The Forensic Accountant's Toolbox
Article, CPA Insider – January 2009
Insurance / Property
Walt Jakl discusses the use of tools for vetting accounting information in damage calculations. He describes the practical application of Benford’s law and how its use aided in negotiation of lost revenue in a loss of earnings case.
As a forensic accountant, much of my practice consists of evaluating damage claims. When working on such claims, we often encounter resistance about the extent of information we request. Because of the demands on short-staffed business owners, we make every effort to minimize the volume and type of documentation requested. Of course, doing so comes with the risk that the limited provided information will result in an inaccurate portrayal of business operations.
Tools of the Trade
CPAs in forensic practice use a variety of tools to evaluate accounting data. Calculations of economic damages include analyses of accounting data, verification of that data, organizing the data into a useful format, calculating the economic loss and presenting conclusions in a clear, concise and supportable manner. Because the use of incomplete, erroneous and/or fraudulent accounting data will invariably result in an erroneous economic loss calculation — and with all of the negative implications and embarrassment that such a result brings — it is imperative for the forensic accountant to assure that provided accounting data presents an accurate reflection of the subject entity’s operations and financial condition. Fortunately for the forensic accountant, there are many tools available to assist with the vetting process. These include documents prepared by the entity or its representatives, such as tax and other compliance returns, general economic data, industry data, and analytical and statistical information. These tools become even more important when the type and amount of information provided is restricted for some reason.
I recently received an assignment that proved just how important some of these tools can be. The claimed loss involved a small family-owned business situated in a strip mall. A delivery truck backed over a fire hydrant several blocks away from the business. In a bizarre sequence of events, when the valve for the fire hydrant was turned off, the resulting pressure surge burst a pipe located on the walkway in front of this business. The resulting flood filled the floor of the business with several inches of water. As a result, the business was closed for some time. The owner made a claim for lost earnings.
By the time that we were assigned the file, the owner had become very upset over a number of issues, real and imaginary, that had cropped up. Her frustration had risen to the point that she was in no mood to cooperate with our request for documents. What was worse, she had decided that she was going to ""stick it to the guilty party"" for the frustration she had endured. We received only a fraction of the information we requested. And that was only after several contentious telephone conversations and the passage of considerable time.
In a matter such as this, we routinely request periodic non-loss time-period sales information supported by some form of source documentation or government-compliance report. Instead, the business owner provided only a spreadsheet with a list of what was represented to have been daily sales for several weeks preceding the event. The average of the reported daily sales tied in to claimed lost sales.
Something didn't seem right about the reported sales. We noticed that there was a preponderance of some numbers — sevens, eights and nines — and few or none of others — zero, one and two. A sample of a few days' worth of daily sales looked like this: $3,837.46, $3,995.69, $3,943.31, $3,499.25 and $3,287.58. It was then that I recalled a tool that I had never had the chance to use before: Benford's Law.
Background on Benford's Law
I first read about Benford's Law in a 1995 article in The Wall Street Journal describing Professor Mark Nigrini’s success at catching financial fraud through the application of Benford's Law. Nigrini, now Associate Professor of Accounting and Information Systems at The College of New Jersey School of Business, had also written an excellent article for the May 1999 Journal of Accountancy. I quickly located the article and reacquainted myself with Benford's Law.
Frank Benford, while working as a physicist at the GE Research Laboratories in the 1920s, discovered that certain pages of a book of logarithms were more worn than others. He noted that the logs of numbers with low digits were consulted more often than logs of numbers with high digits. Later studies, corroborated many times over, revealed that populations of numbers not artificially assigned or constrained to maximum and/or minimum limits followed predictable probability patterns.
Tables included by Nigrini in his article provided the probabilities of digits occurring in the first through fourth positions in numbers. The table demonstrated that regardless of a digit's position in a number, the probability of its occurrence was predictable. People are naturally predisposed to select some numbers more often than others, so when fraudulent data is presented, an analysis using Benford's Law will likely detect it. Knowledge of this phenomenon and its behavior allows auditors to investigate large amounts of data relatively quickly for fraudulent activity. I encourage interested readers to consult Nigrini’s article for a thorough discussion of Benford's Law.
Our analysis of the provided claimed sales history using Benford's Law confirmed our initial suspicion that the numbers were fabricated. In fact, the results were quite dramatic. We discovered that the occurrence of digits was the inverse of what was predicted.
To my surprise, when confronted with our analysis during a meeting of both parties, the business owner put up only a weak protest. After some discussion, during which the insurance carrier notified the owner that in her state, a fraudulent claim could be summarily denied, the owner elected to withdraw the sales numbers she had earlier provided. She readily agreed to provide all of the information we originally requested including copies of sales tax returns and other sales documents containing actual and verifiable sales. The resulting economic loss calculation, while substantially less than originally claimed, was accepted by both parties and a settlement was successfully negotiated.
There are a number of tools that forensic accountants can employ in financial investigations. By choosing the tool most appropriate to the nature of the matter — in this case, Benford's Law — the professional can be assured of an accurate outcome.
As appeared in CPA Insider, January 2009.