Audit Sampling Using Benford's Law: A Review of the Literature with Some New Perspectives

Abstract

ABSTRACT Accounting studies have used the premise that nonconformity to Benford's Law (hereafter, Benford), which gives the expected patterns of the leading digits in numerical data, is a red flag for fraud. This study reviews Benford's Law and divides the accounting applications into five categories. A proposed Benford-based audit sampling method, which selects as the audit sample the set of transactions or balances that needs to be removed from the audit population to leave a remainder that conforms to Benford, is reviewed and reexamined. The finding is that the method, as advocated, can generate large audit samples and that the accuracy rate is questionable, even when known errors are seeded into the data. The study then reviews some new perspectives on using Benford's Law in auditing by reviewing (1) the mathematical bases for expecting Benford conformity, (2) the type of auditee data that are appropriate for Benford-based sampling, (3) various options to limit the sample size, and (4) the limitations of a Benford-based sampling approach. These perspectives draw on some facts related to the way in which the HealthSouth Corporation financial statement fraud was executed.