test coefficient equality across groups using reghdfe

River Huang

Join Date: Mar 2016

Posts: 1908
#1

test coefficient equality across groups using reghdfe

24 Apr 2025, 06:45

Dear all,

I have estimated the following models using reghdfe (after installing reghdfe and ftools via ssc install):

Code:

use "dta\raw", clear global cvars "Size Lev Age Growth TQ NonSoe Top1 BoardSize BoardInd Dual MShare" reghdfe Comp c.Perf c.Digital c.Perf#c.Digital $cvars if Down == 1, absorb(Stkcd Year) cluster(Stkcd) reghdfe Comp c.Perf c.Digital c.Perf#c.Digital $cvars if Down == 0, absorb(Stkcd Year) cluster(Stkcd)

I would like to test whether the coefficient on the interaction term c.Perf#c.Digital is statistically different between the two groups defined by Down (i.e., Down == 1 vs. Down == 0).

Could you please advise on the appropriate method or command to perform this test?

Attached Files

raw.dta (6.05 MB, 1 view)

Ho-Chuan (River) Huang
Stata 19.0, MP(4)
Tags: None

Andrew Musau

Join Date: Oct 2014
Posts: 10102

24 Apr 2025, 07:08

reghdfe is from https://scorreia.com/software/reghdfe/ (FAQ Advice #12). Thanks for the reproducible example.

Code:

use "dta\raw", clear
local cvars "c.Size c.Lev c.Age c.Growth c.TQ c.NonSoe c.Top1 c.BoardSize c.BoardInd c.Dual c.MShare"  
reghdfe Comp c.Perf c.Digital c.Perf#c.Digital `cvars' if Down == 1, absorb(Stkcd Year) cluster(Stkcd)
reghdfe Comp c.Perf c.Digital c.Perf#c.Digital `cvars' if Down == 0, absorb(Stkcd Year) cluster(Stkcd)
gen down0= 0.Down
gen down1= 1.Down
reghdfe Comp i.Down#(c.Perf c.Digital c.Perf#c.Digital `cvars'), absorb(i.down0#i.Stkcd i.down1#i.Stkcd i.down0#i.Year i.down1#i.Year) cluster(Stkcd)
test 0.Down#c.Perf#c.Digital=1.Down#c.Perf#c.Digital

Res.:

Code:

. reghdfe Comp c.Perf c.Digital c.Perf#c.Digital `cvars' if Down == 1, absorb(Stkcd Year) cluster(Stkcd)
(dropped 145 singleton observations)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =     13,354
Absorbing 2 HDFE groups                           F(  14,   2827) =      23.66
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8496
                                                  Adj R-squared   =     0.8087
                                                  Within R-sq.    =     0.1133
Number of clusters (Stkcd)   =      2,828         Root MSE        =     0.3117

                                  (Std. err. adjusted for 2,828 clusters in Stkcd)
----------------------------------------------------------------------------------
                 |               Robust
            Comp | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-----------------+----------------------------------------------------------------
            Perf |   .6840891   .1056485     6.48   0.000     .4769331    .8912452
         Digital |   .0889415   .0338347     2.63   0.009     .0225983    .1552847
                 |
c.Perf#c.Digital |  -.8155218   .2428553    -3.36   0.001    -1.291713   -.3393303
                 |
            Size |   .2449749   .0174767    14.02   0.000     .2107065    .2792432
             Lev |  -.1117888   .0553662    -2.02   0.044     -.220351   -.0032266
             Age |  -.0436835   .0196131    -2.23   0.026    -.0821409   -.0052261
          Growth |  -.0003945   .0113912    -0.03   0.972    -.0227304    .0219414
              TQ |   .0272782   .0047466     5.75   0.000     .0179711    .0365854
          NonSoe |   .0365886   .0432922     0.85   0.398    -.0482989     .121476
            Top1 |  -.0032176   .0970363    -0.03   0.974    -.1934866    .1870515
       BoardSize |   .1230444   .0554504     2.22   0.027     .0143169    .2317718
        BoardInd |   .1048244   .1548339     0.68   0.498    -.1987744    .4084232
            Dual |   .0291436   .0156181     1.87   0.062    -.0014803    .0597676
          MShare |  -.0422034   .0722157    -0.58   0.559    -.1838042    .0993973
           _cons |   8.596347   .4091988    21.01   0.000     7.793989    9.398706
----------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
       Stkcd |      2828        2828           0    *|
        Year |        11           1          10     |
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation

. 
. reghdfe Comp c.Perf c.Digital c.Perf#c.Digital `cvars' if Down == 0, absorb(Stkcd Year) cluster(Stkcd)
(dropped 416 singleton observations)
(MWFE estimator converged in 7 iterations)

HDFE Linear regression                            Number of obs   =      9,827
Absorbing 2 HDFE groups                           F(  14,   2378) =      21.20
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8528
                                                  Adj R-squared   =     0.8052
                                                  Within R-sq.    =     0.1028
Number of clusters (Stkcd)   =      2,379         Root MSE        =     0.3242

                                  (Std. err. adjusted for 2,379 clusters in Stkcd)
----------------------------------------------------------------------------------
                 |               Robust
            Comp | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-----------------+----------------------------------------------------------------
            Perf |   1.163656   .1759658     6.61   0.000      .818594    1.508719
         Digital |   .0651785   .0515074     1.27   0.206    -.0358256    .1661827
                 |
c.Perf#c.Digital |   .2547599   .5477789     0.47   0.642    -.8194138    1.328934
                 |
            Size |   .2315662   .0174485    13.27   0.000     .1973503     .265782
             Lev |  -.0859696   .0623772    -1.38   0.168     -.208289    .0363498
             Age |  -.0460286   .0334524    -1.38   0.169    -.1116274    .0195702
          Growth |  -.0219965   .0092668    -2.37   0.018    -.0401682   -.0038247
              TQ |   .0126919   .0050184     2.53   0.012      .002851    .0225327
          NonSoe |   -.002887   .0491611    -0.06   0.953      -.09929     .093516
            Top1 |  -.1907074   .0961432    -1.98   0.047    -.3792406   -.0021742
       BoardSize |   .1365181   .0744628     1.83   0.067    -.0095007    .2825368
        BoardInd |   .0361779   .1937148     0.19   0.852    -.3436894    .4160452
            Dual |   .0306783    .019108     1.61   0.109    -.0067918    .0681484
          MShare |   .1241965   .0917216     1.35   0.176    -.0556661    .3040591
           _cons |   8.990244   .4208074    21.36   0.000     8.165057    9.815432
----------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
       Stkcd |      2379        2379           0    *|
        Year |        11           1          10     |
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation

. 
. gen down0= 0.Down
(1 missing value generated)

. 
. gen down1= 1.Down
(1 missing value generated)

. 
. reghdfe Comp i.Down#(c.Perf c.Digital c.Perf#c.Digital `cvars'), absorb(i.down0#i.Stkcd i.down1#i.Stkcd i.down0#i.Year i.down1#i.Year
> ) cluster(Stkcd)
(dropped 561 singleton observations)
(MWFE estimator converged in 8 iterations)

HDFE Linear regression                            Number of obs   =     23,181
Absorbing 4 HDFE groups                           F(  28,   2963) =      17.12
Statistics robust to heteroskedasticity           Prob > F        =     0.0000
                                                  R-squared       =     0.8511
                                                  Adj R-squared   =     0.7286
                                                  Within R-sq.    =     0.1088
Number of clusters (Stkcd)   =      2,964         Root MSE        =     0.3763

