Vicarious Liability in Tort

– a search for policy, principle or justification

New South Wales Supreme Court Judges Conference Leura, 25 August 2017

Vicarious liability is something of an anomaly in the law of torts.  It requires A to pay B because C has injured B, where A was not personally at fault.  Nevertheless, it has been with us for a long time and is not peculiar to the common law.  It is also imposed by the courts of civil law systems.  This commonality of approach may not merely reflect a shared source in Roman law.  It may suggest that the idea of an employer being liable for the tortious acts of employees has seemed reasonable, if not compelling, to judges.  This might explain the course taken by the German courts.  Although the German Civil Code (“the BGB”) based the liability of persons such as employers for third parties in fault, the courts managed to evade these provisions in favour of this form of liability.

Recent attempts by some common law courts to explain vicarious liability point to the difficulty which still attends the question of the rationale for this form of liability.  In addressing the question it may be useful to separate out the various methods of explaining or justifying rules which are employed by the courts: policy, which often has a socio-legal base and states a general objective; legal principle, which has emerged from the case law and may be applied generally to determine liability; tests, which state requirements for liability and are also useful to both justify and limit it; and reasons given for the imposition of liability in a particular case which are in the nature of a justification.

It has not been suggested that the early cases contain any express statement of the policy of the law, although an enthusiasm for the imposition of liability on masters for the acts of their servants may be detected.  It may have its roots in unexpressed social policy.

In attempting to identify a justification for its imposition in a particular case the courts have resorted to fictions, maxims and ingenious rationales.  In Hollis v Vabu Pty Ltd it was said that while some rationales may be persuasive to some degree, “given the diversity of conduct involved, probably none can be accepted, by itself, as completely satisfactory for all cases”. No single, unifying or general principle has therefore emerged.

In the Prince Alfred College case it was conceded that “[t]he identification of a general principle for vicarious liability has … eluded the common law for a long time”.  It might also be thought that the extension, over time, of liability to flagrantly unauthorised acts, and to more complex circumstances of wrongful conduct within employment, makes the development of any general principle even more difficult.

This state of affairs persisted for some centuries.  By the beginning of the 20th century the courts had shown no real inclination to identify the real basis for the liability.  Some academic writers did.  The reasons they gave were largely economic:  recourse to solvent defendants and the efficient distribution of loss.  More recently some common law courts appear to have accepted these objectives as reasons for the rule, but not as legal policy.

It may be as well then to go back to where it seems to have started.

Some historical background —Holt CJ

Sir Frederick Pollock is said to have coined the term “vicarious liability” in his correspondence with Oliver Wendell Holmes in 1888.  A liability of this kind in English law is of much earlier origin and is generally attributed to the judgments of Lord Holt CJ in the late 17th century.  According to Wigmore, apart from some cases of the old strict liability (such as a householder’s liability for the escape of a fire started by a servant in the house), the test applied in cases in the 16th and 17th centuries was one of command or consent.  The master’s command excused the servant.  Wigmore says it was when it was sought to limit this rule by requiring that the master must have commanded the very act complained of, that the form of the rule began to change under Holt CJ.  The reaction was, in effect, to extend the rule.

It has been suggested by more than one scholar that Holt CJ was inspired in the approach he took by Roman law.  Holdsworth thought it likely that he was influenced by a combination of Roman law and the medieval principle that masters are liable for their households, including the actions of their servants.  This was also a time of growth in shipping and commercial enterprise and Holt CJ regularly dealt with the admiralty law and law merchant, each of which contained elements of a liability of this kind.

Holt CJ’s judgments do not provide a clear basis in policy or in principle for this new rule as to a master’s liability.  The reasons he gives vary.  Wigmore said of these many justifications that “very often the judicial mind gave up the troublesome task of accurately expressing a reason, and, quite content with the policy of the rule, took refuge, when it came to naming a reason, in a fiction or other form of words”.  Another academic writer said that “the force and generality of the Chief Justice’s language indicate the tendency he had to impose liability on employers”.

In Sir Robert Wayland’s case, Holt CJ said that “the master at his peril ought to take care what servant he employs”.  The idea of being careless in the choice of a servant is thought to derive from Roman law, where shipowners, innkeepers and stable keepers were vicariously liable for damage or theft by their servants on the assumption that the master must have been in some sense at fault for choosing that servant.

This reason no longer has currency.  It seems somewhat inconsistent with a rule of liability which is not fault-based. It is an assumption, or fiction, since “no amount of care in the choice of one’s servant” could exonerate the master .  Moreover, it is difficult to explain the distinction which the law makes (though perhaps not as clearly now) between employees and independent contractors, both of whom are “selected” by the employer and principal respectively.  Nevertheless, traces of the idea of carelessness in the choice of one’s servant can be seen as late as the early 20th century.  In George Whitechurch Ltd v Cavanagh, Lord Roberston regarded it to be relevant that the “innocent third party” had “no voice” in the selection of employee.

