The Australian National Audit Office is not a regulator. It cannot fine anyone, cancel a licence or order a department to change course. What it can do is look: at the accounts of every Commonwealth entity, and, at its own discretion, at whether government programs are run well. It reports what it finds to Parliament. That is the whole of its power, and it is a considerable one.
On the part of the job the law makes compulsory, the ANAO delivers. In 2025-26 it completed all 243 mandated audits of Commonwealth financial statements. Its quality frameworks were found to be operating effectively. The parliamentary committee it reports to, the Joint Committee of Public Accounts and Audit, thanked it "for the high quality of its reports and analysis" (ANAO Annual Report 2025-26, p 16).
The problem is the other part.
Fewer audits than Parliament asked for
Performance audits are where the ANAO examines whether money is being spent well: procurement, grants, Defence projects, service delivery. They are discretionary, and they are what gives the prospect of an audit its force.
For most of the last 25 years the ANAO tabled between 44 and 54 performance audits a year. In 2025-26 it tabled 39, against a target of 38 to 42. Leaving aside 2015-16, when a double dissolution stopped reports being tabled, that is the lowest number in the 25 years of annual reports examined.
The ANAO is candid about why. In 2024-25 it cut its target in-year from 48, "the number the Parliament has requested", to 44, and set 38 to 42 for 2025-26 (Annual Report 2024-25, p 4). In 2025-26 it explained that reducing its deficit "required a reduction in the target number of performance audits" (Annual Report 2025-26, p 2). Its corporate plan for 2026-27 commits to "a historically low number of performance audits".
The office has recorded a net deficit in every year since 2018-19. Parliament's committee has been blunter than the office. In March 2025 it said "The current funding model for ANAO is simply not sustainable". In May 2026 it supported an increase of $10.5 million a year as a priority. The 2026-27 Budget statement for the office reads: "The ANAO has no budget measures in the 2026-27 budget." Funding for the performance audit program falls in 2026-27, and staff numbers fall with it (Portfolio Budget Statements 2026-27).
Those are the government's decisions, not the ANAO's. But the ANAO has absorbed the shortfall in the work that reaches furthest. The Defence Major Projects Report, an annual public account of the military's largest acquisitions since 2008, also ended. The parliamentary committee decided to stop requesting it after the Auditor-General, asked for a way forward, advised in December 2025 that it was no longer fit for purpose, citing Defence's growing non-disclosure of information and the cost of the work. When the Senate asked in April 2026 for it to be reinstated, the Auditor-General replied that this would need a request from the committee and would require reductions in the performance audit program.
Agreed is not the same as done
The ANAO's recommendations are almost always accepted: 96% in 2025-26. Entities also tell it they have implemented them: 92% within two years, on their own account.
The ANAO is the only body Regulator Review has examined so far that checks that kind of claim and publishes the answer. Its follow-up audits of selected entities found that 25% of recommendations entities reported as implemented were "partly implemented or not implemented" (Annual Report 2025-26, p 21). That is a valuable finding, and the office deserves credit for it. It also means the 92% headline overstates, by an amount the ANAO has measured in samples, and the ANAO still marks its implementation target as met on the self-reported figure.
A door that is barely open
The law gives a member of the public no right to request an audit, and that is not the ANAO's fault. What is within its control is how it treats what the public sends. Each audit in progress has an online contribution facility. The ANAO's own wording is that contributors "will not automatically receive feedback" and that "contributions are intermittently monitored". In 25 annual reports it has never published how many contributions it receives, or what became of them, and it has never published a count of complaints about itself. State audit offices in Victoria, Queensland, Western Australia and Tasmania, and the UK National Audit Office, count what the public sends them.
Requests from members of Parliament fare little better. In 2017 Andrew Wilkie MP asked the Auditor-General to audit Centrelink's automated debt recovery. The Auditor-General declined to re-examine it, because the Commonwealth Ombudsman had already reported on it, and said the matters would be considered within a proposed broader audit. The scheme, later known as Robodebt, became the subject of a Royal Commission, which reported in July 2023 and found it was not legal. No ANAO audit of the scheme while it operated has been identified, and when the ANAO audited Services Australia's debt management in 2022-23 it excluded the scheme from scope.
Reporting less about itself
By the standards of the bodies reviewed so far, the ANAO is open about its own performance. It publishes targets and results, admits errors, and has a standing Independent Auditor who audits the office itself.
The 2025-26 annual report nonetheless reports less than the year before. For eight years the report carried a table of every audit request from a member of Parliament and what became of it. In 2024-25 there were ten requests, and none led directly to an audit. The 2025-26 report has no table and no count, although the requests are still listed on the ANAO's website. The performance framework has gone from 17 measures to nine, partly by folding several into one. A revised cost method, applied to all three cost measures, reports the missed performance audit cost measure as 2% over target rather than the 6% rise in dollars. A timeliness target missed by more than half, 48% against 100%, has been lowered to 85%.
The grade
Regulator Review grades the ANAO C: adequate. It does its compulsory job completely and well, its reports are respected, and it is more willing than most to measure itself honestly. It is not graded higher because its discretionary audit program is at a 25-year low outside the dissolution year and below what Parliament asked for, because what the public sends it is neither counted nor reported, and because its reporting about itself went backwards this year.
Much of that is a funding story, and the funding is the government's call. Parliament's own committee has said the arrangement threatens "the robustness of the vital audit function" (JCPAA, 12 May 2026). The government has also not responded to that committee's 2022 review of the Auditor-General's governing Act, four and a half years after it was tabled.
The full report, with every source, is available on this site.