                                       (Std. err. adjusted for 2,964 clusters in Stkcd)
---------------------------------------------------------------------------------------
                      |               Robust
                 Comp | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
----------------------+----------------------------------------------------------------
          Down#c.Perf |
                   0  |   1.163656   .1759295     6.61   0.000     .8186998    1.508613
                   1  |   .6840891   .1056644     6.47   0.000      .476906    .8912723
                      |
       Down#c.Digital |
                   0  |   .0651785   .0514968     1.27   0.206    -.0357947    .1661517
                   1  |   .0889415   .0338398     2.63   0.009     .0225896    .1552934
                      |
Down#c.Perf#c.Digital |
                   0  |   .2547599    .547666     0.47   0.642    -.8190844    1.328604
                   1  |  -.8155218   .2428918    -3.36   0.001    -1.291776    -.339268
                      |
          Down#c.Size |
                   0  |   .2315662   .0174449    13.27   0.000     .1973608    .2657715
                   1  |   .2449749   .0174793    14.02   0.000      .210702    .2792477
                      |
           Down#c.Lev |
                   0  |  -.0859696   .0623644    -1.38   0.168    -.2082515    .0363123
                   1  |  -.1117888   .0553745    -2.02   0.044    -.2203652   -.0032124
                      |
           Down#c.Age |
                   0  |  -.0460286   .0334455    -1.38   0.169    -.1116073    .0195501
                   1  |  -.0436835    .019616    -2.23   0.026     -.082146   -.0052211
                      |
        Down#c.Growth |
                   0  |  -.0219965   .0092648    -2.37   0.018    -.0401627   -.0038303
                   1  |  -.0003945   .0113929    -0.03   0.972    -.0227333    .0219443
                      |
            Down#c.TQ |
                   0  |   .0126919   .0050174     2.53   0.011      .002854    .0225297
                   1  |   .0272782   .0047473     5.75   0.000     .0179699    .0365866
                      |
        Down#c.NonSoe |
                   0  |   -.002887   .0491509    -0.06   0.953    -.0992605    .0934865
                   1  |   .0365886   .0432987     0.85   0.398      -.04831    .1214871
                      |
          Down#c.Top1 |
                   0  |  -.1907074   .0961234    -1.98   0.047    -.3791827    -.002232
                   1  |  -.0032176   .0970509    -0.03   0.974    -.1935115    .1870764
                      |
     Down#c.BoardSize |
                   0  |   .1365181   .0744475     1.83   0.067    -.0094559     .282492
                   1  |   .1230444   .0554588     2.22   0.027     .0143027     .231786
                      |
      Down#c.BoardInd |
                   0  |   .0361779   .1936748     0.19   0.852    -.3435729    .4159287
                   1  |   .1048244   .1548572     0.68   0.499    -.1988141    .4084629
                      |
          Down#c.Dual |
                   0  |   .0306783   .0191041     1.61   0.108    -.0067803    .0681369
                   1  |   .0291436   .0156204     1.87   0.062    -.0014844    .0597716
                      |
        Down#c.MShare |
                   0  |   .1241965   .0917027     1.35   0.176    -.0556109     .304004
                   1  |  -.0422034   .0722265    -0.58   0.559    -.1838227    .0994158
                      |
                _cons |    8.76333    .343901    25.48   0.000     8.089021    9.437639
---------------------------------------------------------------------------------------

Absorbed degrees of freedom:
-------------------------------------------------------+
   Absorbed FE | Categories  - Redundant  = Num. Coefs |
---------------+---------------------------------------|
   down0#Stkcd |      5207        5207           0    *|
   down1#Stkcd |      5207        5207           0    *|
    down0#Year |        22           1          21     |
    down1#Year |        22          22           0     |
-------------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation

. 
. test 0.Down#c.Perf#c.Digital=1.Down#c.Perf#c.Digital

 ( 1)  0b.Down#c.Perf#c.Digital - 1.Down#c.Perf#c.Digital = 0

       F(  1,  2963) =    3.65
            Prob > F =    0.0563

Comment

River Huang

Join Date: Mar 2016

Posts: 1908
#3

24 Apr 2025, 09:09

Dear Andrew,

Thanks so much for this very constructive suggestion.

Ho-Chuan (River) Huang
Stata 19.0, MP(4)
Comment
River Huang

Join Date: Mar 2016

Posts: 1908
#4

28 Apr 2025, 19:48

Dear Andrew,

Could you please take a look at the follow-up question? It is the same as above, but I now use (ssc install) ivreghdfe to deal with endogeneity, aiming to test whether the coefficients of Perf_Digital (c.Perf#c.Digital above) are equal across groups.

