translog production function estimation by sfpanel

[M. Yun]

I have estimated a translog production function using the command sfpanel, model(bc92) and got the output in the attached file. Please help me to interpret the statistics reported, and how I can calculate changes in TFP, technological progress, technological and allocative ineffciencies.

[M. Yun]

Regarding the same estimation above, I got technical efficiency value by the command "predict jlms." The manual explains that this gives estimates of technical efficiency, while "predict u" gives the estimates of technical inefficiency, which is the same explanation as given in Belotti et al 2013. I am confused since Kumbhakar et al (2015: pp 69) mentions that jmls gives the index of technical inefficiency. I am further confused because for my estimation STATA reports exactly the same figures for u and jlms as below. And these figures seem quite extraordinary, as they seem to suggest that either technical inefficiency or efficiency is over 100%.

Variable Obs Mean Std. dev. Min Max
u 22,910 19.23243 1.490071 12.75585 26.66319
jlms 22,910 19.23243 1.490071 12.75585 26.66319