processes for pricing of products written on more than one underlying asset. Our construction is based on a two-factor representation of the dynamics of the asset log-returns. We investigate the properties of the model and introduce a multivariate generalization of some processes which are quite common in financial applications, such as subordinated Brownian motions, jump-diffusion processes and time-changed Levy processes. Finally, we explore the issue of model calibration for the proposed setting and illustrate its robustness on a number of numerical examples.