Why do i have large z test statistics when i ran translog model

[Arunika Kodagoda]

First i ran frontier model and then i did translog method here i have obtained large wald chisquare and large z test statistics my lnv2sig2v is not significant is it a problem what should i do?

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Also, many of us will not open attached files.

We don't even know exactly what estimator and models you ran and how you did the "translog method".

[Arunika Kodagoda]

Dear sir,

Im estimating the production function. first i ran simple regression for OLS then i ran stochastic frontier for my cross sectional data. Finally i did the translog method. In my translog method i assumed the error is half normally distributed but with my data i cant run with the constant because im getting iteration error. therefore i ran with exponential distribution but here i cant estimate the technical inefficiency term im getting iteration error. May i know what i should do. for my cobb douglas production function i didn't get any error. Below is what i obtained for cobb douglas production function technical efficiency model.

Stoc. frontier normal/half-normal model         Number of obs     =         82                                                 Wald chi2(3)      =   30330.19 Log likelihood = -3.5677263                     Prob > chi2       =     0.0000  ------------------------------------------------------------------------------          lny |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval] -------------+----------------------------------------------------------------        lnlab |   1.273875   .0936794    13.60   0.000     1.090267    1.457483    lnorganic |   .2661321   .0434148     6.13   0.000     .1810405    .3512236       lnland |   .0746134   .0456823     1.63   0.102    -.0149223    .1641491 -------------+----------------------------------------------------------------     /lnsig2v |  -2.750859   .1561738   -17.61   0.000    -3.056954   -2.444765     /lnsig2u |  -41.74448   18349.16    -0.00   0.998    -36005.44    35921.96 -------------+----------------------------------------------------------------      sigma_v |    .252731    .019735                      .2168657    .2945277      sigma_u |   8.62e-10   7.90e-06                             0           .       sigma2 |   .0638729   .0099753                      .0443218    .0834241       lambda |   3.41e-09    .019735                     -.0386798    .0386798 ------------------------------------------------------------------------------ LR test of sigma_u=0: chibar2(01) = 0.00               Prob >= chibar2 = 1.000

Stoc. frontier normal/half-normal model         Number of obs     =         82
                                                Wald chi2(3)      =     923.18
Log likelihood =  43.823512                     Prob > chi2       =     0.0000

--------------------------------------------------------------------------------------
                 lny |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
---------------------+----------------------------------------------------------------
lny                  |
               lnlab |   .8047512   .0823812     9.77   0.000     .6432869    .9662154
           lnorganic |   .2251354   .0240381     9.37   0.000     .1780216    .2722493
              lnland |   .2027933   .0287243     7.06   0.000     .1464946    .2590919
               _cons |   1.725193   .1749682     9.86   0.000     1.382262    2.068124
---------------------+----------------------------------------------------------------
lnsig2v              |
               _cons |   -5.07972   .3798877   -13.37   0.000    -5.824287   -4.335154
---------------------+----------------------------------------------------------------
lnsig2u              |
                 age |   .0531842   .0312639     1.70   0.089    -.0080919    .1144602
          experience |  -.0393181   .0240587    -1.63   0.102    -.0864723    .0078362
              hhsize |  -.2061523   .2096913    -0.98   0.326    -.6171396    .2048351
           livestock |    .944724   .4866764     1.94   0.052    -.0091442    1.898592
          occupation |    .664641   .5426226     1.22   0.221    -.3988797    1.728162
              gender |  -.9024016    .520934    -1.73   0.083    -1.923413    .1186101
                  vp |   .5833345   .7466759     0.78   0.435    -.8801234    2.046792
                  sd |  -1.025112   1.305347    -0.79   0.432    -3.583546    1.533322
durationoforganictea |  -.1521679   .0726073    -2.10   0.036    -.2944756   -.0098602
          othercrops |    .948397   .5279737     1.80   0.072    -.0864123    1.983206
   education_dummy_1 |  -10.93334   10.95522    -1.00   0.318    -32.40518     10.5385
   education_dummy_2 |  -4.809491   2.663507    -1.81   0.071    -10.02987     .410886
   education_dummy_3 |  -4.606307   2.669673    -1.73   0.084    -9.838769    .6261563
   education_dummy_4 |  -2.552731   2.570144    -0.99   0.321     -7.59012    2.484658
   education_dummy_5 |  -3.715075   2.630043    -1.41   0.158    -8.869864    1.439713
   education_dummy_6 |  -5.276297   2.328762    -2.27   0.023    -9.840587   -.7120066
   education_dummy_7 |  -7.075948    4.36056    -1.62   0.105    -15.62249    1.470592
   education_dummy_8 |  -.0882475   2.679457    -0.03   0.974    -5.339886    5.163391
            training |   .0993385   .8684734     0.11   0.909    -1.602838    1.801515
---------------------+----------------------------------------------------------------
             sigma_v |   .0788774   .0149823                      .0543591    .1144546
--------------------------------------------------------------------------------------