In Boson v Sandford, liability appears to have been imposed by Holt CJ on the basis of some implied indemnity from the master (“the master ‘undertakes’ for the servant’s care”).  Other cases show that he was simply willing simply to identify the servant with the master or to impute the consequences of a servant’s neglect to him.  He said that an action lay against an employer where his servants ran their cart into another cart causing wine to spill; or ran the cart over a boy injuring him; or where a blacksmith’s employee injured a horse in the course of shoeing.  The reason given for all these examples was that “whoever employs another is answerable for him and undertakes for his care to all that should make use of him”.

It was suggested by Dr Thomas Baty, writing in 1916, that many of the cases decided by Holt CJ could have been determined without resort to this form of liability.  In Hern v Nichols, for example, a merchant sold silk through an agent who misrepresented its quality.  Holt CJ held that it is “more reasonable that he that employs and puts a trust and confidence in the deceiver should be a loser than a stranger”.  The point Baty makes is that the purchaser was sold silk inferior to what the seller had represented, and the purchaser could therefore have brought an action on a contractual warranty.  The case was inextricably bound up with contract.  Dr Baty observed that many of the statements made by Holt CJ were pure dicta.  What he found puzzling was how they acquired the force of law between about 1698 to 1725, as they appear to have done.

A shift in approach

Towards the end of the 18th century there was something of a shift in approach.  Bacon, in his Abridgement, said that the master was answerable for the acts of his servant because the law permitted the master to “delegate the power of acting for him to another”.

This period in the development of the liability is said to have been marked by the judgments of Lord Kenyon, although his language was not uniform either.  In Ellis v Turner he said that the maxim “respondeat superior” applied and the defendants were responsible for the acts of their servant “in those things that respect his duty under them”.

Baty clearly disapproved of the employment of maxims such as “respondeat superior“.  He said they might “roll trippingly off the tongue”, but “they are not arguments”.  In Darling Island Stevedoring and Lighterage Co Ltd v Long, Fullagar J said that maxims such as “respondeat superior” do not really explain vicarious liability.  He thought the doctrine was adopted “as a matter of policy which did not really need to be juristically rationalised, but might perhaps be justified (however illogically) as an extension of the notion of agency”.  Lord Reid described the other maxim which was sometimes used, “he who acts through another does the act himself” (qui facit per alium facit per se), as merely a “fictional explanation” of vicarious liability.

In Bugge v Brown, Isaacs J acknowledged that vicarious liability does not depend “merely on the question of authority, express or implied” but rather on the view that it is “more just to make the person who has entrusted his servant with the power of acting in his business responsible”.  Similarly, Dixon J was later to say that although vicarious liability was “commonly regarded as part of the law of agency”, agency principles did not go as far as it did — to hold the principal generally liable for unauthorised acts committed by an agent.

More recent decisions in Australia and the United Kingdom have observed that, although vicarious liability was traditionally regarded as part of the law of agency, terms such as “agent” have often been used as “statements of conclusion that mark the limits to which vicarious liability is extended” rather than the true explanation of why vicarious liability should be imposed.

In any event, agency-based explanations would not now seem useful to explain the contemporary scope of vicarious liability which may now extend to wrongdoing which constitutes a flagrant breach of the conditions of employment or even intentional wrongdoing, which are the current concerns of the courts.

Control

The fact of an employer’s control was also resorted to as an explanation for imposing vicarious liability. On this view the employer was assumed to be able to control the employee’s work, and direct the employee about what to do and how to do it.  Similar notions can be found in German and French law.

Professor Atiyah noted that the notion of control had been treated as some justification for imposing vicarious liability.  He further noted that the fact of control had been used to explain for whom an employer was liable and why.  But a test which places emphasis on control speaks of other times when control may have been a reality.  In Stevens v Brodribb Sawmilling Company Pty Ltd, it was explained that it was based on different social conditions, in which a person engaging another person to perform work could and did exercise closer and more direct supervision than is possible today, whereas such conditions have since largely disappeared with advances in science and technology.  It has given way to a broader assessment of the totality of the relationship between the parties.  Nevertheless the ability to exercise control may have its place in particular cases.  It is just that it cannot provide the basis for a general principle.

“In the course of employment”

Wigmore understood the cases decided in the period when Lord Kenyon was influential to be the precursors to the test of “in the course of employment”.  From the early 19th century the general test is phrased in the cases as “scope” or “course” of employment and later, and more particularly, “in furtherance of and within the scope of the business with which he was trusted”.

