$X$ and $Y$ are iid r.v.s where $f(x)=\lambda e^{-\lambda x}$, $x>0$. We are given $Z=X/Y$ and asked to find the CDF and density for $Z$.
I tried doing this using the multivariate transformation formula, where
$f_{W_1,W_2}(w_1,w_2)=f_{X_1,X_2}(x_1,x_2)|J|$, but I ended up getting that $f_Z(z)=\frac{1}{(1+z)^2}$, which results in a negative CDF. I was hoping to see a different method of doing this. Any help appreciated.
Edit: Here is my work for obtaining my answer:
$Z=X/Y$; we then let $W=Y$ and $X=ZW$. The Jacobian is then calculated to be $w$. That means we have:
\begin{align} f_Z(z) & =\int_0^\infty f_X(x)f_Y(y)w\,dw \\[10pt] & =\int_0^\infty \lambda e^{-\lambda x}\lambda e^{-\lambda y}w \, dw \\[10pt] & =\lambda^2 \int_0^\infty e^{-(z+1)\lambda w}w\,dw \\[10pt] & =\frac 1 {(z+1)^2} \end{align}