Announcement

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

    Separate analysis of variables in GMM system

    Dear STATA users
    I am analysing the impact on GDP (dependent variable) of 2 independent variables related to education (Grad10 and Grad18). I consider these variables to be predetermined. I apply system GMM using the xtabond2 command. In addition, I use time dummies
    The issue is that when I analyse the two independent variables (Grad10 and Grad18):
    - the parameters are not significant for either of the two variables
    - the constant is not significant either
    - the Hansen value is higher than 3

    On the other hand, if I analyse the independent variables separately, i.e. I run a panel for each variable, with respect to the Grad10 variable.
    - The parameter is significant
    - The constant in the model is significant
    - Hansen's value is between 1-3.
    And with respect to the variable Grad18 (also analysed separately), neither the parameter nor the constant is significant.

    My question is, does the fact that the model calculated exclusively with variable Grad10 is correct add any relevant information?

    Of course, I understand that the significance of the parameters changes when you add variables, but the problem is that adding variables messes up the model, and furthermore, the variable that was significant when analysed independently is no longer significant. In other words, can it get to the point where variables that are significant (when analysed independently) are no longer significant? Sorry if this is a very basic question, but I am a beginner.
    Thank you very much for your help.
    Tags: gmm, panel data, regression, system, xtabond2
  • #2
    Model with two predetermined variables (Grad10 and Grad18):
    Code:
     xtabond2 GDP07 L.GDP07 Grad10 Grad18  i.year,  gmm (l.GDP07  Grad10 Grad18  , collapse lag (1 5))
iv (i.year, eq(level)) small twostep robust nodiffsargan
Favoring speed over space. To switch, type or click on mata: mata set matafavor space, perm.
2014b.year dropped due to collinearity
2016.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.

Dynamic panel-data estimation, two-step system GMM
------------------------------------------------------------------------------
Group variable: iden                            Number of obs      =       266
Time variable : year                            Number of groups   =        38
Number of instruments = 25                      Obs per group: min =         7
F(9, 37)      =     85.12                                      avg =      7.00
Prob > F      =     0.000                                      max =         7
------------------------------------------------------------------------------
             |              Corrected
       GDP07 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       GDP07 |
         L1. |   -.167262   .1005861    -1.66   0.105    -.3710687    .0365448
             |
      Grad10 |   .3723619   .3193583     1.17   0.251    -.2747195    1.019443
      Grad18 |  -.0491527   .3169557    -0.16   0.878     -.691366    .5930606
             |
        year |
       2015  |   .4631921   .3978123     1.16   0.252    -.3428523    1.269236
       2017  |   .5534326   .2484393     2.23   0.032     .0500467    1.056819
       2018  |   .4181025   .2504098     1.67   0.103     -.089276     .925481
       2019  |  -.5242803    .402561    -1.30   0.201    -1.339946    .2913858
       2020  |  -7.133365   .5492023   -12.99   0.000    -8.246155   -6.020576
       2021  |   2.355585   .8906832     2.64   0.012     .5508892    4.160281
             |
       _cons |   2.165791   5.363136     0.40   0.689    -8.700955    13.03254
------------------------------------------------------------------------------
Instruments for first differences equation
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/5).(L.GDP07 Grad10 Grad18) collapsed
Instruments for levels equation
  Standard
    2014b.year 2015.year 2016.year 2017.year 2018.year 2019.year 2020.year
    2021.year
    _cons
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    D.(L.GDP07 Grad10 Grad18) collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =  -2.93  Pr > z =  0.003
Arellano-Bond test for AR(2) in first differences: z =  -1.23  Pr > z =  0.217
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(15)   =  42.54  Prob > chi2 =  0.000
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(15)   =  17.11  Prob > chi2 =  0.313
  (Robust, but weakened by many instruments.)

    Comment

    • #3
      Model with one predetermined variable (Grad10):
      Code:
       xtabond2 GDP07 L.GDP07 Grad10  i.year,  gmm (l.GDP07  Grad10  , collapse lag (1 5))
iv (i.year, eq(level)) small twostep robust nodiffsargan
Favoring speed over space. To switch, type or click on mata: mata set matafavor space, perm.
2014b.year dropped due to collinearity
2016.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.