The test remains with us today as a criterion of liability, but it has its limitations and is sometimes difficult in its application.  In any event it is neither a statement of policy nor a principle.  It has been clung to for some time though, perhaps in the hope that a principle would emerge.

Employer’s benefit

Lord Brougham stated the rule of liability partly on the basis of causation (“by employing him I set the whole thing in motion”) and benefit (“being done for my benefit … I am responsible for the consequences of doing it”).  Glanville Williams was not persuaded that the fact that the employer benefited from the employee’s conduct generally, could provide a basis for liability.  He pointed out that such a proposition was “impossibly wide” in the context of a society based on the division labour, in which all persons are constantly receiving benefit from the work of others.  In Sweeney v Boylan Nominees Pty Ltd, it was said that the fact that some employers profit or benefit from the relevant enterprise is insufficient itself to explain vicarious liability.  Since Lloyd v Grace, Smith & Co it has not been necessary to show that the servant’s wrongful act was undertaken in pursuance of the master’s interests.  Nonetheless, the fact that, generally speaking, an enterprise benefits from the acts of an employee is accorded importance by some common law courts today.

Early 20th century analysis

While the various “rationales” given over the years may be capable of providing explanations in particular cases, they did not articulate the policy reasons behind vicarious liability.  It may however be discerned that underlying these explanations is a concern to ensure compensation to the victim and a sense that the master who empowered the servant ought to bear the cost.

As earlier mentioned, at the beginning of the 20th century the question of the real basis for vicarious liability received the attention of some academic writers.  Amongst the first attempts to explain it were a text by Dr Baty, to whom I have referred, and a paper by Harold J Laski.  They were published in the same year, 1916.

Dr Baty was a legal scholar.  He did not hold back in his assessment of the then modern doctrine of vicarious liability:

“Unknown to the classical jurisprudence of Rome, unfamiliar to the mediaeval jurisprudence of England, it has attained its luxuriant growth through carelessness and false analogy, and it cannot but operate to check enterprise and to penalize commerce.”

He identified nine reasons which had been given for the rule of liability in the case law and was critical of all of them.  In the end he concluded that the real reason for vicarious liability was simply that employers were considered to have deep pockets.  Glanville Williams shared this view.

Laski was a British political theorist and economist and a confidante of O W Holmes Jr and Frankfurter J.  He reasoned that the rationale for vicarious liability was public policy and that the employer is held liable “because in a social distribution of profit and loss, the balance of least disturbance seems thereby best to be obtained”.  He saw employers’ liability and workers’ compensation as similar in character and as reflecting the view that “the needs of the modern state” require the burden of injury to be “charged to the expenses of production”, that is to say the employer who will pass the cost onto the community in the form of increased prices.

Some judges are more frank than others about the search for a solvent defendant able to pay compensation.  Willes J was prepared to say that “there ought to be a remedy against some person capable of paying damages”.  Courts in the United Kingdom regard it as a reason why it is fair and just to impose liability.  The other considerations are:  that the tort will have been committed as a result of activity being taken by the employee on behalf of the employer; the employee’s activity is likely to be part of the business activity of the employer; the employer by employing the employee to carry on the activity will have created the risk of the tort committed by the employee; and the employee will, to a greater or lesser degree, have been under the control of the employer.  It can hardly be denied that vicarious liability serves to distribute loss.  Professor Atiyah argued, in similar vein to Laski, that liability can be distributed over a large section of the community and over time due to insurance because most employers are now corporations.  A contrary argument is that insurance follows liability and cannot be used to create or justify it.

In Scott v Davis, Hayne J acknowledged that whether one speaks of the search for a deep pocket defendant or of loss distribution, it is clear that “considerations of insurance and the relative capacity of employers and employees to pay damages have had a significant influence on the development of vicarious liability, even if they may not provide a unifying or sufficient justification for the rules that have developed”.

Enterprise-created risk and loss distribution

There are echoes of Laski’s views in recent Canadian jurisprudence.  The employer in Bazley v Curry was a children’s foundation which conducted residential care facilities.  The Supreme Court held it liable for the acts of an employee who was a paedophile.  McLachlin J reasoned that employers are able to spread such losses through insurance or higher prices, thereby minimising the “dislocative effect of the tort within society”.

The other reason given for imposing vicarious liability related to the risk created by the enterprise. Her Honour suggested that where the employee’s conduct is “closely tied” to a risk that the employer’s enterprise has placed in the community, the employer may justly be held vicariously liable for it.  This is reminiscent of Sir Frederick Pollock’s idea that responsibility could be based on a business being a dangerous enterprise.  Since the 1990s, French law similarly shifted its focus towards a recognition that an enterprise should bear the risks created by its activities.

