A plaintiff who proves that a defendant breached a legal duty does not automatically succeed. They must also prove that the breach caused their loss. It is here that many claims fail. The primary tool for establishing that causal connection in Australian law is the “but for” test. Under Australia’s civil liability legislation, the causation inquiry has two distinct elements: factual causation and scope of liability. Each deserves careful attention.
What is the two-part test for causation under Australian civil liability law?
Across Australia, the civil liability legislation enacted following the Ipp Report codified causation as a two-part inquiry. To establish that a breach of duty caused particular harm, two requirements must be satisfied:
- that the breach was a necessary condition of the occurrence of the harm (factual causation); and
- that it is appropriate for the scope of the defendant’s liability to extend to the harm so caused (scope of liability).1
The first limb, factual causation, is addressed by the “but for” test. The second limb, scope of liability, is a normative question that asks whether it is appropriate, as a matter of policy and principle, to hold the defendant legally responsible for the harm that factually resulted. Both limbs must be satisfied.
The counterfactual approach to factual causation
The “but for” test asks a deceptively simple question: would the plaintiff have suffered the harm but for the defendant’s breach? If the answer is yes (the harm would have occurred regardless), causation is not established and the claim fails.
The most authoritative modern formulation of the test was given by Edelman J in Lewis v Australian Capital Territory2. The “but for” or counterfactual approach “directs us to change one thing at a time and see if the outcome changes.” The change is the removal of the wrongful act. If the loss would lawfully have occurred without that act, the breach was not necessary for the loss and factual causation is not established. Only the wrongful conduct is removed from the counterfactual; all other facts and circumstances remain constant.
In Lewis v ACT, Mr Lewis had been unlawfully imprisoned for 82 days. His imprisonment was, however, inevitable: even if the defendant had acted lawfully, he would have been imprisoned for the same period. Because the breach made no difference to the outcome, it was not a cause of the loss, and the plaintiff was entitled only to nominal damages.
The Victorian Court of Appeal applied this approach in Mulcahy & Co Accounting Services Pty Ltd v Porter3, a breach of contract claim arising from a breach of confidentiality. Mr Porter had made a confidential offer to acquire a 70 per cent interest in a business known as CBB. His accountant, Mr Mulcahy, owed him ongoing confidentiality obligations but used the form and contents of that offer when preparing a competing bid. The primary judge found that misuse of the letter “was a cause” of Mr Porter’s loss of opportunity to acquire a 40 per cent interest valued at $4.8 million.
The Court of Appeal held that the reasoning was insufficient, on two independent grounds. First, the judge had not positively found that Mulcahy received the Porter offer before the critical 18 December 2017 meeting, having expressly acknowledged it may have arrived after, and the deal was found to have been done at that meeting. Second, and in the alternative, by that date the consortium already possessed from other sources all the commercially significant information needed to make their competing bid, including the independent valuation of the business at $21 million against a purchase price based on $12 million. The breach therefore made no material difference to the outcome.
The Court also confirmed that the traditional “common sense” formulation of causation has been subject to sustained criticism and that the structured “but for” inquiry is now the starting point.
Hypothetical conduct: what would the plaintiff have done?
Factual causation often turns on a hypothetical question: what would the plaintiff have done if the defendant had not breached its duty? This question arises wherever the chain of causation passes through a choice that the plaintiff would have made. A common example is a failure to advise or warn: would the plaintiff have acted differently had the advice or warning been given?
The civil liability legislation in most jurisdictions addresses this question with a two-part rule. Where it is relevant to factual causation to determine what the plaintiff would have done in the hypothetical, the matter is to be determined subjectively in light of all relevant circumstances.4 Certain civil liability Acts contain an additional provision that any statement made by the plaintiff after suffering the harm about what they would have done is inadmissible unless it is against their own interest.5
To the extent that civil liability legislation does not contain such a provision (for instance, in Victoria), courts must nonetheless apply the common law caution to the same effect: a plaintiff’s evidence about what they would have done, given after the event and coloured by hindsight and self-interest, is an inherently unreliable guide to what would actually have occurred. The courts have recognised that objective factors are generally more important than the plaintiff’s subjective testimony in hindsight, and in practice will examine the objective circumstances that prevailed at the time before weighing any subjective account the plaintiff advances against them.
This principle was recently applied by the Victorian Court of Appeal in Selak v National Tiles Co Pty Ltd6. Mr Selak held options over approximately 3.2 million shares in National Tiles, with an exercise price of $393,600 and a face value of approximately $1.4 million. National Tiles was found to have breached the Options Agreement by providing a shareholders’ agreement that was inconsistent with the contractual regime governing the options. Mr Selak did not exercise the options before they expired and sued for his lost gain, contending that, had a compliant agreement been provided, he would have signed it, paid the exercise price, exercised the options and acquired the shares. That counterfactual was not established at trial, and the Court of Appeal upheld that conclusion.
The Court approached causation by examining the objective circumstances that existed at the time. The picture was heavily against Mr Selak. His relationship with Mr Walker, who controlled National Tiles with a 93.88 per cent holding, had completely broken down following board disputes and his removal as a director. Had he exercised the options, Mr Selak would have acquired only a 1.397 per cent shareholding in a company in which he had no ongoing role, while being required to pay $393,600. There was no evidence of either his willingness or capacity to do so. None of his proposals to Mr Walker had involved paying the exercise price: he had not sought to negotiate the terms of the shareholders’ agreement, and by May 2020 he considered it a “very low prospect” that he could extract value from the shares.
