At the recent Praxity FLVS Conference in Las Vegas, I had the privilege of attending sessions led by Dr. Mark Nigrini, one of the world’s leading experts on the theory and application of Benford’s Law.
Benford’s Law gives the expected patterns of the digits in the numbers in tabulated data. The digits are not expected to occur in equal proportions, and there is a large bias towards the lower digits—so much so that nearly one-half of all numbers are expected to start with the digits 1 or 2. These digit patterns were originally discovered by physicist Frank Benford in the early 1930s. Nigrini is a pioneer in applying Benford’s Law to auditing and forensic accounting.
I also had a chance to sit down with Nigrini and talk with him about Benford’s Law applications. This was a very exciting opportunity, as I use Benford’s Law on a regular basis in my data mining work at BKD. Here are a few of the topics we covered.
Clopton: How did you originally become interested in Benford’s Law?
Nigrini: I originally learned of Benford’s Law in a statistical decision theory course in college. There was about half a page of information regarding Benford’s Law in Berger’s statistics book that caught my attention. It struck me that this theory could be applied in an audit to help auditors distinguish between fraudulent and authentic data.
Clopton: What role do you see Benford’s Law playing for accountants?
Nigrini: Benford’s Law is an effective, high-level overview test. It serves as a good test of reasonableness for a data set. Benford’s Law is not too effective in finding the “needle in a haystack”; it is not designed to find one unusual transaction out of a large population. Rather, it helps in identifying data sets that have a large number of unusual duplications.
Clopton: Based on this, I assume Benford’s Law should be used at the start of a project? What do you see as the next steps?
Nigrini: Yes, it should be used at the start of a project. Other forensic analytics tests then should be used to drill deeper to find the actual underlying issues. Examples include the largest subsets, a relative size factor analysis, same-same-same and the same-same-different analysis. In instances where the data may not conform to Benford’s Law, “My Law” could be used to identify changes in the data from period to period.
Clopton: What is My Law?
Nigrini: My Law is used to compare current-period digit patterns with some benchmark other than Benford’s Law. We could, for example, compare Q1 to Q2. This comparison uses the logic of Benford’s Law, but not that law’s expected patterns. Essentially, you are looking for consistent digit patterns from period to period. This is used to identify changes from the organization’s “normal” pattern of activity. We’ve used this before in payroll applications and also in alumni gifts to a college fund.
During the conference, attendees were wowed by Nigrini’s presentation regarding Benford’s Law. Nearly everyone was talking about this topic in the next break.
As for application in forensic accounting, I use the principles from a couple of Nigrini’s books in my work. These books include “Forensic Analytics” and “Benford’s Law.” The Benford’s Law page has some very useful information, history, example data sets and analysis templates.
Benford’s Law is interesting and applicable in forensic accounting. I definitely encourage you to read more and consider how you can apply this in your work.