In State of New South Wales v Lepore Gummow and Hayne JJ suggested that analysis by reference to risk gave no significance to three facts important in that case:  the conduct complained of was intentional conduct; it directly contravened the contract of employment and was contrary to the very core of the task for which the employee was employed; and the employee would not be deterred from engaging in the conduct by the criminal law.  Moreover, they pointed out, reference to risk focuses on how the employee carried out the wrong and may therefore deflect attention from the necessary enquiry as to whether the wrong was done in the course of employment.

It was not said in Bazley v Curry that the creation of risk and distribution of loss were to be understood as policy considerations, although one might have thought loss distribution could qualify as such.  The two major policy considerations which were identified as underlying vicarious liability were the provision of an adequate and just remedy and deterrence of future harm.  One can hardly deny that the former is a worthy aim, though perhaps not a complete explanation of why the employer has to pay.  In relation to the second major policy consideration, it was explained in Bazley v Curry that the imposition of vicarious liability on employers may encourage employers to take steps to reduce risks they have introduced into the community, and therefore to reduce the risk of future harm .  The factors of risk creation and loss distribution appear to have been put forward as justifications for these policies.

There was a further, important qualification to risk creation as a reason for liability.  It was that the employee’s conduct must be “closely tied” to the risk.

Fair and just — closeness of connection

In some recent decisions, the courts of the United Kingdom have focused upon the closeness of the connection between the employment and the wrongful conduct.  In Lister v Hesley Hall Ltd, which had a similar factual context to Bazley v Curry, the House of Lords said that the basis for imposing vicarious liability was that it was “fair and just” to do so.  This was to be determined by reference to the connection between the nature of the employment and the tort of the employee.  The question was whether the torts “were so closely connected with his employment that it would be fair and just to hold the employers vicariously liable”.

In the more recent decision of Mohamud v Wm Morrison Supermarkets Plc, which concerned an unauthorised criminal act of a different kind, it was said that it was necessary to address two enquiries in order to determine liability.  The first asks what functions or “field of activities” have been entrusted by the employer to the employee.  This enquiry is to be approached broadly.  The second is whether there is a sufficient connection between the position for which he was employed and his wrongful conduct “to make it right for the employer to be held liable under the principle of social justice which goes back to Holt”.  The only real criterion for liability then becomes closeness of connection.

Mohamud’s case draws attention to the extent of the connection which is necessary for liability.  In that case the employee serving at the sales counter of a petrol station responded aggressively to a request from a customer and demanded that he leave.  When the customer did so the employee followed him to his car and physically attacked him, twice.  It was held that the employee’s conduct in answering the customer’s request was within the field of activities assigned to him.  It was held that there was a sufficient connection to the employment because there was an unbroken sequence of events when the employee followed upon what he had said to the customer.  It would seem that a causal or temporal connection might suffice.

Concluding observations

Bazley v Curry and the English cases to which I have referred may be seen to have largely accepted the general propositions that a remedy should be provided in most cases and that compensation should be paid by the employer.  If this is so, their views may not be so different from those attributed to Holt CJ.  Indeed the reference to Mohamud’s case to Holt CJ and the “principle of social justice” which, inferentially, informed his thinking may be taken to confirm that this is so.

Even so, it is not asserted by any common law court that the liability is absolute.  In each of these cases the courts say that they require a sufficiently close connection between the tortious act and the employment.  Connection then is a requirement for liability and serves as its only real limit.  It also provides some justification for the imposition of liability.  It would appear to have a much broader scope for the imposition of liability than the test of scope or course of employment.

In two cases following Bazley v Curry, which concerned similar kinds of wrongful conduct, the enquiry might be said to have been directed to the connection between the wrongful act and the employment but it did so by identifying particular features of the employment and the position in which the employee had been placed by his employment vis-à-vis the children.  This is not to say that the provision of an opportunity for the wrongful acts was thought to be sufficient.  What was considered to be crucial was the position of the authority, power, trust or intimacy that was provided to the employee.  The approach has elements of the older ideas of empowerment and authority derived from notions of agency.  It would appear to be capable of applying consistently with the test of course of employment.

Of course these are merely justifications, not principles.  They are intended to persuade the parties and the reader that there is a rational basis for imposing liability.  And they are devised for a particular kind of case; it is to be seen whether they are capable of applying more generally.

A clear basis which might explain the imposition of vicarious liability continues to elude us, although it would seem that it eludes some common law courts more than others.  It may be that the policy behind it is as simple as Baty and Laski say and that similar social ideals motivated Holt CJ.  Even so, since it is not accepted by any common law court that the liability is absolute, there is a need for a guiding principle.  It may be taken from what was said in the majority judgment in the Prince Alfred College case that hope has not been lost that a principle of general application might yet be found.  The majority clearly display that special quality of common law judges about the development of principle:  optimism.

The Hon Susan Kiefel AC 

Chief Justice of Australia