Lords Nicholls, Hoffmann and Rodger also relied on a very similar principle (paras 42, 67 and 168 respectively) and understood the ratio of McGhee as being the same (paras 44, 65 and 168 respectively): “So long as it was not insignificant, each employer’s wrongful exposure of the employee to asbestos dust, and, hence, to the risk of contracting mesothelioma, should be regarded by the law as a sufficient degree of causal connection” (para 42 per Lord Nicholls); “[In McGhee] …the House decided that materially increasing the risk that the disease would occur was sufficient to satisfy the causal requirements for liability… For present purposes, the McGhee principle is sufficient” (paras 65, 74 per Lord Hoffmann); “Following the approach in McGhee I accordingly hold that, by proving that the defendants individually materially increased the risk that the men would develop mesothelioma due to inhaling asbestos fibres, the claimants are taken in law to have proved that the defendants materially contributed to their illness” (para 168 per Lord Rodger).
The justifications for the “McGhee principle”
We think that the House of Lords in Fairchild identified four (overlapping) reasons for adopting the exceptional “McGhee principle”. Below we list these four (overlapping) reasons, then offer a brief assessment of them.
(1) Impossibility
A strong argument in favour of the “McGhee principle” was that to have insisted on the ordinary requirement of proof of causation on the balance of probabilities would have been to have insisted that the claimant do what is scientifically impossible. In the particular circumstances, where the claimants could prove that the employees had been injured by the negligence of one or more of their negligent employers, it seemed particularly harsh to insist that the claimants should lose because the limits of scientific knowledge prevented them from establishing which negligent employer in particular was responsible. Lord Rodger expressly referred (at paras 155 and 170) to the fact that the “McGhee principle” went no further than relieving the claimant from the need to prove the impossible: instead the claimant was required to prove the most that he or she possibly could (i.e. that the defendant’s negligence increased the risk of the harm being suffered).
(2) Otherwise defendants immune and duty empty
The second reason builds on the first reason (that in certain circumstances the ordinary rule would make successful claims impossible), but raises the further point that where impossibility is generic then its effect may be to render a particular duty nugatory. To put the same point in a different way, if none of the people who is owed a particular legal duty will ever be able to prove that a breach of it has caused him or her damage then the legal duty seems useless. Clearly an example of a situation where impossibility is generic is one where the limits of medical knowledge ensure that evidence will never be available to determine whether V was harmed by the risk created by D or by some background risk (including a risk created by a third party). Lord Hoffmann relied on this reason most heavily. He said, at para 62, “a rule requiring proof of a link between the defendant’s asbestos and the claimant’s disease would, with the arbitrary exception of single-employer cases, empty the duty of content. If liability depends upon proof that the conduct of the defendant was a necessary condition of the injury, it cannot effectively exist.” Lord Rodger, at para 155, also relied on the concern that a decision for the defendants might leave the duty “emptied of all practical content”.
Lord Bingham went further still and suggested that immoral defendants who realised that in certain circumstances they could never be held liable would be capable of manipulating this situation. Thus he said, at para 33, “Were the law…[not to make use of the “McGhee principle”], an employer exposing his employee to asbestos dust could obtain complete immunity against mesothelioma (but not asbestosis) claims by employing only those who had previously been exposed to excessive quantities of asbestos dust.”
(3) Balance of fairness
Clearly the risk that the “McGhee principle” carries with it is that negligent defendants may be held liable for harm which they did not in fact cause – perhaps the harm was in fact caused by a single fibre of asbestos whilst working for some other employer. The House of Lords considered, however, that this risk of injustice must be placed alongside the risk of injustice to claimants who might otherwise be left uncompensated despite the possibility (or – in Fairchild – certainty) that they were injured by someone’s negligence. In deciding which party had the better claim to be relieved from the risk of injustice the House of Lords attached particular attention to the fact that it had been established that the defendants were wrongdoers and the claimants innocent.
