Dear all.When we use reghdfe to run pannel fixed-effects model,which kind of Rsq should we report in my paper?Adj R-squared or Within R-sq?
Best!
Raymond Zhang


code:
. webuse grunfeld,clear

. reghdfe invest mvalue kstock,a(company year) cl(company)
(MWFE estimator converged in 2 iterations)

HDFE Linear regression Number of obs = 200
Absorbing 2 HDFE groups F( 2, 9) = 60.08
Statistics robust to heteroskedasticity Prob > F = 0.0000
R-squared = 0.9517
Adj R-squared = 0.9431
Within R-sq. = 0.7201
Number of clusters (company) = 10 Root MSE = 51.7245

(Std. Err. adjusted for 10 clusters in company)
------------------------------------------------------------------------------
| Robust
invest | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
mvalue | 0.118 0.011 10.88 0.000 0.093 0.142
kstock | 0.358 0.048 7.48 0.000 0.250 0.466
_cons | -80.164 22.192 -3.61 0.006 -130.365 -29.963
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
Absorbed FE | Categories - Redundant = Num. Coefs |
-------------+---------------------------------------|
company | 10 10 0 *|
year | 20 1 19 |
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation