Announcement

Collapse
No announcement yet.
X
Collapse
  • Page of 1
  • Filter
  • Time
  • Show
Clear All
new posts
  • #1

    Syntax for Xtabond2

    Hello Everyone,

    I would like to ask some expert in Xtabond2. I have a model that consist of this:

    Variable Dependent and Predetermined: Human Capital Expenditure
    Variabel Independent and Exogenous: Lag demonstration per capita

    Variable demonstration per capita was in lag because it will affect the human capital expenditure in the next period

    Hereby my command Xtabond2 with two-step robust and orthogonal


    Code:
    xtabond2 ln_humancapital_pcpt l.ln_humancapital_pcpt l.totaldemonlabour_governance l.totaldemonstudent_governance l.totaldemonpeople_governance ln_gdppcpt ln_ratiomanugdp10 i.year, gmmstyle(L.(ln_humancapital_pcpt), equation(diff) laglimits(1 .) collapse) gmmstyle(L.(ln_humancapital_pcpt), equation(level) laglimits(1 1)) gmmstyle(L.(totaldemonlabour_governance totaldemonstudent_governance totaldemonpeople_governance), equation(diff) laglimits(1 .) collapse) gmmstyle(L.(totaldemonlabour_governance totaldemonstudent_governance totaldemonpeople_governance), equation(level) laglimits(1 1)) ivstyle(i.year, equation(level)) robust twostep small orthogonal
    Hereby the result

    Code:
     xtabond2 ln_humancapital_pcpt l.ln_humancapital_pcpt l.totaldemonlabour_governance l.totaldemonstu
> dent_governance l.totaldemonpeople_governance ln_gdppcpt ln_ratiomanugdp10 i.year, gmmstyle(L.(ln_
> humancapital_pcpt), equation(diff) laglimits(1 .) collapse) gmmstyle(L.(ln_humancapital_pcpt), equ
> ation(level) laglimits(0 0) collapse) gmmstyle(L.(totaldemonlabour_governance totaldemonstudent_go
> vernance totaldemonpeople_governance), equation(diff) laglimits(1 .) collapse) gmmstyle(L.(totalde
> monlabour_governance totaldemonstudent_governance totaldemonpeople_governance), equation(level) la
> glimits(0 0) collapse) ivstyle(i.year, equation(level)) robust twostep small orthogonal
Favoring space over speed. To switch, type or click on mata: mata set matafavor speed, perm.
2006b.year dropped due to collinearity
2009.year dropped due to collinearity
Warning: Two-step estimated covariance matrix of moments is singular.
  Using a generalized inverse to calculate optimal weighting matrix for two-step estimation.
  Difference-in-Sargan/Hansen statistics may be negative.

