Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: xtabond2 twostep drops variables


From   Gülsün Akın <gg.akin@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: xtabond2 twostep drops variables
Date   Mon, 27 Jun 2011 18:19:35 +0300

Dear Statalist,

I am trying to get GMM results using Stata 9.2. I use the -xtabond2-
command. When I use the -twostep- option (either in difference or
system GMM) many important non-dummy variables drop. I am particularly
worried about this because I understand that the Hansen test results
of other specifications also depend on the two-step regressions.

The Stata output for the one-step and two-step difference GMM cases
are below. "reg4" and "reg2", which are not directly used in the
estimations, are dummy variables. The variables starting with "reg2"
and "reg4" are interction terms of some of my variables (which are not
dummies themselves) with these dummies. "i.qtr" are time dummies.

I would appreciate any comments very much. Thanks a lot for your consideration.

Guzin.

--
G. G. Akin
Department of Economics
Bogazici University


ONE STEP RESULT:

. xi: xtabond2 lnTR_cpi l.lnTR_cpi lnCF lnPK1 lnW_cpi lnAGE lnCQ ys
lnLC1 reg4_lnCF reg4_lnPK1 reg4_lnW_cpi reg2_lnCF reg2_lnPK1 re
> g2_lnW_cpi  i.qtr, gmm(l.lnTR_cpi, lag (1 2) collapse) iv(lnCF lnPK1 lnW_cpi lnAGE lnCQ ys lnLC1 reg4_lnCF reg4_lnPK1 reg4_lnW_cp
> i  reg2_lnCF reg2_lnPK1 reg2_lnW_cpi i.qtr) robust noleveleq orthogonal
i.qtr             _Iqtr_171-195       (naturally coded; _Iqtr_171 omitted)
Favoring speed over space. To switch, type or click on mata: mata set
matafavor space, perm.
_Iqtr_194 dropped due to collinearity
_Iqtr_195 dropped due to collinearity
Warning: Number of instruments may be large relative to number of observations.
Warning: Two-step estimated covariance matrix of moments is singular.
  Using a generalized inverse to calculate robust weighting matrix for
Hansen test.
  Difference-in-Sargan/Hansen statistics may be negative.

