Hunting the Chimaera – vicarious liability after Various Claimants v Morrisons

1. In Mohamud v Morrisons Supermarkets (2016) AC 677, Lord Dyson MR observed that:

“To search for certainty and precision in vicarious liability is to undertake a quest for a chimaera”

2. Whilst absolute certainty and precision may remain unattainable goals in this area of the law, a consistent set of principles, applicable across all cases, is, it is submitted, still well within reach.

3. The Supreme Court decision in Various Claimants v Morrisons Supermarkets (2020) UKSC 12 has been widely publicised. It has been welcomed as a corrective to certain misunderstandings that crept into practice following the earlier decision of the Court in Mohamud (in which a petrol station attendant viciously attacked a customer on the forecourt). The overarching test, derived from Lord Nicholls’s judgment in Dubai Aluminium v Salaam (2003) 1 AC 366, has been restated:

“The court generally has to decide whether the wrongful conduct was so closely connected with acts the employee was authorised to do that, for the purposes of the liability of his employer, it may fairly and properly be regarded as done by the employee while acting in the ordinary course of his employment’

4. The importance of Lord Nicholl’s statement of the law is the emphasis on the course of employment as the limiting factor. Indeed, the decision of the Supreme Court on the facts in Mohamud had been criticised in other jurisdictions for losing sight of this important constraint. In Prince Alfred College v ADC (2016) HCA 37, the High Court of Australia, in a unanimous judgment, made the following observation:

“What occurred after the victim left the kiosk was relevantly unconnected with the employee’s employment. The approach of focusing on any special role that the employer has assigned to the employee and the position in which the employee is thereby placed vis-à-vis the victim, is also designedly different from the approach in Mohamud. This is because the test of vicariously liability, requiring no more than sufficiency of connection – unconstrained by the outer limits of the course or scope of employment – is likely to result in the imposition of vicarious liability for wrongful acts for which the employment provides no more than an opportunity”

5. However, there is some concern that certain comments in Various Claimants v Morrisons suggest a fissure is emerging in the case law, with those cases involving sexual abuse requiring a fundamentally different approach to other types of claim. This concern stems from Lord Reed’s observation (referring to the comments of Lord Nicholls in Various Claimants v Catholic Welfare Society (2013) 2 AC 1) paragraph 23 that:

“the close connection test has been applied differently in cases concerned with the sexual abuse of children, which cannot be regarded as something done by the employee while acting in the ordinary course of his employment. Instead, the courts have emphasised the importance of criteria that are particularly relevant to that form of wrongdoing, such as the employer’s conferral of authority on the employee over the victims, which he has abused”

6. This article will argue that there is no need for a radically different approach in sexual abuse cases. The guiding factors remain the same, and whilst some of those factors may assume more or less importance depending upon the specific facts of the case, the statement of law derived from Dubai Aluminium is of general application.

7. The starting point in this analysis Is Lister v Hesley Hall (2002) 1 AC 215. The decision of the House of Lords is problematic because the four speeches of Lords Steyn, Clyde, Hobhouse and Millett respectively are not uniform in their approach. Nonetheless, whilst their Lordships overruled the earlier Court of Appeal decision in Trotman v North Yorkshire (1999) IRLR 98 (which had rather clumsily attempted to fit sexual abuse into the ‘modes’ of employment test derived from early case law and the ‘Salmond’ formulation), it is tolerably clear that they were not seeking to depart from the ‘course of employment’ as the outer limit of the employer’s vicarious liability. Instead, the correct approach was to define the scope of the relevant employee’s duties broadly rather than narrowly, and thereafter to ask whether the tortious acts were sufficiently closely connected to them. Accordingly, Lord Steyn commented as follows:

“It remains, however, to consider how vicarious liability for intentional wrongdoing fits in with Salmond’s formulation. The answer is that it does not cope ideally with such cases. It must, however, be remembered that the great tort writer did not attempt to enunciate precise propositions of law on vicarious liability. At most he propounded a broad test which deems as within the course of employment “a wrongful and unauthorised mode of doing some act authorised by the master”. And he emphasised the connection between the authorised acts and the “improper modes” of doing them. In reality it is simply a practical test serving as a dividing line between cases where it is or is not just to impose vicarious liability. The usefulness of the Salmond formulation is, however, crucially dependent on focusing on the right act of the employee

[…]

If this approach to the nature of employment is adopted, it is not necessary to ask the simplistic question whether in the cases under consideration the acts of sexual abuse were modes of doing authorised acts. It becomes possible to consider the question of vicarious liability on the basis that the employer undertook to care for the boys through the services of the warden and that there is a very close connection between the torts of the warden and his employment. After all, they were committed in the time and on the premises of the employers while the warden was also busy caring for the children”

