Stochastic Frontier analysis with multiple output

[Jing Zhang Nottingham]

Hello, Everyone!

I am using stata to calculate the efficiency/inefficiency of innovation outputs of different countries. The data includes 63 countries over 16 years, and is unbalanced panel of 838 obs.

I have used stata to calculate the efficiency with one output using "sfpanel". Now I want to include multiple output variables (patent, publication and high-tech export), and inputs (R&D, R&D personnel, etc).

But sfpanel is only for single output, in this case, which command shall I use in stata? Or any other software can do this?

Thank you very much!

Jing Zhang

[Sixten Borg]

Hi Jing,

I asked a very similar question on the Stata list a while ago ( https://www.statalist.org/forums/for...=1505284694717 ), with no response unfortunately. Contrary to what I have heard, I am beginning to think it's not possible to do SFA with multiple outputs in Stata. I did however find that you can do it in R using the Benchmarking package, although I have not done it myself. It's briefly demonstrated how to do it in Bogetoft and Otto, in section 8.2 (Bogetoft, P. and L. Otto (2011). Benchmarking with DEA, SFA, and R. New York, Springer.).

Good luck!
Sixten

[King Duho]

I am doing an SFA study on banks efficiency with one output using stata. I have been trying to obtain the codes to run the analysis but I am finding difficulty. Please I need your assistance.

[Chaojie JIN]

I think it is possible to use sfpanel to do it. Have you employed a translog equation?

[Juan Price Elton]

Dear all: I am new doing efficiency analysis with Stata. I want to estimate production and cost functions for multi output firms using a balanced panel. As I understand xtfrontier is like the frontier command for panel data but still for one single output, am I wrong? It'd be great if any of you can lend me a hand. Many thanks. Best regards, Juan