Code:

use "dta\raw", clear global cvars "Size Lev Age Growth TQ NonSoe Top1 BoardSize BoardInd Dual MShare" ivreghdfe Comp (Digital Perf_Digital = DigitalIV Perf_DigitalIV) Perf $cvars if H_AnalystNum==1, absorb(Stkcd Year) cluster(Stkcd) ivreghdfe Comp (Digital Perf_Digital = DigitalIV Perf_DigitalIV) Perf $cvars if H_AnalystNum==0, absorb(Stkcd Year) cluster(Stkcd)

Ho-Chuan (River) Huang
Stata 19.0, MP(4)
Comment

Andrew Musau

Join Date: Oct 2014
Posts: 10102

29 Apr 2025, 06:13

You just have to interact the endogenous variables and the instruments as well. Note that ivreghdfe is from https://github.com/sergiocorreia/ivreghdfe (FAQ Advice #12).

Code:

use "dta\raw", clear
local cvars "c.Size c.Lev c.Age c.Growth c.TQ c.NonSoe c.Top1 c.BoardSize c.BoardInd c.Dual c.MShare"  

ivreghdfe Comp (Digital Perf_Digital = DigitalIV Perf_DigitalIV) Perf `cvars' if H_AnalystNum==1, absorb(Stkcd Year) cluster(Stkcd)
ivreghdfe Comp (Digital Perf_Digital = DigitalIV Perf_DigitalIV) Perf `cvars' if H_AnalystNum==0, absorb(Stkcd Year) cluster(Stkcd)

gen H_AnalystNum0= 0.H_AnalystNum
gen H_AnalystNum1= 1.H_AnalystNum

ivreghdfe Comp (i.H_AnalystNum#(c.Digital c.Perf_Digital) = i.H_AnalystNum#(c.DigitalIV c.Perf_DigitalIV)) i.H_AnalystNum#(c.Perf `cvars'), ///
absorb(i.H_AnalystNum0#i.Stkcd i.H_AnalystNum1#i.Stkcd i.H_AnalystNum0#i.Year i.H_AnalystNum1#i.Year) cluster(Stkcd)

Res.:

Code:

. ivreghdfe Comp (Digital Perf_Digital = DigitalIV Perf_DigitalIV) Perf `cvars' if H_AnalystNum==1, absorb(Stkcd Year) cluster(Stkcd)
(dropped 175 singleton observations)
(MWFE estimator converged in 7 iterations)

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics robust to heteroskedasticity and clustering on Stkcd

Number of clusters (Stkcd) =      2716                Number of obs =    15990
                                                      F( 14,  2715) =    28.09
                                                      Prob > F      =   0.0000
Total (centered) SS     =  1366.733837                Centered R2   =   0.0974
Total (uncentered) SS   =  1366.733837                Uncentered R2 =   0.0974
Residual SS             =  1233.567058                Root MSE      =     .278

