4
$\begingroup$

It's a Prove/Disprove question.

Given $\mathrm{E}(Y|X)=X$ and $\mathrm{E}(X|Y)=Y$ and both $\mathrm{E}(X^2)$ and $\mathrm{E}(Y^2)$ are finite, then $$P(X=Y)=1$$

If we somehow get $\mathrm{Var}(X-Y)=0$, the above statement could be true.

So, $\mathrm{Var}(X-Y)=\mathrm{Var}(X)+\mathrm{Var}(Y)-2\mathrm{Cov}(X,Y)$

$\qquad\qquad\qquad\quad=\mathrm{E}(X^2)-[\mathrm{E}(X)]^2+\mathrm{E}(Y^2)-[\mathrm{E}(Y)]^2-2\mathrm{E}(XY)+2\mathrm{E}(X)\mathrm{E}(Y)$

Now, $-[\mathrm{E}(X)]^2-[\mathrm{E}(Y)]^2+2\mathrm{E}(X)\mathrm{E}(Y)=0$, since $\mathrm{E}(X)=\mathrm{E}(Y)$.

So, $\mathrm{Var}(X-Y)=\mathrm{E}(X^2)+\mathrm{E}(Y^2)-2\mathrm{E}(XY)$ $-(*)$

The solution given directly says the statement $(*)$ equals zero and hence the result, but \begin{align} \mathrm{E}(X^2) &= \mathrm{E}(X\cdot X) = \mathrm{E}(X\mathrm{E}(Y|X)) \\ \mathrm{E}(Y^2) &= \mathrm{E}(Y\cdot Y)\, = \mathrm{E}(Y\mathrm{E}(X|Y)) \\ \mathrm{E}(X\cdot Y) &= \mathrm{E}(\mathrm{E}(Y|X)\mathrm{E}(X|Y)) \end{align} How are they canceling each other?

$\endgroup$
2
$\begingroup$

Just note that

$\mathbf{E}(X^2)=\mathbf{E}(X\,\mathbf{E}(Y\mid X))=\mathbf{E}(\mathbf{E}(XY\mid X))=\mathbf{E}(XY)$

$\qquad\qquad=\mathbf{E}(\mathbf{E}(XY|Y))=\mathbf{E}(Y\,\mathbf{E}(X\mid Y))=\mathbf{E}(Y^2)$.

$\endgroup$
1
$\begingroup$

Hint: you have not used the hypothesis about $\mathbb{E}[X|Y]$ and $\mathbb{E}[Y|X]$ in your calculations...

$\endgroup$

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.