I. Introduction
In his forthright and, at times, rather eccentric critique written in 1916, Thomas Baty likened vicarious liability to the Upas tree.Footnote 1 The Upas tree (antiaris toxicaria), which is a traditional source of poison for arrows and blow darts, has not surprisingly captured the literary imagination.Footnote 2 With due allowance for hyperbole there is some truth in his characterisation. Writing to Oliver Wendall Holmes, Sir Frederick Pollock would describe Baty's book as “clever paradoxical” and go onto to say that its author was “of the school who think that the law can be reduced to rigorous chess problems”.Footnote 3 In a review of the work in the Law Quarterly Review, Pollock made many of the same points at greater length.Footnote 4 Pollock himself was decidedly more pragmatic. He accused Baty of overlooking the “factor of insurance in the practical aspect of these questions”.Footnote 5 Baty still made a valid point in his general criticism of vicarious liability. Vicarious liability was, and it remains, curiously unsatisfactory. Even within the law of tort, which has proved volatile in several respects, few legal doctrines can be as difficult to pin down as vicarious liability. In Mohamud v WM Morrison Supermarkets plc.,Footnote 6 Lord Dyson recently observed that “To search for certainty and precision in vicarious liability is to undertake a quest for a chimaera”.Footnote 7 He was unconcerned because he thought that “impression is inevitable” when it came to vicarious liability.Footnote 8
In recent decades, the exposure of systemic child abuse has prompted a re-examination of this area of law. The authorities, which are atypical, raise difficult issues about the degree to which an employer ought to be liable for the criminal activity of an employee.Footnote 9 In many of these cases, as in others, there is an allied problem – the extent to which there should be liability for those who are akin to an employee such as an independent contractor.Footnote 10 The history of the subject has received no more than cursory attention in debates about the scope and value of vicarious liability. The omission is important. In Prince Alfred College v ADC, the High Court of Australia has recently observed that a “fully satisfactory rationale for the imposition of vicarious liability” has been “slow to appear in the case law”.Footnote 11 The reasons, in part at least, lay in the past.
The absence of a single clear explanation for vicarious liability might not matter very much were it not for the fact that vicarious liability is frequently central to the question of who pays compensation when a tort is committed. In most instances an insurer indemnifies an employer or corporation.Footnote 12 An employer (or their insurance company) might then, in theory, be able to recover damages on an indemnity against an employee.Footnote 13 In practice, attempts to enforce a master's indemnity against the employee have caused disquiet.Footnote 14 Most modern lawyers, if they think about a rationale at all, would explain vicarious liability using the servant's tort theory under which an employee's duty of care is imputed to the employer and by which the employer becomes liable because of a breach of duty by an employee.Footnote 15 Under the alternative, master's tort theory, an employee's acts rather than their legal duty are attributed to the employer. On this view the employer commits a tort by a breach of their own tortious duty. For the most part both theories produce the same outcome although there are some occasions where the distinction is critical.Footnote 16 For example, under the servant's tort theory, the employer cannot be liable when the employee has a defence to the tort. The master's tort theory has no such difficulties because the employer is liable for the acts of the employee and not their tort.
The reasons behind the failure to settle on a single, clear rationale for vicarious liability are complex. It is sometimes said that the old authorities support the master's tort theory.Footnote 17 They can certainly be read that way, but restraint is required in trying to find concrete answers to modern problems in the very old authorities. A slightly more promising line of inquiry might be to look to the nineteenth century. This was a time, after all, where the foundations of modern tort law were beginning to be established.Footnote 18 Fault became a central component of tortious liability at this time.Footnote 19 At this point vicarious liability, which imposes liability on an employer without fault, clearly required an explanation. Despite these factors, no very coherent explanation was forthcoming. A growing body of legal literature did little to add much clarity. In the period after the Second World War various statutory reforms of tort law presented a new opportunity to consider the basis of vicarious liability. Unfortunately, these decisions added further confusion particularly once non-delegable duties were thrown into the mix. More recent scholarship has tried to locate vicarious liability within broader notions of attribution based on principles of agencyFootnote 20 or enterprise liability.Footnote 21 Many of these ideas, to varying degrees, have some historical pedigree but the modern law is undoubtedly a muddle and the history of the doctrine is part of the problem.
