A tax is a promise. When a government hypothecates a levy — when it tells the public that this particular impost exists for this particular purpose — it makes a representation that the money raised will reach the purpose named. The Queensland Mental Health Levy, introduced on 1 January 2023 to fund mental health services, made exactly such a promise. On 22 May 2026 the Queensland Audit Office reported that the State had built almost nothing capable of keeping it.
The findings are not a matter of tone or interpretation. The Queensland Audit Office found that effective governance arrangements were not established when the levy was introduced. The Auditor-General found no reliable mechanism by which the State could confirm that levy funds were actually being directed to mental health purposes, and no systematic evaluation of whether the money spent delivered the benefits the levy was meant to secure. A separate internal Queensland Health review found that several announced initiatives had not been funded, that levy revenue was applied to programmes characterised as out of scope, and that many projects were significantly delayed with their funding deferred. Media reporting has described the consequence as the mismanagement of hundreds of millions of dollars. That figure is a characterisation, not an audited number, and it should be read as one. The governance failure underneath it requires no exaggeration to be serious.
The two questions a hypothecated levy must answer
Strip away the politics and a hypothecated levy reduces to two questions that any competent system must be able to answer at any time. Did the money go where it was promised? Did spending it achieve what it was for? These are not advanced questions. They are the floor. They are the difference between a fund and a slogan.
On the audited record, Queensland could answer neither. The first question — did the money arrive — requires that funds be ring-fenced and tracked, that someone be able to trace a dollar from collection to a mental health outcome and certify the journey. The absence of a mechanism to confirm that funds reached their purpose is not a paperwork gap. It is the absence of the control that gives the promise meaning. A levy without that control is indistinguishable, in operation, from general revenue wearing a more sympathetic name.
The second question — did it work — requires evaluation: a defined set of intended benefits, indicators capable of registering progress towards them, and an assessment, conducted against those indicators, of whether the benefits materialised. None of that was systematically in place. And here the principle is unforgiving. If you cannot measure it, you did not manage it. Spending that is never evaluated is not investment. It is expenditure with a story attached, and the story is the only evidence it was ever asked to produce.
Unmeasured, ungoverned, by design of omission
This is the textbook failure that a measurement-and-governance discipline exists precisely to prevent. The discipline is not exotic. It is a century old in public finance, older still in any enterprise that has had to account for money it did not raise itself. You name an owner. You ring-fence the funds. You instrument the spending. You evaluate the outcome against the purpose. You do these things at the outset, because retrofitting governance onto money already spent recovers nothing — it merely documents the loss. The failure in Queensland was not that the wrong controls were chosen. It is that the controls were not built. A system was announced and a structure was not.
It is worth being precise about why this matters beyond the ledger. A levy is collected from people who were told what it was for. The mental health levy was sold on a purpose that commands near-universal public sympathy, and that sympathy is exactly why the governance around it should have been tighter, not looser. Causes that are easy to defend are the ones most easily invoked to wave money through. The more unimpeachable the purpose, the greater the obligation to prove the money served it — because the purpose will otherwise be made to do the work that evidence should do. Sympathy is not a control. It is the thing that controls are there to discipline.
What good would have looked like
None of this is hindsight, and the remedy is not mysterious. Good would have looked like four things, all available on day one.
First, ring-fenced funds with named ownership. Levy revenue held separately, traceable from collection to application, with an accountable officer answerable for its direction — not a committee in which responsibility dissolves, but a person who can be asked the question and is expected to have the answer.
Second, a governance structure stood up at inception. A coordinated, whole-of-system arrangement defining who decides what the money funds, against what criteria, with what authority to refuse a use that falls outside scope. Out-of-scope spending is not an accident that befalls a well-governed fund. It is what occurs when nothing is empowered to say no.
Third, leading and lagging indicators. Measures that show, early, whether committed initiatives are actually being delivered, and measures that show, later, whether the intended benefits to Queenslanders materialised. Indicators turn a promise into something falsifiable. A fund that cannot be proved to have failed can never be shown to have worked.
Fourth, independent evaluation beyond the reach of those who ran the programme. Assessment of outcomes conducted by parties with no stake in the verdict, reporting to someone other than the spenders. An organisation that cannot countenance a measurement that could embarrass it has not commissioned a measurement system. It has commissioned applause. The point of independence is that it permits the embarrassing answer, which is the only kind worth paying for.
The current government has signalled that it will tighten oversight and re-focus levy spending on frontline services. That is the correct direction, and the test of it will be ordinary and exacting: ring-fencing, named ownership, indicators, independent evaluation, applied from the first dollar of the next phase rather than narrated after the last. The lesson Queensland has now paid to learn is not that mental health funding is hard to govern. It is that governance is not a feature added to spending once the spending is comfortable. It is the condition on which the spending was ever entitled to be called a fund. Build it first, or do not make the promise.