The statistical analysis of standard-cost deviations

Very little is known about the theoretical statistical properties of accounting numbers. In this paper a probability modelling technique is used to analyse the volume, efficiency, and price deviations – more usually known in the accounting literature as ‘variances’ – of standard costing. Standard-cost deviations are defined as differences between certain types of conditional and unconditional expected costs. The basic statistics required to construct theoretical confidence intervals for these differences are identified, and it is shown how the formulae normally used in their calculation may sometimes be correlated. The modelling technique adopted is based on a theory of accounting measurement which interprets accounting numbers as specialized types of statistics. It illustrates the potential of the theory for the development of a better understanding of the statistical properties of accounting numbers.