Identification of fraudulent financial statements using linguistic credibility analysis

摘要

The strategic use of deceptive language in managerial financial fraud is investigated with linguistic cues extracted from 202 publicly available financial disclosures. Those crafting fraudulent disclosures use more activation language, words, imagery, pleasantness, group references, and less lexical diversity than non-fraudulent ones. Writers of fraudulent disclosures may write more to appear credible while communicating less in actual content. A parsimonious model with Naïve Bayes and C4.5 achieved the highest classification accuracy. Results support the potential use of linguistic analyses by auditors to flag questionable financial disclosures and to assess fraud risk under Statement on Auditing Standards No. 99.