[Limdep Nlogit List] Two Stage Least Squares Fixed Effects Model

[Yee, Jet]

Dear Limdep list members:

 

I have a question concerning the estimator for 

Two Stage Least Squares for the Fixed Effects Model.

 

I used this estimator to estimate a model using state fixed effects and
alternatively using regional fixed effects.  My question involves the
estimates entitled

OLS Without Group Dummy Variables.           

 

I find that these estimates are different for my regional and state
fixed effects models even though they should be the same (a regression
with one common intercept term).

 

My question is why the

OLS Without Group Dummy Variables  

estimates are different.

 

The only change I made is I sorted my data by region (the data is
originally sorted by state) in order to estimate using regional fixed
effects.

 

Any insights would be appreciated.

 

Jet Yee