NSW building defects by the numbers: causes, accountability failures and the evidence gap

Research finding: The public evidence supports a substantial NSW residential-building dispute problem and an underreporting problem among surveyed apartment buildings. It does not currently establish how frequently developer SPVs lead to unrecovered losses, the average legal bill for new-build purchasers, or the percentage of detached new homes with serious defects. Our findings are therefore organized as verified statistics, documented mechanisms, and research questions still requiring case-level verification.

The numbers at a glance

Primary references: NSW Parliament, NCAT 2023–24, NCAT 2024–25, Building Commission 2025 survey, Building Commission dispute figures, IPART HBC review.

1. What exactly counts as a case?

“Defect”, “complaint”, “NCAT application”, “published judgment” and “loss paid by the homeowner” are not interchangeable.

A single building can have multiple defects. One building owner can make more than one complaint. One dispute can produce multiple tribunal decisions (liability, costs, appeal, enforcement). A published judgment may follow years of private discussions. Some matters are settled or withdrawn without a published decision. In the other direction, a tribunal application may concern unpaid progress payments rather than defective work.

Recommendation: Use the underlying residential property dispute as the unit for case-level research, deduplicate associated judgments, and report application counts separately from owners, properties and outcomes.

2. The reporting iceberg: what the data can and cannot show

The 2025 strata research surveyed 520 eligible buildings through strata managers and some self-managed schemes. It reports serious defects in 53% of responding buildings and regulatory reporting by 36% of buildings with serious defects. The 2021 government release about the first survey named cost and lack of awareness as leading reporting barriers.

Expressed intuitively, of 100 surveyed buildings with serious defects, approximately 36 reported to the regulator and 64 did not. This is an internal survey ratio, not a claim that there are 64 hidden defective houses for every 36 reported cases. The denominator is already restricted to buildings identified with serious defects; detached houses were not the study population.

The 2025 research also noted that newer registered buildings in its sample showed lower serious-defect rates than older ones. This matters because alarmist coverage can ignore improvements and treat every building cohort as equally exposed. Building Commission.

What follows: Regulatory reporting misses many serious defects in this surveyed apartment population. What does not follow: Every unreported defect was concealed, unrepaired, uninsured or personally funded by owners.

3. Why the financial cost is especially hard to measure

A defect can generate several distinct costs: identifying it, making it safe, repairing it, proving who is responsible, obtaining a legal remedy and actually recovering payment. Published decisions rarely provide a complete total.

IPART's 2020 review described the potential cost of engineers and legal advice as thousands or tens of thousands of dollars. It reported almost nine months on average for NCAT matters above $30,000; it also observed that homeowners may have to pursue builders who do not comply with orders and then go through additional insurance processes. The timeframe is historical and not an appropriate “average time to settle a 2026 new-build case.” IPART.

A valid future study should record four different dates (first notification, formal dispute, decision, evidence of recovery) and four different monetary amounts (claimed defects, amount ordered, actual legal/expert spend, payment or completed rectification confirmed). Otherwise averages can be meaningless.

4. A causal map: how a construction problem becomes a homeowner loss

Defect or approval issue → detection after sale → responsibility contested → evidence needed → failed informal resolution → formal dispute → order or settlement → payment or rectification → residual loss.

At any point, the chain can fail. For example, the homeowner may run out of time to bring a statutory warranty claim, the correct respondent may be disputed, expert estimates may diverge, the seller may have no assets, or the insurance contract may not yet be triggered.

This is why financial recoverability should be treated as a separate risk dimension from construction quality.

The seven risk-factor families

These factors are candidate mechanisms. Their frequency among developer-resale cases remains unmeasured until judgments and company records are systematically coded.

5. Is an SPV the main problem?

Not proven. An SPV is a common legitimate structure, and corporate limited liability is not unique to property. The deeper question is whether a project's continuing financial obligations are supported by an entity with funds or meaningful security after the sale.

For a defensible SPV statistic, a reviewer must establish company history, ownership and whether the entity was set up for one development. Do not mark a defendant as an SPV merely because its name contains “Developments Pty Ltd”. An insolvent builder is also not necessarily the developer's SPV.

The more actionable question for consumers is: “Who bears the legal obligation, and what assets, guarantees or insurance stand behind it?” Read the accountability guide.

6. What can real cases teach us already?

NCAT's official 2024–25 annual report summarizes two home-building judgments that illustrate different failure mechanisms. They are not a representative sample and neither involved a verified project SPV.

Sader v Renbar Constructions Pty Ltd [2025] NSWCATCD 47. The owners sought $222,315.46 for alleged later-identified defects after earlier Supreme Court proceedings about the construction. NCAT dismissed the new claim because it found the owners knew or reasonably should have known about the defects in the earlier proceedings. It nevertheless considered the merits in the alternative and indicated a waterproofing allegation would otherwise have succeeded. Lesson: knowing about a defect during earlier litigation can have major consequences for later proceedings. The claimed amount is not an award or a proven actual loss.

Fish v Kurmond Homes Pty Ltd; Kurmond Homes Pty Ltd v Polycrete Australia Pty Ltd [2025] NSWCATCD 6. The owner claimed $415,622 concerning finishes and other alleged defects. The tribunal rejected some allegations and awarded $55,999.19 for damage to the polished concrete surface. Lesson: asserted remediation costs can differ sharply from the amount legally established, and responsibility may involve builder and third-party work arrangements.

Both are commissioned-build owner disputes, not examples of unsuspecting purchasers buying completed developer homes. They belong in a comparison group rather than being counted in the target “subsequent purchaser” cohort. Both summaries are recorded in the official NCAT Annual Report 2024–25, pages 41–43. See research case summaries.

7. The quantitative research we still need

A 40-case sample should deliberately separate developer-built properties sold to purchasers (primary cohort) from owner-commissioned builds (comparison cohort) and apartments (context cohort).

Each underlying dispute needs a decision citation, property type, purchaser type, parties' legal identities, known construction and approval facts, insurance issue, disposition, claim amount, remedy, legal costs (if stated), procedural duration and post-order recovery evidence.

The proper denominators are cases with enough evidence to assess the factor. Example: if 10 of 40 cases have sufficient company records to determine SPV status, “four confirmed SPVs out of ten assessable” must not be published as “40% of cases are SPVs”, let alone “40% of NSW developers use SPVs to avoid liability.” Publication should show 4/10 assessable; 30/40 unknown and explain selection bias.

Do not estimate a mean legal fee using only the handful of judgments disclosing costs without prominently noting that disclosure is selective. A more defensible output might be median known cost among disclosed cases, number of observations, and a large uncertainty statement.

Full study methodology and coding rules.

8. What would actually establish an industry-wide financial accountability failure?

We would look for repeated, independently verified combinations of:

1. Documented defect or non-compliance discovered after acquisition.
2. Clearly identified legal obligation owed to the purchaser.
3. Builder/developer refusal, disputed identity or incapacity.
4. Formal order or settlement or supported abandoned claim.
5. Evidence of unrecovered costs, unsuccessful enforcement, inability to fund repairs or an economically rational decision not to litigate.

The public decisions currently support the existence of several mechanisms, not a trustworthy statewide frequency.

A genuinely useful investigation should include successful remedies and counterexamples as well as failures. Building Commission NSW reports that roughly 2,500 disputes are assisted annually and that over 80% are resolved without enforcement action. Those are the regulator's service figures and may reflect a different set of cases and outcomes from NCAT counts. Building Commission.

Frequently asked questions

How many home-building cases went to NCAT in 2024–25?

NCAT recorded 2,409 home-building applications and 2,161 finalizations in that financial year. These include a range of home-building disputes and are not an estimate of defective new homes sold.

What percentage of NSW new-build houses have defects?

A reliable statewide percentage for newly built detached homes sold by developers has not been established from the cited public evidence. The 53% figure is from a defined 2025 survey of apartment buildings and should not be generalized.

How many NSW building disputes involve SPVs?

There is no verified percentage in this report. Case-specific ASIC research is needed to determine project-company structure and distinguish ordinary corporate organization from demonstrable limits on recovery.

How long does NCAT take to resolve building disputes?

IPART reported almost nine months on average for cases above $30,000 using 2020 data. That is historical NCAT resolution time, not the time between discovering a defect and receiving money or completed repairs.

What is the average legal cost of suing a NSW builder?

There is no reliable average established here. IPART described expert and legal expenses of several thousand or tens of thousands of dollars in some disputes. Published decisions often omit actual legal spend.

Does a successful tribunal order prove the homeowner was compensated?

No. Money orders can require registration and enforcement, and repair orders can require follow-up. A study must distinguish a favorable order from confirmed actual payment or completed work.

Do published NCAT judgments represent all building disputes?

No. They are selected published outcomes, and the underlying dispute population includes private resolutions, withdrawals and matters that do not result in reported reasons. Using published cases alone introduces selection bias.

Further reading

• The hidden-cost iceberg: practical pillar article
• Developer SPVs and accountability
• Builder refuses to rectify: rights and enforcement
• Methodology and source register
• Verified case studies from NCAT reporting

Research integrity statement: this is a source-grounded analytical article, not a completed 30–50-case statistical study. Dates, denominators, competing explanations and unknowns are disclosed. It does not allege wrongdoing merely from use of SPVs, insolvency, or the existence of litigation.