March v E & MH Stramare Pty Ltd
High Court of Australia · 1991
March v E & MH Stramare Pty Ltd (1991) 171 CLR 506
A drunk driver hit a truck parked across the middle of a city street at one in the morning. The truck had no business being there. Was the parking a cause of the crash?
WHERE A CIVIL LIABILITY STATUTE IS ENGAGED IT DISPLACES THE COMMON-LAW APPROACH, and the condition is the point. At [44] their Honours say it is sufficient to observe that “in cases where the Civil Liability Act or equivalent statutes are engaged, it is the applicable statutory provision that must be applied”. Engagement does not follow from jurisdiction alone: Civil Liability Act 2002 (NSW) s 3B(1)(b) takes proceedings of the kind referred to in s 11 of the Dust Diseases Tribunal Act 1989 (NSW) outside “the whole Act except sections 15A and 15B and section 18(1)” in its application to s 18(1)(c) losses — so s 5D does not reach those proceedings and the common law this case states still governs their causation, even though parts of the Act continue to apply. Ask whether the statute reaches the claim before reaching for s 5D. At [42] s 5D(1) "divides the determination of whether negligence caused particular harm into two elements: factual causation and scope of liability". At [43] that division "expresses the relevant questions in a way that may differ from what was said by Mason CJ, in March v Stramare (E & M H) Pty Ltd, to be the common law's approach to causation". At [45] the first element "is determined by the ‘but for’ test: but for the negligent act or omission, would the harm have occurred?" And at [55]: "as s 5D(1) shows, the ‘but for’ test is now to be (and has hitherto been seen to be) a necessary test of causation in all but the undefined group of exceptional cases contemplated by s 5D(2)".
SO THE STATUTE SEPARATES WHAT MASON CJ KEPT TOGETHER. His objection was to splitting causation into a but-for question of fact and a separate question of responsibility; s 5D(1) makes exactly that split, and makes the but-for test necessary rather than merely useful. That is a difference in how the questions are put, which is as far as Adeels Palace goes. At [44] their Honours find it unnecessary to examine how far the two approaches diverge in result, which is a limit on how much can be read from the comparison.
WHAT IS NOT CLAIMED HERE: that every jurisdiction has such a provision, or that those which do have it in the same terms. So find out whether a causation provision applies in the jurisdiction you are answering on, and read it, rather than assuming either that one exists or that it matches s 5D.
"COMMON SENSE"
IS THE PART OF THIS CASE THE COURT HAS TURNED AGAINST, and a student should stop using the phrase as if it were the test. Young v Chief Executive Officer (Housing) [2023] HCA 31, GORDON AND EDELMAN JJ at [60]: "[a]lthough causation was once described in this Court as a concept of 'common sense', it has since, and repeatedly, been emphasised that the concept of common sense should be eschewed when applying the principles of causation".
THEIR FOOTNOTES ARE WHY THIS IS MORE THAN TWO JUDGES' PREFERENCE. Footnote 53 to that sentence cites this case -- "March v E & M H Stramare Pty Ltd (1991) 171 CLR 506 at 515" -- so there is no doubt what is being criticised. Footnote 54 collects the criticism: Travel Compensation Fund v Tambree (2005) 224 CLR 627 at 642 [45]; Allianz Australia Insurance Ltd v GSF Australia Pty Ltd (2005) 221 CLR 568 at 596-597 [96]-[98]; Adeels Palace Pty Ltd v Moubarak (2009) 239 CLR 420 at 440 [43]; Comcare v Martin (2016) 258 CLR 467 at 479 [42]; and Tapp v Australian Bushmen's Campdraft & Rodeo Association Ltd (2022) 273 CLR 454 at 470-471 [45]-[46], 487-488 [101]. Adeels Palace appears in that list too.
WHAT IS NOT DOUBTED, and the distinction is the whole point. The Court has not disturbed this case's holding that the but-for test is not the exclusive test of causation, nor its treatment of causation as a question of fact. What is criticised is the LABEL -- deciding causation by invoking common sense instead of stating the principle being applied. And note the limits here: these are the reasons of two Justices of five, in a case about contractual loss under the Residential Tenancies Act 1999 (NT) rather than negligence.
WHAT TO DO IN AN ANSWER: identify the causal question, apply the but-for test as the ordinary negative criterion, and where it does not dispose of the matter say WHY the negligence is or is not treated as a cause -- the intervening act, the scope of the duty, the purpose of the rule -- rather than writing that common sense supplies the answer. Where a civil liability statute applies, its two-element structure governs instead, as Adeels Palace describes.
What happened?
