Given the following random walk model $\Delta Y_t = Y_t - Y_{t-1}=\beta_0+u_t$, where $u_t \sim N(0,\sigma^2_u)$, how do we derive the standard error $\beta_0$ in terms of the estimated $\sigma^2_u$? The model is largely similar to the one given here, except that $t=1,2...,T$ and $Y_0$ is a given. I am unsure of how the degrees of freedom might change in this case.

  • $\begingroup$ If this is a random walk, you errors are iid. You expend one degree of freedeom on $\beta_0$ and that is it. $\endgroup$ Apr 27, 2021 at 17:22
  • $\begingroup$ In that case $SE(\hat{\beta}_0) = \frac{\hat{\sigma_u}}{\sqrt{n-1}}$ then, as per the linked paper? $\endgroup$ Apr 27, 2021 at 17:44


Your Answer

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

Browse other questions tagged or ask your own question.