What do an estate agent, a bank, an individual acting for a company, and the brother you send to the shop to buy a pint of milk have in common? The law's answer is that they are all agents. Agency is now well-established as a distinct subject in its own right. The flexibility of agency has made it an enormously useful tool both in commerce, and outside of it.
But agency's flexibility is also its weakness. Over a century ago, Holmes characterised agency law as the simple-minded result of combining the fiction of identity between principal and agent with common sense.1 In Scott v Davis, Gleeson C.J. described agency as having a “protean nature”.2 Its flexibility comes with a drawback – the concept of agency risks becoming so amorphous that it loses any coherence as a distinct concept. Agency may come to be viewed only as an ad hoc gap-filler. Even worse, agency may be regarded only as a legal fiction – a mere device to hold particular individuals liable for the conduct of others. This is already the view of agency reasoning in vicarious liability, where Lord Wilberforce said in Launchbury v Morgans that describing a person as the agent of another only expresses a conclusion that the principal should be vicariously liable, rather than give a reason for that conclusion.3 But as Seavey once observed:
if we believe with Justice Holmes that “there is no adequate and complete explanation of the modern law, except by the survival in practice of rules which lost their true meaning,” we are left in a very unfortunate position, denied by our belief the ability to rationalize the subject and relying only upon intuition to determine when and to what extent common sense is to be applied. We are condemned to give lip-service to reason in a jargon of wise laws descending from the high priests of the legal cult.4
On the Holmesian view of agency, there appears little unity left to agency law. The threat is that agency law might disintegrate into separate pockets of law, with no unity across these pockets. If we are not careful, we might end up losing the very concept of agency itself.
The aim of this paper is to rescue agency law from these objections. I do so by offering a new definition of agency. Agency relationships, and only agency relationships, are ones where the agent has a power to exercise at least one of the principal's own powers. I call this a “proxy power” for short.5 However, there are multiple reasons justifying why agency relationships arise.
The proxy power definition brings with it three advantages. First, focusing on the agent's proxy power highlights the core idea at the heart of agency: that agency allows a person to act through another in law, not merely have another act on his behalf or for his benefit. In this way it allows principals to multiply their own legal personality in space. That distinctive idea then leads to the second advantage. It helps identify three reasons justifying when an agent has a proxy power: the principal's unilateral manifestation of will, the necessity of protecting the principal's interests, and where the principal is an artificial person who can only unilaterally manifest his will through an agent. Third, the definition not only vindicates prevailing intuitions about agency, but more importantly, gives us a transparent way to decide whether or one's invocation of “agency” is illegitimate or legitimate.
The claim advanced here is one about the English law of agency, though many of its conclusions are likely also to apply with equal force to agency in other common law jurisdictions. However, civilian concepts of agency are different in some important respects.6
The paper is set out as follows. Section II considers the problem of the sceptic's view of agency, while Section III shows that existing definitions are inadequate as they are either under-inclusive, over-inclusive or both. Section IV proposes a new definition of agency, based on the agent's proxy power, drawing from five important features of agency law. Section V shows how this definition illuminates the core idea at the heart of agency. In turn, this definition then helps identify the justifications for when agency relationships arise in Section VI. Section VII then applies the proxy power definition to different relationships, showing how the agent's proxy power vindicates prevailing intuitions about agency and gives us a clear way of identifying whether particular uses of agency reasoning are legitimate or illegitimate, and more importantly – why.
II. Answering the Sceptic's View of Agency
Agency has a deep and pervasive reach into all areas of private law today. It tells us how and when parties can enter into contracts with each other through intermediaries, it identifies some situations when someone can be held liable for the torts of another outside an employment relationship,7 it plays an important role in torts involving misrepresentations,8 it is one way of establishing that the defendant's gain is “at the expense” of the claimant in the law of unjust enrichment,9 and it is a well-known exception to the nemo dat rule in the passing of title to property.10 Yet, could the future of agency be in danger?
There are some hints in the leading English text on agency, Bowstead and Reynolds, that appear at first sight to cast some doubt on whether there could exist agency principles which are generally applicable across the law. As Bowstead and Reynolds say,
The ruling notion of agency law may be said to be that the acts of a person (the agent) authorised, or to be treated as authorised, by another are in certain circumstances to be treated as having the same legal effect as if they had been done by that other (the principal) though this approach has value in imposing some unity on the law applicable to situations where one party represents or acts for another, it should not be taken too literally. It is misleading to assume that a single set of principles, valid for the whole law, relating to the representation of one person by another can be evolved under the title of agency.11
In another paragraph, Bowstead and Reynolds appear to downplay the importance of an independent concept of agency, when they say:
Agency reasoning should therefore be regarded as a tool to be picked up as and when it may be useful. The study and separate treatment of agency as a subject is valuable, but no hidden key is to be found thereby. Indeed in the vast majority of cases problems of agency arise as questions of fact only: and perhaps the majority of cases on agency are useful not as laying down rules, but as providing examples of views that may be taken of facts.12
In some recent scholarly work, an eminent figure in the field and one of the current editors of Bowstead and Reynolds, Professor Peter Watts, has suggested that agency rules attributing acts and knowledge are context-specific and depend on the source of the right and type of the right in issue.13 For instance, attribution in contractual rights and obligations turn on the intentions of the parties to the transaction.14 This will likely differ from attribution within tort, which in turn is likely to differ from attribution within the law of restitution.15 There is a risk that a Holmesian-type sceptic might seize upon this as evidence of, the view that agency might just as well be split up into separate pockets of law, with little loss of conceptual content.
