What are the necessary conditions for groups to be pooled for OLS I have limited duration of sales for multiple regions of a country and need to run pooled regression on it. I want to know the necessary conditions for the data to be pooled. The sales are not divided proportionately and the independent variables too do not have similar variation in the different regions. 
Is it the correct approach? or should I consider FE or RE.. I am limited by the number of time periods of data and hence was considering pooled. let me know.
Thanks 
 A: Wooldrige, in his Econometric Analysis of Cross Section and Panel Data, pps. 4-6, discusses the assumptions of panel data models as follows:

  
*
  
*A stochastic setting
  
*Assumption of random sampling
  
*A population model has been specified
  
*An i.i.d. sample drawn from that population
  

And goes on to note, 1) the virtues of random sampling vs experimental data (which would dictate an FE model), and 2) that the iid assumption is not likely to hold for pooled cross sections. In place of this latter assumption, he introduces independent, not identically distributed (i.n.i.d.) observations...as well as a focus on asymptotic properties of the estimators as opposed to finite sample properties.
Wrt your question of "correctness," I think it's too easy to introduce unnecessarily limiting factors into a statistical analysis. To the best of my knowledge, there is no statistical rulebook with policemen monitoring its implementation. This is a trap many grad students will fall into, many times as a consequence of harsh or even punitive advisors. In other words, there is more than one way to skin a cat and the approach you've proposed is as good as any.
Other references include Lee Cooper's Market Share Analysis, which is somewhat dated (1989) but remains an excellent introduction to this class of models in marketing science. He treats them as attraction models, after theoretical work done by Philip Kotler, as well as variants of choice models. His book is available for free download from his UCLA website at http://www.anderson.ucla.edu/faculty/lee.cooper/MCI_Book/BOOKI2010.pdf 
