Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Testing equality of coefficients between two linear random effects models


From   "Johannes Metzler" <johannes.metzler@campus.lmu.de>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Testing equality of coefficients between two linear random effects models
Date   Thu, 24 Jul 2008 20:48:42 +0200

Dear Stata list users,
 
I have the following situation: I estimate two models with Random Effects
(xtreg, re) over the same sample: 
 
y1_it = c1 + 1*x1_it + e1_it
y2_it = c2 + 2*x1_it + e2_it
 
i.e., the one explanatory variable x1 is the same in both models for all
units, whereas the dependent variable is different. Imagine x1 to be a shock
in time t in municipality i, and y1 for example the share of municipal
inhabitants with characteristic 1, y2 the share of municipal inhabitants
with characteristic 2. 
 
Task: I would like to test if 1=2. 
It it possible, and if so, how?
 
I imagine that the solution would entail estimating both equations in the
same model in order to obtain a common covariance matrix (something along
the lines of
- duplicating all observations
- gen y3=y1 if original, y3=y2 if duplicate
- estimating y3_it = dummy_orig + dummy_duplic + 1*x1*dummy_orig +
2*x2*dummy_dupl + e_it
?)
 
but I do not know how to implement this in a random effects context and
where the caveats are. 
I would greatly appreciate any help.
 
Thank you,
Johannes
 


*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index