[Limdep Nlogit List] random parameter estimation and fixed effects

[Jaap Bos]

Dear all,
 
I am estimating a random parameter frontier. I want to compare estimates - in particular marginal effects - with a fixed effect frontier estimation. 
 
I'm struggling with comparing the results for both specifications. What I try to do therefore is estimate a random parameter frontier with firm dummies. However, that specification (almost) never converges.
 
I know that the random parameter model and the fixed effects model are not nested. Hence: anyone who has ideas on how to choose a preferred specification? 
 
regards, Jaap Bos