Firm fixed effects with reghdfe

Hi statalist community,

I am working with a dataset that looks like this:

Code:

input float(abs_DACC DISCLOSE TACC_2 firm_id)
    .01934835 0  .024187315  1
     .0574966 0   .01029703  1
   .006250922 0   .02406213  1
   .015014593 1 -.032378167  1
 .00020516732 1  -.03928642  1
     .7868658 0 -.032384235  2
     1.394589 0   .03529982  2
     .4186824 1  -.04371227  2
     .0740103 1  -.09670375  2
      .315819 1  -.05156721  2
    .13176967 1   -.0364581  2
     5.007663 0   -2.707317  3
     8.962314 0  -127.03571  3
     11.19606 1  -17.384615  3
    17.230572 1   -26.91398  3
    17.230572 1    19.80435  3
     .2146687 1  .022880916  3
 6.591949e-17 0  .012194293  4
 5.551115e-17 0   .05371979  4
            0 0   .00599256  4
    .05828558 1  -.06137639  4
 .00002448961 1   -.4025554  4
   .000550691 0   .02795735  5
  .0008957673 0   .01801897  5
 .00010497781 1 -.007521968  5
 .00026854477 1   .04847973  5
2.1900884e-17 1 .0006031727  5
   .004420182 0 .0020564343  6
   .032602683 0   .04683008  6
    .01220035 1 -.036919225  6
   .010041647 1 -.008244086  6
   .001055236 1 .0010720822  6
    .08491836 1   -.0839263  6
   .066476375 0  -.09063073  7
    .02260962 0  -.07849443  7
   .018439792 1 -.018218448  7
   .034525417 1  -.10765903  7
    .04779583 1   -.0992876  7
     .1033026 1  -.13874565  7
    .32733145 0  -.06901214  8
    .10863945 0   .10752717  8
    .24321346 1  -.20360465  8
    .06225694 1 -.031496804  8
    .07914343 1   -.1281045  8
   .013619387 1 -.003827409  8
    .01897011 0  -.07032511  9
     .3350872 0  -.11521104  9
     .1689511 1   -.1888067  9
     .9896048 1 -.007527853  9
     1.326567 1   .08559322  9
    .05096782 1 -.033362597  9
    .08315077 0  -.11803018 10
    .06128209 0  -.08649246 10
      .354572 1  -.09273923 10
     .4752698 1  -.07367383 10
  .0041804584 1  -.11853525 10
    .08981778 1  -.05202269 10
    .03780806 0  -.06696881 11
    .01859687 0  -.06523953 11
   .074236915 0  -.11448495 11
    .05487452 1   .07933958 11
   .069816515 1  -.05653808 11
     .0155515 1  -.06308693 11
    1.4021575 0  .010252313 12
    .15850054 0  -.07700977 12
     .8803007 1  .007802302 12
     7.066467 1  -.13872992 12
     .4904179 1  -.17439805 12
    .11286884 1  -.12696882 12
      .012223 0 -.010278583 13
  .0011163342 0 -.008919143 13
   .010542297 1   -.0209406 13
   .005647548 1 -.010411106 13
  .0038684416 1  -.14439845 13
    .03858467 1  .026925163 13
1.7347235e-18 0 -.009141093 14
 6.938894e-18 0  -.04401675 14
2.7755576e-16 1   -.1800494 14
 2.428613e-17 1  -.01925059 14
 6.800116e-16 1  -.06629362 14
1.3877788e-16 1  .019053344 14
   .064121634 1  -.07642028 15
    .05395566 0   -.0760234 16
     .0471172 0  -.10151234 16
   .001719041 1  -.04629297 16
    .03862327 1  .003522699 16
    .05165628 1    -.087208 16
   .009629753 1  -.06834266 16
    .04993461 1 -.029572846 17
    .03946859 1 -.024929015 17
    .07654464 0  -.10461678 18
    .05381111 0  -.08892404 18
    .01776027 0  -.05072735 18
   .030726336 1    -.082152 18
   .064680666 1  -.01432552 18
     .0471099 1  -.02597645 18
     .3535477 0  -.04381294 19
    .29521552 0  -.05642635 19
    .56696486 1 -.010838772 19
    .13899466 1  -.06851529 19

I ran a regression using reghdfe and I noticed that not all of the firm_id's were used. Stata notes that 713 singleton observations are dropped to run the regression but this leaves 764 observations unaccounted for (or am I interpreting this wrong?). Does someone know what is the matter here? Here is some additional information:

Code:

sum firm_id
    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
     firm_id |     24,141    2760.679    1674.156          1       6233

. reghdfe abs_DACC DISCLOSE TACC_2, absorb(firm_id)
(dropped 713 singleton observations)
(MWFE estimator converged in 1 iterations)


HDFE Linear regression                            Number of obs   =     23,428
Absorbing 1 HDFE group                            F(   2,  18670) =      11.20
                                                  Prob > F        =     0.0000
                                                  R-squared       =     0.5093
                                                  Adj R-squared   =     0.3843
                                                  Within R-sq.    =     0.0012
                                                  Root MSE        =     1.6299

------------------------------------------------------------------------------
    abs_DACC |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    DISCLOSE |   .0575487    .023677     2.43   0.015     .0111396    .1039579
      TACC_2 |  -.0005184    .000127    -4.08   0.000    -.0007674   -.0002694
       _cons |   .4833111   .0185127    26.11   0.000     .4470246    .5195977
------------------------------------------------------------------------------



Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
     firm_id |      4756           0        4756     |
-----------------------------------------------------+

Kind regards,
Sarah