Cost driver selection in activity-based costing systems

Activity-based costing (ABC) uses multiple cost drivers when assigning overhead costs to products. In this thesis, the selection of cost drivers is taken into account through two key decisions to achieve an optimal ABC system: which cost drivers to use and how many cost drivers to use. Single objective optimization models are presented that focus on the maximization of overall correlation between cost drivers and cost pools by considering different preferences in the selection process such as maximum number of cost drivers, maximum measurement cost, and maximum number of measurement hours. A multi-objective approach is also developed to provide a better explanation of trade-offs in the cost driver selection process. The accuracy loss of the system is explored by integrating a measurement of accuracy into the multi-objective model. Two potential situations, having invalid cost drivers and correlated cost drivers, which the system developer may encounter in the decision process, are also incorporated into the models. Model use is illustrated through example ABC systems. Results indicate that the methodologies developed in this thesis enable the determination of the most appropriate cost drivers to capture the actual consumption by products of cost pools.