If that view is taken seriously, it then becomes very difficult to defend the continued existence of agency law as an independent subject. These views could be viewed as lending credence to the objections of Holmes and other agency law sceptics that there is no unity to agency law, nor anything distinctive about agency. If so, should we not abandon the idea that agency is a “proper title in the law”? Should we then divide up the contents of books of agency law and distribute them across books on contract law, property law, commercial law and others? The subject of agency would then cease to exist. To these objections we currently have no good answer. We ought to. The integrity of the subject is at stake.
The key to making a start in countering these objections and rescuing agency, I believe, is to offer a successful definition of the concept of agency. Most agency scholars today admit that the “concept of agency is notoriously slippery, and difficult to define”.16 But without a firm definition, it is difficult to get a grasp on the subject matter of agency. Unless we know how far agency extends, it is difficult to draw up general rules of agency which are applicable to all cases of agency. The aim of the rest of the paper is to provide a definition of agency and to show that doing so gives coherence and clarity to the legal concept of agency.
III. Problems with Existing Definitions
Existing definitions of agency have been unsuccessful in providing a definition that is neither under-inclusive nor over-inclusive. Most existing definitions of agency today focus on the presence of up to four features. They are that the agency relationship arises from consent, that the principal-agent relationship involves a Hohfeldian power-liability relationship, that the principal's legal relations can be affected by the agent's acts, and that the agency relationship is fiduciary.17 The four-part definition of agency given by Bowstead and Reynolds includes all of these features.18
Some definitions of agency seek to identify a single feature which exists in all agency relationships. One common such definition is that agency involves a Hohfeldian power-liability relationship. This view was adopted by Dowrick, who thought that “the essential characteristic of an agent is that he is invested with a legal power to alter his principal's legal relations with third parties: the principal is under a correlative liability to have his legal relations altered”.19 A similar view focuses on the ability of the agent to affect his principal's legal relations. One supporter of this view is Fridman. He described agency as “the relationship that exists between two persons when one, called the agent, is considered in law to represent the other, in such a way as to be able to affect the principal's legal position in respect of strangers to the relationship by the making of contracts or the disposition of property”.20 Similarly, in International Harvester Co. of Australia Pty Ltd. v Carrigan's Hazeldene Pastoral Co., the High Court of Australia defined agency as “a word in the law to connote an authority or capacity in one person to create legal relations between a person occupying the position of principal and third parties”.21
The problem with these definitions is that they are over-inclusive. They are too wide and therefore also encompass legal relationships which are not ones of agency. Agency cannot be defined solely by reference to a Hohfeldian power-liability relationship, because such relationships exist in many situations in the law outside agency. For example, the process of offer and acceptance in contract formation can also easily be analysed as involving a Hohfeldian power-liability relationship, with the offeror having a power to make an offer to the offeree, who has a corresponding liability to be made an offer to. Similarly, the legal relations of one party can also be affected by the acts of another outside the law of agency. The best example involves the express trust. A trustee holding property on trust for beneficiaries can affect the legal relations of the beneficiaries if he sells the property within the terms of the trust deed and without breach of duty, since the beneficiaries’ rights are overreached and the purchaser takes free of the beneficiaries’ rights.22 In doing so, the beneficiaries’ legal position against the purchasers changes. Yet, it is well-established that the relationship between trustee and beneficiary is not one of agency.
This is probably why most definitions today rely on more than one of the four features. For example, Seavey defined agency by focusing on consent as the reason why agency relationships arise, the fiduciary nature of the relationship, and the analytical structure of the relationship. In Seavey's words, “Agency is a consensual relationship in which one (the agent) holds in trust for and subject to the control of another (the principal;) a power to affect certain legal relations of the other”.23 The US Restatement, Third on Agency defines agency by relying on consent as the normative reason why agency relationships arise and the fiduciary nature of the relationship.24 McMeel identified the defining characteristics of agency as the consensual nature of the relationship and a Hohfeldian power-liability relationship, though he accepts that this definition is not exhaustive.25 In contrast, Munday defines agency by focusing on the fiduciary nature of the relationship, that the agent alters legal relations between the principal and third parties, and the Hohfeldian power-liability relationship,26 but does not think that consent is part of the essence of agency.27
These definitions suffer from a different problem: they are under-inclusive. They rely on either or both consent or the fiduciary nature of the relationship as part of their definitions, but agency relationships can be non-consensual or non-fiduciary.
Three examples of non-consensual agency can be given. The clearest example is the relationship between a company and the human individuals which act for it, which is frequently described as one of agency.28 At first glance, it might be thought plausible to say that the company consented to an individual A acting for the company as its agent. But someone else must consent on behalf of the company. Let us call this person A2. In turn we might ask how A2 has the power to consent on behalf of the company. This question can only be answered by asking who gave A2 that power, which will be another individual A3 or group of individuals, such as the board of directors, acting for the company. At least under English law, this question is ultimately answered by the company's constitution, including its articles, which allocate the company's powers to particular individuals or different groups of individuals. But the company's constitution itself cannot be described as a product of the company’s consent.29 For a company to be incorporated by registration under modern legislation, it must have articles.30 The company's articles cannot be described as based on the company's consent, since the company itself is created through the process of incorporation, and the company's articles acquire legal force through that act of incorporation. Prior to incorporation, there is simply no company, nor are there articles of the company, since the company does not yet exist.31 The important point is that the agency of individuals acting on behalf of companies does not arise because of the company's consent but rather arises because of allocations of power under the company's constitution, given effect to under the legislation providing for registration by incorporation.
Agency of necessity cases are also difficult to analyse as arising from consent. These are cases where the master of a ship may have powers in cases of emergency to contract for the cargo to be carried forward,32 or sell or hypothecate the cargo whether together with, or separately from the ship.33 These are true cases of agency where the master enters into contracts on behalf of another, binding the principal to a contract with a third party.34 While this relationship is sometimes analysed as ones based on implied consent, the decision of The Choko Star suggests that it is not.35 There, the Court of Appeal held that a shipmaster had no authority to contract for salvage services on behalf of the cargo-owners on the basis of the cargo-owner's implied consent where the requirements of agency of necessity were not met.36 This suggests that agency of necessity and agency arising out of implied actual authority are separate and distinct, and that agency of necessity is not based on implied consent. Rather, as later seen, agency of necessity is justified on the basis that the principal is unable to effectively exercise his powers to appoint an agent and it is necessary in the circumstances to protect the principal's interests by giving the agent agency powers.37
Finally, there are examples of agency relationships arising under statute, which arise irrespective of the principal's consent.38
A similar problem exists for definitions which rely on the fiduciary nature of the relationship, since not all agency relationships are fiduciary.39 Two clear non-fiduciary agency relationships exist. They are cases of agency coupled with an interest, and the receiver of a company. In cases of agency coupled with an interest, the agent can use the authority not solely for the benefit of his principal but for his own benefit as well, so he does not owe fiduciary duties in the exercise of his authority.40 Similarly, a receiver of a company typically acts in the interests of the debenture holder, not the company and does not owe any fiduciary duties to the company.41 But both agency coupled with an interest and the receiver are regarded as true cases of agency.42
Existing efforts to define agency have been unsuccessful. Definitions focusing on one feature alone are over-inclusive, while definitions which rely on consent or the fiduciary nature of agency are under-inclusive. Yet, without a definition of agency, it is difficult to address the sceptic's criticisms levelled at agency. If it is difficult to tell whether a person is an agent, there is the risk that we may wrongly identify particular persons as agents when they are not. If we then apply general agency rules to them, this may then lead to inappropriate outcomes. In turn, this may cast doubt on whether there are general rules of agency law, pointing further in the direction of a process of disintegration and providing further fuel for the agency sceptic.
IV. An Agent's Proxy Power: A New Definition
To nip this potential problem in the bud, I want to put forward a new definition of agency. A relationship is one of agency if the agent A has a proxy power: a power to exercise at least one the principal P's own powers. I do so by examining five of the most important and distinctive sets of rules in agency law and ask what feature or features can explain those rules. Examining these distinctive rules are likely to illuminate the feature or features that unite agency law and what makes it distinctive.
Five sets of rules can be identified. The first set involves rules that P's legal relations with third parties can be affected through A's acts. Second, A's own legal relations with third parties are often not affected through A's acts. This is often described as the rule that A “drops out”. A third important set of rules is captured by the fundamental principle of agency that P cannot be put in a better position if he acts through an agent than if he had acted personally. Fourth, and conversely, another set of rules establish that an agent is able to act for a principal even if he could not do those acts personally on his own behalf. The fifth set of rules concern the operation of sub-agency, where the sub-agent can affect the principal's legal relations but is not regarded as the principal's own agent.
When examined, these five sets of rules show that key to these rules is the idea of the agent's proxy power: a power to exercise at least one of P's own powers in dealing with third parties. It is submitted that it is the presence of this feature of agency that defines agency. A relationship is one of agency if the agent has a proxy power. It is not one if the agent does not have a proxy power. It is the agent's proxy power that unites agency law and makes agency distinctive. Crucially, it is agency law's formal structure that unites it, rather than the reasons why agency relationships arise. Agency relationships may arise for multiple reasons, but all of them possess the same form of a proxy power. Its form is what gives the subject its shape and structure, and its integrity.
A. P's Legal Relations Are Affected through A's Acts
The best-known rule of agency law that when an agent does an act, it is as if the principal does the act himself. This is commonly expressed by the Latin maxim qui facit per alium, facit per se.
