How to compute expectation for the product of squared jointly normal random variables

521 Views Asked by At

Let $X_1$ and $X_2$ be jointly normal random variables with $\mathbb E[X_1] = \mathbb E[X_2] = 0$ and $\operatorname{var}(X_1) = \sigma_1^2$, $\operatorname{var}(X_2) = \sigma_2^2$. The correlation between them is $\rho$.

Find $\mathbb E\left[\left(X_1^2\right)\left(X_2^2\right)\right]$.

I know that $\rho = \frac {\mathbb E[XY]} {\sigma_1 \sigma_2}$, but I'm not sure how this equation changes when $X$ and $Y$ are squared.

Any help would be greatly appreciated!

1

There are 1 best solutions below

0
On

When the $X_1$and $X_2$ are jointly normal the covariance between these two would be $2{\rho}^2$

The proog is here: covariance of two squared (not zero mean) random variables

If you don't like the proof here you an do something like this

The covariance of two squared normal variables

Then it becomes easy

$$\mathop{\mathbb{E}}\left [ {X_{1}}^{2}{X_{2}}^{2} \right ] = Cov \left( {X_1}^2 ,{X_2}^2 \right) + \mathop{\mathbb{E}}\left [ {X_{1}}^{2} \right ]\mathop{\mathbb{E}}\left [{X_{2}}^{2} \right ] = 2{\rho}^2 +{\sigma_{1}}^{2} {\sigma_{2}}^{2} $$

Hope it helps you.