Despite significant regulatory changes, accounting fraud remains a critical problem for any sized business. What makes such white-collar crimes so compelling is that they’re relatively easy to commit, especially against small companies. But you don’t have to accept victimization if you don’t have the resources for a full-scope fraud analysis. Instead, use Benford’s Law.

A commonly-deployed tool among CPAs and certified fraud examiners (CFEs), Benford’s Law is in a nutshell a probabilities matrix. Invented by Frank Benford, he discovered mathematical expectations related to naturally occurring, non-fabricated numbers. Essentially, these numbers should descend in frequency, starting from “1” being the most frequent to “9” being the least.

Benford’s Law is incredibly powerful when attempting to pinpoint accounting fraud and reducing the cost of fraud analysis. This is due to the fact that numbers are building blocks. In order to get to 100, for example, you must first pass 10. Moreover, the number one repeats most often in transcending data sets, such as income statements.

Consider the English language. No matter what you write, you’re more often going to use consonants such as “t,” or vowels such as “a,” as opposed to letters such as “z” or “x.” This applies even if you’re writing about xenophobic zebras. The latter two letters just don’t appear that often in English.

But if you ran a model similar to Benford’s Law on an English writing, and you discovered an unusual amount of these letters, you immediately know that the author was acting deliberately.


Benford’s Law a Powerful Crime Detection Tool

So it is with accounting fraud. Benford’s Law states that the number one should appear as the most-represented digit 30.1% of the time. Practically speaking, especially if you’re doing fraud analysis on a general ledger, the rate is likely much higher.

Keep in mind that you’re looking for broader trends over a large data set. Just because your general ledger shows higher-than-normal instances of specific numbers doesn’t necessarily mean fraud. However, over one year or several years especially, you should see a “waterslide” curve: higher frequencies of ones and twos, and fewer frequencies of eights and nines.

If, however, that doesn’t happen, you should investigate further. For instance, if you have unusually-high occurrences of nines — which is a “rare” number – that’s your initial clue. At that point, drill into which year, quarter, or even month is causing the highest variances against the norm. Then, initiate a concerted investigation into that specified timeframe.

Best of all, Benford’s Law is cheap to implement. You can use a spreadsheet program like Microsoft Excel to get going right away.