Auditor Liability in Singapore and Australia: Where the Limit Sits

What the Court decided, and what it did not

Hin Leong Trading collapsed in April 2020 amid admitted fraud by its controllers, and its liquidators sued the company's long-standing auditor. The negligence alleged was a failure to detect and report material misstatements across the financial years ended 2014 to 2019. Three heads of loss were claimed: US$2.6 billion in trading losses incurred between November 2015 and mid-April 2020, US$90 million in wrongfully declared dividends for FY2017 and FY2018, and the audit fees themselves.

The Court of Appeal struck out the trading losses, and only the trading losses. The theory behind that head of loss was that a proper audit would have put the company into liquidation earlier and stopped the bleeding. The court held the losses were too remote. The claim for dividends, the claim for fees and the negligence claim itself all continue.

The bench matters to the weight of what follows. The appeal was heard by five members of the court, Menon CJ, Chong JCA, Ang Cheng Hock JCA, Hri Kumar Nair JCA and Kannan Ramesh JAD, with the judgment delivered by Ang Cheng Hock JCA. A five-member court is how Singapore signals that it is settling a question of principle rather than deciding a dispute.

The limit has moved out of duty and into remoteness

The doctrinal move is the part that will outlive the facts. The limiting principle at issue comes from the House of Lords decision in South Australia Asset Management Corporation v York Montague Ltd, reported with Banque Bruxelles Lambert SA v Eagle Star Insurance Co Ltd, and it is easy to state: an adviser who supplies information is responsible for the consequences of that information being wrong, not for every consequence of the course of action the client then takes. The illustration, endlessly repeated, is the mountaineer negligently told his knee is fit to climb, who climbs and is injured in an avalanche.

English law treats that principle as a doctrine with a distinct identity, applying in professional negligence and situations analogous to it. That was confirmed by the UK Supreme Court in Manchester Building Society v Grant Thornton UK LLP, decided in parallel with Khan v Meadows. The Singapore Court of Appeal has now rejected that structure. It held that the principle does not belong to causation at all.

Lord Hoffmann had reasoned that the mountaineer's injury was not caused by the doctor's negligence. The Court of Appeal disagreed: once it is accepted that the mountaineer would not have gone on the expedition but for the negligent advice, the breach plainly caused the injury. What the example actually identifies is the scope of the risk the adviser assumed, which is a question about the extent of responsibility, not about causation.

Having taken the principle out of causation, the court put it into contract. The limit is now answered by the ordinary rule on remoteness of damage in Hadley v Baxendale, asking what was within the adviser's reasonable contemplation at the time of contracting. The purpose of the duty, which the English cases treat as the subject of a separate enquiry, becomes one factor among others in that assessment.

Two boundaries the court drew around its own decision

Two boundaries came with the decision, and both matter to anyone assessing exposure. The court expressly declined to define the duty of care by reference to the specific damage suffered, because that folds remoteness back into duty and is inconsistent with the Spandeck framework, which remains Singapore's test for whether a duty exists at all. And it reserved the position where there is no contract, or where the relationship is merely akin to contract, saying it was unnecessary to express a conclusive view. That gap is live for tort-only claims against advisers, which is the ordinary shape of a claim by someone who was never the adviser's client.

The statutory floor set the contractual expectation

The first reason the trading losses failed was the auditor's distance from the trading. On this the court was blunt, and the passage is the one an adviser should keep.

The second reason was statutory, and it is the more transferable of the two. The court observed that the claim came very close to a claim for wrongful trading, because its gist was that the auditor should have acted earlier to stop the company trading and deepening its insolvency. So it asked what the statutory framework would have led the auditor to expect.

Under the Companies Act (Cap 50, 2006 Rev Ed) as in force when the engagement letters were signed, the wrongful trading offence in section 339(3) reached only an officer who was knowingly a party to contracting a debt the company had no reasonable or probable ground of expecting to be able to pay, and section 340(2) made it civilly actionable only on conviction. Fraudulent trading under section 340(1) required a person knowingly party to carrying on the business with intent to defraud creditors or for a fraudulent purpose. Both require actual knowledge. The court did not even decide whether an auditor is an officer capable of incurring that liability; it assumed the point in the claimant's favour and reasoned from the knowledge threshold.

That is a reasoning route worth noticing beyond auditors. Where Parliament has fixed the mental state on which a particular liability attaches, a negligence claim that would impose the same liability on a lower threshold now has to explain why the adviser would have contemplated assuming it.

Where Australia puts the same limit

Australia did not have to choose a doctrinal home for the limit, because Parliament chose one. Since the 2002 civil liability reforms, causation has been split by statute into a factual question and a normative one, and the limit lives in the normative limb.

