chevron-down Created with Sketch Beta.

The International Lawyer

The International Lawyer, Volume 56, Number 2, 2023

The Proper Role of Agency in Corporate Group

King Fung Tsang and Kasper Fan


  • Agency law has, for many years, played an uncertain role in the relationship between parent companies and their subsidiaries.
  • This article aims to understand the relationship's proper role, and the starting point is clear: Salomon v. Salomon & Co. English company law is very much dictated by, and revolves around, this seminal case.
  • At the core of the case is separate legal personality, that is, the principle that companies have a separate legal identity from their shareholders.
The Proper Role of Agency in Corporate Group
Gregory Olsen via Getty Images

Jump to:

I. Introduction

Agency law has, for many years, played an uncertain role in the relationship between parent companies and their subsidiaries. This article aims to understand the relationship’s proper role, and the starting point is clear: Salomon v. Salomon & Co. English company law is very much dictated by, and revolves around, this seminal case. At the core of the case is separate legal personality, that is, the principle that companies have a separate legal identity from their shareholders. Today, the general effect of this principle on corporate groups is that parent companies do not share liability with any of their subsidiaries. But there have been, and remain, exceptional cases where liability must be directed beyond a company itself. One such exception are cases involving the relationship of agency. Unfortunately, nothing about this exception is clear, including its theoretical basis, the proper circumstances under which it will be triggered, the effect of its application, and its interaction with other exceptions.

Around a century ago, Justice Cardozo, one of the most distinguished judges in Anglo-American law, observed that “[t]he whole problem of the relation between parent and subsidiary corporations is one that is still enveloped in the mists of metaphor. Metaphors in law are to be narrowly watched, for starting as devices to liberate thought, they end often by enslaving it.” Perhaps the best known principle voyaging the mists, receiving the most attention to date, is piercing the corporate veil. In fact, Lord Neuberger adopted the same quote in the Supreme Court of the United Kingdom case, Prest v. Petrodel Resources Ltd. This case addressed the English experience on piercing the corporate veil and suggested that it has been unsatisfactory. The Supreme Court of the United Kingdom decided that piercing the corporate veil was overly broad and often applied with much uncertainty, heavily limiting its scope of application, and suggested that other legal principles address corporate liability instead. One such principle specifically mentioned by Lord Sumption, Lord Neuberger, and Lady Hale was agency. In particular, having relegated the piercing doctrine to covering only cases involving evasion of existing liabilities, Lord Sumption sought to explain the shareholder’s liability in Trustor AB v. Smallbone (No. 2) using agency instead. Similarly, Lord Neuberger, who agreed with Lord Sumption’s narrow formulation of the piercing doctrine, redefined Gilford Motor Co. v. Horne as an agency case.

However, neither Law Lords went into detail about how agency should be applied as a principle, nor did they set out any criteria for doing so. Professor Sarah Worthington opines that, despite this lack of clarity, agency “is [still] a much better ‘difficult question’ than the unfocused question [of] ‘should the veil be pierced.’” Yet she too provided no reasoning nor explanation as to why this is the case. Thus, we are left with the all-important question: Is agency the breeze that will finally carry away the mists, or are we merely sweeping all the uncertainty and inconsistency under a new rug and creating another metaphor?

To answer this question, two sub-questions must first be addressed. First, how do we identify agency in a corporate context? To date, both English courts and academics have been unable to provide an answer. Instead, inconsistent and incoherent approaches have constantly confused this area of the law. The authors found it appropriate to lend some reference from the U.S. courts and academics in hopes that their similar criteria for agency shed some light on what a correct test could be. Both jurisdictions share many similarities in their test for agency, so direct comparison can be made between some of the elements. In particular, focus is placed on the role of the principal’s control over its agent and why such control is an essential ingredient in finding agency in the United States but not in the United Kingdom.

Second, where does agency fit amongst the arsenal of principles English law has to tackle parent-subsidiary liability? While the Supreme Court of the United Kingdom has hinted that agency may be a potential alternative to piercing the corporate veil, there are inherent differences between the two principles that seem to make directly substituting one for the other impossible. Besides, there are more than two principles in English law that deal with parent-subsidiary liability. For example, English courts have recently made use of a line of authorities originating from Lord Bingham’s obiter in Lubbe v. Cape Plc, finding that parent companies can be liable to tort victims of their subsidiaries if they voluntarily assume responsibility for the subsidiaries. As such, it is necessary to clarify what role, if any, agency can play in this developing landscape, especially after Prest took piercing out of the picture.

