Melbourne's homelessness

On a winter morning outside Flinders Street Station, a man folds the sleeping bag he has used through four months of cold nights and joins a queue at a nearby outreach van for a hot breakfast. He is one of 78 people the City of Melbourne's homelessness services could name, locate and count sleeping rough within the municipality on a single night in March 2024, a number outreach workers now track week to week, because a decade of policy failure has taught them that a program cannot end what it does not count (Launch Housing, Melbourne Zero By-Name List, 1-Mar-2024).

That count captures only the sharpest edge of a much larger problem. Across Victoria, more than 30,000 people were recorded as homeless on Census night in August 2021, up from around 24,800 five years earlier, a rise of roughly a quarter in half a decade and the steepest increase of any Australian state (approx.; ABS, Estimating Homelessness: Census, 2021, released Mar-2023). Most of that homelessness is invisible: people in severely crowded homes, couch-surfing with relatives, or cycling through boarding houses and short-term crisis accommodation, not sleeping on the street. But it is the same underlying gap between housing need and housing supply that produces both the family doubled up in a relative's lounge room and the man outside Flinders Street Station — and it is the gap the Allan government's homelessness programs, and the housing build behind them, were built to close.

What the government did

The centrepiece of Victoria's response to rough sleeping is From Homelessness to a Home, a program born out of the pandemic. In 2020, as the state entered its first lockdowns, the government moved around 2,000 rough sleepers out of parks, doorways and crisis shelters into hotel rooms to isolate safely — an emergency measure that proved rough sleepers could be housed quickly once government was willing to pay for it (approx.; Premier of Victoria, media release, 2020). It then became a standing $150 million initiative built on the "Housing First" model: moving people directly into stable accommodation with wraparound mental health, drug and alcohol, and case-management support, rather than requiring "housing readiness" first. The program targeted 1,845 households and had housed close to 1,600 within its first two years (approx.; Homes Victoria; Department of Families, Fairness and Housing, program reporting).

That program sits inside a much larger one. The Big Housing Build, launched November 2020 with an original $5.3 billion commitment, is Australia's largest state social-housing program, since grown to more than $8 billion committed, with over 10,800 homes completed or under construction by August 2025 (Homes Victoria, Big Housing Build, progress update, Aug-2025). Inside it, the government earmarked 2,000 units for people living with mental illness and 1,000 for victim-survivors of family violence, two cohorts over-represented in the state's homelessness caseload (Homes Victoria, Big Housing Build community housing partnerships). A separate, older strand is the High-Rise Redevelopment Project: demolishing and rebuilding all 44 of Melbourne's public housing towers, built in the 1950s-70s, which today house around 10,000 residents (approx.; Homes Victoria; Parliament of Victoria, Inquiry into the Redevelopment of Melbourne's Public Housing Towers). The government's case: rebuilding at higher density will eventually house roughly three times as many people on the same sites, adding net social housing even as construction displaces today's tenants.

The evidence

The clearest sign of how concentrated this pressure has become sits in the specialist homelessness service system, the network of crisis, transitional and support services people turn to when they lose housing or fear losing it. In 2024-25, the five local areas with the highest homelessness-service caseloads in the country were all in Victoria, four of them inside greater Melbourne — Melbourne West, Melbourne South East, Melbourne Inner and Melbourne Outer East, together over 40,000 clients in a single year (AIHW, Specialist Homelessness Services Annual Report 2024-25, client geography data, 2026). Victoria also recorded the highest number nationally of clients who were employed but still could not afford stable housing: 9,000 people, over a third of the national total (AIHW, Specialist Homelessness Services Annual Report 2024-25).

