Test of Proportions Screening for the Newcomb-Benford Screen in the Audit Context: A Likelihood Triaging Protocol

Abstract

The basis of the certification audit, in a non-forensic context, is a random sample of sufficient size to create the evidence needed to justify the audit opinion. This has been clearly stated in all of the relevant SAS pronouncements issued by the AICPA and also by the PCAOB through AS 5. However, there has been a dearth of specifics on the critical selection of the set of accounts that are reasonable targets for the statistical sampling needed if extended audit procedures seem warranted. In this paper, we present a simple and validated protocol based upon the digital frequency testing introduced by Newcomb & Benford and popularized in the audit context by Nigrini to identify accounts under audit that seem reasonable candidates for extended procedures testing. The montage of the protocol centers around the parametric test of proportions the equations which were introduced by Nigrini. We validated the logic of the protocol by using a holdback sample. Finally, we have coded the account identification protocol as a Decision Support System in VBA: ExcelO that is available from the authors free without restriction to its use.