The article will be organized as follows: In Section II, the authors discuss the use of agency in the context of a corporate group from the early days of English company law. Section III discusses the three tests for agency that can be identified from English precedents, along with each of their shortcomings. The authors then attempt to formulate a correct test for England in Section IV with the help of U.S. precedents. We argue that the proper test to identify agency must come from general agency principles. A parent company should be liable for the acts of its subsidiary if the subsidiary is authorized, whether through actual or apparent authority of the parent company, under general agency principles. Specifically, control by a parent company over its subsidiary, even if excessive, should not be sufficient to constitute implied authority from the parent company generally. The article concludes with Section V and discusses whether and how the proposed agency test interacts and fits in with other principles, such as piercing the corporate veil and direct liability. The authors believe that there are circumstances where agency is the ideal principle to hold a parent company liable. Cases with express agency agreements or clear representation by the parent company on the apparent authority of its subsidiary are prime examples. On the other hand, agency should not stretch to fill in the shoes left behind by the narrowing piercing doctrine. Contrary to what has been indicated by the Supreme Court of the United Kingdom, cases involving fraud are best dealt with by piercing the corporate veil. To do otherwise would merely substitute one metaphor for another.

II. Background of Agency in a Corporate Group Context

Any discussion that relates to parent-subsidiary liabilities must begin with Salomon v. Salomon & Co. In 1897 when Salomon was decided, agency was a long-established discipline of the law, while company law was still at its infancy. To apply agency as a possible exception to the separate legal personality laid down in the case is, therefore, both natural and understandable. In fact, the House of Lords spent much time in Salomon discussing whether Mr. Salomon’s one-man company was an agent acting on his behalf, thereby making him liable to the creditors. In the past, agency failed to pick up as much traction or attention as piercing the corporate veil. This, as we will see, is no longer the case.

A. Salomon v. Salomon & Co.

In Salomon, Mr. Salomon transferred his successful shoe manufacturing business to a joint stock company consisting of himself and his family. It was undisputed that Mr. Salomon’s wife and children, as well as the other shareholders, were merely nominees of Mr. Salomon. The board of directors also consisted of just himself and two of his children. In short, Mr. Salomon had complete control over the company in question. It came as no surprise, therefore, that, when the company went under, the creditors went after Mr. Salomon.

At first instance, Justice Vaughan Williams held in favor of the creditors based on a finding that the company was an agent of Mr. Salomon. Justice Vaughan Williams held that “[the creditors’] right to [sue Mr. Salomon] would depend on . . . whether the company was a mere alias of the founder or not. In this case it is clear that the relationship of principal and agent existed between Mr. Salomon and the company.” On appeal, Lord Lindley, an expert on commercial law, affirmed that Mr. Salomon should be held liable. He clearly affirmed the general agency principle, stating that “[a] person may carry on business as a principal and incur debts and liabilities as such, and yet be entitled to be indemnified against those debts and liabilities by a person for whose benefit he carries on the business.” But Lord Lindley did not think that application of the agency principle in the corporate context was proper. He felt “a difficulty in saying that the company did not carry-on business as a principal” but did not elaborate further, simply pointing to the fact that “the company must be recognised as a corporation.” In the end, Lord Lindley saw impropriety in Mr. Salomon’s setting up of a one-man company and decided the case against him, likening the company to “a trustee improperly brought into existence by [Mr. Salomon] to enable him to do what the statute prohibits.”

The House of Lords also rejected the agency argument. Lord Halsbury famously stated that “either the limited company was a legal entity or it was not. If it was, the business belonged to it and not to Mr. Salomon. If it was not, there was no person and no thing to be an agent at all.” This was not altered, and should not have been altered, by the fact that the company was under the sole control of Mr. Salomon. Lord Herschell elaborated on this point in his judgment:

As little am I able to adopt the view that the company was the agent of Salomon to carry on his business for him. In a popular sense, a company may in every case be said to carry on business for and on behalf of its share-holders; but this certainly does not in point of law constitute the relation of principal and agent between them or render the shareholders liable to indemnify the company against the debts which it incurs. Here, it is true, Salomon owned all the shares except six, so that if the business were profitable he would be entitled, substantially, to the whole of the profits. The other shareholders, too, are said to have been “dummies,” the nominees of Salomon. But when once it is conceded that they were individual members of the company distinct from Salomon, and sufficiently so to bring into existence in conjunction with him a validly constituted corporation, I am unable to see how the facts to which I have just referred can affect the legal position of the company, or give it rights as against its members which it would not otherwise possess.

