There is more than one route, and they run on different clocks

The first mistake in a defect claim is assuming there is one cause of action and one time limit. In New South Wales there are at least three routes to a remedy against the people who built a defective building, and each has its own requirements, its own defendants and its own deadline. In the ACT the structure is similar but the questions are not the same, and an answer carried across from one jurisdiction to the other is often wrong.

The third of those is the one that changed the landscape, and it is dealt with below at some length because it is both the widest route and the least understood.

The statutory warranties in New South Wales

Section 18B of the Home Building Act 1989 (NSW) implies six warranties into every contract to do residential building work. They are implied by force of the section, so they do not depend on the contract saying anything, and they bind the holder of a contractor licence or a person who was required to hold one.

Section 18B(2) is the part that surprises people. The warranties are not limited to a contract with the owner of the land. They are also implied in the contract between a principal contractor who has contracted to do residential building work and a subcontractor engaged to do that work or part of it. The warranties run down the chain, which is why a head contractor facing a defect claim has a route against the trade that did the work.

The New South Wales clock: six years and two years, from completion

Section 18E is short and it decides most claims. Proceedings for breach of a statutory warranty must be commenced before the end of the warranty period, and the warranty period is:

• six years for a breach that results in a major defect in residential building work; and
• two years in any other case.

Both run from completion of the work to which the warranty relates. Where the work was not completed, section 18E(1)(d) fixes the start as the date the contract was terminated, or if it was not terminated, the date work ceased, or if work never started, the date of the contract.

There is one extension and it is easy to miss. Under section 18E(1)(e), where the breach becomes apparent within the last six months of the warranty period, proceedings may be commenced within a further six months after the period ends. A breach becomes apparent when a person entitled to the benefit of the warranty first becomes aware, or ought reasonably to have become aware, of it.

Section 18E(1A) provides a separate extension for strata work. Where a building bond has been lodged under Part 11 of the Strata Schemes Management Act 2015 (NSW), the two-year period is extended until 90 days after the end of the period for the final inspection report on the work.

The ACT splits the same period on a different question

The ACT also has six years and two years, and it divides them on a different test, so the same defect can fall in a different bucket on either side of the border.

Section 88 of the Building Act 2004 (ACT) implies warranties by force of the section into every contract for the sale of a residential building and every contract to carry out residential building work to which the builder is a party: that the work has been or will be carried out in accordance with the Act; in a proper and skilful way and in accordance with the approved plans; using good and proper materials; with reasonable promptness where no completion date is stated; and fit for a purpose or result made known so as to show reliance.

Two features of the ACT provision are worth noting. Section 88(3) provides that each of the owner's successors in title succeeds to the rights, so a purchaser inherits the warranty. And section 88(2A), inserted with the Property Developers Act 2024 (ACT), adds a warranty by the property developer where the developer arranged the work, additional to and not limiting the builder's warranty. That gives an owner a second solvent defendant in the case where it matters most, which is where the builder has gone.

Section 88(4) says the warranties end at the end of the period prescribed by regulation. The Act does not state the period. Section 38 of the Building (General) Regulation 2008 (ACT) does:

The ACT regulation defines a structural element as a load-bearing component of the building, internal or external, that is essential to the stability of the building or part of it, or a component including weatherproofing forming part of the external walls or roof. A non-structural element is anything else.

That definition does real work. A failure of external weatherproofing is a structural element in the ACT by definition, and therefore on the six-year clock, whatever view anyone takes of how serious it is. In New South Wales the same failure has to be argued as a major defect to get the six years.

The New South Wales statutory duty of care, and why it is the wider route

Part 4 of the Design and Building Practitioners Act 2020 (NSW) created a duty that is not confined to residential work and does not depend on any contract. Section 37(1) provides that a person who carries out construction work has a duty to exercise reasonable care to avoid economic loss caused by defects in or related to the building for which the work is done and arising from the construction work.

Four features make it the route to consider first in New South Wales.

Two further provisions close the obvious escapes. Section 39 provides that the duty cannot be delegated. Section 40 provides that Part 4 applies despite any contract to the contrary made after its commencement, and that no contract can annul, vary or exclude a provision of it: there is no contracting out.

Section 41 confirms the relationship with everything else. Part 4 is in addition to the duties, statutory warranties and other obligations under the Home Building Act, other Acts and the common law, and does not limit them. It is a route alongside the warranties, not a replacement for them.

For owners corporations there is a specific provision. Section 38 provides that an owners corporation is taken to suffer economic loss where it bears the cost of rectifying defects, including damage caused by defects, and that the loss includes the reasonable costs of providing alternative accommodation where necessary. It applies whether or not the owners corporation owned the land when the work was done.

