So this robodebt thing sounds big. What do you need to know?

This piece is all about giving those who haven’t been following the issue a quick overview of the main things driving the public controversy over robodebt. What are the questions that keep coming up? Who is affected? How big is this?

What is a robodebt?

At law, it is the combination of two discredited, blunt ideas:

  1. The crude assumption that annual ATO PAYG returns can be averaged (divided by 26) and substituted for the person’s original reporting, producing a debt. This doesn’t work because the social security law demands calculation based off fortnightly earned (not received) gross income. Earnings vary, people are in and out of work, have complex deductions and working credits.
  2. Reverse onus: the principle that unless the person produces a raft of payslip or financial information, the ATO data constitutes adequate supporting material to establish a debt.

So: from a legal perspective, a robodebt is a debt issued without the required meaningful human intervention by the decision-maker. It is the application of averaged ATO data, without sufficient regard to the standard of certainty or evidence demanded by the law. Yesterday Justice Davies of the Federal Court made a clear finding that in Victoria Legal Aid’s client Deanna Amato’s case, the decision-maker could not validly reach a conclusion a debt existed using the averaged data.

This finding would not have been made prior to 2015, when the Department generally used its formal powers to gather employer evidence prior to issuing debts.

Who is the face of robodebt?

Robodebts are produced as part of the employment income compliance programme. You must have been in and out of work in order to have received a robodebt. Working in the year you got a centrelink payment is an ironclad precondition to receiving one. The number of robodebts is likely in the hundreds of thousands. Statistically the face of robodebt is:

  • Female (55% versus 45%)
  • Under 35 (59%)
  • Not currently on welfare (52% versus 48% who get some kind of centrelink payment)
  • Received newstart (47%) or youth allowance (26%) from 2009 onwards.

These people are being asked to reconstruct their earnings pattern from 7 years ago (and beyond in some cases) when they were in casual, insecure work. Employers might be closed, uncontactable etc. People need help.

Who gets bitten by the flawed assumptions? Well:

  • People who had a little welfare while working inconsistent hours for multiple employers between 2009-. (I have seen files with as many as six employers. All averaged across the year. Heroic stuff)
  • People who work different rates with various forms of loading
  • Students who dropped in and out of hours to try and juggle study
  • Creative people, people in precarious employment of all types.
  • Carers and people with disability who juggle work hours with their own and others’ needs while on payment.

How many of these are there?

Well that’s the $666 million question (that’s the amount recovered so far, there’s actually much more debt raised).

The Department is unable to identify the precise number of files where income averaging has occurred. It will have to check at least 686,000 files to fix the cohort covered by the Federal Court decision: 

  • The Department seems to indicate in its senate submission that around 60% of letter recipients haven’t fully completed the initial ‘review’ by supplying documentation. 
  • Averaged files will correlate reasonably strongly with those who had 10% penalties applied to them for “not engaging” – which occurred in 245,066 cases.
  • There will also be mixed files where the person tracked down some employers/bank statements but not all. The averaging got bunged in like the frog DNA in Jurassic Park. It will be more labourious to find those.
  • They’ll also have to check interim manual files from 2015-2016 despite the media always saying robodebt started in 2016.

I’d say right from the off – at least a quarter of million debts can, even without looking, be dubbed as “averagey”.

But I heard there were robots?

This was never wholly a failure of technology: it was equally a failure to intervene by the decision-makers pressing send on these debts. Technology enabled a 1 million letter system, it mismatched employer names, it prepopulated a crude, boilerplate assumption out of step with modern work. We humans did the rest I’m afraid.

The unifying thread here is: a robotic assumption about averaged data made by humans, facilitated by technology. The Online Compliance Initiative is actually one of the most heavily staffed areas of the Department now. Unfortunately, at the frontline the 1000 labour hire staff spend too much time holding a clipboard with a script, a stopwatch and make evidential demands of people. The Department needs to take responsibility for ensuring the quality of the evidence and making better judgments.

So sorry there’s no lasers. No robots. Terribly robotic approach to complex issues rather. Think Bladerunner not Terminator.

But if people engage won’t the Department help?

Firstly, nothing that follows changes the fact that in an administrative action, it is the person who is seeking a result that must lead adequate supporting material. Individuals have a right to stand behind their origial good faith reporting of they so wish and are not be subjected to reverse onus procedures.

These people sat at their kitchen tables reporting their income six, seven years ago, which was something emphasised in the Federal Court action. The law says that if the Department establishes a lawful debt, and they made a mistake without reasonable excuse, a 10% recovery penalty can be applied. Nowhere in the law is there a permission slip to punish them with an inaccurate debt.

Prior to 2015 the Department would secure employer records prior to debt raising. The biggest dynamic driving robodebt is the decision to discontinue these requests and instead place the onus on the person to hunt the employer.

