I'm a bit confused by the following statement found in "Introductory Econometrics, A Modern Approach" (Wooldridge, 2019):

enter image description here

The phrase seems to imply the assumption is about independence and Normality. But as far as I can tell, $u \sim N(0,\sigma^2)$ already implies complete independence between $u$ and the explanatory variables. Maybe he is just being explicit about the implications of Normality without going onto the details of what implies what, details that might confuse a novice reader (this might be supported by the title of the assumption, which is "Normality" rather than "Independence and Normality").

Still, my point is: $u \sim N(0,\sigma^2)$ implies independence not only for the first two moments of $u$ but for the whole distribution. Is this the case? Am I misunderstanding Wooldridge's point?


2 Answers 2


Stating the distribution of $u$ does not imply independence.

Consider $v = 2u$. Thus, $v \sim \mathcal N (0, 4\sigma^2)$, yet $v$ is not independent of $u$.


$u\sim\mathcal N(0,\sigma^2)$ does not imply independence, as @Firebug stated. But I would add that $u|x_1,x_2,\cdots,x_k\sim\mathcal N(0,\sigma^2)$ does imply independence, since it means that the conditional distribution of $u$ is always the same, regardless of the values of the covariates.


Your Answer

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

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