We see this rule at work in all areas of law. In contract law, an agent A, acting on the principal P's behalf, can enter into a contract on behalf of P with a third party, resulting in the formation of a contract between P and the third party.43 Outside contract law, in the law of torts A can also act as P's agent in making representations to third parties on P's behalf.44 The clearest case is where A makes a representation on behalf of P as P's agent, and the representation turns out to have been negligently made, then P can be liable for the tort of negligent misrepresentation if he assumed responsibility for the statement.45 In unjust enrichment, where A acts on P's behalf in making payments to a third party under the mistaken belief that P owes the third party a debt, P can recover the mistaken payments from the third party in unjust enrichment. These were the facts of the classic case of unjust enrichment, Kelly v Solari.46 Conversely, A can act on P's behalf in receiving money from a third party. If the third party paid under a mistake, it appears that he can bring a claim in unjust enrichment against P, even if the money was never paid over to P by A.47 Crucially, this feature exists even where even in non-consensual agency relationships.48
This fundamental feature of agency law is often used to suggest that that A is exercising a power to act on behalf of P. but this formulation is not specific enough, since it cannot exclude how trusts work. More precisely, what A is doing is exercising a power to exercise one of P's own powers as against third parties. This may be to make or accept offers to enter into contracts (as in the contract cases), to make representations and assume responsibility for them (in the tort cases), or to make or receive payments (in unjust enrichment). In all these cases A is exercising one of P's own powers. This explains why P's legal relations are affected as if the act was done by P himself. It is because P's own powers are exercised, but through A, rather than through P's own body.49
B. A's acts sometimes do not affect A's own legal relations
Another well-known set of rules concern the effect of A's acts on A's own legal relations with third parties. In some cases, A's own acts do not affect A's own legal relations with third parties. A is said to “drop out”.50 The best-known example of A dropping out is where A acts on P's behalf in entering into a contract with a third party. In doing so, A “drops out” and does not enter into a contract with the third party since A does not purport to act in a personal capacity. This is so even where no contract is successfully entered between P and the third party.51 Similarly, in the law of negligent misrepresentations, A typically also drops out. In Williams v Natural Life, the managing director of a company was not liable for negligent misrepresentations he made to a third party because he did not assume personal responsibility for it, and thus owed no duty of care in respect of the statements he made.52
These rules can be explained by a combination of A's proxy power and general rules of private law. In cases of agreements,53 a person only comes under an obligation when he voluntarily assumes the obligation. In these cases, the objective theory of contract formation also applies.54 Where A enters into a contract on behalf of P, he does not undertake personal responsibility for that obligation, objectively viewed. He therefore exercises only his proxy power, but not A's own personal powers to contract. However, if A does undertake personal responsibility for the obligations under the contract, objectively viewed, then he does not drop out. In that case, he exercises both his own power to contract as well as his proxy power to contract for P, so that both A and P come under contractual obligations to the third party.
However, in cases outside of agreements and voluntarily undertaken obligations, whether the agent comes under an obligation does not turn on his voluntary assumption of responsibility, nor does the objective theory of contract apply. In such cases, the cases suggest that the agent does not drop out. For example, this is so for the tort of deceit.55 In these non-agreement cases which do not turn on an assumption of responsibility, when an agent does an act, it appears that he does it both personally and on behalf of another. Since the act is done by A's physical body, it is his act and A cannot say that he is not liable because he did the act only for P.56 Thus, when A intentionally makes false representations, both A and P can be liable for deceit.57
C. P Cannot Be Put in a Better Position by Using an Agent
A third set of agency rules is encapsulated by the maxim that P cannot be put in a better position by using an agent than he would have been in had he acted personally. This has been described as a fundamental principle of agency.58 It manifests itself in different ways across the law of agency but can be most clearly seen in relation to principals who lack powers, or principals who have no legal existence for various reasons.
Four examples can be given. One example involves pre-incorporation contracts entered into by A on behalf of P, a company which was non-existent at the time of formation. In Newborne v Sensolid (Great Britain) Ltd.,59 where the Court of Appeal held that no contract had been formed between P and the third party, as the company was non-existent at the material time.60 Another example involves companies who lack power under their constitutions to do particular acts. These types of cases were historically often described as ones involving the ultra vires doctrine,61 until “ultra vires” was limited to refer only to cases where a company lacked the capacity to do an act under its constitution.62 In these cases, it is clear that no contract is formed between P and the third party through A's act.63 A third manifestation of this principle is where P dies or is dissolved. In these cases, P's legal relations with the third party are not affected by A's acts.64 Fourth, if A purports to act for P in disposing of P's property to a third party, but in the meantime P is found bankrupt and a trustee in bankruptcy is appointed, then A's purported disposition is generally void unless the recipient is a bona fide purchaser.65 This maxim can also be extended to other contexts, such as in determining when agents can act on behalf of a principal with mental incapacity at common law,66 or in imputing an agent's knowledge to his principal.67
In all these cases, the principal cannot be put in a better position by acting through an agent than if he had acted personally. If he would not have been able to affect his own legal position by exercising his powers personally, he cannot do so through the agent either. This feature of agency reflects A's proxy power. In all these cases, the reason why P cannot be put in a better legal position by acting through an agent than if he had acted personally is because A only has a proxy power, a power to exercise P's own powers. If P lacks these powers, A cannot have any greater powers to exercise on P's behalf. This is so regardless of whether the reason why P lacks powers is that he has not yet had any legal existence, his legal existence has come to an end because of death or dissolution, because his powers are limited by virtue of his constitution (in cases of incorporated bodies), or because of the event of bankruptcy. In all these cases, the underlying principle is the same, and it is explained by the agent's proxy power.
