I am trying to prove a Levy process "renews itself" at stopping time, which implies the strong Markov property of Levy process. And I have a difficulty to verify the following result:

Let $X$ be a Levy process and $T$ be a bounded stopping time. Show $$\frac{E[e^{iuX_{T+t}}]}{E[e^{iuX_{T+s}}]}=E[e^{iuX_{t-s}}],\quad t>s.$$

First I can't use $X_{T+t}-X_{T+s}$ is independent of $\mathcal{F}_{T+s}$ and has the same distribution as $X_{t-s}$ since this is the result I am trying to prove. Hence I think I should prove it from the property of Levy process.

By definition $E[e^{iuX_{T+t}}]=\int_\Omega e^{iuX_{T(\omega)+t}(\omega)}\,dP$. Also I know $E[e^{iuX_{t}}]/E[e^{iuX_{s}}]=E[e^{iuX_{t-s}}]$ for any $t>s$. But I don't know how to connect the desired result to this identity.


It is not difficult to show hat

$$T_j(\omega) := k 2^{-j} \qquad \text{for} \, \, \omega \in \{(k-1) 2^{-j} \leq T < k 2^{-j}\}, k \geq 1,$$

defines a sequence of discrete-valued stopping times such that $T_j \downarrow T$. By the right-continuity of the sample paths of $(X_t)_{t \geq 0}$, this implies by the dominated convergence theorem

$$\mathbb{E}e^{i u X_{T+t}} = \lim_{j \to \infty} \mathbb{E}e^{iu X_{T_j+t}} = \lim_{j \to \infty}\sum_{k \geq 1} \mathbb{E} \left( e^{iu X_{k 2^{-j}+t}} 1_{\{T_j = k 2^{-j}\}} \right). $$


$$X_{k 2^{-j}+t} = \big( X_{k2^{-j}+t}-X_{k2^{-j}+s} \big) +X_{k 2^{-j}+s}$$

and conditioning on $\mathcal{F}_{k2^{-j}+s}$ we get by the independence and stationarity of the increments

$$\mathbb{E}e^{iu X_{T+t}} = \mathbb{E}e^{iu X_{t-s}} \lim_{j \to \infty} \underbrace{\sum_{k \geq 1} \mathbb{E}(e^{iu X_{k 2^{-j}+s}} 1_{\{T_j = k 2^{-j}\}})}_{\mathbb{E}e^{iu X_{T_j+s}}}.$$

Using again the dominated convergence theorem, we conclude

$$\mathbb{E}e^{iu X_{T+t}} = \mathbb{E}e^{iu X_{t-s}} \mathbb{E}e^{iu X_{T+s}}.$$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.