st: Robust-RE and OLS(robust, cluster):

[Stefano Li Pira <stefano.lipira@gmail.com>]

Dear Statlist,
I am doing a random-effect analysis on a unbalance panel data with 13
firms, and 4 to 22 observation per firms. The number of observation
are 181.

I'm trying to measure to estimate the two-way and three-way interaction

1) I have chosen the RE analysis because I have a time-invariant
variable, so I could not compare the FE and the RE with the standard
Hausmann procedure.
can I use a different procedure to compare them?
is there a procedure to verify the small variability  that, I think,
should confirm my choice of the RE even without the time-invariant
variable justification?

2) I decided to compare the RE-robust with an OLS(robust and cluster
errors) but the BreuschPagan result is not comprehensible. Is there
another test I can use?

Breusch and Pagan Lagrangian multiplier test for random effects

        Zmarket_share_t1[idfirm,t] = Xb + u[idfirm] + e[idfirm,t]

        Estimated results:
                         |       Var     sd = sqrt(Var)
                ---------+-----------------------------
       Zmarket~1 |          1              1
                      e |   .0198728       .1409709
                      u |          0              0

        Test:   Var(u) = 0
                             chibar2(01) =     0.00
                          Prob > chibar2 =   1.0000


Thanks

Stefano Li Pira
PhD Student
Department of Management
Università Ca'Foscari di Venezia
Fondamenta San Giobbe - 873 Cannaregio
30121 Venezia
Italy
E-mail: stefano.lipira@unive.it - stefano.lipira@gmail.com

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/