(4) International parity
This reason can be presented by quoting from Lord Bingham: “Whether by treating an increase in risk as equivalent to a material contribution, or by putting a burden on the defendant, or by enlarging the ordinary approach to acting in concert, or on more general grounds influenced by policy considerations, most jurisdictions would, it seems, afford a remedy to the plaintiff. … In a shrinking world (in which the employees of asbestos companies may work for those companies in any one or more of several countries) there must be some virtue in uniformity of outcome whatever the diversity of approach in reaching that outcome” (para 32).
(5) Assessment of the reasons
The second and third reasons seem to be the pivotal ones. The first reason – impossibility – is more a statement of fact that a reason and the fourth reason supports a solution rather than the solution of the “McGhee principle”. The reasons that carry the day are the second and third, with their suggestion that on balance it is better to put the burdens of impossibility on the defendant when the alternative is to leave the duty “empty”. As we will see below when we consider the scope of the “McGhee principle” the third reason is not sufficient without the second: it is not the case that the House of Lords thought that it is always better on balance to make negligent defendants suffer any problems involved in proving causation. Thus it seems that so far as the reasons given by the House of Lords justify the “McGhee principle”, they operate cumulatively.
The scope of the “McGhee principle”
(1) Wilsher v Essex Area Health Authority
The scope of the “McGhee principle” can best be explored by considering how the members of the House of Lords treated the facts of Wilsher v Essex Area Health Authority [1987] QB 730, [1988] AC 1074. Four of their Lordships in Fairchild (Lord Nicholls being the exception) expressly stated (at paras 22, 70, 118 and 149) that the “McGhee principle” should not be extended to the facts of Wilsher. (As you may recall Wilsher involved a child who developed a serious eye condition (RLF). The defendant had negligently subjected the child to excess oxygen. But the medical evidence was that although excess oxygen could have caused the RLF, the child also suffered from four other conditions implicated as possible causes of RLF, and it could not be said that it was more probable that the excess oxygen had caused the RLF than that some other agent had caused it.) It is worth working out why their Lordships thought that the facts of Wilsher do not fall within the proper scope of the “McGhee principle”, because it seems that in Wilsher it would have been impossible for the claimant to have proved any more than that the defendant’s negligence increased the risk of RLF.
Lord Bingham’s explanation is that “It is one thing to treat an increase in risk as equivalent to the making of a material contribution where a single noxious agent is involved, but quite another where any one of a number of noxious agents may equally probably have caused the damage” (para 22). Lord Hutton offered a similar analysis, but in terms of “risks” rather than “agents” (para 118). Lord Rodger offered a more detailed analysis bridging the language of “risks” and “agents”: “the claimant must prove that his injury was caused by the eventuation of the kind of risk created by the defendant’s wrongdoing. … The principle does not apply where the claimant has merely proved that his injury could have been caused by a number of different events, only one of which is the eventuation of the risk created by the defendant’s wrongful act or omission. Wilsher is an example. … This will usually mean that the claimant must prove that his injury was caused, if not by exactly the same agency as was involved in the defendant’s wrongdoing, at least by an agency that operated in substantially the same way” (para 170). Lord Hoffmann, however, whilst supporting the view that Wilsher falls outside of the scope of the principle, expressly rejected a “single noxious agent” limit as unprincipled (para 72). Instead he reiterated his view that “the concepts of fairness, justice and reasonableness underlie the rules which state the causal requirements of liability for a particular form of conduct” (para 56) and suggested that the causal requirements in cases involving employers’ duties to protect employees against the risk of disease (where the duty might be empty if the usual causation rules were used) might be different from duties to take reasonable care in treating patients, not least because of the political and economic consequences of imposing a heavier burden of liability on the National Health Service (para 129).
(2) Assessment of single agent / single risk approach
Is there any moral difference between a case where A is at risk of developing dermatitis from brick dust and B makes that 20 per cent more likely by not providing showers, and a case where A is at risk of developing RLF from four possible causes and B makes RLF 20 per cent more likely by creating a fifth possible cause? Unless there is a moral difference then it hardly seems worth going any further into trying to establish what counts as a single agent or a single risk.
Comparing two hypothetical cases might test for whether a moral difference is intuitive:
Case (A): Two hunters, D1 and D2, negligently shoot at V and only one bullet strikes him; it is not possible to tell whose bullet.
