Sydney and Melbourne's top-end homes are down more than 10%: how far the housing downturn runs while the RBA priooritises inflation

Australia's most expensive homes are leading its property downturn. In the top quarter of the market, values have now fallen 10.7% from their peak in Sydney and 10.5% in Melbourne, according to Cotality's September charts. The national index has dropped for five straight months, August's 0.9% fall the steepest since the market tipped over, with prices down in roughly 93% of suburbs. This is the correction the Reserve Bank of Australia was always going to tolerate: house prices in the country's two priciest cities are the last thing its inflation battle protects.

Conventional wisdom holds that a housing downturn feeds on itself. Higher mortgage costs curb demand, demand pulls prices down, falling prices shrink household wealth, and diminished wealth softens spending, which curbs demand again. The loop looks self-reinforcing, and markets are treating it that way. The slight twist that matters is where the loop is biting first. The biggest declines are at the expensive end, the segment where buyers are least dependent on debt. That tells you the mechanism doing the work right now is not the credit one that torments bankers, but the wealth one that unsettles retailers.

The numbers bear this out on one side of the ledger and defy it on the other. Retail is feeling the downturn already: gross retail profits fell 5.2% in the three months to June, the largest quarterly decline since 2024, and big listed retailers warn that on-the-ground sentiment is worse than official statistics imply. Household spending recorded its first fall since September 2024. A housing slide that erodes the perceived wealth of the country's top spenders lands on consumer-discretionary earnings now, not after a lag.

Bank credit quality is another story entirely, and it has not yet turned. House prices rising 18% over four years left less than 1% of mortgagors in negative equity; the median borrower rebuilt prepayment buffers during 2025's rate cuts; and mortgage arrears, while steady, have yet to reflect the two rate rises delivered in March and May. The share of housing loans more than three months in arrears sits around pre-pandemic levels. Over 90% of non-performing housing loans are still "well secured", meaning the collateral covers the debt. None of this is the texture of imminent bank losses.

So the puzzle hangs on which channel reaches the central bank first. The RBA is not going to pivot because a few millionaires lose a tenth of their paper wealth; it will move only if the loop starts showing up where it is paid to care — unemployment and lending stress. On that score the loop is still quiet. The cash rate sits at 4.35%, inflation at 3.5% remains above the 2-3% target, and the market's own economists expect more tightening, not a pause: NAB calls a September hike to 4.6%, Commonwealth Bank has moved its call to November for the same move. The next decision lands on September 29th.

This is where the "amplified by forecasters" reading of the downturn becomes discriminating. Housing expectations have upped their pessimism as the fall gathers speed — National Australia Bank now projects 10% peak-to-trough falls in Sydney and Melbourne, Commonwealth Bank 12-13% — and the market has already front-run the consequences in bank share prices. Since late February NAB has fallen 23%, Westpac nearly 15%, ANZ 11%, and even Commonwealth Bank, the fortress, is down. Yet the actual credit losses the shares are discounting remain hypothetical. Macquarie has cut bank earnings forecasts by only 1-4% through 2028.

That gap — prices sliding, credit stress delayed — is the instructive part for an investor. Mortgages make up roughly 60% of the big four banks' combined loan books, against 40-50% for global peers, so the housing cycle is nearly all there is to their credit economics. But the top-end fall now underway is concentrated in the most lightly leveraged borrowers, the ones the banks least fear, while the leveraged end that would stress them — investor loans, which were falling by up to half year on year — is retreating precisely because leverage has become unaffordable. A downturn led by cash-rich sellers is a weak engine of mortgage losses.

The thesis that this is accelerating faster than anyone models therefore stands or falls on three specific observations, and it is currently winning on roughly one of them. Price declines are not moderating; they are broadening. Mortgage stress is not yet in the data, which is how a lag looks, not how containment looks. And the central bank shows no sign of a dovish pause — rather the opposite. The loop is running through spending and asset wealth, and until it prints in jobs or arrears, the RBA's inflation priority gives it room to keep the correction running. The gain to Australian banks from a dovish turn is the whole game, and the timeline to that turn depends on the credit leg finally waking up. So far the evidence says the pain is reaching the consumer before it reaches the lender — and that a central bank indifferent to a 10% fall in top-end values is a deliberate feature, not an oversight.