Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: 2-way Fixed Effects model


From   Ulrich Kohler <kohler@wzb.eu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: 2-way Fixed Effects model
Date   Wed, 05 Sep 2012 10:03:48 +0200

Another way would be

. iis company
. xtreg y x* i.county, fe
. predict u, u

Variable u then holds the coefficients of the dummy-variables for
company centered at their mean. Note that this is just another way to
access the dummy variable coefs. Austins remark concerning the
reliability of the coefs remains to be true, therefore. 

I don't know waht the "dummy variable trap" is.

Uli

Am Dienstag, den 04.09.2012, 17:19 -0700 schrieb Lucas:
> Austin,
> 
> Thanks for your note.  I know I can do this for a one-way OLS fixed
> effects model with the -noconstant- option. However, in a two-way
> fixed effects model this should produce a dummy variable.  That's what
> I expect with the reg command.  Am I mistaken?  Also, does the iv2reg
> prevent the dummy variable trap?
> 
> Sam
> 
> On Tue, Sep 4, 2012 at 10:13 AM, Austin Nichols <austinnichols@gmail.com> wrote:
> > Lucas <lucaselastic@gmail.com> :
> > Just use -regress- e.g.
> >   reg y x* i.company i.county
> > but you will need many obs per company-cluster pair to trust those coefs.
> >
> > Also see -ivreg2- on SSC for one implementation of 2-way cluster-robust SE
> > [ http://www.stata.com/meeting/boston10/boston10_baum.pdf ]
> > and compare SEs on x vars (not SE on fixed effects)
> > using OIM, het-robust, and cluster-robust SEs.
> >
> > On Tue, Sep 4, 2012 at 1:02 PM, Lucas <lucaselastic@gmail.com> wrote:
> >> Dear Statalisters,
> >>
> >> I thought this was a simple issue, estimating a 2-way fixed effects
> >> model, but, alas, I cannot figure it out. I see several stata and
> >> user-written commands, but I am not certain any of them will work in
> >> my situation.
> >>
> >> My data is that I have dozens of companies crossed with dozens of
> >> jurisdictions (e.g., counties).  For substantive reasons I need to
> >> obtain the coefficients for each company and for each county (which
> >> makes areg not quite what I need).
> >>
> >> FYI, there is no time element here, so lags, AR processes, and such
> >> are not relevant.
> >>
> >> Any guidance anyone can offer is greatly appreciated.
> >>
> >> Sam
> > *
> > *   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/
> *
> *   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/


*
*   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   |   Site index