II. Vicarious Liability: A Short History
In Majrowski v Guy's and St. Thomas’ NHS Trust,Footnote 22 Lord Nicholls observed: “In times past this ‘employer's tort’ analysis of vicarious liability had respectable support in England. But since then your Lordships’ House has firmly discarded this basis in favour of the ‘employee's tort’ approach.” Trying to understand the nature of vicarious liability in the Middle Ages is an exercise in piecing together disparate fragments.Footnote 23 There are dangers in judges reading modern interpretations into those authorities.Footnote 24 There are several reasons why looking for a single coherent explanation for vicarious liability before the nineteenth century is difficult. Jurors sat at the heart of the system of civil litigation and were allowed to determine questions of liability under the cloak of proof.Footnote 25 Legal doctrine was largely hidden behind jury verdicts of guilty or not guilty.Footnote 26 The decline of the jury began in the mid-eighteenth century but the boundary between fact and law, where the former was a matter for a jury and the latter a matter for a judge, only really hardened in the nineteenth century when judges in civil cases almost entirely asserted their dominance over juries.Footnote 27 A second reason for caution when looking at the old authorities relates to the way that the law was structured. Whereas the modern law is founded on an action in negligence the mediaeval law was built around the forms of action.Footnote 28 The way in which the liability was framed was relevant to the way that older action of trespass interacted with the newer action of trespass on the case.Footnote 29
Early examples, in which a master was liable for a servant, occur in the context of actions of trespass on the case based on customs of the realmFootnote 30 one of most important of which concerned the escape of a domestic fire.Footnote 31 In the course of argument in Beaulieu v Finglam in 1401 it was said that “It would be against all reason to put blame or fault on a man where there is none in him; for his servant's negligence cannot be said to be his own doing”. Markham J. was unsympathetic:
A man is bound to answer for his servant's act, as for his lodger's act, in such a case. For if my servant or lodger puts a candle on the wall and the candle falls into the straw and burns the whole house, and also my neighbour's house, in this case I shall answer to my neighbour for the damage which he has suffered.Footnote 32
The same principles soon began to apply in the absence of a custom of the realm. In Caunt's Case,Footnote 33 an action on the case for deceit was brought against both a master and his servant for warranting a butt of rumney wine sound, when it was “unwholesome and unsuitable”. It emerged in the pleading that a servant had sold the wine. Martin J. explained that:
But if your servant, with your collusion and by your command, sells some unwholesome wine, the buyer shall have an action against you; for it is your own sale. If the case is that you did not command your servant to sell the wine to this plaintiff, then you may say that you did not sell it to the plaintiff.
The idea here was that an employer should be liable for an act that they had commanded. The precise scope of a command could cause disagreementFootnote 34 but the general idea reflects the view that a servant stood in the shoes of a master. This idea is rather closer to the modern law of agency in appearance than it is to vicarious liability.
By the late seventeenth century principle of authority began to replace that of command. Authority is a more malleable concept. It is easier to say that a servant acted with a master's authority (which encompasses a greater range of conduct), than it is to say that they acted under his command. It is possible to authorise something without commanding it. As Holt C.J. explained in Tuberville v Stamp Footnote 35 when a fire spread:
[I]f the defendant's servant kindled the fire in the way of husbandry, and proper for his employment, though he had no express command of his master, yet his master shall be liable to an action for damage done to another by the fire; for it shall be intended that the servant had authority from his master, it being for his master's benefit.Footnote 36
Command and authority were still sometimes used interchangeablyFootnote 37 which is perhaps not surprising as they could easily cover some of the same conduct. Authority was closer to vicarious liability in a modern sense. It is more difficult to conceptualise a servant acting within the master's authority as standing in the master's shoes in the same way as one commanded to do a thing. A series of decisions Holt C.J. made clear that, in so far as there was a rationale for the master's liability, it remained unchanged – the master was responsible for the act of his servant.Footnote 38 For a century or more afterwards, judges stressed that a master was liable for the acts of his servant acting with his authority.Footnote 39
How liability was imposed in practice was no doubted determined by how juries of the time understood the employment relationship. None of this amounts to a very carefully worked out justification for vicarious liability, although there is a strong hint in some judgments of Holt C.J. that he saw it as desirable that an employer was liable for his employee if the alternative was that the injured party went uncompensated.Footnote 40 Before the nineteenth century, the employment relationship was highly regulated,Footnote 41 but at the same time servants were often seen as members of the family unit.Footnote 42 Given the very different nature of the employment relationship, some care needs to be taken in using the old authorities to support a particular rationale for the modern law.
