Abstract
Using Microburbs' proprietary property-level public housing dataset, we count the number of public housing properties at precise distances from 100,000 repeat sales across NSW, Victoria, Queensland, South Australia, and Western Australia (2015–2023).
A single public housing property within 10 metres is associated with 4.2% slower annual growth. On an $800,000 home, that is $33,600 per year in forgone growth. The first public housing neighbour has the largest impact; subsequent properties add diminishing drag.
Combining counts at multiple radii explains 8.4% of growth variance beyond entry price and hold period. This is the first study in Australia to measure the effect at the individual property level.
Contents
Key Findings
- About 4% at 10 metres. A single public housing property next door is linked to roughly 4% slower annual growth (n=283). On an $800,000 home, that is about $32,000 per year in forgone growth.
- The first one has the biggest impact. Going from 0 to 1 public housing within 100m is associated with 3.8% slower growth. Going from 1 to 50+ adds another 2.7%.
- Two scales at work. Close radii (10-500m) capture the direct neighbour effect. Wide radii (5-100km) capture regional socioeconomic context. Both carry independent information about growth.
- Multi-radius outperforms any single radius. Combining counts from 500m to 100km explains 8.4% of growth variance beyond entry price, compared to 6.3% for the best single radius (100km).
- The pattern holds across five states. The growth gap between zero-PH and high-PH areas ranges from 7% (South Australia) to 27% (Victoria).
Property-Level Data: Why It Matters
No government in Australia publishes which specific properties are public housing. The Census reports social housing proportions at the area level. Microburbs has built a proprietary dataset of 138,000 identified public housing properties across five states.
This enables a type of analysis that area-level data cannot support. The Census can report that a neighbourhood is 20% public housing. It cannot distinguish between "public housing is next door" and "public housing is 800 metres away in the same statistical area." Our data shows that difference is worth 4% per year in capital growth.
The Multi-Radius Effect
Cost of the first public housing property at each distance
| Distance | No PH | 1 PH | Growth cost | Sample |
|---|---|---|---|---|
| 10m (next door) | +7.8%/yr | +3.6%/yr | -4.2% | 283 |
| 25m | +7.8%/yr | +5.2%/yr | -2.5% | 742 |
| 50m | +7.8%/yr | +4.4%/yr | -3.4% | 1,132 |
| 100m | +7.9%/yr | +4.1%/yr | -3.8% | 1,662 |
| 200m | +8.1%/yr | +4.0%/yr | -4.1% | 2,839 |
| 500m | +8.6%/yr | +4.7%/yr | -4.0% | 4,875 |
Growth by number of PH within 100m
| PH within 100m | Annual growth | Sales |
|---|---|---|
| 0 | +7.9% | 95,036 |
| 1-3 | +4.0% | 2,090 |
| 4-10 | +5.6% | 1,397 |
| 11-50 | +4.9% | 1,302 |
| 50+ | +1.4% | 175 |
How predictive power scales with radius
We assessed how much additional growth variance is explained by public housing count at each radius, beyond entry price and hold period:
| Radius | Growth variance explained (single radius) |
|---|---|
| 100m | +0.1% |
| 500m | +0.7% |
| 2km | +1.1% |
| 5km | +2.1% |
| 10km | +2.6% |
| 25km | +3.4% |
| 50km | +4.8% |
| 100km | +6.3% |
| All radii combined | +8.4% |
The predictive power of public housing count increases monotonically with radius and does not plateau. At 100km, the count is effectively measuring "how many public housing properties are in your metro region." This captures the same socioeconomic information as income, education, and family structure data from the Census. At close range (10-200m), the count captures the hyperlocal effect of specific properties on immediate neighbours. Combining both scales provides the most complete picture.