Section 5D(1)(a) is entirely factual and does nothing but apply the but-for test. Section 5D(1)(b) is entirely normative, and section 5D(4) directs the court, in determining the scope of liability, to consider whether or not and why responsibility for the harm should be imposed on the negligent party. The label differs from Singapore's and the work is the same: filtering out losses that are factually caused yet should not be compensated.

The Australian Capital Territory provision is section 45 of the Civil Law (Wrongs) Act 2002, and although it is the same rule the drafting is not identical. It speaks of a decision rather than a determination, of the happening rather than the occurrence of the harm, and it uses a colon and semicolons where the New South Wales provision uses a dash. Section 45(3) carries the equivalent of section 5D(4). Quote the provision of the jurisdiction you are in, because a court reading a submission that quotes the other one will notice.

Australia absorbed the risk principle rather than rejecting it

It is often assumed that Australia turned its back on the English scope of duty cases. It did not. The High Court has placed the risk principle inside the statutory scope of liability limb, and adopted the mountaineer example while doing so.

The footnotes to that passage cite Banque Bruxelles at 213 and at 214, and the High Court's own decision in Kenny & Good Pty Ltd v MGICA (1992) Ltd. The English authority was not distinguished. It was adopted, and relocated.

The paragraph immediately before it is the one that matters most in practice, because it tells an Australian practitioner what a court must actually do with section 5D(1)(b) in a case that is not covered by precedent.

Wallace v Kam remains current. It was applied by the Full Federal Court, on these very paragraphs, in Hassan v Minister for Home Affairs on 22 April 2025.

Kenny & Good, and the sentence that cuts both ways

Kenny & Good is where the Australian and Singaporean routes almost meet, and it needs stating carefully. A valuer negligently overvalued a Hunters Hill property, a lender advanced 65 per cent of the valuation, the borrower defaulted, and the loss was deepened by a falling market. The High Court, sitting as Gaudron, McHugh, Gummow, Kirby and Callinan JJ, dismissed the valuer's appeal with costs on 17 June 1999, so the valuer answered for the whole loss. Their Honours gave separate reasons.

McHugh J took the route Singapore has now taken. He identified the governing principle as the rule formulated by Alderson B in Hadley v Baxendale, held that an adviser is liable only for losses flowing naturally from the breach or within reasonable contemplation, and said that as a general rule those do not include the consequences of market declines. His statement of where the question belongs could have been written in 2026.

On the facts the valuer still lost, and the reason is worth getting right because it is easy to overstate. The valuation report recommended the property as suitable security for the investment of trust funds to the extent of 65 per cent of the valuation for a term of three to five years. McHugh J treated the valuation as having represented that the property would have sufficient value to enable a lender to recover the sum advanced and interest at any time during the next five years, so the loss flowed directly and naturally from it. That was his Honour's reasoning, not a warranty claim. The lender never sued to obtain the benefit of that statement as a contractual warranty and never complained of the representation embodied in it. And Gaudron J read the same clause the other way.

So the drafting lesson runs in both directions, and that is the useful part. A sentence about future suitability can extend an adviser's exposure by bringing later losses within contemplation, and the same sentence can confine it by fixing a period and a percentage beyond which the adviser did not undertake anything. Which way it runs depends on how it is written. What it cannot be is neutral.

The quiet finding in this comparison is that the path the Singapore Court of Appeal adopted in 2026, routing the limit through Hadley v Baxendale, is a path a member of the High Court of Australia mapped in 1999.

The Australian auditor case, and the reasoning Singapore dismantled

On the specific question of an auditor's liability for a company's continued trading losses, the leading Australian authority is a 1987 decision of the New South Wales Court of Appeal. In Alexander v Cambridge Credit Corporation Ltd the auditor was negligent, the company traded on from 1971 to 1974, and it claimed roughly $145 million, being the deterioration in its net asset position over that period. The court divided two to one against the company. Mahoney JA held that although the loss was in the broadest sense a result of the breach, that did not suffice: allowing the company to remain in existence exposed it to all of the dangers of being in existence, which is not the same as causing the losses. McHugh JA rested his conclusion on practical common sense. Glass JA dissented. The English Court of Appeal followed the same route in Galoo Ltd v Bright Grahame Murray.

The outcome in Alexander is the outcome in Hin Leong. The reasoning did not survive. The Singapore Court of Appeal gave three reasons for rejecting common sense causation: it is analytically indeterminate, it conceals the real reasoning, and it commits a category error. On the first it pointed out that Alexander proves the objection, because the dissenting judge thought common sense pointed the other way.

Then it used McHugh J's own change of mind against the case he had decided. Less than five years after Alexander, sitting in the High Court in March v E & M H Stramare Pty Ltd, McHugh J said that reliance on labels of this kind obscured the reality that the court was making a policy judgment about the extent of a defendant's responsibility.

