Government Senators' Dissenting Comments

Government Senators disagree with significant elements of the Committee’s majority interim report, offering the following additional dissenting comments.
Government Senators also note that this is the second interim report tabled by the Committee and that this dissenting report should be reviewed in conjunction with the first interim report dissenting comments.

Overarching concerns

It is noted with some concern by Government Senators that the majority report only selectively incorporates evidence the Committee has received during its inquiries into Centrelink’s Income Compliance Program (Program). Frequently, where there is evidence from Commonwealth agencies, and in particular Services Australia or the then Department of Human Services on a point of concern being explored in the report, that evidence is not presented, or the issue is not considered in context. This approach is inexplicable noting the volume of evidence the Committee has received from the Commonwealth both during the current inquiry and previously in 2017, and this has led to basic errors.
For example, the majority report includes multiple inaccurate assertions that averaging of ATO income data was used in the identification of discrepancies before seeking confirmation from an individual as well as to subsequently calculate debts. This is not the case, and a completely separate methodology for the identification of discrepancies was as outlined by Services Australia in its evidence to the Committee in 2017.1
The Committee has also in practice adopted a position of accepting evidence from groups and individuals, including assertions and hearsay, without ever having examined or tested these claims directly, or in the context of other evidence provided to the Committee. The Committee has done this by, amongst other things:
raising claims by individuals about the experience and impact of receiving initial discrepancy letters, without noting that this was not a new feature of compliance activity established by the Program;
raising claims by individuals about the experience and impact of the review/appeal process, without noting that this was not a new feature of compliance activity established by the Program; and
not exploring the impact on the Program of the decision by significant numbers of people to simply not engage in the processes or not even open letters sent to them by Centrelink.
This approach has seen the Committee include in its report a number of misleading blanket statements and oversimplifications about events that have transpired as part of the Program, and the numbers of those so impacted and how they were impacted.
This approach is disappointing given the need for accuracy given the public debate on this topic.

Legal insufficiency

Whilst it is now widely understood that it is insufficient to rely on the averaging of ATO income data as the sole basis for raising a debt, this does not mean the Program overall is either ‘illegal’ or ‘unlawful’. The Government’s decision to no longer rely solely on income averaging to raise a debt but rather to seek additional proof points was announced by the Hon. Stuart Robert MP, Minister for Government Services on 19 November 2019. The legal insufficiency of relying on averaging along was acknowledged by the Minister on 29 May 2020 and confirmed by evidence during the Hearing 31 July 2020.2
Government Senators note the Government has worked quickly to rectify these issues. Refunds for those who had repaid monies on debts raised solely or partly using averaged income data commenced in July 2020, and Government Senators note the majority of refunds are expected to be paid by November 2020.

History of Income Compliance

Since the first Senate inquiry in 2017, the consistent advice from agencies in the social services portfolio has been:
Australia has a targeted social security system where recipients are subject to eligibility, income and assets tests, so assistance is provided to those who need it most;
compliance activity is a fundamental activity in administering the Social Security system. This has been the case over many decades under a range of integrity measures including the Program;
compliance measures ensure people receive correct payments and help them to meet their obligations. This includes checking discrepancies, identifying potential overpayments, and appropriately recovering debts. This protects the integrity and sustainability of the welfare system;
people have always been asked to clarify discrepancies; and
the Income Compliance Program which commenced on 1 July 2015 was designed on the basis that averaging of ATO income data was sufficient to determine a debt where other information was not available, or people did not engage with the (then) Department of Human Services.3
Government Senators note that the first data matching program was established in 1990, enabled by the then Labor government’s Data-matching Program (Assistance and Tax) Act 1990.
Throughout the history of the various data matching programs, the actual data matching process itself did not change.
Key features of income compliance are longstanding and not new. Of the three core features of the Program – data-matching, income averaging and online systems/portals – only the online portal was a new feature, with data-matching and income averaging being historical features of compliance.
Even prior to the 2015 commencement of the Program, ATO information has been averaged to calculate debts in the absence of other information. When employers did not respond or were not able to be located, decisions were made on the best available information, including the averaging of ATO information.4 As noted by the majority report, random samples of debts raised in 2009 and 2011 found that this process had occurred in around 20 per cent of cases.5
Importantly, Government Senators note that the Program has not had any debt targets or quotas for debt-raising:
Compliance staff are not required to finalise a prescribed number of reviews each week. Staff are not required to meet targets for debt raising.6