Ten years, and then nothing

Both jurisdictions stop everything at ten years, and the longstop does not care when the defect was discovered.

In New South Wales, section 6.20 of the Environmental Planning and Assessment Act 1979 provides that a civil action for loss or damage arising out of or in connection with defective building work or defective subdivision work cannot be brought more than ten years after the date of completion. Completion is the date an occupation certificate is issued authorising occupation, or where no certificate is required, a compliance certificate; failing that, the date of a required final inspection by a certifier; failing that, the date the building was first occupied or used.

In the ACT, section 142 of the Building Act 2004 provides that a building action may not be brought more than ten years after the day a certifier gave the certificate of completion; or where there is none, the day of the last inspection in the course or on completion of the work; or where there is neither, the day the building was first occupied or used. Section 142(3) preserves any shorter limitation period under another territory law, so the ten years is a ceiling and not an entitlement.

The practical point is that the ten-year longstop and the warranty periods are different things doing different jobs. The warranty period can expire long before the longstop, and expiry of the warranty period ends that route whether or not ten years have passed.

In the ACT, each defendant pays only its share

Section 141 of the Building Act 2004 (ACT) applies proportionate liability to a building action. A defendant found liable is liable only for the amount of damages the court considers just having regard to the extent of that defendant's responsibility for the loss, and that liability is limited to the amount for which judgment is given against it even if another Act or rule of law provides otherwise. A defendant held liable for a proportionate part does not contribute to the damages apportioned to anyone else, and does not indemnify them.

The consequence for a plaintiff is that suing the most solvent defendant and leaving it to sort out contribution does not work. If responsibility is shared between a builder, a certifier and a designer, and two of them are gone, the one that remains pays its share and no more. So the identification of every potentially responsible party, and their solvency, is part of the assessment at the start of the matter rather than a detail for later.

What to do when a defect appears

The sequence matters more than the speed.

1. Fix the date of completion. Every clock in this article runs from it, and it is a documented date: the occupation certificate, the certificate of completion, the final inspection, or first occupation. Find it before doing anything else, because it tells you how much time you have and therefore how much deliberation you can afford.
2. Get the defect inspected and recorded. By someone whose report could be used later. Photographs with dates, and an opinion on cause rather than only on appearance, since the cause is what determines who is responsible.
3. Identify everyone who might be liable, not just the builder. In New South Wales the duty of care reaches designers, product suppliers, project managers and anyone with substantive control. In the ACT proportionate liability makes this decisive rather than merely useful.
4. Consider rectification by the original builder. It is frequently the cheapest outcome for an owner, and under a contract's own defects regime it may be a step that has to be offered.
5. Then choose the forum. Tribunal, court, or the contract's own dispute procedure. They differ on cost, speed and whether legal costs are recoverable, and the choice should follow the value of the claim rather than the seriousness of the defect.

Where this article stops

The position stated is as at 27 August 2026, read on that date from the official registers.

Start with what the debtor has, not with what you are owed

The amount of the invoice tells you what is at stake. It tells you nothing about what you will get. The first hour of work on an unpaid debt should go into the debtor, not the debt, because every remedy below costs money and each one is worth using only against a counterparty that can pay.

Four things are worth knowing before anything is sent:

• Is it still trading? A company search shows whether it is under external administration, whether it has been deregistered, and whether its directors have changed recently. A recent change of director on a company that has stopped paying is a signal.
• Has anyone registered against it? A search of the Personal Property Securities Register shows who has security over its assets, and therefore who is ahead of you.
• Is anyone else on the hook? A director's guarantee, a parent company guarantee or a co-obligor turns an empty company into a solvent target. This is in the paperwork you already have, or it is nowhere.
• Are you the only one chasing? Court listings and the notices published by ASIC show whether other creditors have already commenced. If a winding up application is on foot, your position changes completely and you should not be spending money on a demand of your own.

A judgment against a company with no assets is an expensive piece of paper. We would rather establish that in the first week than bill a client to discover it in the sixth month.

The letter of demand still resolves most debts

It is the cheapest step and it has the highest strike rate, which is why it comes first. A demand that sets out the contract, the invoices, the amounts and a date for payment does most of the work in most matters, because a great many non-payments are not disputes at all. They are cash-flow decisions about who gets paid this month, and a debtor deciding that order pays the creditor who looks most likely to escalate.

A demand also improves your position later. A court considering costs looks at how each side behaved before proceedings began, and a creditor who set the case out properly, attached the documents and gave a reasonable time to pay is in a materially better position than one who went straight to a claim form.

