The Wealth Effect Runs in Reverse

Spring's second weekend printed 58.5% on 1,594 auctions (1,197 reported: 704 cleared, 493 uncleared), the highest preliminary read since 3 May (60.2%), up 5.8 points on last week's 52.7% and the largest one-week rise since February. The book grew 9% on the week to its largest since late June, and the composition strengthened with the headline. 440 homes sold under the hammer against 253 prior to auction, lifting the hammer's share of cleared sales to 62% from 55%, while withdrawals fell to 190 (11.9% of scheduled) from 229 (15.7%) and passed-ins to 303 (25% of reported) from 309. Every capital improved. The My Housing Market read (Dr Andrew Wilson) rose to 51.9% from 46.1%, describing 'some early signs of strengthening auction market activity to begin the spring selling season, although certainly coming from a low base in most capitals'. The same weekend a year ago cleared 74.8% on a preliminary 2,455 auctions (69.0% final on 2,402): volume is down a third year-on-year and clearance 16 points. Last week's 52.7% preliminary finalised at 49.3% (Sydney 52.5%, Melbourne 54.6%, Brisbane 26.0%, Adelaide 36.2%, Canberra 37.1%), a 3.4-point revision, so this print should settle near 55%, which would still be the best final since May.

City by city

Sydney firmed to 59.6% on 557 scheduled (394 reported, 235 cleared), its best preliminary since early May. The composition is still negotiation-led, with 124 sold prior against 107 under the hammer and 4 after, but withdrawals eased for a second week to 92 (16.5% of scheduled) from 105 (20.1%) and 127 (24.6%) before that, and 67 passed in. On the price side nothing new landed this week: Cotality's August index has Sydney down 1.4% to $1,222,718, 4.7% over the quarter and 7.1% below the February peak, with houses (-1.8%) falling more than four times as fast as units (-0.4%). The ABS added the state-level view on Tuesday: NSW's dwelling stock lost $92.9 billion (2.0%) in the June quarter, the deepest fall of any state. A year ago Sydney cleared 74.7% preliminary.

Melbourne lifted to 63.3%, its best since 29 March (63.5%), on 713 scheduled, the largest book since mid-June and larger again than last week's 662; 577 reported, 365 cleared, 265 under the hammer against 96 prior, 147 passed in, and just 65 withdrawn (9.1%, the lowest share of any major). Eight straight weeks in the mid-50s-to-60s band on rising volume, now at prices 1.1% lower again in August ($786,718, -3.9% for the quarter, -4.7% annual, roughly 6% below the November 2025 cyclical peak) is a demand floor. The economic backdrop is softer than the auction floor: Victoria's dwelling stock fell $44.3 billion (1.6%) in the June quarter and Victorian business conditions dropped 13 points in the August NAB survey. Gross yields at 4.0% (units 5.1%) are the highest of the eastern capitals. Our four-signals framework still has one of four buy conditions met, and the buy waits on the credit signal. A year ago Melbourne cleared 75.2% preliminary.

Brisbane snapped back to 41.6% from 24.8% (168 scheduled, 125 reported, 52 cleared with 19 prior and 33 at auction, 56 passed in, 17 withdrawn), the largest one-week improvement of any capital, but still one in three passed in and the supply swing is intact: total listings are up more than 24% on a year ago and new listings are still rising, the opposite of Sydney. Values fell 1.0% in August to $1,080,142, 2.7% below the May peak, with July revised from -0.6% to -1.2%. Queensland was among the states where the ABS still recorded rising dwelling values in the June quarter. A year ago Brisbane cleared 74.0% preliminary.

