Accounting for complexity in costing high technology manufacturing

A new type of ‘costing unit’ is proposed in the form of an adjusted measure of plant output that takes into account the increases in complexity of operations with rapid and continuous changes in products and process technologies. Use of this costing unit makes it possible to conduct meaningful analyses of past and present plant performance using historical costs, and to project future cost behaviour even in the presence of ongoing technological changes. Employing a simple power function, tests conducted in a wafer fabrication plant of a semiconductor manufacturing company show that this costing unit yields better statistical fits and predictions than are obtainable from output data that are not similarly adjusted. The power functions used for the different cost components are extended to include Cobb-Douglas forms in a simple recursive regression system that makes it possible to relate overall cost behavior to these underlying components.