Alice,
The coefficients are the same but the SEs are different because of a
standard trap in fixed effects estimation - the SEs have to be
adjusted for the degrees of freedom used up by the fixed effects.
Have a look at any standard econometrics book, where it should be
covered, or indeed in the Stata manual entry for -xtreg,fe-.
--Mark
From: "Alice(Libo) Sun" <[email protected]>
To: <[email protected]>
Subject: st: Panel FE vs. pooled OLS with demeaned variables
Date sent: Tue, 20 Sep 2005 10:03:26 -0500
Send reply to: [email protected]
> Dear Statalist,
>
>
> I need to run a two-step regression similar to the fixed effect model.
> (Xtreg, fe).
>
> I first replicate the fixed effect results by running pooled OLS with
> demeaned variables (demean by groups). Yet, I got same coefficients, but
> relative larger t statistics.
>
> I attached the results for the two approaches as follows. Any suggestion
> will be appreciated.
>
> Thanks,
>
>
> Alice
>
>
> ---------------------------------------------------------------------
> . regress dqratios ddiv_bkrgs ddiv_bkpts dsizes dwlvgs debitdags dtos
>
> Source | SS df MS Number of obs =
> 839
> -------------+------------------------------ F( 6, 832) =
> 5.71
> Model | 9.47964286 6 1.57994048 Prob > F =
> 0.0000
> Residual | 230.110177 832 .276574732 R-squared =
> 0.0396
> -------------+------------------------------ Adj R-squared =
> 0.0326
> Total | 239.58982 838 .285906706 Root MSE =
> .5259
>
> ----------------------------------------------------------------------------
> --
> dqratios | Coef. Std. Err. t P>|t| [95% Conf.
> Interval]
> -------------+--------------------------------------------------------------
> --
> ddiv_bkrgs | -.0790218 .0611639 -1.29 0.197 -.1990756
> .0410319
> ddiv_bkpts | .1574348 .0536925 2.93 0.003 .0520462
> .2628235
> dsizes | .2569522 .0672621 3.82 0.000 .1249289
> .3889755
> dwlvgs | -.0851709 .0287239 -2.97
> 0.003 -.1415507 -.0287911
> debitdags | .0556536 .0272819 2.04 0.042 .0021041
> .1092031
> dtos | -.0469219 .0285672 -1.64 0.101 -.1029942
> .0091505
> _cons | -1.21e-09 .0181562 -0.00 1.000 -.0356374
> .0356374
> ----------------------------------------------------------------------------
> --
>
> . xtreg qratios div_bkrgs div_bkpts sizes wlvgs ebitdags tos, fe
>
> Fixed-effects (within) regression Number of obs =
> 839
> Group variable (i) : permno Number of groups =
> 142
>
> R-sq: within = 0.0396 Obs per group: min =
> 1
> between = 0.0375 avg =
> 5.9
> overall = 0.0466 max =
> 9
>
> F(6,691) =
> 4.74
> corr(u_i, Xb) = -0.1086 Prob > F =
> 0.0001
>
> ----------------------------------------------------------------------------
> --
> qratios | Coef. Std. Err. t P>|t| [95% Conf.
> Interval]
> -------------+--------------------------------------------------------------
> --
> div_bkrgs | -.0790218 .0671148 -1.18 0.239 -.2107952
> .0527515
> div_bkpts | .1574349 .0589164 2.67 0.008 .0417582
> .2731115
> sizes | .2569522 .0738062 3.48 0.001 .1120409
> .4018635
> wlvgs | -.0851709 .0315185 -2.70
> 0.007 -.1470545 -.0232873
> ebitdags | .0556536 .0299363 1.86 0.063 -.0031234
> .1144305
> tos | -.0469219 .0313466 -1.50 0.135 -.1084679
> .0146242
> _cons | -5.08e-10 .0199227 -0.00 1.000 -.0391163
> .0391163
> -------------+--------------------------------------------------------------
> --
> sigma_u | .9147399
> sigma_e | .57707052
> rho | .71531722 (fraction of variance due to u_i)
> ----------------------------------------------------------------------------
> --
> F test that all u_i=0: F(141, 691) = 10.99 Prob > F =
> 0.0000
>
>
>
> ----------------------------------------------------------------------------
> ------
>
>
>
>
>
>
>
>
>
>
> *
> * 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/
Prof. Mark E. Schaffer
Director
Centre for Economic Reform and Transformation
Department of Economics
School of Management & Languages
Heriot-Watt University, Edinburgh EH14 4AS UK
44-131-451-3494 direct
44-131-451-3296 fax
44-131-451-3485 CERT administrator
http://www.sml.hw.ac.uk/cert
*
* 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/