Regardless of the volume of fraud in the public sector, it’s clear that a good many cities and states are attacking the problem. And we have no quibbles with that. Up to a point. Sometimes, it may cost more to find the fraud than to suffer it. That said, any fraud that can be eliminated is good for taxpayers.
Some of the most helpful antifraud approaches are reasonably straightforward. Among the most common are requiring two people to sign a check rather than one; having bank statements reconciled by someone other than the person who writes the checks; and making sure that people log into their computers via individualized passwords, rather than providing shared access. “That is an inexpensive -- probably free -- way to prevent fraud,” says David Denby, deputy state auditor of Mississippi.
Then there are purely data-based approaches. One such is derived from something called Benford’s Law. Bear with us for a moment on this, because it may seem complicated at first, but it’s really very cool. Benford’s Law demonstrates that when you’re looking at a long list of numbers from most real-life sources of data, the first digit is going to be 1 about a third of the time, and the frequency of other digits in that position declines as they get higher. A computer run of a department’s books that discovers this law isn’t being followed can lead directly to a potentially worthwhile investigation.
One approach to stomping out fraud that we’ve seen blossoming are fraud hotlines. The Los Angeles Controller’s office has a fraud hotline that is now accessible in more than 150 languages, a change made following a few recent incidents in which city employees were busted for defrauding non-English speakers. In Mississippi, many of the calls that come in on the state’s hotline don’t merit investigation for a variety of reasons. But the ability to register complaints anonymously is still very valuable. “We have calls from people in counties who know that if their name was revealed, it could cost them their job,” Denby says.
One example of a useful hotline -- intended for employees only -- can be found in Virginia. It’s run by the Division of State Internal Audit and receives more than 700 calls per year. That’s up from about 400 not too long ago, thanks to better promotion.
It accomplishes a lot. About 35 percent of those calls lead to substantiated findings of fraud (following investigations). As Tim Sadler, audit manager for the division, points out, the benefits go beyond simply identifying instances of fraud. The tips also help avoid future problems. For instance, 10 percent of the tips, though they don’t specifically uncover instances of existing fraud, lead to recommendations for actions, such as a change in internal controls.
Moreover, Sadler adds that the existence of the hotline itself is a deterrent: When people think they might be caught, they might be less likely to commit fraud. He recalls a 2004 case when a hotline-originated investigation led to the discovery that three state officials had used state funds to buy supplies for an African safari. The state recovered the funds, but as Sadler notes, this also helped the state improve agency operations to stop similar, if less exotic, abuses from coming up in the future.
The hotline seems to pay for itself. In fiscal 2010, identified savings from substantiated allegations made to the hotline were $1.27 million, about twice the cost of the hotline.
As useful as hotlines can be, they’re not a panacea. “The current whack-a-mole approach of waiting for the phone to ring to report a fraud is not a systemic way of reducing fraud losses,” says Chris Swecker, former assistant director of the FBI and a fraud consultant with Chris Swecker Enterprises. “Most government programs have some sort of integrity programs to catch fraudulent activities, but it is a one-fraud-at-a-time approach. The detection is not real time -- it is always pay and chase.”
A few more thoughts from Swecker:
- Go after people who make a living out of defrauding the government first. Then go after the folks who are opportunists.
- Governments should use data analysis software for real-time detection processes that aren’t siloed by agency, program type or fraud type. This broad approach would “enhance” the antifraud efforts already in place.
- In using databases to detect fraud, confidentiality and privacy issues can make life difficult. “You have to jump through hoops as with any other data-sharing actions.”
