Let $(X_t, Y_t)$ be a two-dimensional $(\mathscr{F}_t)$-Brownian motion started from 0. We set, for every $t \geq 0$ $$\mathscr{A}_t = \int_0^t X_s dY_s - \int_0^t Y_s dX_s$$ (Lévy's area). My goal is to show $E[e^{i\lambda\mathscr{A}_r}] = \frac{1}{\cosh(\lambda r)}$ ($\lambda \in \mathbb{R}$).

I already know the following, which I believe should be helpful. (They were previous parts of the same question from Le Gall's book, Brownian Motion, Martingales, and Stochastic Calculus.)

First, I know that $E[e^{i\lambda\mathscr{A}_t}] = E[\cos(\lambda \mathscr{A}_t)]$.

Second, I know that $\langle \mathscr{A}, \mathscr{A} \rangle_t = \int_0^t (X_s^2 + Y_s^2) ds$.

Third, I know that, if $$Z_t = \cos(\lambda \mathscr{A}_t)$$ $$W_t = -\frac{f'(t)}{2} (X_t^2 + Y_t^2) + f(t)$$ and if $f(t) = -\log\cosh(\lambda(r-t))$, then $Z_t \exp(W_t)$ is a martingale. I also know what the canonical decompositions of these processes are and that $\langle Z, W \rangle_t = 0$. (See this question.)

The fact that $Z_t \exp(W_t)$ allows me to compute $E[Z_t \exp(W_t)] = E[Z_0 \exp(W_0)] = \frac{1}{\cosh(\lambda r)}$. This should be relevant but I don't know how to relate this back to $E[\cos(\lambda \mathscr{A}_t)] = E[Z_t]$.

Any thoughts?

  • $\begingroup$ Note that your "goal" has a typo: $t$ and $r$ should be the same. Which is a hint... $\endgroup$ – Nate Eldredge Mar 3 '18 at 23:26
  • $\begingroup$ You are correct. $\endgroup$ – cgmil Mar 4 '18 at 5:09

You're practically done. The statement $E[Z_t \exp(W_t)] = \frac{1}{\cosh(\lambda r)}$ is true for any $t$; in particular it's true when $t=r$. But $f(r) = f'(r) = 0$ so $W_r = 0$ and hence you have $E[Z_r] = \frac{1}{\cosh(\lambda r)}$ as desired.

  • $\begingroup$ Ah, way easier than I thought, but thanks for this. $\endgroup$ – cgmil Mar 4 '18 at 5:09

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.