8. Subsequently, when sitting as a member of the Privy Council in Attorney General of Jamaica v Bernard (2005) IRLR 398, Lord Steyn observed that the decision in Lister:

“did not come out of the blue. On the contrary, it was a development based on a line of decisions of high authority dating from Lloyd v Grace, Smith & Co [1912] AC 716 where vicarious liability was found established in cases of intentional wrongs. Lister is, however, important for a number of reasons. It emphasised clearly the intense focus required on the closeness of the connection between the tort and the individual tortfeasor’s employment. It stressed the need to avoid terminological issues and to adopt a broad approach to the context of the tortious conduct and the employment”

9. Critically, therefore, Lister was not a case concerned only with the correct approach to claims arising from sexual abuse. It was intended to be, and should be regarded as, applicable to all claims. Lord Clyde recognised this in his speech:

“Cases which concern sexual harassment or sexual abuse committed by an employee should be approached in the same way as any other case where questions of vicarious liability arises. I can see no reason for putting them into any special category of their own.”

10. And, in Bernard (which involved a non-sexual assault by a police officer), Lord Steyn put the matter beyond doubt:

“While the facts of Lister are very different from the circumstances of the present case, the principles enunciated in Lister are of general application to intentional torts”.

11. One of the most important principles which has arisen in the sexual abuse cases is the creation of risk. It is a concept which was most notably deployed in the seminal Canadian cases (Bazley v Curry (1999) 174 DLR 45 and Jacobi v Griffiths (1999) 174 DLR 71) which were so heavily relied upon Liste However, the creation of risk is emphatically not a concept unique to the sexual abuse cases. Rather, it is part of core reasoning underpinning the use of the ‘course of employment’ as the outer-limit of vicarious liability. Accordingly, in Lister itself, Lord Hobhouse explained the position as follows:

[…] Atiyah, Vicarious Liability in the Law of Torts … suggested, at p 171: “The master ought to be liable for all those torts which can fairly be regarded as reasonably incidental risks to the type of business he carries on.” These passages are not to be read as confining the doctrine to cases where the employer is carrying on business for profit. They are based on the more general idea that a person who employs another for his own ends inevitably creates a risk that the employee will commit a legal wrong. If the employer’s objectives cannot be achieved without a serious risk of the employee committing the kind of wrong which he has in fact committed, the employer ought to be liable. The fact that his employment gave the employee the opportunity to commit the wrong is not enough to make the employer liable. He is liable only if the risk is one which experience shows is inherent in the nature of the business. 

66. While this proposition has never, so far as I am aware, been adopted in so many words as a test of vicarious liability in any of the decided cases, it does I think form the unspoken rationale of the principle that the employer’s liability is confined to torts committed by an employee in the course of his employment” (emphasis added)

12. Since Lister, there have been numerous examples of Courts at the highest levels deploying the ‘creation of risk’ test as a cornerstone of their analysis in cases far removed from the context of sexual abuse. A few examples may be instructive.

13. In Bernard (above) the Privy Council (Lord Steyn giving the judgment) was concerned with a police officer who shot a member of the public in the head when he refused to hand over a telephone to allow the officer to make a call. Lord Steyn stated as follows:

“Moreover, one must consider the relevance of the risk created by the fact that the police authorities routinely permitted constables like Constable Morgan to take loaded service revolvers home, and to carry them while off duty. The social utility of allowing such a licence to off duty policemen may be a matter of debate. But the state certainly created risks of the kind to which Bingham JA made reference. It does not follow that the using of a service revolver by a policeman would without more make the police authority vicariously liable. That would be going too far. But taking into account the dominant feature of this case, viz that the constable at all material times purported to act as a policeman, the risks created by the police authorities reinforce the conclusion that vicarious liability is established”.

14. By contrast, the Court of Appeal in Graham v Commercial Bodyworks (2015) ICR 655 concluded that there was no vicarious liability where a mechanic, as a prank, sprayed thinning agent over a colleague’s overalls and lit a cigarette lighter nearby, causing serious burns. Longmore LJ cited at length from the decision in Bazley v Curry before stating as follows:

“Although all this was, no doubt, said with sex abuse cases primarily in mind, it is a useful general statement of the position and justifies an inquiry into the question whether there is a close connection between the creation or enhancement of a risk and the wrong that accrues therefrom. On the facts of this case I would, by reference merely to the passage quoted above, conclude that, although the defendant employers did create a risk by requiring their employees to work with thinning agents, it is difficult to say that the creation of that risk was sufficiently closely connected with Mr Wilkinson’s highly reckless act of splashing the thinner onto Mr Graham’s overalls and then using a cigarette lighter in his vicinity. It is only the first of McLachlin J’s five factors that is present in this case. The other factors tell against the imposition of liability. The wrongful act did not further the employer’s aims; there was no friction or confrontation inherent in the employer’s enterprise and such intimacy as there was likewise had no connection with that enterprise; it is inappropriate to talk either of power conferred on Mr Wilkinson in relation to Mr Graham or any particular vulnerability of Mr Graham to the wrongful exercise of such power”.

