1
$\begingroup$

Given an unbiased estimator $\hat \theta_n$ of a parameter $\theta$, if the estimator has small variance (approaching $0$ as $n\to\infty$), it seems reasonable to expect that the estimator is consistent (i.e. that $\hat \theta_n$ converges in probability to the constant $\theta$).

Is that actually true?

$\endgroup$
1
  • 2
    $\begingroup$ yes. it follows from Markov's inequality $\endgroup$ – Anthony Quas Feb 6 '12 at 23:37
4
$\begingroup$

Markov's inequality says that for a non-negative random variable $X$ with expected value $\mu$, $$ \Pr(X>a) \le \frac \mu a. $$ (E.g. no more than $1/15$ of the population can have more than $15$ times the average income (assuming all incomes are non-negative), etc.)

So $$ \Pr(|\hat \theta_n - \theta| > \varepsilon) = \Pr((\hat\theta_n-\theta)^2>\varepsilon^2) \le \frac{\sigma_n^2}{\varepsilon^2} \to 0\text{ as }n\to\infty $$ where $\sigma_n^2=\operatorname{var}(\hat\theta_n)$.

$\endgroup$

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.