That's *NOT* an inequality, it's an equality. There's a clear "$=$" sign between $\mathbb{E}X$ and the $\int$.
– Glen_bOct 18 '13 at 19:43
2
Use Fubini. Then, pause. Then, ask yourself why the argument doesn't work, in general, if $X$ can take both positive and negative values.
– cardinalOct 18 '13 at 20:01
1
See discussion of the result [here](http://stats.stackexchange.com/questions/18438/does-a-univariate-random-variables-mean-always-equal-the-integral-of-its-quanti) and [here](http://math.stackexchange.com/questions/64186/intuition-behind-using-complementary-cdf-to-compute-expectation-for-nonnegative) and [here](http://math.stackexchange.com/questions/64186/intuition-behind-using-complementary-cdf-to-compute-expectation-for-nonnegative). I've only heard it expressed as 'the expectation is the integral of the survival function' ... rather than with any particular name.
– Glen_bOct 18 '13 at 20:04
1
Incidentally, for continuous random variables, you can show it in two lines using integration by parts.
– Glen_bOct 18 '13 at 20:14