I am trying to understand what I am misunderstanding when computing the variance of the OLS estimator.
Using the formula for the variance:
V($\hat\beta$) = E($\hat\beta$$\hat\beta^T$) - E($\hat\beta$)^2 = E($\hat\beta$$\hat\beta^T$) - $\beta$^2.
But the first term here gives me the variance I expect which is $\sigma^2(X^TX)^-1$ and im not sure why I have the extra term