Given random variables like below:

$Y \sim \operatorname{Gamma}(a + 1, 1)$

$U_0 \sim \operatorname{Unif}(0,1)$

$U = U_0^\frac1a$

If $Y$ and $U_0$ is independent,

How can I proof $X=YU \sim \operatorname{Gamma}(a, 1)$ ?

I tried to solve this problem with theorem $f_{U,V}(u,v) = f_{X,Y}(h_1(u,v), h_2(u,v))|J|$

But I'm confusing what should be preimage.


Find the moment generating function of $YU$. Use independence to make it equal to MGF of $Y$ times MGF of $U$. Look up the gamma MGF and uniform MGF and then times them together, and see that it is also gamma.

  • $\begingroup$ Are you sure $MGF(XY) = MGF(X)MGF(Y)$? I cannot see how you could untangle exponents under integral even in the case of independent $Y$ and $Y$. It is usually stated as a fact for sum of $X$ and $Y$. $\endgroup$ – Severin Pappadeux May 21 at 22:36

Again, like in If $Y\sim\operatorname{Beta}(a,1-a)$ and $Z\sim\operatorname{Exp}(1)$, then $YZ\sim\operatorname{Gamma}(0,1)$? we will be using Mellin transform to get image of the product, using the fact, that image of the product is the product of Mellin images. For $U(0,1)^{\frac{1}{a}}$ we could easily reconstruct PDF

$$ PDF_X(x) = a x^{a-1} 1_{0<x<1} $$

Its Mellin image is

$$ M(X) = \frac{a}{s+a-1} $$

For Gamma-distribution

$$ M(Y) = \frac{\Gamma(s+a-1+1)}{\Gamma(a+1)} $$


$$ M(XY) = \frac{\Gamma(s+a-1+1)}{s+a-1} \frac{a}{\Gamma(a+1)} = \frac{\Gamma(s+a-1)}{\Gamma(a)} $$

using well-known property of Gamma-function. Which is obviously could be transformed back to

$$ PDF(x|XY) = M^{-1}\{\frac{ \Gamma(s+a-1) }{\Gamma(a)} \} = \frac{ \exp(-x) x^{a-1} }{\Gamma(a)} 1_{x>0} $$


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.