Attorneys can run into fraud in a wide range of settings — from criminal and civil cases to questionable valuations and damages calculations. And with the overwhelming amounts of financial data being churned out by businesses every day, detecting such wrongdoings might seem like looking for a needle in a haystack.
Fortunately, qualified fraud experts can zero in on fraud, and produce reliable substantiating calculations, by using a number of data analysis techniques.
Data Analysis Advantages
Data analysis provides the ability to review and find relationships between pieces of data — even when such financial information is spread across multiple business systems and
applications. Advanced technologies allow savvy data analytics experts to mine years’ worth of material for signs of fraud, such as suspicious matches between names and addresses.
Most data analysis technologies are capable of keeping comprehensive logs of the various activities performed. Doing so can be critical when a fraud investigation ends up in court.
Fraud experts can employ one or more of a variety of data analysis techniques, including:
- Calculation of statistical parameters — for example, averages or standard deviations — to identify outliers,
- Stratification of figures to identify suspiciously high or low figures,
- Classification of data by, for example, geography, to identify patterns,
- Duplicate testing to identify duplicate invoices or transactions of, for example, payments or claims,
- Gap testing to identify missing values in sequential data such as that contained in purchase orders, and
- Entry date validation to identify suspicious or inappropriate times for data entry.
These analyses can be conducted on an ad hoc basis for a specific investigation or continuously as an automated procedure aimed at both detecting and deterring fraud.
Several data analysis techniques rely on Benford’s Law, which relates to the frequency with which digits in some types of real-world data (including many types of financial information) appear. For example, numbers beginning with 1, 2 or 3 are more likely than those starting with 4 through 9. When fraud perpetrators fabricate data, though, their understanding of random numbers usually doesn’t correspond with reality. Believing all nine digits are equally probable, they tend to select a relatively equal distribution of numbers beginning with 1 through 9 — a misstep easily uncovered by data analysis.
It’s important to remember that data analysis techniques often stop short of actually proving fraud. Statistical deviations, for example, might occur for nonfraudulent reasons. But thorough, properly conducted analyses can lay out a more-focused road map for further investigation of suspicious events or transactions.
If you have questions about this, or any other matter, visit www.warrenaverett.com to learn more about our Forensic, Valuation and Litigation Support Services.