DHS discovers $1.7bn in welfare fraud through data matching

The Department of Human Services has discovered $1.7 billion in welfare fraud through new data matching processes that have finally made it cost effective for the department to chase the overpayments.

Senior bureaucrats today told Senate estimates that prior to recent advances in technology, it would have cost the department more to go after these debts than it would make in collecting them.

The DHS and the Australian Taxation Office routinely match their respective data sets, and in recent years uncovered 1.1 million instances where around 800,000 welfare recipients reported a different income to the tax man than they did to human services.

DHS Secretary Kathryn Campbell told Senate estimates it was finally becoming cost effective to go after these cases and recoup an estimated $1.7 billion in retrospective overpayments.

“At some points it was going to cost us more to act on that information,” she said. "The cost of doing that has now decreased."

The 2015 budget laid out a plan for the department to return $1.7 billion to government coffers over five years by investing in better fraud detection and debt recovery capabilities.

DHS business integrity manager Mark Withnell told the hearing that about $10 million of the $40 million set aside for the anti-fraud campaign would pay for new analytics technologies. Much of the remainder will be invested in staff to carry out investigations and debt recovery.

“We know there are cases we need to look into,” he said. “We need to spend the $10 million so we can cost effectively deal with them."

He said existing analytics tools used by the department can already tell the difference between intentional fraud and simple human error - often the difference between a fake identity and a misplaced decimal point, he said.

Just last week, the DHS announced a partnership up with the Department of Immigration to compare records of people claiming Australian visas as the spouse of a citizen or permanent resident, to catch out any ‘fake couples’ reporting as singles for Centrelink purposes.

DHS today also flagged greater use of analytics for compliance activities like matching study loads to those reported by student welfare recipients, and automating the exchange of information with AUSTRAC when it comes to welfare recipients with unexplained wealth.

Under current Commonwealth data matching guidelines, agencies need to publicly gazette plans for any data matching exercise that compares the records of more than 5000 people.

While the new compliance campaign is not reliant on the $1 billion back-end overhaul of Centrelink’s payments system, the upgrade will help, officials said.

Withnell said the detection systems his team currently use draw batches of payments data out of the core Centrelink system intermittently.

The Centrelink infrastructure is so unstable, he said, that if he “hits” the system too frequently it can cause serious problems. The new infrastructure - which is roughly seven years away - will be able to do these operations in real time.