Two and a half years ago I criticised the failure of the Department to put in place any kind of policy where they would ever go directly to employers. After an inexplicable delay of 18 months, they finally published their policy last November, but it does not feature in their front of house communications or letters. After shouting for this tiny, marginal protection for a couple of years, we saw that it was poorly drafted and reactive only. The department starts from the premise getting information on behalf of people is exceptional only with inability to contact employers or disability listed as “a factor” for consideration. This is a ‘mercy’ discretion, not a commitment to support.

Then buried in a footnote to the second Ombudsman report on the scheme, I saw that the department had only requested records from employers for affected people 570 times. This seemed incredible given the scale of the system. I actually struggled to believe it at first. I told journalists it might be a typo. It was true. At the senate inquiry the Department updated us and stated that it has only ever got 1000 payslips for people despite having formidable statutory information powers. As most people will have multiple employers this means that the number of people who got helped to get payslips is the hundreds. In a 686,000 debt, 1.1 million “income review” system.

Paul Farrell’s 7.30 reports were a devastating portrayal of the failure to actually assist on the frontline. We watched vulnerable people, people with disability, grieving families living quietly with their debts. People’s vulnerability did not change, reporting did not change, what changed was the onus that was put on people, the people who represent us walked away from going direct to source – employer records.

But can’t people just upload bank statements?

Firstly, nothing I’m about to say affects the fact that the practical onus to adduce adequate evidence supporting a debt is on the Commonewalth. As the Federal Court ordered yesterday, individuals have a right to stand behind their original good faith reporting if they so wish and not be subjected to reverse onus procedures. End of.

As a teacher of administrative law, the most helpful thing I can do, without patronising anyone, is to underline the complexities of the files and the legal requirements here.

To know someone’s entitlement on any particular fortnight requires us to establish their pattern of earnings, how that might reduce their working credits (yep, see), take into account deductions. None of that appears on a bank statement, which only shows received net income. If your employer payed you erratically, without discipline, if you were given catch up payments? Ding ding ding you ping the robodebt software and even if you have bank statements you are on course for one heck of a contestable reconciliation process.

Bank statements cost money. They are only retained for certain periods of time, which can be shortened when account close. Enough from me: meet Victoria Legal Aid’s client Ken O’Shea, he’s the bringer of truth on this issue.

What about the 113,000 debts that already were changed after they were issued?

Please never forget those who worked so hard to do the Department’s maths for it. Many of them sacrificed health, money, productivity and the joys of life in a Kafkaesque process. People should speak out for that cohort as victims here, particularly as they are not the focus of legal actions. Read the greatest and truest piece on robodebt: Robert Skinner’s “How I fought off my robodebt”. It has even more power today.

Administrative law and the judiciary are not this society’s moral keepers. But what happened yesterday at least means the unfairness some people suffered can never ever be disappeared or denied. The Australian government has just admitted that it sent one million letters to its citizens asserting a bald right which simply did not exist. In a currently unknown number of instances, it actually populated that crude assumption into life changing debt decisions. People like Mr Skinner, whose time and stress are being called an ‘efficiency’, are the only reason the number isn’t 100% of all debts issued.

In claiming the right to average data, the department placed a moving treadmill, an imbalance of power at the centre of all interactions with it. Hundreds of thousands of Australians were denied the quiet dignity which should always accompany even tough government decisions. This has been incredibly damaging for people’s trust and relationship with a department so central to work, families and care in this country.

This is taxpayers money, Dr O’Donovan?

Whenever I hear this line, I think of my students who first got me involved in this issue. Upset, asking for extensions, trying to look after themselves, trying to make rent, with this averaging treadmill on them. Government never counts and values people’s time and lost productivity in its policy costings. So much of this was counterproductive. I saw firsthand how this policy endangered the most important things in our society: health and the idea that the best of us should get their shot. My students will never get that vital time in their lives back.

The current budget numbers reflect ordinary Australians’ status, their fear, their vulnerability. Any accurate business case needs to actually audit a sample of the files for accuracy, and inbuild front up information gathering and improved evidence. Too many of these debts have been shaped by the resources, the misunderstanding of ordinary people, rather than objective evidence.

I would give anything to see the cost modelling for the various ways evidence can assembled and people supported. At a rather sparky senate inquiry hearing, I asked for these numbers to be released so Australians could know the monetary value of different options. Nothing doing. Hard to have a debate about economics when you don’t have costings of different approaches such as gathering employer evidence.

Again, getting a robodebt is linked to one event: transitioning in and out of work, or working a little while on payments. It’s amazing how just the words “centrelink debt” conjure up a shortcut cultural images. Stopping robodebts is about supporting our hardworking young people. And er, they pay tax?

What can I do?

Listen and talk about people’s stories. Evidence shows privileged people like me are useless at actually changing things.

Tell everyone about their rights: Victoria Legal Aid’s incredible resource is here.

Unlawful debts are not starting points, at law they have no status. This Department needs to identify these files. It then needs to stop, listen and finally justify itself. It needs to submit to what it has always resisted: a full principled audit of a sample of its files, and a full independent expert review of its compliance system. Remediation of past mistakes should be carried out by qualified staff.


Darren O’Donovan