His subjective evidence further undermined the counterfactual he sought to establish. He described entering into the offered shareholders’ agreement as “throwing good money after bad” and repeatedly emphasised the importance of protections against minority and marketability discounts. However, a shareholders’ agreement that complied with the Options Agreement could still permit such discounts. The Court regarded this as further reason to doubt that Mr Selak would in fact have entered into the counterfactual agreement on which his causation case depended. The appeal was dismissed. The decision illustrates that where a plaintiff’s own evidence, when assessed against the objective circumstances, is inconsistent with the counterfactual they seek to establish, retrospective assertions about what they would have done will not be sufficient to prove factual causation.
When does the “but for” test not apply? The material contribution exception
There are, however, limited circumstances in which the “but for” test is inadequate. This occurs where the nature of the evidence makes it impossible to prove that the defendant’s breach was a necessary condition of the harm, even though it is clear that the breach materially contributed to the harm or the risk of harm.
Two well-known examples are the UK decisions of Fairchild v Glenhaven Funeral Services Ltd & Ors7 and Bonnington Castings Ltd v Wardlaw.8
In Fairchild, employees were exposed to asbestos by multiple employers and later developed mesothelioma. Medical science could not determine which exposure had triggered the disease. Because the claimants could not prove on the balance of probabilities that any particular employer’s breach caused the illness, the House of Lords held that it was sufficient that each employer had materially increased the risk of the disease.
In Bonnington Castings, an employee contracted pneumoconiosis after inhaling silica dust from both negligent and non-negligent sources. It was impossible to identify which particles caused the disease. The House of Lords held that the negligent source need only have materially contributed to the disease.
The High Court considered both decisions in Amaca Pty Ltd v Booth9. Mr Booth had been exposed to asbestos from products manufactured by two related defendants and later developed mesothelioma. Medical evidence established that his disease was caused by asbestos exposure, but it was impossible to determine which defendant’s products were responsible. The High Court upheld the finding of causation against both defendants, accepting that where the nature of the causal mechanism makes it impossible to identify which source of exposure caused the harm, it is sufficient that the defendant’s breach materially contributed to the risk of the disease occurring. The case confirmed that “but for” causation remains the primary test, but that in circumstances analogous to those in Fairchild and Bonnington Castings, material contribution to risk will suffice.
How do Australia’s Civil Liability Acts deal with these ‘exceptional cases’?
The Ipp Report recognised that cases such as Fairchild and Bonnington Castings expose an evidentiary gap in the ordinary “but for” analysis. In response, Australia’s civil liability legislation generally permits courts, in appropriate or exceptional cases, to find factual causation even where the breach cannot be established as a necessary condition of the harm.10
Courts have been reluctant to expand these provisions beyond circumstances analogous to Fairchild and Bonnington Castings. They do not apply merely because a plaintiff cannot prove causation on the available evidence. Rather, they are directed to situations where the nature of the case makes proof of “but for” causation impossible. Absent some analogy with those exceptional categories, courts have generally insisted upon proof of causation through the ordinary counterfactual analysis.
Key considerations when assessing causation in a claim
Proving a breach of duty is only the first step. The more demanding question is often whether the breach made any difference. The High Court confirmed in Adeels Palace Pty Ltd v Moubarak11 that proof of breach does not prove causation. In that case, a shooting occurred at a restaurant that had no security. The plaintiffs argued the restaurant was negligent in failing to employ security guards. The High Court held that causation was not established: there was no evidence that the presence of a security guard would have prevented the attack. The omission was a breach; it was not a cause.
That requirement applies whether the claim sounds in tort, contract or statute. Causation is a distinct issue from breach, and early investigation should address both. Many claims that survive a breach analysis will not survive a causation analysis.
The counterfactual must be scrutinised carefully against the objective evidence available at the time. Every assumption embedded in the plaintiff’s version of events should be tested. Where causation depends on what the plaintiff would have done, certain civil liability Acts render post-harm self-serving statements inadmissible, and even at common law, objective facts carry greater weight than a plaintiff’s retrospective account of their own hypothetical choices.
1 5D(1) of the Civil Liability Act 2002 (NSW), s 11(1) of the Civil Liability Act 2003 (Qld), s 51(1) of the Wrongs Act 1958 (Vic), and their equivalents in other jurisdictions.
2 (2020) 271 CLR 192.
3 [2025] VSCA 261.
4 Civil Liability Act 2003 (Qld) s 11(3)(a), Civil Liability Act 2002 No 22 (NSW) s 5D(3)(a), Wrongs Act 1958 (VIC) s 51(3), and equivalent provisions in other jurisdictions.
5 This rule appears in s 5D(3)(b) of the Civil Liability Act 2002 (NSW), s 11(3)(b) of the Civil Liability Act 2003 (Qld).
6 [2026] VSCA 118.
7 [2002] UKHL 22.
8 [1956] AC 613.
9 (2011) 246 CLR 36.
10 5D(2) of the Civil Liability Act 2002 (NSW), s 11(2) of the Civil Liability Act 2003 (Qld), s 51(2) of the Wrongs Act 1958 (Vic), and their equivalents in other jurisdictions.
11 (2009) 239 CLR 420.
This article may provide CPD/CLE/CIP points through your relevant industry organisation.
The material contained in this publication is in the nature of general comment only, and neither purports nor is intended to be advice on any particular matter. No reader should act on the basis of any matter contained in this publication without considering, and if necessary, taking appropriate professional advice upon their own particular circumstances.