Adelaide rebounded to 55.6% from 34.6% on 78 scheduled: 54 reported (69% collected against 58% last week), 30 cleared (5 prior, 23 under the hammer, 2 after), 20 passed in, only 4 withdrawn. Last week's scare was the sample, as flagged: the 34.6% preliminary finalised at 36.2%, and this week's read puts Adelaide back above the combined line. Values fell 0.8% to $937,207 in August, 1.6% below the May peak, still the shallowest correction of the majors, and South Australia's dwelling stock rose in the June quarter on the ABS measure. Regional SA (+0.6% in August, +2.3% for the quarter, +11.4% annual) remains the only rest-of-state market in the country still rising. A year ago Adelaide cleared 75.4% preliminary.

Canberra improved to 46.9% on 32 results (57 scheduled, 15 cleared, 8 passed in, 9 withdrawn); values fell 1.1% in August to $864,998. Perth reported 15 results (7 cleared), a private-sale market where the auction read carries no signal. Values fell 0.8% in August to $999,987, back under $1 million, and WA business conditions fell 13 points in August, the joint-largest fall of any state. No Tasmanian auctions were recorded this week.

What the tape says

Two weekends into spring, buyers are turning up at these prices: every capital improved, the hammer out-sold the negotiation by the widest margin of the season, and vendors withdrew fewer homes than in any week since winter. That does not make the correction over. The year-ago gap is 16 points, finals will take three to four points off, and the improvement came on a book still a third smaller than last year's. The stronger clearance and the weaker economic data belong to the same story: vendors are cutting reserves because the alternative is sitting on stale stock into a hike, and that is where the buyer's leverage sits this spring. The next tests are the 29 September RBA decision and the size of the pre-Grand-Final book in Melbourne.

The week moved the rate debate from markets to institutions. On Monday 8 September Westpac abandoned its hold call: chief economist Luci Ellis wrote that 'the likelihood of an additional rate hike has risen enough to make a November hike (+25bp to 4.6 per cent) the base case again', citing 'the growing evidence of a more resilient household sector' in the national accounts and internal data, and a larger-than-expected impetus from data-centre construction. Westpac does not think a September move is the most likely outcome, expecting the Board to wait for the September-quarter CPI and revised forecasts in November, and pencils three 25-basis-point cuts from August 2027. That leaves all four majors forecasting a hike: NAB at the 28–29 September meeting (with November as its risk case for a second move), CBA, ANZ and Westpac in November. None publishes a cut in 2026.

The same evening Deputy Governor Andrew Hauser told ABC's 7.30 that inflation is 'the one big problem' left in the economy, that the public is 'furious' about a price level 10 to 20 per cent higher than it should have been since COVID, and that the question before the nine-member Board is 'have we done enough or is more needed'. He named three upside risks (the Middle East, an AI-driven global upswing, and Australia's weak supply capacity) and declined to call a September hike inevitable, stressing genuine disagreement on the Board. On housing he was explicit that the Bank is looking through prices: values remain about 3% above a year ago and 50% above the start of the decade, the supply shortage puts a floor under them, and they are 'part of the story', not the target.

The data that landed around those two statements pointed the other way. The ABS reported the first fall in the total value of Australia's dwelling stock since September 2022: down $34.1 billion (0.3%) to $12.7 trillion in the June quarter, with the mean dwelling price down 0.7% to $1.1 million; NSW (-2.0%, -$92.9 billion), Victoria (-1.6%, -$44.3 billion) and the ACT fell while the other states still rose. The Westpac–Melbourne Institute consumer sentiment index slumped 5.2% to 84.4 in September, erasing August's gain, on fuel prices and rate fears; assessments of family finances against a year ago fell 9.2%, the 'time to buy a dwelling' sub-index fell 10.7% to 85.5, and house-price expectations (110.3) sit at their lowest in several years. NAB's August business survey recorded conditions at -1, down five points and negative for the first time since August 2020, with confidence at -8, profitability down ten points to a post-COVID low, construction among the sectors leading the fall, and Victoria and WA each down 13 points; purchase-cost growth (2.3% quarterly) is running well ahead of product-price growth (0.8%), so the pressure is on margins rather than on prices. Roy Morgan's July mortgage-stress read added the household side: 32.5% of mortgage holders 'at risk', 1.79 million people and up 180,000 in a month, the highest share in 18 years and within three points of the 2008 record.

