ivreghdfe exporting F statistics from first stage

[Adrienne Wold]

Hello,
I am running the following specification using ivreg

eststo: ivreghdfe Y (endogenous endogenous2 = Z Z2 ), a(year) first savefirst savefprefix(st1)
estadd scalar fst = `e(rkf)': st1endogenous
estadd scalar fst2 = `e(rkf)': st1endogenous2

I am trying to export the first stage F-stat for my two instruments, which are displayed in output as below:


But when I etreturn list, I cannot find them. Where are the first-stage F statistics stored that I could use to export?

[Andrew Musau]

ivreghdfe and estout are from SSC, as you are asked to explain (FAQ Advice #12). See #6: https://www.statalist.org/forums/for...-stage-outcome

[Adrienne Wold]

Thanks for the reminder, Andrew. Yes stout and ivreghdfe are from SSC, using Stata 17.

And thanks for the link to the related post. However, it's not the statistics stored in e(cdf) or e(sttat) that I would like to export, which are 11560 and 3.3 respectively, but the values shown in the screenshot above. I'm still not understanding where those values might be stored.

[Andrew Musau]

The idea is to reproduce the statistics from the first-stage(s) if they are not included in the stored results. At some point, you need to try and understand what the code is doing.

webuse nlswork, clear
ivreghdfe ln_w age c.age#c.age not_smsa (tenure = union south), absorb(id) first
g sample = e(sample)
local excluded= e(exexog)
reghdfe `e(instd)' `e(insts)' if sample, absorb(`e(absvars)')
test `excluded'

Res.:

. ivreghdfe ln_w age c.age#c.age not_smsa (tenure = union south), absorb(id) first
(dropped 673 singleton observations)
(MWFE estimator converged in 1 iterations)

First-stage regressions
-----------------------


First-stage regression of tenure:

Statistics consistent for homoskedasticity only
Number of obs =                  18334
------------------------------------------------------------------------------
      tenure |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       union |   .3896861   .0706568     5.52   0.000     .2511901    .5281821
       south |  -.4296172   .1349122    -3.18   0.001    -.6940618   -.1651726
         age |   .0863031   .0343238     2.51   0.012     .0190241    .1535821
             |
 c.age#c.age |   .0039115    .000549     7.12   0.000     .0028353    .0049876
             |
    not_smsa |  -.3552488   .1269318    -2.80   0.005    -.6040507   -.1064468
------------------------------------------------------------------------------
F test of excluded instruments:
  F(  2, 14868) =    21.29
  Prob > F      =   0.0000
Sanderson-Windmeijer multivariate F test of excluded instruments:
  F(  2, 14868) =    21.29
  Prob > F      =   0.0000



Summary results for first-stage regressions
-------------------------------------------

                                           (Underid)            (Weak id)
Variable     | F(  2, 14868)  P-val | SW Chi-sq(  2) P-val | SW F(  2, 14868)
tenure       |      21.29    0.0000 |       52.50   0.0000 |       21.29

Stock-Yogo weak ID F test critical values for single endogenous regressor:
                                   10% maximal IV size             19.93
                                   15% maximal IV size             11.59
                                   20% maximal IV size              8.75
                                   25% maximal IV size              7.25
Source: Stock-Yogo (2005).  Reproduced by permission.
NB: Critical values are for Sanderson-Windmeijer F statistic.

Underidentification test
Ho: matrix of reduced form coefficients has rank=K1-1 (underidentified)
Ha: matrix has rank=K1 (identified)
Anderson canon. corr. LM statistic       Chi-sq(2)=52.35    P-val=0.0000

Weak identification test
Ho: equation is weakly identified
Cragg-Donald Wald F statistic                                      21.29

Stock-Yogo weak ID test critical values for K1=1 and L1=2:
                                   10% maximal IV size             19.93
                                   15% maximal IV size             11.59
                                   20% maximal IV size              8.75
                                   25% maximal IV size              7.25
Source: Stock-Yogo (2005).  Reproduced by permission.

