1
$\begingroup$

Let $X_n^m$ be a double sequence of integrable random variables such that $X_n^m \rightarrow_n X^m$ in probability, where $X^m$ is another sequence of integrable random variables. Assume that such convergence is uniform in $m$ and that $X_{n}^m \rightarrow_n X^m$ in $L^1$.

Does it follow that $X_{n}^m \rightarrow_n X_m$ in $L^1$ uniformly in $m$?

Thanks

$\endgroup$

1 Answer 1

1
$\begingroup$

It does not follow from the assumptions: take $X_{m,n}:=mn\cdot \chi_{(0,n^{-2})}$ on $(0,1)$ endowed with Lebesgue measure. We have for $\varepsilon\lt 1$ and each integer $n$, $$\sup_{m\in\mathbb N}\mathbb P\{|X_{m,n}|\gt \varepsilon \}=n^{-2},$$ hence there is uniform convergence in probability to $0$. We also have $\mathbb E[X_{m,n}]=\frac mn$, hence we have convergence in $\mathbb L^1$ to $0$, but this convergence is not uniform in $m$.

$\endgroup$

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.