Estimating Strata Levies from Listing Descriptions
Abstract
We present a method for estimating quarterly strata levies for Australian apartments and townhouses using information extracted from property listing descriptions. From 6.8 million listing descriptions, we identified 1.66 million observations of strata levy amounts across 111,100 unique properties, then filtered to a high-confidence subset of 49,234 properties in 19,270 buildings. We only publish a building where multiple independent listings from different agents or time periods report consistent levy amounts (within 10% of each other). Strata levies are primarily a building-level attribute, not a unit-level one: within a qualifying building, different units typically report similar quarterly amounts. For buildings without agent-disclosed data, a separate model estimates the levy based on location, building type, and amenities.
Key Findings
- 49,234 properties across 2,119 suburbs have high-confidence levy estimates, covering all eight states and territories.
- The typical Australian apartment pays $774 per quarter ($3,096 per year) in strata levies. This ranges from $288/quarter in South Australia to $1,269 in the Northern Territory.
- Strata levies consume about 16% of gross rental income nationally. In the Northern Territory, this rises to 25%. In South Australia, it is just 7%.
- Building amenities are the strongest driver of levy variation. Each additional amenity (pool, gym, concierge, sauna) adds roughly $80 to $100 per quarter. A building with seven or more amenities pays about 80% more than one with none.
- Strata levies do not predict capital growth. Across 702 suburbs with both levy and sale price data from 2020 to 2025, there is no meaningful relationship.
- Victorian terminology differs. Victorian agents use “Owners Corporation Fees” rather than “strata levies” or “body corporate fees”. Our extraction handles both terms, though Victorian coverage remains lower than other states.
Methodology
Data collection
We analysed 6.8 million property listing descriptions from a comprehensive Australian property database. The listings cover all states and territories and span from 2019 to early 2026. We focused on strata-eligible property types: apartments, units, townhouses, villas, studios, and duplexes.
Levy extraction
Strata levy amounts were identified in listing descriptions using structured text analysis. Australian agents commonly include levy information in standard formats, and we convert all amounts to a consistent quarterly basis.
To mitigate extraction noise, we use the median of all extractions for each property rather than the most recent, and exclude properties where repeated extractions show high internal variation.
Building identification
Properties were grouped into buildings using address-based matching. This approach identifies over 43,000 unique buildings, of which more than 27,000 contain two or more observed units.
Quality filtering
We apply multiple quality filters before publishing an estimate, including outlier detection at the building level and noise removal for inconsistent properties. Only buildings that pass all filters are included in the published dataset.
Results
National coverage
| State | Properties | Typical quarterly levy | Annual equivalent | % of gross rent | Note |
|---|---|---|---|---|---|
| NSW | 24,064 | $854 | $3,416 | ~16% | |
| QLD | 11,514 | $800 | $3,200 | ~15% | |
| WA | 6,837 | $634 | $2,536 | ~17% | |
| SA | 2,419 | $288 | $1,152 | ~7% | |
| ACT | 1,634 | $752 | $3,008 | ~16% | |
| VIC | 1,257 | $612 | $2,448 | ~8% | Low coverage |
| NT | 1,200 | $1,269 | $5,076 | ~25% | Small sample |
| TAS | 309 | $375 | $1,500 | ~8% | Small sample |
Internal consistency
We only publish buildings where independent listings agree. Within the published dataset, different units in the same building report levy amounts that are almost identical. This is not a prediction tested against official strata records. It is a measure of how consistently agents disclose the same figure for the same building, across different listings, agents, and years.
Accuracy varies by building type. Small, homogeneous blocks (2 to 20 units, no mixed property types) achieve the highest accuracy. Large towers with diverse unit types (studios through to penthouses) are less predictable because the building median does not account for differences in unit size and lot entitlement.
Concrete examples
To illustrate the level of precision achievable, consider two buildings from the published dataset:
128 Charlotte Street, Brisbane City (QLD): 31 units observed, typical levy $1,819 per quarter ($7,275 per year). Most units pay between $1,577 and $2,139. The building has a pool and gym.
160 Rundle Mall, Adelaide (SA): 17 units observed, typical levy $166 per quarter ($662 per year). Most units pay between $142 and $212. For comparison, the typical Sydney (NSW) apartment costs $854 per quarter, roughly five times more.
Amenity premiums
We extracted building amenity information from listing descriptions and measured the median levy for buildings with and without each amenity.
| Amenity | Typical levy with amenity | Premium over no-amenity buildings ($760/q) |
|---|---|---|
| Sauna | $1,380/q | +$620 (+82%) |
| Concierge | $1,267/q | +$507 (+67%) |
| Lift | $1,230/q | +$470 (+62%) |
| Gym | $1,194/q | +$434 (+57%) |
| Pool | $1,129/q | +$369 (+49%) |
Each additional amenity adds approximately $80 to $100 per quarter. A building with no detected amenities (typically a small walk-up block) pays around $760 per quarter. A building with seven or more amenities pays around $1,368 per quarter, an 80% premium.
Strata levies and capital growth
We tested whether suburbs with higher or lower strata levies experienced different rates of unit price growth from 2020 to 2025. Across 702 suburbs with sufficient data, we found no meaningful relationship. Suburbs with the lowest levies slightly outperformed those with the highest, but this appeared to reflect differences in property type and location rather than a causal levy effect.
Defence Against Criticism
“The data comes from what agents choose to disclose”
True. Our dataset reflects buildings where agents included levy information in listing descriptions. This means the sample is not representative of all strata properties. However, for the properties we do cover, the estimates are internally validated: the same building produces consistent levy amounts across multiple independent listings by different agents at different times. The consistency of these independent observations is the basis for our confidence thresholds.
“How reliable is the extraction?”
Individual extractions carry some noise. Our production estimates use the median of all extractions for each property and exclude properties with high internal variation. After these cleaning steps, the vast majority of properties with multiple listings produce consistent levy amounts across their listing history. The remaining inconsistent properties are excluded from the published dataset.
“Within-building variation means unit-level estimates are unreliable”
Correct for large, diverse buildings. We address this by only publishing estimates for buildings where the internal variation is below 10%. For these buildings, every unit pays a similar amount regardless of size or floor level. For buildings with high variation (typically large towers with mixed unit types), we do not publish a point estimate.
Limitations
- No ground truth validation. All accuracy metrics are measured against our own extracted levy amounts, not against actual levy notices. Independent validation against official strata documents would strengthen these findings.
- Coverage is uneven. NSW accounts for 49% of the dataset. Victoria, despite being the second-largest apartment market, accounts for only 3%. Coverage depends on agent disclosure practices, which vary by state and agency.
- Temporal currency. Strata levies increase over time (approximately 2% per year based on same-building observations). Estimates based on older listings may understate current levies.
- No causal claims. Amenity premiums describe associations, not causation. A building with a pool pays more in strata, but the pool is not necessarily the cause of the higher levy. Other correlated factors (building age, height, location) may contribute.
Conclusion
Strata levies are a significant and often underestimated holding cost for apartment investors. The typical Australian apartment pays $3,096 per year in strata levies, consuming about 16% of gross rental income. This varies from $1,152 per year in South Australia to $5,076 in the Northern Territory.
By extracting levy information from listing descriptions and cross-validating within buildings, we can provide indicative estimates for 49,234 properties across 2,119 suburbs. These are based on what agents disclosed, not official strata records, and should be treated as a screening tool rather than a substitute for formal due diligence.
This data is now available on Microburbs property and suburb reports, giving buyers and investors an early indication of strata costs before they commit to an inspection or due diligence process.