D. P Can Act Through A Where A Has Limited Powers
The agent's proxy power also helps explain another set of agency rules that P's legal position can be affected by A's acts even where A had no power to do that act personally. This is the converse of the principle that P cannot be put into a better position by acting through an agent.
Three examples of this rule can be given. If A is a minor, he or she has limited powers to contract on his or her own behalf, save for contracts of necessaries, but he can still act for P in entering into contracts on behalf of P, so that P enters into a contract with the third party through A.68 Another example involves married women. Before 1882, married women had no powers to own property and had no powers to contract personally,69 but they could still act as their husband's agent in entering into contracts to purchase goods.70 A third example of more commercial relevance involves the passing of title to goods. The core rule is that a principal is bound by dispositions of property made by his agent, acting within the scope of the agent's actual and apparent authority.71 P, the owner of goods, may wish to sell the goods through an agent, A. Only P has title to the goods, not A. However, if A sells the goods to a third party, acting as P's agent and with the authority to sell the goods, then A's acts in selling the goods can be sufficient to transfer title in the goods from P to the third party.72 This is even though A himself does not have title to the goods, and thus has no power to give good title to the third party.
The agent's proxy power can explain this feature of agency as well. In each case, since what A is doing is exercising a power to exercise the principal's own powers, it does not matter that A has no power to do the act personally. In the case of married women, it does not matter that A has no power to contract personally or hold property in her own name, and in the case of the passing of title, it does not matter that A cannot give good title by virtue of his own title. What matters is that P had powers to do those acts. If he did, then A can have a proxy power and can affect P's legal relations by exercising it. All this seems to require is that he can understand the nature of the act that he is doing and that he can perform the act required.73 This is all that is required for him to be able to exercise his proxy power.
Finally, the agent's proxy power also explains the operation of sub-agency.
The term “sub-agency” is sometimes used to describe two different situations, both of which allow the principal's legal relations with third parties to be affected through the sub-agent's acts.74 The first situation is where A, P's agent, appoints the supposed sub-agent S in substitution for or in addition to A himself.75 The second situation is where A himself appoints his own agent S to exercise A's powers, which include A's powers to exercise P's own powers. In both cases, P's legal position can be affected by the acts of S.76 Bowstead and Reynolds consider only the second case a true case of sub-agency, because S is not P's agent, but A's agent. The idea of the agent's proxy power helps explain why.
In the first case, S himself becomes P's agent, not A's agent. A and S are really co-agents.77 S here has a proxy power to exercise P's own powers, not A's. If A dies, S continues to be able to affect P's legal position. In contrast, in the second case, S is A's agent, not P's agent.78 S has no proxy powers to exercise P's own powers without A's interposition; he only has powers to exercise A's own powers. This means that if A dies, S can no longer affect P's legal position, since S only has a power to exercise A's own powers, and A's death means that A no longer has any powers.
The agent's proxy power thus helps explain why only the second situation involves sub-agency. It also explains how it is that S can affect P's legal relations, even though S is not P's agent. It is because S is exercising a power to exercise A's own powers, which include a power to exercise P's own powers. These results are true even for cases of non-consensual agency, such as that involving individuals acting for companies.79 Just as with the other features examined earlier, sub-agency too can be explained by the agent's proxy power.
The agent's proxy power can explain these five important sets of rules of agency. It is submitted that agency relationships can be defined by the existence of an agent's proxy power, since the agent's proxy power is reflected in and underpins some of most important features of agency law. In order to determine whether the relationship is one of agency, we should ask whether the agent has a proxy power. If he does, then the relationship is one of agency, and these five features of agency will apply. If not, then the relationship is not one of agency.
V. The Distinctive Idea behind Agency
Defining agency based on the idea of an agent's proxy power is not just a technical exercise in classification. Above all, defining agency by reference to the agent's proxy power is important because it is the agent's proxy power that captures the distinctive idea at the heart of agency: that agency allows one person to act through another. This is how agency allows a person to “multiply his legal personality in space”.80 By focusing on this distinctive idea, we can then see when and why agency relationships arise.
The agent's proxy power gets to the heart of the legal concept of agency. Agency is often described broadly as involving cases where one person acts for another, so that his acts can be identified with the acts of the person for whom he is acting. Another similar way of expressing this idea is that one person may act on the behalf of the other. But neither of these expressions are quite precise enough, since in law, many people can be said to act for another or on behalf of another in some sense. Fiduciaries can be said to act for and on behalf of another, yet not all fiduciaries are agents. In a sense, even persons who provide services to another can be said to act for the other person.
What is truly distinctive about the legal concept of agency is the idea that through agency, one person can act through another. This feature is exactly what the agent's proxy power focuses on. Generally, our legal responsibility for our acts is circumscribed by the acts of our own physical body and the consequences of those bodily movements. Agency is an exception to this. It allows the legal relations of the principal to be affected by the acts of another person, the agent. The reason why the principal's legal relations with third parties are affected through the agent's acts is that in doing the act, the agent is exercising a power to exercise the principal's own powers. Through the agent's acts, the principal himself acts, because it is the principal's own powers to deal with third parties that have been exercised. The great contribution of agency was to allow people other than the principal to exercise the principal's own powers. It was the recognition of the agent's proxy power. This is how agency allows a person to “multiply his legal personality in space”.81 Agency does so by allowing him to act through the body of another.