At about 1.00am on 15 March 1985 the plaintiff was driving south in Frome Street, Adelaide, at more than 60 km/h, in the lane nearest the centre of a six-lane road. The trial judge found it unlikely that his blood alcohol was below .18, and that it substantially impaired his judgment of speed and distance, his eye functions, and his co-ordination and reaction times.
The second defendant had parked a tray-top truck along the centre line, rear facing southbound traffic, with about half of it in the plaintiff's lane, so as to load bins of fruit and vegetables by forklift. The truck had its lights on. The middle of the street was the most convenient place to load from, and he had parked there at night for that purpose "for as long as he could remember". The plaintiff's car struck the rear of the tray with enough force to shift the truck about three metres.
The trial judge apportioned responsibility seventy per cent to the plaintiff and thirty per cent to the defendants. A majority of the Full Court of South Australia held that the parking was negligent but had not caused the accident: "[t]he sole effective cause of the accident was, in their Honours' view, the appellant's own negligence." The action was dismissed.
What did the Court decide?
The appeal was allowed with costs, and the trial judge's judgment for the plaintiff restored. The order sets aside the Full Court's orders "other than the order dismissing the cross-appeal and the order that the plaintiff pay the costs of the cross-appeal", and in lieu orders that the appeal to that court be dismissed with costs — so the apportionment stood and the cross-appeal's disposal did not move.
Five judges wrote separately and all five allowed the appeal. On the test for causation they divided four to one.
Proposition
What is the principle?
The "but for" test is not the exclusive test of causation in negligence. Mason CJ: "I do not accept that the 'but for' (causa sine qua non) test ever was or now should become the exclusive test of causation in negligence cases." What was the cause of an occurrence is "a question of fact which 'must be determined by applying common sense to the facts of each particular case'", and the test "applied as an exclusive criterion of causation, yields unacceptable results".
It still does real work as a negative criterion, but not even there without qualification. Deane J: "unqualified acceptance of the 'but for' test as even a negative or exclusionary test of causation for the purposes of the law of negligence would lead to the absurd and unjust position that there was no 'cause' of an injury in any case where there were present two independent and sufficient causes of the accident in which the injury was sustained."
McHugh J held the opposite on the test and allowed the appeal anyway: "the preferable course is to use the causa sine qua non test as the exclusive test of causation", with every other limiting rule "recognised as a policy-based rule concerned with remoteness of damage and not causation".
Why does this case matter?
Because of what it let the plaintiff keep. On the Full Court's reasoning a drunk driver who runs into an unlawfully parked truck recovers nothing, because his own negligence is the only cause worth naming. That reasoning belongs to a world before apportionment legislation, where identifying one "effective cause" decided the whole case. Once damages can be apportioned, the question of whether the defendant's negligence is a cause is separate from how much of the loss the plaintiff must bear, and collapsing the two is how a defendant who created the danger escapes entirely.
The other reason is the shape of the argument it licenses. Mason CJ's answer to the Full Court was not that the truck driver's negligence passed a test — it was that his wrongful act "created a situation of danger, the risk being that a careless driver would act in the way that the appellant acted", and that the duty existed to protect motorists from exactly that risk. An intervening careless act does not break the chain of causation where the defendant's wrongful conduct "has generated the very risk of injury resulting from the negligence of the plaintiff or a third party and that injury occurs in the ordinary course of things". Both halves carry weight: the risk must be the very one that eventuated, and the injury must have come about in the ordinary course.
Exam and application relevance
Start from the statute, and find out whether there is one. Most Australian jurisdictions have legislated on causation in negligence, and where a civil liability provision applies it is the provision you work from rather than this case — but the coverage is not uniform and the exclusions matter. Say which regime you are in before you state a test.
Where the common law applies, do not import the statute's structure into it. Mason CJ expressly criticises the commentators' division of causation into a question of fact settled by the but-for test and a separate question of legal responsibility: that approach "places rather too much weight on the 'but for' test to the exclusion of the 'common sense' approach which the common law has always favoured", and implies "that value judgment has, or should have, no part to play in resolving causation as an issue of fact". At common law it is one question, and the value judgment is inside it.
So use the but-for test as an aid rather than a stage. Ask whether the harm would have occurred anyway — which will usually dispose of it, subject to the case the test cannot handle, where two independent acts were each sufficient. Then say, on these facts, why the defendant's negligence should or should not be regarded as a cause. "Novus actus interveniens" and "sole effective cause" are conclusions, and an answer that reaches for one instead of explaining why the defendant's negligence did or did not generate the risk that eventuated has skipped the work.
Check your understanding
A defendant negligently creates a hazard, and the plaintiff is injured only because he then behaves carelessly himself. Why does that carelessness not necessarily break the chain of causation — and what would have to be different for it to?