Hi everyone this is my first question here. I have the following question on my exercise sheet: let $B$ be a brownian motion, $\phi$ a progressively measurable function such that $E \int_0^T \phi_t dt \lt\infty $, $\forall T>0$. Set $X_t:= \int_0^t \phi_s dB_s$. Compute $\left<X,B\right>_t$, i.e. the quadratic covariation of $X_t$ and $B_t$. (the space is endowed with a filtration $\left(F_t\right)_{t\geq 0}$ and the brownian motion is adapted w.r.t. to it)

$\boldsymbol{First}$ $\boldsymbol{attempt}$: let $\Pi_n $:=$\{0=t_0<\ldots<t_n= t \}$ be a partition of $[0,t]$. $$\left<X,B\right>_t:= \lim_{|\Pi_n|\to0}\sum_{i=0}^{n-1}[X_{t_{i+1}}-X_{t_{i}}][B_{t_{i+1}}-B_{t_{i}}]=\\ \lim_{|\Pi_n|\to0}\sum_{i=0}^{n-1}X_{t_{i+1}}[B_{t_{i+1}}-B_{t_{i}}]-X_{t_{i}}[B_{t_{i+1}}-B_{t_{i}}]. $$ The second term should converge to $\int_0^t X_s dB_s$ in $L^2$, but i don't know what to do with the first term since it cannot converge to the Ito integral of $X$ w.r.t. the brownian motion.

$\boldsymbol{Second}$ $\boldsymbol{attempt}$: I know that $$\left<X,B\right>_t= \frac{1}{4} (\left<X+B\right>_t-\left<X-B\right>_t).$$ Now, since $\left<X+B\right>_t$ is the unique process s.t. $\left(X_t+B_t\right)^2-\left<X+B\right>_t$ is a martingale, if i compute $E[X_t^2+B_t^2+2X_tB_t|F_s]$, $s<t$ i will be done.

I know that $E[B_t^2|F_s]=B_s^2+t-s$.

My strategy for $2E[X_tB_t|F_s]$ is to consider the simple process $\phi^{Q}_t=\sum_{k=0}^{m-1}\phi_{t_{k}}\mathbb{1}_{[t_{k},t_{k+1}]}(t)$, where Q is a partition of $[0,t]$ and to compute $2E[\sum_{k=0}^{n-1}\phi^Q_{t_k}\left(B_{t_{k+1}}-B{t_k}\right)B_t|F_s]$ but i cannot proceed further.

Since this is homework i ask only for a hint


1 Answer 1


There is the following general statement which is helpful for the second approach

Let $\psi$ be a progressively measurable function such that $\mathbb{E}\int_0^T \psi(t)^2 \, dt < \infty$ for all $T>0$. Then the quadratic variation of $$Y_t := \int_0^t \psi_s \, dB_s \tag{1}$$ equals $$\langle Y \rangle_t = \int_0^t \psi^2(s) \, ds.$$

Hint: Write $$X_t + B_t \qquad \text{and} \qquad X_t -B_t$$ as a stochastic integral of the form $(1)$ and apply the above statement to calculate $\langle X+B \rangle$ and $\langle X-B \rangle$, respectively.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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