Hi I'm currently working through past exam questions and am stuck with the following question:

Random variables $X_1$, $X_2$ and $X_3$ are identically distributed, with common mean $\mu$, common variance $\sigma^2$, and common pairwise correlation $\rho$ (i.e. $corr(X_i,X_j) =$ $\rho$ for all i $\neq$ j). Let Y = $X_1$ - $2X_2$ + $X_3$.

a) Fin the expectation and variance of Y.

b)Find the correlation between Y and $X_1$

Am I wrong in assuming that the expectation of Y would simply be $E(X_1) - E(2X_2) + E(X_3)$ given that E(X+Y)=E(X)+E(Y). Am I completely off? Any hints would be greatly appreciated.

  • $\begingroup$ Yes, you are on the right track with the expectation. You should also know that $E(aX) = aE(X)$. For the variance of Y, remember the formula $Var(aX + bY) = a^2Var(X) + b^2Var(Y) + 2abCov(X, Y)$. $\endgroup$ – James Jun 11 '15 at 11:03

You're right. Note though that since those variables are identically distributed, $E(X_1) = E(X_2) = E(X_3) = \mu$ so in the end $E(Y) = 0$.

For the variance you have a different formula (which you have to prove:)

$$Var(X_1 + X_2) = Var(X_1) + Var(X_2) + 2\text{Cov}(X_1, X_2)$$

Use this formula to find the variance of $Y$.

  • $\begingroup$ Thanks for a swift response! Am I correct in thinking that the: Var($X_1$ - $2X_2$ + $X_3$) = Var($X_1$) + 4*Var($X_2$) + Var($X_3$) - 4Cov($X_1$,$X_2$) + 2Cov($X_1$,$X_3$) - 4Cov($X_2$,$X_3$) ? $\endgroup$ – orwellian Jun 11 '15 at 11:57

You are correct about the expectation of $Y$ so you find easily that $\mathbb EY=0$.

Advice on the rest:

First solve it for special case $\mu=0$ and $\sigma=1$.

Then realize that in the general case (co)variances are multiplied with $\sigma^2$ and that correlations are not affected.


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.