Victorian Labor’s collapse is not only about debt, crime or infrastructure blowouts, but a delayed reckoning from communities that felt broken by lockdowns and ignored by the recovery.

They broke us, our kids’ mental health, and went back to their high vis vest.

That is how one participant in a recent RedBridge focus group described the Victorian government. She was not asked about the pandemic. She was asked about the state she lives in.

It is worth sitting with the sentence, because it contains an entire theory of what went wrong. There is her injury and there is a second injury done to her children. A set of injuries that are a manifestation of grief turning into resentment, which is the return to normal by a government that decided the recovery would be built out of concrete and not by wrapping their arms around a bruised community.

Melbourne holds a record no city wanted. Two hundred and sixty-two days of lockdown across six separate closures, the longest cumulative total in the world. That number has been argued over for five years, almost always as a question of whether it was justified. That is the wrong argument because the lockdowns are over. What matters now is what the state did with the people it had asked to endure them, and it’s within this world where we can clearly see why the Victorian Labor government is firmly up against the ropes, in round two against Mike Tyson in his prime.

It’s up against the ropes because of what it chose to build versus rebuild and what it selected to ignore. What it did? Level crossings, tunnels, rail loops, road duplications. A pipeline measured in the tens of billions and a workforce in high-visibility clothing. It was legible, it was announceable, and it was the thing the government already knew how to do. What it was not, was a response to what had actually happened to Victorians. The rebuilding of the social fabric of the state.

Almost every explanation offered for this Victorian Labor government’s predicament has been an explanation of the obvious. Debt. Crime. The Suburban Rail Loop. The Commonwealth Games. The CFMEU. Each of these is real and each of them is a symptom. The disease is older, the polling shows when it presented, and the demographic breakdowns show who is carrying it straight to the ballot box in November, with determination and anger.

When it broke

Labor did not begin bleeding after the 2022 election. It held the line through all of 2023 and then lost almost everything in a single year.

In December 2023, 13 months into the term and two months after Daniel Andrews resigned, RedBridge had Victorian Labor on 37 per cent, roughly its election result. There is no decay in that data. Whatever Victorians felt about the pandemic in 2023, it had not yet reached vote intention and there is a clear reason why (more of that later).

Then suddenly everything changed. Thirty-six in March 2024, 35 in June, 31 by August, 30 by spring. It never returns to the mid-thirties. In June 2026 it sits at 26, level with the Coalition on 26 and behind One Nation on 27, in a sample of more than 5,500 Victorians. This week, it has collapsed to 23, a historic low.

The two-party numbers underneath are worse than the primaries suggest. The Coalition leads Labor 54 to 46. Labor leads One Nation by only 52 to 48. A party that governed Victoria for 11 of the last 12 years is now within four points of losing a two-party contest to Pauline Hanson, almost running third on primaries and, on a very bad November day, could end up with as little as 20 MPs in the lower house.

Why 2024, and not 2022

Two things happened in the window, and neither of them is a scandal.

The first is that Daniel Andrews left. This is not a point about competence and it is not a point about his successor. It is a point about psychology. Andrews had been the single human container for everything Victorians felt about the pandemic, the loyalty and the rage in equal measure. While he remained, the feeling stayed attached to a person and to a period. When he went, it detached and went looking for the institution. He was not holding up the vote. He was holding down the reckoning, and his departure kicked off a personal reckoning by millions of Victorians.

The second is when the bill arrived. These lockdowns were clearly very expensive and in the 2023 budget the state introduced the Covid Debt Repayment Plan, a 10-year levy running to 2033. The payroll component started on 1 July 2023. The land component started on 1 January 2024, and it did so by cutting the land tax threshold from $300,000 to $50,000. More than 328,000 Victorians received a land tax assessment who had never received one before. You do not need to be a psephologist to work out the electoral impact of this surprise bill.

Note what happened here. It is not a tax with a neutral name. It is an invoice, addressed personally, with the word Covid printed on the memo, arriving in letterboxes or email, through the first half of 2024, payable until 2033. It is the moment the state converted a shared emergency into an individual debt.

