Dear all:
I have a basic question regarding fixed effects modelling (but I can't find an answer in any econometrics textbooks). I am trying to run a tobit model that regresses individual poverty upon a host of individual-level controls as well as the state poverty rate.
I would like to estimate separate models for each year (I have several repeated cross-sections) and for the pooled data, (then estimate the marginal effects), in order to explore the effect of the state poverty variable over time.
In the model for the pooled data, I have added fixed effects for time and for each state. However in the yearly models, I have removed state-level dummies because their inclusion alongside the state poverty rate would induce multicollinearity.
My question is the following: Is it legitimate to run a model on the pooled data using fixed effects (for time and each state), and then to compare the coefficient on the state poverty variable with coefficients derived from cross-sections for specific years (estimated without state dummies, since their inclusion would induce multicollinearity)?
Many thanks in advance.
Regards,
Brian
____________________________________________________________________________________
Cheap talk?
Check out Yahoo! Messenger's low PC-to-Phone call rates.
http://voice.yahoo.com
*
* 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/
