Let $X\sim N(\mu_X,\sigma_X^2)$ and $Y\sim N(\mu_Y,\sigma_Y^2)$ and $Cov(X,Y)=\sigma_{XY}$. Define $Z=X+Y$.

I know that $E[X|Z=z]=\mu_X + \frac{\sigma_X^2+\sigma_{XY}}{\sigma_X^2+2\sigma_{XY}+\sigma_Y^2}(z-\mu_X-\mu_Y)$.

But what is $E[X|Z\leq z]$?

Can I just integrate over all possible realizations of $Z\leq z$ and use linearity to do the following?

$$ \begin{align*}E[X|Z\leq z] & = \mu_X + \frac{\sigma_X^2+\sigma_{XY}}{\sigma_X^2+2\sigma_{XY}+\sigma_Y^2}(E[Z|Z\leq z]-\mu_Z)\\ & = \mu_X + \frac{\sigma_X^2+\sigma_{XY}}{\sqrt{\sigma_X^2+2\sigma_{XY}+\sigma_Y^2}}\frac{\phi(\frac{z-\mu_Z}{\sigma_Z})}{\Phi(\frac{z-\mu_Z}{\sigma_Z})} \end{align*}$$


Yes, using the law of total expectation, \begin{align} E(X|Z\le z) &=E(E(X|Z)|Z\le z) \\&=E(a + bZ|Z\le z) \\&=a + bE(Z|Z\le z). \end{align}


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.