Dynamic panel-data estimation, two-step system GMM
------------------------------------------------------------------------------
Group variable: bps_2009                        Number of obs      =      2287
Time variable : year                            Number of groups   =       378
Number of instruments = 39                      Obs per group: min =         1
F(13, 377)    =  30263.64                                      avg =      6.05
Prob > F      =     0.000                                      max =         8
----------------------------------------------------------------------------------------------
                             |              Corrected
        ln_humancapital_pcpt | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-----------------------------+----------------------------------------------------------------
        ln_humancapital_pcpt |
                         L1. |    .116792   .0632612     1.85   0.066    -.0075971    .2411811
                             |
 totaldemonlabour_governance |
                         L1. |  -.0711599   .0961021    -0.74   0.459    -.2601232    .1178033
                             |
totaldemonstudent_governance |
                         L1. |  -.0078265   .0221416    -0.35   0.724     -.051363      .03571
                             |
 totaldemonpeople_governance |
                         L1. |  -.0302341   .0402158    -0.75   0.453    -.1093095    .0488413
                             |
                  ln_gdppcpt |   .3884126    .264069     1.47   0.142    -.1308199    .9076452
           ln_ratiomanugdp10 |  -.4803554   .1529133    -3.14   0.002    -.7810253   -.1796856
                             |
                        year |
                       2007  |  -.3207633   .0490133    -6.54   0.000    -.4171371   -.2243895
                       2008  |  -.1513371   .0221617    -6.83   0.000    -.1949132   -.1077611
                       2010  |   .0001135   .0203818     0.01   0.996    -.0399627    .0401897
                       2011  |   .1900164   .0236246     8.04   0.000      .143564    .2364689
                       2012  |   .2231411   .0363974     6.13   0.000     .1515738    .2947085
                       2013  |   .2940117   .0579161     5.08   0.000     .1801325    .4078908
                       2014  |   .4031246   .0692231     5.82   0.000     .2670128    .5392364
                             |
                       _cons |   9.693363     1.2348     7.85   0.000     7.265405    12.12132
----------------------------------------------------------------------------------------------
Instruments for orthogonal deviations equation
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/8).(L.totaldemonlabour_governance L.totaldemonstudent_governance
    L.totaldemonpeople_governance) collapsed
    L(1/8).L.ln_humancapital_pcpt collapsed
Instruments for levels equation
  Standard
    2006b.year 2007.year 2008.year 2009.year 2010.year 2011.year 2012.year
    2013.year 2014.year
    _cons
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    D.(L.totaldemonlabour_governance L.totaldemonstudent_governance
    L.totaldemonpeople_governance) collapsed
    D.L.ln_humancapital_pcpt collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =  -2.56  Pr > z =  0.010
Arellano-Bond test for AR(2) in first differences: z =  -0.54  Pr > z =  0.591
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(25)   =  37.71  Prob > chi2 =  0.049
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(25)   =  15.99  Prob > chi2 =  0.915
  (Robust, but weakened by many instruments.)

Difference-in-Hansen tests of exogeneity of instrument subsets:
  GMM instruments for levels
    Hansen test excluding group:     chi2(21)   =  14.20  Prob > chi2 =  0.861
    Difference (null H = exogenous): chi2(4)    =   1.78  Prob > chi2 =  0.775
  gmm(L.ln_humancapital_pcpt, collapse eq(diff) lag(1 .))
    Hansen test excluding group:     chi2(18)   =  10.47  Prob > chi2 =  0.916
    Difference (null H = exogenous): chi2(7)    =   5.52  Prob > chi2 =  0.596
  gmm(L.ln_humancapital_pcpt, collapse eq(level) lag(0 0))
    Hansen test excluding group:     chi2(24)   =  15.79  Prob > chi2 =  0.896
    Difference (null H = exogenous): chi2(1)    =   0.20  Prob > chi2 =  0.653
  gmm(L.totaldemonlabour_governance L.totaldemonstudent_governance L.totaldemonpeople_governance, co
> llapse eq(diff) lag(1 .))
    Hansen test excluding group:     chi2(5)    =   6.24  Prob > chi2 =  0.283
    Difference (null H = exogenous): chi2(20)   =   9.74  Prob > chi2 =  0.973
  gmm(L.totaldemonlabour_governance L.totaldemonstudent_governance L.totaldemonpeople_governance, co
> llapse eq(level) lag(0 0))
    Hansen test excluding group:     chi2(22)   =  14.94  Prob > chi2 =  0.865
    Difference (null H = exogenous): chi2(3)    =   1.05  Prob > chi2 =  0.790
  iv(2006b.year 2007.year 2008.year 2009.year 2010.year 2011.year 2012.year 2013.year 2014.year, eq(
> level))
    Hansen test excluding group:     chi2(18)   =   8.64  Prob > chi2 =  0.967
    Difference (null H = exogenous): chi2(7)    =   7.34  Prob > chi2 =  0.394
    I was wondering whether I perform the syntax for Xtabond2 right?

    Thank you very much.
    Last edited by Reza Mahardika; 23 Jun 2022, 09:36.
    Tags: None
  • #2
    I cannot detect any obvious problem. The following presentation might be useful:
    https://www.kripfganz.de/stata/

    Comment

    • #3
      Thank you very much Sebastian for your comment! I will check your presentation.

