Treasury eyes expanded data access for agencies in fight against fraud
Amid a governmentwide push to root out fraud in federal programs, the Treasury Department is exploring ways to broaden agency access to certain payment verification datasets.
Speaking at a FedInsider webinar this week, Justin Marsico, assistant commissioner for fraud prevention and financial integrity at the Treasury Department’s Bureau of Fiscal Service, detailed internal efforts to strengthen Do Not Pay, a resource it provides to agencies and states that manage federal programs.
Pulling from myriad public-sector datasets, the online tool allows its users to verify the eligibility of people or organizations seeking federal money, providing a secure way to detect risk or anomalies before making an award or certifying a payment.
Marsico said Treasury has engaged in “a lot of administrative work” to acquire those datasets, while ensuring that the right cybersecurity controls are in place. But there’s an appetite for more information, including access to tax information covered by section 6103 of the federal tax code.
“We have been working with IRS to scope out some potential legislative proposals to make it possible for some federal agencies to get access to validate [Employer Identification Numbers] or to act to validate income level using that data,” Marisco said. “We haven’t cleared the hurdles on that yet, but it’s definitely on our radar.”
Authorized by 2019’s Payment Integrity Information Act, Do Not Pay has already been expanded since its creation, but Treasury officials believe there’s more room to grow. Appearing before the House Oversight Government Operations Subcommittee earlier this month, Treasury official Renata Miskell said the system has been significantly “underutilized.”
Do Not Pay lacks “sufficient authority to access key federal databases that could detect the most common drivers of improper payments, namely verifying identity, financial status and death,” Miskell, the agency’s deputy assistant secretary for accounting policy and financial transparency, told lawmakers.
“One of the pieces that we are missing is the ability to ping authoritative federal databases to confirm a pay ID, such as a tax identification number or Social Security number,” she added. “We already received the data, we just can’t verify it. So there are a number of databases that would help.”
Access to more government datasets aligns with Marsico’s stance that the data-sharing tool “is at its heart … a governmentwide collaboration effort,” the “centralized” nature of which makes identifying fraud and “improper payment detection really easy.” The whole-of-government approach means more eyes on insights or trends “that are not visible to individual programs,” he said.
In addition to considering new datasets, Treasury has looked at artificial intelligence and machine-learning tools to bolster Do Not Pay’s efficiency. Marsico said they’ve already found some “new, sophisticated techniques” that are in production and effective, while others are still in an exploratory phase.
There’s also ongoing work with the Social Security Administration’s Death Master File, Marsico noted. Allowing agencies to check “against death-related data sources” at the time of payment, he added, could be another big development for Do Not Pay and the government’s overarching fight against fraud.
Historically, the government has pursued an approach “where the priority is like, just get the money out the door. We’ll figure out if it’s improper, and kind of claw back later,” Marsico said.
“What we are trying to do,” he continued, “is push detection earlier into the life cycle, so we are identifying problems right after they occur — or hopefully before they occur.”
