Hausman Test - Time Constant Variable?

[joep]

Hi,

I used the Hausman Test Option in Eviews in a Random Effects Model to decide whether a Random Effects or a Fixed Effects Model is better. The test returns a p-value = 1. So the Hausman Test suggests to prefer the Random Effects Model. I was a little bit surprised to get a p-value equal to one. So I have two questions:

First:
I included a dummy variable, which is constant over time (more specific: Dummy Variable for the location of the individuals; the location cannot change). Is it possible to use the Hausman Test with a variable which is constant over time? I am not sure about this, because in a Fixed Effects Model including such a variable is not possible.

Second:
In a similar model I am also running the Hausman Test but this time without any time constant variables. The p-value of the Hausman Test is also equal to one. Is it likely to receive a p-value=1 in a Hausman Test or does this suggest any errors or false specifications?


Thanks for your help!

[joep]

In the meantime I think I figured out the answer for my first question. Eviews automatically drops the time invariant variables and estimates a new random effects model without these variables and afterwards compares the new model with the fixed effects model.