3

The problem is as follows:

Let $(W_t)$ be a Brownian Motion, $\alpha>0$, and $\tau = \inf\{t>0 : W_t \geq \alpha\}$ be the First Exit Time.

Compute $\mathbb{E}(\tau)$.

I am aware that the result is not finite, however I am having trouble showing that the integral does not converge.

This problem has answers here. However they use results I haven't seen or understand, such as Wald's Identities.

What I have done so far:

$$ \mathbb{P}(\tau \leq t) = \mathbb{P}(W_t^* \geq \alpha) = 2 \cdot \mathbb{P}(W_t \geq \alpha) = 2(1- \Phi(\alpha / \sqrt{t})) $$

Where $W_t^* = \underset{0\leq s \leq t}{\sup} W_s$.

$$ f(t) = \frac{\partial}{\partial t} \mathbb{P}(\tau \leq t) = 2\phi\left(\frac{\alpha}{\sqrt{t}}\right)\left(\frac{\alpha}{2 \sqrt{t^3}}\right) = \frac{\alpha}{\sqrt{2\pi t^3}} e^{-\frac{1}{2t}\alpha^2} $$

I am stuck here:

$$ \mathbb{E}(\tau) = \int_0^{\infty} t \cdot \frac{\alpha}{\sqrt{2\pi t^3}} e^{-\frac{1}{2t}\alpha^2} dt \overset{?}{=} \infty $$

Any help would be appreciated.

FISR
  • 151

1 Answers1

5

I do not know whether your expression is correct, but $$\begin{align} &\int_0^{\infty} t \cdot \frac{\alpha}{\sqrt{2\pi t^3}} e^{-\frac{1}{2t}\alpha^2} dt \\ \ge &\int_1^{\infty} t \cdot \frac{\alpha}{\sqrt{2\pi t^3}} e^{-\frac{1}{2t}\alpha^2} dt \qquad\text{ shortening the integral}\\ \ge &\int_1^{\infty} t \cdot \frac{\alpha}{\sqrt{2\pi t^3}} e^{-\frac{1}{2}\alpha^2} dt \qquad\text{ since } e^{-\frac{1}{2t}\alpha^2} \text{ is an increasing function of } t\\ = &\frac{\alpha}{\sqrt{2\pi}} e^{-\frac{1}{2}\alpha^2}\int_1^{\infty} \frac{1}{\sqrt{t}} dt\\ = &+\infty \end{align}$$

Henry
  • 169,616