Stochastic Frontier Models

Abstract

The stochastic frontier model was first proposed by Aigner, Lovell and Schmidt (1977) and Meeusen and van den Broeck (1977) in the context of production function estimation. The model extends the classical production function estimation by allowing for the presence of technical inefficiency. The idea is that, although the production technology is common knowledge to a group of producers, efficiency in using that technology in the production process may vary by producers, with the degree of efficiency depending possibly on factors such as experience, management skills, and so on. Given the technology, fully efficient producers may realize the full potential of the technology and obtain the maximum possible output for given inputs, while less efficient producers see their output fall short of the maximum possible level. Therefore, the underlying technology defines a frontier of production, and actual outputs observed in the data may fall below the frontier because of the presence of technical inefficiency.