Every bubble looks like progress while it inflates. Many investors have been wondering whether there is an AI bubble or a Space bubble in the stock market. However, the West’s longest existing boom has been in people who live inside their head for a career – and four pressures are now closing on it at once.
My grandfather was a civil engineer. He built things for a living – first in wartime England, then in Tasmania – and in the whole of his working life there was no such creature as a middle manager. There were engineers, and there were the men who dug or put together, and above them all a thin layer of owners, whether public or private. The thick, heavy slab of individuals whose job it is to manage other people who in turn manage other people had not so far been invented.
My parents went to university in Australia in the late 1960s, when – as my former boss keeps reminding me – going to university was a faintly eccentric thing to do; relatively few did at the time. Within two generations both oddities – the middle manager and the graduate – became unremarkable. We treated them as progress. Indeed, even today it is hard to think of them as otherwise… They do, in fact, form a bubble which will come to represent this era in time, and the relative decline of these two unremarkable elements of the modern age is precisely what should worry us.
For 50 years, the great growth industry of the Western world has not been mining, energy, or manufacturing, or even finance. It has been the production of people who work with symbols rather than things: the graduate, the manager, the official. Call it the symbolic economy – the side of the modern workforce which processes information, administers, credentials, advises, reports.
It expanded on an unexamined assumption: that cognitive labour would stay scarce, and so valuable, indefinitely. On that assumption an entire civilisation rebuilt its idea of the good life. Get the degree, join the firm, climb the ladder, and if all else failed there was always the security of the expanding state. The trouble with that assumption is the trouble with attitudes behind every bubble. It was true until it wasn’t.
Four pressures are now converging on the symbolic economy, and it is their convergence, not any one of them, that makes a crash rather than a correction.
The first is overproduction. Tony Blair gave the industry its final impetus with his pledge to send half of all young people to university, but the instinct was universal across the West: the degree was sold as a lounge pass to the symbolic economy, and governments facilitated the printing of these passes however possible.
The problem is arithmetical. A credential that confers status only works if not everyone holds it; a credential everyone holds is merely the new minimum, and confers nothing. We have spent 30 years manufacturing aspirants faster than the economy manufactures the chairs they were promised.
Peter Turchin calls the result elite overproduction, and the bill is arriving.
Undergraduates in America told a 2026 survey that they expected to earn $80,000 in their first year out of university; the jobs on offer paid more like $56,000 – a shortfall of 30 per cent. Nor does the disappointment fade with experience; it compounds. The same students expected roughly $145,000 a decade on, against a real mid-career figure nearer $95,000.
Worse, more than two in five recent graduates are now working in jobs which never required a degree in the first place. It is like buying a lounge pass for the airport only to discover when you arrive that the lounge is already full. ‘You’ll have to join the waitlist.’ So you end up back at the restaurants and bars everyone else is using.
The second pressure is the one few examine, because it hides inside the word ‘career’. A management career is a promise resting on the logic of geometry. A pyramid can only lift everyone toward its apex if you keep building new pyramids – new firms, new divisions, new layers needing new supervisors. For decades growth obligingly supplied them.
Now, the movement has reversed. Organisations merge, flatten, delayer, consolidate; the apex slots disappear even as the queue of the qualified lengthens behind them. The stalled manager is not unlucky. Like some of our family friends, he is standing at the wrong position in a shape that has stopped expanding outward and started folding in. There is no promotion at the top of a stable pyramid, let alone one that is shrinking.
Indeed the data has started to arrive: an American study of 1.3 million mid-career professionals, published in May, found a quarter hit a ceiling before their peak earning years – and that the single most stall-prone field of all was public administration, the credentialed bureaucracy itself, with nowhere left to promote anyone.
The third pressure is political, and it is the oldest. To one set of parties the public wing of the symbolic economy is not a workforce but a swamp – a blob, a self-interested administrative class that votes for its own expansion and calls it service. The conviction that government is too big is now shared by new forces such as MAGA, Reform, and One Nation, as well as by a substantial part of the legacy right.
They have tried to act on it before, and mostly failed. Howard trimmed and the numbers crept back. In Washington DC, Trump and Musk launched the most aggressive assault on the bureaucracy in living memory. One Nation talks about it here. What earlier efforts lacked, every time, was an argument that wasn’t taken as just ideological – because ideology, when in government, invites a fight.