Check the security you may already have

Before commencing anything, find out whether you can simply take the goods back. A supplier who sells on retention of title terms, and who has registered on the Personal Property Securities Register, may have a much faster route to value than any proceeding: the goods themselves.

The trap is that the clause alone is not enough. An unregistered interest can be worth nothing at exactly the moment it matters, which is when the customer goes into administration. We have written that half of the subject up separately, and it is the piece most commonly got wrong by suppliers exporting into Australia:

The PPSA trap: what German exporters must know about retention of title.

The general point holds beyond retention of title. Security, a guarantee, a right of set-off or a lien is worth checking before you spend money litigating, because each of them puts you ahead of the creditors who are about to be in the queue with you.

The statutory demand: what it is, and what it is for

A statutory demand is not a debt recovery tool. It is an insolvency tool that happens to recover debts. It does not order anyone to pay, and it produces no judgment. What it does is create a presumption of insolvency that lets you apply to wind the company up, and the prospect of that is what makes companies pay.

It is available only against a company. An individual debtor is a different regime under different legislation and nothing in this section applies to them.

Section 459E of the Corporations Act 2001 (Cth) sets the requirements. The demand may relate to a single debt or to several, each of which must be due and payable, and whose amount or total must be at least the statutory minimum. It must specify the debt and its amount, or the total; require payment, or security, or a composition to the creditor's reasonable satisfaction within the statutory period; be in writing and in the prescribed form; and be signed by or for the creditor. Unless every debt is a judgment debt, the demand must be accompanied by an affidavit verifying that the debt is due and payable.

The Act uses two defined terms and states neither figure. Both are set by the Corporations Regulations 2001 (Cth):

Both provisions carry an alternative figure, $20,000 and six months, for a company eligible for temporary restructuring relief. Those do not apply to any demand served on or after 1 August 2021, so for present purposes they are spent.

What non-compliance actually does

Failure to comply does not make the company liable to pay. It makes the company presumed insolvent. Section 459F(1) provides that where the period for compliance ends and the demand is still in effect and has not been complied with, the company is taken to fail to comply. Section 459C(2)(a) then requires the Court to presume that the company is insolvent if it failed to comply during or after the three months ending on the day the winding up application was made.

That presumption is the whole mechanism. Insolvency is otherwise a matter of proof, and proving it from outside a company is difficult and expensive. The statutory demand converts it into something the company has to disprove.

Section 459C(2) lists other triggers for the same presumption, and they are worth knowing because they may already have happened without your doing anything: execution on a judgment returned wholly or partly unsatisfied, or the appointment of a receiver over property subject to a circulating security interest.

The twenty-one days that cannot be extended

This is the single most consequential deadline in the area, and it catches competent people every year. A company served with a statutory demand that wants to challenge it must apply to set it aside, and section 459G(2) provides that the application "may only be made within" the statutory period after the demand is served.

Section 459G(3) then defines what making the application means, and it is two things, not one. Within that same period, an affidavit supporting the application must be filed with the Court, and a copy of the application and the affidavit must be served on the person who served the demand. Filing without serving is not an application under the section. Neither is serving without filing.

The High Court settled in 1995 that the period cannot be extended. In David Grant & Co Pty Ltd v Westpac Banking Corporation the company argued that the general power in section 1322(4)(d) of the Corporations Law, which allows a court to extend the period for doing any act, could be used to extend the 21 days. The Court rejected it unanimously.

The Court's reasoning was that Part 5.4 is a scheme for resolving solvency quickly, that it contains its own express powers to extend time where Parliament intended them, and that a later and more specific provision attaching a limitation to a particular class of application is not overridden by an earlier general one.

Two practical consequences follow, one for each side. If you have received a statutory demand, the clock started on service and the only safe assumption is that nothing will save you if you miss it. Get advice in the first days, not the third week. If you have served one, the deadline is your advantage and it is worth calculating the date precisely rather than approximately.

There is one piece of relief in the other direction. Where an application is made in accordance with section 459G, the period for complying with the demand runs until seven days after that application is finally determined or otherwise disposed of, and the Court may extend it further. So a company that files and serves in time buys itself the whole of the proceeding.

When a statutory demand is the wrong tool

A statutory demand is for a debt that is not genuinely in dispute. Used against a disputed debt it usually fails, costs the creditor money, and can be characterised as an abuse of process.

