Key takeaways

In 2019, the Federal Court ruled on a class action against the Commonwealth. It found that the government had been illegally raising debts for historical social support payments.

It did this through a A$1.2 billion welfare savings program from the 2015-16 federal budget. This program was later labelled Robodebt.

Robodebt automated comparisons of historical fortnightly social security records. It matched them with assumed income for the same period, based on tax data. When a discrepancy was found, a debt was raised – also automatically. The method was inaccurate and illegal.

According to new research from the Australian National University (ANU), Robodebt was neither a one-off nor an error of judgment. It was a culmination of decades of institutional change. Specifically, the annual expansion of welfare compliance measures to meet savings targets.

The research identified three institutional arrangements that will have to change to prevent another Robodebt:

• The Commonwealth budget process: At the budget’s outset, priorities are set by the prime minister, the treasurer and the finance minister. Then, their corresponding departments negotiate with line agencies to ensure that all budgets align with these goals. For decades, these goals have included demands for improved efficiency and reduced spending. In practice, this means that every year, agencies like Services Australia must find savings and offset any new spending.

• The divide between policy and service delivery: This has unintended consequences. In the social services portfolio, policymaking is the remit of the Department of Social Services (DSS). However, Services Australia – formerly the Department of Human Services (DHS) – delivers those services. Every year, each function has to find its own savings. And an agency only responsible for service delivery has just two ways to meet targets: efficiency gains and stricter compliance.

• Data-matching-driven compliance: The use of data-matching has grown steadily for decades, and been gradually automated, as it’s considered the ‘least harmful’ path to savings. Despite the rigidity of service delivery, data-matching has kept savings in reach. However, until Robodebt, it was informed by human decision-makers. The automation of the whole process – from calculation through to debt notice – was new.

The findings indicate that the Royal Commission’s recommendations aren’t sufficient to prevent another Robodebt.

They also explain why Services Australia dedicates millions to fraud, compliance and debt activity, despite the fact that the Royal Commission found that the scale welfare fraud is “miniscule.”

By the time Robodebt emerged, the question wasn’t ‘should we do more compliance?’. It was, ‘what compliance should we do next?’

Institutional factors didn’t make Robodebt inevitable. And they don’t excuse lawbreaking. But they help explain how it happened.

To prevent another Robodebt, leaders have to go beyond the recommendations of the Royal Commission. They have to rethink how the Commonwealth budget process interacts with the division in social services. This approach would be in line with the latest evidence.