0
$\begingroup$

enter image description here

enter image description here

Above is the capture from the Econometrics slide from "Hill,Griths and Lim (2018) Principles of Econometrics".

I have no problem the see the blue font sentence separately, it's all about the numerator and denominator issues. But, if I think deeply about its intuitions, it bugs me.

The first blue sentence could interpret as more coverage of $x_i$ can reduce the variance

then I expand the equation, larger $\sum(x_i-\bar{x})^2=\sum x_i^2-n\bar{x}^2$ means Larger the value of$\sum x_i^2$. It's reduce variance, and this contradict to second blue sentence Clearly second blue is right, so what's go wrong with my reasoning about the larger $\sum x_i^2$ .

Please enlighten me, Thank you so much

$\endgroup$
1
  • 1
    $\begingroup$ Welcome to economics.se, please next time do not include picture of texts or equations you should rewrite them so the whole question is searchable $\endgroup$
    – 1muflon1
    Oct 11, 2021 at 13:27

2 Answers 2

3
$\begingroup$

You should be able to confirm the claim after dividing the numerator and denominator by $\sum x_i^2$. Then, ceteris paribus, a greater $\sum x_i^2$ increases the denominator and reduces the variance of the random variable $b_1$.

$\endgroup$
3
$\begingroup$

I don't like the blue sentences. More rigorously, it should state that, if denominator is fixed, an increase in numerator gives a larger variance.

That is, if $\sum_{i=1}^{n}x_{i}^{2}$ increases, meanwhile, $\bar{x}$ increases too so that the denominator remains the same, then variance becomes smaller. Without any statement about the denominator (more precisely, about $\bar{x}$), you don't know whether the variance becomes larger or smaller. What if $\bar{x}=0$, will increase in $\sum_{i=1}^{n}x_{i}^{2}$ affect the variance?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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