Hello,
I have run an xtabond regression and have proceeded to test for auto correlation using estat abond
I believe my results are the wrong way round.
For AR1, i have a prob>z figure of 0.0264 and for AR2 i have 0.0001
From what i can figure, that means that there is auto correlation in both lag 1 and 2 errors
I believe that is fine for lag 1, but not for lag 2, how can i over come this?
Thank you