Pre requisites to apply a stochastic frontier model

[Thierry Coulibaly]

Hi everyone. I'm a little confused about the pre requisites to have before applying a stochastic frontier model. In the book A practitioner's Guide to Stochastic frontier Analysis using Stata, they present some codes relative to hnormal, tnormal and exponential distribution. And the only 2 pre requisites they present there is i) the test to know if the residuals of the ols regression are left skewed ii) test of the presence of inefficiencies. Now, I want to know if there is no other pre requisites before applying these models. I also want to know if we must just assumed that the distributions are hnormal tnormal or exponential or if there are some tests to do. To finish I want to know how to specify that my production function is a cobb douglas or a translog function. I'm using Stata 14. Thanks for the reply