Accordingly, the combination of the following factors cannot constitute agency between a company and its shareholders: (i) carrying on business by the company on behalf of shareholders, (ii) legal ownership of essentially all shares by one particular shareholder, (iii) the entitlement of all economic benefits of the company, and (iv) absolute control over decision-making of the company. The House of Lords did not, however, rule out the possibility of finding agency between parents and their subsidiaries. It merely found no agency relationship existed despite Mr. Salomon’s full control over the company. This was a negative definition of agency, as it provided a clear example of what did not constitute agency in a corporate context. Unfortunately, the House of Lords did not put forth a positive test for agency to show under what circumstances one could establish an agency relationship between Mr. Salomon and his company. As a result, subsequent courts were left scrambling for an answer.

B. Express Agency

In 1921, Rainham Chemical Works Ltd. v. Belvedere Fish Guano Co. provided the House of Lords another opportunity to tackle the agency question. In this case, the company was deemed an agent of the shareholders and was, therefore, subject to tortious liability arising from an explosion. Lord Buckmaster confirmed that Salomon did not prohibit finding agency between shareholders and a company, stating that “it may be established by evidence that in its operations [a subsidiary] does not act on its own behalf as an independent trading unit, but simply for and on behalf of the people by whom it has been called into existence.” The Salomon creditors simply failed to prove this. Beyond that, Lord Buckmaster paved a positive pathway for those who seek to prove an agency relationship. He held that the language of one of the agreements between the shareholders and the company made “[the company] agents for the vendors.” The presence of the express agency agreement reconciles Rainham and Salomon because, in Rainham, the House of Lords was clear that the agency agreement did not arise out of the fact that the defendants “occupied the position of governing directors, [and had] control of the business of the company conferred on them by . . . the articles of the association.” The agency relationship arose purely out of the express agency agreement, which was wholly independent from any finding of shareholding or control over a business. In other words, the defendants would have been liable, even if they did not own any shares of the company, provided that there was a valid express agency agreement. The finding in Rainham that express agency agreements can form valid agency relationships reduced the uncertainty surrounding agency. The question for courts then became whether agency can be found in a corporate group context absent an express agreement. Cases that followed Rainham attempted to answer this question. These attempts can be summarized into three approaches, which we will now explore.

III. The Three Current Approaches

English courts have long attempted to establish a set of criteria for finding agency relationships in the corporate context. An established criteria is necessary for agency to serve any meaningful role in bettering parent-subsidiary liabilities. These attempts are reflected in three leading cases.

A. Smith, Stone & Knight Ltd. v. Lord Mayor of Birmingham

Smith, Stone & Knight Ltd. v. Lord Mayor of Birmingham marked an early attempt by the English courts to establish a test for agency in a corporate group context that does not involve express agency. In this case, the parent company sought to receive statutory compensation from the Birmingham government for the compulsory purchase of a piece of land held by the parent company’s subsidiary. The parent company argued that the subsidiary was a separate legal entity and did not act as its agent in holding the piece of land on its behalf. Justice Atkinson decided that the key to determining the existence of agency was to ask whether the subsidiary carried on the business of the parent company. He formulated a six-factor test to answer this:

  1. Were the profits treated as the profits of the parent?
  2. Were the persons conducting the business appointed by the parent company?
  3. Was the parent company the head and the brain of the trading venture?
  4. Did the parent company govern the venture and decide what should be done and what capital should be embarked on the venture?
  5. Did the parent company make the profits by its skill and direction?
  6. Was the parent company ineffectual and in constant control?

The six factors focused on the amount of control exercised by the parent over the operations of the subsidiary. The court concluded that agency existed between the companies, and, as such, the parent was entitled to compensation because the subsidiary was never really in operation and was merely treated by the parent company as one of its many departments. This “control” focused test was widely cited by subsequent cases, but it suffered from various problems.

First, Smith, Stone & Knight may be inconsistent with Salomon. As seen above in both Salomon and Rainham, control alone has been insufficient to establish an agency relationship. Thus, one may argue that the six factors point to a type of control which exceeds that of “ordinary control” and is more akin to “excessive control.” This is especially the case because the subsidiary never actually carried on its own business, as required by the first factor. It was, for all intents and purposes, nothing more than a shell company that never entered into operation. This day-to-day control of the company may arguably be a form of control which exceeds that discussed in Salomon. But one must not forget that Mr. Salomon also substantively had full control of the company as well. While there were other shareholders in his company, they were nothing more than mere nominees. This casts doubt over the six-step test’s assertion that control of a company’s day-to-day activities would suffice to establish agency, as this definitely was not the case in Salomon. There must be something more, in addition to control, or even excessive day-to-day control, for a finding of agency to be consistent with Salomon.

