If $X$ is a continuous random variable taking non-negative values only, how do I show that $$\mathbb{E}(X)=\int_0^{\infty}[1-F_X(x)]dx$$

It's easy to give a proof if we assume the density function $f_X$ exists (use Fubini's Theorem to interchange the integrations). However, I got stuck in the general case. What if $X$ does not have a density function?


marked as duplicate by Nate Eldredge, PhoemueX, Did probability Sep 27 '15 at 19:51

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • $\begingroup$ "X has a density function" is usually the definition of "X is a continuous random variable". $\endgroup$ – Nate Eldredge Sep 27 '15 at 17:47
  • 1
    $\begingroup$ @NateEldredge: The definition I know only requires that the distribution of $X$ assigns zero measure to all singletons (i.e. the distribution function is continuous (not necessarily absolutely continuous)). $\endgroup$ – PhoemueX Sep 27 '15 at 19:09
  • $\begingroup$ Yeah, an example of a continuous CDF with no PDF is the Cantor staircase. $\endgroup$ – Did Sep 27 '15 at 19:54