III. The Nineteenth Century: A Failure to Find a Rationale
By the nineteenth century, it began to be said that an employer's liability turned on whether the employee was acting in the course of employment. Despite the new terminology, which is still used to this day, the basic justification was still the same: an employer was liable because the employee's act was imputed to them rather than the employee's duty of care. A principle that was more definitively recognisable as the master's tort theory came to be emphasised. For example, Lord Chelmsford L.C. in a well-cited passage on a Scottish Appeal, Bartonshill Coal Co. v McGuire Footnote 43 explained:
It has long been the established law of this country that a master is liable to third persons for any injury or damage done through the negligence or unskilfulness of a servant acting in his master's employ. The reason of this is, that every act which is done by a servant in the course of his duty is regarded as done by his master's orders, and consequently is the same as if it were the master's own act, according to the maxim, Qui facit per alium facit per se. Footnote 44
Another statement along these lines can be found in Hutchinson v York, Newcastle and Berwick Railway,Footnote 45 where Baron Alderson held that “The principle upon which a master is in general liable to answer for accidents resulting from the negligence or unskilfulness of his servant, is, that the act of his servant is in truth his own act”.Footnote 46 He continued that “whatever the servant does in order to give effect to his master's will may be treated by others as the act of the master: Qui facit per alium, facit per se”.Footnote 47
As well as judges, many legal writers of the period came to a similar conclusion. Charles Smith expressed himself with some brevity: “A master is ordinarily liable to answer in a civil suit for the tortious or wrongful acts of his servant, if those acts are done in the course of his employment in his master's service: the maxims applicable to such cases being, Respondeat superior, and … Qui facit per alium facit per se.”Footnote 48
In books as varied as law student crammer,Footnote 49 Underhill's treatise on tort law,Footnote 50 to Holland's avowedly theoretical, The Elements of Jurisprudence,Footnote 51 the maxim qui facit per alium facit per se was used to explain the reason a master was liable for his servant. To modern eyes stating “let the master answer” or “he who acts through another, acts himself” are poor explanations. At the time however the use of legal maxims was a popular techniqueFootnote 52 and maxims were still regarded as important enough to merit an entire treatise.Footnote 53 Even those writers who avoided maxims and instead drew parallels with agency were still making the same point that an employer was liable for the acts of his employee.Footnote 54
Despite the words of Lord Chelmsford L.C. and Baron Alderson there was still room for ambiguity.Footnote 55 This was clear from a further mid-nineteenth description by Willes J. in Barwick v English Joint Stock Bank Footnote 56: “The master is answerable for every such wrong of the servant or agent as is committed in the course of the service and for the master's benefit, though no express command or privity of the master be proved.”Footnote 57
In this passage Willes J. emphasised the servant's wrong as opposed to their act. He also seems to hint at something like the modern idea of enterprise liability.Footnote 58 Willes J. did not develop a concrete principle of enterprise liability on this occasion. But there is more than a suggestion of similar ideas in older authorities that, when activity benefits an enterprise, the enterprise should carry the risk of any damage caused by employees.
In Bush v Steinman,Footnote 59 Rooke J. had said: “He who has work going on for his benefit, and on his own premises, must be civilly answerable for the acts of those whom he employs.”Footnote 60 On that occasion the owner of a house was liable when a servant of a sub-contractor employed by a builder left lime in the road, causing the plaintiff's carriage to overturn. The relationship between the defendant and the person causing the damage was more tenuous than the usual one between a master and servant. Whilst agreeing with his colleagues, that the owner was liable, Eyre C.J. was clearly uneasy about the outcome. As he pointed out, the owner “appeared to be so far removed from the immediate author of the nuisance, and so far removed even from the person connected with the immediate author”.Footnote 61 For a time the courts were at least content to apply something like enterprise liability to occupiers of land rendering them liable for independent contractors but this line of authority was finally rejected in Reedie v London and North Western Railway Co. Footnote 62 A broader principle of enterprise liability did not catch on with judges at this time, although it would become attractive to lawyers at various points subsequently.Footnote 63 The episode shows a degree of fluidity when it comes to finding a rationale for vicarious liability which was mirrored in the broader debate of the period.
IV. Searching for a Rationale for Vicarious Liability
By the final decades of the nineteenth century, it was obvious to many legal writers that the orthodox explanations were inadequate. Sir Frederick Pollock argued that respondeat superior, “is a dogmatic statement, not an explanation”; qui facit per alium facit per se was also said to be deficient because it does not cover acts of servants that are unauthorised but still within the course of employment.Footnote 64 He points out if the master was to be blamed for showing a lack of care in employing a servant who commits a tort then it should follow that if he had shown all due care he ought to be absolved of liability, which was clearly not the case. Pollock's favoured explanation was closer, but not identical to, the modern idea of enterprise liability: “I am answerable for the wrongs of my servant or agent, not because he is authorised by me or personally represents me, but because he is about my affairs, and I am bound to see that my affairs are conducted with due regard to the safety of others.”Footnote 65
Pollock took up the subject of vicarious liability again in his more philosophical work, Essays in Jurisprudence and Ethics.Footnote 66 Having once more criticised the standard justifications he pointed out that: “In the case of injury suffered through a servant's negligence, the servant, generally speaking, cannot pay, and the master can; and the feeling that compensation ought to be had somewhere jumps at the master's liability.”Footnote 67
