Expected Value and Variance of a process (Itô's lemma, brownian motion)

733 Views Asked by At

I have a short question regarding this problem:

$B$=$B_t$ is standard Brownian motion
Process $q$ = $q_t$ defined by $q_t$ = $B_t^2$-t

Show that $E$[$q_t$]=$0$ and $Var$[$q_t$]=$2t^2$

What am I looking for here? I tried some different approaches but I'm not really sure how to solve this. I've been trying since 3-4 days which is rather frustrating considerung that the solution will probably have like 2 lines :D

Well anyways, I'm not necessarily looking for the solution straight away but if anyone could guide me into the right directiong that would be immensly helpful! Thanks

1

There are 1 best solutions below

2
On BEST ANSWER

What do you know about a Brownian motion? $B_t = B_t - B_0$ is an increment so its $\mathcal{N}(0,t)$ distributed

So $E[B_t] = 0$ and $Var(B_t) = E[B_t^2] - E[B_t] = E[B_t^2] = t$

So you have: $E[q_t] = E[B_t^2] - t = t-t=0$

Because you know the distribution of $B_t$ you know it's density $f_{B_t}$ and can calculate $E[g(B_t)]$ for any measurable function by $$\int_{-\infty}^\infty g(x) f_{B_t}(x) dx$$

Reacognising $Var(q_t) = Var(B_t^2 - t) = Var(B_t^2) = E[B_t^4] - E[B_t^2]^2 = E[B_t^4] - t^2$ gives you the variance by calculation $E[B_t^4]$ what you can do as described above setting $g(x) = x^4$ (or checking Wikipedia for the 4th moment of a normal rv)