Growth by PH concentration (% of nearby properties that are PH)
Using PH % (proportion of properties within a radius that are public housing) rather than raw count normalises for density:
| PH % within radius | 200m | 500m | 1km | 2km |
|---|---|---|---|---|
| 0% (no PH) | +8.1%/yr | +8.6%/yr | +9.1%/yr | +9.4%/yr |
| 0-5% | +3.6%/yr | +4.4%/yr | +5.3%/yr | +6.1%/yr |
| 5-15% | +4.6%/yr | +4.6%/yr | +4.7%/yr | +5.0%/yr |
| 15-30% | +5.1%/yr | +4.8%/yr | +5.0%/yr | +5.4%/yr |
| 30%+ | +5.0%/yr | +5.2%/yr | +5.6%/yr | +5.9%/yr |
The most consistent finding across all radii: the jump from 0% to any PH is larger than the jump from low to high PH concentration. The 0-5% band shows lower growth than 15-30% because it captures inner-city high-density areas (units near CBD) that grew more slowly across the board during this period, regardless of PH. Controlling for property type and distance to CBD would likely flatten this anomaly. The core finding — that the presence of any PH is linked to roughly 4% slower annual growth — is robust across sample sizes from 283 (10m) to 4,875 (500m).
Maroubra case study: property-level view
Maroubra has 1,388 identified public housing properties concentrated around the Coral Sea Park estate. Properties with no PH within 200m grew 6.5%/yr. Those with any PH within 200m grew 5.4%/yr. The within-suburb variation is the finding that only property-level data reveals.
469 repeat sales and 1,388 PH properties were analysed in this suburb.
National Comparison (2020 to 2022)
| State | Growth: zero PH areas | Growth: 20%+ PH areas | Gap |
|---|---|---|---|
| NSW | +35.7% | +21.7% | 14.0% |
| Victoria | +46.2% | +19.4% | 26.8% |
| Queensland | +41.3% | +29.2% | 12.1% |
| South Australia | +35.0% | +27.9% | 7.1% |
| Western Australia | +25.9% | +9.0% | 16.9% |
Suburb Examples
Bonnyrigg (30% PH, Sydney): $640,000 to $830,000 (+30%). Active redevelopment converting PH to private. Outperforms its PH band by 8%.
Woodridge (high PH, Brisbane): $218,000 to $385,000 (+77%). Affordability demand overwhelmed the PH drag.
Flemington (high PH, Melbourne): $850,000 to $825,000 (-3%). Concentrated towers, no redevelopment. Prices fell while Melbourne rose 46%.
Maroubra (15% PH, Sydney): median grew 19%, but streets near Coral Sea Park estate grew far less. The within-suburb variation is the finding that only property-level data reveals.
Defence Against Criticism
"Wider radii just measure affluent vs non-affluent areas"
Correct. At 50km+, public housing count is a proxy for regional socioeconomic character. But this is precisely the point: public housing concentration at every scale, from next door to city-wide, carries predictive information about growth. The local effect (10-200m) and the regional effect (5-100km) are complementary signals.
"The 10m sample is too small"
The 10m result (4.2% cost) is based on 283 properties with exactly one PH next door. The 100m result (3.8% cost) is based on 1,662 properties. The 500m result (4.0% cost) is based on 4,875. The effect is consistent across sample sizes, which strengthens the finding.
"Correlation is not causation"
We demonstrate a consistent, predictive relationship across five states, multiple radii, and multiple time periods. We do not claim public housing causes slower growth. We claim it is a measurable signal that investors can use.
Limitations
- Repeat-sale analysis only. Properties sold once are excluded.
- Sales data covers 2015-2023. Limited post-2023 data.
- Does not control for renovations, views, or property-specific improvements.
- Tasmania, Northern Territory, and ACT are not included.
Conclusion
Public housing proximity is associated with slower capital growth at every scale tested. A single property next door is associated with 4.2% slower annual growth. Combining public housing counts from multiple radii explains 8.4% of growth variance beyond entry price and hold period.
This analysis is only possible with property-level public housing data. Census area averages cannot distinguish between "next door" and "800 metres away." In our data, that distinction is associated with a difference of about $33,600 per year on an $800,000 home.