Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Say I have the expression for $\mathrm{Cov}(X^2,Y)>0$, where both $X$ and $Y$ are known non-negative random variables. What if anything can we say about $\mathrm{Cov}(X,Y)$ without having to derive it from scratch? I think we can say that $\mathrm{Cov}(X,Y)>0$, but can we say anything more, given the expression for $\mathrm{Cov}(X^2,Y)$?

share|cite|improve this question

I think it can say very poor . For example if you take $Y=X^2$, then $Cov(X^2,Y)=1$. But if you take positive values for X then $Cov(X,Y)>0$, and for negative values for X, $Cov(X,Y)<0$. Posibly if you take any type of values, $Cov(X,Y)\cong 0$.

share|cite|improve this answer
Thanks, but note that I wrote: both X and Y are known non-negative random variables. – Omri May 31 '12 at 14:48
I was trying to find a counterexample by picking Y=c/X + X$^2$ +D. I am thinking that since we have so much freedom picking a distribution for X Let D be a large enough constant so that Y is always > 0 Then we can allow C to be negative. Part of Y is negatively correlated with X and X$^2$ and the other part is positively correlated. So I figured that I could find a distribution for X and constants C and D such that Y>0 covarince of Y with X$^2$ is positive but Cov(X, Y)<0. I was not able to do this with D=0 but think that it may be possible with large D. – Michael Chernick May 31 '12 at 17:24

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.