This looks less like an explanation than a statement of the obvious. Pollock recognised that his own reasoning was not entirely satisfactory. Taken to its logical conclusion it would mean that a master could be liable for losses caused by behaviour that he has expressly forbidden provided the servant was in breach of his duty and acting in the course of his employment.Footnote 68 The idea that the employer was better able to pay has some intuitive appeal and Pollock was still not alone in mentioning it.Footnote 69 Pollock's own position was a little more nuanced. He combined something akin to enterprise liability with the idea that an employer had a moral duty for the acts of their employees.Footnote 70
One of the reasons that vicarious liability had come to be seen as problematic was identified by Pollock in his note in the Law Quarterly Review which pointed out that a doctrine which had grown up during very different social and economic conditions was “incongruous” in an age of “industrial individualism”.Footnote 71 Pollock was not alone in seeking a novel explanation for vicarious liability. John Henry Wigmore, the American evidence scholar, looked for answers in the history of the doctrine.Footnote 72 Providing a critique was rather easier than providing an alternative explanation. Baty produced the most radical proposal of the era. Beyond the erroneous claim that vicarious liability was an invention of Holt C.J.,Footnote 73 his argument is rather difficult to unravel. Baty seems to object to the idea that an employer could be liable for an employee simply because they act in the course of their employment. Rather he suggests that there should be no liability on the part of the employer unless there is an “invitation of confidence” between the plaintiff and the employer.Footnote 74 An equally novel approach taken by Harold Laski made no mention of a master or servant tort theory and, whilst he saw value in vicarious liability in an article in the Yale Law Journal in 1916,Footnote 75 he argued that a new “social interpretation” was needed. He found this in the idea that “The promotion of social solidarity is an end it is peculiarly incumbent upon the law to promote, since its own strength and even life, depends upon the growth of that sentiment”.Footnote 76 Laski's support of this version of vicarious liability was a reflection of his criticism of laissez-faire ideology, and he would comment “it becomes increasingly evident that society cannot be governed on the principles of commercial nihilism”.Footnote 77
Laski was a political theorist rather than a lawyer. His friend, the American jurist Oliver Wendell Holmes, added a note of caution in a letter to Laski when he said that “there should be a limit to be found in values to the public of the life, limb, or what not damaged”.Footnote 78 Laski's theory was likely to be too idealistic for most lawyers, but it does make the point that the existing and long established rationale for vicarious liability were no longer seen as adequate. Other writers looked for a social explanation. The future US Supreme Court justice William O. Douglas, agreed with Laski that the issue was one of “economic and social factors” but rather than promoting social solidarity he thought that vicarious liability was explained by an allocation of risk.Footnote 79
The absence of a single convincing explanation was one indication that vicarious liability rested on insecure and shifting doctrinal foundations. Baty went to some trouble to identify all the different reasons put forward for vicarious liability. He came up with nine different justification from which he deduced that “A doctrine that is accounted for on nine different grounds may reasonably be suspected as resting on no very firm basis of policy”.Footnote 80 Like Pollock in his Essays, Baty would conclude that the real reason for an employer's liability rested on nothing stronger than that “damages are taken from a deep pocket”.Footnote 81 Other writers were less worried by the absence of a justification. Thomas Beven would describe respondeat superior as a formula and not a reasonFootnote 82 down to the final edition of his treatise in 1908.Footnote 83 Whilst he did not take the maxims too seriously, he still chose to adopt a master's tort approach.
Although the old maxims began to be abandoned, lawyers were not always very clear to distinguish liability for a tort, and liability for an act. Willes J. and Pollock both talk about a servant's “wrong”. Oliver Wendell Holmes was curiously imprecise. There are two instances in his book The Common Law,Footnote 84 in which he mentions the subject – on one occasion he suggests that a master is liable for the servant's wrong, and on the other the servant's act.Footnote 85 In the leading treatise, by Clerk and Lindsell, the authors state that “the employer is liable for all torts committed by the party employed”.Footnote 86 Yet, a few pages earlier the older familiar formula appears with an emphasis on the act rather than the tort of the employee.Footnote 87 At best this seems to suggest that the authors put little weight on the distinction between wrongs and acts, and by extension had no set notion of a master or servant's tort theory in mind. The same approach using these two formulas was evident in later editions of the same work.Footnote 88
In the opening paragraph of the section of his treatise on master and servant, John Salmond, wrote:
A master is liable for any tort committed by his servant while acting in the course of employment …. Its rational justification is to be found in the presumption that the negligence and other torts of a servant in the execution of his master's business are either actually authorised by the master, or, at least, are the result of some want of care on the master's part in the choice of competent servants or in the superintendence and control of their work.Footnote 89
Salmond gave no explanation for abandoning the traditional view in favour of the idea that an employer was liable for the torts of an employee rather than their acts. Salmond's wording was unambiguously supportive of a servant's tort analysis. Yet, somewhat perplexingly in his book on legal theory, Salmond wrote that vicarious liability meant that “one man is made answerable for the acts of another”.Footnote 90 This phrase is redolent of the master's tort theory, The same formula was used down to the last edition of Jurisprudence, for which Salmond was responsible in 1924.Footnote 91 All of this invites the question whether Salmond appreciated the different meanings to which his language could give rise, or whether he was actually committed to a master or servant tort analysis and was just imprecise with the language that he used. It seems likely that, like other writers of the time, he was doubtful about the existence of any coherent theory and thought that there was little to choose between a master's or servant's tort analysis.
Writing 30 years after Salmond, Percy Winfield would unequivocally favour a servant's tort theory: “A is liable for the tort of B committed against C, though A is no party to the tort. B himself is of course usually liable.”Footnote 92 At the same time he would admit that “a scientific reason for the rule is hard to find”.Footnote 93 He continued: “It seems to be based on a mixture of ideas – that the master can usually pay while the servant cannot; that a master must conduct his business with due regard to the safety of others; that the master by employing the servant has ‘set the whole thing in motion’.”Footnote 94
The confusing debate about the true basis of vicarious liability in the legal literature of the time unsurprisingly made little impression on the case law. There is still no indication that judges placed any significance on a master's or servant's tort theory any more than Clerk and Lindsell did.Footnote 95
The absence of judicial comment has another explanation. For the most part it was not something that it was necessary for judges to address given either the master's or servant's tort theory produced the same outcome. It is difficult to be definitive but in those few situations where it produced a different outcome, judges seemed content to adopt a master's tort analysis without explicitly admitting that they were doing so. Dyer v Munday Footnote 96 illustrates the point. The plaintiff's lodger had failed to pay an instalment on a hire-purchase agreement and, in the process of removing the lodger's furniture, the defendant's employee assaulted the plaintiff. The employee was not liable because of legislation, which prevented someone who had suffered a criminal sanction from facing a civil action in the same matter.Footnote 97 The employee had a defence to a claim in tort, yet the Court of Appeal saw no difficulty in finding the defendant's employer liable. All three members of the Court of Appeal stressed that an employer could be liable for a criminal act of an employee; provided it was in the course of employment. Lord Esher M.R. and Rigby L.J. explained that the statute, which protected the employee, did not apply to the employer as a matter of statutory interpretation.Footnote 98
Outside specific legislation, there were other situations in which the employee's defence failed to prevent an employer's tortious liability. As it stood at the time, the law meant that spouses could not be liable to each other in tort.Footnote 99 This potentially causes a problem when one spouse (the employee) injured another and the injured party brings a claim against an employer or principle. In Smith v Moss,Footnote 100 an injured wife successfully brought an action against her mother in law, who was the master for the purposes of the action, when she was injured by her husband's driving. Charles J. held that “I cannot conceive that, if a husband, while acting as agent for somebody else, commits a tort, which results in injury to the wife, the wife is deprived of her right to recover against the principal who is employing the husband as agent”.Footnote 101 There was no attempt to justify this outcome by reference to master's or servant's tort theory. Rather Charles J. reasoned that the negligent driver was both a husband and an agent, and that his status as the former did not preclude a claim against a principal.
By the late 1930s, when Winfield was writing, legal authors had begun to reject traditional explanations for vicarious liability. Judicial silence on the subject does not necessarily correlate to enthusiasm for orthodoxy as opposed to acquiescence with it. During the second half of the twentieth century the master's tort theory would be eclipsed by the servant's tort theory of liability. In the immediate post-war period however, despite what some legal writers were advocating, the master's tort theory still held sway.
V. The Master's Tort Theory Reasserted
In the period after the Second World War, judges became more explicit in adopting a master's tort theory in those cases where the employee could call on a defence. Broom v Morgan Footnote 102 was the first detailed examination of the impact of marriage on tort liability between spouses. The plaintiff and her husband were employed by the defendant to run his public-house. The wife was injured when she fell through a trap door which her husband had negligently left unprotected. At first instance, Lord Goddard, restated the master's tort theory: “The master is just as much liable as though he commits the tort himself because the servant's act is his act.”Footnote 103 Yet, as Lord Goddard conceded, Smith v Moss aside, there was very little authority on the precise question of vicarious liability and spouses. It is perhaps telling that he was forced to rely on an American decision of Cardozo C.J. in finding the employer liable.Footnote 104 The defendant appealed. The Court of Appeal dismissed the appeal and found the employer liable.Footnote 105 Although the judgments vary in detail, all three judges adopted a version of the traditional analysis. Singleton L.J. quoted from Lord Chelmsford in Bartonshill Coal Co. v McGuire, and again repeated the standard maxim, qui facit per alium, facit per se.Footnote 106 Hodson L.J. supported a master's tort analysis and referred to Lord Chelmsford.Footnote 107 Denning L.J. gave a slightly different account of the master's tort theory. His version aligned more closely with the earlier, albeit fragmented, support for enterprise liability: “It is the sound moral reason that the servant is doing the master's business, and it is the duty of the master to see that his business is properly and carefully done.”Footnote 108 The Court of Appeal clearly felt uncomfortable about using the status of the victim as a wife in barring a claim against the employer. As Singleton L.J. explained: “there is no reason, either in law or in common sense, why they (the employer) should be given an immunity which springs in the case of husband and wife from the fiction that they are one, and the desire that litigation between husband and wife should not be encouraged.”Footnote 109
A statute in 1962 meant spouses could be liable to each other in tort and therefore this situation ceased to cause difficulties.Footnote 110 Actions against the employer of a spouse were not the only occasion when judges in the immediate post-war period would reiterate the master's tort analysis. In Twine v Bean's Express Ltd.,Footnote 111 an employee of the defendant carried an unauthorised passenger in the defendant's van who was killed as a result of the employee's negligent driving. The employer was held to not be liable. The plaintiff was a trespasser, and the employee was not authorised to carry passengers. The decision is unsatisfactory on several levels,Footnote 112 but it does contain an unequivocal endorsement of the master's tort theory. Uthwatt J. stated: “The law attributes to the employer the acts of a servant done in the course of his employment and fastens upon him responsibility for those acts.”Footnote 113 The Court of Appeal did not consider the correctness of this analysis, but dismissed the appeal on the grounds that the driver was not acting in the course of his employment.Footnote 114
A third group of decisions concern the doctrine of “common employment”. The origins of the rule are obscure,Footnote 115 but it was not contested by the mid-nineteenth century.Footnote 116 It meant that an employee was prevented from bringing a claim against an employer for an injury caused to him by another employee. The common employment doctrine marked a significant inroad into the principles of vicarious liability. Whilst the implications were serious for those involved, after 1880 they were not quite as far reaching due to the rise of statutory liability for workplace injury.Footnote 117 There were a series of significant exceptions to the prohibition.Footnote 118 By the late 1930s, the whole doctrine of common employment was increasingly seen as unsatisfactory. Lord Wright, in Wilson & Clyde Coal Company Ltd. v English, went as far as to suggest that it was “well-established, but illogical”.Footnote 119 Academic writers were also critical of the principle. William Robson wrote that “The time has clearly come when the doctrine of common employment should be abolished”.Footnote 120 One difficulty recognised by Lord Macmillan was that common employment could potentially come into conflict with vicarious liability: “If a servant is injured by the negligence of a fellow-servant acting within the scope of their common employment, the former doctrine would impose liability on the master, while the latter doctrine would exculpate him.”Footnote 121
On the facts, the common employment doctrine was held not to apply because the cause of the injury was due to flaws in the safe system of working by which the mine operated, and the agent was performing the duties of the owner. This was not a situation where one employee injured another. This case fell within the category of a non-delegable duty. A year later in Radcliffe v Ribble Motor Services Ltd.,Footnote 122 the House of Lords would again consider the common employment doctrine and decide that it did not apply. One of the defendant's coach drivers negligently knocked down and killed another of the defendant's drivers. The accident occurred on a public street. The plaintiff widow succeeded in a claim. The common employment doctrine was once again criticised. Lord Macmillan said that “its original ratio has long ceased to be regarded as tenable”.Footnote 123 Although he regretted the existence of the common employment rule, Lord Wright emphasised that as something settled by the House of Lords it could only be changed by legislation.Footnote 124 On the facts, the doctrine was held not to apply. The employment was not common because the employees were said not to be engaged in “common work”. The fact that they were working for the same employer was not enough. There must be a “common object” to their activities.
The entire rationale for these decisions was to avoid applying the common employment doctrine and, by extension, to refuse to limit vicarious liability. In 1948, the rule was finally abolished.Footnote 125 Three years earlier the defence of contributory negligence was also liberalised.Footnote 126 One consequence of these changes was to focus judicial attention on the relationship between employees and the extent to which an employer might be vicariously liable free from the shackles of common employment and the old style contributory negligence defence. In Jones v Staveley Iron and Chemical Co. Ltd.,Footnote 127 one worker was injured by another. One question to be addressed was the standard of care to be expected by a skilled worker and the way that translated to vicarious liability for an employer. In the Court of Appeal, Denning L.J. gave one of the last major endorsements of the master's tort theory: “He acts by his servant; and his servant's acts are for this purpose, to be considered as his acts. Qui facit per alium facit per se …. If he takes the benefit of a machine like this he must accept the burden of seeing that it is properly handled.”Footnote 128
He goes on to cite Broom v Morgan to the effect that an employer is liable even if the employee is immune from liability. Denning L.J. combined the traditional maxim with something akin to enterprise liability. Romer L.J. said nothing about immunity but can equally be said to adopt a master's tort analysis:
[I]t is well settled that a master is liable for the acts of his servants, if done in the course of their employment, on the principle qui facit per alium facit per se. If an employer employs a crane driver to operate a crane, it is the employer himself who, in the eye of the law, is operating it, though he is doing so through the person of the employee.Footnote 129
Hodson L.J. seemed more inclined towards a servant's tort analysis: “In my opinion, fault must be attributed to the crane driver, and the question then arises whether this amounts to negligence.”Footnote 130
In dismissing an appeal, the House of Lords took the opportunity to support a servant's tort analysis.Footnote 131 Lord Reid said of Denning L.J.’s use of the maxims that “it is rarely profitable and often misleading to use Latin maxims in that way” and that “I do not think that they add anything”.Footnote 132 He stressed that in order for an employer to be liable the servant must be negligent.Footnote 133 Lord Reid suggests that Romer and Hodson L.JJ. both “appear to … base their judgment on the crane driver having been negligent”.Footnote 134 This is a servant's tort analysis. Lord Morton was of the same view. In fact, it is far from clear that Romer L.J. did subscribe to this analysis. On the question of immunity, he held that Denning L.J.’s words are “in my view, incorrect as applied to a case where the liability of the employer is not personal but vicarious. In such a case if the servant is ‘immune’, so is the employer”.Footnote 135 Lord Tucker's speech is more ambiguous. He states that “The present is a simple straightforward case of a master's responsibility for the acts of his servant done in the course of her employment”.Footnote 136 Yet he also seems to base his conclusions on the negligence of the crane driver.Footnote 137
VI. The Rise of a Servant's Tort Analysis
As Glanville Williams pointed out, Jones v Staveley Iron and Chemical Co. Ltd. did not “settle the question beyond all doubt”.Footnote 138 The remarks were obiter and aspects of the reasoning are not entirely satisfactory. It was still the first major judicial endorsement of a servant's tort analysis. The same approach was also reflected in statute. The Crown Proceedings Act 1947s. 2(1)(a) declared that the Crown was liable in “respect of torts committed by its servants or agents”. It is difficult to give a definitive reason for the decline of the master's tort theory. It is unlikely that one factor alone was at play. The master's tort theory had been around for a long time and was not killed off by a single decision. The backstory is more complicated. Winfield, in his influential treatise on tort, had adopted a servant's tort analysis in the late 1930s. He would still concede that a “scientific basis” of vicarious liability is hard to find. The sixth edition of Winfield's treatise, edited by T. Ellis Lewis,Footnote 139 contained a new section addressing the nature of vicarious liability which Denning L.J. singled out in a book review in which he commented that “The true basis of the master's liability has been under much discussion lately”.Footnote 140
Ellis Lewis began by saying that the true basis of vicarious liability “still waits final determination” before explaining that “the doctrine is based on public policy or sound necessity rather than deduction from legal principle”.Footnote 141 The demise of the common employment doctrine was one reason why vicarious liability had come to the fore, but it was also part of a wider movement in the 1950s in which the courts were making a conscious effort to modernise vicarious liability. The older terminology of master and servant fell out of use,Footnote 142 maxims were regarded with suspicion,Footnote 143 and the nineteenth-century law was increasingly seen as ill-suited to more modern employment relationships.Footnote 144
Another important factor which helped to ensure that judges began to systematically address vicarious liability was highlighted by academic commentators at the time,Footnote 145 but has received little notice in the intervening decades. In a series of judgments, Denning L.J. sought to elide vicarious liability in which one person is liable for another with the principle of a non-delegable duty. He did so by using the old maxim qui facit per alium facit per se. Using this analysis, the employer was liable not because the employee's tort or act was imputed to them but because they had themselves committed a tort. A passage from Broom v Morgan Footnote 146 illustrates the point:
My conclusion on this part of the case is, therefore, that the master's liability for the negligence of his servant is not a vicarious liability but a liability of the master himself owing to his failure to have seen that his work was properly and carefully done. If the servant is immune from an action at the suit of the injured party owing to some positive rule of law, nevertheless the master is not thereby absolved. The master's liability is his own liability and remains on him notwithstanding the immunity of the servant.Footnote 147
In Jones v Staveley Iron and Chemical Co. Ltd.,Footnote 148 Denning L.J. used the same analysis:
The employer is made liable, not so much for the crane driver's fault, but rather for his own fault committed through her. He puts her in charge of a great machine which can cause much damage if it is not properly handled. He must see that reasonable care is used in the handling of it so that it does not cause damage. No matter whom he employs to handle it, he must ensure that a proper standard of care is obtained.Footnote 149
This approach was not entirely novel. It is a little like the way that the House of Lords in Wilson & Clyde Coal Company Ltd. v English Footnote 150 sidestepped the common employment doctrine. On that occasion, the House of Lords stressed that the employer was liable for failing to provide a safe system of work. Denning L.J. seems to be pointing towards the idea of a non-delegable duty. It is no coincidence that Denning L.J. argued elsewhere that hospitals were liable for injuries suffered by patients using a non-delegable duty.Footnote 151 Although Denning L.J. was careful to distinguish this form of liability from vicarious liability,Footnote 152 his approach risks collapsing the two. In cases of a non-delegable duty and vicarious liability the employer is liable without fault on their part. Although similar, the two are not identical.Footnote 153 A non-delegable duty is a form of primary liability. The employer is the one with the duty. The role that non-delegable duties ought to play in the modern law is contested.Footnote 154 Denning L.J.’s intervention added to the conceptual confusion. The way that he tried to develop a form of primary liability expressed as a non-delegable duty was broader than anything that had gone before.
