I am baffled about the following problem:

Let $(B_t)$ be a standard Brownian motion. Let $$ \tau:= \inf\{ t \geq 0 :B_t = x \} \wedge \inf\{ t \geq 0 :B_t = -y \}$$ be a stopping time, where $x,y >0$. I am eager to know why the stopped process $(B_{t \wedge \tau})_{t \geq 0}$ is U.I..

Moreover, we know that $(\tilde{B}_t := B^2_t -t)_{t \geq 0}$ is a martingale. But the book also claims that $(\tilde{B}_{t \wedge \tau \wedge n})_{t \geq 0}$ is U.I., for any $n \in \mathbb{N}$. Why is that?

  • $\begingroup$ What does U.I. mean? Uniformly integrable? $\endgroup$
    – Math1000
    Dec 22 '14 at 2:04
  • $\begingroup$ That is what I think, Math1000. $\endgroup$ Dec 22 '14 at 2:22
  • $\begingroup$ Richard: 1. Which parts of the answer were escaping you when you asked this question? 2. Since the answer does not solve the part about $\tilde B$, I guess that you managed to solve it yourself. Is that correct? $\endgroup$
    – Did
    Dec 22 '14 at 6:10
  • $\begingroup$ yeah, I was just thinking about proving uniform integrability by some other methods. I overlooked the possibility of proving boundedness. $\endgroup$
    – Richard
    Dec 23 '14 at 0:04

We know that $\mathbb{E}[\tau] < \infty$ and $B_{t \wedge \tau}$ is a martingale. It is also easy to see that $-y \leq B_{t \wedge \tau} \leq x$. Any bounded martingale is uniform integrable. Hence, $B_{t \wedge \tau}$ is UI martingale.


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.