      Comment

      • #4
        Hi Sebastian, I am sorry I want to ask again. So I want to put the second lag of dependent variable and in the model, all variable is endogenous expect time dummies and profileration of district


        Code:
        xtabond2 ln_socialassistance_pcpt L.ln_socialassistance_pcpt L2.ln_socialassistance_pcpt  L.totaldemonlabour_governance L.totaldemonstudent_governance L.totaldemonpeople_governance l.ln_ratiomanugdp10 l.poverty_rate i.bpk_audit  i.profileration i.year, gmm(L.ln_socialassistance_pcpt L2.ln_socialassistance_pcpt L.totaldemonlabour_governance L.totaldemonstudent_governance L.totaldemonpeople_governance L.ln_ratiomanugdp10 L.poverty_rate i.bpk_audit , lag(1 4) eq(diff) collapse) gmm(L.ln_socialassistance_pcpt L2.ln_socialassistance_pcpt L.totaldemonlabour_governance L.totaldemonstudent_governance L.totaldemonpeople_governance l.ln_ratiomanugdp10 l.poverty_rate i.bpk_audit , lag(1 4) eq(level) collapse)  iv(i.year i.profileration, eq(level)) robust twostep small orthogonal
        My code is like this, and I put the second lag of dependent variable in the gmm-style instrument. Did I do it right? Or should I just use the first lag of dependent variable in the instrument? Thank you so much. Hereby the result

        Code:
        Dynamic panel-data estimation, two-step system GMM
------------------------------------------------------------------------------
Group variable: bps_2009                        Number of obs      =      2665
Time variable : year                            Number of groups   =       490
Number of instruments = 113                     Obs per group: min =         1
F(20, 489)    =  11762.17                                      avg =      5.44
Prob > F      =     0.000                                      max =         6
----------------------------------------------------------------------------------------------
                             |              Corrected
    ln_socialassistance_pcpt | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-----------------------------+----------------------------------------------------------------
    ln_socialassistance_pcpt |
                         L1. |    .515031   .0373566    13.79   0.000     .4416317    .5884302
                         L2. |   .1273237   .0324987     3.92   0.000     .0634693     .191178
                             |
 totaldemonlabour_governance |
                         L1. |  -.4141795   .2981657    -1.39   0.165    -1.000024    .1716646
                             |
totaldemonstudent_governance |
                         L1. |  -.0821481   .0316809    -2.59   0.010    -.1443956   -.0199006
                             |
 totaldemonpeople_governance |
                         L1. |   .0843339   .0956623     0.88   0.378    -.1036261    .2722939
                             |
           ln_ratiomanugdp10 |
                         L1. |  -.0972017   .0942801    -1.03   0.303    -.2824458    .0880423
                             |
                poverty_rate |
                         L1. |   .0134834   .0095122     1.42   0.157    -.0052064    .0321732
                             |
                   bpk_audit |
                          2  |  -.0758461   .1147817    -0.66   0.509    -.3013722    .1496801
                          3  |  -.1860249   .2248165    -0.83   0.408    -.6277504    .2557007
                          4  |  -.2313119   .1661175    -1.39   0.164    -.5577041    .0950802
                          5  |   .0448973   .1379185     0.33   0.745    -.2260888    .3158834
                          7  |  -.0284265   5.014315    -0.01   0.995    -9.880689    9.823836
                          8  |  -.1824176   12.41633    -0.01   0.988    -24.57835    24.21352
                             |
               profileration |
                          2  |   -.239869   .1137174    -2.11   0.035    -.4633041   -.0164339
                          3  |  -.2663717   .1111554    -2.40   0.017     -.484773   -.0479705
                             |
                        year |
                       2009  |   .4931072   .1082158     4.56   0.000     .2804817    .7057326
                       2010  |  -.0796944   .0917183    -0.87   0.385     -.259905    .1005162
                       2011  |   .1658735   .0743814     2.23   0.026     .0197268    .3120202
                       2012  |    -.80685   .0923656    -8.74   0.000    -.9883324   -.6253675
                       2013  |  -.0189647   .0660597    -0.29   0.774    -.1487606    .1108313
                             |
                       _cons |    3.39348   .4106321     8.26   0.000     2.586659    4.200301
----------------------------------------------------------------------------------------------
Instruments for orthogonal deviations equation
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/4).(L.ln_socialassistance_pcpt L2.ln_socialassistance_pcpt
    L.totaldemonlabour_governance L.totaldemonstudent_governance
    L.totaldemonpeople_governance L.ln_ratiomanugdp10 L.poverty_rate
    1b.bpk_audit 2.bpk_audit 3.bpk_audit 4.bpk_audit 5.bpk_audit 6.bpk_audit
    7.bpk_audit 8.bpk_audit 9.bpk_audit) collapsed
Instruments for levels equation
  Standard
    2007b.year 2008.year 2009.year 2010.year 2011.year 2012.year 2013.year
    2014.year 1b.profileration 2.profileration 3.profileration
    _cons
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    DL(1/4).(L.ln_socialassistance_pcpt L2.ln_socialassistance_pcpt
    L.totaldemonlabour_governance L.totaldemonstudent_governance
    L.totaldemonpeople_governance L.ln_ratiomanugdp10 L.poverty_rate
    1b.bpk_audit 2.bpk_audit 3.bpk_audit 4.bpk_audit 5.bpk_audit 6.bpk_audit
    7.bpk_audit 8.bpk_audit 9.bpk_audit) collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =  -7.81  Pr > z =  0.000
Arellano-Bond test for AR(2) in first differences: z =  -0.49  Pr > z =  0.625
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(92)   = 114.10  Prob > chi2 =  0.059
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(92)   =  89.68  Prob > chi2 =  0.549
  (Robust, but weakened by many instruments.)