The House of Lords in Jones v Staveley Iron and Chemical Co. Ltd. was clear in resisting the attempt by Denning L.J. to reshape employer's liability for the acts of their employees. It did so by insisting that the case was properly one of vicarious rather than primary liability. Lord Morton explained: “Cases such as this, where an employer's liability is vicarious, are wholly distinct from cases where an employer is under a personal liability to carry out a duty imposed upon him as an employer by common law or statute.”Footnote 155 Both the master's and servant's tort theories are examples of vicarious liability, but it is easier to draw a sharp line between vicarious liability and primary liability when a servant's tort analysis is used. Once an employee is required to commit a tort, there was no possibility of conflating his tort with that of the employer.
The reason that the House of Lords was resistant to extending the scope of primary liability, expressed as a form of non-delegable duty, was never fully articulated. One likely fear was that it would involve extending the scope of an employer's liability. The fear of indeterminate liability for employers was something that Lord Denning himself would comment on in the context of the liability for hospitals.Footnote 156 The same concern has continued to be reflected in modern attempts to explain the imposition of non-delegable duties using the idea of “assumption of responsibility”.Footnote 157
VI. Conclusion
In the late 1960s, Patrick Atiyah wrote that “On the whole it seems doubtful whether much is to be gained by examination of the ‘true’ basis of vicarious liability”.Footnote 158 On one level, this statement is true. The master's or servant's tort theory usually gives the same outcome as far as liability is concerned but the question of the accepted analysis has a significance beyond itself. It goes to the nature of liability in tort more generally. Once tortious liability was characterised by fault then vicarious liability was increasingly difficult to justify. Despite the struggle to find a coherent justification, vicarious liability has proved to be remarkably resilient.
A master's tort analysis can be constructed from the older cases, although the way jurors reached decisions about liability makes it difficult to know how these notions were applied in practice. The emergence of a much larger body of legal literature in the nineteenth century did little to introduce much clarity. In part, this was because there was no well-developed theoretical structure for tort law – in contrast to Will theory which, however imperfectly, provided a framework for contract law.Footnote 159 It was extremely difficult to find a theory that fitted all of the authorities on vicarious liability. Ever since pragmatic or policy-based justifications for vicarious liability have proved to be attractive. In Imperial Chemical Industries Ltd. v Shatwell,Footnote 160 Lord Pearce made the point that
The doctrine of vicarious liability has not grown from, any very clear, logical or legal principle but from social convenience and rough justice. The master having (presumably for his own benefit) employed the servant, and being (presumably) better able to make good any damage which may occasionally result from the arrangement, is answerable to the world at large for all the torts committed by his servant within the scope of it.Footnote 161
As late as the mid-1970s, in Launchbury v Morgans,Footnote 162 Viscount Dilhorne and Lord Pearson were still repeating the maxim, qui facit per alium, facit per se. Footnote 163 Unlike their nineteenth-century counterparts, the fact that Lord Pearson talked about the negligence of the driver does not suggest that he was actually adopting the master's tort theory.Footnote 164 The House of Lords on this occasion was anxious to set down the limits on vicarious liability outside an employment relationship, and more specifically to close off an attempt by Lord Denning M.R. in the Court of Appeal to impose liability on the owner for accidents caused by drivers of the family car. By this stage the servant's tort theory was no longer questioned.
It is easy enough to see how various reforms of tort law thrust vicarious liability before the courts. It is more difficult to explain why, in a relatively short period, the courts began to adopt a servant's tort analysis. One reason may be that the idea was gaining ground in the legal literature. Another, perhaps more plausible, reason may lay in the fact that a servant's tort theory draws a clearer distinction between primary and vicarious liability because it is clear that the servant breaches the duty of care. It may be no coincidence that Lord Denning favoured a master's tort analysis. In part, the servant's tort analysis may have been a judicial reaction to his attempt to develop a primary liability expressed as something akin to a non-delegable duty. His analysis not only ran the risk of blurring the boundary between primary and secondary liability, but of creating open-ended liability. This becomes a problem once non-delegable duties start to be used outside judicial attempts to subvert the common employment doctrine, or other similarly narrow exceptional circumstances and become a general doctrine.Footnote 165
Because so many explanations for vicarious liability are possible, this area of the law has been particularly subject to changes in fashion. Sometimes this is cyclical. Just recently appellant judges have turned to enterprise liability once more.Footnote 166 It has proved easier for lawyers to treat vicarious liability as a distinct body of doctrine, which is difficult to explain. Little attempt has been made to explain where it fits within a broader theory of tort law. Stevens, who advocates a master's tort theory, is a modern exception in this regard.Footnote 167 In fact the master's tort theory hasn't completely disappeared from view.Footnote 168 The choice between a master's or servant's tort theory matters less than it used to because there are fewer instances where a choice makes a difference in practice. Yet, from the point of view of coherence, something closer to a satisfactory explanation is only ever likely to be forthcoming when lawyers can give a proper account of the survival of vicarious liability in a fault-based system of tort law.
