Cost Control with Imperfect Parameter Knowledge

A variety of decision models for cost variance investigations have been suggested in the accounting literature. However, few of these models explicitly have recognized that users of these models seldom have perfect knowledge of the model parameters, particularly parameters such as the average cost when "out of control." In this paper, the issue of parameter uncertainty in cost control is addressed in two ways. First, the expected cost (or loss) arising from misestimating the parameters is estimated using two methods (numerical approximation and simulation). Then a model comparison scheme is introduced for using reported costs to make inferences about the parameters of the cost process, resulting in a "learning model" by which a manager may use an investigation to find out about the cost process, as well as to correct an out-of-control situation.