Dynamic panel-data estimation, two-step system GMM
------------------------------------------------------------------------------
Group variable: iden                            Number of obs      =       266
Time variable : year                            Number of groups   =        38
Number of instruments = 19                      Obs per group: min =         7
F(8, 37)      =    107.47                                      avg =      7.00
Prob > F      =     0.000                                      max =         7
------------------------------------------------------------------------------
             |              Corrected
       GDP07 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       GDP07 |
         L1. |  -.1664337   .1009655    -1.65   0.108    -.3710093    .0381419
             |
      Grad10 |   .4382939   .1631216     2.69   0.011     .1077781    .7688096
             |
        year |
       2015  |   .4318397   .2790518     1.55   0.130    -.1335729    .9972524
       2017  |   .4857418   .2909114     1.67   0.103    -.1037007    1.075184
       2018  |   .2755004   .3608951     0.76   0.450    -.4557425    1.006743
       2019  |  -.5783644   .4291867    -1.35   0.186    -1.447979    .2912506
       2020  |  -7.294489   .6461618   -11.29   0.000    -8.603737   -5.985241
       2021  |   2.173442   .8908683     2.44   0.020     .3683715    3.978513
             |
       _cons |   1.357833   .6592576     2.06   0.047     .0220506    2.693616
------------------------------------------------------------------------------
Instruments for first differences equation
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/5).(L.GDP07 Grad10) collapsed
Instruments for levels equation
  Standard
    2014b.year 2015.year 2016.year 2017.year 2018.year 2019.year 2020.year
    2021.year
    _cons
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    D.(L.GDP07 Grad10) collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =  -3.23  Pr > z =  0.001
Arellano-Bond test for AR(2) in first differences: z =  -1.28  Pr > z =  0.200
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(10)   =  40.89  Prob > chi2 =  0.000
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(10)   =  14.03  Prob > chi2 =  0.172
  (Robust, but weakened by many instruments.)

      Comment

      • #4
        Model with one predetermined variable (Grad18):
        Code:
        xtabond2 GDP07 L.GDP07 Grad18  i.year,  gmm (l.GDP07  Grad18  , collapse lag (1 5))
iv (i.year, eq(level)) small twostep robust nodiffsargan
Favoring speed over space. To switch, type or click on mata: mata set matafavor space, perm.
2014b.year dropped due to collinearity
2016.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.

Dynamic panel-data estimation, two-step system GMM
------------------------------------------------------------------------------
Group variable: iden                            Number of obs      =       266
Time variable : year                            Number of groups   =        38
Number of instruments = 19                      Obs per group: min =         7
F(8, 37)      =     84.53                                      avg =      7.00
Prob > F      =     0.000                                      max =         7
------------------------------------------------------------------------------
             |              Corrected
       GDP07 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       GDP07 |
         L1. |  -.1135327   .1123855    -1.01   0.319    -.3412473    .1141819
             |
      Grad18 |   .0449471   .1829238     0.25   0.807    -.3256917    .4155859
             |
        year |
       2015  |   .2608682   .2821142     0.92   0.361    -.3107493    .8324858
       2017  |   .7053932   .2723121     2.59   0.014     .1536365     1.25715
       2018  |   .5548799   .2901237     1.91   0.064    -.0329664    1.142726
       2019  |  -.2013023   .3072198    -0.66   0.516    -.8237887    .4211841
       2020  |  -6.671747   .6351008   -10.51   0.000    -7.958584   -5.384911
       2021  |   3.157226   .9401639     3.36   0.002     1.252273    5.062179
             |
       _cons |    2.22481   2.471956     0.90   0.374    -2.783849     7.23347
------------------------------------------------------------------------------
Instruments for first differences equation
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/5).(L.GDP07 Grad18) collapsed
Instruments for levels equation
  Standard
    2014b.year 2015.year 2016.year 2017.year 2018.year 2019.year 2020.year
    2021.year
    _cons
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    D.(L.GDP07 Grad18) collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =  -2.86  Pr > z =  0.004
Arellano-Bond test for AR(2) in first differences: z =  -1.37  Pr > z =  0.169
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(10)   =  16.71  Prob > chi2 =  0.081
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(10)   =  13.58  Prob > chi2 =  0.193
  (Robust, but weakened by many instruments.)

        Comment

        Previous Next
        Working...
        X