When Does the Housing Correction End

Everyone is asking when this ends. It is the wrong question to start with, because the answer depends on something almost nobody is tracking: where the money that left established housing actually went.

Get that wrong and every recovery forecast is built on a cohort that is not coming back in the way the model assumes.

So start there.

Ray White chief economist Nerida Conisbee published open home attendance data on 7 September showing something the price indices have not caught up with: premium markets stabilising while demand weakens across more affordable, investor-heavy ones.

Her figures, and they are hers. Ray White tracks around 13,000 open homes a week, which makes it one of the few genuinely timely reads on buyer activity in the country.

National attendance fell from around 4.5 people per open in January to close to 2.0 by July, and has since edged back to around 2.2. Sydney is averaging 2.3, roughly 0.2 above its pre-Budget level and up 0.31 over the past eight weeks. Melbourne is at 2.2, back to about where it sat before the Budget, up 0.17 over the same period. Brisbane, Adelaide and Perth remain well below pre-Budget levels.

Underneath the national number the split is sharper. The biggest gains since the Budget are in premium markets: Sydney's Eastern Suburbs up 0.8 attendees per open, with Melbourne Inner South, North Sydney and Hornsby, and Melbourne Inner East each up around 0.6. The biggest falls are in affordable ones: Adelaide South down 2.9, Perth North West down 2.4, Perth North East down 1.7, Cairns down 1.4.

Prices are starting to agree. Cotality has national values down 3.3 per cent over the three months to August, but in August itself Eastern Suburbs prices rose 1.1 per cent, North Sydney and Hornsby 1.0 per cent, and Ryde, the Inner West and the Northern Beaches around 0.8 per cent. Annual declines are still 6 to 8 per cent.

Her own caveats belong attached to the numbers: "an early shift rather than a full recovery", "a fragile recovery", and a warning that a deterioration in the rate outlook could slow the momentum.

Her explanation is that the Budget's negative gearing and capital gains changes reduced the appeal of established housing to investors, and the effect is largest where investors were the largest share of demand.

That is correct. But the sequence needs stating carefully, because read quickly her data looks self-contradictory — the expensive end fell first and is now recovering, while the cheap end held up and is now falling.

Both are true. They are two different events, with two different causes, and separating them is the whole point.

Two waves, and they are not the same downturn

The first wave hit the top, and it was a rates event. Values peaked in March and upper-quartile stock drove the steepest losses from there — large loans, the most serviceability-sensitive borrowers in the country, three rate rises this year. That is the wave I wrote about on 2 September. Nerida's data now shows it bottoming: the premium markets that fell hardest over the past year are the ones posting August price rises and recovering attendance, because owner-occupiers are returning after a 6 to 8 per cent annual adjustment.

The second wave is hitting the bottom, and it is a tax event. Affordable, investor-heavy markets held up better on price through all of that. They are deteriorating now — Adelaide South, Perth's north, Cairns — and the timing tracks the Budget rather than the rate cycle.

The premium end fell because of what a bank would lend.

The affordable end is falling because of who is allowed to claim a loss.

Same downturn on the chart, opposite ends of the market, unrelated mechanisms.

The two do overlap, and it would be tidy-minded to pretend otherwise: the rate-driven fall had already broadened into most capital city suburbs by July, which is what forced CBA's hand. What is new in Nerida's data is a second, separate withdrawal, concentrated precisely where investors were the largest share of the buyer pool.

Which is the part her piece does not follow — because it is not what she set out to write.

Investor money has drained out of established housing in the markets where investors were the biggest share of the bidding: Adelaide South, Perth's north, Cairns, and the affordable investor-heavy catchments like them around the country. That is what her attendance figures are measuring.

It did not evaporate. It went somewhere.

That is this article.

A second count, from a different firm

Sydney auctioneer Damien Cooley told the AFR Property Summit on 8 September that auctions are averaging 2.1 bidders, down from more than seven a few months ago, and that buyer sentiment is at "an absolute all-time low".

Two independent counts of how many people are standing in the room, measured different ways by different firms, both landing near two. That is worth more than either number alone.

The scale, from Ray White's own chart

Ray White's three-year chart of four-week rolling attendees per open, sourced from NurtureCloud and current to 29 August, puts this year against the two before it. The readings below are taken off the chart rather than a published table, so treat them as approximate.

The annual peak is February, not January — January is the holiday trough, and any commentary treating it as a high point has the shape of the year wrong. February 2024 is the highest point on the chart at around 4.65. February 2026 sits near 4.5. February 2025 nearer 4.0.

