All,
I have been working on a panel dataset, looking at the relationship
between inflation and openness, but am slightly confused at to the
results I have been finding. When running OLS and Fixed-Effects (FE)
estimations completely different results are found. However, when the
i() option for the FE estimation is set to the year (which it
shouldn't be; it should be the country code), the panel data results
are almost exactly equal to OLS.
Am I making an error in my use of Stata, or is this a kind of
'pooling' (when setting i() to the year), implying that the 'correct'
value is as found by the 'correct' FE estimation. I have not been
able to find any direct information on this in the Stata documentation
or the net in general - any comments that you can give will be much
appreciated.
The three estimations are:
. xtreg lninf open if [various statements], fe i(period)
Fixed-effects (within) regression Number of obs = 1727
Group variable (i): period Number of groups = 21
R-sq: within = 0.0327 Obs per group: min = 70
between = 0.0167 avg = 82.2
overall = 0.0323 max = 86
F(1,1705) = 57.67
corr(u_i, Xb) = 0.0026 Prob > F = 0.0000
------------------------------------------------------------------------------
lninf | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
open | -.9897034 .1303229 -7.59 0.000 -1.245313 -.7340938
_cons | -1.800335 .0537737 -33.48 0.000 -1.905804 -1.694865
-------------+----------------------------------------------------------------
sigma_u | .19099431
sigma_e | 1.2090179
rho | .02434839 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(20, 1705) = 2.08 Prob > F = 0.0034
. xtreg lninf open if [various statements], fe i(code)
Fixed-effects (within) regression Number of obs = 1727
Group variable (i): code Number of groups = 86
R-sq: within = 0.0017 Obs per group: min = 12
between = 0.0809 avg = 20.1
overall = 0.0323 max = 21
F(1,1640) = 2.73
corr(u_i, Xb) = -0.3667 Prob > F = 0.0984
------------------------------------------------------------------------------
lninf | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
open | .4950169 .2993937 1.65 0.098 -.0922174 1.082251
_cons | -2.315556 .1061136 -21.82 0.000 -2.523688 -2.107423
-------------+----------------------------------------------------------------
sigma_u | .9055351
sigma_e | .89717541
rho | .50463719 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(85, 1640) = 18.02 Prob > F = 0.0000
. reg lninf open if [various statements]
Source | SS df MS Number of obs = 1727
-------------+------------------------------ F( 1, 1725) = 57.61
Model | 85.2610149 1 85.2610149 Prob > F = 0.0000
Residual | 2552.98501 1725 1.47999131 R-squared = 0.0323
-------------+------------------------------ Adj R-squared = 0.0318
Total | 2638.24603 1726 1.52853188 Root MSE = 1.2165
------------------------------------------------------------------------------
lninf | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
open | -.9919216 .1306868 -7.59 0.000 -1.248243 -.7356004
_cons | -1.799565 .053978 -33.34 0.000 -1.905434 -1.693696
*
* 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/