Section 459H applies where the Court is satisfied that there is a genuine dispute about the existence or amount of the debt, or that the company has an offsetting claim. The Court then calculates a substantiated amount by subtracting the offsetting total from the admitted total, and the demand is varied or set aside accordingly. The threshold is deliberately low: the company does not have to prove its dispute, only to show that one genuinely exists.

Section 459J provides two further grounds. The Court may set a demand aside where a defect in it will cause substantial injustice unless it is set aside, or where there is some other reason why it should be. Section 459J(2) makes clear that a defect alone is not enough: the Court must not set a demand aside merely because of a defect.

Winding up, and what it actually returns

The application to wind up is the point of the demand, and it is also the point at which most creditors stop wanting it. A liquidator is appointed, the company's assets are realised, and the proceeds are distributed. An unsecured trade creditor is at the back of that queue, behind the secured creditors, behind the costs of the liquidation itself, and behind employee entitlements.

So the honest position is this. Winding up is excellent leverage and it is a poor recovery mechanism. Most statutory demands are paid rather than litigated, and that is where the value sits. Where the company genuinely cannot pay, the application produces a liquidation in which you may recover very little, and you will have funded the exercise for the benefit of every other creditor.

That arithmetic is worth doing before the demand is served rather than after it is ignored, because a creditor who serves a demand and then does not want to follow through has spent money to reveal that it will not escalate.

The money you were paid can be taken back

Being paid is not always the end of the matter. Where the company is later wound up, a liquidator can recover payments the company made to a creditor in the period before the winding up, on the basis that they were an unfair preference.

Section 588FA provides that a transaction is an unfair preference where the company and the creditor are parties, and the transaction results in the creditor receiving more in respect of an unsecured debt than it would have received if the transaction were set aside and it proved for the debt in the winding up. Section 588FE(2) makes such a transaction voidable where it is an insolvent transaction entered into during the six months ending on the relation-back day, or after that day but on or before the day the winding up began.

There is an important qualification in section 588FA(3) for ongoing trading relationships. Where the transactions are, for commercial purposes, an integral part of a continuing business relationship such as a running account, and the level of net indebtedness rises and falls across a series of transactions, all of those transactions are treated as a single transaction. The question then is the net effect across the relationship, not the individual payments. For a supplier who kept trading, that is usually a considerably better position than a payment-by-payment analysis.

The director's exposure is a separate question, and it changes the negotiation

A director who lets a company incur debts while it is insolvent can be personally liable for them. Section 588G applies where a person is a director at the time the company incurs a debt, the company is insolvent at that time or becomes insolvent by incurring it, and there are reasonable grounds at that time for suspecting insolvency.

Section 588GA provides the safe harbour. The liability does not apply where, after the person starts to suspect the company may become or be insolvent, they start developing one or more courses of action reasonably likely to lead to a better outcome for the company, and the debt is incurred in connection with such a course of action or in the ordinary course of business. The protection ends at the earliest of the person ceasing to take that course of action, the course of action ceasing to be reasonably likely to lead to a better outcome, or the appointment of an administrator or liquidator. The note to the subsection records that the director bears the evidential burden.

A creditor does not enforce section 588G directly; a liquidator does. But it matters to a creditor for a practical reason. A director who understands that continuing to trade while insolvent is a personal exposure has a strong reason to deal with your debt rather than let it sit, and a strong reason to take advice early. It is one of the few points at which a company's problem becomes an individual's problem, and it changes how negotiations go.

Unpaid tax carries a further personal exposure for directors, on its own timetable and under different legislation. We have not set out that regime here and it should not be assumed to work the same way.

Where the debtor is overseas, or you are

Cross-border debts change the sequencing rather than the remedies. A foreign supplier owed money by an Australian company has the whole of the above available to it, and the statutory demand in particular is often a surprise to counterparties used to a jurisdiction with no equivalent. It does not require the creditor to be in Australia.

What changes is the groundwork. Service on an overseas party, the governing law and jurisdiction clauses in the contract, whether an Australian judgment will be recognised where the assets are, and whether registration on the Personal Property Securities Register was ever done are all questions that should be answered before a strategy is chosen rather than after.

Where the debt sits under a foreign law or the assets sit in a foreign country, we work with local counsel in that jurisdiction. That is a real limit and it is better stated at the start.

The order, and what each step is for

The sequence is the advice. Almost every mistake in debt recovery is a step taken out of order: a demand sent before anyone checked whether the debtor had assets, a statutory demand used on a disputed debt, a winding up application funded by a creditor who did not want a liquidation.

Steps two and three are cheap and are almost always worth taking. Steps four to six cost real money and each should be a decision made on what the search in step one turned up.