Dynamic panel-data estimation, one-step difference GMM
------------------------------------------------------------------------------
Group variable: code                            Number of obs      =       284
Time variable : qtr                             Number of groups   =        17
Number of instruments = 37                      Obs per group: min =         0
Wald chi2(36) =     58.82                                      avg =     16.71
Prob > chi2   =     0.010                                      max =        23
------------------------------------------------------------------------------
             |               Robust
    lnTR_cpi |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    lnTR_cpi |
         L1. |   .9822867     .07386    13.30   0.000     .8375236     1.12705
        lnCF |  -.0564652   .0801848    -0.70   0.481    -.2136246    .1006942
       lnPK1 |  -.0413206   .0425474    -0.97   0.331    -.1247121    .0420709
     lnW_cpi |  -.3810669   .1706384    -2.23   0.026    -.7155121   -.0466217
       lnAGE |   -.434488   .5253184    -0.83   0.408    -1.464093    .5951171
        lnCQ |  -.0709354   .0404604    -1.75   0.080    -.1502363    .0083656
          ys |    .630057    2.63359     0.24   0.811    -4.531685    5.791799
       lnLC1 |  -.0079554   .0115534    -0.69   0.491    -.0305996    .0146888
   reg4_lnCF |  -.0320389   .0724118    -0.44   0.658    -.1739633    .1098856
  reg4_lnPK1 |   .0405702   .0365559     1.11   0.267     -.031078    .1122183
reg4_lnW_cpi |  -.0728926   .1708264    -0.43   0.670    -.4077062    .2619209
   reg2_lnCF |   .1198198   .1259726     0.95   0.342    -.1270819    .3667215
  reg2_lnPK1 |   .0189572   .0455336     0.42   0.677     -.070287    .1082015
reg2_lnW_cpi |   .4085474      .1934     2.11   0.035     .0294904    .7876044
   _Iqtr_172 |  -1.777399   7.929361    -0.22   0.823    -17.31866    13.76386
   _Iqtr_173 |  -1.981734    8.53677    -0.23   0.816     -18.7135    14.75003
   _Iqtr_174 |  -1.975342   8.129462    -0.24   0.808    -17.90879    13.95811
   _Iqtr_175 |  -.5954979   2.279078    -0.26   0.794    -5.062409    3.871413
   _Iqtr_176 |   .1194124   .8342819     0.14   0.886     -1.51575    1.754575
   _Iqtr_177 |  -.0788134   1.008014    -0.08   0.938    -2.054485    1.896858
   _Iqtr_178 |  -.3560044   1.995979    -0.18   0.858    -4.268051    3.556043
   _Iqtr_179 |  -.1091355   1.044778    -0.10   0.917    -2.156864    1.938593
   _Iqtr_180 |   .1215488   .6757695     0.18   0.857    -1.202935    1.446033
   _Iqtr_181 |   .3284245   1.329039     0.25   0.805    -2.276445    2.933294
   _Iqtr_182 |   .4557443   1.753742     0.26   0.795    -2.981526    3.893015
   _Iqtr_183 |   .2580215   .9571964     0.27   0.787    -1.618049    2.134092
   _Iqtr_184 |   .4823346   1.195847     0.40   0.687    -1.861483    2.826152
   _Iqtr_185 |   .4044741   1.082867     0.37   0.709    -1.717906    2.526854
   _Iqtr_186 |  -.0925303   .2665519    -0.35   0.728    -.6149624    .4299017
   _Iqtr_187 |  -.1039607   .4834029    -0.22   0.830    -1.051413    .8434916
   _Iqtr_188 |   .1631972   .2023393     0.81   0.420    -.2333805    .5597748
   _Iqtr_189 |  -.0748803    .511241    -0.15   0.884    -1.076894    .9271337
   _Iqtr_190 |   .4133159   .9470381     0.44   0.663    -1.442845    2.269476
   _Iqtr_191 |   .3120399    1.51166     0.21   0.836    -2.650759    3.274838
   _Iqtr_192 |   .1291133    .946743     0.14   0.892    -1.726469    1.984696
   _Iqtr_193 |   .1759939   .7766529     0.23   0.821    -1.346218    1.698206
------------------------------------------------------------------------------
Instruments for orthogonal deviations equation
  Standard
    FOD.(lnCF lnPK1 lnW_cpi lnAGE lnCQ ys lnLC1 reg4_lnCF reg4_lnPK1
    reg4_lnW_cpi reg2_lnCF reg2_lnPK1 reg2_lnW_cpi _Iqtr_172 _Iqtr_173
    _Iqtr_174 _Iqtr_175 _Iqtr_176 _Iqtr_177 _Iqtr_178 _Iqtr_179 _Iqtr_180
    _Iqtr_181 _Iqtr_182 _Iqtr_183 _Iqtr_184 _Iqtr_185 _Iqtr_186 _Iqtr_187
    _Iqtr_188 _Iqtr_189 _Iqtr_190 _Iqtr_191 _Iqtr_192 _Iqtr_193 _Iqtr_194
    _Iqtr_195)
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/2).L.lnTR_cpi collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =  -2.68  Pr > z =  0.007
Arellano-Bond test for AR(2) in first differences: z =   1.65  Pr > z =  0.099
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(1)    =   0.14  Prob > chi2 =  0.712
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(1)    =   0.00  Prob > chi2 =  1.000
  (Robust, but can be weakened by many instruments.)


TWO-STEP RESULT:

. xi: xtabond2 lnTR_cpi l.lnTR_cpi lnCF lnPK1 lnW_cpi lnAGE lnCQ ys
lnLC1 reg4_lnCF reg4_lnPK1 reg4_lnW_cpi reg2_lnCF reg2_lnPK1 re
> g2_lnW_cpi  i.qtr, gmm(l.lnTR_cpi, lag (1 2) collapse) iv(lnCF lnPK1 lnW_cpi lnAGE lnCQ ys lnLC1 reg4_lnCF reg4_lnPK1 reg4_lnW_cp
> i  reg2_lnCF reg2_lnPK1 reg2_lnW_cpi i.qtr) robust noleveleq orthogonal twostep
i.qtr             _Iqtr_171-195       (naturally coded; _Iqtr_171 omitted)
Favoring speed over space. To switch, type or click on mata: mata set
matafavor space, perm.
_Iqtr_194 dropped due to collinearity
_Iqtr_195 dropped due to collinearity
Warning: Number of instruments may be large relative to number of observations.
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 difference GMM
------------------------------------------------------------------------------
Group variable: code                            Number of obs      =       284
Time variable : qtr                             Number of groups   =        17
Number of instruments = 37                      Obs per group: min =         0
Wald chi2(36) =  77562.07                                      avg =     16.71
Prob > chi2   =     0.000                                      max =        23
------------------------------------------------------------------------------
             |              Corrected
    lnTR_cpi |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    lnTR_cpi |
         L1. |  (dropped)
        lnCF |   .2990594    .051845     5.77   0.000     .1974452    .4006737
       lnPK1 |  -.1315095   .0310195    -4.24   0.000    -.1923067   -.0707124
     lnW_cpi |  (dropped)
       lnAGE |  (dropped)
        lnCQ |   -.007913   .0369965    -0.21   0.831    -.0804249    .0645989
          ys |  -.1966009     .02193    -8.96   0.000    -.2395829   -.1536189
       lnLC1 |  -.0749398   .0102445    -7.32   0.000    -.0950185    -.054861
   reg4_lnCF |   -.024488   .0073436    -3.33   0.001    -.0388811   -.0100949
  reg4_lnPK1 |  (dropped)
reg4_lnW_cpi |  (dropped)
   reg2_lnCF |   -.279656   .0223123   -12.53   0.000    -.3233873   -.2359247
  reg2_lnPK1 |   .2014782   .0220579     9.13   0.000     .1582455    .2447109
reg2_lnW_cpi |  (dropped)
   _Iqtr_172 |   .0827462   .1281865     0.65   0.519    -.1684947    .3339871
   _Iqtr_173 |  (dropped)
   _Iqtr_174 |   .1755964   .1089994     1.61   0.107    -.0380386    .3892314
   _Iqtr_175 |  (dropped)
   _Iqtr_176 |  -.3964868   .0684031    -5.80   0.000    -.5305544   -.2624192
   _Iqtr_177 |  (dropped)
   _Iqtr_178 |   .0708974   .0462206     1.53   0.125    -.0196933    .1614882
   _Iqtr_179 |  (dropped)
   _Iqtr_180 |  (dropped)
   _Iqtr_181 |  (dropped)
   _Iqtr_182 |  (dropped)
   _Iqtr_183 |  (dropped)
   _Iqtr_184 |  (dropped)
   _Iqtr_185 |  (dropped)
   _Iqtr_186 |  -.4241229   .0529373    -8.01   0.000     -.527878   -.3203677
   _Iqtr_187 |  -.1603154   .0285739    -5.61   0.000    -.2163192   -.1043115
   _Iqtr_188 |  -.0144077   .0755751    -0.19   0.849    -.1625321    .1337168
   _Iqtr_189 |  -.1993144   .0948881    -2.10   0.036    -.3852917   -.0133371
   _Iqtr_190 |  (dropped)
   _Iqtr_191 |  (dropped)
   _Iqtr_192 |  (dropped)
   _Iqtr_193 |   -.064516   .0305566    -2.11   0.035    -.1244058   -.0046262
------------------------------------------------------------------------------
Instruments for orthogonal deviations equation
  Standard
    FOD.(lnCF lnPK1 lnW_cpi lnAGE lnCQ ys lnLC1 reg4_lnCF reg4_lnPK1
    reg4_lnW_cpi reg2_lnCF reg2_lnPK1 reg2_lnW_cpi _Iqtr_172 _Iqtr_173
    _Iqtr_174 _Iqtr_175 _Iqtr_176 _Iqtr_177 _Iqtr_178 _Iqtr_179 _Iqtr_180
    _Iqtr_181 _Iqtr_182 _Iqtr_183 _Iqtr_184 _Iqtr_185 _Iqtr_186 _Iqtr_187
    _Iqtr_188 _Iqtr_189 _Iqtr_190 _Iqtr_191 _Iqtr_192 _Iqtr_193 _Iqtr_194
    _Iqtr_195)
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/2).L.lnTR_cpi collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =  -2.73  Pr > z =  0.006
Arellano-Bond test for AR(2) in first differences: z =   0.80  Pr > z =  0.423
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(1)    =   0.14  Prob > chi2 =  0.712
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(1)    =   0.00  Prob > chi2 =  1.000
  (Robust, but can be weakened by many instruments.)
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index