The transmission to borrowers is unchanged. Lenders assess new loans at the contract rate plus APRA's 3-percentage-point buffer, so a move to 4.60% lifts assessment rates from about 9.4% toward 9.65% and trims borrowing capacity another 2–3%, into a market where sales are already about 15% below last year and investor lending fell 10.2% by value in the June quarter. Lenders have nonetheless kept trimming short fixed rates, a sign their own curves see the cycle ending rather than extending. For anyone buying this spring the position is unchanged: underwrite at 4.85%, bid off today's capacity, and read the 2027 cut profile as the earliest the credit catalyst can arrive. The August monthly CPI lands on 30 September, the day after the decision.

No new vacancy print this week. SQM's July read stands, with the August release due mid-month: national vacancy held at 1.3% for a third month, with the two-speed split intact. Sydney (1.7%), Melbourne (1.7%) and Canberra (1.8%) keep loosening gradually; Adelaide, Perth and Hobart hold 0.6%, Brisbane 0.9%, and Darwin at 0.3% remains the tightest rental market in the country. Five of eight capitals sit below 1%. When the August print lands, the things to check are whether the vacant-dwelling count keeps rising (it did in June and July even as the headline held) and whether Sydney and Melbourne move through 1.8%.

The rent side is the one series still moving in investors' favour, and this week's yield guide on the blog sets out how to read it. Gross yield is rent divided by value, and it has expanded for two reasons at once: national values are 3.6% below their March peak while rents rose 5.7% over the year to August, taking the national gross yield to 3.79%, the highest since September 2019, and the combined-capitals yield from 3.34% in December to 3.6%. The distinction that matters is which half of the ratio is doing the work. In Perth, Adelaide, Darwin and Brisbane's unit market the expansion is mostly rent-led: income the owner will receive, on vacancy below 1%. In Sydney and, to a lesser extent, Melbourne it is mostly price-led: a smaller capital base under the same rent, which says more about the asset than about the income. On an 80% interest-only loan at 6.50% with holding costs of 1.2% of value, a leveraged buyer needs roughly 6.4% gross to cover costs before tax (about 5.1% at 60% LVR), which is why the sub-$650K unit corridors (Cannington and Logan Central at 4.7–5.6%, Salisbury at 5.2%) keep leading the Top 10 rotation while the Sydney house market, at 3.3%, does not appear on it. August's seasonal dip in asking rents (houses down month-on-month in six of eight capitals) has not changed the annual picture: growth still runs from +4.3% in Canberra to +11.2% in Darwin.

The wealth effect runs in reverse

For five months the housing downturn was a property story: an index falling, a clearance rate sliding, a set of forecasts being revised. On Tuesday 8 September it became an economic one. Within a few hours the ABS reported that the total value of Australia's dwelling stock fell $34.1 billion (0.3%) to $12.7 trillion in the June quarter, the first quarterly decline since September 2022, with NSW down $92.9 billion (2.0%) and Victoria $44.3 billion (1.6%) while the smaller states still rose; the Westpac–Melbourne Institute index recorded a 5.2% slump in consumer sentiment to 84.4, with the 'time to buy a dwelling' sub-index down 10.7% to 85.5 and house-price expectations at their lowest in several years; and NAB's business survey showed conditions falling five points to -1, negative for the first time since August 2020, with construction, mining and manufacturing leading the fall and Victoria and Western Australia each down 13 points. Roy Morgan's July read of mortgage stress (32.5% of mortgage holders 'at risk', 1.79 million people, the highest share in 18 years) added the household side.

