Thu Nov 3

The Robodebt scheme failed tests of lawfulness, impartiality, integrity and trust

Written by Adam Graycar, Professor of Public Policy, University of Adelaide

The Robodebt royal commission is currently hearing evidence of tremendous hardship ^[1] inflicted on people by a government that appeared to have little concern for the people its actions affected.

The bureaucratic process was malign, and it harmed and stigmatised welfare recipients.

Despite questions about the scheme’s legality, the program recovered about $750 million from around 380,000 people by a process called income averaging ^[2]. This involved sending debt notices that came as a surprise to the recipients who had virtually no options to challenge these notices.

The whole process was outrageous. It led to severe trauma among many of the poorest people in the country, mental health episodes and some reported suicides ^[3].

When the matter finally came to court as a result of a class action involving about 400,000 people, the government made a $1.2 billion dollar settlement ^[4] in 2020, but did not admit liability. It was absurd politics and absurd financial management.

What was the main driver? Was it recovery of money, or something else? The royal commission will answer these questions.

Irrespective of the final outcome, Robodebt lacked integrity and projected the government’s deficit of trust towards its most vulnerable citizens.

This wasn’t an example of artificial intelligence gone mad. Automation in public administration is inevitable and can bring great benefits. AI expert professor Anton van den Hengel, wrote in an email to the authors of this article:

Automation of some administrative social security functions is a very good idea, and inevitable. The problem with Robodebt was the policy, not the technology. The technology did what it was asked very effectively. The problem is that it was asked to do something daft.

Centrelink took income data from the Australian Taxation Office (ATO) and matched it with social welfare recipients’ income as self-reported to Centrelink. These income data were averaged, and though income fluctuated making recipients eligible for welfare payments, many were issued with debt notices.

The onus of proof was firmly placed on welfare recipients to prove their innocence.

There’s evidence many public servants were uneasy with the process ^[6]. Evidence is also coming out that government lawyers cautioned about its legality ^[7].

The ATO also advised the Department of Social Services the scheme was “unlawful ^[8]”.

Evidence is being heard now by the royal commission regarding the scheme’s lawfulness. This isn’t new, as law professor Terry Carney wrote in 2018 ^[9]^[10]

there can only be a debt if another provision [of the Social Security Act] creates it. There is no relevant provision ‘automatically’ creating a debt just because data-matching shows a discrepancy…

Carney also pointed to the flawed arguments in the government’s legal defence of debt averaging practices and the effective reversal of the onus of proof onto welfare recipients.

Impartiality is a central feature of government, but Robodebt had a disproportionate impact on welfare recipients. The complexity of the welfare system, and the lack of access – particularly in remote areas – ensured many were effectively unable to challenge debt notifications they received from Services Australia. People with the technical means to engage with a large bureaucracy were often able to challenge the notices, but those who weren’t were left out in the cold.

Integrity and trust in government is fundamental to a civil society and one in which legitimacy is accorded to actions of government, even though they may be unpopular or harmful to some people.

In this case, the fundamental principles of governance – a shared view between citizens and the state – fell short of normal expectations. The policy undermined the trust between government and the people, which resulted from an inability to establish a system to correctly identify and review debts owed to the government.

Governments are often concerned about the diminishing trust citizens have in them. Yet Robodebt sent a clear message to Australians that their government did not trust them.

Overall the scheme lacked integrity. It was a malign policy – and without even going into issues of whether it was designed to be malign or whether it became malign over time – and it set a very poor example for the conduct of government.

The lessons for the future are that lawfulness, impartiality, integrity and trust should underpin all government actions. These are so obvious that it seems superfluous to state them the way we have. But hopefully this will be one enduring lesson from Robodebt.