Page 1 of 1

De-Trending AND Time Dummy Variables ...

Posted: Fri May 22, 2009 10:40 am
by TheRocK
Hello everybody,

I have yet another econometric question. I am conducting a cross-section fixed effects regression. Amongst other variables, I have one variable which is showing a linear trend and I have three time period dummy variables which state whether the market is in a boom or bust phase. Do I need to de-trend the trending variable using an auxilliary regression prior to using it in the main regression, or are the time period dummy variables taking care of that? I de-trended the variable, and the subsequent regression is giving me nice results. Am just wondering if my approach was correct ...

Thank you very much !

All the best

Kai