The more useful comparison is not the spikes. It is spring 2025, which ran at roughly 3.8 rising to about 3.95 and held there for something like four months. February peaks collapse within weeks. A sustained run gives a campaign room to work in, and on that chart the best selling window of the past three years was last spring, not any February.

Against that: around 3.8 in September 2025, against 2.2 now. Down roughly 40 per cent year on year, with 2026 tracking below both prior years for the entire year.

So the recovery being reported is around 0.2 off the lowest base in three years.

That does not contradict Nerida — she says fragile, she says early shift, and she is right on both. It puts the scale in proportion, using her own firm's chart.

Which investors left, exactly

The reform is narrower than the coverage of it, and the narrowness is the whole story.

From 1 July 2027, an investor who buys an established residential property after 7:30pm on 12 May 2026 can no longer offset a rental loss against salary — the loss is quarantined against rental income or a future capital gain. Established property bought before that date is fully grandfathered. A new build is fully exempt, whenever it is bought. Under the capital gains reset, new builds also keep a choice no other property gets: retaining the legacy 50 per cent discount as an alternative to the new indexation-plus-minimum-tax model.

Favourable treatment twice over, in the same Budget.

So this is not an investor exodus from property. It is an exodus from one category of property, and a deliberate push toward another.

Which is why the second wave is landing where it is. The withdrawal is not spread evenly across the country — it lands in proportion to how much of the local buyer pool was investors in the first place. Where they were a small share, losing them costs a few bidders at the back of the room. Where they were a large share — Adelaide South, Perth's north, Cairns — losing them removes the bid.

Those markets are not falling because they ran out of buyers generally. They are falling because one specific cohort, which happened to be carrying an outsized share of the bidding there, has a legislated reason to stop.

The investor was not taxed out of the market. The investor was redirected.

The question that decides what happens next is whether the destination exists.

The destination has a two-and-a-half year queue in front of it

Before finance, the mechanics of actually getting a new dwelling built.

NEX Building managing director Andrew Helmers, whose group operates McDonald Jones Homes and Arden Homes, laid out the timeline at the same summit. A customer typically buys unregistered land. That takes around 18 months to come to title. Approvals through local authorities run another 60 to 100 days. The build itself takes about 100 days.

Two and a half years of waiting before construction starts.

The Productivity Commission's own data has the median approval time at 146 business days in Victoria and 18 calendar days in South Australia — an eight-fold spread between states for the same administrative step. Its chairwoman Danielle Wood said in April that houses, townhouses and apartments now take 40 per cent longer to build than they did 15 years ago.

The cost base has not stopped moving either. Helmers put increases as high as 9 per cent in some states. Built Living's Leanne Boyle put forward escalation at 3 to 5 per cent. Metricon chief executive Brad Duggan described residential trades being pulled into data centre construction, with "$50-an-hour differentials between a third-year apprentice electrician working on a residential site compared to a data centre" — and said the momentum his business was building in the first half of the year "has evaporated in the last three months".

Housing Minister Clare O'Neil put the same point from the other side: projects that stacked up on the economics of seven or eight years ago are not getting built today.

None of that is new to anyone in construction. What is new is what happened to the funding.

The finance wall

This is the part that was missing from the story a week ago.

Bathla — a developer with a 14,000-home pipeline and 2,500 half-built homes — collapsed into administration in late August owing $3.4 billion, with the majority of its debt funded by 40 non-bank lenders. Receivers have been appointed to at least 17 of its projects.

Andrew Schwartz runs Qualitas, a non-bank lender that funds property development — the kind of money a builder uses when a bank will not write the loan. His firm looks at roughly one development funding deal a day, which makes him about as close to the coalface of development finance as anyone in the country. He said the cost of that debt had already moved 30 to 50 basis points higher before Bathla collapsed, and that lenders now expect to be compensated for the risk that there are more like it.

Mirvac chief executive Campbell Hanan said his company is fielding calls from private credit-backed syndicates frustrated by rising holding times and costs. Victoria Hardie, whose firm HMC Capital runs a commercial property lending book, put it plainly from the lender's side: the risk/reward equation has swung back toward lenders, who can now get a better return for the same risk than twelve months ago.

The AFR's summary of the summit was blunter than any of them: the elephant in the room is feasibility, projects signed off six weeks ago now look very different, and the numbers no longer stack up.

Then the delivery data. Balmain chief executive Andrew Griffin, one of Bathla's lenders, put completions at 173,400 homes in the year to March 2026, against the 240,000 annual run-rate the National Housing Accord requires. The National Housing Supply and Affordability Council had the 1.2 million-home target running about 18 months behind before any of this.

Put the two halves together.

