Stochastic bounds for Lévy processes. (Q1879836)

Let \(X=(X_t : t \geq 0)\) be an arbitrary Lévy process which Lévy measure satisfies \(\Pi(R)>0\). For a fixed interval \(I\), which contains zero and has \(\Pi(I^c)>0\), write \(\tau_n\) for the time at which the \(n\)th jump in \(X\) occurs, whose value lies in \(I^c\). Using the associated random walk \(S:=(X(\tau_n) : n \geq 0)\) one has the natural stochastic bounds \[ I_n:=\inf_{\tau_n \leq t < \tau_{n+1}}X_t \leq X_t \leq \sup_{\tau_n \leq t < \tau_{n+1}}X_t =:M_n. \] The author finds a representation for the random variables \(M_n\) and \(I_n\) which allows him to draw conclusions about the asymptotic behavior of Lévy processes from the corresponding results for random walks in a simpler way.