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

The wikipedia definition of correlation is $$ \rho_{X,Y} = \frac{\mathrm{Cov(X,Y)}}{\sigma_X\sigma_Y} $$ where $X,Y$ are random variables with standard deviations $\sigma_X, \sigma_Y$ respectively. Of course this is nonsense if $X=0$ almost surely. Just to make sure, we define $\rho_{X,Y} =0$ in this case, correct? Or is there a broader definition of $\rho_{X,Y}$ which does not involve division by $\sigma_X$?

share|cite|improve this question
up vote 0 down vote accepted

I have only encountered the correlation between non-degenerate random variables. The definition i'm aware of:

Let $X$ and $Y$ be real non-degenerate random variables from $\mathcal{L}^2(P)$. Then the correlation between $X$ and $Y$ are defined to be $$ \rho(X,Y)=\frac{\mathrm{Cov}(X,Y)}{\sqrt{\mathrm{Var}(X)\mathrm{Var}(Y)}}. $$

Note that $X$ is degenerate if and only if $\mathrm{Var}(X)=0$. I suppose the important property of the correlation is the following:

Let $X$ and $Y$ be non-degenerate random variables from $\mathcal{L}^2(P)$. Then the following are equivalent:

  1. $\rho(X,Y)\in\{-1,1\}$,

  2. There exists $\alpha,\beta\in\mathbb R$ such that $Y=\alpha+\beta X$ a.s.

By allowing degenerate random variables into the definition with $\rho=0$, the above result is no longer true. If $X=c_1$ and $Y=c_2$ a.s. then $Y=\alpha+\beta X$ a.s. with $\alpha=0$ and $\beta=\frac{c_2}{c_1}$. I would probably just stick with the definition for non-degenerate variables.

share|cite|improve this answer

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.