------------------------------------------------------------------------------
             |               Robust
        Comp | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
     Digital |   .0362972   .0462587     0.78   0.433    -.0544086     .127003
Perf_Digital |  -.6027751   .2210178    -2.73   0.006    -1.036155   -.1693949
        Perf |   .8928023   .1107296     8.06   0.000     .6756796    1.109925
        Size |   .2402219   .0158028    15.20   0.000     .2092352    .2712085
         Lev |  -.0562848   .0575666    -0.98   0.328    -.1691636     .056594
         Age |   .0008869   .0214733     0.04   0.967    -.0412188    .0429926
      Growth |  -.0218931   .0096426    -2.27   0.023    -.0408007   -.0029855
          TQ |   .0178361   .0047083     3.79   0.000     .0086039    .0270684
      NonSoe |  -.0252304   .0455556    -0.55   0.580    -.1145575    .0640968
        Top1 |  -.1403694     .09235    -1.52   0.129    -.3214528     .040714
   BoardSize |   .1838664   .0554432     3.32   0.001     .0751512    .2925816
    BoardInd |   .2430504   .1516834     1.60   0.109    -.0543762     .540477
        Dual |    .049042   .0156845     3.13   0.002     .0182871    .0797968
      MShare |  -.0259193   .0692166    -0.37   0.708    -.1616419    .1098032
------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic):             67.308
                                                   Chi-sq(1) P-val =    0.0000
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic):             3571.116
                         (Kleibergen-Paap rk Wald F statistic):        113.028
Stock-Yogo weak ID test critical values: 10% maximal IV size              7.03
                                         15% maximal IV size              4.58
                                         20% maximal IV size              3.95
                                         25% maximal IV size              3.63
Source: Stock-Yogo (2005).  Reproduced by permission.
NB: Critical values are for Cragg-Donald F statistic and i.i.d. errors.
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments):         0.000
                                                 (equation exactly identified)
------------------------------------------------------------------------------
Instrumented:         Digital Perf_Digital
Included instruments: Perf Size Lev Age Growth TQ NonSoe Top1 BoardSize
                      BoardInd Dual MShare
Excluded instruments: DigitalIV Perf_DigitalIV
Partialled-out:       _cons
                      nb: total SS, model F and R2s are after partialling-out;
                          any small-sample adjustments include partialled-out
                          variables in regressor count K
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
       Stkcd |      2716        2716           0    *|
        Year |        11           1          10     |
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation

. ivreghdfe Comp (Digital Perf_Digital = DigitalIV Perf_DigitalIV) Perf `cvars' if H_AnalystNum==0, absorb(Stkcd Year) cluster(Stkcd)
(dropped 504 singleton observations)
(MWFE estimator converged in 8 iterations)

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics robust to heteroskedasticity and clustering on Stkcd

Number of clusters (Stkcd) =      1529                Number of obs =     7074
                                                      F( 14,  1528) =     9.93
                                                      Prob > F      =   0.0000
Total (centered) SS     =  542.4841762                Centered R2   =   0.0638
Total (uncentered) SS   =  542.4841762                Uncentered R2 =   0.0638
Residual SS             =  507.8623462                Root MSE      =    .2684

------------------------------------------------------------------------------
             |               Robust
        Comp | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
     Digital |   .1243627   .0767049     1.62   0.105    -.0260954    .2748208
Perf_Digital |  -2.075021   .5444572    -3.81   0.000    -3.142984   -1.007059
        Perf |    .437495   .1141209     3.83   0.000     .2136449    .6613452
        Size |   .2023082   .0222079     9.11   0.000      .158747    .2458694
         Lev |  -.1104748   .0687273    -1.61   0.108    -.2452846     .024335
         Age |  -.0531308   .0354666    -1.50   0.134    -.1226993    .0164376
      Growth |  -.0133474   .0091982    -1.45   0.147    -.0313898     .004695
          TQ |   .0217212   .0062831     3.46   0.001     .0093968    .0340455
      NonSoe |    .076801    .052988     1.45   0.147    -.0271359    .1807378
        Top1 |   .0027556   .1273316     0.02   0.983    -.2470076    .2525188
   BoardSize |   -.028521   .0764924    -0.37   0.709    -.1785623    .1215202
    BoardInd |  -.3445723   .2046194    -1.68   0.092    -.7459368    .0567923
        Dual |  -.0026342   .0211331    -0.12   0.901    -.0440872    .0388187
      MShare |   .2834581   .1243561     2.28   0.023     .0395315    .5273848
------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic):             12.525
                                                   Chi-sq(1) P-val =    0.0004
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic):             1329.960
                         (Kleibergen-Paap rk Wald F statistic):         25.538
Stock-Yogo weak ID test critical values: 10% maximal IV size              7.03
                                         15% maximal IV size              4.58
                                         20% maximal IV size              3.95
                                         25% maximal IV size              3.63
Source: Stock-Yogo (2005).  Reproduced by permission.
NB: Critical values are for Cragg-Donald F statistic and i.i.d. errors.
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments):         0.000
                                                 (equation exactly identified)
------------------------------------------------------------------------------
Instrumented:         Digital Perf_Digital
Included instruments: Perf Size Lev Age Growth TQ NonSoe Top1 BoardSize
                      BoardInd Dual MShare
Excluded instruments: DigitalIV Perf_DigitalIV
Partialled-out:       _cons
                      nb: total SS, model F and R2s are after partialling-out;
                          any small-sample adjustments include partialled-out
                          variables in regressor count K
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
       Stkcd |      1529        1529           0    *|
        Year |        11           1          10     |
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation

.
. gen H_AnalystNum0= 0.H_AnalystNum

. gen H_AnalystNum1= 1.H_AnalystNum

.
. ivreghdfe Comp (i.H_AnalystNum#(c.Digital c.Perf_Digital) = i.H_AnalystNum#(c.DigitalIV c.Perf_DigitalIV)) i.H_AnalystNum#(c.Perf `cvar
> s'), ///
> absorb(i.H_AnalystNum0#i.Stkcd i.H_AnalystNum1#i.Stkcd i.H_AnalystNum0#i.Year i.H_AnalystNum1#i.Year) cluster(Stkcd)
(dropped 679 singleton observations)
(MWFE estimator converged in 9 iterations)

IV (2SLS) estimation
--------------------

Estimates efficient for homoskedasticity only
Statistics robust to heteroskedasticity and clustering on Stkcd

Number of clusters (Stkcd) =      2976                Number of obs =    23064
                                                      F( 28,  2975) =    17.69
                                                      Prob > F      =   0.0000
Total (centered) SS     =  1909.218013                Centered R2   =   0.0879
Total (uncentered) SS   =  1909.218013                Uncentered R2 =   0.0879
Residual SS             =  1741.429405                Root MSE      =    .2751

---------------------------------------------------------------------------------------------
                            |               Robust
                       Comp | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
----------------------------+----------------------------------------------------------------
     H_AnalystNum#c.Digital |
                         0  |   .1243627   .0766477     1.62   0.105    -.0259253    .2746506
                         1  |   .0362972   .0462728     0.78   0.433    -.0544328    .1270272
                            |
H_AnalystNum#c.Perf_Digital |
                         0  |  -2.075021   .5440513    -3.81   0.000    -3.141776   -1.008267
                         1  |  -.6027751   .2210854    -2.73   0.006    -1.036271   -.1692792
                            |
        H_AnalystNum#c.Perf |
                         0  |    .437495   .1140358     3.84   0.000      .213898    .6610921
                         1  |   .8928023   .1107634     8.06   0.000     .6756216    1.109983
                            |
        H_AnalystNum#c.Size |
                         0  |   .2023082   .0221913     9.12   0.000     .1587962    .2458201
                         1  |   .2402219   .0158076    15.20   0.000      .209227    .2712168
                            |
         H_AnalystNum#c.Lev |
                         0  |  -.1104748   .0686761    -1.61   0.108    -.2451322    .0241826
                         1  |  -.0562848   .0575842    -0.98   0.328    -.1691937    .0566242
                            |
         H_AnalystNum#c.Age |
                         0  |  -.0531308   .0354402    -1.50   0.134    -.1226206     .016359
                         1  |   .0008869   .0214799     0.04   0.967      -.04123    .0430038
                            |
      H_AnalystNum#c.Growth |
                         0  |  -.0133474   .0091913    -1.45   0.147    -.0313694    .0046746
                         1  |  -.0218931   .0096456    -2.27   0.023    -.0408058   -.0029804
                            |
          H_AnalystNum#c.TQ |
                         0  |   .0217212   .0062784     3.46   0.001     .0094108    .0340316
                         1  |   .0178361   .0047097     3.79   0.000     .0086015    .0270708
                            |
      H_AnalystNum#c.NonSoe |
                         0  |    .076801   .0529485     1.45   0.147    -.0270183    .1806203
                         1  |  -.0252304   .0455695    -0.55   0.580    -.1145814    .0641206
                            |
        H_AnalystNum#c.Top1 |
                         0  |   .0027556   .1272366     0.02   0.983    -.2467251    .2522363
                         1  |  -.1403694   .0923782    -1.52   0.129    -.3215011    .0407623
                            |
   H_AnalystNum#c.BoardSize |
                         0  |   -.028521   .0764354    -0.37   0.709    -.1783926    .1213505
