A gmm command that works in Stata 13 returns "could not evaluate equation 1" in Stata 14.1

[Anat Tchetchik]

Dear all,
I have just upgraded to Stata/SE 14.1. I work on the attached data file. The following gmm command which works in my colleague's Stata 13 version returns an error

could not evaluate equation 1
r(498);

The command is:

#delimit;
gmm (ln(marketshare)- {B0}-{xb: agri_activity extreme_activity kidsfac agri_sale food_sales guided_act seating
 access uppergalilee}+{alpha}*(ln(total_price)-{sigma}*(ln(region_ms)))),instruments(accommod extreme_activity
agri_activity kidsfac established_attr food_sales agri_sale access guided_act atract_3region 
activity_size  agri_work fam_work agri_fac uppergalilee)twostep winit(unadj,indep);

Any ideas what is the problem?
Best,
Anat

[Richard Williams]

Are you sure this exact same code and data ran in Stata 13? I get the same error you got in Stata 13 and in Stata 14. That leads me to suspect something has been lost between your colleague's machine and your machine. Or that s/he is incorrect in saying s/he got it to run. Double check with your colleague, e.g. send your colleague your code and see if it runs.

[Anat Tchetchik]

Thank you, I will double check!

[Richard Gates (StataCorp)]

Anat,

The variable marketshare does not exist. It is ln_marketshare. This applies to the other logged variables. Unfortunately, gmm captures the error and issues its own vague message leaving you a bit clueless what the problem is. I also found that initial estimates are necessary.

I used regress to get crude initial estimates.

. regress ln_marketshare agri_activity extreme_activity kidsfac ///
>         agri_sale food_sales guided_act seating access uppergalilee ///
>         ln_total_price ln_region_ms

      Source |       SS           df       MS      Number of obs   =        71
-------------+----------------------------------   F(11, 59)       =     84.56
       Model |  232.638538        11   21.148958   Prob > F        =    0.0000
    Residual |  14.7565583        59  .250111158   R-squared       =    0.9404
-------------+----------------------------------   Adj R-squared   =    0.9292
       Total |  247.395096        70  3.53421566   Root MSE        =    .50011

------------------------------------------------------------------------------
ln_markets~e |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
agri_activ~y |  -.4145576   .1466714    -2.83   0.006    -.7080464   -.1210689
extreme_ac~y |   .2966711   .2841567     1.04   0.301    -.2719252    .8652674
     kidsfac |   .0578384   .1386673     0.42   0.678    -.2196342     .335311
   agri_sale |   .4625761   .1769551     2.61   0.011     .1084898    .8166625
  food_sales |   .2258728   .1515022     1.49   0.141    -.0772825     .529028
  guided_act |  -.1803722   .1562924    -1.15   0.253    -.4931125    .1323681
     seating |   .0015671   .0007517     2.08   0.041      .000063    .0030711
      access |   .1355426   .1178812     1.15   0.255    -.1003372    .3714224
uppergalilee |   1.389592   .1669886     8.32   0.000     1.055448    1.723735
ln_total_p~e |    .096056   .1117153     0.86   0.393    -.1274858    .3195978
ln_region_ms |   .8064377   .0503264    16.02   0.000     .7057348    .9071407
       _cons |  -7.010594   .7674078    -9.14   0.000    -8.546173   -5.475014
------------------------------------------------------------------------------

.
. mat b = e(b)

. mat li b

b[1,12]
    agri_activ~y  extreme_ac~y       kidsfac     agri_sale    food_sales
y1    -.41455764      .2966711     .05783837     .46257611     .22587279

      guided_act       seating        access  uppergalilee  ln_total_p~e
y1    -.18037218     .00156709     .13554258     1.3895916     .09605602

    ln_region_ms         _cons
y1     .80643774    -7.0105935

. local k = colsof(b)

. mat b0 = b[1,`k']

. mat b1 = b[1,`--k']

. mat b2 = -b[1,`--k']

. mat b= b[1,1..`--k']

. local stripe : colnames b

. mat coleq b = xb

. mat colnames b0 = B0:_cons

. mat colnames b1 = alpha:_cons

. mat colnames b2 = sigma:_cons

. mat b = (b0,b,b2,b1)