This idea is the key to understanding most features of the law of agency without the need for fictions. The reason why the principal's legal relations are affected through the agent's acts is that the agent has a proxy power to exercise the principal's own powers. It is unnecessary to explain this result by relying on the fiction that identities of the principal and agent merge together.82 Agency involves only the simple idea that legal persons can act through other legal persons. It is this idea that is at the heart of agency and that is distinctive to agency. The mechanism of acting through another is through the agent's proxy power.
VI. Justifying when Agency Relationships Arise
This core idea that agency allows principals to act through others, not merely for or on behalf of others, then provides another contribution: it helps us identify the justifications for when agency relationships arise. In particular, what we need are justifications for when one person should have a power to exercise the powers of another. It is submitted that English law today recognises only three justifications for the creation of proxy powers.
Generally, legal persons are regarded as small-scale sovereigns over their own powers to act,83 so that it should be the principal alone who decides whether or not to exercise his own powers to make offers in forming contracts, to transfer property to another, or to make representations. But it would be impractical and inconvenient for the principal if he could only exercise those powers through his own body. Thus, one way in which the law could facilitate social and economic activity is by recognising in the principal the power to create a proxy power in others.84 This is exactly what the law did.
Most agency relationships therefore arise where the principal unilaterally manifests his will to give an agent a proxy power.85 In most cases, the principal will manifest his will to the agent, consenting to give the agent a proxy power, and the agent will also consent to have a proxy power to act on behalf of the principal. But the agent's consent is not necessary, since the principal can create proxy powers by a unilateral act such as a deed.86 Most agency relationships arise by way of the principal's unilateral manifestation of will, both in the commercial and non-commercial sphere.
Because the principal generally has control over his own powers, including whether to create proxy powers, the general rule is that no proxy powers will be created without the principal manifesting his will to do so.87 But there are two instances where proxy powers will arise by operation of law because the principal is unable to effectively manifest his will one way or another: where it is impossible to contact the principal and it is necessary to protect the principal's interests, and where the principal is an artificial legal person who cannot exercise his own will at all without the use of proxy powers.
First, a proxy power will arise by operation of law where the principal does have the powers to grant the agent a proxy power but is unable to effectively exercise those powers, and it is necessary to protect the principal's interests that the agent has a proxy power. This is justified because the principal's control over his own powers would be of little practical value if his interests in his property or bodily integrity were already lost by the time he could effectively exercise control over those powers.
The case of the shipmaster in agency of necessity falls into this category. The agent's proxy power to sell the cargo arises by operation of law, so as to protect the principal's interests in circumstances where he cannot effectively grant the agent a proxy power but where his interests would be harmed if nothing was done. This explains the strict requirements that must be met in order for agency of necessity to arise, particularly that it must be impossible or at least extremely impracticable for the agent to communicate with the principal,88 not merely difficult.89 Where the principal could be communicated with or if the action was not necessary to protect the interests of the principal, then the shipmaster should seek the principal's instructions, waiting to do so if necessary.90
Another example of this category is found under statute in the Mental Capacity Act 2005. A may have proxy powers to pledge the credit of a mentally incapax person P in order to pay for P's care and treatment.91 In this case, A's proxy power arises under the statute because it is necessary to protect P's best interests by allowing him to obtain care and treatment and because P is incapable of exercising his own powers to pledge his credit to pay for his own treatment because of his mental incapacity. These requirements are enshrined in the Act itself.92
Second, proxy powers will arise where the principal may be an artificial legal person which cannot exercise its own will without the use of proxy powers. The most common cases today involve companies, though this category also comprises other artificial legal persons. Companies, like all artificial legal persons, can only act through human individuals.93 Because companies can only act and make decisions through the acts of human individuals acting for the company, companies cannot unilaterally manifest their will to exercise powers to create proxy powers without already having individuals who have proxy powers. This is why it is a necessary condition for the incorporation of a company that the company's powers are allocated to some human individuals acting for the company, whether under the company's constitution and then given effect through the incorporating statute or conferred directly under the relevant incorporating statute.94 Without these proxy powers arising by operation of law through the incorporating statute, the principal simply could not exercise its own will to do anything, even to create proxy powers.
VII. Applying the Definition: Legitimate and Illegitimate Uses of Agency
The proxy power definition has a third advantage. When it is applied to relationships familiar to agency lawyers, it largely reaches conclusions that agency scholars and courts already reach. Its virtue is that does so in a clearer and simpler way than existing definitions. It avoids the problems of weighing up the presence or absence of multiple characteristics, such as the presence of fiduciary duties or the principal's control over the agent, to decide whether the relationship is one of agency or not.95 On the proxy power definition, a relationship is one of agency when the agent has a proxy power. It is not if the structure of the relationship simply does not give rise to any proxy power or because giving the agent a proxy power cannot be justified. This provides a clear way of identifying whether particular uses of agency are legitimate or illegitimate.
