Inner-urban apartment approvals collapsed everywhere in early 2023 and have since rebuilt across every mainland capital — not just in Melbourne, where the effect is largest, but in comparable form in Sydney, Adelaide and Perth. Extending the APN Codex approvals series back to July 2022 makes this a documented three-city pattern rather than a single-city anecdote.
A pattern, not an outlier
Comparing new dwelling approvals in the January–March 2023 trough against the July–September 2025 national peak, a distinct cluster of inner and middle-ring local government areas across four capitals recorded growth well above the 41% national average.
Melbourne supplies the largest number of LGAs in this cluster — Yarra, Darebin, Banyule, Maribyrnong and Maroondah all recorded growth between 176% and 588% — but the pattern is not Melbourne-specific. Sydney’s Canada Bay recorded the single largest gain in the national dataset at any base size worth treating as reliable, rising from 75 to 1,232 approved dwellings. Adelaide’s Unley (+569%) and Perth’s Vincent (+515%) and South Perth (+347%) sit in the same range as the Melbourne cluster, on comparable base sizes.
One LGA is deliberately excluded from the chart on methodological grounds: North Sydney’s approvals rose from 4 to 102 dwellings across the same window, a percentage change so large (+2,450%) that it is more a function of the tiny starting base than a comparable signal to the LGAs above. It points in the same direction as the rest of the cluster but should not be read as a like-for-like figure alongside Canada Bay or Yarra.
A caveat the underlying data demands
The trough-to-peak figures above compare two single quarters, and at this cluster of LGAs that comparison needs a qualification the headline percentages do not show on their own. Individual apartment and medium-density developments are approved as single large jobs, and a handful of such approvals can dominate one LGA’s entire quarterly count. Canada Bay’s own quarterly path illustrates this directly: 75 dwellings in the Q1 2023 trough, but no smooth climb from there — 571 in Q2 2023, back down to 83 in Q3 2023, a mix of quarters in the 60–400 range through 2024, then 1,232 in the Q3 2025 quarter used as this series’ peak, followed immediately by 179 in Q4 2025 and 84 in Q1 2026. The trough-to-peak comparison is real and both endpoints are correctly sourced, but Q3 2025 for Canada Bay looks more like a single large approval landing in that specific quarter than a sustained tripling of the underlying market.
Summing the ten-LGA cluster by quarter, rather than looking at any single LGA, reduces this effect without eliminating it.
Even at cluster grain, quarter-to-quarter swings remain large — from a low of 830 in Q4 2023 to a high of 4,639 in Q3 2025, a range that ten LGAs summed together do not fully smooth out. What the cluster chart does show, and what a single-LGA comparison cannot, is a genuine shift in the floor: the lowest quarters within 2023 sat at 830–883, the lowest quarters within 2024 rose to 1,039–1,312, and the lowest quarter within 2025 (Q2, excluding the Q3 spike) sat at 1,579. Each year’s weakest quarter has been stronger than the last. That progression is a more defensible basis for describing a genuine level shift in this cluster than the single trough-to-peak comparison alone, even though the trough-to-peak figures remain accurate as reported.
What is driving it
Each of the LGAs in this cluster is characterised by a housing stock weighted toward apartments and medium-density product rather than detached houses — the building typology most exposed to financing conditions, construction cost inflation and pre-sale thresholds during 2022–23, and also the typology most prone to the single-large-project lumpiness described above, since one apartment tower can represent hundreds of dwellings approved in a single month. The rising quarterly floor through 2023–2025 is consistent with an apartment-specific cycle running underneath the broader national approvals trend, distinct from the greenfield detached-housing growth corridors that make up the rest of the national recovery. This dataset does not measure financing conditions or pre-sale thresholds directly, so that explanation should be read as a plausible account consistent with the observed pattern, not a conclusion this data proves on its own.
