xtreg with no normality of dependent variable

[Paul Olz]

Hello everyone,

Unfortunately, I can not find clear information about the following problem. I would like to conduct a regression with the xtreg command, although my dependent variable is not normally distributed. Is this possible? And if yes, do you have a reliable source for this action?

Thank you very much in advance!
Paul

[Joao Santos Silva]

Dear Paul Olz,

The dependent variable is never required to be normally distributed, so that is not a problem. However, if your variable is bounded (e.g., non negative or between 0 and 1) you may want to use methods that take that into account.

Best wishes,

Joao

[Paul Olz]

Perfect, thank you very much for your quick answer!!!

[Joro Kolev]

Normality of the error term is only required for small sample analysis. Panel data methods are justified on large sample grounds, and you do not need normality of the error term here.