Controlling for autocorrelation in a regression


New Member
Hi all

I have a panel data; I ran the Hausman test and it showed a significant p-value, thus I conduct my regression by using 'xtreg' command + fixed effects, vce (robust).

As I understand 'robust' option is for heteroscedasticity. I checked my variables for multicollinearity and did not find any problems.
Also I tested for autocorrelation and the result indicates that I do have serial correlation in my data.

Now I would like to control for autocorrelation, but I am stuck on this issue. What is the right way to run regression and control for serial correlation in Stata? Does this issue drastically change the output?

Many thanks for any advice!