Andrew Paul Klein

I. The numbers

In the 2024–25 financial year, 4,812 older Australians died while waiting for home care funding they had already been approved to receive. That is roughly thirteen people a day, every day, for a year.

In the 2025–26 financial year, the figure rose to 6,436. An average of seventeen a day. An increase of almost 34 per cent on the year before.

These are not estimates. They are departmental figures, published by the health department and reported by the ABC in October 2026, after the government had failed to comply with two Senate orders to produce them. The Shadow Health Minister, Anne Ruston, said what the families already knew: “Everybody on the wait list is a person; they’re not a statistic or a number”.

But they are a statistic. That is precisely the problem.

The queue itself was almost 154,000 people — not the 100,191 the government reports, but the figure you get when you calculate the waitlist the way it was calculated before November 2025, when the method was changed and people receiving interim funding were removed from the count. The median wait between applying for aged care and receiving it is ten months. Two hundred and ninety-seven days.

And the deaths are not new. In February 2021, a senator rose in the chamber and said: “Twenty-eight thousand people died waiting for a home-care package. That is a great shame and a great stain on the social compact that people believe they have with the government that they’ve paid taxes to for all of their working life”. By 2022, the figure had passed 50,000 since July 2017. The man who would become the Minister for Health and Ageing called it “a national disgrace”.

He was right then. The question this essay asks is why being right then did not stop it happening now.

II. The chain

A death becomes a statistic through a sequence of decisions, each of which is defensible in isolation and none of which is made by anyone who intends the death. The aged care chain is the best documented, so let us trace it.

The settlement

The Howard government did not set out to create a system in which people die waiting. It set out to govern well, and it defined governing well as fiscal management. The retrospective on the Howard years puts it precisely: “competence equals virtue” became the governing moral claim, and “responsible financial management” became “an almost spiritual category of debt repaid, AAA restored, the Future Fund created, taxes reformed”.

The consequence was not cruelty. It was a frame. Structural outcomes became “weather” — things that happen to a country rather than things a government does. A drought is not a policy failure. A waitlist is a system pressure. Once outcomes are weather, nobody is responsible for them, and the question “is this working?” stops being a question anyone is required to answer.

The capacity

The public service that would have monitored the waitlist, modelled the demand, and told the Minister the truth about the trajectory was being hollowed out at the same time, for reasons that had nothing to do with aged care.

Capability reviews — the instrument by which departments examined their own weaknesses — were discontinued in 2016. The Thodey Review of 2019, described as the most substantive review of the APS since 1976, found the service “not performing at its best” and “not ready for the big changes and challenges that Australia will face between now and 2030.” Its “greatest concern” was “the hollowing out of strategic policy skills,” and it identified the cause directly: the use of consultants “to perform work that has previously been core in-house capability”.

John Quiggin’s explanation is the one that matters. “Much of the knowledge required for policy management is tacit and shared by the professional and vocational employees who actually do the work.” When the work is outsourced, the knowledge goes with it. The department retains the contract and loses the capacity to evaluate what the contract delivers.

By 2021, the Senate’s APS Inc report found Australia had “arguably the heaviest reliance on consultants for advisory and management work” across all Anglophone countries. The machinery that could have seen the aged care crisis coming had been sold for parts.

The instrument

Into this vacuum came the algorithm.

The Integrated Assessment Tool — the questionnaire completed by a clinician — was recommended by the Royal Commission, tested on 22,000 people, developed through three clinical working groups, and published with a trial report. It is a legitimate instrument.

The algorithm that scores it was never recommended by the Royal Commission. It had no live trial. The department could not name the clinicians who signed off on it, and refused a formal Order for the Production of Documents. A prototype, believed to have been designed by a consultancy, was taken in-house, modified and validated by public servants, and launched on 1 November 2025.

Professor Kathy Eagar, who designed the classification system used to fund residential aged care, reverse-engineered it. Her findings: the vast majority of the more than 500 questions do not feed into the algorithm; the clinical notes recorded in more than 100 text boxes cannot be read by it; frailty, cognitive problems, mental health and carer fatigue enter “far too late” to affect funding. A one-point difference at a particular juncture is worth almost $40,000 a year.

The assessor records. The algorithm decides. The record — the observation, the person’s own account of their life, the carer’s exhaustion — is collected and then not used.

This is the travertine mould. The wood decays; the calcium carbonate keeps its shape. The shape is accurate. It is not the thing. And the apparatus reports the cavity with great precision, and the precision is read as truth.

The response

When the ABC’s Four Corners reported the deaths in August 2026, the Minister was asked about it on radio. “We know that, unfortunately, for as long as this system has been in place, tragically, people will die on a wait list for surgery, for home care, for entry to an aged care facility,” Mark Butler said. “Every death while waiting for any procedure, whether it’s an aged care procedure or a hospital procedure, is a tragedy and a very strong signal we’ve got to do better”.

It is a compassionate statement. It is also a statement that describes the deaths as a feature of the system rather than an output of it — as though the waitlist were a natural phenomenon, like weather, and the government’s role were to lament it rather than to end it.