For an Australian practitioner the position is comfortable in result and exposed in reasoning. The outcome in Alexander is orthodox and is unlikely to be disturbed. Its ratio rests on a mode of reasoning that its own author qualified within five years, that the High Court has since said will ordinarily be of limited utility unless it can be unpacked and explained, and that a five-member Singapore Court of Appeal has now analysed and rejected at length. A modern Australian claim against an auditor for trading losses should be run through section 5D(1)(b) with the normative reasoning stated openly, not through an appeal to common sense.

The case that helps claimants, and the wound test

There is one recent English decision holding an auditor liable for a company's continuing losses. In AssetCo plc v Grant Thornton UK LLP the auditor failed to detect that the business was sustainable only on the strength of dishonest representations, and was held liable for losses flowing from two loss-making contracts. The liquidators in Hin Leong relied on it. The court rejected the analogy on two grounds, and the distinction it drew is the most useful part of the judgment for anyone assessing exposure.

The test that emerges is whether the undetected problem is a continuing source of loss. Where the negligence leaves an identifiable, ongoing haemorrhage, such as a loss-making contract that would have been terminated, the resulting losses can fall within the adviser's responsibility. Where the losses are discrete year-on-year outcomes driven by market movements and management decisions, they do not. In Hin Leong each year's losses came from new trades entered into in that year, and the company had in fact turned a profit from derivatives trading in two of the six years in question.

The second ground was a pleading point, and it is the one to take away. The liquidators had pleaded that derivatives trading was inherently complex and volatile, which the court read as a plea that the trading was capable of generating large losses. They had not pleaded that it was likely to. That distinction defeated the analogy on the pleadings alone, without any evidence and without any finding about the trading.

Two Australian features with no Singapore counterpart

A claim against an auditor for economic loss caused by a failure to take reasonable care is ordinarily an apportionable claim. Part 4 of the Civil Liability Act 2002 (NSW) applies to a claim for economic loss in an action for damages, whether in contract, tort or otherwise, arising from a failure to take reasonable care. Where it applies, section 35(1)(a) limits a concurrent wrongdoer's liability to the proportion of the loss that the court considers just having regard to the extent of that defendant's responsibility.

That machinery reaches the auditor cases squarely, because the other wrongdoers in them are usually the people who committed the fraud. In Hunt & Hunt Lawyers v Mitchell Morgan Nominees Pty Ltd the High Court held, by majority, that solicitors who had negligently drafted a mortgage were concurrent wrongdoers together with the fraudsters who had induced the lender to advance the money, so the solicitors' liability was apportioned rather than falling on them in full.

The reverse does not follow, and this is where the point is usually stated too loosely. Section 34A denies the benefit of apportionment to a concurrent wrongdoer who intended to cause the loss or who caused it fraudulently. So the fraudster answers in full while the negligent adviser answers for a share. Professional standards legislation can cap an auditor's exposure further where the firm is a member of an approved scheme.

Neither feature has a Singapore counterpart in this context. A loss that survives the remoteness limit remains recoverable from the auditor in full. The practical consequence is that the strike-out application does more work in Singapore than it does here, because it is the main instrument for cutting down a very large head of loss, and there is no apportionment waiting downstream to do it instead.

What this changes in your engagement letters

The engagement letter now does more work in Singapore than it did. If the limit is fixed by what was within the adviser's reasonable contemplation at the time of contracting, the recitals describing the purpose and scope of the engagement become the primary evidence of that contemplation. They are no longer administrative front matter. They are the document the remoteness argument will be run on.

How the pleading differs, and your review checklist

Pleading practice diverges between the two jurisdictions. In Singapore a defendant adviser pleads remoteness, and a claimant must plead the facts that put the loss within contemplation at the time of contracting. In New South Wales and the Australian Capital Territory the same argument is a scope of liability point under section 5D(1)(b), supported by reasons of legal policy, and in a novel case the court is obliged to explain the answer in those terms. Running a Singapore claim on Australian statutory language, or an Australian claim on Hadley v Baxendale alone, will misfire.

The strike-out risk for large consequential heads of loss has increased. Hin Leong confirms that a court will determine this limit summarily where the facts can be assumed in the claimant's favour. A claimant who cannot articulate why the specific loss was within the adviser's contemplation should expect that head of loss to be tested early rather than at trial.

For an adviser, the practical reading of Hin Leong is not that auditors have been given an immunity. It is that the largest heads of loss in these claims now have to be justified at the pleading stage by reference to what the engagement actually was. For a company pursuing an adviser, the reading is that the engagement letter it signed years ago will be doing more of the work than the negligence it can prove.

This article is provided for general information purposes only and does not constitute legal advice. Specialised legal counsel should be sought for specific fact patterns.