Between 2009-10 and 2016-17, the number of social security fraud prosecutions in Australia fell by 80% as a proportion of Centrelink’s customer base. This is remarkable considering national dialogue continues to focus on punitive approaches to welfare fraud. The Department of Human Services, which absorbed Centrelink in 2011, has even reiterated its “zero-tolerance” approach to the issue.
My research suggests this swift and dramatic decline in the welfare fraud prosecution rate is tied to fundamental shifts in the department’s approach to prosecutions. This has led to a less punitive culture among its Serious Non-Compliance staff.
Legal and policy changes
Several legal and policy changes have contributed to the declining prosecution rate. This includes the impact of two High Court cases – one from 2011 and another in 2013 – on so-called “omission cases”. These are cases where the allegation of fraud relates to a customer’s failure to inform the department of a change in circumstances (an omission), rather than any deliberate misrepresentation or dishonesty on the recipient’s part.
In the 2011 case, the court ruled that because there was no legal obligation for Centrelink customers to inform the department of changed circumstances, failing to fulfil this duty could not constitute fraud – or, more specifically, the offence of “obtaining financial advantage” under the Criminal Code.
The government sought to preempt this decision by swiftly passing a law creating such an obligation, making it retrospective to the year 2000. But, the retrospectivity of this law was successfully challenged in the 2013 High Court case.
Feeling lonely or just want to make new friends? Come join the MDM Club for free. Chat in a safe and tolerant space, created especially for the disability community.
These cases’ combined effect was to prevent the department from prosecuting “omission cases”, where the relevant omission had occurred prior to August 4, 2011 – the date on which the law requiring welfare recipients to inform the department of changes in their circumstances was passed.
After this date, such omissions constituted fraud and so could be prosecuted. But these two High Court decisions effectively reduced the potential pool of cases that could be prosecuted – at least in the short term.
At around the same time, a new requirement was introduced making the department responsible for preparing full briefs of evidence, including witness statements, for each welfare fraud case it wished to refer for prosecution.
Previously, the department only needed to prepare a short-form brief of evidence, which amounted to little more than a statement of facts. According to the department, this change:
… impacted the (prosecution) referral numbers as significantly as the High Court decisions.
Together, these changes impacted on the number and types of cases the department could pursue. But they only tell part of the story.
I conducted in-depth research on welfare fraud, including interviews with Serious Non-Compliance staff at the department during 2014 and 2015. Other factors emerged that were far more influential in reducing welfare fraud prosecutions over the long term.
From prosecuting everything to targeting serious fraud
Until 2011, Centrelink compliance functions, including prosecutions, were subject to quantitative targets. For example, in 2000-01, Centrelink was required to refer 4,000 cases for prosecution and generate A$708.8 million in savings by recovering welfare overpayments.
To meet these organisation-wide targets, Centrelink developed targets for individual investigators. For example, in 2008-09, some investigators were required to complete 96 investigations and refer at least six cases of suspected fraud.
According to compliance staff I interviewed, these requirements led to a focus on achieving targets rather than selecting the most appropriate cases for prosecution.
Janice, an investigator who had been with the department (and its predecessors) for more than 17 years, described the approach:
… we used to just prosecute, prosecute, prosecute. Quite frankly, it used to be about a number, it used to be about a benchmark.
In 2011, these targets were scrapped, reducing the pressure on staff to make up prosecution numbers. Since then, a more strategic approach to prosecutions has emerged – one focused on serious and unambiguous cases of fraud. As Henry, a member of the Serious Non-Compliance team, explained:
The directors (in the DHS) at least do know that it’s not a numbers game. It’s finding people that legitimately need to answer for their actions in front of the court rather than getting a referral to the court for the sake of achieving a number.
My research suggests this new approach has also underpinned the emergence of a less punitive culture among compliance staff, where staff consider the appropriateness and impacts of prosecution in each case.
One investigator explained the shift as:
… more about prevention and intervention than it is about the end result. We’re more conscious of what we’re doing, and the impacts as well.
Similarly, as Bruce explained:
I don’t believe in a heavy-handed approach. If someone is defrauding the Commonwealth, until we actually determine that, they’re customers. They’re not crooks … The culture has changed.
The wider reform debate
The general approach to welfare compliance in Australia continues to be punitive and underpinned by a suspicion of people who use welfare. The recent passage of a bill that will introduce a punitive demerit system for non-compliance is testament to this.
Nevertheless, the approach to social security prosecutions has undergone significant change. The removal of quantitative targets, combined with a more strategic approach to prosecutions, has contributed to a sustained reduction in the welfare fraud prosecutions rate.
This has occurred despite, rather than because, of changes in federal government policy. It serves as a small reminder that punitive welfare measures can be effectively contested and sensibly reformed, even as the government continues to campaign to “get tough” on welfare non-compliance.