. mat li b

b[1,12]
              B0:           xb:           xb:           xb:           xb:
           _cons  agri_activ~y  extreme_ac~y       kidsfac     agri_sale
r1    -7.0105935    -.41455764      .2966711     .05783837     .46257611

              xb:           xb:           xb:           xb:           xb:
      food_sales    guided_act       seating        access  uppergalilee
r1     .22587279    -.18037218     .00156709     .13554258     1.3895916

           sigma:        alpha:
           _cons         _cons
r1    -.09605602     .80643774

.
.
. gmm (ln_marketshare - {B0}-{xb: agri_activity extreme_activity kidsfac ///
>         agri_sale food_sales guided_act seating access uppergalilee} + ///
>         {alpha}*(ln_total_price-{sigma}*(ln_region_ms))),            ///
>         instruments(accommod extreme_activity agri_activity kidsfac    ///
>         established_attr food_sales agri_sale access guided_act        ///
>         atract_3region activity_size  agri_work fam_work agri_fac      ///
>         uppergalilee) twostep winit(unadj,indep) quickderivatives from(b)
(71 real changes made)

Step 1
Iteration 0:   GMM criterion Q(b) =   18.27112  
Iteration 1:   GMM criterion Q(b) =  .18611487  
Iteration 2:   GMM criterion Q(b) =  .18611415  
Iteration 3:   GMM criterion Q(b) =  .18474342  
Iteration 4:   GMM criterion Q(b) =  .18231836  
Iteration 5:   GMM criterion Q(b) =  .17909891  
Iteration 6:   GMM criterion Q(b) =  .17529522  
Iteration 7:   GMM criterion Q(b) =  .17452223  
Iteration 8:   GMM criterion Q(b) =  .17014215  
Iteration 9:   GMM criterion Q(b) =  .16354259  
Iteration 10:  GMM criterion Q(b) =   .1612697  
Iteration 11:  GMM criterion Q(b) =  .14942856  
Iteration 12:  GMM criterion Q(b) =  .13906626  
Iteration 13:  GMM criterion Q(b) =  .11323153  
Iteration 14:  GMM criterion Q(b) =  .04754047  
Iteration 15:  GMM criterion Q(b) =  .04705605  
Iteration 16:  GMM criterion Q(b) =  .04694532  
Iteration 17:  GMM criterion Q(b) =  .04694532  

Step 2
Iteration 0:   GMM criterion Q(b) =  .09280402  
Iteration 1:   GMM criterion Q(b) =  .07194506  
Iteration 2:   GMM criterion Q(b) =   .0717963  
Iteration 3:   GMM criterion Q(b) =   .0717963  

GMM estimation

Number of parameters =  12
Number of moments    =  16
Initial weight matrix: Unadjusted                 Number of obs   =         71
GMM weight matrix:     Robust

------------------------------------------------------------------------------
             |               Robust
             |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
B0           |
       _cons |   -15.2408   3.267991    -4.66   0.000    -21.64595   -8.835657
-------------+----------------------------------------------------------------
xb           |
agri_activ~y |  -.0440558   .3300331    -0.13   0.894    -.6909088    .6027972
extreme_ac~y |   .0947773   1.733157     0.05   0.956    -3.302148    3.491703
     kidsfac |   .2349969   .4291657     0.55   0.584    -.6061524    1.076146
   agri_sale |    .406263   .5941873     0.68   0.494    -.7583227    1.570849
  food_sales |   .2133124   .7378262     0.29   0.772      -1.2328    1.659425
  guided_act |  -.1470461   .3913638    -0.38   0.707    -.9141051     .620013
     seating |   .0063738   .0040694     1.57   0.117     -.001602    .0143496
      access |    .579207   .3273026     1.77   0.077    -.0622943    1.220708
uppergalilee |   .5264446   .3828372     1.38   0.169    -.2239026    1.276792
-------------+----------------------------------------------------------------
      /alpha |    -1.0601   1.089973    -0.97   0.331    -3.196407    1.076208
      /sigma |  -.2599064   .2663586    -0.98   0.329    -.7819596    .2621469
------------------------------------------------------------------------------
Instruments for equation 1: accommod extreme_activity agri_activity kidsfac
    established_attr food_sales agri_sale access guided_act atract_3region
    activity_size agri_work fam_work agri_fac uppergalilee _cons

I hope this helps.

Rich

[Anat Tchetchik]

Richard, thank you so much!!! it helped! I'm so happy!

[Richard Williams]

Great catch by Richard G. I would still check with the original source of the code and data, because if it was running before then something must have gotten zapped along the way.