Case (B) One hunter (an expert who only rarely misses), D3 negligently thinks that V is a bear and shoots at his V’s head. At the time, V – who is blind – was approaching a cliff edge negligently left unfenced by D4. After D3 fires his gun, V falls over the cliff and his body is washed out to sea and is never recovered. In the aftermath, it is impossible to determine if D3’s bullet killed V and his body then toppled over the cliff, or V fell over the cliff so that the bullet passed over his head: D3’s evidence is that V disappeared a moment after he pulled the trigger.
Does it seem easier to make D1 and D2 liable in case (A) than to make D3 and D4 liable in case (B)?
If we do go further then there will obviously be problems in individuating risks and agents. Imagine a case where V is employed to word-process for 75 per cent of the working day and to twist the top off old pill-bottles and dispose of the contents for 25 per cent of the working day. The typing conditions meet all reasonable standards, but it is negligent for the employer to have set the top-twisting task without providing a top-twisting strain-relief tool. If V develops repetitive strain disorder and experts cannot say whether the reasonable typing or the negligent twisting more probably caused this then will the “McGhee principle” apply? Or, to put the same question differently, are different types of strain from different tasks parts of a single agent / risk or separate agents / risks? Note that we cannot classify risks in terms of the result they cause because it must have been assumed that in Wilsher the agents / risks were different even though they were agents / risks leading to the same result (RLF).
Again, suppose V is screened for breast cancer and D, her doctor, negligently fails to spot that she has breast cancer. The cancer goes untreated and V eventually dies from it. Medical evidence shows that had V’s breast cancer been spotted when she was screened, there was a 40% chance that the cancer could have been treated successfully. On normal principles, we would not say that D’s negligence caused V to die – on the balance of probabilities, V would still have died from cancer even had D treated her with a reasonable degree of care and skill. But could we apply the “McGhee principle” here to say that D’s negligence contributed to V’s death? It is not easy to see how the cancer and the failure to assess and treat the cancer could be treated as separate agents leading to V’s death, if in McGhee the dust and the failure to provide showers to wash it off counted as a “single noxious agent” causing McGhee’s dermatitis.
Of the three Lords who rely on some version of the single agent / single risk approach, only Lord Rodger seems to appreciate the problems that await the courts in using it. But his answer to these problems was merely to suggest (without explanation) that it might be necessary to supplement this approach with a further principle distinguishing between: (a) cases where D negligently creates a risk similar to a risk negligently created by another (Fairchild); (b) cases where D negligently creates a risk similar to a risk innocently created by himself or herself (McGhee); and (c) cases where D negligently creates a risk similar to a risk innocently created by another or by natural occurrence (the breast cancer screening example) (para 170). Lord Rodger did not explain, however, why this further distinction would be attractive.
In the light of the discussion above it is difficult to conclude that the majority of the House of Lords managed to identify a “sufficiently specific” principle to guide future cases. Indeed it would not be an exaggeration to say that the majority relies on a distinction which lacks both moral validity and practical utility.
(3) Assessment of Lord Hoffmann’s approach
Lord Hoffmann’s approach is one which insists that there can be different “causal requirements” in different contexts depending on what is fair, just and reasonable. He said this was illustrated by Reeves v Commissioner of Police of the Metropolis [2000] 1 AC 360 (where it was held that a prisoner’s decision to commit suicide in custody will not break the chain of causation between the prison authority’s breach of its duty to take reasonable steps to ensure that the prisoner did not commit suicide and the prisoner’s death – see McBride and Bagshaw, Tort Law, p 486) and Environment Agency v Empress Car Co (Abertillery) Ltd [1999] 2 AC 22 (where it was held that a defendant could still commit the strict criminal offence of “causing” polluting matter to enter water when the polluting matter was released from the defendant’s property by the deliberate intervention of an unknown person). We agree that these cases illustrate that what is fair, just and reasonable plays a role in shaping the law on causation. But we think it is important to note that neither of the cases Lord Hoffmann relies on goes as far as Fairchild: neither suggested that a defendant can be held liable in tort even if it is not proved that his behaviour had any effect on the claimant. Fairchild takes that extra step.
