2
$\begingroup$

Hello Dear fellows,

I thank you in advance for your help and ideas.

I have just read an article and want you to help me understand the rational behind a part of it.

We have two processes $v_t$ and $V_t$

such that:

$dv_{t}=\omega v_{t} dW_{t}$ $W_{t}$ is Brownian and $V_{t}=\sqrt{\frac{1}{T} \int_0^T {v_{t}}^{2}\,dt}$

We know that $v_{t}=(law)v_{0}\exp(\omega z \sqrt{T}-\frac{\omega^{2}T}{2})$ $z$ is a gaussian with mean 0 and std 1.

The next statement is the one that i did not fully undestand the way they authors did have it :

They state $V_{T}=(law)\exp(\omega z_{1} \sqrt{\frac{T}{3}}-\frac{\omega^{2}T}{6})$ ou $z_{1}$ is a gaussian with mean 0 and std 1 and $\rho (z,z_{1})=\frac{\sqrt{3}}{2}$

I have tried to approximate $V_{T}$ by $\frac{1}{T} \int_0^T v_{t}\,dt$ (We know that $\frac{1}{T} \int_0^T v_{t}\,dt \leq V_{T}$) But I did not obtain any pertinent result this way.

Does someone have an idea how to prove the law equality above?

Many thanks

$\endgroup$
2
  • 1
    $\begingroup$ Hi AlGoRiS, Could you provide the reference of the paper your problem is coming from ? Regards $\endgroup$
    – The Bridge
    Jul 2, 2010 at 11:14
  • 2
    $\begingroup$ The "ou" is French for "where"? Is there a "where" missing in the expression for $v_t$ and for $dv_t$ as well. You should check your equations. $\endgroup$
    – Peter Shor
    Aug 18, 2010 at 11:12

1 Answer 1

1
$\begingroup$

This is wrong. Let for simplicity $\omega = 1$. Write $$ E[V_T^2] = \frac 1T \int _0^T E[\exp(2W_t-t)]dt = 1/T\int_0^Te^{t} dt = (e^{T}-1)/T. $$ This is inconsistent with their claim.

(In fact, by considering expectations of greater powers of $V_T$, one can see that $V_T$ cannot be log-Gaussian with any parameters.)

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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