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.

Does consistency imply asymptotically unbiasedness? I know the statement doesn't work in the other direction. My guess is it does, although it obviously does not imply unbiasedness.

I think it wouldn't be too hard if one digs into measure theory and makes use of convergence in measure. But I have a gut feeling that this could be proved with only elementary probability theory.

But can anybody give me a proof? Thanks!

share|improve this question
add comment

1 Answer

up vote 2 down vote accepted

Let $Y_n$ take the value $0$ with probability $(n-1)/n$ and take the value $n$ with probability $1/n$. Then $(Y_n)$ is a consistent sequence of estimators for zero but is not asymptotically unbiased: the expected value of $Y_n$ is $1$ for all $n$.

If we assume a uniform upper bound on the variance, $\mathrm{Var}(Y_n-X)\leq \mathrm{Var}(Y_n)+\mathrm{Var}(X)<C$ for all $n$, then consistency implies asymptotic unbiasedness. Let $Y_1,Y_2,\dots$ be a consistent sequence of estimators for a random variable $X$. This means that for all $\epsilon>0$ the probability of the event $|Y_n-X|>\epsilon$ tends to zero as $n\to\infty$. We want to show that $Y_1,Y_2,\dots$ is asymptotically unbiased: that the expected value of $Y_n-X$ tends to zero as $n\to\infty$. This follow from Cauchy-Schwarz for example: $$\mathbb E[|Y_n-X|]\leq \epsilon+\sqrt{\mathbb E [|Y_n-X|^2]\mathbb{P}[|Y_n-X|\geq\epsilon]}\to 0.$$

It is also useful to note that by Chebyshev's inequality, if the variance tends to zero then asymptotic unbiasedness implies consistency.

share|improve this answer
add comment

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.