I was hoping someone might be able to help.
I am trying to run a true fixed effects translog cost frontier in STATA using the sfpanel command on some panel data. I have two outputs (Y1 Y2) and two input prices (X1 X2) as well as 5 environmental variables (Z1 Z2 Z3 Z4 Z5) which I would like to test putting in the inefficiency term and a time trend "t" . The problem is that my results come out insignificant with values in the elasticities and cost efficiencies which are not correct given previous litterature results and economic theory. (I have tested both just logged variables and de-meaned all my logged variables including the Z's and the results are both times insignificant).
The general translog looks like this:
sfpanel lnTC = lnY1 lnY2 lnX1 lnX2 1/2(Y1)^2 1/2(y2)^2 1/2(x1)^2 1/2(x2)^2 lnY1lnX1 lnY1lnX2 lnY2lnX1 lnY2lnX2 lnX1lnX2 t
The steps I took for the coding are:
//declare data set as panel
xtset company year
//Constrained Estimation;
//Set constraints for Translog
//all betas on the input prices must sum to 1
constraint define 1 lnX1 + lnX2 = 1
//all betas on the cross output with input prices must sum to 0
constraint define 2 lnY1lnX1 + lnY1lnX2 + lnY2lnX1 + lnY2lnX2 = 0
*all betas on the cross input prices and inputs prices must sum to 0
constraint define 3 lnX1lnX2= 0
*sfpanel Model true fixed effects: Half normal dist, environmental variable modelling in the inefficiency term;
sfpanel lnY1 lnY2 lnX1 lnX2 1/2(Y1)^2 1/2(y2)^2 1/2(x1)^2 1/2(x2)^2 lnY1lnX1 lnY1lnX2 lnY2lnX1 lnY2lnX2 lnX1lnX2 t, cost model(tfe) distribution(hnormal) usigma( Z1 Z2 Z3 Z4 Z5) constraints (1 2 3)
//find cost efficiency scores
predict e, jlms
If anyone could offer any help I would very much appreciate this. Perhaps there is something wrong in my code or in the way I set up my variables?
I am trying to run a true fixed effects translog cost frontier in STATA using the sfpanel command on some panel data. I have two outputs (Y1 Y2) and two input prices (X1 X2) as well as 5 environmental variables (Z1 Z2 Z3 Z4 Z5) which I would like to test putting in the inefficiency term and a time trend "t" . The problem is that my results come out insignificant with values in the elasticities and cost efficiencies which are not correct given previous litterature results and economic theory. (I have tested both just logged variables and de-meaned all my logged variables including the Z's and the results are both times insignificant).
The general translog looks like this:
sfpanel lnTC = lnY1 lnY2 lnX1 lnX2 1/2(Y1)^2 1/2(y2)^2 1/2(x1)^2 1/2(x2)^2 lnY1lnX1 lnY1lnX2 lnY2lnX1 lnY2lnX2 lnX1lnX2 t
The steps I took for the coding are:
//declare data set as panel
xtset company year
//Constrained Estimation;
//Set constraints for Translog
//all betas on the input prices must sum to 1
constraint define 1 lnX1 + lnX2 = 1
//all betas on the cross output with input prices must sum to 0
constraint define 2 lnY1lnX1 + lnY1lnX2 + lnY2lnX1 + lnY2lnX2 = 0
*all betas on the cross input prices and inputs prices must sum to 0
constraint define 3 lnX1lnX2= 0
*sfpanel Model true fixed effects: Half normal dist, environmental variable modelling in the inefficiency term;
sfpanel lnY1 lnY2 lnX1 lnX2 1/2(Y1)^2 1/2(y2)^2 1/2(x1)^2 1/2(x2)^2 lnY1lnX1 lnY1lnX2 lnY2lnX1 lnY2lnX2 lnX1lnX2 t, cost model(tfe) distribution(hnormal) usigma( Z1 Z2 Z3 Z4 Z5) constraints (1 2 3)
//find cost efficiency scores
predict e, jlms
If anyone could offer any help I would very much appreciate this. Perhaps there is something wrong in my code or in the way I set up my variables?
Comment