Dear sir,

Recently i've been trying using SFA to determine the efficiency of Gov't Spending. Im using Greene true random effect (TRE) with the guidance from Belloti "Stochastic frontier analysis using Stata". Sfpanel

There are several question that i would ask
1. What is frontier variable? can i treat it as control variable?
2. Is my syntax is true? because i have no idea if this is true or not
sfpanel lny lnx1 lnx2 (Treated as control variable), model(tre) distribution(hnormal) usigma(lnu1 lnu2 (Govt Spending)) difficult rescale nsim(100) simtype(genhalton)