Second, it may be argued that some of the six factors go beyond control. For example, factors (1), (4), and (5) specifically focus on the profit making and financing of a company. This can be explained by the six-factor test’s origin. Justice Atkinson admittedly derived the test from a series of tax cases, particularly The Gramophone & Typewriter, Ltd. v. Stanley. The issue in these tax cases was whether profit made by the overseas subsidiaries of English parent companies would be subject to taxation in England. It was determined that the appropriate test for taxation purposes was whether the subsidiary was “carrying on the business” of its English parents. In fact, the language of “carrying on the business” in Justice Atkinson’s test seems to be derived from the relevant tax statute. Accordingly, there is a concern that this financial-heavy focus was a result of the taxation-specific context and may not be applicable to general agency in a corporate context. This is consistent with the fact that Justice Cozens-Hardy’s opinion in Gramophone clearly confined the test to taxation and taxation alone:

I do not doubt that a person in that position may cause such an arrangement to be entered into between himself and the company as will suffice to constitute the company his agent for the purpose of carrying on the business, and thereupon the business will become, for all taxing purposes, his business.

Therefore, Justice Atkinson’s direct application of the six-factor test necessarily extends its initial intended applicability. Further, this extension to a context beyond taxation may be inappropriate as it inaccurately reflects its original intent. The inquiry of whether a foreign company should be subject to taxation by the nation where its parent company resides is markedly different from the inquiry of whether a private creditor shall be entitled to hold the parent liable for the acts of the subsidiary. The former inquiry is a public law question, while the latter is strictly a private law question. When a court is presented with a public law question in the context of a statute, it will naturally consider the public policies behind the legislation. For example, a question on foreign taxation may involve the consideration of comity. In fact, the same can also be said about the statutory setting of Smith, Stone & Knight itself. Since the issue was whether statutory compensation could be claimed by the parent as principal of the subsidiary, it is unclear whether Justice Atkinson intended for the test to apply beyond the ambits of Smith, Stone & Knight. It is also fair to ask if Justice Atkinson’s decision was influenced by his sympathy towards the parent company in this particular statutory context. After all, it could not have predicted that the piece of land held by the subsidiary would be under compulsory purchase by the government when the parent placed the land in the hands of the subsidiary. The same sympathy could not be found nor extended to the creditors in Salomon. As Lord Macnaghten put it, “they [had] only themselves to blame for their misfortunes.” Salomon’s creditors should have sought securities from the company when they entered into the transaction, having known that they were dealing with a limited company. There was a self-help possibility in Salomon but not Smith, Stone & Knight.

Third, this agency test has widely been regarded as an alternative, or even akin, to piercing the corporate veil. Even if it is assumed that it is a form of piercing, which the authors disagree, its theoretical basis is comparatively far less solid. Traditionally, it is third parties who wish to pierce the veil, seeking to make parents liable for their subsidiaries’ actions. But this six-factor test was borne out of a context which reverses the status quo. Here, the parent wished to bypass the asset partition between the parent and the subsidiary in order to obtain statutory compensation. It is this reversal that attracted much criticism from the Land Tribunal in the subsequent case of Woolfson & Another v. Glasgow Corp.:

[I]n the absence of a situation of principal and agent, for a court of law to entertain a plea of that kind from a company itself or from its shareholders or owners would in the ordinary way be a legal novelty; for such a proposition would seem to represent a fundamental attack upon the basic principles of incorporation of a limited company, namely that a company’s memorandum of association binds the company in its relations with third parties; that it contains the fundamental conditions upon which alone a company is allowed to be incorporated; that these conditions are introduced for the benefit of the creditors and the outside public as well as the shareholders; and that anyone whether a shareholder or an outsider who has dealings with a registered company must be taken to have notice of the memorandum.

It is one thing for third parties to look behind the doctrine of separate legal entity to obtain a just outcome. But it is entirely different for a company itself to assert that it should be able to go around the doctrine to obtain an advantage. Essentially, this would allow companies to have their cake and eat it too, enjoying limited liability while also having the flexibility to bypass any obstacles potentially presented by a separate legal personality. This could very well work against the entire rationale of piercing, which is to cure injustices that the doctrine of separate legal personality creates.

As such, the six-factor test of Smith, Stone & Knight cannot be the test for agency. It is a test developed for a very particular purpose, and it risks providing a loophole for companies to bypass the clear prohibition of contravening separate legal personality simply by having substantial control over the subsidiary.

B. Adams v. Cape Industries Plc

In 1990, the Court of Appeal took a deep dive into the relationships within corporate groups in Adams v. Cape Industries Plc. Cape Industries was an English company that operated an asbestos factory in Texas. When employees in the United States got sick from exposure to asbestos, they sued Cape and obtained judgment from a Texas court. In the English court, the issue was whether the Texas judgment could be enforced in England. Enforceability depended on, among other things, whether Cape was “present” in the United States according to the English private international law rule. The Court of Appeal decided that this presence could be established through a subsidiary either by piercing the corporate veil or agency.

