Re: st: difference between -xtreg, fe- and -areg, absorb- when adding the cluster option

[Ryan Kessler <ryan.kessler.stata@gmail.com>]

Note #2 from http://www.kellogg.northwestern.edu/faculty/matsa/htm/fe.htm:
 When clustering, areg reports cluster-robust standard errors that
reduce the degrees of freedom by the number of fixed effects swept
away in the within-group transformation; xtreg reports smaller
cluster-robust standard errors because it does not make such an
adjustment.  xtreg’s approach of not adjusting the degrees of freedom
is appropriate when the fixed effects swept away by the within-group
transformation are nested within clusters (meaning all the
observations for any given group are in the same cluster), as is
commonly the case (e.g., firm fixed effects are nested within firm,
industry, or state clusters). See Wooldridge (2010, Chapter 20).

Note xtreg's -dfadj- option:

sysuse auto, clear
gen time = _n
xtset rep78 time
areg mpg price, absorb(rep78) vce(cluster rep78)
xtreg mpg price, fe vce(cluster rep78)
xtreg mpg price, fe vce(cluster rep78) dfadj

Ryan Kessler

On Mon, Jan 14, 2013 at 5:11 PM, natalie rebolledo
<na.rebolledo.02@gmail.com> wrote:
> Dear statalist users,
>
> I'm trying to run a regression that accounts the influence of clusters
> on the standard errors. I have been researching and I saw that there
> are at least to basic solutions: one would be running an OLS with the
> cluster option (-reg, cluster (clustervar)-) and a second option would
> be -xtreg, fe i(groupvar)-. I have lots of clusters and observations
> so I'm not too much worried about the assumption of assymptotic
> consistency of this models.
> The thing is that after my research I decided to go for the fixed
> effect model and found out that this is equivalent to do -areg,
> absorb(clustervar)-,
> but I obtained different results between the two when I add the
> cluster option, what I mean is that,
>
> -xtreg, fe i(clustervar)- and -areg, absorb(clustervar)-  are
> equivalent, but when I run
> -xtreg, fe i(clustervar) cluster(clustervar)- and - areg,
> absorb(clustervar) cluster(clustervar)-
> I found the same coefficients but different standard errors, why is
> that? shouldn't the standard errors be equal too?
> and also, between those two models which one is the best to account
> for the presence of clusters in my dataset?
>
>
> Thanks in advance,
> Natalie Rebolledo
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/