‘Robodebt’ and automation

Government Senators strongly disagree with the incorrect characterisation and use of the term ‘robodebt’ to refer to the program.
The term ‘robodebt’ was first used in the media and by certain groups/people in late 2016/early 2017 in relation to debts that involved a portal which was referred to as the Online Compliance Intervention (OCI).
There is no program known as ‘robodebt’, and the OCI process that initiated the term was discontinued and superseded by the Employment Income Confirmation process from February 2017. However, the term ‘robodebt’ has continued to be used in various ways in the media and social media, causing significant confusion.
The term ‘robodebt’ has been used inconsistently to refer to a range of difference time periods:
the initial Online Compliance Intervention portal that commenced operation from 1 July 2016;7
just the three iterations of Online Compliance Intervention (OCI), Employment Income Confirmation (EIC) and Check and Update Past Income (CUPI) processes (i.e. as outlined in the initial Class Action Statement of Claim);8 and
the entire income compliance programme from 1 July 2015 to the present (i.e. as referenced in the class action further amended Statement of Claim).9
The ‘robodebt’ term has also been used when referencing a range of different aspects of the Program, including:
to describe the use of data-matching itself: '…Now, the current Government announced compliance campaigns 2015, 2016, they started introducing robodebt – that's the use of the algorithm to data match ATO data with Department of Social Security data';10
to reference the use of data matching to send discrepancy notices seeking information from individuals (e.g. applicants’ statement of claim which refer to these as being a robodebt notification);11
to describe the debts being refunded under the Income Compliance Program;12 and
to describe all debts raised under the Income Compliance program (irrespective of whether averaging was used to actually raise a debt).13
Government Senators note this high degree of inconsistency and agree with the observation of the majority report that use of the ‘robodebt’ term has included any debt or compliance letter received from Centrelink, not just those issued through the Income Compliance Program.
The OCI portal implemented as part of the Program allowed people to review and update or confirm income information online with part of the compliance process automated. However, introduction of an online portal did not in any way 'automate' debt raising – the implication implicit in the 'robodebt' slogan that decisions are automated is entirely incorrect.
Discrepancies identified through data-matching for the purpose of initiating a review are not identified automatically and are also not identified by averaging. Government Senators note key evidence of Service Australia in its submission, where it states, 'human checks and balances built into the case selection process to ensure only those most likely to have an overpayment are selected for a review'.14
In its evidence, Services Australia sought to highlight the continuing involvement of staff in all aspects of the Program, particularly under EIC and CUPI, stating:
The involvement of customers and staff is evident at every stage of the review process […]
Staff are involved in all parts of the process, from the selection of who is reviewed, through to assisting customers to complete their review, to calculating overpayments, recovering debts, and the review and appeals process […]
The Agency has invested in staffing to support customers. The Commonwealth Ombudsman’s April 2019 report acknowledged that with "direct access to a well-staffed dedicated hotline" customers have "improved accessibility to compliance officers" and that this has "bolstered the fairness of the process".15
Services Australia’s submission further outlines that the Commonwealth Ombudsman’s April 2019 report found that 'the process for income compliance is now fairer and more transparent and that improvements made to communication with customers, both through our letters and the online system, meant the process was more robust with regards to procedural fairness and useability.'16
Contrary to some of the evidence presented to the Committee, the Program has not required individuals to 'mount a defence' or engage with complex mathematical calculations – they have simply been asked to clarify and update the income they received while in receipt of an income support payment.17

Program improvements

Government Senators accept the evidence presented to both the current inquiry and the Committee’s inquiry in 2017 regarding the shortfalls, particularly in the early iterations of the Program, in relation to communications with customers at key stages of the compliance process.
There were clearly initial implementation issues when the program commenced communicating with former customers where the then Department of Human Services did not have updated contact details. However, Government Senators note that those issues have been comprehensively addressed. For example, following early changes to the Program, Services Australia’s evidence to the Committee noted:
(a) review process does not commence without the person undertaking some action. This includes receipting their initial notification of discrepancy through registered mail or myGov or going online to commence the review process via the online system, or phoning the dedicated phone line specified in their letter.18
Government Senators note that the design of the program has been informed by, and has benefited from, extensive consultation. According to Services Australia:
The current system has been developed in response to the feedback we have received from our customers, our staff, stakeholder groups, the two Commonwealth Ombudsman’s reports into the implementation of the online system, and the previous Senate Inquiry.19
According the Services Australia, evidence suggests that the fairness and usability of the latest iteration of the online system – the Check and Update Past Income (CUPI) system. Services Australia reports:
...there has been an increase in people choosing to complete their reviews online with no or minimal active assistance from staff. Currently about two thirds of reviews are commenced online, and in about a third of reviews, customers are choosing to either complete their review fully online or with some assistance from our staff.20