On agency, Justice Slade formulated a new test: whether “a representative of the overseas corporation has for more than a minimal period of time been carrying on the overseas corporation’s business in the other country at or from some fixed place of business.” Although the test, like the one in Smith, Stone & Knight, determines whether the principal’s business was being carried out, it is different in two major ways. First, due to the nature of the case and its ties with private international law, the test places heavy emphasis on the need for a fixed place of business, so that it would satisfy the territorial basis requirement of the “presence” rule. Second, a fortiori of the territorial basis requirement, the court picked out very different factors to form the elements of the test for agency:

(a) whether or not the fixed place of business from which the representative operates was originally acquired for the purpose of enabling him to act on behalf of the overseas corporation; (b) whether the overseas corporation has directly reimbursed him for (i) the cost of his accommodation at the fixed place of business; (ii) the cost of his staff; (c) what other contributions, if any, the overseas corporation makes to the financing of the business carried on by the representative; (d) whether the representative is remunerated by reference to transactions, e.g. by commission, or by fixed regular payments or in some other way; (e) what degree of control the overseas corporation exercises over the running of the business conducted by the representative; (f) whether the representative reserves (i) part of his accommodation, (ii) part of his staff for conducting business related to the overseas corporation; (g) whether the representative displays the overseas corporation’s name at his premises or on his stationery, and if so, whether he does so in such a way as to indicate that he is a representative of the overseas corporation; (h) what business, if any, the representative transacts as principal exclusively on his own behalf; (i) whether the representative makes contracts with customers or other third parties in the name of the overseas corporation, or otherwise in such manner as to bind it; (j) if so, whether the representative requires specific authority in advance before binding the overseas corporation to contractual obligations.

There are certainly overlaps between the Adams and Smith, Stone & Knight tests. The degree of control remains a relevant factor (factor (e) in particular), as do the financial arrangements of the company (factors (b), (c), and (d)). But there is one significant factor in the Adams test that had not previously appeared: authority of the principal (factors (g), (h), (i), and (j)). The Court of Appeal noted that the list of factors were not exhaustive and that even the lack of authority would not (in all circumstances) mean that there could not be an agency relationship. Every case must individually examine all relevant circumstances. Nonetheless, the court highlighted that the general principle stated by Lord Justice Pearson in F. & K. Jabbour v. Custodian of Israeli Absentee Prop. stands, namely that, “in the case of an agency, the principal test to be applied in determining whether the corporation is carrying on business at the agency is to ascertain whether the agent has authority to enter into contracts on behalf of the corporation without submitting them to the corporation for approval.”

In Adams, Justice Slade added that:

On the authorities, the presence or absence of such authority is clearly regarded as being of great importance one way or the other. A fortiori, the fact that a representative, whether with or without prior approval, never makes contracts in the name of the overseas corporation or otherwise in such manner as to bind it must be a powerful factor pointing against the presence of the overseas corporation.

This added element of authority improves upon the Smith, Stone & Knight test. It was mentioned earlier that mere control cannot establish an agency relationship, as this would be irreconcilable with the result in Salomon. Authority seems to provide that additional element beyond control that could justify establishing an agency relationship, especially in those “one man company” cases like Salomon. Not only has authority always been an essential feature of traditional agency law, but it incorporates into the test the ability to determine whether the subsidiary has “permission” to represent the principal to the outside world on the specific transaction in question, which is something control alone cannot detect or quantify. The addition was a step in the right direction, but this test is not unflawed. In fact, the Court of Appeal’s new agency test (being more recent and created by a higher court) is rarely cited or applied by subsequent courts. The reasons are threefold.

First, much like Smith, Stone & Knight, the test in Adams arose out of a specific statutory context, this time to do with the enforcement of a foreign judgment. In English private international law, it is conventionally thought that the English courts are under an obligation to give effect to a foreign judgment that is rendered on a jurisdictional basis, if acceptable to English law. One of such acceptable bases is the presence of the defendant within the territory of the nation that renders the judgment. The importance of territoriality is evidenced by the emphasis of the test on the fixed place of business. In addition, the reference to “carrying on business” in the test, like Smith, Stone & Knight, came from statutory language instead of general agency law. This again raises concerns as to whether this test is appropriate for agency in general or whether it was designed specifically to deal with a specified situation. In some of the cases Justice Slade used to derive the test, the existence of agency was not even disputed. What was disputed in those cases was whether the presence of agency was sufficient to constitute territorial presence. Foreign companies would not be subject to a country’s jurisdiction if they carry out their business through an agent instead of by an agent. It is, therefore, doubtful whether it is a test to determine the existence of an agency relationship, given that, in many cases, the relationship was assumed or accepted for the true purpose of determining presence.