The fourth pressure is the new one, and it changes the calculus for the other three. The impact of artificial intelligence should still take time, but its real significance is already here: it does two things other factors don’t. It works on the need for human mental processing, rather than on the supply of workers. And it hands the politician a new argument.
Washington has just run the experiment for us. The Department of Government Efficiency (DOGE) delivered one of the largest peacetime cuts to the federal workforce in decades – something close to a tenth of the administration, gone inside a year. Yet federal spending kept rising – revealing almost no break where the cutting began.
The reasons are twofold. Salaries are no more than a rounding error compared with transfers, pensions and, increasingly, government interest. And across the agencies managers quietly rehired, because the work was still sitting there when the workers had gone. That is the lesson the next reforming government will absorb. You cannot shrink the state by removing people while leaving their work behind; the work simply summons the people back.
Artificial intelligence is the variable that breaks the pattern. If the task itself can be done by a machine, as in the era of Victorian industrialisation, then the specific task stays cut.
Which brings me to the second, sharper claim. The crash will fall across the whole symbolic economy – the graduate without the salary, the manager without the promotion – but it will fall hardest on the public servant. The official sits at the purest node of the bubble: the part least disciplined by any market, most swollen by debt, most resented by a major part of the political spectrum, and most exposed to machines, because rules-based processing of information is exactly the work artificial intelligence learns first.
Consider the fiscal backdrop. In Victoria, employee costs now consume more than a third of the Victorian budget, against roughly five per cent federally. State debt has risen more than seven times in a decade. Federally, the public service has just reached a record size – and it’s still growing.
Of course we can argue that cutting officials is unlikely to balance the books. The latest exercise in Washington suggests as much. The debt is driven by spending other than the salaries of clerks. But debt need not fix the accounting to offer permission. It is the licence – and artificial intelligence is the alibi.
The next minister from One Nation will not stand up and say he or she thinks the bureaucracy is hopeless or interfering, which is an opinion others will surely dispute – not least within the bureaucracy itself. She will simply argue that technology can now do the work, that the private sector has already started pursuing similar opportunities, that to refuse is not prudence but protectionism. ‘What is the argument to offer special protection to the intellectual worker?’ she might ask. ‘Are they more deserving than the manual labourer?’
That question, dressed as an embrace of modernity rather than a denial of the task itself, is very hard to answer in the traditional vocabulary. Would we then see Luddism in the government – among the very people who used to laugh at the Luddites in the textile industry two centuries ago?
Of course, the AI machines may not be as good as the sales pitch; the competence of artificial intelligence is still a wager, not yet entirely a demonstrated fact. But in this argument the weapon confers legitimacy rather than assures the result. Indeed, the public service has buried every reformer who has ever marched on it.
However, the thesis is that this time is indeed different, because for the first time in the modern age the actual demand for human mental work is slowing down.
Nothing grows forever, and the symbolic economy that was sold to us as the shape of the future was in truth the shape of a single 50-year expansion that is now passing its apotheosis. When an expansion so large reprices, it risks creating something that is politically combustible: a large, credentialed, downwardly mobile class promised status as well as comfort and then handed an ordinary stall. They were offered a lounge – but ended up sitting like everybody else on the hard seats in the main concourse.
History has a settled view of what disappointed, overproduced elites do next. They do not go quietly. They go looking for a new banner. This has two ironic results. The first is that the conservative party reaches for its sharpest intellectuals: Turnbull here, Sunak in Britain. Is Taylor any different?
The second, deeper irony belongs to the side which built this bubble in the first place: the centre-left. Labor is the flagship party of the symbolic economy – the graduate, the city professional, the official are at once its base and its self-portrait. They rejected the worker for the intellectual; so much that the worker has had to dress up as an intellectual to feel acceptable. So the crash does not merely cost the party votes, it dissolves the world that made the Labor party what it currently is.
The displaced graduate, who is quite certain he or she has nothing in common with the displaced tradesman, is going to discover they are standing in the same queue.
None of this is only an Australian story, though Australia is a good place to read it in real-time. It is Blair’s graduates in Britain, the swamp in Washington, the swollen ministries of Paris and Berlin. The symbolic economy was the proudest output of the modern Western age following the loom and the railway. We are about to learn what happens as it deflates.
