xtabond2 Hansen test

[Giorgio Carlini]

Hello I'm using system GMM to estimate TFP growth (gtfp_100) at regional level .
I've a problem with the Hansen test, which is clearly too high (1.000), I guess.
I suspect this is due to instrument proliferation. So I tried to limit the instruments using collapse. Also I get the same results when using only (1 2) lags for the equation in difference. Any suggestions, please.

xtabond2 gtfp_100  l(1).ltfp_rel_lev95 ///
l(1).(   S dettotoccrfl  netai psii  netae     psie   ///
   branchps  rspri_stock_va  totroadskms   cs_14pesos) ///
tau1995-tau2012   ///
if cod_reg<21   , ///
iv(l(1).(   S dettotoccrfl  netai psii  netae     psie  ///
   branchps  rspri_stock_va  totroadskms   cs_14pesos) ///
tau1995-tau2012  , eq(diff)) ///
iv(l(1).(   S dettotoccrfl  netai psii  netae     psie   ///
   branchps  rspri_stock_va  totroadskms   cs_14pesos) ///
tau1995-tau2012  , eq(lev)) ///
gmm(l(1).ltfp_rel_lev95 , laglimits(1 .) collapse  eq(diff) )   ///
gmm(l(1).ltfp_rel_lev95 , laglimits(0 1) collapse  eq(lev) )   ///
h(3) rob   ar(3)

Dynamic panel-data estimation, one-step system GMM
------------------------------------------------------------------------------
Group variable: cod_reg                         Number of obs      =       320
Time variable : anno                            Number of groups   =        20
Number of instruments = 54                      Obs per group: min =        16
Wald chi2(26) =   1839.21                                      avg =     16.00
Prob > chi2   =     0.000                                      max =        16
--------------------------------------------------------------------------------
               |               Robust
      gtfp_100 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
---------------+----------------------------------------------------------------
ltfp_rel_lev95 |
           L1. |  -.2348444   .0606115    -3.87   0.000    -.3536407   -.1160481
               |
             S |
           L1. |    .019605   .0120456     1.63   0.104    -.0040039    .0432139
               |
  dettotoccrfl |
           L1. |  -.2225202   .1280743    -1.74   0.082    -.4735413    .0285009
               |
         netai |
           L1. |  -.0201765   .0176353    -1.14   0.253    -.0547411    .0143881
               |
          psii |
           L1. |   .0576091   .0654523     0.88   0.379    -.0706749    .1858932
               |
         netae |
           L1. |  -.0480485   .0209141    -2.30   0.022    -.0890394   -.0070575
               |
          psie |
           L1. |   -.170621   .1062258    -1.61   0.108    -.3788197    .0375778
               |
      branchps |
           L1. |   .0445189   .0261953     1.70   0.089     -.006823    .0958607
               |
rspri_stock_va |
           L1. |   .0015889   .0008631     1.84   0.066    -.0001028    .0032806
               |
   totroadskms |
           L1. |   .0095032   .0181984     0.52   0.602    -.0261651    .0451714
               |
    cs_14pesos |
           L1. |   -.140023   .0930925    -1.50   0.133     -.322481    .0424351
               |
       tau1997 |    .081572   .0248708     3.28   0.001      .032826    .1303179
       tau1998 |   .0765403   .0236921     3.23   0.001     .0301045     .122976
       tau1999 |   .0820527    .023021     3.56   0.000     .0369324     .127173
       tau2000 |   .0906023   .0229646     3.95   0.000     .0455925     .135612
       tau2001 |   .0773837   .0238363     3.25   0.001     .0306653     .124102
       tau2002 |   .0514515   .0216221     2.38   0.017     .0090731      .09383
       tau2003 |    .047534   .0194083     2.45   0.014     .0094945    .0855735
       tau2004 |   .0584287   .0162254     3.60   0.000     .0266274      .09023
       tau2005 |   .0508587   .0146411     3.47   0.001     .0221627    .0795548
       tau2006 |   .0504148   .0134201     3.76   0.000     .0241118    .0767178
       tau2007 |   .0464896   .0121259     3.83   0.000     .0227233    .0702559
       tau2008 |   .0233039    .010413     2.24   0.025     .0028949     .043713
       tau2009 |  -.0103899   .0071914    -1.44   0.149    -.0244849    .0037051
       tau2010 |   .0320077   .0055658     5.75   0.000      .021099    .0429164
       tau2011 |   .0195995   .0040825     4.80   0.000      .011598     .027601
         _cons |  -.1010425   .1680118    -0.60   0.548    -.4303397    .2282547
--------------------------------------------------------------------------------
Instruments for first differences equation
  Standard
    D.(L.S L.dettotoccrfl L.netai L.psii L.netae L.psie L.branchps
    L.rspri_stock_va L.totroadskms L.cs_14pesos tau1995 tau1996 tau1997
    tau1998 tau1999 tau2000 tau2001 tau2002 tau2003 tau2004 tau2005 tau2006
    tau2007 tau2008 tau2009 tau2010 tau2011 tau2012)
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/17).L.ltfp_rel_lev95 collapsed
Instruments for levels equation
  Standard
    L.S L.dettotoccrfl L.netai L.psii L.netae L.psie L.branchps
    L.rspri_stock_va L.totroadskms L.cs_14pesos tau1995 tau1996 tau1997
    tau1998 tau1999 tau2000 tau2001 tau2002 tau2003 tau2004 tau2005 tau2006
    tau2007 tau2008 tau2009 tau2010 tau2011 tau2012
    _cons
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    DL(0/1).L.ltfp_rel_lev95 collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =  -3.21  Pr > z =  0.001
Arellano-Bond test for AR(2) in first differences: z =  -0.09  Pr > z =  0.927
Arellano-Bond test for AR(3) in first differences: z =   0.27  Pr > z =  0.785
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(27)   =  95.16  Prob > chi2 =  0.000
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(27)   =   0.00  Prob > chi2 =  1.000
  (Robust, but weakened by many instruments.)