Debt raising

Government Senators note that debt raising is difficult, and that no one likes receiving a debt or being involved in any process that could lead to a debt outcome. This sentiment is echoed in the extensive emotive evidence received by the Committee related to aspects of the debt and review processes under the Program that were unchanged from previous practices.
While accepting some issues with implementation, the purpose of the Program is to protect the integrity of Australia’s welfare system. As noted previously in its evidence to the Committee, Services Australia has continuously strived to ensure its processes fair and transparent, and that staff are well trained to help customers navigate the compliance and debt process.
Government Senators note significant enhancements brought in through the later iterations of the Program, particularly through the rollout of the CUPI system, with greatly enhanced quality of communication and confirmation of its receipt by an individual before any action was taken. Key evidence from Services Australia was telling:
The use of registered post and read receipted email mean that we know that around 30 per cent of people do not engage with the process at all, and another 30 per cent of people who start the process, do not finish it.21
This includes individuals who did not raise any issues with Services Australia and who did not seek assistance. This evidence highlights the fundamental difficulties involved in debt-raising and debt processes, where some individuals simply do not want to engage under any circumstances.
Government Senators note the academic research leveraging economics and psychology that has occurred into 'information avoidance' by individuals in everyday life.22 This includes in relation to financial matters and health, where individuals choose not to obtain information that is freely available through responses such as physical avoidance, selective inattention, biased interpretation of information and some forms of forgetting. One form of this approach is the Ostrich Effect, where individuals 'bury their head in the sand' or ignoring negative or unpleasant information due to a Fear Of Finding Out (FOFO).
Government Senators note that information avoidance may have influenced decisions by significant numbers of individuals who had discrepancies identified as part of the Program and their subsequent non or part engagement with the compliance process. Government Senators further note that within one month of its operation, over 77 000 former customers, approximately 50 per cent of those eligible for a refund, had completed the online tasks through the dedicated portal to update and confirm their details to facilitate their refund without difficulty.23

Impacts and mental health

Government Senators note that there have been apologies from both the Government and senior officials for any hurt or harm caused by income compliance.24
Mental health and suicide are very delicate areas and there are any number of factors which can contribute. As such, talking about suicide requires sensitivity and the avoidance of unfounded conclusions.
Services Australia’s evidence to the Committee highlights that it assists people facing difficult situations every day and that they have an extensive social worker network to support people in times of crisis and vulnerability.
Services Australia’s evidence reiterated that it has taken advice from the National Suicide Prevention Adviser, the Suicide Prevention Taskforce and Mindframe’s guidelines for communicating about suicide.25 This Committee has also received the benefit of a briefing from experts in relation to mental health issues.
The expert advice is that public discussion not treated with care and sensitivity could inadvertently place undue distress on susceptible people and present suicide as an option for dealing with problems.26
Government Senators note the clear and consistent evidence from Services Australia to the Committee that it is incorrect to interpret Centrelink customer death statistics as suicides, or to attribute media reporting around suicides to the Program. Government Senators reject the repeated attempts by opposition Senators to attempt to make these links, including in relation to cases where there is no allegation of exposure to procedural unfairness as part of the Program or the use of averaging of ATO income data. This is misleading, irresponsible, and has the potential to cause harm to vulnerable people.
Government Senators make the general observation that public commentary about the Program has been riddled with misinformation. Suggestions that debts are pursued by law enforcement or that debts give rise to adverse credit reports are simply false and serve to confuse and frightened people.

Public Interest Immunity issues

As mentioned in the first Interim report dissenting comments, it is the view of Government Senators that the claim of public interest immunity made by the Hon Stuart Robert MP, on 11 February 2020,27 and subsequently reiterated on 31 July 2020 and 13 August 2020 are valid.28
Government Senators note there are currently significant active legal proceedings underway and the pleadings in a class action brought against the Commonwealth arising from the conduct of the Income Compliance Program. In his letter to the Committee of 13 August regarding a claim public interest immunity, the Minister for Government Services stated relevantly:
...the claims of the applicants in the class action include a negligence claim for damages. Possible additional claims for exemplary damages and misfeasance in public office have also been foreshadowed.
The applicants have said, in public hearings of the litigation, that the legal knowledge of the Commonwealth or Commonwealth officials at particular times would be relevant to determination of these claims. Information about the cost, timing and provider of relevant legal advice relates directly to that issue. While any evidence given to the Committee would be covered by Parliamentary Privilege and not admissible in the Federal Court to prove the matters to which the evidence relates, the disclosure of this information under these specific circumstances could potentially prejudice the Commonwealth’s position in the litigation. It would therefore not be in the public interest for the information to be publicly disclosed.
I have considered whether the possible harm to the public interest could be avoided if information about relevant legal advice was provided to the Committee as in camera evidence. Even if evidence is taken in camera, the evidence may later be published. I have decided that, given the potential magnitude of the applicants’ claims in the class action, there may still be harm to public interest if the relevant information was provided as in camera evidence.
On the basis of the comments from the Minister, Government Senators believe there is a clear case for the claiming of public interest immunity over relevant information in relation to legal advice and Cabinet deliberation.
This view was upheld by the Federal Court when it considered documents sought from the Commonwealth as part of discovery in the class action. The Court upheld the Commonwealth’s claims of certain documents being covered by legal professional privilege, and a number of additional documents being covered by public interest immunity as a result of their connection to Cabinet’s decision making.29

