Antw: Re: Re: st: Clustered standard errors: Insufficient observations

["Jost Heckemeyer" <Heckemeyer@zew.de>]

ok, thanks, I think I got it, if I tell stata to cluster within say, 50
countries, the DoF is 50-1, and if in addition, I estimate xtreg with
firm fixed effects (15,000 firms), then there is not sufficient DoF to
correctly estimate the model parameters, mse etc. with 15000 fixed
effects. Is this the idea? Thanks again, Stas.

>>> Stas Kolenikov  14.05.11 14.56 Uhr >>>
On Sat, May 14, 2011 at 1:45 PM, Jost Heckemeyer  wrote:
> The "generally recommened" referred to the example I gave, i.e. I have
> cross-country data where several 100s and 1000s of firms are clustered
> within several dozens of countries, so clustering and the country
level
> would indeed make sense, as you write it, Stas.  Still, it does not
work
> although and I cannot see any reason why it doesn t.

The technical reason is that your degrees of freedom is the number of
clusters. If you have fixed effects, you consume all of these degrees
of freedom just by the fixed effects/country dummies, and you don't
have anything left to produce standard errors on your coefficients of
interest.

-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.

*
*   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/

Zentrum für Europäische Wirtschaftsforschung GmbH (ZEW), L7,1 68161
Mannheim 
Sitz der Gesellschaft: Mannheim Amtsgericht Mannheim HRB 6554
Aufsichtsratsvorsitzender: Gerhard Stratthaus MdL, Finanzminister a.D.
Geschaeftsfuehrer: Prof. Dr. Dr. h.c. mult. Wolfgang Franz, Thomas Kohl
Centre for European Economic Research L7,1 68161 Mannheim Germany
Seat of the Company: Mannheim Local Court Mannheim HRB 6554
Chairman of the Supervisory Board: Gerhard Stratthaus MdL, Minister,
ret.
Executive Directors: Prof. Dr. Dr. h.c. mult. Wolfgang Franz, Thomas
Kohl
--------------------------------------------------------------------------------
*
*   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/