Second, on top of the specific context, the wording of the Adams test limits its application to overseas corporate groups, presenting even more evidence in favour of a limited test instead of a generally applicable one. Enforcement of foreign judgments would clearly trigger international considerations, such as international comity, which has no place or relevance in issues revolving around the domestic commercial setting between private parties.

Third, authority is a step in the right direction, but the fact that company law is now only partially aligning with traditional agency law seems irrelevant. While authority is an essential condition under agency law, the Court of Appeal merely labeled authority as an important but non-essential factor without providing much reasoning as to why there is a need to depart from traditional agency law principles. Presumably, this is due to the application of agency to a corporate context, but no justification or reasoning is provided to this effect. This is again due to the fact that the test stemmed from a case irrelevant to traditional agency law. The disparity, put on display in Adams, showed a clear need to bridge company law and agency law, or at least justify any necessary differences—something that the courts had not done in the past.

C. Yukong Line Ltd. of Korea v. Rendsburg Investment Corp. of Liberia & others (No.2) (The Rialto)

In J. H. Rayner (Mincing Lane) Ltd. v. Dept. of Trade and Indus., the House of Lords rejected Lord Sumption’s agency argument. Two significant points were borne out of the case. First, it confirmed the House of Lords’ long-held view since Salomon that a high level of control cannot alone constitute agency. In line with Lord Herschell’s judgment quoted above, the court reiterated that “whether a corporation acts directly on the instructions of its members, who constitute the directorate, or indirectly because of the members’ control in general meeting, makes no difference in principle.” This is another a negative definition. The more important point is the court’s direct reference to authority as the pathway to agency, which is not limited only to express authority, as in Rainham. It states that “[t]he rights of creditors against the members, if any, depend solely on the creation between the members and the [company] of the rights and duties which, in domestic law, are created by the authority which, as a matter of law, is conferred on the [company].” The court did not elaborate on how to establish such authority, but this renewed emphasis on authority is further developed by this next case.

In Yukong Line Ltd. of Korea v. Rendsburg Inv. Corp. of Liberia & Others (No. 2) (The Rialto), a shipowner brought an action for wrongful repudiation by a company. The repudiation was due to the market falling and the company’s inability to renegotiate the charter price. It was also revealed that the defendant company had transferred funds to another company on the day of the repudiation. An agency argument was advanced in an attempt to hold the director of the two companies liable, claiming that the repudiating companies contracted and acted as the agent for the director. Justice Toulson made two points when addressing the agency issue. First, he criticized Smith, Stone & Knight. Second, he stated that the traditional test for general agency law should be applied despite the corporate background of the case.

On Smith, Stone & Knight, Justice Toulson disagreed that the test of agency should depend on whether a company was carrying on its own business or that of its owner. This was mainly due to the inconsistency with Salomon and the different results they would lead to. In his opinion:

[I]f the company was a legal entity independent of its members, it followed that the business belonged to it and not to Mr. Salomon. It was nothing to the point that it acted on the direction of Mr. Salomon and for his benefit. Something quite different would need to be established in order to show that the company, in law an entity independent of its owner, was acting in respect as agent for its owner, the necessary requirement being to show that the relationship of agency was intended to be created.

He correctly pointed out that following the six-factor test of Smith, Stone & Knight meant “Salomon’s case would surely have been decided differently.”

Justice Toulson, therefore, believed the correct test must be the general test of agency formulated by Lord Pearson in Garnac Grain Co. Inc. v. H.M.F. Faure & Fairclough Ltd:

The relationship of principal and agent can only be established by the consent of the principal and the agent. They will be held to have consented if they have agreed to what amounts in law to such a relationship, even if they do not recognise it themselves and even if they have professed to disclaim it . . . . But the consent must have been given by each of them, either expressly or by implication from their words and conduct.

According to Justice Toulson, this principle applies irrespective of the corporate context some of these cases were borne from. In particular, he lent support from Atlas Maritime Co. S.A. v. Avalon Maritime Ltd. (The Coral Rose), where Lord Justice Staughton stated:

[a] . . . subsidiary . . . which will operate with the parent’s funds and on the parent’s directions but not expose the parent to liability, may not seem to some the most honest way of trading. But it is extremely common . . . [t]o hold that it creates an agency relationship between the subsidiary and the parent would be revolutionary doctrine.

As mentioned above, Adams brought the idea of authority into the mix, but it never made it a strict requirement. The general test of agency clearly has a higher threshold, requiring intent between principal and subsidiary to form such a relationship. Any such intention would essentially grant authority to the agent to act on the principal’s behalf. These observations are clearly correct and were cited by a number of courts.