15. In this regard, the decision in Mohamud may be regarded as something of an outlier. In dealing with the argument that the creation of risk is an important factor, Lord Toulson was relatively dismissive:

“…the court is not required in each case to conduct a retrospective assessment of the degree to which the employee would have been considered to present a risk. As Immanuel Kant wrote: “Out of the crooked timber of humanity, no straight thing was ever made.” The risk of an employee misusing his position is one of life’s unavoidable facts”.

16. However, it is submitted, with all due respect, that this was not the correct question. The issue is not whether the ‘employee would have been considered a risk’ in hindsight, but rather whether the broad nature of the employee’s job, and the surrounding circumstances of their working environment, can be said to have given rise to the inherent or reasonably incidental risk of deliberate wrongdoing. In Bernard, the provision of firearms to officers plainly did create this risk, and it was sufficiently connected with the officer’s use of that firearm to assert his authority over the member of the public using the phone. In Graham, there was a low-level background risk created by using dangerous thinning agents for certain jobs, but the deliberate prank was too far removed from and had no sufficient connection with that risk to justify the imposition of vicarious liability.

17. Mohamud was perhaps a borderline decision on its facts. It is difficult not to have some sympathy with Christopher Clarke LJ’s observation in the Court of Appeal that:

“On the judge’s findings, that for reasons of his own Mr Khan, having declined Mr Mohamud’s request — for reasons of his own which are unexplained — decided to do what he did; and went out on to the fore court to assault Mr Khan, when it was no part of his job to do anything in respect of drivers there who were behaving perfectly properly. If Morrisons were liable it would mean that in practically every case where an employee was required to engage with the public, his employer would be liable for any assault which followed on from such an engagement. That appears to me to be a step too far”

18. Mohamud is perhaps best understood, following the clarification provided by Various Claimants v Morrisons, as a case where, although Mr Khan’s job carried little inherent risk of intentional wrongdoing, there was plainly some, since he was in a customer-facing role and able to interact one-on-one in the relatively sparse confines of a petrol station forecourt. However, as Lord Reed explained, the assault grew out of the initial argument in the kiosk. The risk, whilst small, crystallised without interruption and there was no clear stage at which it could be said that Mr Khan shifted from a professional to personal capacity, such as to remove him from the ordinary course of his employment.

19. It is instructive to return to the modern source of the ‘risk creation’ theory to search for a more coherent analysis. The decision of the Canadian Supreme Court in Jacobi v Griffiths (decided on the same occasion as Bazley v Curry) is often overlooked in favour of its better known companion. However, it is a masterful analysis of the law in this field. The abuser in that case worked as a program director at a youth club. He abused two children that he met in the club, but the abuse took place outside working hours in his own home. The Court held that the club was not vicariously liable. The Majority’s judgment contains the following particularly illuminating passage:

“The key to this case, in my view, is that the Club’s “enterprise” was to offer group recreational activities for children to be enjoyed in the presence of volunteers and other members.  The opportunity that the Club afforded Griffiths to abuse whatever power he may have had was slight.  The sexual abuse only became possible when Griffiths managed to subvert the public nature of the activities.  The success of his agenda of personal gratification, which ultimately progressed to sex acts, depended on his success in isolating the victims from the group.  The progress from the Club’s program to the sexual assaults was a chain with multiple links, none of which could be characterized as an inevitable or natural “outgrowth” of its predecessor:

(1)   The Club provided Griffiths with the opportunity to work with children.

(2)   While it was undoubtedly part of Griffiths’ job to develop a positive rapport with the children, the relationship envisaged by the Club had no element of intimacy comparable to the situation in Children’s Foundation.

(3)   While Griffiths might come into occasional physical contact with children by reason of his job, e.g., steadying a child on a piece of gym equipment, the authorized “touching” had no more to do with parenting, nurture or intimacy than could be said of a normal adult reaching out to steady a child who, e.g., tripped over a carpet.