This section applies the definition to four categories of relationships: relationships which are regarded as being on the fringe of central agency principles because they possess some but not all of the typical feature of agency, other well-known tripartite relationships such as the bailment and the trust, two-party relationships involving the internal relationship between principal and agent only, and relationships where although an agent might theoretically have a proxy power, there is no justification for such a proxy power.
A. “Fringe” Agency Relationships
The proxy power definition helps us understand some relationships which are often regarded as on the fringe of central agency principles. Despite lacking all the features of a typical agency relationship, these relationships are still frequently described by courts and commentators as ones involving legal agency. Focusing on the proxy power tells us why they may often be true agents.
The “canvassing” or “introducing” agent is one such example. He is simply hired to introduce parties but does not make contracts on the parties’ behalf,96 and is generally regarded as being on the fringe of central agency principles used by the common law.97 The introducing agent might be regarded as a true agent on the basis that he owes fiduciary duties.98 But on this explanation, it is difficult to see why a trustee cannot also be described as his beneficiaries’ agent simply because he owes fiduciary duties. The proxy power model provides a better explanation for why the canvassing agent will frequently be a legal agent: he will typically have some limited proxy powers to act on behalf of the principal at least to make representations to the prospective counterparty. If he does have proxy powers to act for the principal, even if only in making representations, then he will be an agent. If he does not have any powers to act for the principal, he will not. This explains at least some important cases where the principal of a canvassing agent has been held liable in tort when its agent makes false representations.99
Another example involves irrevocable agency. These involve cases where the agent has a relevant interest of his own in the exercise of his authority, so that the principal cannot revoke the agent's authority at any time.100 An example is where a creditor is given a power of attorney to sell the principal's land so as to satisfy a debt owed by the principal to the creditor.101 Bowstead and Reynolds regard the use of agency reasoning in the case of irrevocable agency as inappropriate on the basis that it is purely a device used to create property or security interests102 and “is not really within the bounds of general agency reasoning”.103 Despite this, Lord Sumption clearly regarded irrevocable agency as involving “true” agency in the recent decision of Angove's Pty Ltd. v Bailey. In his words, “there is no principled reason why a true agent employed on his principal's affairs should not also be regarded as having a personal interest in the exercise of his authority sufficient to make it irrevocable”.104 The proxy power definition tells us why these cases are still ones of true agency. It is because the agent has a proxy power.105
B. Agency, the Trust and Bailment
The proxy power also tells us why some relationships, such as the trust and the bailment, are not ones of agency. It is widely accepted that agency is conceptually different from the trust or the bailment.106
Just like with agency, beneficiaries’ legal relations with third parties can also be affected through the trustee's acts. The difference between agency and trusts lies in the agent's proxy power, which the trustee lacks. An express trustee typically has powers to transfer title to the trust property to some third party or to sell the trust property and receive the proceeds of sale. But the trustee's powers to do these acts does not arise because the trustee has a proxy power, but because he has legal title to the trust property.107 Similarly, a trustee has powers to enter into contracts with third parties to the trust, but these powers to do so stem from his own status as a legal person,108 and are not powers to exercise the beneficiaries’ own powers to enter into contracts.
Similarly, the agent's proxy power also helps distinguish bailments from agency. Bailments and agency are also widely recognised as being conceptually distinct relationships.109 A bailment relationship is one where the bailer passes possession in tangible personal property to the bailee. The bailee has no powers to exercise the bailer's own powers in relation to the bailed property. The powers that the bailee has are either those he has by virtue of his status as a legal person, or those which he has by virtue of having possession of the property. They do not involve proxy powers, but agency does.
In both bailment and trusts, the relationships centre around the holding of property. The trustee or bailee's powers stemming from their legal title to the property or to their possession of the property, not by virtue of another person. In contrast, agency relationships need not involve any property at all, but merely involve one person acting through another via the agent's proxy power.
C. Two-Party Relationships
Just as the structure of the trust and bailment tell us why a relationship cannot be one of agency because no proxy power exists, there is also no proxy power in two-party relationships. These relationships thus cannot be ones of agency. This is because the agent's proxy power always envisages there being three parties: the agent himself who has a power to exercise the principal's own powers, the principal's powers which can be exercised with relation to third parties, and a third party against whom the principal's powers can be exercised.
This explains why those who do acts for the benefit of another in cases of necessity and then seek reimbursement (similar to the Roman doctrine of negotorium gestio) are not now regarded as true agents.110 Bowstead and Reynolds now regards these cases as better belonging in the law of restitution.111 Regardless of whether this is correct, the proxy power definition explains why these cases are not ones of agency. The person who seeks reimbursement had no proxy power.112
It also explains why “indirect representation” does not involve agency. In these cases, a principal appoints an intermediary to deal on his behalf on the understanding that the intermediary will deal in his own name as principal in any dealing with a third party.113 These cases are sometimes referred to as ones of agency, because the internal rights and duties of the principal and intermediary in these cases are the same as that of those in classic cases of legal agency.114 But the intermediary typically has no proxy powers to deal with a third party on the principal's behalf, so they are not legal agents, but at best only functionally similar intermediaries.115
D. Non-Justifiable Agency Relationships
Finally, defining agency by reference to the agent's proxy power provides a way of explaining why some uses of agency reasoning are inappropriate in some other more difficult cases. In these cases, although the agent could have a proxy power, there is no justification for the creation of that proxy power. The idea of the agent's proxy power helped us identify when the creation of an agent's proxy power can be justified. From there, we also can see when it is not justified. Three examples, arising in a variety of different contexts, are discussed.
