Heavy tail distribution in terms of the mean residual hazard function (proof)

48 Views Asked by At

I've been trying to prove the next characterization and looked in several books. But all of their proofs seem a little poor. In fact, many of them say that is easy and let the proof as an exercise. But I can't figure a nice way to prove it yet:

Let $X>0$ be a random variable with probability distribution function $F(x)$ and $\overline{F}(x):=1-F(x)$. We define the "mean residual hazard function" as:

$$ \mu_F(x):=\mathbb{E}(X-x|X>x)=\int_x^\infty\frac{\overline{F}(y)}{\overline{F}(x)}dy $$

Proof that if $\displaystyle\lim_{x\to\infty}\mu_F (x)=\infty$, then $X$ is heavy-tailed.

Remember, $X$ is heavy tailed if its moment generating function $\left(\hat{m}_F (s)\right)$ satisfy:

$$ \hat{m}_F (s) = \infty \hspace{1cm} \forall s>0. $$

The statement comes from the book: "Stochastic Processes For Insurance And Finance", page 68, Tomasz Rolski

Does anyone has any idea of how to prove it?