1
$\begingroup$

Let $(B_t)_{t \geq 0}$ be a standard Brownian motion. The reflection principle tells us that $\sup_{u \leq t} B_s$ has the same distribution as $|B_t|$. Let $H_a$ be the first hitting time of $a$ by $|B_t|$, and $T_a$ the first hitting time of $a$ by $B_s$.

I believe that the first hitting time of $a$ by $B_s$ is equal to the first hitting time of $a$ by $\sup_{u \leq t} B_s$. Why then, is $\mathbb{E}[H_a]$ finite and $\mathbb{E}[T_a]$ infinite? Are the underlying distributions not equal in each case due to the reflection principle?

Thanks.

$\endgroup$
2
$\begingroup$

The reflection principle states that

$$M_t := \sup_{s \leq t} B_s \quad \text{and} \quad |B_t|$$

have the same distribution for each fixed $t$; it does not say that the processes $(M_t)_{t \geq 0}$ and $(|B_t|)_{t \geq 0}$ are equal in distribution. In fact, we cannot expect that the two processes are equal in distribution; this follows for instance from the fact that $(M_t)_{t \geq 0}$ has non-decreasing sample paths whereas the sample paths of $(|B_t|)_{t \geq 0}$ fail to be monotone.

Since the stopping times $T_a$ and $H_a$ clearly depend on the whole sample path of Brownian motion, the reflection principle cannot be used to deduce that $H_a$ and $T_a$ have the same distribution. (For instance if we want to know whether $T_a(\omega)=t$ for some $\omega$ and $t$, then it does not suffice to know the value $B_t(\omega)$; we need to know the path $(B_s(\omega))_{s \leq t}$.)

$\endgroup$
  • $\begingroup$ Thank you, this makes a lot of sense. $\endgroup$ – Daven May 26 '18 at 17:13
  • $\begingroup$ @Daven You are welcome. $\endgroup$ – saz May 26 '18 at 17:14

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.