Exporting First Stage Results of 2SLS Regression Using XTIVREG

Farhan Hasnat

Join Date: Dec 2021
Posts: 91

Exporting First Stage Results of 2SLS Regression Using XTIVREG

03 Nov 2022, 12:29

I am trying to export the results of first stage regression in a word file table. However, using Outreg2 i am only able to get the second stage regression results

Below is my regression results and the code i ran to extract results to word file

Code:

outreg2 using Reg2Event2.doc, append stat(coef tstat) ctitle(2SLS, AEM) addstat("F-statistic", e(F), "R-Squared", e(r2_b)) addtext(Industry Fixed Effect, YES, Year Fixed Effect, NO) dec(4) tdec(2)

Code:

xtivreg ABS_DA_w POST_REG INBD POSTREG_X_REM ROA_w Size MTB_x_w LEV_w NOA_X_w (REM_PROXY_w    =    Lag_ABS_DA),    fe    first    small

First-stage within regression

Fixed-effects (within) regression               Number of obs     =      1,315
Group variable: id                              Number of groups  =        281

R-squared:                                      Obs per group:
Within  = 0.4889                                         min =          1
Between = 0.8082                                         avg =        4.7
Overall = 0.6542                                         max =          8

F(9,1025)         =     108.93
corr(u_i, Xb) = 0.4034                          Prob > F          =     0.0000


REM_PROXY_w  Coefficient  Std. err.      t    P>t     [95% conf. interval]

POST_REG    .0186762   .0080652     2.32   0.021       .00285    .0345024
INBD    .0783671   .0643577     1.22   0.224    -.0479208    .2046551
POSTREG_X_REM    .6598204   .0222133    29.70   0.000     .6162317    .7034092
ROA_w   -.1308316   .0420391    -3.11   0.002    -.2133242    -.048339
Size    .0271795   .0140415     1.94   0.053    -.0003739    .0547329
MTB_x_w   -.0133533   .0042623    -3.13   0.002    -.0217171   -.0049895
LEV_w    -.017941   .0324372    -0.55   0.580    -.0815918    .0457098
NOA_X_w    .0118952   .0056894     2.09   0.037      .000731    .0230594
Lag_ABS_DA    .0226871   .0434657     0.52   0.602    -.0626047    .1079789
_cons   -.4367363   .2118211    -2.06   0.039    -.8523888   -.0210837

sigma_u   .11201445
sigma_e   .10246778
rho   .54442229   (fraction of variance due to u_i)

F test that all u_i=0: F(280, 1025) = 3.53                   Prob > F = 0.0000

Fixed-effects (within) IV regression            Number of obs     =      1,315
Group variable: id                              Number of groups  =        281

R-squared:                                      Obs per group:
Within  =      .                                         min =          1
Between = 0.0018                                         avg =        4.7
Overall = 0.0000                                         max =          8

F(290,1025)       =      0.43
corr(u_i, Xb) = -0.7818                         Prob > F          =     0.9196


ABS_DA_w  Coefficient  Std. err.      t    P>t     [95% conf. interval]

REM_PROXY_w    -3.09999   6.098128    -0.51   0.611    -15.06623    8.866251
POST_REG    .0653209   .1154681     0.57   0.572      -.16126    .2919018
INBD    .2044777   .5204369     0.39   0.694    -.8167657    1.225721
POSTREG_X_REM    2.040223   4.023804     0.51   0.612    -5.855611    9.936058
ROA_w   -.6007786   .7939438    -0.76   0.449    -2.158719    .9571623
Size    .1016277   .1747267     0.58   0.561    -.2412353    .4444907
MTB_x_w   -.0389572   .0808277    -0.48   0.630    -.1975638    .1196495
LEV_w    -.031798   .1456086    -0.22   0.827     -.317523     .253927
NOA_X_w    .0354542   .0755283     0.47   0.639    -.1127536    .1836621
_cons   -1.526531   2.788213    -0.55   0.584    -6.997789    3.944727

sigma_u   .36017545
sigma_e   .32614935
rho    .5494557   (fraction of variance due to u_i)

F test that all u_i=0: F(280,1025) =     0.09            Prob > F    = 1.0000

Instrumented: REM_PROXY_w
Instruments: POST_REG INBD POSTREG_X_REM ROA_w Size MTB_x_w LEV_w NOA_X_w
Lag_ABS_DA

Any help would be helpful. Thank you

Tags: None

Andrew Musau

Join Date: Oct 2014
Posts: 10275

04 Nov 2022, 03:27

outreg2 is from SSC, as you are asked to explain. Perhaps just run the first stage using xtreg.

