14
$\begingroup$

Let $X$ and $Y$ be real-valued random variables defined on the same space. Let's use $\phi_X$ to denote the characteristic function of $X$. If $\phi_{X+Y}=\phi_X\phi_Y$ then must $X$ and $Y$ be independent?

$\endgroup$
4
  • 4
    $\begingroup$ No, as the case $X=Y=$ standard Cauchy shows. $\endgroup$
    – Did
    May 1, 2013 at 7:58
  • $\begingroup$ @Did why this is a counterexample? $\endgroup$
    – Connor
    Dec 12, 2016 at 13:19
  • $\begingroup$ @Connor Did you try to compute the characteristic function of a standard Cauchy random variable? If you did, this should be clear. If you did not, what are you asking? $\endgroup$
    – Did
    Dec 12, 2016 at 13:43
  • $\begingroup$ I flagged Connor's question as a duplicate but ended up expanding your comment to an answer, @Did $\endgroup$
    – Therkel
    Dec 12, 2016 at 14:02

2 Answers 2

23
$\begingroup$

As user75064 already pointed out, the answer is "no". However, there is the following result:

Let $X,Y$ be $\mathbb{R}^d$-valued random variables. Then the following statements are equivalent.

  1. $X,Y$ are independent
  2. $\forall \eta,\xi \in \mathbb{R}^d: \mathbb{E}e^{\imath \, (X,Y) \cdot (\xi,\eta)} = \mathbb{E}e^{\imath \, X \cdot \xi} \cdot \mathbb{E}e^{\imath \, Y \cdot \eta}$

i.e. if the characteristic function of the random vector $(X,Y)$ equals the product of the characteristic function of $X$ and $Y$, then $X$ and $Y$ are independent (proof).

$\endgroup$
2
  • $\begingroup$ the small letter that you have in the exponent is the imaginary number $i$, right? $\endgroup$ May 27, 2016 at 20:25
  • $\begingroup$ @CarlosMendoza Yes.... $\endgroup$
    – saz
    May 28, 2016 at 5:33
6
$\begingroup$

No. The property in your post is called subindependence, and it is strictly weaker than independence. (Note that some people use the term "subindependent" as a synonym for "uncorrelated".) In addition to the references given in Wikipedia, you can find an example in this short note. Unfortunately it's behind a paywall. The example consists of two random variables with joint pdf $$h(x,y)=f(x)f(y)(1+\cos x\cos 3y)$$ where $$f(x)=C\left(\int_0^{1/2} \exp(1/(4s^2-1))\cos (sx)\,ds\right)^2$$ with appropriate normalizing constant $C$.

$\endgroup$

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.