Ten Per Cent Won't Do It

The Commonwealth Bank has downgraded its housing outlook again, and this time there is nowhere left standing. A capital city average decline of 10%, peak to trough, driven by higher rates and the Budget's tax changes.

Worth pausing on how quickly that position moved.

Four forecasts, one calendar year

In March, CBA had national dwelling prices rising 5% over 2026. After the May Budget it moved to 3%. In June it went to flat, with Sydney down 6% and Melbourne down 7% — but Brisbane, Perth and Adelaide still growing.

Now it is a 10% correction, and every capital is in it.

That is a fifteen-point swing in six months on the same year, from Australia's largest home lender, with the best view of loan applications in the country.

I am not raising that to score a point. The revisions are the story, because of what they say about the lag.

The forecast was chasing an index that was itself behind

National values peaked in March and have fallen five months in a row. By the end of August they were 3.6% below that peak.

But the monthly prints have been revised down almost every time. July was first published as a 0.7% national fall. It was revised to 1.2% — nearly double. Cotality's own commentary flagged that August's 0.9% may well be revised materially larger over coming months. Their research director noted the largest downgrades were occurring in Perth and Brisbane.

So the forecast was lagging the index, and the index was lagging itself.

Both of them are lagging the thing that actually matters, which is what a contract is being signed at this week. A hedonic index is built on settled sales. A property that goes under contract in September settles in November and enters the data after that.

Brisbane, and the shape of the fall

Brisbane dwelling values fell 1.0% in August after 0.6% in July, putting the city 2.7% below its May peak. The August fall is close to double July's, so the decline is accelerating rather than steadying.

The city-wide number hides what is actually happening underneath it.

This correction started at the top and it started months ago. Inner city and upper quartile stock has driven the steepest losses throughout. Over the three months to July, Brisbane's upper quartile values fell while the lower quartile actually rose. Nationally the gap was wider still — upper quartile down more than 3% over the same period, lower quartile marginally up.

For most of this year that made it a segment story. Expensive stock, large loans, the most rate-sensitive end of the market. Easy to look at from a suburb where nothing was moving and conclude it had nothing to do with you.

What changed is July and August.

The correction broadened. By July more than three-quarters of capital city suburbs were recording falls over the prior three months, and Brisbane and Adelaide had posted their first consecutive monthly declines. Cotality described the downturn as having become genuinely national for the first time in this cycle.

That is what forced CBA's hand. In June the bank still had Brisbane, Perth and Adelaide growing, because at that point the falls were concentrated at the top of Sydney and Melbourne. Two months later there is nowhere left standing. The forecast did not move because the top end fell further. It moved because the fall stopped being confined to the top end.

A narrow correction is a segment story. A broadening one is a market story.

And the broadening is now reaching mid-market stock in the outer suburbs. Agents writing contracts in the outer north are reporting four-bedroom houses transacting around 6% below where they sat in May — roughly twice the city-wide rate. That figure is what is being said in listing appraisals this week rather than a published statistic, and it should be read as such.

That stock is not the top of the market and it is not the bottom. It is where a buyer needs a large loan without a large income, which makes it the most serviceability-exposed segment there is. It is also the only part of the market where a new build genuinely competes with an established home two streets away.

The lower quartile is still holding. It is the last thing to go, and it will go last because it is where everyone trading down eventually lands.

Where we are in the cycle

If the peak was March and CBA is right about the destination, roughly a third of the fall has happened.

Stockland put the usual cycle at 12 to 18 months and said in August that the market was about six months in. On that reading, the trough lands somewhere between the middle and the end of 2027 — which is also when CBA expects prices to stabilise and lift.

A further rate rise makes it the longer end of that range. NAB has the cash rate reaching 4.85% by November.

And a rate rise does something worse than extend the timeline. It moves the target at the same time. A higher cost of capital lifts the yield an investor requires before a deal clears, so the correction needed grows while the time available to deliver it also grows.

The part nobody is running

Here is the question that decides whether a correction is a reset or just a slow bleed: at the end of it, does the investor come back?

Brisbane houses carry a gross rental yield of around 3.2%. The modelling that produced the widely quoted 32% figure put the minimum yield an investor needs, once negative gearing is quarantined, at about 5.15%.

Take CBA's full 10% correction and run it to completion. On the same basis, that lifts the Brisbane house yield to roughly 3.6%.

Still around 30% short of the hurdle.

That is my arithmetic on the published yield, not CBA's and not Ray White's, and it should be read as an illustration rather than a forecast. But the direction is not in doubt.

A 10% correction covers about a sixth of the distance between where investors are and where they would need to be. It is a serious event for every vendor in the country and it does not, on its own, bring a single investor back into established housing.

The remainder has to come through rents. Which is why the rental data matters more than the price data right now — and why a segment where rents have stopped moving is the thing to watch, not the headline percentage.

What this changes for a principal

Two things.

A comparable from June is not evidence. If the largest lender in the country has moved the same year's forecast from plus five to minus ten in six months, and the index has been revised down every month, then a settled sale from three months ago is a historical record rather than a guide to price. That is doubly true this quarter, because June sits on the wrong side of the broadening. A comparable struck before July describes a market where the correction had not yet arrived in most suburbs. Vendors will bring those comparables to the appraisal. They are describing a market that no longer exists.

And a city median is no better. With the top end down several times the city-wide figure and the bottom still rising, the average describes almost nobody. The only useful comparison is the same price band in the same catchment, struck in the last six weeks.

And the correction does not end the buyer problem. There is a temptation to tell a vendor to hold until the market bottoms and the buyers return. On these numbers the investor buyer does not return at the bottom of a 10% fall. The pool that comes back first is owner-occupiers, whose applications fell far less than investors' after the Budget, and who are constrained by serviceability rather than by tax.

Pricing to the market that exists is not pessimism. It is the only way a listing sells in a market where the data everyone is quoting is describing last quarter.

No principal works through their pipeline position in a comment section. If you are quietly running that number, send me a DM.

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.