Issue with Levin Lin Chu unit root test and strongly balanced data

[Micha]

Hi everyone,

I am trying to apply xtunitroot llc on a strongly balanced panel dataset without gaps (according to xtset). Nevertheless, the test fails with the error message that llc requires a strongly balanced dataset.

Do the commands have different definitions of strongly balanced? Or might it be an issue that some values are empty, i.e. "."?

Any help would be most appreciated.

Thanks and best,
Micha

[Scott Merryman]

Or might it be an issue that some values are empty, i.e. "."?

Yes, it checks the balance over the restricted sample.

[Micha]

Dear Scott,

thanks a lot for your reply! Is there a pragmatic way of solving this? I assume replacing "." would not be a solution as it would influence the results of the regressions. I was focussing on LLC as it seems to be the only panel unit root test, that has H1: the variable is stationary in all panels, which i am looking for

Any help would be most appreciated!

Thanks and best,
Micha