A historical example involves the agency of married women. Before 1882, married women did not have powers to hold property in their own name, nor to enter into contracts in their own name. They were thus regarded as acting as their husband's agent when they purported to buy goods.116 The justification for giving a married woman a proxy power was the agent's necessity, since a married woman would otherwise be unable to support herself if her husband refused to do so. After statutory changes gave married women powers to act in their own name, this category of agency slowly faded from history.117 It is submitted that creation of proxy powers on the grounds of the agent’s necessity cannot be justified since it is incompatible with the idea that the principal is entitled to control over the exercise of his own powers, nor is it defensible on the grounds of protecting the principal's interests. It was an inappropriate use of agency, used only to escape the consequences of the married woman's limited powers.
A more modern example involves the use of agency in the context of motor cars to hold the owner of a car liable in tort for the negligent driving of the driver. It may be that some cases can be regarded as legitimate applications of true agency,118 but most cannot be. This principle has been used where a husband driver negligently drove a car owned by his mother, injuring his wife, a passenger in a car, so that the husband was regarded as driving the car as the mother's agent.119 Even if it were theoretically possible to explain the husband as having a proxy power to drive the car for the mother, there is no justification for doing so since there was no unilateral manifestation of will that the driver was exercising a proxy power to drive the car for the owner, no situation of necessity, nor is the owner an artificial legal person who conferred the power on the driver. There was therefore no defensible justification for why the agency relationship arose on those facts. Agency then was relied on only to escape the effects of a statutory provision that gave the husband a defence against the wife for committing a tort120 or the availability of insurance.121 Today this problem would likely not arise after changes in the availability of insurance.
A final example is apparent authority. The US Restatement Third describes apparent authority as a true case of agency where the agent has a proxy power.122 But how can this proxy power be justified? It cannot be justified on the grounds of necessity nor does it apply only to artificial legal persons. Perhaps it could be defended on the basis that the principal unilaterally manifests his will to create a proxy power to the third party but not to the agent. However, this argument is problematic because it is hard to see how the agent can exercise a proxy power when he does not know that he has one. Even more problematically, there are many cases where the principal clearly manifests his will to the agent that the agent does not have authority to do particular acts. In that case, the principal appears to have unilaterally manifested his will differently both to the agent and the third party. The mere unilateral manifestation of will to the third party cannot justify conferring a proxy power on the agent when it would be contrary to the unilateral manifestation of the principal's will to the agent. This analysis suggests that there is no justification for the creation of a proxy power in the agent under the doctrine of apparent authority. In contrast, the English law of agency regards apparent authority as based on an estoppel by representation but does not regard it as a true form of agency.123 On this view, it only estops the principal from denying the agent's authority, but does not confer a proxy power on the agent. In this respect, the English position appears to be the better one.
Applying the proxy power definition of agency to some familiar categories of relationships shows that it largely reaches the intuitions of agency lawyers and the reasoning of courts. Its added virtue is that it can explain not just when, but more importantly the different reasons why agency reasoning is appropriate or inappropriate. In some cases, the agent simply has no proxy power at all, because the structure of the relationship between the principal and agent does not give rise to one. In some other cases, the agent could theoretically have a proxy power, but there is no justification for the agent's proxy power. Those situations are ones where agency reasoning is inappropriate, since the relationship is not one of agency.
On the sceptic's view, agency risks disintegrating into separate pockets of law with nothing unifying them. If we do not take care to arrest this trend, we might eventually lose the very concept of agency itself. To fend off the sceptical view, the central aim of this paper has been to offer what is hoped to be a successful definition of the concept of agency – a relationship is one of agency where an agent has a proxy power: a power to exercise at least one of the principal's own powers.
This definition, if accepted, brings with it other added payoffs. Focusing on the agent's proxy power highlights the distinctive idea at the core of agency: that people can act through others in the law, thereby expanding their legal personality in space. This distinctive idea helps identify the three reasons justifying when an agent has a proxy power: the principal's unilateral manifestation of will to grant the agent a proxy power, where it is necessary to grant the agent a proxy power in situations of emergency so as to protect the principal's interests in his bodily integrity or property, and where the principal is an artificial person who can only unilaterally manifest his will through an agent.
Moreover, the proxy power definition vindicates prevailing intuitions about agency, and gives us a transparent way to decide whether or one's invocation of “agency” is illegitimate or legitimate. When applied to relationships familiar to agency lawyers, it largely reaches conclusions that agency scholars and courts already reach. The added virtue of the proposed definition is that it does so in a clearer and simpler way than existing definitions. A relationship is one of agency when the agent has a proxy power. It is not if the agent has no proxy power, either because the structure of the relationship does not give rise to any proxy power or because giving the agent a proxy power cannot be justified.