Two-tiered Stochastic Frontier Analysis

[Gaetano Francesco Coppeta]

Hello to everyone,

I am working on application of the two-tier stochastic frontier (Polachek and Yoon 1987). I was wondering if anyone could give me suggestions on how to perform the analysis with Stata.

Browsing the internet, looking for information, I came across some papers that make mention of some ado files, but I couldn't find it.

I hope any of you can help me

Gaetano

[Meera Ancy Vincent]

Hi,

I am currently working with the two-tier stochastic frontier model. I am aware of the sftt command for cross-sectional data (Lian, Y., Liu, C., & Parmeter, C. F. (2023). Two-tier stochastic frontier analysis using Stata. The Stata Journal, 23(1), 197-229). However, I am unsure how to apply this framework to panel data.

I would greatly appreciate any assistance you can provide.

Thank you!

[George Ford]

https://journals.sagepub.com/doi/pdf/10.1177/1536867X231162033

search sftt

[Meera Ancy Vincent]

Thank you for sharing the link. However, I have encountered challenges in incorporating bootstrap estimation into the scaling 2TSF model. Specifically, I attempted the following Stata commands:

sftt y x1 x2, sigmau(u1 u2 u3) sigmaw(w1 w2) vce(bootstrap, reps(1000) seed(123))

and

bootstrap, reps(1000) seed(123): sftt y x1 x2, sigmau(u1 u2 u3) sigmaw(w1 w2)

However, these commands return errors such as "vcetype bootstrap not allowed" or "varlist required".

As I am relatively new to frontier models and Stata, I would greatly appreciate your guidance on the correct approach to implementing bootstrap estimation within the sftt framework for panel data.
Could someone please guide me on how to apply Bootstrap? I look forward to any insights you can share. Thank you

[George Ford]

to bootstrap, you need the "scaling" option