Thank you very much for your response.  Is long-differencing supported by gretl as well?  I am trying to compare estimators for fixed effects, first differences, and long differences to check for the possibility of errors-in-variables.

Allin Cottrell wrote: 
On Tue, 30 Nov 2010, Yangbo Du wrote:

  
I am attempting to estimate a panel model consisting of 28 stacked time
series of 40 periods each; the pure rate of time preference (delta) is
the dependent variable.  Is there a way of running an instrumented
variables regression with fixed effects?  I have attached the data file
for your reference and appreciate any assistance.  Thank you very much.
    

You can use the "tsls" command.  Allow for fixed effects by
including group dummies (see the command "genr unitdum", section
5.4 in the Gretl User's Guide) or by first-differencing.

Allin Cottrell

_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users