With data analytics, program managers can take a more active role in preventing fraud and abuse at their agencies.
Managers can use analytics to spot and prevent fraud several ways, according to panelists at ACT-IAC’s July 20 Fraud and Abuse Forum. They can not only uncover fraud and improper payments and prevent future fraud, but they can also run cost-benefit analyses to determine the best tech investments by leveraging data matching, predictive analytics and risk assessments
Linda Miller, now a director at Grant Thornton, encouraged “agencies to think about fraud from a risk-based perspective,” and to proactively use risk assessment.
“You don’t know how much fraud you have because you’re likely not catching most of it,” said Miller, who led the development of the framework while she was assistant director for the Government Accountability’s Forensic Audits and Investigative Services team.
The solution, she said, is “to figure out how much fraud you have,” through a “fraud risk assessment at the front end that’s thorough and comprehensive.” Risk assessment is also critical in making the right tech investments for future fraud prevention, she added.
A January GAO report noted that governmentwide improper payments climbed to $144 billion in fiscal year 2016. The distinction between improper payments and fraud is an important one, as administrative errors can dramatically affect those numbers. Data matching can help identify potential instances of intentional fraud.
Deloitte’s Dave Mader said at the event that when he was controller at the Office of Management and Budget, he helped create different categories for agencies “to start to deconstruct their improper payments so that we could better understand which of those payments was out-and-out fraud.” This was part of an effort, he said, along with the Fraud Reduction and Data Analytics Act, to help agencies “better understand their data” and “better understand that there are different treatments that would be required, based on that analysis.”
For Full Story, Please click here.