An Effective and Efficient Analytic Technique: A Bootstrap Regression Procedure and Benford's Law

Analytic procedures allow auditors (or forensic accountants) to identify the presence of unexpected relationships among financial data, including unusual transactions or events, and to develop the scope of audit planning or investigation (Corbett and Clayton 2006, Bhattacharjee and Machuga 2004, Asare and Wright 2001).The effectiveness of analytic techniques e.g., trend analysis, ratio analysis, or a regression analysis depends on the precision of expectations developed (Kenyon and Tilton 2006).However, there is a caveat associated with these expectations. Specifically, Hitzig (2004) notes that “…because the expectation is specified by the auditor, and because the expectation’s parameters are estimated from data that may be misstated [or contain fraudulent transactions], there is a risk that reasonable expectations are incorrect… ” (p. 34). As such, the risk of failing to detect the presence of suspicious trends, patterns, or anomalies within the financial data likely exists. This problem is referred to herein as the risk of Type II error. On the other hand, analytic procedures infer propriety of individual transactions based on the reasonableness of aggregated account balances or classes of transactions (Stringer and Stewart 1986, p. 16). The percentage of misdiagnosis associated with these procedures ranges from 15 % to 50 % (Kogan et al. 2010) and research indicates that the use of daily or monthly disaggregated data improves th