Difference-in-Hansen tests of exogeneity of instrument subsets:
  GMM instruments for levels
    Hansen test excluding group:     chi2(36)   =  45.81  Prob > chi2 =  0.127
    Difference (null H = exogenous): chi2(56)   =  43.86  Prob > chi2 =  0.881
  gmm(L.ln_socialassistance_pcpt L2.ln_socialassistance_pcpt L.totaldemonlabour_governance L.totalde
> monstudent_governance L.totaldemonpeople_governance L.ln_ratiomanugdp10 L.poverty_rate 1b.bpk_audi
> t 2.bpk_audit 3.bpk_audit 4.bpk_audit 5.bpk_audit 6.bpk_audit 7.bpk_audit 8.bpk_audit 9.bpk_audit,
>  collapse eq(level) lag(1 4))
    Hansen test excluding group:     chi2(36)   =  45.81  Prob > chi2 =  0.127
    Difference (null H = exogenous): chi2(56)   =  43.86  Prob > chi2 =  0.881
  iv(2007b.year 2008.year 2009.year 2010.year 2011.year 2012.year 2013.year 2014.year 1b.profilerati
> on 2.profileration 3.profileration, eq(level))
    Hansen test excluding group:     chi2(85)   =  75.51  Prob > chi2 =  0.760

        Comment

        • #5
          It does not make much of a difference whether you add the 2nd lag to the instruments or not. Most of the instruments created for the first and for the second lag will overlap and drop out.

          gmm(L.Y L2.Y, lag(1 4) eq(diff) collapse)
          will create instruments L(1/4).L.Y L(1/4).L2.Y for the transformed model. After removing redundant instruments, this is in compact form L(2/6).Y.

          gmm(L.Y, lag(1 4) eq(diff) collapse)
          will create instruments L(1/4).L.Y. Consolidating the lag operators, this is L(2/5).Y.
          https://www.kripfganz.de/stata/

          Comment

          • #6
            Ah I see, thank you very much Sebastian . Your comment already saved my research. I am very grateful for your help

            Comment

            • #7
              Hi Sebastian Kripfganz, i am sorry to bother you again. I would like to ask question regarding predetermined variable. Basically I treat all RHS variable as predetermined variable because theoretical review. My research is about the impact of demonstration on fiscal spending and since the fiscal spending is already decided in t-1, therefore I use all rhs variable in t-1, except the fiscal spending because it is in t-1 and t-2.

