Article Abstract:
New ways of analyzing dollar-unit samples have been suggested for auditing use recently. Results are presented of a simulation study that contrasted four of these new techniques. Also offered are additional simulations on optional forms of two of these methods. Evidence on the relative size of the upper error bounds and reliability for each method is offered for 2,160 combinations of confidence levels, sample sizes, and types of accounting populations. Results offer a comprehensive evaluation of the capabilities of each of the techniques, with all four methods demonstrating both weaknesses and strengths. The Multinomial Dirichlet bound is very reliable, but as formulated it is also very conservative and not capable of utilizing understatement errors. The two Cox and Snell bound versions tested showed considerable variation in reliability over a range of confidence levels.
User Contributions:
Comment about this article or add new information about this topic:
Article Abstract:
The relative performance of two auditing procedures for dollar-unit sampling, the quasi-Bayesian method developed by McCray (1984) and a modified version of the moment method developed by Dworin and Grimlund, are tested. Data include simulation studies of 96 representative accounts receivable and inventory populations. Results indicate nominal confidence levels for both methods, and tighter bounds for both methods than those provided by the Stringer procedure. The modified moment (MM) method is easier to use than the quasi-Bayesian (QB) method. The MM method provides a tighter confidence bound at low error rates; the QB method provides tighter bounds for populations with high error rates.
User Contributions:
Comment about this article or add new information about this topic:
Article Abstract:
Research on market risk assessment and to determine whether funds and cash flows possess information beyond accrual earnings is described. Cash flow risk measures were found to provide better explanation of variability in market betas than were provided by the earnings risk measure. Earnings betas do not provide better explanatory power than either funds or cash flow betas. Accrual earnings information appears to be a subset of cash flows information.
User Contributions:
Comment about this article or add new information about this topic: