Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Suppose a random variable $X_n$ converges in distribution to X and $E[X]< \infty$,
Is this statement true or is there any condition to satisfy this?

$ plim_{n\to \infty}Var(X_n) = Var(X) $ (or $lim_{n\to \infty}Var(X_n) = Var(X)$ )

In fact, I would like to know whether the (unconditional) variance of A=$\sqrt n(\hat\beta -\beta )$ converges in probability to the variance of the asymptotic distribution of A for the OLS estimator $\hat\beta$ in the linear regression model with stochastic regressors when the regressor X and the error term are independent.

e.g) $y_i = x_i\beta + e_i$ , $e_i$ ~ i.i.d $( 0, \sigma^2) $, $plim{{\sum x_i^2} \over n}=E(x_i^2)<\infty$ , and x and e are independent.

I thought that Var(A) = $\sigma^2 E[({{\sum x_i^2} \over n})^{-1} ] $ and A $\to^D $ N(0 , $\sigma^2 [E(x_i^2)]^{-1}$)
I would like to know whether $plim_{x\to \infty}Var(A)=\sigma^2 (E(x_i^2))^{-1}$

share|improve this question

1 Answer 1

up vote 0 down vote accepted

No, in general, $X_{n}$ converges in disn to $X$, and $E \vert{X}\vert < \infty$ does not imply that $Var(X_{n})$ converges to $Var(X)$. You also need to have uniform integrability of $\vert X_{n}\vert^{2}$

share|improve this answer
    
Thanks for the explanation. I have another question. Doesn't Var($X_n$) converge to Var(X) in the example above also? I thought Var(A) $\to$ $\sigma^2$ $(E(x_i^2))^{-1}$ if Var(A)<$\infty$. But I don't know whether it's true or not. –  wikizero Sep 29 '12 at 6:54
    
You would have to see whether the sequence of random variables $n(\hat{\beta}-\beta)^{2}$ is uniformly integrable or not. –  Ken Dunn Sep 29 '12 at 9:03

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.