Code:

webuse nlswork, clear
xtset idcode year
xtivreg ln_w age c.age#c.age not_smsa (tenure = union south), fe first
est sto second
xtreg `e(instd)' `e(insts)' if e(sample), fe
est sto first
outreg2 [first second] using myfile.rtf, replace nocons

Res.:

Code:

. xtivreg ln_w age c.age#c.age not_smsa (tenure = union south), fe first

First-stage within regression

Fixed-effects (within) regression               Number of obs     =     19,007
Group variable: idcode                          Number of groups  =      4,134

R-squared:                                      Obs per group:
     Within  = 0.3019                                         min =          1
     Between = 0.0578                                         avg =        4.6
     Overall = 0.1289                                         max =         12

                                                F(5,14868)        =    1285.83
corr(u_i, Xb) = -0.1871                         Prob > F          =     0.0000

------------------------------------------------------------------------------
      tenure | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         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
       union |   .3896861   .0706568     5.52   0.000     .2511901    .5281821
       south |  -.4296172   .1349122    -3.18   0.001    -.6940618   -.1651726
       _cons |  -2.554764   .5293374    -4.83   0.000     -3.59233   -1.517197
-------------+----------------------------------------------------------------
     sigma_u |  3.1334845
     sigma_e |  2.5869529
         rho |  .59467598   (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(4133, 14868) = 6.40                 Prob > F = 0.0000

Fixed-effects (within) IV regression            Number of obs     =     19,007
Group variable: idcode                          Number of groups  =      4,134

R-squared:                                      Obs per group:
     Within  =      .                                         min =          1
     Between = 0.1304                                         avg =        4.6
     Overall = 0.0897                                         max =         12

                                                Wald chi2(4)      =  147926.58
corr(u_i, Xb) = -0.6843                         Prob > chi2       =     0.0000

------------------------------------------------------------------------------
     ln_wage | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
      tenure |   .2403531   .0373419     6.44   0.000     .1671643    .3135419
         age |   .0118437   .0090032     1.32   0.188    -.0058023    .0294897
             |
 c.age#c.age |  -.0012145   .0001968    -6.17   0.000    -.0016003   -.0008286
             |
    not_smsa |  -.0167178   .0339236    -0.49   0.622    -.0832069    .0497713
       _cons |   1.678287   .1626657    10.32   0.000     1.359468    1.997106
-------------+----------------------------------------------------------------
     sigma_u |  .70661941
     sigma_e |  .63029359
         rho |  .55690561   (fraction of variance due to u_i)
------------------------------------------------------------------------------
 F test that all u_i=0: F(4133,14869) =     1.44          Prob > F    = 0.0000
------------------------------------------------------------------------------
Instrumented: tenure
 Instruments: age c.age#c.age not_smsa union south

Click image for larger version

Name: Capture.PNG
Views: 1
Size: 12.5 KB
ID: 1687989

Last edited by Andrew Musau; 04 Nov 2022, 04:03.

Comment

Farhan Hasnat

Join Date: Dec 2021

Posts: 91
#3

05 Nov 2022, 11:53

Andrew Musau Thank you Andrew for your kind help. This works perfectly as the run the regressions.
Comment
radwa saleh

Join Date: Jan 2018

Posts: 24
#4

02 Sep 2023, 06:50

how to report 2nd stage R-squared in this case?
Comment

Andrew Musau

Join Date: Oct 2014
Posts: 10275

02 Sep 2023, 07:29

Code:

ssc install estout, replace

Code:

webuse nlswork, clear
xtset idcode year
xtivreg ln_w age c.age#c.age not_smsa (tenure = union south), fe first
estadd scalar r2= e(r2_w)
est sto second
xtreg `e(instd)' `e(insts)' if e(sample), fe
est sto first
outreg2 [first second] using myfile.rtf, replace nocons

Comment

radwa saleh

Join Date: Jan 2018

Posts: 24
#6

03 Sep 2023, 15:41

thank you so much, that helped me much. however when trying to make sure of the numbers exported, the coefficients are the same however the t-tests are different. the ones reported in the first stage xtivreg command is different from that reported by "xtreg `e(instd)' `e(insts)' if e(sample), fe" why is that ?. moreover, if i want to report the postestimation results from the xtoverid,nois how to add this to the outreg2 . your help is much appreciated.
Comment
radwa saleh

Join Date: Jan 2018

Posts: 24
#7

03 Sep 2023, 16:07

I specifically want to export the Hansen j test the weak identification and the under identification and finally the f statistic of the first stage. i found your suggested code of the return list then outreg2 however this reported only the hansen j statistic along its chi-square.
Comment
Andrew Musau

Join Date: Oct 2014

Posts: 10275
#8

03 Sep 2023, 22:04

#4 gives a general approach: https://www.statalist.org/forums/for...-stage-outcome. If you can extract one statistic, you can extract them all.
Comment
radwa saleh

Join Date: Jan 2018

Posts: 24
#9

04 Sep 2023, 02:04

and what about the differences in the t-statistics obtained from the first stage regression using the XTIVREG and that one "xtreg `e(instd)' `e(insts)' if e(sample), fe"
Comment

Andrew Musau

Join Date: Oct 2014
Posts: 10275

#10

04 Sep 2023, 04:24

I do not spot any differences (see highlighted). Elaborate on your point.

Code:

webuse nlswork, clear
xtset idcode year
xtivreg ln_w age c.age#c.age not_smsa (tenure = union south), fe first
xtreg `e(instd)' `e(insts)' if e(sample), fe

Res.:

Code:

. xtivreg ln_w age c.age#c.age not_smsa (tenure = union south), fe first

First-stage within regression

Fixed-effects (within) regression               Number of obs     =     19,007
Group variable: idcode                          Number of groups  =      4,134

R-squared:                                      Obs per group:
     Within  = 0.3019                                         min =          1
     Between = 0.0578                                         avg =        4.6
     Overall = 0.1289                                         max =         12

                                                F(5,14868)        =    1285.83
corr(u_i, Xb) = -0.1871                         Prob > F          =     0.0000

------------------------------------------------------------------------------
      tenure | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         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
       union |   .3896861   .0706568     5.52   0.000     .2511901    .5281821
       south |  -.4296172   .1349122    -3.18   0.001    -.6940618   -.1651726
       _cons |  -2.554764   .5293374    -4.83   0.000     -3.59233   -1.517197
-------------+----------------------------------------------------------------
     sigma_u |  3.1334845
     sigma_e |  2.5869529
         rho |  .59467598   (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(4133, 14868) = 6.40                 Prob > F = 0.0000

Fixed-effects (within) IV regression            Number of obs     =     19,007
Group variable: idcode                          Number of groups  =      4,134

R-squared:                                      Obs per group:
     Within  =      .                                         min =          1
     Between = 0.1304                                         avg =        4.6
     Overall = 0.0897                                         max =         12

                                                Wald chi2(4)      =  147926.58
corr(u_i, Xb) = -0.6843                         Prob > chi2       =     0.0000

------------------------------------------------------------------------------
     ln_wage | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
      tenure |   .2403531   .0373419     6.44   0.000     .1671643    .3135419
         age |   .0118437   .0090032     1.32   0.188    -.0058023    .0294897
             |
 c.age#c.age |  -.0012145   .0001968    -6.17   0.000    -.0016003   -.0008286
             |
    not_smsa |  -.0167178   .0339236    -0.49   0.622    -.0832069    .0497713
       _cons |   1.678287   .1626657    10.32   0.000     1.359468    1.997106
-------------+----------------------------------------------------------------
     sigma_u |  .70661941
     sigma_e |  .63029359
         rho |  .55690561   (fraction of variance due to u_i)
------------------------------------------------------------------------------
 F test that all u_i=0: F(4133,14869) =     1.44          Prob > F    = 0.0000
------------------------------------------------------------------------------
Instrumented: tenure
 Instruments: age c.age#c.age not_smsa union south

. 
. xtreg `e(instd)' `e(insts)' if e(sample), fe

Fixed-effects (within) regression               Number of obs     =     19,007
Group variable: idcode                          Number of groups  =      4,134

R-squared:                                      Obs per group:
     Within  = 0.3019                                         min =          1
     Between = 0.0578                                         avg =        4.6
     Overall = 0.1289                                         max =         12

                                                F(5,14868)        =    1285.83
corr(u_i, Xb) = -0.1871                         Prob > F          =     0.0000

------------------------------------------------------------------------------
      tenure | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
         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
       union |   .3896861   .0706568     5.52   0.000     .2511901    .5281821
       south |  -.4296172   .1349122    -3.18   0.001    -.6940618   -.1651726
       _cons |  -2.554764   .5293374    -4.83   0.000     -3.59233   -1.517197
-------------+----------------------------------------------------------------
     sigma_u |  3.1334845
     sigma_e |  2.5869529
         rho |  .59467598   (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(4133, 14868) = 6.40                 Prob > F = 0.0000

.

Comment

radwa saleh

Join Date: Jan 2018

Posts: 24
#11

04 Sep 2023, 04:41

I have managed to solve the problem of the difference in the t-statistic. my question now is how to export the postestimation results after the xtoverid,nois. the link you have suggested is working on xtivreg2 not xtivreg. when trying the ereturn list, it did not include the Hansen J and the overidentification tests. any suggestions is much appreciated?
Comment
Andrew Musau

Join Date: Oct 2014

Posts: 10275
#12

04 Sep 2023, 04:50

https://www.statalist.org/forums/for...overid-command
Comment
radwa saleh

Join Date: Jan 2018

Posts: 24
#13

04 Sep 2023, 05:36

I am sorry for asking too many questions, but i have been stuck in this for couple of days. how to export the Kleibergen-Paap rk LM statistic,Cragg-Donald Wald F statistic,Kleibergen-Paap rk Wald F statistic and the f statistic of the first stage. another question is related to the xtivreg2 command, its easier because it can directly export the postestimation results. however i faced two problems with it. the first one is related to the number of observations and clusters as they are different from that reported in xtivreg. i tried the commands you have suggested in one of the forums(using xtreg to get the first stage) yet it doesn't yield the same results. the other question is related to the constant and the R-squared, for the constant using xtivreg2 its not reported in the first or 2nd stage. as for the r-squared of the first stage its also not reported. any help is appreciated?
Comment

Announcement

Exporting First Stage Results of 2SLS Regression Using XTIVREG

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment