Vicarious liability as risk-through-placing

Introduction

When is a defendant organisation vicariously liable to a claimant who is injured as a result of the tort of someone who does work for the defendant? The cases tell us that the vicarious liability doctrine consists of two rules, but those two rules have changed over time. Once, the first rule required the individual tortfeasor (who wrongfully caused the claimant’s injury) to be the defendant’s employee, and the second rule required that the wrong be committed in the course of the tortfeasor’s employment, in that it was an unauthorised mode of doing an authorised act. Now, both rules are expressed at a higher level of generality—the first rule is that there must be a sufficiently close relationship between the defendant and the individual tortfeasor (who wrongfully caused the claimant’s injury), and the second rule is that there must be a sufficiently close connection between that wrongful conduct and the defendant–tortfeasor relationship. The present two-rule formula was neatly summed up by Lord Reed SCJ in Cox v Ministry of Justice: ‘Vicarious liability in tort is imposed upon a person in respect of the act or omission of another individual, because of his relationship with that individual, and the connection between that relationship and the act or omission in question’. ¹

Unfortunately, the cases also proffer many justifications for the vicarious liability doctrine. These justifications are pitched at a high level of generality. This is unsatisfactory; a doctrine that has several imprecise justifications is unstable, because a court can ‘reason-shop’, invoking one or more of these vague justifications for its decision. If we don’t know why we have vicarious liability, the doctrine will surely run amok.

This paper examines four justifications for the doctrine that are found in the cases. They are: (a) effective compensation through spreading the cost of the claimant’s injury; (b) deterring the defendant from inadequately supervising potential tortfeasors who do work for them; (c) the defendant should pick up the burden of an organisational or business relationship which he had undertaken for his own benefit; and (d) the defendant, by placing the tortfeasor in the position in its organisation that it did, had created or significantly enhanced the risk of the claimant’s injury.

A justification, to qualify as such, has to explain why vicarious liability is imposed in some situations but not in others. So something that cannot explain why vicarious liability was imposed or not imposed in a significant number of cases should be rejected as a justification. On this basis, only the fourth justification (justification (d)) should be embraced, and the other three (justifications (a) – (c)) rejected.

This paper first explains why the first three justifications should be rejected. It then explains why the fourth justification is the sole satisfactory justification for the vicarious liability doctrine. Finally, it suggests three implications of accepting the fourth justification as the sole justification.

Rejecting three of the four justifications

The first justification for the vicarious liability doctrine is: vicarious liability exists to ensure that the claimant receives effective compensation by spreading the cost of his injury. This justification has two parts: (i) effective compensation; and (ii) spreading the cost of the injury which was caused by the wrongdoing. The first part, effective compensation, focuses on the plight of the defendant—the tortfeasor being (almost always) financially worthless, someone else has to foot the bill or the claimant’s injury goes uncompensated.

But who should step in and compensate the otherwise pitiful claimant? That’s where the second part comes in. The second part tells us that the defendant should foot the bill because it is not too painful for it to do so, since it can spread the bill through insurance or higher prices to its customers (or lower bonuses for its staff).

The first justification is regularly invoked in the cases. Lord Phillips SCJ said in 2012: ‘The policy objective underlying vicarious liability is to ensure, insofar as it is fair, just and reasonable, that liability for the tortious wrong is borne by a defendant with the means to compensate the victim. Such defendants can usually be expected to insure against the risk of such liability, so that this risk is more widely spread’. ² A 2017 ruling that a local social services authority was vicariously liable for sexual abuse committed by foster parents has been explained as being substantially ‘motivated by a desire that the burden of the claimant’s misfortune be distributed more widely’. ³ In the Canadian case of Bazley v Curry, McLachlin J said that vicarious liability ‘improves the chances that the victim can recover the judgment from a solvent defendant’, ⁴ and that the defendant organisation is ‘often in the best position to spread the losses through mechanisms like insurance and higher prices, thus minimizing the dislocative effects of the tort within society’. ⁵

So far, we have established that (1) because the tortfeasor is financially worthless, we want to make someone besides the tortfeasor pay for the claimant’s injury; and (2) we want that someone to be able to pay for the injury and then spread that cost through insurance or by making many other people bear little parts of it. These two things do not naturally lead us to the two rules of vicarious liability. The two rules tell us that the defendant is vicariously liable if it has a close relationship with the tortfeasor and there is a close connection between the wrongful conduct and the defendant–tortfeasor relationship. They, and the cases, do not restrict liability to big and wealthy corporate defendants that can spread the cost of the injury; vicarious liability has been routinely imposed on non-profit organisations. And even if they did, the first justification does not adequately explain the two-rule structure of the vicarious liability doctrine. Because of this first justification, we could have concluded that the Government should pay for the injury; after all, the Government is excellent at spreading the cost through taxation and transfer payments.

Thus, the first justification for vicarious liability—vicarious liability exists to ensure that the claimant receives effective compensation by spreading the cost of his injury—cannot justify the doctrine’s two rules. The consequences of imposing vicarious liability on the facts of a particular case may be effective compensation and (sometimes) spreading the cost of the injury. But we should not mistake consequence (the result of imposing vicarious liability) for justification (a good reason for imposing vicarious liability).

The second justification for the vicarious liability doctrine is: vicarious liability exists to deter defendants from inadequately supervising potential tortfeasors who do work for them. Thus, McLachlin J noted in Bazley that imposing vicarious liability on the tortfeasor’s employer will incentivise the employer to better administer and supervise its employees. ⁶ This justification was accepted without explanation by the Court of Appeal in Gravil v Carroll. ⁷ But it has not found favour in the Supreme Court. In Armes v Nottinghamshire CC, the Supreme Court noted that, unlike in Canada, deterrence has not been a ‘prominent justification’ for the imposition of vicarious liability in the British case law. ⁸

The deterrence justification should be rejected for several reasons.

First, defendants that are not guilty of inadequate supervision have been found vicariously liable in cases where the claimant’s injury is caused by a single deliberate or negligent act of the tortfeasor.

Poland v John Parr And Sons is a stark example that involves a single deliberate act. ⁹ There, the defendant was vicariously liable for its carter striking the claimant. The carter struck the claimant as he mistakenly believed that the claimant was trying to steal the defendant’s sugar. Clearly, in the circumstances, there was no inadequate supervision of the defendant that had to be deterred and the judgement makes no reference to any such thing; indeed, the inadequacy discussed was with respect to the judgement of the carter. ¹⁰

Likewise for single negligent acts. In Century Insurance Co Ltd v Northern Ireland RTB, ¹¹ Davison was employed to drive his employer’s petrol tanker and deliver petrol to garage proprietors. While filling the petrol storage tank of a garage proprietor, Davison lighted a cigarette and threw away the lighted match, which ignited some material on the garage floor and caused a fire. Again, there was no inadequate supervision to be deterred and the judgement makes no reference to the same. We can imagine the absurdity of insisting that Davison’s employer supervise Davison to prevent him from doing something so obviously dangerous.

Secondly, one would imagine, and the case law confirms, that where the claimant’s injury is caused by the tortfeasor’s multiple deliberate acts over time, deterrence can possibly be invoked as a justification for imposing vicarious liability. In these cases, courts believe that such acts could have been stopped earlier if the defendant had adequately supervised the tortfeasor. But even here the deterrence justification rings hollow, for four reasons.

The first reason: courts do not tell us clearly why inadequate supervision has to be deterred in some cases but not in others. Contrast two Canadian Supreme Court cases: Bazley and KLB v British Columbia. ¹² In Bazley, a non-profit organisation that ran a home for emotionally troubled children was found vicariously liable for the sexual abuse of those children by a substitute parent employed by the organisation. McLachlin J noted that vicarious liability was justifiably imposed on the non-profit organisation so as to deter it from inadequate supervision of the wrongdoing employee. ¹³

In KLB, the claimants were physically abused by foster parents and sued the Government, which knew that the foster parents had a history of abusing the children in their care. McLachlin CJ, writing on behalf of herself and seven other judges, held that the Government was not vicariously liable for the abuse. Imposing vicarious liability, the court believed, would not deter inadequate supervision, because the Government just could not supervise foster parents on a day-to-day basis; stricter monitoring not being a ‘real option’. ¹⁴

The discomfort that arises is this: why should inadequate supervision be deterred in Bazley but not in KLB? The defendant in Bazley being a non-profit organisation with limited resources, surely day-to-day supervision was no less unrealistic in Bazley than in KLB? Arbour J in KLB thought that the Government should be vicariously liable for the foster parents’ abuse, as this would incentivise the Government to inform foster children to tell their social workers if there was abuse in the home. ¹⁵ The other eight judges in KLB did not explain why they were not convinced by Arbour J’s argument.

The second reason: it is unclear what a defendant must do to satisfy the court that its supervision of potential tortfeasors was adequate. Bazley tells us that to escape vicarious liability the defendant must have been ‘as scrupulously diligent as it might feasibly have been’ in employee supervision. ¹⁶ This brings to mind Mason J’s observation, in a case concerning non-delegable duty, that the ‘traditional common law response to the creation of a special danger is not to impose strict liability but to insist on a higher standard of care in the performance of an existing duty’. ¹⁷ So what is being deterred is not supervision that is unreasonably lax, but supervision that, while reasonable, is not as exacting as scrupulous diligence feasibly demands. Another way of putting it is that vicarious liability incentivises feasibly scrupulous diligence. The problem is that the defendant does not know how to meet this standard.

The cases have not provided guidance as to how a defendant is to meet the feasibly scrupulous diligence standard. Surely deterrence is not served if organisations do not know what is sought to be deterred and thus to be avoided. It gets worse. The feasibility descriptor implies that a non-profit organisation with limited resources can escape vicarious liability if it meets a lower standard of supervision than a large and wealthy corporation that performs the same activities. Yet it has been observed that in Canada, following Bazley and Jacobi v Griffiths, ¹⁸ non-profit employers enjoy no immunity from vicarious liability by virtue of their status as such. ¹⁹ The UK position is the same—a defendant’s non-profit status is irrelevant. ²⁰

The third reason: if we try hard enough, we could perhaps cobble together an answer to the question—what must a defendant do to satisfy the court that its supervision of potential tortfeasors is adequate?—but that answer expects the defendant to do the impossible. Failing to reasonably supervise employees is an explicit source of negligence liability for the defendant ²¹ ; other sources of negligence liability include the defendant authorising the wrongdoing, ²² where there is a systems failure, ²³ and where the defendant occupier failed to choose a reasonably competent independent contractor to perform a dangerous activity on his premises. ²⁴ In Bazley, McLachlin J noted that the type of inadequate supervision that vicarious liability sought to deter was a supervision that escapes negligence liability either because the harm was not foreseeable at the time of the negligence, or because the defendant had complied with the industry standard then prevailing. ²⁵ The defendant’s supervision is not negligent because, as Paula Giliker noted, ‘negligence requires the court to judge the conduct of the institution by the standards which prevailed at the time of the tort and not those applied today’, so that without ‘clear evidence (or a concession by the defendant) that a reasonable employer would have identified [the tortfeasor] as a potential abuser, liability will be difficult to establish’. ²⁶ Thus, the trial judge in Blackwater v Plint, which involved the sexual assault of aboriginal children by a dormitory supervisor in a school jointly operated by the two defendants, held that both defendants were not negligent because ‘the children had not been very clear in reporting the abuse and the adults to whom they reported did not realize the children were talking about sexual abuse, an almost unthinkable idea at the time’. ²⁷

In short, if the law seeks to do something by imposing vicarious liability on a defendant who failed to meet a supervision standard that only comes into existence decades later at trial, that something is surely not deterrence. Obviously, nobody can be deterred from failing to meet a standard that he has no way of ever knowing. Put another way, imposing vicarious liability cannot incentivise defendants to do x when he has no way of knowing what x is. Defendants are left chasing their tails to meet a standard that by definition always exists in the future.

The fourth (weak) reason: even a good defendant may not escape vicarious liability. Suppose the school in Blackwater somehow meets the feasibly scrupulous diligence standard of supervising potential tortfeasors. Suppose this supervision allows the timely detection and immediate expulsion of the dormitory supervisor after he had committed one act of sexual assault against the claimant. Would this absolve the defendants in Blackwater? One suspects not—Poland reminds us that a defendant can be vicariously liable for a single deliberate act of its employee. If this is correct, deterrence of inadequate supervision turns out to be insignificant; vicarious liability may be imposed regardless of the adequacy of the defendant’s supervision.

That a good defendant may nevertheless attract vicarious liability diminishes the significance of deterrence as a justification for vicarious liability. But this fourth reason is weak because one suspects that in certain circumstances a good defendant (that met the standard) will escape vicarious liability while a bad defendant (that did not meet the standard) under the exact same circumstances will not. If one compared (i) the number of hypothetical circumstances under which a good defendant is not found vicariously liable (y); with (ii) the number of hypothetical circumstances under which a bad defendant is not found vicariously liable (z), one suspects that y will be greater than z.

Taking stock, the deterrence justification for imposing vicarious liability should be rejected. Where the claimant’s injury is caused by a single deliberate or negligent act of the tortfeasor, the cases never discuss deterrence. Where the claimant’s injury is caused by the tortfeasor’s multiple deliberate acts over time, the deterrence justification is invoked, but this rings hollow for four reasons. The four reasons are in fact symptoms of one overarching problem: the deterrence justification gives the wrong impression that the defendant must be deterred from doing something wrong, but the vicarious liability doctrine has always insisted that whether the defendant committed a wrong is irrelevant. ²⁸ To concurrently assert that vicarious liability (i) deters defendants from committing wrongs; and yet (ii) is blind to whether the defendants are at fault (ie negligent), leads to unbearable tension. Blackwater illustrates. In Blackwater the Canadian Supreme Court upheld the trial judge’s decision to impose vicarious liability 75:25 as between the two defendants who ran the school for aboriginal children. It reasoned that ‘vicarious liability … implies fault’ since ‘vicarious liability is imposed on someone who was in a position to have supervised and thus to have prevented the occurrence of the harm’. ²⁹ This prompted Jason Neyers to remark, with respect to the 75:25 apportionment, that ‘someone who was held liable without fault, could be more at fault than another person, who was held liable without fault’. ³⁰

In 1999, Binnie J cautioned in Jacobi that overemphasizing deterrence would blur the line between vicarious (not-fault-based) and negligence (fault-based) liability. ³¹ Time has shown the prescience of this warning. Indeed, the time has shown that we should reject deterrence altogether as a justification for imposing vicarious liability.

The third justification for the vicarious liability doctrine was stated by Lord Phillips SCJ in the CCWS case: ‘vicarious liability was a doctrine designed for the sake of the claimant, imposing a liability incurred without fault because the employer was treated at law as picking up the burden of an organisational or business relationship which he had undertaken for his own benefit’. ³² Lord Reed SCJ in Armes called this the ‘most influential’ justification in the British case law for the imposition of vicarious liability. ³³ This justification is not restricted to cases involving deliberate wrongs like the CCWS case and Armes. It can and has been deployed to explain many cases where the tortfeasor’s negligence while performing work for the defendant caused the claimant’s injury. ³⁴

The problem with this justification is its vagueness—it can be interpreted in three different ways, and all three are problematic.

The first interpretation of Lord Phillips SCJ’s words in the CCWS case envisages an implied bargain between the defendant and tortfeasor: because the tortfeasor's work benefits the defendant, the defendant promises the tortfeasor that, if the tortfeasor while doing this work inadvertently injures the claimant, the defendant instead of the tortfeasor will in practice be the party that compensates the claimant. In other words, this interpretation, which Jason Neyers calls the indemnification theory, ³⁵ envisages an exchange of benefits. The tortfeasor employee benefits the defendant employer with his work and his loyalty. In exchange, the employer benefits the employee by standing ready to meet claims by those who suffer an injury caused by the employee’s negligence while discharging that work. The indemnification theory should be rejected as it generally fails to explain the cases where defendants were held vicariously liable for deliberate wrongs committed by tortfeasors who were doing work for them.

In most cases, the indemnification theory cannot logically cover deliberate wrongs by the employee. How can we say that the defendant agreed to be answerable in damages for claims by those injured by the deliberate wrongs of its employee? We can expect a defendant agreeing to relieve its employee from the spectre of financially ruinous negligence liability, if such liability stemmed from inadvertence in the course of the employee’s work (and practically, the employee may not be subject to financially-ruinous negligence liability in any real sense if the employee were a worthless defendant). We generally cannot expect a defendant to agree to pay damages for its employee’s deliberate criminal acts. Surely a defendant, if asked to compensate victims of child sexual abuse committed by its employee warden in its boarding house, ³⁶ would retort: ‘I’m equally shocked! These disgusting acts are the complete antithesis of what he was supposed to do!’

The failure of the indemnification theory to account for the courts imposing vicarious liability in deliberate-wrong cases is a good enough reason to reject the theory. The old test for vicarious liability, the Salmond test, was discarded in Lister for exactly the same reason. Prior to Lister, the Salmond test dictated that in most cases, the imposition of vicarious liability on the defendant turned on whether its employee’s wrong was an unauthorised mode of doing an act that it had authorised. ³⁷ Though it was not always clear whether an employee’s conduct was an unauthorised mode of doing an authorised act, ³⁸ the test could sensibly be applied to make a defendant vicariously liable for its employee’s negligence. Generally, however, the test could not be sensibly applied to make the defendant vicariously liable if its employee’s wrong was deliberate. ³⁹ Deliberate sexual abuse of children provided a stark example of the test’s inadequacy. As Lady Butler-Sloss commented in 1998, ‘in the field of serious sexual misconduct, [it is] difficult to visualise circumstances in which an act of the [abuser] can be an unauthorised mode of carrying out an authorised act’. ⁴⁰

In 2002, the British public was rocked by revelations of widespread child abuse in North Wales children’s homes. ⁴¹ The Law Lords, confronted with child sexual abuse committed by a warden of a boarding house in Lister, felt that the vicarious liability doctrine had to be released from the strictures of the Salmond test. The test was discarded and replaced with the close connection test: vicarious liability would be imposed if the warden’s acts of sexual abuse ‘were so closely connected with his employment that it would be fair and just to hold the employers vicariously liable’. ⁴² On the facts of Lister, the close connection test was fulfilled and vicarious liability was imposed.

If the House of Lords found it necessary to discard the Salmond test in Lister, a fortiori the indemnification theory should be rejected. This is because we could conceivably read the Salmond test as applying to some deliberate wrongs, but this escape route is unavailable to the indemnification theory. Claire McIvor has argued that the more specifically the authorised act (in the Salmond test) is defined, the easier it is for the claimant to establish that the deliberate wrong was an unauthorised mode of performing that authorised act (and so succeed in establishing vicarious liability). ⁴³ In Trotman, ⁴⁴ the deputy headmaster of a special school sexually assaulted a mentally handicapped pupil while on a school trip to Spain. The court held that the school was not vicariously liable for the assault because his conduct was ‘far removed from an unauthorized mode of carrying out a teacher’s duties on behalf of his employer’ and was ‘a negation of the duty of the council to look after children for whom it was responsible’. ⁴⁵ McIvor has pointed out that this conclusion hinges on a non-specific description of the deputy headmaster’s authorised act as ‘caring for a handicapped teenager while on a foreign holiday’. Under this description, the assault would obviously be far removed from an unauthorised mode of performing the authorised act, since caring and abusing are irreconcilable. But if the authorised act was defined more specifically as ‘to share a room with the boy at night’, it becomes easier to construe the sexual assault as an unauthorised mode of carrying out this act, ⁴⁶ and so impose vicarious liability. The Law Lords in Lister could have used this method so that the Salmond test could be applied to deliberate wrongs generally. But this escape route is closed to the indemnification theory. And this only strengthens the case for rejecting it.

The inadequacy of the indemnification theory is exacerbated by the fact that vicarious liability cases that involve deliberate wrongdoing are common. Besides the child sexual abuse cases, ⁴⁷ vicarious liability has also been imposed in cases that involve assaulting customers ⁴⁸ and colleagues, ⁴⁹ defrauding customers, ⁵⁰ stealing customer property ⁵¹ and punching a member of the opposite team in the course of a rugby match. ⁵² Guards and policemen who deliberately shot claimants for no good reason have also attracted vicarious liability for their employers. ⁵³

In short, the indemnification theory should be rejected because it generally fails to explain why defendants can be vicariously liable for deliberate wrongs committed by tortfeasors who do work for them. Lord Steyn remarked that the Law Lords in Lister discarded the Salmond test so as to do justice in cases of deliberate wrongdoing. ⁵⁴ The indemnification theory should share the same fate.

The second interpretation, which we will call the benefit theory, takes Lord Phillips SCJ’s words in the CCWS case at face value: the defendant should foot the bill arising from the tortfeasor’s wrongfully performed work, because the tortfeasor’s work benefits the defendant.

This interpretation is really of ancient origin. ⁵⁵ However, it expresses a conclusion rather than an argument—we are not told why, merely because the defendant benefits from the tortfeasor performing work for it, it must logically compensate a claimant who suffers injury caused by the tortfeasor’s deliberate or careless conduct. If there were an exchange of promises between society and the defendant, that might provide a solution. But we cannot generate an implied society-defendant bargain here. The benefits flow from the tortfeasor to the defendant (the tortfeasor doing work for the defendant). Reciprocity demands that the defendant benefit the tortfeasor in some way. However, under the benefit theory, the defendant benefits society at large (by compensating the claimant). The benefits flow to the wrong party! (No benefit flows to the tortfeasor because even absent vicarious liability, the tortfeasor does not foot the bill in any real sense, as he is a worthless defendant.)

Another reason for rejecting the benefit theory is that it is too blunt a tool for explaining why vicarious liability is imposed in some cases but not in others. Consider four cases: Bazley, Lister, Jacobi and KLB. All involved defendant organisations were interested in advancing the welfare of children. In all four cases individuals assigned that work by the defendants deliberately abused certain children. Yet vicarious liability was imposed in the former two cases but not the latter two. Under the benefit theory, all four defendants should have borne the cost of the wrong as, ignoring the wrong committed, they had benefitted from the tortfeasors’ work for them.

The third and final interpretation of Lord Phillips SCJ’s words in the CCWS case de-emphasises the idea that the defendant benefits from the tortfeasor doing work for it. Instead, a macro perspective is taken and what is examined is the relationship between the defendant organisation and society at large. This examination finds us an implied defendant-society bargain: the defendant should bear the cost of the tortfeasor’s wrong as the price exacted by society in exchange for being allowed to operate in society to advance its interests. This interpretation, which we will call the quid pro quo theory, has two problems.

The first problem with the quid pro quo theory is that it assumes that society is justified in exacting a price (namely, vicarious liability) from the defendant organisation, in exchange for granting that organisation the freedom to operate in society to advance its interests. Vicarious liability, so the quid pro quo theory goes, is the price the organisation has to pay to operate in society. The problem is that no justification is proffered for why an organisation has to pay to operate in society. After all, individuals are sovereign; we do not conduct our affairs only at the largesse of society. And if we are not blameworthy in conducting our affairs, we generally do not ‘pick up the burden’ if someone gets injured as a result. One would expect an organisation, being a group of individuals pursuing common interests, to be treated likewise; ie it should not pay for injury to society (and more specifically claimants) in the absence of blameworthy conduct. The quid pro quo theory does not explain why organisations should be treated differently from individuals.

The second problem with the quid pro quo theory is that it necessarily implies that any and all wrongdoing by any member of the defendant organisation ⁵⁶ should attract vicarious liability for the defendant. But that is not the case. For example, Bazley and Jacobi both involved employees who sexually abused children while working in organisations interested in advancing the welfare of children. Under the quid pro quo theory, vicarious liability should have been imposed in both cases as the price for the defendants’ freedom to operate in society and advance their interests. But vicarious liability was imposed in Bazley and not in Jacobi.

We pause to take stock. A justification for vicarious liability, to qualify as such, has to explain why vicarious liability is imposed in some situations but not in others. In this Part, we explored three of the justifications for vicarious liability proffered in the case law. The first two of these three justifications are (a) effective compensation through spreading the cost of the claimant’s injury; and (b) deterring the defendant from inadequately supervising potential tortfeasors who do work for them. Both do not explain the cases and should be rejected. The third justification is that the defendant should pick up the burden of an organisational or business relationship that he had undertaken for his own benefit. This third justification can in turn be interpreted to mean three different things, which we have called the indemnification theory, the benefit theory and the quid pro quo theory. We explained why these theories are inadequate in explaining the cases. In the next Part, we explore the fourth justification and show why it explains the cases and thus should be embraced.

The fourth justification: risk-through-placing

The fourth justification for the vicarious liability doctrine is: vicarious liability exists to ensure that the defendant bears the cost of the tortfeasor’s wrong, and it should bear that cost because, by placing the tortfeasor in the position in its organisation that it did, it had created or significantly enhanced the risk of the claimant’s injury. As Lord Phillips SCJ stated in CCWS: ‘vicarious liability is imposed where a defendant, whose relationship with the abuser put it in a position to use the abuser to carry on its business or further its own interests, has done so in a manner which has created or significantly enhanced the risk that the victim or victims will suffer the relevant abuse. ⁵⁷ Another account is given by Lord Nicholls:

Whether an act or omission was done in the ordinary course of a firm’s business cannot be decided simply by considering whether the partner was authorised by his co-partners to do the very act he did. The reason for this lies in the legal policy underlying vicarious liability. The underlying legal policy is based on the recognition that carrying on a business enterprise necessarily involves risks to others. It involves the risks that others will be harmed by wrongful acts committed by agents through whom the business is carried on. When these risks ripen into loss, it is just that the business should be responsible for compensating the person who has been wronged. ⁵⁸

We will call this justification—that the defendant is vicariously liable because, by placing the tortfeasor in the position in its organisation that it did, it had created or significantly enhanced the risk of the claimant’s injury—the risk-through-placing justification.

The first limb of this justification—placing the tortfeasor in the position in its organisation that it did—is not to be understood as the defendant being negligent in appointing a tortfeasor who was unsuitable for the task at hand, due to a lack of skill or defect in character or both. A direct negligence claim would lie against such a defendant; vicarious liability is not engaged. Rather, the phrase ‘by placing the tortfeasor in the position in its organisation that it did’ refers to the privileges and demands of that particular position in the defendant’s organisation; the entire description of that position, so to speak. The characteristics of the tortfeasor who occupies that position are disregarded.

This first limb then shades into the second—the defendant’s ‘placing’ action thereby creating or significantly enhancing the risk of the claimant’s injury. Again, the tortfeasor’s characteristics are not what creates or significantly enhances the risk of the claimant’s injury. And for the same reason: a direct negligence claim would lie against the defendant if it had so placed an unsuitable tortfeasor in its organisation, precisely because placing that unsuitable individual would have created or significantly enhanced the risk of the claimant’s injury. Rather, the risk-through-placing justification is somewhat subtler: vicarious liability is justified on the basis that the defendant organisation had a position in its organisation to fill, and by virtue of filling that position (with the tortfeasor or for that matter anyone else) it simultaneously created or significantly increased the risk of the claimant’s injury. In other words, the attention is always on the privileges and demands of the position that the defendant organisation ‘put out there’. No attention is paid to the suitability of the particular tortfeasor in relation to exercising those privileges and meeting those demands.

The risk-through-placing justification adequately explains the cases—it can explain why vicarious liability is imposed in some cases but not in others. Therefore, the risk-through-placing justification should be embraced. I explain.

The cases on vicarious liability can be divided into three waves. In all three waves, the vicarious liability doctrine has been expressed by a two-rule formula (and both rules must be fulfilled for vicarious liability to be imposed); the first prescribing the necessary features of the tortfeasor–defendant relationship, and the second prescribing the necessary connection between that relationship and the tortfeasor’s wrong (that resulted in the claimant’s injury). Precisely how the two rules were expressed changed from one wave to the next—when one of the rules was found to have been pitched at too low a level of generality—but the formula’s two-rule structure has never changed. All three waves of cases on vicarious liability can be explained by the risk-through-placing justification.

The first wave of cases embraced the Salmond test: vicarious liability was imposed only if (i) the tortfeasor was the defendant’s employee; and (ii) the wrong that the tortfeasor committed was in the course of employment in that it was an unauthorised mode of doing an authorised act. ⁵⁹

The Salmond test can be explained by the risk-through-placing justification—ie the defendant by placing the tortfeasor in the position in its organisation that it did (with the accompanying privileges and demands of that position), created or significantly enhanced the risk of the claimant’s injury. In other words, we have the Salmond test in order to get to the same result that we would have gotten if vicarious liability were imposed only where the risk-through-placing justification was engaged.

The first part of the Salmond test—requiring the tortfeasor to be the defendant organisation’s employee (and many first-wave cases laid down the features of an employment relationship) ⁶⁰ —is reflected in the risk-through-placing justification’s requirement that the defendant place the tortfeasor in the position in its organisation that it did. Obviously, an employee is someone to whom the defendant organisation assigned a position and the privileges and demands that went with that position.

The second part of the Salmond test—requiring an unauthorised mode of doing an authorised act—is also reflected in the risk-through-placing justification. The requirement for an ‘authorised act’ corresponds to the necessity of the defendant placing the tortfeasor in a position with the accompanying privileges and demands of that position. The ‘unauthorised mode’ requirement corresponds to the risk that someone in that position, in exercising those privileges or meeting those demands, does so negligently, thereby creating or significantly enhancing the risk of injury to others. ⁶¹

Indeed, many first-wave cases insist on the wrong being close (a) in time and space to employment; and (b) to the tortfeasor’s work function (ie purpose of employment), before vicarious liability could be imposed. ⁶² We could say that this thereby limits vicarious liability to situations where the privileges and demands of the position (and not the subjective characteristics of the tortfeasor) necessarily create or significantly enhances the risk of the injury to others. So again, applying the Salmond test yields the same results as if vicarious liability were imposed only where the risk-through-placing justification was engaged.

Century Insurance ⁶³ illustrates how the result under the Salmond test reflects the result under the risk-through-placing justification. As discussed, Davison’s job was to drive his employer’s petrol tanker and deliver petrol to garage proprietors, and one day while filling the storage tank of a garage proprietor, Davison lighted a cigarette and threw away the lighted match, causing a fire that ultimately damaged property.

The House of Lords held that Davison’s acts were within his course of employment. This conclusion can be explained on the basis of the risk-through-placing justification: in employing someone to fill the storage tanks of garage proprietors with petrol from its petrol tank, the employer had increased the risk of injury to others if such filling was performed negligently. As Lord Wright said: ‘The negligence is to be found by considering the time when and the circumstances in which the match is struck and thrown down’. ⁶⁴

So much for the first wave. In the second wave of vicarious liability cases, the second limb of the Salmond test was replaced by the close connection test. Under the Salmond test, the requirement that the wrong be an ‘unauthorised mode of doing an authorised act’ rendered a defendant vicariously liable for the careless but (generally) not deliberate wrong of its employee. ⁶⁵ This inadequacy was tolerated for a while; some cases resorted to principles of bailment ⁶⁶ or agency ⁶⁷ or both ⁶⁸ in holding defendant organisations liable for property or monetary loss caused by their employees’ deliberate wrongs.

All that changed, however, when courts were confronted with a swath of cases involving claimants who were sexually abused as children by employees of schools and care homes. In Lister, the Law Lords realised that the ‘unauthorised mode of doing an authorised act’ had pitched the issue at too low a level of generality. They abstracted from the Salmond test to a more general principle—vicarious liability may be imposed for an employee tortfeasor’s wrong, deliberate or not, if there was a sufficiently close connection between the wrong and his employment. ⁶⁹

This abstraction exercise was not a radical move; May J had, 25 years earlier, extracted from the Salmond test the more general principle of closeness between wrong and employment. ⁷⁰ With the close connection test, the general immunity of defendants for the deliberate wrongs committed by their employees in the course of employment was no more. Indeed, courts in the second wave remarked that some first wave cases, in which vicarious liability was not imposed because the tortfeasor committed a deliberate wrong, would have been decided differently under the close connection test. ⁷¹

The close connection test has been endorsed and consistently applied in cases involving employees who commit not just child sexual abuse but also other types of deliberate wrongs. ⁷² As mentioned earlier, these cases have involved employees assaulting customers and colleagues, defrauding customers, stealing customer property, punching a member of the opposite team in the course of a rugby match, and deliberately shooting claimants for no good reason (where the employee is a guard or policeman). ⁷³ The close connection test is also used in Canada. ⁷⁴

The close connection test can be explained by the second limb of the risk-through-placing justification—ie the defendant created or significantly enhanced the risk of the claimant’s injury. In other words, we have the close connection test in order to get to the same result that we would have gotten if vicarious liability were imposed only where the defendant had increased the risk of the claimant’s injury by ‘putting out there’ a position (occupied by the tortfeasor) with its accompanying privileges and demands, There are two ways we know this.

First, the cases show us that a close connection is found precisely where the defendant organisation increased the risk of the claimant’s injury by ‘putting out there’ a position (occupied by the tortfeasor) with its accompanying privileges and demands. There are two reasons we know this.

The first reason: some cases tell us so explicitly. ⁷⁵ In fact, Bazley explicitly identifies as fulfilling the close connection test situations of created or significantly enhanced risk in the context of deliberate wrongs—namely, where the defendant had conferred on the tortfeasor power over a claimant who was vulnerable, the tortfeasor had the opportunity to abuse that power, and the tortfeasor’s deliberate wrong related to friction, confrontation or intimacy involved in the defendant’s enterprise. ⁷⁶ Australia similarly adopts a risk-centric approach—although the explicit language of ‘close connection’ is not adopted, Prince Alfred College tells us that, for vicarious liability to be imposed, the defendant must have created or significantly enhanced the risk of the tortfeasor committing the wrong by placing him in a position that gives ‘occasion for the wrong’, having reference to ‘authority, power, trust, control and the ability to achieve intimacy with the victim’. ⁷⁷

The second, albeit indirect, reason: the cases consistently tell us that vicarious liability can only be imposed for deliberate wrongs if the wrongs were reasonably incidental to the purpose for which the tortfeasor was employed. ⁷⁸ This is arguably the same as saying that it is only the materialisation of risks that arise from the defendant organisation ‘putting the position out there’, with its accompanying privileges and demands, that can attract the imposition of vicarious liability.

There is a second way to know that the function of the close connection test is to get us to the same result that we would have gotten if vicarious liability were imposed only where the defendant had increased the risk of the claimant’s injury by ‘putting out there’ a position (occupied by the tortfeasor) with its accompanying privileges and demands. It is this: (a) different outcomes ensue when the close connection test is applied in ostensibly similar second-wave cases; and (b) these disparate outcomes correspond to the degree by which the defendant’s conferral of privileges and imposition of demands on the position occupied by the tortfeasor increased the risk of the claimant’s injury; the greater the risk, the more likely the close connection test is fulfilled. Thus, the defendant organisation was vicariously liable for child sexual abuse in Bazley but not in Jacobi, because the tortfeasor in Bazley was conferred greater power over the victim, thus leading to a higher risk of abuse. ⁷⁹ Similarly, KLB is distinguishable from Lister and Bazley because the deliberate wrongs were committed on the defendant’s premises in the latter two cases but not in KLB; ie the tortfeasor in the latter two cases (but not KLB) abused the privilege of access and control over the defendant’s premises in carrying out their nefarious acts. ⁸⁰

Mohamud illustrates how the result under the close connection test reflects the result under the risk-through-placing justification. Mr Khan was employed by the defendant to serve customers at a petrol station. The claimant approached Mr Khan to ask if it was possible to help him print some documents from a USB stick he was carrying. Mr Khan, who was behind the counter, responded abrasively. When the claimant protested being spoken to in such a manner, Mr Khan used foul language to order the claimant to leave. The claimant returned to his car. Mr Khan followed him to his car, told him never to come back, and then physically attacked him.

The Supreme Court held that Mr Khan’s employer was vicariously liable for the claimant’s injuries. Lord Toulson, with whom the other members of the court agreed, reasoned that the close connection test was fulfilled, stating:

[I]t was Mr Khan’s job to attend to customers and to respond to their inquiries. His conduct in answering the claimant’s request in a foul-mouthed way and ordering him to leave was inexcusable but within the ‘field of activities’ assigned to him. What happened thereafter was an unbroken sequence of events. ⁸¹

The same result applies under the risk-through-placing justification. Vicarious liability is engaged because the defendant had ‘put out there’ the position of a customer-facing officer in the petrol station. And the position’s privileges and demands created a risk that the customer-facing officer would verbally abuse and physically assault a customer, as happened in Mohamud. ⁸² (Indeed, ‘putting out there’ customer-facing officers creates a myriad of risks, as can also be seen in Morris, where the employee of a furrier took in a customer’s mink stole for cleaning and absconded with it.) ⁸³ The risk that materialised must be incidental to the employment. Thus, Mr Khan’s employer would probably not be vicariously liable if Mr Khan had, while on vacation, physically assaulted a tourist on the same tour bus. The assault would not be closely connected to Mr Khan’s employment. Neither would it be the materialisation of a risk that the defendant created by ‘putting out there’ a customer-facing position. As the Court of Appeal recently stated:

It is well-established that merely providing the opportunity for wrongdoing is not sufficient without more to give rise to vicarious liability, absent a holding out of someone in [the tortfeasor’s] position as having authority to act for the defendant sought to be made vicariously liable. ⁸⁴

So much for the second wave of vicarious liability cases. We are now in the third wave. The second wave allowed courts to impose vicarious liability on defendant employers for the deliberate wrongs of their employees. But this still meant that defendants could escape vicarious liability for wrongs committed by non-employees.

All that changed with the CCWS case. The tortfeasors in the CCWS case were lay brothers of the defendant religious Institute that assigned them to teach in the school where the claimant was housed and educated as a child. The claimant sought to make the defendant vicariously liable for sexual abuse he suffered at the hands of the lay brothers while he was at the school. Although the tortfeasors were not employees of the defendant, Lord Phillips SCJ held that the defendant was vicariously liable, because being lay brothers the tortfeasors were ‘so much a part of the work, business or organisation’ of the defendant that they were akin to employees. ⁸⁵

Just as in Lister, here was an abstraction from the ‘employee requirement’ to a more general principle—vicarious liability may be imposed in relation to not only tortfeasors who were employees, but also tortfeasors who were like employees in that they were part of the defendant’s organisation. With the liberalisation of the ‘employee requirement’, courts could now impose vicarious liability for the wrongs of tortfeasors who stood in non-employment but close relationships with defendants, as seen in Cox (prisoner and prison authority) and Armes (foster parent and local social services authority).

To test whether a tortfeasor was part of the defendant’s organisation obviously reflects the first limb of the risk-through-placing justification; ie the defendant placed the tortfeasor in the position in its organisation that it did. These two statements mean the same thing; the defendant assigning to an individual a position in its organisation, with that position’s privileges and demands, necessarily implies that the individual is a part of the defendant’s organisation.

To conclude, the fourth justification for the vicarious liability doctrine, also called the risk-through-placing justification, states: vicarious liability exists to ensure that the defendant bears the cost of the tortfeasor’s wrong, and it should bear that cost because by placing the tortfeasor in the position in its organisation that it did, with that position’s privileges and demands, it had created or significantly enhanced the risk of the claimant’s injury. We have also clarified that the focus is always on the attendant risks of the position that the defendant ‘put out there’; consequently, no attention should be paid to the characteristics of the tortfeasor who was placed in that position. ⁸⁶

The risk-through-placing justification is the only one of the four justifications proffered in the case law that adequately explains the three waves of vicarious liability cases. Consequently, it should be embraced and the other three justifications rejected. As Lord Reed JSC noted in Cox, vicarious liability focuses ‘upon the business activities carried on by the defendant and their attendant risks’ and ‘reflects prevailing ideas about the responsibility of businesses for the risks which are created by their activities’. ⁸⁷

We have seen that the cases can be explained by the risk-through-placing justification. But is this justification theoretically attractive, even if we ignored consequentialist arguments? While the question is beyond this paper’s scope, I suggest that the risk-through-placing justification can derive some ‘legal coherence attraction’ from the fact that it is another instantiation of the law linking no-fault liability with risk.

The risk-through-placing justification tells us that a defendant to be vicariously liable must have created or significantly enhanced the risk of the claimant’s injury. This link between no-fault liability and risk-creating-or-enhancing conduct is not new in the law. It is not uncommon for the law to impose no-fault liability on defendants that create unusual risks which unfortunately materialise. Prominent scholars call this phenomenon outcome-responsibility, and there is a wide literature on the subject. ⁸⁸ Thus, no-fault liability for damage stemming from ‘abnormally dangerous activities’ ⁸⁹ on the defendant’s land, culminating in an escape of dangerous substances from that land and consequent damage to its surroundings, has been justified on the basis that we can ‘tolerate the activity on condition that it pay its way regardless of whether it is carried on carelessly or not’. ⁹⁰

Christine Beuermann’s work on no-fault liability for the torts of others also appears to link no-fault liability with risk. In her book, ⁹¹ Beuermann painstakingly examines the cases on both vicarious liability and non-delegable duty and suggests that the only relationships (apart from the foster care relationship) which generally give rise to no-fault liability for the tort of another are the employment relationship, the school relationship and the agency relationship. ⁹² Beuermann then explains why this is so: in all three relationships, there is a risk that (a) the defendant abuses its authority in putting an employee in a position that the employee unduly prioritises compliance with the defendant’s directions over other relevant considerations like safety, and this ultimately leads to a claimant’s injury; or (b) someone conferred authority by the defendant abuses that authority. ⁹³

Beuermann’s central thrust is unmistakable—vicarious liability is imposed or a non-delegable duty is found, in cases where there is a potential abuse of authority either by the defendant or someone conferred authority by the defendant. Arguably, this provides implicit support for the risk-through-placing justification. After all, another way of describing a potential abuse of authority is to say that placing the tortfeasor in a position in the defendant’s organisation, with that position’s privileges and demands, created or significantly enhanced the risk of injury to the claimant.

Therefore, the ‘legal coherence attraction’ of the risk-through-placing justification can be expressed as such:

Vicarious liability is a form of no-fault liability.

Implications of embracing the Risk-Through-Placing justification

What are the implications of embracing the risk-through-placing justification as the justification for the vicarious liability doctrine? Three suggestions follow.

First, the five factors laid down in the CCWS case (CCWS Factors) should be clarified so that only tortfeasors who are part of the defendant’s organisation (having been placed there by the defendant), and not independent contractors engaged by the defendant, may trigger vicarious liability for the defendant. The CCWS Factors are used to determine whether the tortfeasor is part of the defendant’s organisation. They thus reflect the risk-through-placing justification’s insistence that the defendant’s risk creation/enhancement must result from the defendant placing the tortfeasor in the position in its organisation that it did. The five factors (CCWS Factors) are ⁹⁵ :

The defendant is richer or better insured than the tortfeasor.

Factor (d)—the defendant created the risk of the wrong being committed—should also be discarded. It confuses the first rule of vicarious liability with the second. Risk creation or enhancement is exclusively the concern of the second rule of vicarious liability. But the CCWS Factors go toward only the first rule, which is concerned exclusively with prescribing the necessary defendant–tortfeasor relationship (under which the defendant can be said to have created or significantly enhanced the risk of the wrong being committed). Keeping factor (d), which goes towards the second rule, needlessly compromises the two-rule structure of the vicarious liability formula.

This leaves us with factors (b), (c) and (e). Collectively, they require the wrong to be committed in the course of activities undertaken for the defendant, where such activities are part of the business activity or regular concern of the defendant, ⁹⁸ by a tortfeasor who was to some extent under the defendant’s control. Confusion arises because these three factors are pitched at such a high level of generality that potentially any tenuous link between the defendant and the tortfeasor can fulfil the first rule. Vicarious liability for independent contractors has long been anathema to the vicarious liability doctrine; thus ‘the general rule is that a defendant will be liable for the negligent act of a servant committed in the course of his employment but not for the negligent act of an independent contractor’. ⁹⁹ Under the CCWS Factors, however, liability for the negligence of independent contractors, long thought to be the exclusive province of the non-delegable duty doctrine, ¹⁰⁰ is suddenly within the reach of vicarious liability.

Why is this so? Tortfeasors who are independent contractors can fulfil factors (b), (c) and (e) of the CCWS Factors. An independent contractor who commits a wrong while working for a defendant could be said to be performing activities undertaken for that defendant. One can also envisage that defendant outsourcing one of its many ‘business activities’ or ‘regular concerns’ to this particular independent contractor. Further, the independent contractor could be said to be under the defendant’s control, in that the latter could direct what the former did in terms of things like work scope, specifications and/or timelines.

Because factors (b), (c) and (e) are pitched at such a high level of generality, James Plunkett warned in 2016 that ‘it now is difficult to see any justification for refusing to extend vicarious liability to independent contractors’. ¹⁰¹ And indeed that warning proved prescient. In the 2018 Court of Appeal case of Barclays Bank plc v Various Claimants, ¹⁰² vicarious liability was imposed on a bank for acts of sexual assaults committed by an independent contractor doctor while conducting medical examinations on the bank’s employees and prospective employees, where such examinations were prerequisites for continued employment or offers of employment. The court held that the effect of the CCWS Factors was that defendants could be vicariously liable for the wrongs of those they had engaged as independent contractors. ¹⁰³ Whether the first rule of vicarious liability was fulfilled was to be determined exclusively by reference to the CCWS Factors, and on the facts, all five factors were fulfilled. Factor (b) was fulfilled: the medical examinations were undertaken on behalf of the bank since the examinations ‘tended to ensure fit entrants’ (ie employees) who were ‘able to give long service to the Bank’. ¹⁰⁴ Factor (c) was fulfilled: ‘there could hardly be a clearer example’ of the bank’s business activity than ‘the selection of suitable employees for a responsible institution in the service sector’. ¹⁰⁵ Factor (e) was fulfilled because the bank was ‘directional in identifying the questions to be asked and the physical examination to be carried out by the doctor for the purpose of completing the [bank’s] templated form’. ¹⁰⁶

How can factors (b), (c) and (e) be clarified? The answer is by interpreting them in light of the risk-through-placing justification. The risk-through-placing justification says that the defendant should bear the cost of the tortfeasor’s wrong because it placed the tortfeasor in the position in its organisation that it did (and this created or significantly enhanced the risk of the tortfeasor committing the wrong). The first rule of vicarious liability reflects this idea by prescribing that the tortfeasor be part of the defendant’s organisation; he must be ‘so much a part of the work, business or organisation’ of the defendant that he was akin to an employee. In other words, factors (b), (c) and (e), although expressed using general words, are meant to tell us whether the tortfeasor is part of the defendant’s organisation. The interpretative key lies in the fact that Lord Phillips SCJ, in reforging the ‘employee requirement’ into the ‘part of organisation’ requirement in the CCWS case, drew heavily from Ward LJ’s judgment in E v English Province of Our Lady of Charity. ¹⁰⁷ That case concerned a visiting priest who sexually abused the claimant when the claimant was in a children’s home in the bishop’s diocese. Himself relying on an article by Richard Kidner, ¹⁰⁸ Ward LJ deemed the tortfeasor priest as the bishop’s employee, so that vicarious liability was attached to the defendant’s diocesan trust through the bishop. Ward LJ explained this conclusion on the basis that (i) the priest by ministering to the souls of the faithful had a central role in meeting the Roman Catholic Church’s business (ie its regular concern) of spreading the word of God (called the organisation test) ¹⁰⁹ ; (ii) the priest was wholly integrated into the organisational structure of the church’s enterprise (called the integration test) ¹¹⁰ ; the priest was not an entrepreneur; he was not taking the appropriate risks and enjoying the resulting profits (called the entrepreneur test) ¹¹¹ ; and (iv) the priest was accountable to the bishop, who could tell him what to do and not do but not how to do it (called the control test). ¹¹²

Integration, of course, merely describes the state where the tortfeasor is part of the defendant’s organisation. In the CCWS case, the integration test was discarded; instead, the organisation, entrepreneur and control tests would all serve to answer the question: is the tortfeasor part of (ie integrated into) the defendant’s organisation? Seen in this light, factors (b), (c) and (e) correspond to the entrepreneur, organisation and control tests, respectively.

Factor (b)—the wrong was committed in the course of activities that the tortfeasor undertook on behalf of the defendant—refers to the entrepreneur test. The words ‘undertook on behalf of the defendant’ should be understood as ‘undertook not on his own account’; ie not as an entrepreneur.

Factor (c)—the activity was likely part of the business activity of the defendant—refers to the organisation test. The business activity of the defendant should be understood as ‘the regular concern and core objective of the defendant’.

Lastly, factor (e)—the tortfeasor was to some extent under the defendant’s control, meaning that the defendant could direct what the tortfeasor did, not how he did it—obviously refers to the control test.

The three tests—entrepreneur, organisation and control—are not immune from criticism. For one, they overlap substantially; a specific fact often goes towards more than one of them. ¹¹³ Thus, an individual who does not have mandated working times is by definition subject to less control by the defendant. But this also indicates that the individual is in business on his own account, because someone who is in business on his own account has the freedom to decide when to work.

Further, one of the tests—the control test—is fraught with uncertainty. Although the current definition is provided by the CCWS case and Cox, ¹¹⁴ scholars have put forward many other definitions. Phillip Morgan has said that control has two meanings: (i) whether the tortfeasor is part of the defendant’s team (which he calls day-to-day control); and (ii) the defendant’s right to direct the tortfeasor as to how to perform the task or achieve a stipulated end. ¹¹⁵ Peter Watts has his own version of the two meanings of control: (i) ‘micro-control’, whose presence indicates that the tortfeasor is part of the defendant’s organisation but whose absence is a neutral fact; and (ii) the duty of obedience and loyalty owed by employees to employers. ¹¹⁶ And Kidner has also suggested two other meanings of control: (i) the degree to which the tortfeasor is accountable to his employer; ie to what extent he is subject to managerial procedures of his employer in relation to matters such as how he arranges his work and his use of assets; and (ii) the extent to which the tortfeasor is allowed to control himself. ¹¹⁷ It is obvious that all these definitions substantially overlap.

The lesson here is that it is preferable to do away with the three tests and instead view each specific fact as either going towards or against the tortfeasor being ‘part of the defendant’s organisation’, and then weigh all the facts on both sides of the line to make a finding on the issue. But this is not a major issue; the courts do not seem to have trouble reaching just and sensible outcomes when they use the three tests. Being relatively harmless they can be tolerated.

To recap, the CCWS Factors are meant to determine when the first rule of vicarious liability is fulfilled. But the presence of factor (d), and factors (b), (c) and (e) being pitched at such a high level of generality, have caused the five factors to lose their original purpose. Thus, in Cox, Lord Reed SCJ described factors (b), (c) and (d) as ‘inter-related’, in that they collectively reflect the essential idea that ‘the defendant should be liable for torts that may fairly be regarded as risks of his business activities’. ¹¹⁸ Coupled with the downplaying of the significance of control in Cox and Armes, ¹¹⁹ the result is a conversion of the CCWS Factors to now serve predominantly the second risk-creation/enhancement rule of vicarious liability. Barclays Bank, with its ‘ground-shifting … refusal to place any weight on the fact that the doctor could properly be categorised as an independent contractor’, ¹²⁰ is the high and perhaps inevitable point of that conversion.

Thankfully there is some resistance in the Court of Appeal. ¹²¹ The Supreme Court has also overruled Barclays Bank, stating that (a) individuals who carry on business on their own account do not attract vicarious liability for the defendant organisation they perform work for; and (b) the use of the CCWS Factors should be confined to doubtful cases. ¹²² Notwithstanding the Supreme Court’s decision, the foregoing discussion is still of value—for the reasons above, factors (a) and (d) should be removed, and factors (b), (c) and (e) clarified to correspond to the entrepreneur, organisation and control tests, respectively. Only then will the three factors once again serve the ‘part of the organisation’ requirement of the vicarious liability doctrine.

The second implication of embracing the risk-through-placing justification is that we now have a good reason for rejecting agency reasoning in vicarious liability. Agency law and vicarious liability share many similarities. Both are concerned with defining the circumstances under which an individual’s act may be attributed to an organisation. ¹²³ Both doctrines share a regard for whether the tortfeasor is part of the defendant’s organisation. Some cases can attract arguments based on agency and vicarious liability, especially in the context of misappropriation of funds and property. ¹²⁴ Indeed, Anthony Gray would subject the vicarious liability doctrine to agency reasoning; he argues that a defendant should only be vicariously liable if the tortfeasor had actual or apparent authority to act. ¹²⁵

Despite all this, the case law has rejected making vicarious liability a subset of agency law. It was remarked in Rose that vicarious liability was ‘not a principle which derives from a critical or refined consideration of other concepts in the common law, for example, …the concept of agency…’, and that vicarious liability is not imposed on the basis of the servant’s authority. ¹²⁶ McKendrick has noted that ‘little progress has been made in the case law in invoking vicarious liability in the case of agents, except in the rather anomalous cases concerning the liability of vehicle owners when the vehicle is being driven by somebody else and the liability of a trade union for the acts of its officials and shop steward’. ¹²⁷ Francis Reynolds and Tan Cheng Han have remarked that the liability imposed in McKendrick’s driving cases (where it was sought to make the owner of a car liable for the acts of others whom he allowed to drive) was ‘always difficult to account’. ¹²⁸ More crucially, since defendants do not generally authorise their agents to commit deliberate wrongs, Gray would not have vicarious liability imposed for deliberate wrongdoing. But the cases show that that horse has bolted.

Why is this so? The risk-through-placing justification provides an answer. Although both agency law and vicarious liability tell us when an individual’s act may be attributed to an organisation, they employ different reasoning. Agency approaches the attribution question through the lens of actual or apparent authority of the tortfeasor and reliance on such authority by the claimant. Vicarious liability, as this paper suggests, employs altogether different reasoning: a defendant is vicariously liable because it placed someone in its organisation thereby creating or significantly enhancing the risk of the claimant suffering injury; authority and reliance are irrelevant to vicarious liability.

The third implication of the risk-through-placing justification is that courts should stop referring to ‘enterprise risk’ as a justification for imposing vicarious liability. ¹²⁹ This phrase should be rejected as it seems to have no settled meaning, so that it is routinely invoked to encompass many different ideas. In 2003, Simon Deakin explained the enterprise risk approach as placing liability upon an enterprise for dealing with the risks it creates, and proffered two justifications for it: (a) spreading the cost of the injury which was caused by the wrongdoing (through insurance or otherwise); and (b) deterrence, since the enterprise through the powers of managerial coordination is better able to limit the risk. ¹³⁰ In 2015, David Tan said that enterprise risk posits that ‘any institution would be vicariously liable for the acts of its agents where the risk of injury could be said to be closely associated with the wrong that occurred, and it would be just, fair and reasonable that the entity that engaged in, and profited from or gained a non-economic benefit from, particular activities should internalise the full cost of operations, including potential torts’. ¹³¹ This was echoed by James Plunkett in 2016; the enterprise risk approach is ‘the idea that where a body uses another person to advance their interests, and thereby introduces an inherent risk of injury to others, if the body is to reap the rewards of doing so, it is only fair that they also accept the consequences when those risks materialise—they must take the bad with the good’. ¹³²

According to Deakin, Tan and Plunkett, under the enterprise risk approach, vicarious liability is imposed because the defendant enterprise’s activity creates a risk of injury. This is no different from the second limb of the risk-through-placing justification. However, Deakin, Tan and Plunkett differ as to why this risk creation in itself (without fault) justifies the imposition of vicarious liability. Deakin invokes cost-spreading and deterrence reasons. Plunkett and Tan, on the other hand, invoke the same reasoning found in the quid pro quo theory. (This paper has criticised all three reasons as inadequate justifications for the vicarious liability doctrine.) Tan also details the socio-economic benefits of embracing the enterprise risk approach: it would ‘increase social welfare’, ‘lower social marginal costs’, ‘lead to the employment of … more responsible agents with an attendant increase in loss-avoidance effort’, and ‘lead to a decrease in the number of injuries’. ¹³³ If there is no agreement amongst scholars as to exactly what reasons for imposing vicarious liability are encompassed by the words ‘enterprise risk’, we can hardly expect the courts to do better. Therefore, unless courts clarify exactly what reasons are encompassed by the words ‘enterprise risk’, it would be better for courts to stop using those words.

Conclusion

The importance of the vicarious liability doctrine will continue to grow because, as Milsom remarked in a different context, ‘modern technology makes fault harder to locate, and capable of causing damage out of all proportion to the fault itself’. ¹³⁴ For better or worse, Lord Hughes SCJ’s argument in Armes that vicarious liability should be confined to exceptional situations because it is a no-fault liability (and which led to his dissent in that case) ¹³⁵ appears to be losing ground.

A justification for vicarious liability must explain why vicarious liability is imposed in some situations but not in others. Unfortunately, the cases have not matched the increasing prominence of the doctrine with a sufficiently clear justification for it. As Longmore LJ asked: ‘Is it that the law should impose liability on someone who can pay rather than someone who cannot? Or is it to encourage employers to be more vigilant than they would be pursuant to a duty of care? Or is it just a weapon of distributive justice?’ ¹³⁶ Mohamud’s answer—‘social justice’ ¹³⁷ —is hardly useful, since social justice means all things to all people. We are reminded of Lord Hobhouse’s caution in Lister: ‘[l]egal rules have to have a greater degree of clarity and definition than is provided by simply explaining the reasons for the existence of the rule and the social need for it’. ¹³⁸ Indeed, many learned scholars have pondered on the justifications of the vicarious liability doctrine. ¹³⁹

Without a clear justification for the vicarious liability doctrine, it will run amok. This paper examined justifications for the doctrine that are found in the cases: (a) effective compensation through spreading the cost of the claimant’s injury; (b) deterring the defendant from inadequately supervising potential tortfeasors who do work for them; (c) the defendant should pick up the burden of an organisational or business relationship which he had undertaken for his own benefit; and (d) the defendant, by placing the tortfeasor in the position in its organisation that it did, with that position’s privileges and demands, had created or significantly enhanced the risk of the claimant’s injury.

The paper then argued that the sole justification for the vicarious liability doctrine is the fourth one, which we called the risk-through-placing justification. This justification alone adequately explains the vicarious liability cases. It is also theoretically defensible, even if we ignored consequentialist arguments, as it can be seen as another instantiation of the law linking no-fault liability with risk. The risk-through-placing justification states: vicarious liability exists to ensure that the defendant bears the cost of the tortfeasor’s wrong, and it should bear that cost because by placing the tortfeasor in the position in its organisation that it did, with that position’s privileges and demands, it had created or significantly enhanced the risk of the claimant’s injury. This justification focusses on the attendant risks of the position that the defendant ‘put out there’, and pays no attention to the characteristics of the tortfeasor who is placed in that position.

Finally, this paper showed that there are three implications flowing from embracing the risk-through-placing justification. First, the CCWS Factors should be clarified so that only tortfeasors who are part of the defendant’s organisation, and not independent contractors engaged by the defendant, may trigger vicarious liability for the defendant. Second, we should reject agency reasoning in vicarious liability. Third, courts should stop using the words ‘enterprise risk’.

Many issues surrounding the vicarious liability doctrine remain unresolved; in particular: (i) whether public perception of the defendant–tortfeasor relationship should matter in determining whether the tortfeasor was part of the defendant’s organisation ¹⁴⁰ ; and (ii) whether a parent company can be vicariously liable for the torts of its subsidiary ¹⁴¹ or a tortfeasor that was its subsidiary’s employee. Perhaps the risk-through-placing justification can provide a more stable foundation on which these issues may be discussed and resolved.