Testing for serial correlation.

[Alejandro Cuadros]

Hey guys!
I am running a fixed effects regression and wanted to test for serial correlation. I used the commands attached in the following picture.
Can someone explain why I can't run this test? (I am very new to state and econometrics so I would appreciate taking it under consideration)
Thanks a lot for the help

[Joro Kolev]

You are not running a fixed effects regression.

From the error it seems that the test is not applicable for panel data.

For testing first order autocorrelation after fixed effects, check the user written -xtserial-, which implements a test proposed in Wooldridge, Econometrics of Cross Sectional and Panel Data.

[Alejandro Cuadros]

Thanks Joro.

[Alejandro Cuadros]

I tried with the xtserial but I don't know how to interpret the results. How do I know that there is autocorrelation?

[Joro Kolev]

It is written at the bottom: the null hypothesis of no autocorrelation is rejected at any level. So you have autocorrelation.