Let $(\Omega, \mathcal{F},P)$ be a probability space and $(X_n)$ be a positive-valued sequence of random variables on $\Omega$. We assume that $(X_n)$ converges in probability to the random variable $X$. Is it true that $$E(X)\leq \liminf_{n\to+\infty}E(X_n)$$

Thanks a lot for the help in advance!

  • $\begingroup$ Fatou's lemma does not require the convergence of the sequenze as $\liminf X_n$ converges a.s. to some r.v. Y. A fortiori $X=Y$ a.S.. $\endgroup$ – Kolmo Jan 29 '16 at 14:39

Yes, it's true. By the very definition of $\liminf$, there exists a subsequence such that

$$\liminf_{n \to \infty} \mathbb{E}(X_n ) = \lim_{k \to \infty} \mathbb{E}(X_{n_k}). \tag{1}$$

Since, by assumption, $X_{n} \to X$ in probability (hence in particular $X_{n_k} \to X$ in probability), we can choose a further subsequence of $(X_{n_k})_{k \in \mathbb{N}}$ which converges almost surely to $X$. Now the claim follows directly from Fatou's lemma and $(1)$.

  • $\begingroup$ Great! Passing from convergence in probability to (a subsequential) a.e. convergence is a part I missed. Thanks a lot! $\endgroup$ – Nikolaos Skout Jan 29 '16 at 14:01
  • $\begingroup$ I think in (1) you wanted to write somewhere liminf or not? And did you use the positivity of the $X_n$'s? Thank you. $\endgroup$ – Jimmy R. Jan 29 '16 at 14:07
  • 1
    $\begingroup$ I am sure on (1) saz meant $\varlimsup$ instead of $\lim$ on the left side of the equality. By the way, the positivity of the $X_n$'s is used, when applying Fatou's lemma. Thanks for pointing them out Jimmy. $\endgroup$ – Nikolaos Skout Jan 29 '16 at 14:16
  • 1
    $\begingroup$ @JimmyR. Yeah, exactly, thanks... $\endgroup$ – saz Jan 29 '16 at 15:24

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.