st: question on random effects estimator and linked employer -employee data

[Simone Moriconi <simone.moriconi@eco.unipmn.it>]

Dear Statalist experts,

I'm trying to estimate the wage premium produced by firm level 
collective agreement using a cross-sectional linked employer-employee 
dataset.

In the data wages are at the worker level while the variable for 
collective agreement is at the firm level. Since data are cross 
sectional I cannot add firms fixed effects; thus I cannot control for 
firms unobservables (I control of course for all observable 
characteristics of firms) and my estimates are affected by the presence 
of firm level unobserved heterogeneity.

The question is the following one: can I use a random effects approach - 
i.e. exploit the linked employer-employee nature of the data (since I do 
not have time variation) and treat firms unobserved characteristics as 
random?

I think that commands xtmixed or gllamm perform this in Stata?

Thanks a lot in advance for any suggestion.

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