FRONTIER Procedure

The FRONTIER procedure analyzes stochastic frontier production or cost models. These are linear models with an error term that includes two components: a symmetrically distributed component that allows for random shocks of the production or cost, and a systematic positive component that adjusts for technical or cost inefficiency. The inefficiency component of the error term can be from a half-normal, exponential, or truncated-normal distribution.

Stochastic frontier production models were first introduced by Aigner, Lovell, and Schmidt (1977) and Meeusen and van den Broeck (1977). Kumbhakar and Lovell (2000) provide a good introduction to stochastic frontier analysis. Sickles and Zelenyuk (2019) provide detailed information about this topic.

For a detailed introduction to the stochastic frontier models that PROC FRONTIER supports, see the section Details: FRONTIER Procedure.