The mechanism is old, and it is now visible

Housing feeds the economy through four channels: the wealth effect on consumption, transaction activity and the services that depend on it, borrowing capacity, and residential construction. NAB Economics puts the wealth effect at roughly 0.1–0.2% of consumption for every 1% change in housing wealth over time, which means a 10% national correction, CBA's base case and within the range of ANZ's and NAB's own, takes about a percentage point off consumption growth over a couple of years. The transaction channel is already measurable: turnover fell to 4.1% of the dwelling stock in July from 5.2% in 2024, and the lost sales alone are estimated to have subtracted 0.10–0.15 percentage points from June-quarter GDP, on the way to the 0.4-point drag of 2018–19 if the downturn runs as long. Housing credit growth, running near 8% a year, is projected to slow toward 3% by late 2027 as loan sizes shrink with borrowing capacity. Construction is the lagging channel: a detached house now takes about 45 weeks to build, half again longer than before the pandemic, so the large pipelines in Queensland, WA and SA will hold completions up for a while, and NSW and Victoria, with thinner pipelines and the deepest price falls, will feel the slowdown first. This is the reverse of the boom that ran from 2023 to early 2026, when rising values, high turnover and easy equity release added to spending. It is also, as the RBA's deputy governor was careful to say this week, what restrictive monetary policy is supposed to do.

The paradox of the week

The same Tuesday, Westpac became the last of the big four to forecast a rate rise, with a 25-basis-point hike to 4.60% in November again its base case on 'the growing evidence of a more resilient household sector' and larger-than-expected spillovers from the data-centre boom, and Andrew Hauser told 7.30 that the question before the Board on 28–29 September is 'have we done enough or is more needed', naming the Middle East, an AI-driven global upswing and weak domestic supply capacity as the risks that could keep inflation above target. He did not call a September hike inevitable; he did say the public is 'furious' about a price level 10 to 20% higher than it should have been, and that the Bank would rather finish the job. On house prices he was explicit about looking through them: values are still about 3% above a year ago and 50% above the start of the decade, the supply shortage puts a floor under them, and they are 'part of the story', not the target. The Bank is weighing a fourth hike against an economy in which business conditions are negative, buyer sentiment has fallen a tenth in a month and household wealth has started to shrink.

What it means for the cycle

Our read is that the two sides reconcile through timing. A hike delivered into a weakening economy does more work per basis point than one delivered into a strong one, because it lands on households whose wealth is already falling and whose stress is already at an 18-year high, which is why the banks that forecast a hike also forecast the cycle turning within a year: Westpac pencils three cuts from August 2027, and none of the majors publishes a cut in 2026. That first cut is the credit catalyst our four-signals framework has been waiting for, and it arrives because the economy softens. Two things follow. First, the reverse wealth effect is a negotiating tool this spring: vendor expectations lag the data by months, buyer sentiment has fallen faster than prices, and every listing, selling-time and clearance measure still favours the bidder; this weekend's 58.5% settles near 55% on finals. Second, the same dynamic argues against waiting for a bottom nobody can time: the floor, when it comes, will be set by the credit signal turning, and by then the stale stock that gives buyers their leverage today will have cleared. Underwrite at 4.85%, hold growth at zero for twelve months as the stress case, and let the yield expansion, 3.79% nationally and the highest since 2019, carry the hold.

The risks to the read

Two would change it. A September hike followed by a second in November, which NAB flags as a risk case, would deepen the consumption channel faster than the housing supply floor could offset it, and the 2022–23 comparator (a 9.1% peak-to-trough fall) would stop being the ceiling. And a continued climb in distressed listings in Queensland, WA and SA, up 25–51% on a year ago from low bases, would turn a demand-led correction into something with a supply push behind it. Neither is in the data yet; both are why the four signals decide when to buy.

The week ahead

SQM's August vacancy and asking-rent releases are due mid-month; Cotality's finals for this weekend publish midweek and will show whether 58.5% holds above 55%; the August labour-force data and the RBA's 28–29 September decision follow, with the August monthly CPI the day after. Melbourne's auction book shrinks sharply for the AFL Grand Final weekend, so the next clean read on the combined capitals is likely to be the first weekend of October.