Let $(X,Y)$ be a bivariate random variable, where $X$ is an exponential r.v. with mean $1$. Also let $ \operatorname{Cov}(X,Y) = -2$, $E[Y]=-2$, and $ \operatorname{Var}(Y) = 4$. Find the cdf of $Y$.

All I can get is $E[XY]=-4$ and $E[Y^2]=8$. How can these condition get the cdf of $Y$?

  • 2
    $\begingroup$ I don't think it is possible to determine the cdf of $Y$ from the given information. $\endgroup$ – Kabo Murphy Jan 14 at 5:31

The key observation is that $$ -\sqrt{\text{Var}(X)}\sqrt{\text{Var}(Y)}=-2=\text{Cov}(X,Y). $$This holds since $E[X^2]=\int_0^\infty x^2e^{-x}dx = \Gamma(3)=2$ and $\text{Var}(X)=E[X^2]-[EX]^2 = 1$. By Cauchy-Schwarz inequality (and its equality condition) it holds that $Y-EY=c(X-EX)$ almost surely for some constant $c\in\mathbb{R}$. We can see that $c=-2$ from $\text{Cov}(X,Y)=-2\text{Var}(X)$. This gives the distribution of $Y=-2X$ as follows. $$ P(Y\le y)=P(-2X\le y)=P(X\ge -\frac{y}{2})=\begin{cases}1,\quad y\ge 0\\e^{\frac{y}{2}},\quad y<0\\ \end{cases}. $$

  • $\begingroup$ could you explain how you use Cauchy Schwarz inequality to get the equation Y−EY=c(X−EX)? $\endgroup$ – clement Jan 14 at 23:29
  • $\begingroup$ @YibeiHe Sure. If you are not familiar with the argument, I'll directly show that. Let $X'=X-EX$ and $Y'=Y-EY$. Observe that $$ E[(Y'+2X')^2]=\text{Var}(Y)+4\text{Cov}(X,Y)+4\text{Var}(X) = 4-8+4=0.$$ This implies $Y'+2X'=0$ almost surely and the claim is proved. $\endgroup$ – Song Jan 14 at 23:33

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.