Some ideas:
a) use -regress- with time and industry dummies to have both types of fixed effects
b) if -hausman- suggested the fixed effects, why not removing the industry dummies? Then you could create an industry identifier and cluster the standard errors on it
c) industry x dummy interactions let you evaluate how the industry effects change over time (I have seen something like this latter idea, but I can't recall where).
Nicola
At 02.33 15/02/2007 -0500, lamia chourou wrote:
>Dear statlist users,
>I have a panel data model that includes time invariant
>variables (dummy variables controlling firm industry).
>The Hausman test suggests using the fixed effect
>model. As this model cannot estime time invariant
>variables, I tried to use Hausman Taylor model however
>all my time invariant variables are exogenous. I would
>like to know wich model is suitable in this case?
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