(4)   Griffiths enticed each child to his home to cultivate a one-on-one relationship.  The Club activities did not require the Program Director to be alone with a child off Club premises and outside Club hours.  Such a practice was explicitly prohibited after 1988.

(5)   Griffiths established his own bait of home attractions, such as video games, that had nothing to do with Club activities.  It was not part of his job to entertain children at home after hours.

(6)   Unlike the situation in Children’s Foundation, the appellants’ mother was a parental authority interposed between the assailant and his victims.  She gave permission to the children to go to Griffiths’ home.  No doubt, knowing of Griffiths’ job at the Club, she did not regard him as a stranger or as a threat.  Nevertheless, it must have been evident to a reasonably cautious parent that Griffiths’ home entertainment was not part of the Club’s program.

(7)   Once the children were drawn into his home-based activities, Griffiths gradually increased the level of intimacy, initially with Randy and subsequently with Jody, in terms of banter and sexually suggestive talk.  This was not only unauthorized, it was antithetical to the moral values promoted by the Club.

(8)   Eventually, when Griffiths saw his chance, he committed the assaults.

81 If it was the law that the provision of “opportunity” was enough to render the employer liable, then the progression from step 1 to step 8 might be seen as a series of  “but-for” opportunities sufficient to impose liability.  But that is not the law.  I accept that “but for” the opportunity created by Griffiths’ employment at the Club, it is unlikely these assaults would have occurred in the way that they did.  As pointed out by McLachlin J. in Children’s Foundation (at para. 37), however, the relevant nexus, if it exists, is between the job-related conduct at step one and Griffiths’ criminal assault at step eight.  It is not enough to postulate a series of steps each of which might not have happened “but for” the previous steps.  Where, as here, the chain of events constitutes independent initiatives on the part of the employee for his personal gratification, the ultimate misconduct is too remote from the employer’s enterprise to justify “no fault” liability.  Direct liability would attach, of course, if the employer could be found derelict in respect of any of its own responsibilities towards these children.  However, this appeal has been argued on the assumption that there is no such fault on the part of the employer”

20. The reasoning in Jacobi is particularly resonant because it bears a close resemblance to the approach recently adopted by the Supreme Court in Various Claimants v Morrisons. In particular, it is strikingly similar Lord Reed’s analysis of the facts in Mohamud:

“Read in context, Lord Toulson’s comments that there was “an unbroken sequence of events”, and that it was “a seamless episode”, were not directed towards the temporal or causal connection between the various events, but towards the capacity in which Mr Khan was acting when those events took place. Lord Toulson was explaining why, in his view, Mr Khan was acting throughout the entire episode in the course of his employment. When he followed the motorist out of the kiosk and on to the forecourt, he was following up on what he had said to the motorist in the kiosk. He ordered the motorist to keep away from his employer’s premises, and reinforced that order by committing the tort. In doing so, he was “purporting to act about his employer’s business”. As Lord Toulson said, “this was not something personal”.

21. It is submitted that both Lord Reed in Various Claimants, and the majority of the Canadian Supreme Court in Jacobi, were articulating the same concept. Namely that the creation of risk and the presentation of an opportunity exist at different ends of the same spectrum. An opportunity to commit a tort may, if accompanied by certain factors, properly be characterised as an inherent risk. Those factors are necessarily non-exhaustive, but may include the provision of dangerous weapons, training or instructions to use physical contact/force, intimacy, privacy, time alone with the victim or, in cases of theft, with the object which is stolen. Where that risk exists, and it is significant, it may imbue subsequent acts with the essence of the tortfeasor’s job or role, such that they can properly be regarded as carried out in the course of his or her employment even if they are relatively remote in time or place from the torfeasor’s primary work environment. However, the weaker the risk created, and the closer it is to a mere opportunity to begin with, the more likely it will be that each subsequent action dilutes the connection with the tortfeasor’s job or field of activities, such that the ultimate tortious act becomes increasingly ‘personal’, and therefore too far removed to be regarded as having been undertaken in the course of employment.

22. In the final analysis there is little doubt, as Lord Nicholls observed in Dubai Aluminium v Salaam, that:

“The crucial feature or features, either producing or negativing vicarious liability, vary widely from one case or type of case to the next. Essentially the court makes an evaluative judgment in each case, having regard to all the circumstances”

23. However, it is hoped that the analysis above demonstrates that creation of risk is an integral part of the ‘course of employment’ test and therefore remains a cogent factor which unites the ‘sexual abuse’ cases with others. Whilst it may not ever be possible to capture ‘the chimaera’ it can, it is submitted, be described and understood in a way which is useful to practitioners in this interesting and dynamic field of the law.