Lord Hoffmann’s principal explanation for taking the extra step is that otherwise the duty to protect employees against asbestos would be “empty”. But he also acknowledges (para 62) that this is untrue because if the claimant was only exposed to asbestos by one employer during his working life then causation could be proved. Indeed we think that any employer who exposed the claimant to more than half of his cumulative exposure could be held liable, because “more likely than not” that employer wrongfully exposed the claimant to the fibre or fibres that caused his or her condition. Further, as Lord Bingham pointed out even if the duty to protect employees from asbestos did not help victims of mesothelioma it would still help employees who developed asbestosis (which is thought to be caused by accumulation of asbestos inhaled). Given this, the employer’s duty would not be “empty”, at least not in the same way as a duty to take reasonable steps to ensure that V did not commit suicide might be “empty” if V’s decision to commit suicide always broke the chain of causation between the breach of that duty and V’s death. Certainly, it is hard to see how the duty would be any more “empty” than the duty to use reasonable skill and care not to expose the child to excess oxygen in Wilsher. To summarise, our first concern about Lord Hoffmann’s approach is that he provided no clear guidance on when a duty will be considered “empty” because of the ordinary causation rules.
A second concern about Lord Hoffmann’s approach is the risk that it might spark chaos (or at least lead to uncertainty!) whilst a host of unpopular causation rules are challenged and re-examined. Lord Hoffmann stressed that his intention was merely to justify a refinement of the rules relating to causation in a particular context rather than to replace all the rules with “a discretion to adapt causal requirements to the individual case” (para 60). But it seems reasonable to predict that Fairchild will lead to an increase in challenges to settled causation rules, with those challenges often relying on ill-defined assertions of what is “fair, just and reasonable”.
How important is Fairchild?
In practical terms the outcome of the case was very important. In 2001, 5,000 people died in the United Kingdom from asbestos-related diseases (mainly mesothelioma) and the number is estimated to rise towards 10,000 deaths per year by 2010. Although the Fairchild decision will not assist them all, it was estimated that there were 500 claims awaiting the outcome, and that the total cost of the decision to the UK insurance industry could be between six and eight billion pounds.
But is the case important in terms of development of tort law principles? The most obvious answer to this is: “No, not really”. If the House of Lords had decided that damages can always be awarded in proportion to risk negligently created (rather than in proportion to harm caused) or that once negligence is proved the burden of disproving causation is on the defendant, then the decision would have been very important. But instead the House of Lords merely decided that an exceptional principle, the “McGhee principle”, could be applied in certain limited circumstances. Further, as we have set out above, the House of Lords defined those limited circumstances narrowly. Consequently, unless a future court relaxes these limits, then – with the exception of the backlog of other mesothelioma claims – the Fairchild decision will only affect a tiny proportion of the tort claims that come before the courts each year.
Would a decision in favour of the defendants have been “deeply offensive to instinctive notions of what justice requires and fairness demands”?
In our opinion the answer to this question depends on whether one considers tort law as the only method of achieving justice and fairness. We share the sense that it would be grossly unfair if those suffering from mesothelioma were left without medical and financial support. But we are less convinced than Lord Nicholls that it is just to make the few employers who are still in business liable in tort for the full cost of the problem (although we accept that the intricacies of employers’ liability insurance mean that the employers will not themselves pick up the bill). We think that a lot could be said in favour of a legislative solution involving a compensation package funded by those industries (mainly the construction industry) which exposed employees to asbestos, those insurers who offered cover against the risks and by the State. Such a package would have the advantages that it could cover victims of mesothelioma who can identify no solvent former employers (including victims of environmental asbestos, such as those living near production facilities, and victims who cannot establish where they were subjected to asbestos), and could be put in place without any distortion of ordinary tort law. It would also avoid some of the costs and delays involved in adversarial legal claims. But the House of Lords clearly had no power to impose a legislative solution. Instead the House of Lords did what it could. And it seems to be widely thought that Fairchild is “a victory for justice and fairness” (as 92 MPs claimed in an Early Day Motion on 16 May).