And the contrast is stark. In 2022, as Shadow Minister, Butler said of the same deaths: “This is a national disgrace“. In 2025, as Minister, he delayed the release of 83,000 home care packages by four months, linking the release to the delayed commencement of the Aged Care Act. Providers had told a Senate inquiry they were ready to deliver from 1 July. Departmental officials told the same inquiry there were no technical reasons the packages could not be rolled out. The only thing between 20,000 older Australians and the help they urgently needed, one report noted, was a nod from the Minister.

The nod took five days. In that time, the government talked nonsense about providers wanting the delay, and was shellacked by crossbenchers until it moved.

This is not the conduct of people who do not care. It is the conduct of a system in which caring is not the variable that determines the decision.

III. The mechanism

Why does this happen? Not because the people are bad. Because the incentives reward something other than outcomes.

The political science literature calls it blame avoidance. The foundational work — Hood’s The Blame Game, Weaver’s studies of the politics of blame — describes officeholders structuring institutions to minimise the traceability of failure. Delegation to arm’s-length agencies, complexity, multiple veto points, the diffusion of responsibility: these are not accidents. They are rational responses to the incentive structure of democratic accountability.

The problem is that a system optimised for the avoidance of blame is not optimised for the production of outcomes. It is optimised for the production of defensibility. And defensibility and care are not the same thing.

The social psychology literature names the second mechanism. Albert Bandura’s theory of moral disengagement describes the eight ways individuals and institutions distance themselves from the human consequences of their decisions: moral justification, euphemistic labelling, advantageous comparison, displacement of responsibility, diffusion of responsibility, disregarding or distorting the consequences, dehumanisation, and attribution of blame.

Steve Davies, who has applied the framework to the Australian Public Service, puts it bluntly: “It is absolutely beyond doubt that moral disengagement has literally infected the Australian Public Service. Over the past decade that infection has been normalised and escalated”. Davies also draws on Chris Argyris’s work on organisational defensiveness — the routines organisations develop that prevent them from learning about their own failures.

This is the answer to the question of why the system will not examine its own failures. It has routines that prevent it. The capability reviews were discontinued. The Thodey recommendations were paused. The consultants were hired to do the thinking. The algorithm cannot read the notes. The appeal is heard by the department that built the system. Each of these is a defensiveness routine, and together they form a closed loop.

The institutional theory literature names the third mechanism. Meyer and Rowan’s concept of decoupling describes how organisations adopt formal structures that legitimise them while their actual operations remain unchanged. Brunsson’s term is organised hypocrisy: the formal commitment serves the legitimacy function, and the operational core continues as before.

This is the best frame for the gap between the Aged Care Act’s Statement of Rights and the algorithm that cannot read clinical notes. The Act is real. The rights are real. The Statement of Rights is a formal structure that legitimises the system. And the operational core — the scoring, the queue, the delay — continues as before.

The rights are not directly enforceable. The Commission notes this. The Act establishes a framework in which rights exist and cannot be acted on. That is not a failure of implementation. It is the design.

IV. The disciplines

There is a fourth mechanism, and it is the one that closes the loop across generations.

The disciplines that train people to see structural harm — sociology, politics, social policy, gender studies — have been defunded. Not by a decision to suppress them, but by a funding model that made their survival a matter of entrepreneurial competition.

The Job Ready Graduates package of 2020 increased student contributions in sociology, politics, history, philosophy, communications and journalism by 117 per cent, from $6,684 to $14,500, and now $16,992. The Innovative Research Universities analysis estimated a reduction in university base funding of $813 million in 2024. Low-SES commencements in the humanities and social sciences declined by 19.7 per cent. The Universities Accord concluded that JRG had “failed” and needed “urgent remediation” — and two years later the sector was still waiting for clarity.

The consequence is not that sociology disappeared. It is that the pipeline of people trained to analyse institutional failure narrowed. The Australian Academy of the Humanities describes “decades of cuts to humanities and creative arts programs, enrolments, research support and staffing” eroding “our national capabilities“. The disciplines that teach students “to understand power, inequality, and the ways that our institutions function and respond to social change” are the ones being asked to justify their economic value in a market that has already decided what counts.

A system that cannot train people to see structural harm will not see structural harm. And a system that has defunded the disciplines that produce that capacity has arranged, without deciding, that the next generation of public servants will be less able to evaluate the system they inherit.

V. The loop

Here is the synthesis, offered as inference rather than established fact.

The system is not indifferent. Indifference is the output of a structure that has been optimised for other things.

1. The Howard settlement defined competence as fiscal management, which made structural outcomes into weather — nobody’s responsibility.

2. The APS shed the capacity to analyse its own failures — capability reviews discontinued, tacit knowledge outsourced, strategic policy hollowed out.

3. The political competition became a contest over image, because the economic frame was shared and only the presentation was contestable.

4. The disciplines that train people to see structural harm were defunded, not by a decision to suppress them but by a funding model that made their survival a matter of entrepreneurial competition.

5. The measurement instruments that would make deaths accountable were never built — no national domestic violence death data framework, no funding mapping, a waitlist whose calculation method changed.