              For example, I treat first lag of total demonstration variable (L.totaldemons_governance) as predetermined variable. Therefore, I use the L.totaldemons_governance as instrument with first and maximum second lag, hereby the code that I made

              Code:
              xtabond2 ln_socialassistance_pcpt L.ln_socialassistance_pcpt L2.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt L.poverty_rate L.ln_ratiomanugdp10  L.i.bpk_audit i.profileration i.year, gmm(L.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt L.poverty_rate L.ln_ratiomanugdp10  L.i.bpk_audit, lag(1 2) eq(diff) collapse) gmm(L.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt L.poverty_rate L.ln_ratiomanugdp10 L.i.bpk_audit, lag(1 2) eq(level) collapse) iv(i.year i.profileration , eq(level)) robust twostep small orthogonal artest(3)
              My understanding is, with lag (1 2) as instrument, it means that "using the first lag and second lag of the first lag of demonstration variable as instrument" right? Or should my interpretation is "using the second lag and third lag of the first lag of demonstration variable as instrument"? I am sorry if my language confusing you.

              Hereby the result. Thank you very much

              Code:
              . xtabond2 ln_socialassistance_pcpt L.ln_socialassistance_pcpt L2.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt L.poverty_rate L.ln_
> ratiomanugdp10  L.i.bpk_audit i.profileration i.year, gmm(L.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt L.poverty_rate L.ln_rati
> omanugdp10  L.i.bpk_audit, lag(1 2) eq(diff) collapse) gmm(L.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt L.poverty_rate L.ln_rat
> iomanugdp10 L.i.bpk_audit, lag(1 2) eq(level) collapse) iv(i.year i.profileration , eq(level)) robust twostep small orthogonal artest(3)
Favoring space over speed. To switch, type or click on mata: mata set matafavor speed, perm.
1bL.bpk_audit dropped due to collinearity
1b.profileration dropped due to collinearity
2007b.year dropped due to collinearity
2008.year dropped due to collinearity
2014.year dropped due to collinearity
Warning: Two-step estimated covariance matrix of moments is singular.
  Using a generalized inverse to calculate optimal weighting matrix for two-step estimation.
  Difference-in-Sargan/Hansen statistics may be negative.

Dynamic panel-data estimation, two-step system GMM
------------------------------------------------------------------------------
Group variable: bps_2009                        Number of obs      =      2351
Time variable : year                            Number of groups   =       466
Number of instruments = 52                      Obs per group: min =         1
F(19, 465)    =   2762.00                                      avg =      5.05
Prob > F      =     0.000                                      max =         6
------------------------------------------------------------------------------------------
                         |              Corrected
ln_socialassistance_pcpt | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------------------+----------------------------------------------------------------
ln_socialassistance_pcpt |
                     L1. |   .5162112   .0470008    10.98   0.000     .4238509    .6085715
                     L2. |   .1193599   .0367344     3.25   0.001     .0471738    .1915459
                         |
  totaldemons_governance |
                     L1. |   -.085512   .0574624    -1.49   0.137    -.1984301    .0274062
                         |
              ln_gdppcpt |
                     L1. |   1.431008   .5087517     2.81   0.005     .4312712    2.430746
                         |
            poverty_rate |
                     L1. |    .045899   .0262043     1.75   0.081    -.0055945    .0973926
                         |
       ln_ratiomanugdp10 |
                     L1. |  -.1618793    .270246    -0.60   0.549    -.6929339    .3691754
                         |
             L.bpk_audit |
                      2  |   .2516471   .1787279     1.41   0.160    -.0995672    .6028614
                      3  |   .2577831    .250493     1.03   0.304    -.2344554    .7500217
                      4  |   .1814778   .2491964     0.73   0.467    -.3082127    .6711683
                      5  |   .2175038   .2202405     0.99   0.324    -.2152861    .6502936
                      7  |   .2499165   .3977855     0.63   0.530    -.5317633    1.031596
                      8  |  -3.060321   10.47693    -0.29   0.770    -23.64832    17.52768
                         |
           profileration |
                      2  |  -.2892703   .2551492    -1.13   0.257    -.7906586    .2121179
                      3  |  -.3781023   .2371174    -1.59   0.111    -.8440567    .0878521
                         |
                    year |
                   2009  |   .7237284    .182889     3.96   0.000     .3643371     1.08312
                   2010  |  -.0017756   .1401468    -0.01   0.990    -.2771752     .273624
                   2011  |   .2823996   .1209671     2.33   0.020     .0446897    .5201095
                   2012  |  -.7830696   .1067195    -7.34   0.000    -.9927819   -.5733574
                   2013  |   .0707747    .079812     0.89   0.376    -.0860622    .2276115
                         |
                   _cons |  -21.58386   9.018366    -2.39   0.017    -39.30566   -3.862063
------------------------------------------------------------------------------------------
Instruments for orthogonal deviations equation
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/2).(L.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt
    L.poverty_rate L.ln_ratiomanugdp10 1bL.bpk_audit 2L.bpk_audit 3L.bpk_audit
    4L.bpk_audit 5L.bpk_audit 7L.bpk_audit 8L.bpk_audit) collapsed
Instruments for levels equation
  Standard
    2007b.year 2008.year 2009.year 2010.year 2011.year 2012.year 2013.year
    2014.year 1b.profileration 2.profileration 3.profileration
    _cons
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    DL(1/2).(L.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt
    L.poverty_rate L.ln_ratiomanugdp10 1bL.bpk_audit 2L.bpk_audit 3L.bpk_audit
    4L.bpk_audit 5L.bpk_audit 7L.bpk_audit 8L.bpk_audit) collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =  -7.91  Pr > z =  0.000
Arellano-Bond test for AR(2) in first differences: z =  -0.99  Pr > z =  0.321
Arellano-Bond test for AR(3) in first differences: z =   0.45  Pr > z =  0.651
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(32)   =  46.32  Prob > chi2 =  0.049
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(32)   =  24.93  Prob > chi2 =  0.809
  (Robust, but weakened by many instruments.)