For translpg production function

Stoc. frontier normal/half-normal model         Number of obs     =         82
                                                Wald chi2(9)      =   1.55e+12
Log likelihood =  25.822511                     Prob > chi2       =     0.0000

------------------------------------------------------------------------------
         lny |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
        lab2 |  -.0551643   .0000434 -1271.57   0.000    -.0552493   -.0550793
       land2 |   .0569984   7.00e-06  8137.36   0.000     .0569847    .0570121
        org2 |   .0687043   4.92e-06  1.4e+04   0.000     .0686946    .0687139
     labland |  -.0581108   .0000105 -5552.95   0.000    -.0581313   -.0580903
      laborg |  -.2283966   .0000281 -8129.83   0.000    -.2284517   -.2283416
     landorg |  -.1388188   4.65e-06 -3.0e+04   0.000    -.1388279   -.1388097
       lnlab |   2.431192    .000091  2.7e+04   0.000     2.431014     2.43137
   lnorganic |   .0649848    .000047  1382.67   0.000     .0648926    .0650769
      lnland |   1.095407   .0000267  4.1e+04   0.000     1.095355    1.095459
-------------+----------------------------------------------------------------
    /lnsig2v |  -38.37038   393.1166    -0.10   0.922    -808.8648    732.1241
    /lnsig2u |    -2.0814   .1561738   -13.33   0.000    -2.387495   -1.775305
-------------+----------------------------------------------------------------
     sigma_v |   4.66e-09   9.15e-07                      2.3e-176    9.5e+158
     sigma_u |   .3532073   .0275809                      .3030833    .4116209
      sigma2 |   .1247554   .0194835                      .0865684    .1629424
      lambda |   7.59e+07   .0275809                      7.59e+07    7.59e+07
------------------------------------------------------------------------------
LR test of sigma_u=0: chibar2(01) = 14.20              Prob >= chibar2 = 0.000

But here i didnt get the output without supressing the constant term therefore i ran with the exponential distribution.

Stoc. frontier normal/exponential model         Number of obs     =         82
                                                Wald chi2(9)      =     759.43
Log likelihood =  19.813519                     Prob > chi2       =     0.0000

------------------------------------------------------------------------------
         lny |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       lnlab |  -.4966434   2.073942    -0.24   0.811    -4.561495    3.568208
   lnorganic |  -.4182501   .4761111    -0.88   0.380    -1.351411    .5149106
      lnland |   1.059041   .4903156     2.16   0.031     .0980403    2.020042
        lab2 |   .1430166   .5366499     0.27   0.790    -.9087978    1.194831
       land2 |   .0403039   .0615121     0.66   0.512    -.0802577    .1608654
        org2 |    .044659   .0381807     1.17   0.242    -.0301739    .1194918
     labland |  -.2700881   .1711921    -1.58   0.115    -.6056186    .0654423
      laborg |   .0758639   .2220813     0.34   0.733    -.3594076    .5111353
     landorg |  -.0324216   .0648854    -0.50   0.617    -.1595946    .0947513
       _cons |   5.197133   2.440417     2.13   0.033      .414003    9.980263
-------------+----------------------------------------------------------------
    /lnsig2v |  -5.015822    .685764    -7.31   0.000    -6.359895   -3.671749
    /lnsig2u |  -3.130248   .3746372    -8.36   0.000    -3.864524   -2.395973
-------------+----------------------------------------------------------------
     sigma_v |   .0814382   .0279237                      .0415878     .159474
     sigma_u |    .209062   .0391612                      .1448203    .3018013
      sigma2 |   .0503391    .013768                      .0233543    .0773239
      lambda |   2.567125   .0614051                      2.446774    2.687477
------------------------------------------------------------------------------
LR test of sigma_u=0: chibar2(01) = 2.18               Prob >= chibar2 = 0.070

.
But with the exponential i cannot estimate the technical inefficiency model.

May i know where i have gone wrong and what i should do?

Thanks in advancce