                         1  |   .1838664   .0554602     3.32   0.001     .0751222    .2926106
                            |
    H_AnalystNum#c.BoardInd |
                         0  |  -.3445723   .2044668    -1.69   0.092     -.745483    .0563385
                         1  |   .2430504   .1517298     1.60   0.109    -.0544556    .5405563
                            |
        H_AnalystNum#c.Dual |
                         0  |  -.0026342   .0211173    -0.12   0.901    -.0440403    .0387718
                         1  |    .049042   .0156893     3.13   0.002     .0182789     .079805
                            |
      H_AnalystNum#c.MShare |
                         0  |   .2834581   .1242634     2.28   0.023     .0398073    .5271089
                         1  |  -.0259193   .0692378    -0.37   0.708    -.1616781    .1098394
---------------------------------------------------------------------------------------------
Underidentification test (Kleibergen-Paap rk LM statistic):              0.000
                                                   Chi-sq(1) P-val =    1.0000
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic):             2170.853
                         (Kleibergen-Paap rk Wald F statistic):          0.000
Stock-Yogo weak ID test critical values:                       <not available>
------------------------------------------------------------------------------
Hansen J statistic (overidentification test of all instruments):         0.000
                                                 (equation exactly identified)
------------------------------------------------------------------------------
Instrumented:         0b.H_AnalystNum#c.Digital 1.H_AnalystNum#c.Digital
                      0b.H_AnalystNum#c.Perf_Digital
                      1.H_AnalystNum#c.Perf_Digital
Included instruments: 0b.H_AnalystNum#c.Perf 1.H_AnalystNum#c.Perf
                      0b.H_AnalystNum#c.Size 1.H_AnalystNum#c.Size
                      0b.H_AnalystNum#c.Lev 1.H_AnalystNum#c.Lev
                      0b.H_AnalystNum#c.Age 1.H_AnalystNum#c.Age
                      0b.H_AnalystNum#c.Growth 1.H_AnalystNum#c.Growth
                      0b.H_AnalystNum#c.TQ 1.H_AnalystNum#c.TQ
                      0b.H_AnalystNum#c.NonSoe 1.H_AnalystNum#c.NonSoe
                      0b.H_AnalystNum#c.Top1 1.H_AnalystNum#c.Top1
                      0b.H_AnalystNum#c.BoardSize 1.H_AnalystNum#c.BoardSize
                      0b.H_AnalystNum#c.BoardInd 1.H_AnalystNum#c.BoardInd
                      0b.H_AnalystNum#c.Dual 1.H_AnalystNum#c.Dual
                      0b.H_AnalystNum#c.MShare 1.H_AnalystNum#c.MShare
Excluded instruments: 0b.H_AnalystNum#c.DigitalIV 1.H_AnalystNum#c.DigitalIV
                      0b.H_AnalystNum#c.Perf_DigitalIV
                      1.H_AnalystNum#c.Perf_DigitalIV
Partialled-out:       _cons
                      nb: total SS, model F and R2s are after partialling-out;
                          any small-sample adjustments include partialled-out
                          variables in regressor count K
------------------------------------------------------------------------------

Absorbed degrees of freedom:
---------------------------------------------------------------+
           Absorbed FE | Categories  - Redundant  = Num. Coefs |
-----------------------+---------------------------------------|
   H_AnalystNum0#Stkcd |      4245        4245           0    *|
   H_AnalystNum1#Stkcd |      4245        4245           0    *|
    H_AnalystNum0#Year |        22           1          21     |
    H_AnalystNum1#Year |        22          22           0     |
---------------------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation

Comment

River Huang

Join Date: Mar 2016

Posts: 1908
#6

29 Apr 2025, 06:27

Thank you very much, Andrew. I'll give it a try.

Ho-Chuan (River) Huang
Stata 19.0, MP(4)
Comment

Announcement

test coefficient equality across groups using reghdfe

Comment

Comment

Comment

Comment

Comment