Last week Westpac dropped its May 2026 housing update. That makes eight major 2026 calls now public — CBA, Westpac, NAB, ANZ, KPMG, Domain, SQM Research and PropTrack. For Sydney they range from -3% to +7%. That's a ten-point spread on the country's biggest market. Whose call do you trust?
| Forecaster | 2026 call |
|---|---|
| Westpac IQ Housing Forecast Update, 26 May 2026 | -3% |
| ANZ Research Bluenotes, April 2026 | Small fall (~-1 to -2%) |
| CBA Reported via APU, Q1 2026 | Flat to slight fall |
| SQM Research (base) Boom & Bust Report, Nov 2025 | +3 to +6% |
| NAB Reported via APU, Q1 2026 | Modest positive (~+3-5%) |
| PropTrack 2026 outlook commentary | +5 to +7% |
| KPMG Residential Outlook, Jan 2026 | +5.8% |
| Domain Forecast Report FY26 | ~+7% (median ~$1.9M) |
Range: roughly -3% to +7%. Spread: 10 percentage points.
That isn't a normal forecaster disagreement. That's the market telling you no one knows.
Perth 2026 calls range from +9% (Westpac) to +13% (KPMG) to +12-16% (SQM). Brisbane runs +7% to +13%. Even Melbourne — supposedly the easy "weak" call — has KPMG at +6.8% and Westpac at -4%. An eleven-point spread. When forecasters cluster, you can lean on the consensus. When they split this wide, the consensus number is fiction.
Two recent reminders of how badly the consensus can miss inflection points:
Late 2022, predicting 2023
The big four banks went into 2023 forecasting national house price falls in the 5-8% range. By December 2023, CoreLogic's national index was +8.1%, with Sydney up double-digits. The consensus was directionally wrong by roughly 13-15 percentage points.
Early 2024, predicting 2025 Perth
Most major 2024 forecasts had Perth growing 4-10% in 2025. Actual: Perth's median jumped 9.9% in the December 2025 quarter alone — pushing past $1M for the first time and finishing the year up roughly 15-16% YoY. The miss was about 8-10pp.
The pattern is consistent: forecasters under-call momentum, over-rely on mean reversion, and miss inflection points by a wide margin. That's not a knock on the forecasters. Forecasting is hard. But it does mean the headline number you read in the paper is rarely the number you should anchor on.
Pull at least three numbers for any city you care about: long-run growth (CoreLogic 10-year), suburb-level actuals from the last 12 months, and the average across the eight forecasters above. If a property pitch leans on a single bank's forecast, that's the weakest part of the case.
Westpac revised its Sydney 2023 call from -8% to +7% across the year. A bank that is rapidly revising in one direction is telling you the current call is unstable. Track changes month-on-month, not the snapshot.
A national or capital-city forecast averages across hundreds of submarkets. The same year Sydney falls -3% at the headline level, a specific corridor can grow 15%+. The Central Coast LGA, Liverpool LGA and outer-Penrith ring all grew double-digits in the year ending May 2026 — even as some Sydney inner-east suburbs went sideways. If you can't trust the headline, work one level down.
—
One question
Which forecaster's 2026 call have you actually read end-to-end — not just the press summary, but the underlying methodology? Reply with the source. I'm compiling a side-by-side dispersion chart for next week and will send it back to anyone who replies.
Go one level down
City forecasts are noise. Suburb data is signal.
Run a free Estait report on any Australian suburb — actual median, yield, DOM, vacancy, growth, zoning, and an AI investment brief.
Run a free report →Forecast figures sourced from each institution's most recent public publication: KPMG (Jan 2026), Westpac IQ (May 2026), ANZ Bluenotes (April 2026), SQM Research Boom & Bust Report (Nov 2025), Domain Forecast Report FY26, PropTrack 2026 outlook commentary, and Australian Property Update coverage of CBA/NAB Q1 2026 economics notes. Historical actuals from CoreLogic Home Value Index and Domain December 2025 House Price Report.
Forecast figures sourced from public publications as of 2026-05-30. Not financial advice.
Get the next issue free
Subscribe to The Estait Signal →