Aggregation, Specification and Measurement Errors in Product Costing

Recent attention has focused on the design of cost systems that improve measurement of product costs. Much of this work has been classified under the general heading of Activity-Based Costing (ABC). There has been little systematic analysis, however, of why an ABC system with multiple cost pools, activity drivers and allocation bases generates more accurate product costs. The intuitive argument rests on the belief that multiple cost pools and multiple activity drivers better reflect the cause and effect relation between overhead resource consumption and products. Our analysis reveals the existence of trade-offs attributable to specification error, aggregation error, errors in measurement of overhead costs and errors in measurement of product-specific units of allocation bases. Our research provides some guidance for implementing ABC systems. There are two principal results of our analysis. First, partially improving specification of cost allocation bases and increasing the number of cost pools in a costing system can actually increase specification and aggregation errors. Second, reductions in specification and aggregation errors from more disaggregated and better specified costing systems may increase measurement errors and hence errors in product costs. We assume that firms do not know actual product costs, but rather implement new cost systems by identifying better cost drivers and increasing the number of cost pools under an implicit assumption that refinements in the cost system lead to improved accuracy of product cost numbers. This assumption is warranted in many cases. However, our results suggest that such incremental refinements in the cost system may actually cause product cost errors to increase. Therefore, a firm cannot assume that refining its cost system will always lead to more accurate product costs.