Last week, I spoke at the Fourth Annual Fraud & Forensic Accounting Conference in Atlanta. Thanks to all who attended the session, provided examples of real-world fraud cases and asked insightful questions. If you missed the event, the theme of my presentation was Predicting Fraud in Small Business. Given that small businesses suffer a large proportionate share of the total burden when it comes to occupational fraud, I posed the question, “What if fraud could be predicted?” I blogged about this topic last week, specifically addressing “why are small businesses so susceptible to fraud” and “solutions available to address the issue of errors and fraud“. In this wrap-up post, I’ll offer some insights on how fraud might actually be predicted.
Mr. Marks, by mandate of the District of Columbia Precrime Division, I’m placing you under arrest for the future murder of Sarah Marks & Donald Dubin that was to take place today, April 22 at 0800 hours & four minutes. – Chief John Anderton (from the movie Minority Report)
In reality, the notion of predicting errors and fraud is not as far-fetched as you may think. In fact, predictive analytics are used to address similar problems today. The best recognized example may come from the field of consumer finance where characteristics like payment history and credit utilization are used to assess the likelihood of a person defaulting on a mortgage or credit card agreement. A credit score is a number representing the creditworthiness of a person or the likelihood that person will pay their debts. It has shown to be quite predictive of risk which has helped to lower the cost of providing credit while making credit more widely available to consumers.
With this example in mind, consider how the same methods applied to different data may serve in identifying small companies that are at risk for errors and fraud. Much like payment history is an indicator of creditworthiness, characteristics like industry classification or transaction volume may be useful to predict the likelihood of financial irregularities. Take even further, an index based on error and fraud data could even be used to benchmark a company against it’s peers, similar to how FICO scores are used to measure a person’s creditworthiness.
The concept is really about finding that “ounce of prevention”, as the ability to detect errors and fraud using predictive techniques could provide benefit to the millions of small businesses fueling the U.S. economy. It’s also the nature of a research project AuditMyBooks is about to launch with the support of the National Science Foundation. If you are interested in the topic, I’d love to hear from you.
- CP Morey
