Seemingly Unrelated Regression (SUR) with Panel Data

[Stefan Wimmer]

Dear Stata Users,

I want to estimate a system of equations, consisting of a cost function and multiple input-demand functions. The input-demand functions do not add any new parameters, therefore I need constraints indicating identical parameters across functions.

Taking no account of the panel characteristic of my data, I can estimate:

  constraint define 1 [xinput1]_b[input1] = [cost]_b[input1_squared]
  constraint define 2 ...
  ...
   
  sureg (cost: C = X X X) (xinput1: Qinput1 = X X X) (xinput2: Qinput2 = X X X) (xinput3: Qinput3 = X X X), isure constraints (1-36)

I know that the xtsur command can handle panel data, but - as far as I know - does not offer the constraint option. Please let me know if I am wrong.
I also read about the CMP command and tried:

  cmp(Cost​ = X X X) (Qinput1 = X X X) ​(Qinput2 = X X X) (Qinput3= X X X), indicators(1 1 1 1) vce(cluster firm_ID)  level(95)

However, I am not sure how to add constraints to this equation. Also I am not sure if vce(cluster) really represents panel data.

Is there a better way to account for panel data in a SUR equation system?

I would really appreciate if somebody could help me out in this matter.

Thanks a lot,
Stefan

[Marilyn Parada]

Hi Stefan, I would like to know if you could solve seemingly unrelated regression with panel data and constraints, because I have a system of equations with constraints and the comand XTSUR only allows me solve the system without constraints,
Thanks