This is not simply demand outrunning a system that otherwise works well. An estimated 8,600 Victorians sleep rough for at least one night over a year, far more than any single-night count like the one outside Flinders Street Station can capture, because rough sleeping is often episodic rather than constant (approx.; AIHW estimate, cited in Victorian Auditor-General's Office, Victoria's Homelessness Response, tabled 17-Sep-2020). That 2020 audit examined the $45 million Homelessness and Rough Sleeping Action Plan, the program that preceded and fed into From Homelessness to a Home, and found the department could not demonstrate it was reducing rough sleeping at all: only 12 people had been housed through purpose-built modular units, just 22 per cent of allocated flexible brokerage funding had been spent, and only 22 per cent of clients had a completed case plan against a departmental target of 90 per cent (Victorian Auditor-General's Office, Victoria's Homelessness Response, tabled 17-Sep-2020). Five years later, the strain has not eased: specialist homelessness services nationally logged around 129,000 unmet requests for help in 2024-25, up from 110,000 the year before, an increase the AIHW attributes largely to worsening demand in Victoria, Queensland and Western Australia; six in ten were simply for a bed for the night, refused because none was available (AIHW, Specialist Homelessness Services Annual Report 2024-25).

Victoria's own auditor found the state's flagship rough-sleeping plan could not demonstrate it was reducing rough sleeping at all — and five years on, the services that replaced it are turning people away at a rising rate.

The critique

The clearest test of whether Housing First is a funded commitment or a slogan came in the government's 2022-23 budget, when From Homelessness to a Home's annual funding was cut by around $43 million, roughly 78 per cent of its previous allocation (approx.; Council to Homeless Persons, 3-May-2022, citing Victorian Budget Papers 2022-23). Council to Homeless Persons — the sector's peak advocacy body, a clearly labelled secondary source rather than an independent auditor — warned the cut would strip support from around 1,440 people still in the program and reverse the gains the government itself had been advertising. The government partially reversed course in 2023-24, allocating $134 million over four years back into homelessness support, and rolled the funding forward again in 2025-26 (Victorian Budget Papers, 2023-24 and 2025-26). But a program cut by three-quarters, restored under pressure, then renewed year to year rather than funded on a multi-year, indexed basis, is not one that tenants, caseworkers or the teams counting rough sleepers on the CBD's streets can plan around.

The larger structural problem is that even a fully funded Housing First program cannot succeed without somewhere to house people, and on that measure Victoria remains near the back of the field. Despite the Big Housing Build, Victoria's social housing stock sits at around 3.1 per cent of all housing, against a national average of about 4.5 per cent, and the 2025-26 budget made no new commitment to lift it, despite Infrastructure Victoria and a bipartisan parliamentary inquiry recommending at least 60,000 more social housing dwellings over the next decade (approx.; VCOSS budget analysis, citing AIHW and Productivity Commission housing-stock data, 2025-26). A state can run an effective program for the thousands it reaches each year and still fall further behind in aggregate, if the total stock of social and supportive housing does not grow faster than the caseload documented above.

The tower displacement effect

The clearest collision between the government's housing ambition and its homelessness record sits inside the tower program itself. The 44 towers due for demolition and rebuilding are not new housing stock in the way the rest of the Big Housing Build is; they are existing homes, and every resident moved out of one, whatever the eventual quality of the replacement, is a household placed at real risk of homelessness in the years between the old block coming down and the new one opening. A parliamentary inquiry found in 2025 that the government had withheld the bulk of its own cost-benefit analysis for demolition over renovation, producing only a fraction of the documents the Legislative Council had ordered it to release (Parliament of Victoria, Legal and Social Issues Committee, Inquiry into the Redevelopment of Melbourne's Public Housing Towers, tabled 2025). Independent research published January 2026 sharpened the question: a study of one Brunswick tower, where 97 per cent of residents are aged over 55, found renovating and adding new homes on the same site could cut embodied carbon emissions by up to 44.5 per cent compared with demolition and rebuilding, at broadly comparable cost — precisely the kind of site-by-site case the government has not published its own version of (university research, RMIT, Jan-2026, clearly labelled secondary source).

None of this means demolition is the wrong call on every site; some towers may genuinely be beyond economic repair. But a government running a Housing First program on the premise that stable housing is the precondition for everything else cannot then relocate thousands of its longest-standing and most vulnerable tenants, many of them elderly, without publishing the evidence for why the disruption is necessary. The same by-name lists that track rough sleepers in the CBD showed 612 people still active across Melbourne's expanded outreach areas by the end of 2024-25, including people in temporary and respite accommodation (approx.; Launch Housing, Advance to Zero data, 2024-25) — a reminder of how quickly people cycle between insecure housing and the street when stable-home supply does not keep pace, in either direction.