Some FKG inequalities for stochastic processes

Abstract

This paper is interested in proving correlation inequalities of the FKG-type for various stochastic processes in continuous time. The pivotal tool which yields these correlation inequalities is an approximation with (possibly conditioned) Markov chains and random walks. In particular, we prove FKG inequalities for Lévy processes, Bessel processes and several conditioned Brownian processes. As a side result, we also provide a necessary and sufficient condition for a random walk distribution in $\mathbb{Z}$ to satisfy the well-known ``FKG lattice condition''.

15 pages