Weak-instrument-robust inference
Tests of joint significance of endogenous regressors B1 in main equation
Ho: B1=0 and orthogonality conditions are valid
Anderson-Rubin Wald test           F(2,14868)=   126.26     P-val=0.0000
Anderson-Rubin Wald test           Chi-sq(2)=    311.39     P-val=0.0000
Stock-Wright LM S statistic        Chi-sq(2)=    306.19     P-val=0.0000

Number of observations               N  =      18334
Number of regressors                 K  =          4
Number of endogenous regressors      K1 =          1
Number of instruments                L  =          5
Number of excluded instruments       L1 =          2

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

Estimates efficient for homoskedasticity only
Statistics consistent for homoskedasticity only

                                                      Number of obs =    18334
                                                      F(  4, 14869) =    74.14
                                                      Prob > F      =   0.0000
Total (centered) SS     =  1119.473248                Centered R2   =  -4.2766
Total (uncentered) SS   =  1119.473248                Uncentered R2 =  -4.2766
Residual SS             =  5907.007821                Root MSE      =    .6303

------------------------------------------------------------------------------
     ln_wage |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      tenure |   .2403531   .0373419     6.44   0.000     .1671583    .3135478
         age |   .0118437   .0090032     1.32   0.188    -.0058037    .0294912
             |
 c.age#c.age |  -.0012145   .0001968    -6.17   0.000    -.0016003   -.0008286
             |
    not_smsa |  -.0167178   .0339236    -0.49   0.622    -.0832123    .0497767
------------------------------------------------------------------------------
Underidentification test (Anderson canon. corr. LM statistic):          52.345
                                                   Chi-sq(2) P-val =    0.0000
------------------------------------------------------------------------------
Weak identification test (Cragg-Donald Wald F statistic):               21.286
Stock-Yogo weak ID test critical values: 10% maximal IV size             19.93
                                         15% maximal IV size             11.59
                                         20% maximal IV size              8.75
                                         25% maximal IV size              7.25
Source: Stock-Yogo (2005).  Reproduced by permission.
------------------------------------------------------------------------------
Sargan statistic (overidentification test of all instruments):           1.691
                                                   Chi-sq(1) P-val =    0.1935
------------------------------------------------------------------------------
Instrumented:         tenure
Included instruments: age c.age#c.age not_smsa
Excluded instruments: union south
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 |
-------------+---------------------------------------|
      idcode |      3461           0        3461     |
-----------------------------------------------------+

. 
. g sample = e(sample)

. 
. local excluded= e(exexog)

. 
. reghdfe `e(instd)' `e(insts)' if sample, absorb(`e(absvars)')
(MWFE estimator converged in 1 iterations)

HDFE Linear regression                            Number of obs   =     18,334
Absorbing 1 HDFE group                            F(   5,  14868) =    1285.83
                                                  Prob > F        =     0.0000
                                                  R-squared       =     0.6832
                                                  Adj R-squared   =     0.6094
                                                  Within R-sq.    =     0.3019
                                                  Root MSE        =     2.5870

------------------------------------------------------------------------------
      tenure |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       union |   .3896861   .0706568     5.52   0.000     .2511901    .5281821
       south |  -.4296172   .1349122    -3.18   0.001    -.6940618   -.1651726
         age |   .0863031   .0343238     2.51   0.012     .0190241    .1535821
             |
 c.age#c.age |   .0039115    .000549     7.12   0.000     .0028353    .0049876
             |
    not_smsa |  -.3552488   .1269318    -2.80   0.005    -.6040507   -.1064468
       _cons |  -2.498797   .5297226    -4.72   0.000    -3.537118   -1.460475
------------------------------------------------------------------------------

Absorbed degrees of freedom:
-----------------------------------------------------+
 Absorbed FE | Categories  - Redundant  = Num. Coefs |
-------------+---------------------------------------|
      idcode |      3461           0        3461     |
-----------------------------------------------------+

. 
. test `excluded'

 ( 1)  union = 0
 ( 2)  south = 0

       F(  2, 14868) =   21.29
            Prob > F =    0.0000

.