But Justice Toulson did not expand on how this test should apply in a corporate context. Unlike Smith, Stone & Knight and Adams’s clear attempts to customize and tailor an agency test for a corporate situation, Justice Toulson imply cited Lord Pearson’s general test for agency and, in a rather brief manner, concluded that agency was not found in the present circumstances, as no intention to conclude agency was displayed. This lack of detail has made it difficult for subsequent courts to follow the test.

For example, in The Government of Sierra Leone v. Edward Ormus Sharincton Davenport & Others, the court adopted the reasoning of Yukong and made Lord Pearson’s formulation in Garnac Grain Co. Inc. v. H.M.F. Faure & Fairclough Ltd the test. However, without an express agency agreement, the court struggled to distinguish control from authority in determining whether there was implied authority in the case. It sought to rely on the unreported case of Hyundai Engineering and Construction Co Ltd v. HRH Jefri Bolkiah. In that case, it was found that ownership and control of the company by a shareholder was “relevant but not by themselves sufficient.” That is correct, but what more would it take to infer agency? The court in Hyundai simply referred to the “the pleaded facts that [the company] neither owned the land on which the marina was being built [which was the subject matter of the contract] nor had any other commercial interest in the performance of the contract.” Reasoning like that went full circle back to Smith, Stone & Knight, which the court emphatically rejected itself. Recall that the lack of operation of the subsidiary in Smith, Stone & Knight was insufficient to find agency. At best, the absence of activities or other commercial purpose is indicative of a high level of control, beyond that ordinarily exercised by an average shareholder. Unsurprisingly, the court ended up deciding the case on essentially the same vague reasoning. Like Hyundai and Smith, Stone & Knight, it pointed to the lack of activities by the company, as well as certain statements made by the shareholder. Similarly, in Abbar v. Saudi Economic & Development Co (SEDCO) Real Estate Ltd, the court denied finding agency purely based on control, relying on the same reasoning in Yukong, but it failed to elaborate what would constitute “the essential element of consent to the relationship of principal and agent which is necessary to a finding that the relationship has been established,” which the court found lacking in the case. These cases show that, without more concrete guidance, Yukong has not been able to replace Smith, Stone & Knight to become the authoritative and sole test for agency in a corporate context.

Another potential issue with Justice Toulson’s finding in Yukong is his opinion on one-man companies. Stressing the importance of intention in forming an agency relationship, he said that “ordinarily, the intention of someone who conducts trading activities through the vehicle of a one-man company will be quite the opposite.” This suggestion that one-man companies cannot ordinarily form the requisite intent poses two questions. First, whether forming an agency relationship requires intention, and, if so, second, whether one-man companies can ever be party to an agency relationship. Justice Toulson did not elaborate on the correct method of applying the agency test of Lord Pearson. But we may be able to find some answers in Lord Pearson’s words. While it was indeed said that agency must be formed by an agreement, it was also stated that such an agreement can arise even if “[the parties] do not recognise it themselves and even if they have professed to disclaim it.” In other words, this agreement does not necessarily have to be an express one but can be an implied one, inferred from all relevant circumstances. Hence, while the general purpose of a one-man company may be to separate liability between the company and the individual, there can be cases where there is sufficient evidence pointing to the fact that a one-man company may have been acting on their principal’s behalf, even if there was never a clearly written agreement to evidence such intention. The answer must be that a one-man company can be party to an agency relationship. Because agency can be implied, a one-man company does not need to have its own “capacity” to agree or form an intent. It is trite law that “all persons including minors and other persons with limited or no capacity to contract on their own behalf, but excluding the profoundly insane, are competent to act or contract as agents.” In addition, as an English scholar put it, despite Lord Pearson’s speech, “[c]onsent does not lie at the heart of every agency.” Otherwise, it would be difficult to explain how apparent authority can constitute agency, given that the concept stems from estoppel, that is, a relationship between the third party and the principal but not the one between the principal and the agent. It is, therefore, argued that a one-man company should be no different; they too can perform the functions of an agent. Thus, our principle of agency must be able to cover these situations, instead of categorically excluding them simply because an express agreement between the principal and its one-man company cannot be found. This conclusion is consistent with case law over the years, in particular Prest, where Lord Sumption found that a one-man company in Trustor AB v. Small Bone (No. 2) was an agent. Lord Neuberger in the same case also labelled the “one-man company” in Gilford Motor v. Horne as an agent to the controller, even though he was neither a shareholder nor a director of the company. Given the above, it is doubtful that an agreement or express intent must be shown in order to form an agency relationship, and, therefore, it is possible for a one-man company to act as, or make use of, an agent.

Accordingly, Yukong laid an important foundation for our test. There is no question that using agency law as the blueprint for agency in a corporate context was long overdue. But, in doing so, Justice Toulson still left plenty of questions unanswered. These issues will be clarified by providing a revamped test below.

