Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

How do you show the following process is a martingale? My notes say it is a martingale by I can't work it out.

$$ E[e^{\sigma B(t) - \frac{\sigma ^2 t}{2}} | \mathscr{F}(s)] $$

I tried to multiply and divide by $e^{\sigma B(s) - \frac{\sigma ^2 s}{2}}$ but got stuck.

share|improve this question
add comment

1 Answer

The $\sigma^2t$ term is deterministic hence one can move it out of the conditional expectation. The $B(t)$ term is $(B(t)-B(s))+B(s)$. The $B(s)$ term is $\mathscr F_s$-measurable hence one can move it out of the conditional expectation. The $B(t)-B(s)$ term is independent of $\mathscr F_s$ hence one can integrate it. To summarize, $$ \mathbb E(\mathrm e^{\sigma B(t)-\sigma^2t/2}\mid\mathscr F_s)=\mathbb E(\mathrm e^{\sigma\cdot (B(t)-B(s))})\,\mathrm e^{\sigma B(s)}\,\mathrm e^{-\sigma^2t/2}. $$ The next step is to compute $\mathbb E(\mathrm e^{\sigma\cdot (B(t)-B(s))})$. To do that, one needs to identify the distribution of $B(t)-B(s)$, and maybe you can proceed from there.

Edit: To be honest, I fail to grasp how one can ask questions about Brownian martingales and be unable to compute basic Gaussian integrals... but here we go: the random variable $\sigma\cdot (B_t-B_s)$ is centered normal with variance $\sigma^2\cdot (t-s)$ and, for any standard normal random variable $X$ and any real number $u$, $$ \mathbb E(\mathrm e^{uX})=\int_{-\infty}^{+\infty}\frac1{\sqrt{2\pi}}\mathrm e^{ux}\mathrm e^{-x^2/2}\mathrm dx=\int_{-\infty}^{+\infty}\frac1{\sqrt{2\pi}}\mathrm e^{u^2/2}\mathrm e^{-(x-u)^2/2}\mathrm dx=\mathrm e^{u^2/2}. $$ In particular, for $u=\sigma\cdot \sqrt{t-s}$, $$ \mathbb E(\mathrm e^{\sigma\cdot (B(t)-B(s))})=\mathrm e^{\sigma^2(t-s)/2}. $$

share|improve this answer
    
Ok so its normally distributed with mean 0 and var t-s. How do you solve $E[e^{\sigma (B(t) - B(s))}]$? –  riotburn Dec 20 '12 at 7:36
    
You know the distribution of $B(t)-B(s)$ but you do not know how to compute $E(e^{\sigma(B(t)-B(s)})$? –  Did Dec 20 '12 at 7:41
    
$\int e^{\sigma (x-y)} \frac{1}{\sqrt {2 \pi (t-s)}} e^{-\frac{(x-y)^2}{2(t-s)} }dy$ ? Honestly not sure, the exponential is throwing me off. –  riotburn Dec 20 '12 at 7:53
    
What you write has little to do with the question. See Edit. –  Did Dec 20 '12 at 8:35
    
thanks for your help but I fail to grasp why you fail to grasp that someone asking for help on a math help forum is having trouble with a math question... –  riotburn Dec 20 '12 at 8:46
show 3 more comments

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.