Given a bivariate normal distribution $(X_1,X_2)$ with parameters $\mu_1,\mu_2,\sigma_1,\sigma_2,\rho$ where $\rho\neq 0$.

I want to find $E(X_1X_2)$. Now $E(X_1X_2):=\int_{-\infty}^\infty \!\int_{-\infty}^\infty \! x_1x_2f(x_1,x_2) \, \mathrm{d}x_2\mathrm{d}x_1$ where $f(x_1,x_2)$ is the bivariate normal density function. Is there a short way to evaluating this integral?

  • $\begingroup$ Is integration required? Or can you use another method? $\endgroup$ – grand_chat Apr 17 '18 at 22:00
  • $\begingroup$ Is there another method? $\endgroup$ – Sid Caroline Apr 17 '18 at 22:51

If integration is not required, you could obtain $E(X_1X_2)$ by plugging into the formulae: $$ \operatorname{Cov}(X_1, X_2)=E(X_1X_2)-E(X_1)E(X_2)\tag1 $$ and $$ \rho = \frac{\operatorname{Cov}(X_1,X_2)}{\sqrt{\operatorname{Var}(X_1)\operatorname{Var}(X_2)}}\tag2$$


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.