Let $X$ be real valued random variable such that $\mathbb{E}(X) = 0$. Denote by $X^{+}$ the positive part of $X$, i.e. $X^{+} = \max(0, X)$. I am trying to show that $$\mathbb{E}(X^{+}) \leq \sqrt{\text{Var}(X)}$$.

Is my reasoning correct/could it use any improvements?

Step 1: $\mathbb{E}(X^{+}) \leq \mathbb{E}(|X|)$, and therefore it suffices to show that $\mathbb{E}(|X|)^{2} \leq \text{Var}(X)$.

Step 2: Since $\mathbb{E}(X) = 0$ we have $\text{Var}(X) = \mathbb{E}(X^{2})$ so it suffices to show that $\mathbb{E}(|X|)^{2} \leq \mathbb{E}(X^{2})$.

Step 3: By Jensen's inequality, $\mathbb{E}(|X|)^{2} \leq \mathbb{E}(|X|^{2}) = \mathbb{E}(X^{2})$

  • 1
    $\begingroup$ Just to add on: The inequality is strict since $x^2$ is strictly convex. $\endgroup$ – rookie Jul 19 '17 at 13:48

$\color{green}{\checkmark}$ Looks good to me.

Perhaps show Step $0$: $\left|\begin{align}\Bbb E(X^+) & = \Bbb E(X\mid X\geq 0)\;\Bbb P(X\geq 0)\;+\;\;\Bbb E(0\mid X<0)\;\;\Bbb P(X<0) \\ \Bbb E(|X|) & = \Bbb E(X\mid X\geq 0)\;\Bbb P(X\geq 0)+\Bbb E(-X\mid X<0)\;\Bbb P(X<0) \\ \hline \therefore \Bbb E(X^+) & \leq \Bbb E(|X|)\end{align}\right.$


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.