Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a question about calculating covariances of local martingales. Suppose $M_1$ and $M_2$ are local martingales. Put $M = M_1+M_2$. Is there a nice way to calculation $[M]$ in terms of $[M_1]$ and $[M_2]$?

I feel that if $B_1$ and $B_2$ are independent Brownian motion, and $M_1 = \int f(B_1(s),B_2(s),s) \, dB_1(s)$ and $M_2 = \int g(B_1(s),B_2(s),s) \, dB_2(s)$, then $[M] = \int f^2 + g^2 \, dt$.

share|cite|improve this question
@ Minke : your are implicitly supposing that $B_1$ and $B_2$ are independent, you need to add a "quadratic covariation term" to be general. Regards – TheBridge Feb 27 '12 at 8:38

To answer this question one must understand the covariation of $M_1$ and $M_2$ (see page 44 of these notes). We denote the covariation of continuous local martingales $L$ and $N$ by $[L,N]$ and, under suitable conditions, we have

$$[L,N]_t = \lim_{n\to\infty}\sum_{k=0}^{\lfloor 2^nt\rfloor}(L_{(k+1)2^{-n}}-L_{k2^{-n}})(N_{(k+1)2^{-n}}-N_{k2^{-n}})$$

So, in particular, $[M]_t=[M,M]_t$. Covariation of processes behaves in much the same way as the covariance of random variables. For example:

  • $[M_1]_t\geq 0$, but $[M_1,M_2]_t$ can be negative.
  • If $M_1$ and $M_2$ are independent then $[M_1,M_2]_t=0$, but the reverse needn't hold.
  • $[,]$ is bilinear. So $[M]=[M_1]+[M_2]+2[M_1,M_2]$. So we need to know $[M_1,M_2]$ in addition to $[M_1]$ and $[M_2]$ to determine $[M]$.
share|cite|improve this answer

Your Answer


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.