st: Re: comparison of OLS and fixed-effect regression coefficients

[Kit Baum <baum@bc.edu>]

But if the test that Martin mentions rejects its null, you have  
biased and inconsistent coefficients in the OLS regression.

What you can do to jointly test _both_ coefficients for differences is

webuse grunfeld
reg invest mvalue kstock
est store ols
xtreg invest mvalue kstock, fe
est store fe
hausman fe ols, sigmaless

The Hausman test in this context considers a consistent estimator  
(fe) vs a counterpart which would be efficient if the constraints  
that dummies were excluded from the fe model were appropriate. A  
strong rejection is based on the differences in the point and  
interval estimates.

Kit

Kit Baum, Boston College Economics and DIW Berlin
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html


On Feb 2, 2008, at 2:33 AM, Wencke wrote:


> This is not what I was looking for. I would like to compare the
> coefficients of 2 specific dependent variables and not the whole
> regression model.
> Let's assume the following model that I estimate twice, one OLS and  
> one FE:
> y =  b1x1 + b2x2 + ... + bnxn
>
> Then, I would like to test both b1 coefficients of x1 if they are
> significant different from each other.
> Hypothesis:  b1(OLS) ~= b1(FE).
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/