Difference-in-Hansen tests of exogeneity of instrument subsets:
  GMM instruments for levels
    Hansen test excluding group:     chi2(25)   =   0.00  Prob > chi2 =  1.000
    Difference (null H = exogenous): chi2(2)    =   0.00  Prob > chi2 =  1.000
  gmm(L.ltfp_rel_lev95, collapse eq(diff) lag(1 .))
    Hansen test excluding group:     chi2(11)   =   0.00  Prob > chi2 =  1.000
    Difference (null H = exogenous): chi2(16)   =   0.00  Prob > chi2 =  1.000
  gmm(L.ltfp_rel_lev95, collapse eq(level) lag(0 1))
    Hansen test excluding group:     chi2(25)   =   0.00  Prob > chi2 =  1.000
    Difference (null H = exogenous): chi2(2)    =   0.00  Prob > chi2 =  1.000
  iv(L.S L.dettotoccrfl L.netai L.psii L.netae L.psie L.branchps L.rspri_stock_va L.totroadskms L.cs_14pesos tau1995 tau1996 tau1997 tau1998 tau19
> 99 tau2000 tau2001 tau2002 tau2003 tau2004 tau2005 tau2006 tau2007 tau2008 tau2009 tau2010 tau2011 tau2012, eq(diff))
    Hansen test excluding group:     chi2(6)    =   0.00  Prob > chi2 =  1.000
    Difference (null H = exogenous): chi2(21)   =  -0.00  Prob > chi2 =  1.000
  iv(L.S L.dettotoccrfl L.netai L.psii L.netae L.psie L.branchps L.rspri_stock_va L.totroadskms L.cs_14pesos tau1995 tau1996 tau1997 tau1998 tau19
> 99 tau2000 tau2001 tau2002 tau2003 tau2004 tau2005 tau2006 tau2007 tau2008 tau2009 tau2010 tau2011 tau2012, eq(level))
    Hansen test excluding group:     chi2(13)   =   0.00  Prob > chi2 =  1.000
    Difference (null H = exogenous): chi2(14)   =   0.00  Prob > chi2 =  1.000

[Connie Gao]

May create a sample that cod_reg<21 first.
Make sure number of instruments smaller than number of groups. Your current number of IVs is 54, which is larger than number of groups (n=20).

[CODE]
xtabond2 gtfp_100 l(1).ltfp_rel_lev95 l(1).(S dettotoccrfl netai psii netae psie branchps rspri_stock_va totroadskms cs_14pesos) tau1995-tau2012, gmm(ltfp_rel_lev95 , laglimits(2 4) collapse) iv(l(1).(S dettotoccrfl netai psii netae psie branchps rspri_stock_va totroadskms cs_14pesos) tau1995-tau2012 , eq(level)) twostep robust small h(3) rob ar(3)/CODE]

[Giorgio Carlini]

Thanks a lot Connie,
I've tried to follow your suggestions. However, creating the sample that cod_reg<21 and using your code does not change the results. Hansen test still equal 1.000
By the way, I've used the one-step system GMM because the two-step produce omitted variables. Also, some suggestion on this would be useful.
Regarding the instrument proliferation, with the dataset used and the specification model it is almost impossible to reduce the number of instrument below N, i.e., 20.
Any help, please.