6. The result is a system in which harm is produced without perpetrators, because each function that could attribute responsibility has been either outsourced, discontinued, or rendered unfalsifiable.

The loop closes because each element reinforces the others. A government that cannot measure outcomes cannot be blamed for them. An opposition that shares the economic frame cannot offer an alternative. A public service that has lost the capacity to evaluate cannot tell either of them that they are wrong. And a defunded sociology cannot train the next generation to point it out.

This is why the deaths do not produce accountability. Not because anyone decided they didn’t matter, but because the apparatus that would have to register their mattering has been dismantled — and dismantling it required no conspiracy, only a series of defensible decisions made by ordinary people doing their jobs.

VI. What it says about a society

Now the question the morning was reaching for.

When deaths become a statistic, something has already happened before the first death is counted. A society has built an instrument that can register the death and cannot register the loss. It has a number and not a name. It has a waitlist and not a person. It has a funding level and not a life.

The statistic is not a failure of compassion. It is a success of the apparatus. The apparatus was built to count, and it counts. It was not built to mourn, and it does not mourn. And because it cannot mourn, the question of whether it should stops being askable inside the system that produced the number.

This is what the anthropologists of bureaucracy have been telling us for a century, and what the aged care algorithm demonstrates in its purest form: the instrument does not merely describe the world; it decides what is describable. A death that can be counted is a death the system can respond to. A death that cannot be counted — because the person was not on the list, because the category did not exist, because the note was in a text box the formula could not read — is a death that, for the purposes of the apparatus, did not happen.

The statistic is therefore not neutral. It is a moral claim disguised as an arithmetic one. It says: this is the unit that matters. This is the granularity at which we are prepared to be responsible. And everything finer than that granularity — the person, the family, the particular grief — is below the resolution of the instrument, and therefore below the threshold of institutional concern.

A society that counts its dead in this way is not a society that does not care. It is a society that has arranged its caring at a scale where caring does not require action. It can mourn at the level of the aggregate, because the aggregate does not have a face. It can express regret at the level of the trend, because the trend does not have a name. And it can continue, because the instrument reports that the number is within the range of what the system can bear.

That is what deaths becoming statistics says about a society. Not that it is cruel. That it has built a way of knowing that is structurally incapable of registering what it has done.

VII. The question of government

So what is the role of government in Australia in 2026?

The evidence of this essay suggests that it is currently three things, and none of them is what the word “government” is supposed to mean.

It is a purchaser — it buys services, sets targets, and measures outcomes.

It is a data collector — it gathers personal information at a scale and granularity that would have been unimaginable a generation ago.

It is a reputational manager — it spends money on image, communication and brand, and it has been documented doing so by its own auditors.

It is not, increasingly, the thing that knows what a person needs. That function has been displaced to the algorithm, to the provider, to the market, to the consultant. And because it has been displaced, the government can no longer be blamed for what it does not know — which is, of course, the point.

The question is whether this is what a government should be. And the answer is not a matter of ideology. It is a matter of whether a democracy can survive a system in which no one is responsible for the outcomes that matter most.

A government that is reactive and self-limiting — that responds to crises it did not foresee and cannot explain, and that limits its own capacity to foresee and explain — is not a small government. It is a blind government. It is not exercising restraint. It is exercising incapacity, and calling it restraint.

A proactive government would do four things the present system does not.

It would measure what matters, not what is easy — which means qualitative intake, participatory standing, and correction by the person measured, at the cost of the speed and control the procurement model exists to maximise.

It would retain the capacity to evaluate — which means the capability reviews come back, the consultants are used for advice and not for thinking, and the tacit knowledge stays inside the institution that is accountable for it.

It would build the instruments that make deaths accountable — a national domestic violence death data framework, a funding map that tracks where the money goes, a waitlist calculated in a way that cannot be changed without a public explanation.

And it would fund the disciplines that train people to see what it cannot — not because sociology is economically productive, but because a democracy that cannot analyse its own institutions is a democracy that cannot correct itself.

None of this is expensive. All of it costs speed and control. That is why it does not happen.

VIII. The ledger and the name

The excavators at Abric Romaní know what they are holding. A cavity, honestly described, its limits stated. They do not present the shape as the thing. They do not report thirteen wooden objects and call it the whole of Neanderthal life. They say: this is what survived, and here is what a mould cannot show.

That is the standard. Not perfection. Honesty about what the instrument can and cannot register, and a refusal to let the cavity stand in for the thing that decayed.

The aged care system has 4,812 deaths in one year and 6,436 in the next. It has a waitlist of 154,000 people. It has an algorithm that cannot read the notes that describe what a person needs. It has a Minister who calls each death a tragedy and a system that processes them as a trend.

The statistic is not the problem. The statistic is the symptom. The problem is that the apparatus has been built in a way that makes the statistic the only thing it can see — and that the apparatus has been built that way by people who were doing their jobs.

The next time you are handed a figure about a person — a score, a rating, a funding level, a risk assessment, a waitlist position — ask two questions.

What did the instrument measure?

And what could it not see?

Then ask the third question, which is the one that matters.

Who decided that what it could not see was not important?

Andrew Paul Klein