Common Fraud Detections Methods

Abstract

The literature related to the detection of fraudulent financial reporting exhibited different methods employed to detect fraud. These methods include auditor’s analytical procedures, statistical models, digital, textual and data-mining models. These methods attempt to detect fraudulent financial reporting using financial and non-financial variables as proxies (indicators) for misrepresentation or omission of material facts (amount or disclosure) in the financial reporting. The misrepresentation and omission of material facts or disclosure are the two important indicators of fraudulent financial reporting (Goel & Gangolly, 2012). To detect financial statements fraud, researchers and practitioners employed quantitative, qualitative, and mixed methods for both financial and non- financial variables as proxies/ indicators of fraud (red flags) using statistical analysis, digital, textual and data-mining analysis. The accuracy of a fraud-detection method varies for each method based on the type of variables employed, and the type of analysis applied. In addition to auditor’s Analytical procedures, researchers and practitioners have developed multiple models (based on the quantitative and qualitative components of the company’s financial reporting) to predict financial statement fraud-risk and to classify companies financial reporting to fraudulent or non-fraudulent one. These methods include discriminant analysis models, statistical models (Dechow, 2011), digital analysis (Hsieh & Lin, 2013), data-mining models (Lin, Chiu, Huang, & Yen, 2015; Zhou & Kapoor, 2011), and textual (linguist) mining models (Throckmorton, Mayew, Venkatachalam, & Collins, 2015).