Remedying Robodebt: The Unanswered Questions

With the onset of coronavirus crisis, all Australians have recognised that we must to come together to protect those most in need. We have rushed to support unprecedented measures to build resilience and encourage the solidarity needed to push through unique challenges. As I’ve discussed elsewhere, the speedy reset of our social security system reflects a long-delayed and needed recognition that it represents a critical piece of national infrastructure. There remain, however, areas we must fix so we can generate the community needed to fight the challenges facing us. One is the urgent need to remedy robodebt.

In a show stopping leak ten days ago, the Guardian obtained a ministerial briefing admitting that hundreds of millions may have been unlawfully recovered from ordinary Australians. The government’s legal advice admits that debt amounts reflecting fortnightly averaged data will need to be refunded. While all centrelink debts have been frozen for six months, there is still no word on how we will remedy the online compliance initiative’s painful legacy. Those with averaged debts are amongst those most affected by the current unfolding crisis. Young workers in and out of casual employment. The carers and single parents with variable income. They are still facing an unknown future liability, and, if they are currently receiving a payment, cannot get an advance to assist cash flow.

We are, again, at a profoundly serious juncture for this programme. These now identified debts, should not be held out as starting points for conversation or reassessment. Unlawfully calculated debt amounts have no legal status and should be returned. Promptly refunding already identified robodebts is the most efficient and healing action.

Have we found all the unlawful debts?

The Guardian report finally provided us with a preliminary, estimated number of how many debts are covered by the Amato litigation. The Ministerial brief estimates the unlawfully calculated debts at 449,500.

Since November 19th, 649 staff had been manually clicking into 594,636 files and checking if they are crudely averaged in whole or in part. This expensive process was required as Services Australia never tracked the use of the unlawful averaging technique. Somewhat incredibly, we still have no idea if this count is finished or what stage it is at. It may now be suspended so staff can administer the unprecedented wave of Jobseeker applications. What the Guardian report claims, however is that averaging featured in 73% of the checked robodebt files. That 15,000 Australians, not knowing their legal position, “accepted” the ATO data being applied to their records. This was far beyond initial estimates. It speaks to how people were overwhelmed by the process.

The 594,636 starting point figure needs to be analysed by Parliament. It may be too low given what we know about the Department’s debt raising:

  • Two trial cohorts from 2014-2015 need to be checked.
  • In the first Ombudsman report, Services Australia admitted that since the 1980s, it felt it had the right to average where there were no records available. Generally, prior to robodebt, employers were asked to provide the records to centrelink. In 2015 that changed. What is unknown however, is how often files were averaged when no records were found, or when the company had gone out of business. Someone needs to question the Department on this unknown cohort. Has anyone recently successfully challenged an averaged debt earlier than 2014-2015?
  • It needs to be confirmed that debts raised under the “interim manual” programme in 2015-2016 have been checked. While “robodebt” entered the public arena when the full automation went live in 2016, averaging was a core feature from 2015.

To identify all robodebts you need to find and add all the above cohorts. You then subtract the debts we know have already been zeroed (63,397) and some ‘bank interest’ debts. Fundamentally, we don’t yet know if the 594,636 checks are sufficient to find all the debts raised through the unlawful method.

One governance lesson from all this is that all legislation should contain a data reporting schedule which provides for the collection and publication of key data points. The opportunity cost of using 649 staff to ‘count’ must never happen again. Those staff could otherwise have been available in the traumatic early days of the current crisis. We must urgently find the people who paid back or are paying back averaged amounts. We cannot compound things by waiting for vulnerable people to call.

What is the Government’s plan? Why aren’t the debts being paid back?

The government previously refused to tell the senate how many debts were found and how it would deal with the found debts. It claimed public interest immunity on the basis these immutable facts would somehow prejudice the settlement negotiations (not the legal arguments!) in the pending class action. The notion that democratic oversight should be suspended for such pragmatic leverage is deeply concerning. The Guardian report has now blown the doors off that purported justification. The now published legal advice indicates the Gordon Legal class action is likely to succeed in its unjust enrichment argument. At any other time, the leak would have triggered a political storm.

Since November, those who have called and queried their robodebt (often using this excellent Victoria Legal Aid resource) have been parked until an unknown date. As the legal advice in the report admits, the interest bill attaching to any unlawfully calculated debt can accumulate. The Guardian reports that the government still wants to wait until victims of averaging self-identify themselves. Given that only 1.6% of robodebts were ever independently reviewed, the idea that people will self-identify their averaged debt is unsustainable. Why would refunds take upwards of a year when the presence of averaging is confirmed? This unannounced process of an unknown nature was to commence at the wholly arbitrary date of July. Now we don’t even know if that process will go ahead.

Another significant aspect was the brief’s proposal to ‘reassess’ 80,000 debts out of the “in scope” cohort. This might be the first grudging, too limited admission that a full process of reassessment would be highly inefficient. These debts are aging everyday as underlying pay and banking records disappear and businesses close. Once a file has to be “worked” rather than just blankly averaged, the putative business case of the programme collapses. Returning the debts immediately, reducing the possible interest bill, is the hard-headed action that is needed. But where is it? Frozen debts continue to press down on people’s lives at this very stressful time.

Can we find out where we are, and what caused all this?

After four long years battling the programme, we all naturally reacted to acknowledgment of a potential refund with a wave of relief and emotion. There is however, a long way to travel before we address the hurt. Learning these details through a leak chimed with the engrained, institutional defensiveness that critics have been met with for the past four years. There is no clarity about the scope and nature of any refund. Public scrutiny of the federal budget is being unfairly impaired by the now defunct public immunity claim. We need clarity. For instance:

  • The headline amount of $555 million appears to be an estimate of the fraction of ‘in scope’ debts that have already paid back. But what is the total value of the 449,500 debts?
  • What are the costings for Services Australia unknown, year long ‘reassessment’ process? What proportion of victims does the government believe will “self-identify”? What is the estimated interest that might accrue where amounts are held that long?
  • Will the government commit to proactively notifying all those it has identified of the class action? How could it instruct its lawyers to make submissions to the court opposing this, despite its legal advice?
  • How are the forward estimates being revised to reflect the fact it can no longer reflexively average files but has to get evidence?

There has also still been no revision of the Guide to Social Security Law to reflect the need for Services Australia itself to gather evidence and properly calculate any debt. Instead individuals calling up citing the Amato litigation were still being asked chase hard pressed employers for evidence rather than getting the debt quashed promptly.

Finally, we must not disappear the need to restore trust to those who administered their own debt. Over one million letters were sent to Australians, asserting rights which the Government did not have. Many vulnerable people went through the trauma and expense of chasing historical documents under the shadow of asserted ‘debt amounts’ which were far higher than what they owed. There is no legal remedy for them – their experience needs to be marked.

An apology remains appropriate, and answers are still needed. Why was the programme not perceived as a risk in 2016, 2017, 2018, 2019? What discussions took place as critics, lawyers and those affected spoke out? In the event the class action settles, these questions will never be answered. The dynamics which drove the greatest single failure in public administration in the past decade may never be fully known. This might reflect the normal course of politics. Regardless of your political persuasion, however, it would be deeply unhealthy for us to not learn the lessons from this debacle. Post immediate crisis, a full independent inquiry into our social security system is needed.

Darren O’Donovan