To Catch Fraud, Just Do the Math

Numbers play a large role in my career as a turnaround authority. When I am trying to save a company from financial ruin, I spend a lot of time looking at the books. All those hours I spent in math class have paid off.

My favorite use of math is uncovering fraud, which costs this country an estimated $300 billion a year. So a recent article about fraud in WSJ caught my eye — “Accountants Increasingly Use Data Analysis to Catch Fraud.”

Seems that auditors at KPMG discovered that the numbers at a national call center weren’t adding up. The operators there each issued more than 10,000 refunds a year and were authorized to issue checks for up to $50.

The firm decided to analyze some data and applied something called Benford’s Law. When they did so, they determined there were too many fours in the refund checks. So what is Benford’s Law and what’s wrong with the number 4?

Also called the First-Digit Law, it refers to the frequency of digits appearing first in a list of numbers. It seems that the number one occurs first 30 percent of the time, with each successive number occurring less and less. The number 9 appears first less than 5% of the time.

This holds true whether you’re talking about street addresses or stock prices. That’s probably why we are advised not to start a passcode with the number 1.

The phenomenon was first mentioned by the American astronomer Simon Newcomb in 1881. Then physicist Frank Benford ran all sorts of data that included things like the surface area of 335 rivers, sizes of U.S. populations and numbers listed in an issue of Reader’s Digest. He found the phenomenon to hold true in a variety of situations. The law was then named after him.

So the accountants applied Benford’s Law, and took a look at the first number of the refunds issued by the operators. When they hit the number four they saw a huge spike in the number of refunds. Can you guess what was happening?

Turns out some of these operators, less than a dozen, were issuing fraudulent refunds at the $40 level, knowing they wouldn’t be checked.

These refunds totaled several hundred thousand dollars. And the only way they were caught was by applying Benford’s Law.

Whenever there is a set limit like this, I always encourage people to ask their auditors to look below it.

I tell the story in my book, “How Not to Hire a Guy Like Me: Lessons Learned from CEO’s Mistakes” of a CFO at a manufacturing company in Wilmington, DE. He knew the level at which auditors would review expense reports was $5,000. So what did he do? He wrote himself lots of checks in the mid-$4,000 range.

Part of my review process at a company that is losing money is to look at expense reports at the senior level. I became really suspicious when this CFO dragged his feet in getting me the information I requested. I learned why when I found $180,000 of recent fraud.

Here’s a tip I offer anyone looking for fraud in a business. If you have any type levels set up like this, make sure you or your auditors are sampling below the limit of that transaction to make sure you are not getting defrauded. That seems to be a sweet spot for those committing fraud. And you may want to add Benford’s Law to your fraud detection toolbox.

One thought on “To Catch Fraud, Just Do the Math

  1. Pingback: Funny, But True: Two Cases of Too Many Zeroes – The Turnaround Authority

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s