Frontier production functions: Improved procedures for estimation and application

Document Type


Publication Date



This paper develops improved estimating techniques for frontier production functions and measures of technical efficiency. A clustering analysis procedure is introduced to isolate subsets of data observations on the basis of qualitative considerations. A dominance test is formulated to screen out data observations which will not influence the frontier production function. A computational algorithm is presented to efficiently implement these procedures. These techniques are applied to the steam-electric generating industry and discussed.

Publisher's Statement

© 1976 Published by Elsevier Ltd. Publisher’s version of record: http://dx.doi.org/10.1016/0305-0548(76)90002-2

Publication Title

Computers & Operations Research