The tax reform points investor capital at new builds. New builds require a developer to fund a project through a two-and-a-half year queue, at a cost of debt that just repriced upward, in a market where the end price is falling and building costs are still climbing. The channel the policy redirects investors into is the channel that has just become hardest to finance.

That is the answer to where the money went. A meaningful part of it has not gone anywhere, because there is nothing built for it to go into — and less coming than there was in June.

The metric being read as a recovery can move on supply alone

One more thing to resolve before anyone treats 2.2 as a floor.

Attendees per open is a ratio. Attendees on top, open homes underneath. It rises if more buyers turn up. It also rises if fewer homes are opened.

Nothing in the 7 September piece addresses listing volumes. Every figure in it is demand or price. That is not a criticism of the analysis; it is the boundary of what it measures.

If withdrawn stock is being held back rather than brought to market — and vendors waiting rather than listing has been the consistent theme of auction commentary since winter — then part of the move from 2.0 to 2.2 is the same buyers spread across fewer opens.

I am not asserting that is what happened. I do not have the listings series alongside it, and it was not published with one. What I am saying is that the two possibilities behind that number are not remotely the same business:

One of those is worth hiring for. The other is worth staffing carefully through.

The line that should end the "wait for the bottom" conversation

Morgan Stanley chief economist Chris Read gave the summit three markers for the end of the slump: transaction volumes picking up, buyers becoming confident that rate cuts are coming, and the gap between borrowing costs and rental yields narrowing.

Then he said the part that matters more than all three:

"If you look at the 18 months after a housing trough, the majority of a pick-up in activity is driven by investors."

Set that against a reform that has quarantined negative gearing on established property and redirected investors into a segment that cannot currently be financed at scale.

The cohort that normally leads the recovery has been legislated out of the stock that makes up most listings, and pointed at stock that is not being built. That is a structural statement about the shape of the next upswing, not a forecast about its timing — and it is the same conclusion I reached from the yield arithmetic on 2 September. Read gets there from the transaction data. I got there from what an investor needs a property to yield. Two different routes, same destination.

Meanwhile the rate side is not helping. RBA assistant governor Sarah Hunter told the summit that inflation is "top priority right now" and that the board "may well have to raise interest rates" — with the trimmed mean stuck at 3.6 per cent and bond traders pricing roughly a 70 per cent chance of a rise on 29 September. Hunter also confirmed the RBA expects housing construction to fall in 2027 and 2028, and named the sectors already wearing it: real estate agents, conveyancers, removalists.

That last part is not commentary about the market. It is a central bank naming your industry as a transmission channel.

What this means for an agency

Three things, and none of them are forecasts.

You have fewer bidders on established investor stock, and not evenly across the country. In the affordable, investor-heavy areas, the buyer who was setting the top of the bidding now has a reason written into law to stay home. You will feel that as longer days on market, and as offers coming in under the last comparable sale — well before any price report picks it up.

That money may never come back through your door. An investor who takes the exemption buys new — from a builder, a display village, a project marketer. There is no listing for you to win, no commission for you to write, and the property does not come back on the market for years. The rental management goes wherever the builder sends it. Same investor, same suburb, same money — someone else's settlement.

Attendance is not commission. People walking through open homes do not pay wages. What pays is a listing signed, then a contract signed, then a settlement that actually banks — and between the first of those and the last sit 60 to 90 days where the wages, the rent and the software still go out on time, every time.

So a busy open home and money in the bank are two different things. You can see the first one this Saturday. You will not see the second until the month it either turns up or it doesn't.

There is a fourth, and it is big enough to get its own article next week. John McGrath told the summit he is already fielding approaches from agents wanting to sell their rent rolls, and expects a sharp fall in agent numbers before this is over.

The rent roll is the only part of an agency that pays you every month whether you list anything or not, and the only part someone else will pay real money to take over. When a lot of them come up for sale at once, the industry's one lasting asset changes hands — and whether you are the one buying or the one selling comes down to what your bank will back, not how good your year was.

Which is where Andrew Griffin's line about Bathla belongs. Explaining the collapse to his own investors, the lender wrote:

"Balance sheets don't fail because a market moves; they fail because there was no slack left when it did."

That was written about a developer with a $3.4 billion book. It applies without a word changed to an agency with eleven staff and a rent roll facility.

No principal works through their funding position in a comment section. If you are quietly running that number, send me a DM — or put fifteen minutes in the diary directly: calendly.com/peergroupadvisory/15min

---

John King is a Fellow of the Institute of Public Accountants and an Authorised Credit Representative. Peer Group Advisory works with boutique real estate agency principals on forward cash flow, structuring and acquisition funding.