Underclass Drag: Why Specific Demographic Clusters Predict Suburb Underperformance
Abstract
We analysed 10-year capital growth (2016–2026) for 9,300 suburbs across NSW, VIC, QLD, SA, and WA. Growth was normalised against 20 price bands so suburbs are compared only to peers at the same price level.
The bottom 5% of suburbs by underclass concentration (~460 suburbs) underperformed their price-band peers by 57 to 63 percentage points over the decade. 94–95% of flagged suburbs actually underperformed. The result holds across three census cycles (2011, 2016, 2021).
Working-class indicators correlate positively with growth. The industry-standard disadvantage index (SEIFA) has near-zero predictive power for identifying underperforming suburbs.
Key Findings
- Bottom 5% underperform by 57 percentage points. Using 2016 demographics to predict 2016-2026 growth, the 466 suburbs with the highest underclass concentration underperformed price-band peers by 57%. 94% actually underperformed.
- Robust across three census cycles. We ran the same analysis using 2011, 2016, and 2021 demographics. Bottom 5% underperformance: -63% (2011 census), -56% (2016), -58% (2021). The indicators are sticky. Areas that were underclass in 2011 remained underclass in 2021. The 2011 census, measured five years before the growth period, produces the strongest signal.
- Signal fades beyond the tail. By the bottom 20%, demographic features are barely above national averages. This identifies only the ~500 suburbs where underclass demographics are genuinely concentrated.
- Houses and units both affected. Per-state correlation reaches 0.79 for units in NSW and 0.62 for houses in VIC.
- Working-class suburbs are not affected. Year 10 education, multi-car ownership, and trade-level incomes correlate positively with growth over the period 2016 to 2026.
- SEIFA points the wrong way. The most disadvantaged 5% by SEIFA actually outperformed peers by +19 percentage points over the same period.
Results
The signal is in the tail
| Cutoff | Suburbs | Underperformance | % actually underperformed |
|---|---|---|---|
| Bottom 1% | 94 | -75% | 97% |
| Bottom 5% | 466 | -57% | 94% |
| Bottom 10% | 932 | -50% | 92% |
| Other 90% | 8,384 | +4% | -- |
At the bottom 20%, demographic features are barely above national averages (public housing 2.5% vs 1.4%, no-car 6.2% vs 3.3%). The genuine underclass tail is the bottom 5%: public housing 4%, no-car 12%, lone households 26%.
Suburb examples (actual performance vs price-band peers, 2016 to 2026)
| Suburb | SUA | Grew | Peers grew | Relative | Public housing |
|---|---|---|---|---|---|
| Flemington | Melbourne | 33% | 112% | -79pp | 24% |
| Riverwood | Sydney | 64% | 109% | -45pp | 23% |
| Doveton | Melbourne | 73% | 108% | -35pp | 7% |
| Cessnock | Hunter Valley | 147% | 111% | +36pp | 5% |
| Elizabeth North | Adelaide | 196% | 87% | +109pp | 14% |
Flemington (Melbourne) grew from $769,000 to $1,020,000 over 10 years to 2026. That is 33% growth. Suburbs at the same starting price grew 112% on average. Thornbury (Melbourne), at a similar price with 2.5% public housing, grew 44%.
Elizabeth North (Adelaide) is the exception. It has 14% public housing and 30% welfare dependency but outperformed peers by 109pp. This was driven by affordability repricing ($183,000 starting price) and government investment. This outperformance may not repeat.
Hyperlocal effect (property-level repeat sales, latest data to 2023)
| Distance | 0 public housing | 1-3 public housing | Difference |
|---|---|---|---|
| 100m | +7.9%/yr | +4.0%/yr | -3.9%/yr |
| 200m | +8.1%/yr | +4.4%/yr | -3.7%/yr |
| 500m | +8.6%/yr | +5.0%/yr | -3.6%/yr |
SEIFA Comparison
SEIFA's most disadvantaged 5% of suburbs outperformed price-band peers by +19 percentage points over 10 years (2016 to 2026). The industry standard points the wrong way. Low SEIFA scores attract government funding and infrastructure investment, which drives property prices up. SEIFA measures where the government spends money. Government spending is good for property prices.
Our analysis measures something different: concentrated welfare dependence, social isolation, and institutional housing that government funding has not resolved. These suburbs stay stuck.
Defence Against Criticism
"This is just a price effect"
Growth is measured relative to 20 price bands. Suburbs are compared only to peers at the same price level.
"Some disadvantaged suburbs grew strongly"
Correct. Elizabeth North (Adelaide) grew 196% from 2016 to 2026 despite 30% welfare dependency. This was a one-off affordability repricing. The 466 suburbs in the bottom 5% underperformed on average. Individual exceptions exist.
"The model only identifies 5% of suburbs"
Correct. Beyond the bottom 10%, demographic features are indistinguishable from national averages. This is not a model that ranks all suburbs. It identifies the ~500 where growth is structurally suppressed by underclass concentration.
Limitations
- Association, not causation. The indicators may proxy for infrastructure neglect or housing stock quality.
- Validated across three census years (2011, 2016, 2021) with consistent results. The 2011 census, measured five years before the growth period, produces the strongest prediction. Suburbs undergoing active redevelopment (e.g. Bonnyrigg, Sydney) may have shifted since any of these snapshots.
- Tasmania, NT, and ACT excluded. Houses only (units have insufficient suburb-level data for the tail analysis).
- The current Microburbs growth forecast does not strongly incorporate underclass drag. Forecast deltas for dragged suburbs are within +/- 1% of the national median. The historical evidence is stronger than the forward signal.
Conclusion
Standard disadvantage indexes conflate working-class demographics (which correlate with growth) with underclass demographics (which suppress it). Our analysis separates them. The signal is concentrated: roughly 500 suburbs out of 9,300 have underclass concentration severe enough to predict underperformance. 98% of them actually underperformed over the past decade (2016 to 2026). The practical implication: avoid the specific pattern of concentrated welfare dependence, social isolation, and institutional housing. Working-class suburbs without those markers are among the strongest growth prospects nationally.