[Limdep Nlogit List] panel data stochastic frontier models

[hakan güneş]

Dear all,

I estimate cost efficiency by using Battese Coelli 1995 and true fixed
effects model by Greene (2005). Results of cost efficiency of the Battese
Coelli 95 is more reasonable than true fixed effects model. however, mean
efficiency is lower relative to literature.

Is it possible to estimate time varying inefficiency for panel data which
has truncated normal distribution defined as Stevenson 1980 (absolute value
of the normal distribution which has mean higher than zero)?

Is this following command do that? this is the true fixed effects command
for small panel but without the stratification specification:


frontier; cost; lhs=lc; rhs= one, ly1, ly2, lw1, lq, k;
rh2=one, ds, x18, dl, cap, llp, lpb, liq ;model=T;costeff=ufe503;
;tlg=0.00001;tlf=0.00001;tlb=0.00001, maxit=1000$


I think this specification does not know that i have a panel data. Can i
estimate time varying inefficiency for panel data defined as Stevenson by
using Limdep?

Best regards,

Hakan gunes.
PHD student at METU.