That is precisely the trigger the disaster recovery literature identifies. Collective trauma does not move in a straight line. It moves through an emergency phase, then a phase of solidarity in which people rally around the response or the flag, and then a phase of disillusionment set off by the withdrawal of support and by the realisation that the damage is not temporary. Disillusionment is not caused by the disaster. It is caused by what comes after it. Kai Erikson’s distinction matters here too. Aftermaths with a human agent behind them run longer and angrier than natural ones, because there is someone to hold responsible. A lockdown is a policy.

On that reading, November 2022 was not the reckoning. It was the tail of the solidarity phase, and Labor’s read of that phase was as if it was the verdict. It was a catastrophic error. The reckoning was always going to arrive during the disillusionment and eventually anger phase, at around 2024 to 2025.

Who is carrying it

The June 2026 breakdowns settle the question of what kind of divide this is, and the answer is not the one usually offered.

It is not fundamentally geographic. Labor polls 29 in inner metropolitan Melbourne and 28 in the outer suburbs. On its own vote, the gap is a rounding error.

It is educational, and the gradient is enormous.

Among Victorians who did not complete year 12, Labor polls 20 and One Nation polls 40. Among Victorians with a university degree, Labor polls 31 and One Nation 17. That is a 23-point swing on each party across a single variable, and it is the widest cleavage anywhere in the survey.

Education is doing a job here that has nothing to do with schooling. It is the cleanest available proxy for whether your job could be done from a spare room in 2020. Degree holders worked from home on full pay. People with trade qualifications and people without year 12 either kept turning up in person or lost the income entirely. Five years later, those two groups are voting as if they lived through different events, because they did.

The generational pattern points the same way. Gen X and, to a certain extent, older Millennials – the cohort that was parenting school-aged children through two years of remote learning – sits at 23 for Labor and 33 for One Nation, the worst combination of any generation under 60. Younger Millennials and Gen Z, who were mostly not managing a locked-down household with children in it, sit at 30 for Labor.

What geography actually shows

Geography still matters, but not in the way the seat-by-seat commentary assumes. The difference between inner and outer Melbourne is not the size of Labor’s vote. It is where the anti-Labor vote has gone.

In inner Melbourne the alternative to Labor is still the Coalition on 29 and the Greens on 15. One Nation is on 20. The two-party system is intact. In the outer suburbs, One Nation is on 28, ahead of the Coalition on 24. In rural Victoria it is on 34.

So, the inner suburbs are having an argument about who should govern. The outer suburbs and the regions are having a different argument about whether the people who govern are worth anything at all. That is the geographic finding, and it is a more serious one than a swing.

There were two lockdowns, not one

The central analytical error, in government and in most of the commentary since, is treating the lockdown as a single shared experience. It was not. It divided along a line that ran straight through the metropolitan map, and the line was whether your job could be done from a spare room.

In the inner suburbs, lockdown was a salary that kept arriving, a mortgage that got cheaper to service, and an asset that appreciated while you were confined to it. It was tedious and it was lonely. It was not, for most, financially destabilising. It also in part explains why this part of Melbourne has continued to enjoy some level of personal wealth stability as the country slammed into an inflationary crisis.

Further out (with an exception of inner-urban public housing towers), lockdown was the closure of the income itself. It was also enforcement. Curfews, the five-kilometre radius, postcode-based restrictions, and the hard lockdown of nine public housing towers that the Victorian Ombudsman later found breached the Charter of Human Rights and Responsibilities. That is the difference between experiencing the state as an inconvenience and experiencing it as a force.

These are also the households where remote schooling failed worst. More children, fewer rooms, fewer devices, and parents who were not home to supervise because their work was classified as essential. The learning loss and the mental health cost did not fall evenly. They fell where the housing was smallest and the families largest. The Gen X and older Millennial number in the table above is what that looks like five years later.

The exception that has to be explained

There is one finding in the June survey that cuts against the argument as it is usually told, and it should be dealt with directly rather than left for someone else to raise.

Migrant Victoria is where Labor is strongest. Among Victorians born overseas, Labor polls 35 and One Nation 20. Among the Australian-born, Labor polls 24 and One Nation 29. In households where a language other than English is spoken, Labor polls 32 and One Nation 22. On the raw numbers, diverse Victoria is not leading the revolt. It is the last thing holding the government up.

That is awkward, because these are precisely the communities that could not work from home, that carried the highest rates of self-employment, that were policed hardest, and that were locked down by postcode. Every measure of exposure says they should be the angriest. The vote says they are the most loyal.