Refunds and Income Compliance Moving Forward

The integrity of the welfare system is important to all Australians because of the size and scope of social welfare, this equals more than $180 billion a year.
The government has previously noted that Australians owe approximately $5 billion across more than 900 000 debts.30 Government Senators therefore note that income compliance must remain a core part of measures to ensure the integrity of the welfare system.
On 19 November, the Government announced that it will no longer rely on averaging of ATO income data as the sole basis for raising a debt, and that other proof points will be sought going forward.
On 29 May 2020, the Government announced it would refund monies repaid by individuals on debts solely raised relying wholly or partly on averaged ATO income data, commencing from 1 July 2020. As of 14 August, 79 per cent of debts raised in this manner had been reduced to zero and refunded, with the value of refunds paid $568 million.31
Government Senators note that, in future, the near real-time reporting of income data by the ATO through Single Touch Payroll (STP) will make it easier for welfare recipients to check and update their income details each fortnight against what they are actually paid to prevent overpayments occurring in the first place.
Senator Wendy Askew
Senator Hollie Hughes

  • 1
    Senate Community Affairs References Committee, Better Management of the Social Welfare System initiative, Public Hearing, 18 May 2017.
  • 2
    Ms Campbell, Secretary, Department of Social Services, Committee Hansard, 31 July 2020, p. 19.
  • 3
    Ms Campbell, Secretary, Department of Social Services, Committee Hansard, 31 July 2020, p. 2; Ms Michelle Lees, Deputy Chief Executive Officer, Customer Service Delivery, Services Australia, Committee Hansard, 31 July, pp. 2–3 and Committee Hansard, 17 August 2020, p. 5.
  • 4
    Commonwealth Ombudsman, Centrelink’s automated debt raising and recovery system, April 2017.
  • 5
    Services Australia, Random sample of 500 debts conducted by Services Australia for the Minister (LEX 55417), Document released under FOI, 13 July 2020.
  • 6
    Services Australia, Submission 20, pp. 13–14.
  • 7
    Asher Wolf, 'Robodebt was an algorithmic weapon of calculated political cruelty', Canberra Times, 1 June 2020.
  • 8
  • 9
  • 10
    The Hon Bill Shorten MP, speaking at a Press Conference with Gordon Legal, 17 September 2019.
  • 11
  • 12
    Jordan Hayne & Matt Doran, 'Government to pay back $721m as it scraps Robodebt for Centrelink welfare recipients', ABC News, 29 May 2020.
  • 13
    Emily McPherson, 'Hundreds of thousands still stuck in Centrelink robodebt limbo', Nine News, 30 June 2020.
  • 14
    Services Australia, Submission 20, p. 6.
  • 15
    Services Australia, Submission 20, p. 5.
  • 16
    Services Australia, Submission 20, p. 9.
  • 17
    Name Withheld, Submission 53.
  • 18
    Services Australia, Submission 20, p. 7.
  • 19
    Services Australia, Submission 20, p. 5.
  • 20
    Services Australia, Submission 20, p. 9.
  • 21
    Services Australia, Submission 20, p. 8.
  • 22
    R Goldman, D Hagmann and G Loewenstein, 'Information Avoidance', Journal of Economic Literature, vol. 55, no. 1, 2017,
  • 23
    Ms Lees, Services Australia, Committee Hansard, 17 August 2020.
  • 24
    Ms Campbell, Secretary, Department of Social Services, Committee Hansard, 31 July 2020, p. 2.
  • 25
    Mindframe, Guidelines for communicating about suicide, (accessed 26 August 2020).
  • 26
    Mindframe, Guidelines for communicating about suicide, (accessed 26 August 2020).
  • 27
    Senate Community Affairs Reference Committee, Centrelink’s compliance program – Interim report, Government Senators’ dissenting comments.
  • 28
    The Hon Stuart Robert MP, Public Interest Immunity Claim, 29 July 2020, 13 August 2020.
  • 29
    Prygodicz & Ors v Commonwealth of Australia, Federal Court of Australia, VID1252/2019, 13 August 2020.
  • 30
    The Hon Stuart Robert MP, House of Representatives Hansard, 29 July 2020, p. 5575.
  • 31
    Ms Lees, Services Australia, Committee Hansard, 17 August 2020.

 |  Contents  |