enter image description here

I know, that the variance of OLS $\beta = \sigma^2 (X^TX)^{-1}$. Then I did calculations of $ (X^TX)^{-1}$ and I need the inverse in order to complete my proof. But the inverse of 3*3 matrix is a pain(especially in general case).

So I need a hint.Is there another way to prove the statement above? Or the only option is taking painful inverse of 3*3 matrix?


One hint is that $N^{-1} \sum_i (X_i^TX_i)$ converges to $\mathbb{E}(X^TX)$, which is almost the variance. Recall that in the scalar case, $V(X) = \mathbb{E}(X^2) - (\mathbb{E}(X))^2$. And since your variables are demeaned, you know something about their expectation.

Next, $X$ has three components: the first is a constant so it has zero variance and zero covariance with the other two. So there are only four relevant entries in the $3\times 3$ matrix. And when you know the correlation, $r$, you can get a little further.


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.