The explanation is not contentment. It is that the exit is blocked by One Nation. One Nation is the vehicle carrying the disillusioned vote in Victoria, and One Nation is not available to a Turkish or Vietnamese or Indian small business owner in Melbourne’s north, whatever they think of the government that closed their shop. The grievance is real and it has nowhere to go. A blocked exit is not loyalty. It’s a red flag that it can be a lot worse if some other party opens that exit.

How hard the votes are

The softness data tells you which way this is travelling.

Among voters who describe their choice as solid, One Nation polls 33 and Labor 24. Among those who describe it as very soft, One Nation polls 17 and Labor 30. Labor’s vote is disproportionately made of people who are not sure. One Nation’s is disproportionately made of people who are.

That is the wrong way around for a government seeking a fourth term and the wrong way around for anyone treating the One Nation number as a protest that will evaporate in the booth.

How little attention was being paid

There is a detail from the 2022 campaign that captures the level of attention these communities were receiving. Greenvale, a seat in the north-west that also contains one of the poorest suburbs in Victoria, Meadow Heights, was allocated $60,000. Not $60 million. Sixty thousand. It is understood the allocation was a clerical error and the money had been intended for a different electorate.

Whether or not the error was ever corrected is beside the point. A party that had just presided over the longest lockdown in the world went to an election having budgeted, by accident, $60,000 for one of the suburbs that had absorbed the most of it. That is not malice; it is something worse politically, which is the absence of any thought, or empathy.

The vest becomes the symbol

This is where the CFMEU enters, and it enters symbolically rather than causally. The government did not know what would later be alleged about the union’s Victorian branch. That is not the argument.

The argument is one of political economy. The recovery was routed through a single sector. That sector’s workforce was among the few that never had to stay home. Its governance then collapsed in public view, with allegations serious enough to place the branch into administration.

For a small business owner in Broadmeadows or St Albans who lost two years of trading and watched their children fall behind, the high-visibility vest stopped being workwear. It became the emblem of who the state decided to rebuild around, and of who it did not.

The scandals of the last two years all attach to the concrete. The blowouts are Big Build blowouts. The debt is infrastructure debt. The union at the centre of the allegations is the construction union. They are not a rival explanation to the sentence in that focus group I mention at the start but rather, they are its documentation.

Trauma does not swing. It exits

If this were ordinary incumbency fatigue, Labor’s losses would be sitting with the Coalition. They are not. The Coalition is on 30 in our latest AFR poll, ahead of Labor and One Nation, but the lead can be halved within the margin of error. The Coalition is still struggling in the outer suburbs of Melbourne and hence, 11 years out of office, facing a government in this condition, and barely beating it, suggests something else.

That is the behavioural signature of the argument. People who believe the state broke them do not transfer their loyalty to the alternative manager of the state. A swing voter is someone choosing between two governments. Someone who has concluded that government itself did this to them is not swinging. They are leaving. They in part already did this in 2022 but back at that state election, the vote did not have a central focus; instead it sprayed across numerous other minor party candidates and independents.

It is the same movement visible federally, with the same internal structure. A founding cohort with long-standing grievances, and a much larger late-arriving cohort that has been pushed rather than pulled. In Victoria, the pushing was done between March 2020 and October 2021, and the invoice for it arrived in January 2024.

What was actually on the table

The choice in 2022 was never build or do not build. Infrastructure was needed and much of it will outlast the argument against it.

The choice was what the recovery was for. A state that had just run the longest lockdown in the world had, at that moment, an unusually clear mandate to do something unprecedented. Mental health care and support, at the scale of the damage rather than just appearing on someone’s ministerial title. Catch-up tutoring held in place for a decade rather than a term. Debt relief for the businesses that were closed by order. Something addressed to the people who had carried the cost.

It went to concrete instead but as I have previously written about, the injury compounded by the reality that that concrete was selectively poured. Five years on, the person in that focus group can still describe the whole thing in a single sentence, which is the surest sign that nothing has been resolved. She was not asked about the pandemic. She was asked about the state, and she answered about the pandemic, because for her they have never been separate things.

The views expressed in this article may or may not reflect those of Pearls and Irritations.