A long term analysis of stochastic theta methods for mean reverting linear process with jumps

In this paper a relative analysis of moments reversion of the class of theta methods is provided for an stochastic differential equation with Poisson-driven jumps. We first determine under which conditions the first and second moments revert to steady state values. Second, we consider two different classes of implicit theta methods; theta-Euler method, and compensated theta-Euler method, and derive closed-form expressions for the conditional and asymptotic means and variances of considered methods. We provide a full analysis about the possibility to find methods able to replicate such long-terms quantities. Finally, to verify our theoretical results numerical experiments are given.