3
$\begingroup$

I am trying to understand a Lemma in Olav Kallenberg's book "Foundations of Modern Probability" (Lemma 26.19 in the second edition or 23.19 in the first edition).

The part of the lemma that I do not understand goes as follows. Let $M$ be a (not necessarily continuous) local martingale, $a \in \mathbb{R}$ and define $X = e^{M - a[M]}$. The author then claims that the following is a consequence of Itô's formula: $$X^{-1}_- \cdot X = M - (a-1/2)[M]^c + \sum \left\{e^{\triangle M - a(\triangle M)^2 }- 1 - \triangle M\right\}.$$ In trying to prove this, I proceeded as follows. Putting $f(x) = \log(x)$, we have $f(X) = M - a[M]$ and then, by Itô's formula, $$\begin{aligned} M - a[M] = f(X) &= f'(X_-) \cdot X + \frac{1}{2} f''(X_-) \cdot [X]^c + \sum\left\{\triangle f(X) - f'(X_-)\triangle X\right\} \\ &= X^{-1}_- \cdot X -\frac{1}{2} X^{-2}_- \cdot [X]^c + \sum \left\{\triangle (M - a[M]) - X^{-1}_- \triangle X \right\} \end{aligned}$$ or, rearranging, $$X^{-1}_-\cdot X = M - a[M] + \frac{1}{2} X^{-2}_- \cdot [X]^c - \sum \triangle M + a \sum \triangle [M] + \sum X^{-1}_- \triangle X.$$ Now, I noticed the following:

$\bullet\; [M] - \sum \triangle [M] = [M]^c.$

$\bullet\;$ $\sum X^{-1}_- \triangle X = \sum \left( e^{\triangle M - a \triangle [M]} - 1\right) = \sum \left( e^{\triangle M - a (\triangle M)^2} - 1\right)$.

Using these, the above becomes $$X^{-1}_-\cdot X = M - a[M]^c + \frac{1}{2} X^{-2}_- \cdot [X]^c + \sum \left\{e^{\triangle M - a(\triangle M)^2 }- 1 - \triangle M\right\}.$$ So what seems to be missing is the equality $$X^{-2}_- \cdot [X]^c = [M]^c.$$ Can anyone clarify why this is true or otherwise where I have gone wrong?

$\endgroup$

1 Answer 1

2
$\begingroup$

I think your last equality is true: in your notation, $[M]^c=[M-a[M]]^c=[f(X)]^c=[f´(X_{-})\cdot X]^c=[(X_{-})^{-1} \cdot X]^c=(X_{-})^{-2} \cdot [X]^c$ using your Ito expansion of $f(X)$. Hope it helps.

$\endgroup$
0

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.