Hello,

I am struggling to perform a system GMM analysis. I am new to this method and have read the basic papers. When I try the xtabond2 command, however, the output does not present any values for AR(1) or AR(2).

My model generally investigates the effect of inequality on economic growth from 1990 to 2015:
ΔGDP = GDPt-1 + ginit-1 + bottom inequalityt-1 + top inequalityt-1 + redistributiont-1 + average investmentt + human capitalt-1
ΔGDP is taken over a 5 year period (all lags are a 5-year period (lagged values for 1990 are taken from 1985)).


Code:
input float(growthlny llny lgininet lbottominq ltopinq lredabs avrginv lhumancapital) byte(y1990 y1995 y2000 y2005 y2010)
   .2323642  8.362949 40.1  2.894325   3.27451   .7  17.03367  7.45 1 0 0 0 0
   .7023568  8.595314 42.1         3 4.3650794   .5 17.654509  7.88 0 1 0 0 0
   .1649471   9.29767 43.7 3.5292554  4.053504   .4 18.318789  8.34 0 0 1 0 0
-.008170462  9.462618 45.7  4.419355   3.89781  -.3 14.904442  8.55 0 0 0 1 0
   .2285496  9.454447 43.9  2.843137 3.3793104   .2  17.81613  9.26 0 0 0 0 1
  .03858224  9.682997   40       2.5 3.2689655  1.3 16.188004  9.48 0 0 0 0 0
  .09274411  7.105482 31.3 1.7181275 2.1588407  5.3 15.993094   2.4 1 0 0 0 0
  .08115608  7.198225 32.1   1.79661 2.2759433  4.9 17.934032  2.84 0 1 0 0 0
-.068234116  7.279382 33.4 1.7838453 2.7066326  4.4   22.2395  3.29 0 0 1 0 0
  .07602675  7.211147 34.5 1.8057803  2.731114  4.1 24.803436  3.69 0 0 0 1 0
   .4567258  7.287174 35.1 1.7796804  2.727389  4.2 26.195526  4.19 0 0 0 0 1
The code I use is:
Code:
xtabond2 growthlny llny lgininet lbottominq ltopinq lredabs avrginv lhumancapital y*, gmmstyle (llny lgininet lbottominq ltopinq, collapse) ivstyle(y*, equation (level)) nodiffsargan robust orthogonal small
Which gives me the following output:
Code:
xtabond2 growthlny llny lgininet lbottominq ltopinq lredabs avrginv lhumancapital y*, gmmstyle (llny lgininet lb
> ottominq ltopinq, collapse) ivstyle(y*, equation (level)) nodiffsargan robust orthogonal small
Favoring space over speed. To switch, type or click on mata: mata set matafavor speed, perm.
Warning: Two-step estimated covariance matrix of moments is singular.
  Using a generalized inverse to calculate robust weighting matrix for Hansen test.

Dynamic panel-data estimation, one-step system GMM
------------------------------------------------------------------------------
Group variable: cid                             Number of obs      =       167
Time variable : time                            Number of groups   =        28
Number of instruments = 26                      Obs per group: min =         5
F(12, 27)     =     22.26                                      avg =      5.96
Prob > F      =     0.000                                      max =         6
-------------------------------------------------------------------------------
              |               Robust
    growthlny |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
         llny |  -.9889005   .3025729    -3.27   0.003    -1.609729   -.3680721
     lgininet |  -.0025041   .0240529    -0.10   0.918    -.0518566    .0468484
   lbottominq |  -.0023967    .051481    -0.05   0.963     -.108027    .1032335
      ltopinq |  -.0146415   .0978276    -0.15   0.882    -.2153671    .1860842
      lredabs |  -.0355697   .0614711    -0.58   0.568    -.1616979    .0905586
      avrginv |   .0379343   .0144294     2.63   0.014     .0083276     .067541
lhumancapital |   .1521571   .1023877     1.49   0.149    -.0579252    .3622393
        y1990 |  -.3122944   .2985934    -1.05   0.305    -.9249575    .3003687
        y1995 |  -.2628505   .2342049    -1.12   0.272    -.7433994    .2176983
        y2000 |   -.253974   .1783147    -1.42   0.166    -.6198454    .1118975
        y2005 |  -.2029251   .1485477    -1.37   0.183    -.5077197    .1018696
        y2010 |  -.0632347   .0876716    -0.72   0.477     -.243122    .1166526
        _cons |   7.341048    2.27664     3.22   0.003     2.669769    12.01233
-------------------------------------------------------------------------------
Instruments for orthogonal deviations equation
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    L(1/25).(llny lgininet lbottominq ltopinq) collapsed
Instruments for levels equation
  Standard
    y1990 y1995 y2000 y2005 y2010
    _cons
  GMM-type (missing=0, separate instruments for each period unless collapsed)
    D.(llny lgininet lbottominq ltopinq) collapsed
------------------------------------------------------------------------------
Arellano-Bond test for AR(1) in first differences: z =      .  Pr > z =      .
Arellano-Bond test for AR(2) in first differences: z =      .  Pr > z =      .
------------------------------------------------------------------------------
Sargan test of overid. restrictions: chi2(13)   =   3.79  Prob > chi2 =  0.993
  (Not robust, but not weakened by many instruments.)
Hansen test of overid. restrictions: chi2(13)   =  11.17  Prob > chi2 =  0.597
  (Robust, but weakened by many instruments.)
I have checked other entries asking similar questions but was unable to find a solution. As far as I understand, I do have enough time periods for the AR values to be computed.
The instruments here are still very rough and need to be reconsidered but as I am just starting to work with this method, I would first like to find out where I made a mistake that leads to missing AR-values.

Thank you for your help!