Let us consider a process $(X_t)_{t \geq 0}$ which follows a Geometric Brownian Motion (GBM) $-$ assuming both $\mu>0$ and $\sigma>0$:

$$ \begin{align} & dX_t = \mu X_tdt+\sigma X_tdW_t \\[3pt] & X_0 = x_0 \end{align}$$

Letting $w>0$, we define the hitting time of the Brownian motion $W_t$ as:

$$ \tau_W = \min\{t \geq 0: W_t=w\}$$

Letting $\lambda>0$, it is well known that the Laplace transform of the hitting time is given by:

$$ \mathbb{E}\left[e^{-\lambda\tau_W}\right] = e^{-w\sqrt{2\lambda}}$$

Now, letting $x>0$, we define the hitting time of the GBM $X_t$ as:

$$ \tau_X = \min\{t \geq 0: X_t=x\}$$

I am looking for a closed-form formula for the Laplace transform of $\tau_X$ with respect to the drift term:

$$ \mathcal{L}_{\mu}(\tau_X)=\mathbb{E}\left[e^{-\mu \tau_X}\right] $$

Building on the proof for the Laplace transform of the plain Brownian motion, up to now I have tried the following. Given:

$$ X_t = x_0e^{\left(\mu - \frac{\sigma^2}{2}\right)t+\sigma W_t}$$

I have defined the following drifted Brownian motion:

$$ B_t= W_t+\left(\mu - \frac{\sigma^2}{2}\right)\frac{t}{\sigma} $$

So that:

$$ \tau_X = \min\{t \geq 0: B_t=\frac{1}{\sigma}\log \frac{x}{x_0}\}$$

Letting $\tau = \tau_X$, I then have considered the auxiliary process $-$ which is a martingale:

$$ Y_t = e^{\sigma B_t - \mu t} = e^{\sigma W_t - \frac{\sigma^2}{2}t}$$

As well as the stopped process $Y_{t \wedge \tau}$. Given $Y_t$ is a martingale, so does the stopped process, from which we can conclude that:

$$ \mathbb{E}[Y_{t \wedge \tau}] = \mathbb{E}[Y_0] = 1$$

From here, analysing the behaviour of $Y_{t \wedge \tau}$ when $t \rightarrow \infty$ and using the dominated convergence theorem, I got:

$$ \lim_{t \rightarrow \infty}\mathbb{E}[Y_{t \wedge \tau}] = \mathbb{E}\left[\lim_{t \rightarrow \infty}Y_{t \wedge \tau}\right] = \mathbb{E}\left[\mathbf{1}_{\{\tau \, < \, \infty\}}\left(e^{\sigma B_{\tau} - \mu \tau}\right)\right] = 1$$

Thus by definition of $\tau$:

$$ \mathbb{E}\left[\mathbf{1}_{\{\tau \, < \, \infty\}}e^{- \mu \tau}\right] = \frac{x_0}{x} $$

However, I get stuck here as I cannot get rid of the indicator function.

Any hints/ideas on how to continue from here? Alternatively, is there any general closed-form formula for the Laplace transform of the GBM hitting time in all generality?

Any reference would be greatly appreciated.

  • 1
    $\begingroup$ Assuming $\mu>0$, the indicator takes care of itself because $e^{-\mu\tau}=0$ on $\{\tau=\infty\}$. $\endgroup$ Jul 1, 2017 at 16:27
  • $\begingroup$ @JohnDawkins so are you saying I can simply state that: $\mathbb{E}[\mathbf{1}_{\{\tau \, < \, \infty\}}e^{-\mu \tau}] = \mathbb{E}[e^{-\mu \tau}] $? $\endgroup$ Jul 1, 2017 at 16:53
  • $\begingroup$ @JohnDawkins There is something that bugs me, I have the impression that then I can both show that $\mathbb{E}[e^{-\mu \tau}] = x_0/x$ as you said, as well as $\mathbb{E}[\mathbf{1}_{\{\tau \, < \, \infty\}}] = x_0/x$ by letting $\mu \rightarrow \infty$ and invoking some theorem like dominated or monotone convergence. Where is the problem here? $\endgroup$ Jul 1, 2017 at 16:57
  • 1
    $\begingroup$ Don't forget that $\tau$ itself depends on $\mu$! $\endgroup$ Jul 1, 2017 at 17:11
  • $\begingroup$ Oh oh oh, true that! I hadn't realised it. $\endgroup$ Jul 1, 2017 at 17:13

1 Answer 1


Let $W$ be standard Brownian motion and $Y$ geometric Brownian motion, i.e., for $t\geq 0$ $$ Y_{t}=y_{0}\exp (\alpha t+\sigma W_{t}), $$ where $y_{0}\in \mathbb{R}_{++}$, $\alpha \in \mathbb{R}$, and $\sigma \in \mathbb{R}_{++}$.

The Handbook of Brownian Motion by Borodin and Salminen says that for any $y\geq y_{0}$ and $r>0$ $$ \Bbb{E}[e^{-r\tau _{y}}]=\left(\frac{y_{0}}{y}\right)^{\kappa}, $$ where $\tau_y$ is the time of the first transition to $y$ and $$ \kappa =\left(\sqrt{\alpha ^{2}+2r\sigma ^{2}}-\alpha \right)\sigma ^{-2} $$ is a strictly positive constant.

Which I think means you are close.


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.