"Robodebt" is the label applied to the Australian government initiative starting in 2016 to increase recoveries of "overpayments" made to social security recipients. There have since been two reports by the Commonwealth Ombudsman and two Senate inquiries (one ongoing). In November 2019, the Federal Court ruled that central features of the program were unlawful. Following a class action on behalf of those affected, the government has now conceded and will refund $720 million to 470,000 individuals. The total budgetary cost (including refunds plus cancelled future savings) is about $1.5 billion, not including additional administrative costs involved in recalculating "debts", and potential compensation. The class action is continuing and submissions to the Senate point to the personal suffering of the individuals and families affected.
How did a government come to the decision to develop an unlawful policy, and why did it continue with it for nearly four years?
A properly independent inquiry like a royal commission, with the power to compel witnesses and require evidence to be produced, not just a Senate committee inquiry, is needed. Not only to clarify what happened and who should be held accountable, but to identify lessons for the future.
What is most striking is that the central failure - the calculation of overpayments and the raising of debts were unlawful - should have been apparent from inception. Under the Social Security Act, income support payments are based on income earned in the fortnight for which payments are made. Averaging - which was applied to nearly 70 per cent of those who were told they had debts - involved dividing annual income reported to the Tax Office by 26, as if recipients earned exactly the same amount each fortnight. Common sense tells us that this would not apply to people who moved on or off payments part-way through the financial year or to people whose earnings vary from fortnight to fortnight. Not only is the latter situation increasing as casual work has become more common, but social security policy for the last 30 years has actively encouraged clients to supplement benefits with part-time and casual work.
The Commonwealth can hardly claim that it wasn't warned. The legal basis has been questioned since public complaints started in late 2016. In July 2017, Peter Hanks QC gave the National Lecture on Administrative Law at the Administrative Law National Conference in which he criticised the Commonwealth for its new method of raising and recovering what "Centrelink chose to describe as 'debts'". Between April and September 2017, Professor Terry Carney - who served for nearly 40 years as a member of the Social Services and Child Support Division of the Administrative Appeals Tribunal and its predecessor Tribunal - made five judgments that there was no debt where the Department of Human Services calculated overpayments applying averaged annual income to specific shorter periods. This may well be the tip of an iceberg of unpublished adverse AAT decisions.
Were the APS Values of 'committed to service' and 'impartiality' upheld, and was the obligation to act strictly according to the law met by the administrative processes used?
DHS could have appealed these judgments to the next level of the Administrative Appeals Tribunal where rulings would have been published, but chose not to - suggesting that it was not confident of its position. In an article published in 2018, Professor Carney argued that the failure of a person to disprove the possibility of a debt is not a legal foundation for a debt. This was also pointed out by Peter Hanks in his 2017 address, and it was the conclusion reached by the Federal Court in 2019.
In May 2019, DHS completely dropped a debt against an individual (the Masterton case) which, if tested in court, may have led to robodebt being found unlawful. Again in the Amato case in September 2019, DHS dropped the debt (indeed, the debt of $3,200 plus a 10 per cent penalty was recalculated by DHS as an underpayment of $480), but the case continued in the Federal Court, ultimately leading to the present situation.
A second clear failure of proper administration is the apparent reversal of the onus of proof. The Social Security Act requires the Commonwealth to prove that a debt exists, but the DHS letters and processes placed the onus of proof on the clients (in the Amato case, the client first found out about the "debt" when it was recovered from a tax refund owed to her). The processes used do not reflect "model litigant" requirements, particularly for treating vulnerable and poorly informed citizens.
Who should be held accountable?
As Richard Mulgan has pointed out in this paper (The Public Sector Informant, July 7) the convention is that the relevant ministers should be held to account. At the time of the initial development of robodebt, the Human Services Minister was Marise Payne and the Social Services Minister was Scott Morrison; more recently Stuart Robert has taken the main role in defending the government's actions. The failure to correct the situation for nearly four years adds weight to the responsibility of the ministers involved; ministers should also be held responsible for creating a general view of welfare recipients as welfare cheats and the welfare system as a burden to society. Malcolm Turnbull has recently indicated his regrets about the introduction of the scheme and that he was sorry about the distress it had caused.
But it also seems likely that such a technical measure originated from within the public service. It would also have been scrutinised by departments other than DHS and the Department of Social Services through the budget process. Were ministers adequately advised of the legal issues involved? How were the estimates of savings made? How were the risks assessed?
The public service should also take some of the responsibility for administration. Were the APS Values of "committed to service" and "impartiality" upheld, and was the obligation to act strictly according to the law met by the administrative processes used? Did anyone attempt to warn ministers, if not why not, and were any who did so properly protected?
A royal commission could shed light on future policy and administration issues, some going beyond social security.
The appropriate use of modern technology is one such issue. DSS/DHS has long been an early adopter of information and communications technology. As the number of clients increased it developed increasingly sophisticated administrative systems, including to ensure compliance. From Stratplan in the 1970s to data matching in the 1980s and 1990s, ICT systems helped to ensure more accurate payments, identifying both overpayments and underpayments, these being reported publicly in annual reports and program performance statements through the 1990s. While DHS' more recent ICT advances have been built on now dated legacy systems, a major project is under way to replace and enhance them.
Such uses of technology are essentially neutral: they improve the efficiency of the payments system, supporting greater accuracy in the implementation of the policies set in the legislation. But as they apply algorithms to determine payment eligibility in large numbers of cases, important issues arise including as to who is the decision-maker and what rights of review clients should have, including rights to access a human decision-maker. Poorly designed, the systems can also lose their "neutrality", this seeming likely with underpayments no longer being publicly reported.
Such systems also facilitate increasingly complex eligibility criteria, or "system driven policy reform". The danger here is that inadequate attention may be given to appropriate program design or sensible overall policy design. Robodebt is not the only example here: the cashless debit card and the increasingly complex system of means tests and eligibility rules have been enabled by technology but not always introduced with adequate attention to the policy implications.
A broader lesson here may be renewed recognition that simplicity in tax and transfers design is important, not just for administrative efficiency but for public understanding and for supporting desired behaviours, such as incentives to work and to save, and for enhancing policy coherence.
Depending on the findings, a royal commission might also identify lessons for the APS and the appropriate degree of independence it should exercise in administering programs like social security. Should Centrelink be re-established as a statutory authority rather than just (part of) an executive agency under a ministerial department? Should the sort of processes recommended by the Thodey Review for secretary and agency head appointments and terminations be considered again? These are all major issues that should not be considered in a partisan way. A royal commission would allow them to be properly explored.
- The authors are from the ANU Crawford School of Public Policy. They worked in DSS in the 1970s and 1980s; Peter Whiteford and David Stanton were also there in the 1990s.