Coefficients in pool
Posted: Thu Jul 31, 2014 11:33 am
I have estimated a pool model (essentially price with a demand proxy and supply for 61 locations). I have done it with no fixed effects, with time FE, with cross section FE and with both. The coefficients for the no FE and time FE have a cross section correction of -0.9995. For the other two models, it is closer to -0.5. Theory suggests -1 as the correlation but I would not expect anything so close in an estimation, and the two results stand out. Can you explain?