Improving fraud risk assessments through analytical procedures

Abstract

This study incorporates concepts from accounting and criminology literatures to develop a model of financial statement variables that provides researchers and auditors with information about the likelihood of fraudulent financial reporting. This study is also one of the first to test whether the predictive ability of fraud indicators has changed over time. Game theory suggests that if fraud firms consistently manifest similar characteristics then auditors will isolate those fraud indicators and react to them. The results show that accruals, market-to-book ratio, and lack of a Big Four auditor are the most influential fraud risk indicators. Fraud firms generally have higher abnormal market returns in the year of the fraud and larger market-to-book ratios compared to nonfraud firms. This finding suggests that managers who commit fraud generally attempt to preempt bad earnings news rather than react to bad news in the form of low company stock price. The results do not suggest a fundamental shift in fraud risk indicators. However, while the lack of a Big Four auditor was not significant in the 1970's and 1980's, it was significant in the 1990's. Fraud firms in the 1970's and 1980's appear to have more debt relative to the industry but are not more profitable. Frauds in the 1990's appear to be more profitable relative to the industry but do not have more debt. This finding may be due to more stock options and other performance-based incentives in the 1990's and a greater emphasis on beating analysts' expectations.