Difference-in-Hansen tests of exogeneity of instrument subsets:
  GMM instruments for levels
    Hansen test excluding group:     chi2(10)   =  11.99  Prob > chi2 =  0.286
    Difference (null H = exogenous): chi2(22)   =  12.94  Prob > chi2 =  0.935
  gmm(L.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt L.poverty_rate L.ln_ratiomanugdp10 1bL.bpk_audit 2L.bpk_audit 3L.bpk_audit 4L.
> bpk_audit 5L.bpk_audit 7L.bpk_audit 8L.bpk_audit, collapse eq(diff) lag(1 2))
    Hansen test excluding group:     chi2(10)   =   6.56  Prob > chi2 =  0.766
    Difference (null H = exogenous): chi2(22)   =  18.37  Prob > chi2 =  0.684
  gmm(L.ln_socialassistance_pcpt L.totaldemons_governance L.ln_gdppcpt L.poverty_rate L.ln_ratiomanugdp10 1bL.bpk_audit 2L.bpk_audit 3L.bpk_audit 4L.
> bpk_audit 5L.bpk_audit 7L.bpk_audit 8L.bpk_audit, collapse eq(level) lag(1 2))
    Hansen test excluding group:     chi2(10)   =  11.99  Prob > chi2 =  0.286
    Difference (null H = exogenous): chi2(22)   =  12.94  Prob > chi2 =  0.935
  iv(2007b.year 2008.year 2009.year 2010.year 2011.year 2012.year 2013.year 2014.year 1b.profileration 2.profileration 3.profileration, eq(level))
    Hansen test excluding group:     chi2(25)   =  20.54  Prob > chi2 =  0.718
    Difference (null H = exogenous): chi2(7)    =   4.39  Prob > chi2 =  0.734

              Comment

              • #8
                Let's first ignore the fact that you specified the orthogonal option. Without that option, you would use lags 1 and 2 of the first lag of total demonstration as instruments, i.e. lags 2 and 3 of total demonstration. Your first interpretation would be correct.

                Now, at the risk of confusing you even more, with the orthogonal option everything becomes more complicated. With that option, xtabond2 internally shifts the observations by one period. When you specify lags 1 and 2 as instruments, these are now actually lags 0 and 1 of the first lag of total demonstration. This is still what it is supposed to be when the first lag of total demonstration is predetermined, because the valid lag order depends on whether you consider first differences or forward-orthogonal deviations. See slide 70 of my 2019 London Stata Conference presentation.
                https://www.kripfganz.de/stata/

                Comment

                • #9
                  Thank you for all your help Sebastian Kripfganz , I am grateful for all your helps

                  Comment

                  Previous Next
                  Working...
                  X