reghdfe – noabsorb

[Chan Ge]

Dear all,

May I have help with interpreting a piece of information in reghdfe outputs.
The code and header relevant information is pasted below.
My questions are: (1) when I specify noabsorb, shouldn't I expect "Absorbing 0 HDFE group" regarding the output in red? (2) Regardless of the reporting, this line of code specifies an OLS with no fixed effects, and Stata executes so, right?
Thank you.

. reghdfe DVAR varlist, noabsorb vce(robust)
(MWFE estimator converged in 1 iterations)

HDFE Linear regression                            Number of obs   =        500
Absorbing 1 HDFE group                            F(  18,    481) =       4.05
                                                  Prob > F        =     0.0000
                                                  R-squared       =     0.1306
                                                  Adj R-squared   =     0.0980
                                                  Within R-sq.    =     0.1306
                                                  Root MSE        =     0.0568

--------------------------------------------------------------------------------
               |               Robust
          DVAR | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
---------------+----------------------------------------------------------------

[Sebastian Schirner]

Hey Chan,

I do not know the answer to (1). Regarding (2), you are correct. You can run the following code for a basic linear regression and you get the same result:

reg DVAR varlist, vce(robust)

Best,
Sebastian

[Chan Ge]

Thank you for attention, Sebastian. Yes, I got the same results using reg, vce(robust). It was just that the information highlighted in red looked puzzling.
A background of my question: I output the results of alternative fixed effect options in one table for comparison, and expected 0, 1, 2 for the line of N_hdfe.
This post might be served as a corroboration for the expected N_hdfe=0.

[Carlo Lazzaro]

Chan:
the -noabsorb- option actually absorbs _cons, that counts as 1 HDFE:

. reghdfe price mpg, noabsorb
(MWFE estimator converged in 1 iterations)

HDFE Linear regression                            Number of obs   =         74
Absorbing 1 HDFE group                            F(   1,     72) =      20.26
                                                  Prob > F        =     0.0000
                                                  R-squared       =     0.2196
                                                  Adj R-squared   =     0.2087
                                                  Within R-sq.    =     0.2196
                                                  Root MSE        =  2623.6529

------------------------------------------------------------------------------
       price | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         mpg |  -238.8943   53.07669    -4.50   0.000    -344.7008   -133.0879
       _cons |   11253.06   1170.813     9.61   0.000     8919.088    13587.03
------------------------------------------------------------------------------

. mat list e(dof_table)

e(dof_table)[1,5]
       Categories   Redundant   Num Coefs      Exact?     Nested?
_cons           1           0           1           0           0

.

[Chan Ge]

Thank you for clarifying the puzzle, Carlo.
May I ask a follow-up question? Is there a reference manual illustrating the items in the dof_table, such as when and what does it mean when the reporting says "redundant" versus. "coefs".
Below is an example of specifying a FE.
Thank you.

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
    ioSecTRG |        15          15           0    *|
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation

. mat list e(dof_table)

e(dof_table)[1,5]
            Categories   Redundant   Num Coefs      Exact?     Nested?
1.ioSecTRG          15          15           0           0           1

[Carlo Lazzaro]

Chan:
as -reghdfe- is a community-contributed module, the only available resource to delve into its detail is the related help file.

[Chan Ge]

Chan:
as -reghdfe- is a community-contributed module, the only available resource to delve into its detail is the related help file.

Got it. Thank you Carlo.