Implementation issues associated with the construction of an activity-based costing system in an engineering components manufacturer

A number of researchers have documented how to construct an activity accounting system from first principles (for a comprehensive listing of published activity-based costing cases see Innes and Mitchell [l]). For instance Cooper [2] in a Harvard Business School case study entitled Schrader Bellows discusses in detail the cost driver identification process and the two stage tracing process from cost centres to cost pools and then onto products. Brimson [3] has produced a monograph entitled Activity Accounting which covers in detail the method of how to construct an activity accounting system. This paper appraises the use of an activity-based costing (ABC) system some time after its initial construction. It was felt that a detailed micro analysis of a single (engineering components) manufacturer would be useful in the process of ascertaining exactly how ABC impinges on the management process. To date, some of the discussion of ABC has been based on the presentation of general intuitive ideas rather than detailed analysis of implementation or logical model construction. It is hoped that the case study analysis of an implementation presented here, will provide the focus necessary to understand the logic of when and how ABC is useful. Scapens [4] discusses the general role of case study methods in management accounting research. ABC analysis demonstrates that one should not attempt to artificially restrict the decision-making environment to one in which only volume related decisions are appropriate. In modern manufacturing environments it is not just the volume of products that matter. When production capacity constraints exist careful management of the production process is required. When production capacity constraints and multiple products exist this will be described here as the management of complexity. It is possible to view overhead costs, such as, for example, setup costs as drivenlexplained by the number of setups required to support production. Given some defined production volume over a