IV. The Proposed Test

It is our opinion that none of the three approaches is satisfactory, but we believe that each test offers some important considerations. By combining the teachings in these tests together, we will attempt to propose a new test for agency in the corporate context. We will first summarize the key issues under English law and then examine U.S. authorities to derive a proposed test that works on a practical level.

A. The Current State of English Law

In the spirit of Justice Toulson in Yukong, the starting point of the test must be the principles of agency law. This is most clearly set out in Bowstead & Reynolds on Agency, the classic English work on agency:

(1) The relationship of principal and agent may be constituted—

a. by the conferring of authority by the principal on the agent, which may be express, or implied from the conduct or situation of the parties, and may or may not involve a contract between them;

b. retrospectively, by subsequent ratification by the principal of acts done on the principal’s behalf.

(2) A person may be liable under the doctrine of apparent authority in respect of another who is not that person’s agent at all; or may be estopped as against a third party from denying the existence of an agency relationship.

The significance of authority can clearly be seen in (1)(a) above. This was captured in Adams as being important and later in Yukong as being essential. To fully return to agency law, it is necessary to examine whether the subsidiary has authority, actual or apparent, according to traditional agency law principles. The editors of Bowstead & Reynolds also agree that when companies are agents for each other or for shareholders, authority “would be required to be proved by normal [agency] criteria.”

According to general agency law, an express agency agreement would undoubtedly form an agency relationship despite the corporate group setting, as illustrated in Rainham. The same can be said for any sort of express ratification by the principal after the subsidiary has carried out some act on behalf of the principal, as stated in (1)(b). No written agreement is necessary either way. In fact, The Government of Sierra Leone v. Edward Ormus Sharincton Davenport & Others may be argued as a case on express authority despite the lack of an express agency agreement. The court found that the shareholder of the company conceded the existence of the agency relationship between himself and the company, which could be seen as a ratification rather than implied actual authority. The difficulty lies in the less clear-cut cases, namely those agency situations founded on implied authority under (1)(a) or apparent authority under (2). Arguably, the test for agency must be able to clearly decide these two difficult scenarios in order to be successful.

Salomon provides a good set of facts to illustrate both the simplicities and difficulties the general agency test presents, even though authority was never examined by the Salomon court. But, if we assume for the purpose of illustration that Mr. Salomon entered into an express agency agreement with his company to transact with the creditors on his behalf, there is no question that the creditors could sue Mr. Salomon as the principal of his company. While this is straightforward, apparent authority is more difficult. It refers to situations where the principal (individual shareholder or parent company) has made representations to the counterparty which suggest that an agent (subsidiary) has authority to represent the principal and the reliance of the representation by the counterparty is reasonable. Even though authority may not have actually been conferred, the principal would still be bound in this case. Say, for instance, Mr. Salomon represented to the creditors that they were contracting with him, when, in fact, they were contracting with the company. The House of Lords should have no hesitation in finding agency in that case as well. Although Mr. Salomon had no intention to grant authority to the company to contract on his behalf, his misrepresentation had the effect of conveying this message to the third-party creditors, thereby satisfying the test for apparent authority. No such situation existed in Salomon, although, as Lord Macnaghten remarked:

[T]he unsecured creditors . . . have only themselves to blame for their misfortunes. They trusted the company . . . because they had long dealt with Mr. Salomon and he had always paid his way; that they had fair notice that they were no longer dealing with an individual, and they must be taken to have been cognizant of the memorandum and . . . the articles of association.

On principle, the finding of apparent authority in a parent-subsidiary relationship is sound. The real difficulty lies in the lack of precedent in this regard under English law.

The most difficult question is under what circumstances implied authority can be established, particularly in situations where neither an express agency agreement nor apparent authority are present. Naturally, the court must focus on the facts and circumstances of the case to see if authority can be inferred from the conduct of parties. From our review of Smith, Stone & Knight, we can see that a huge factor that courts focus on in seeking to find implied authority is control. Thus, the biggest uncertainty that arises out of the English approaches is whether control exercised by the parent company over the subsidiary can constitute implied authority for an agency relationship. As it stands in English law, Salomon has decided from the outset that ordinary control would not suffice to create agency. Nor, after Smith, Stone & Knight and its reception, does excessive control suffice to create agency. This was clearly stated in Yukong. Attempts to link control and implied authority also failed in The Government of Sierra Leone v. Edward Ormus Sharincton Davenport & Others. Yet control seems to be the only factor that has consistently been present in most of the cases we have discussed, and, therefore, whether control plays an important role in the identification of implied authority appears to be the key issue.

This is where comparison with the U.S. system may come in handy. While U.S. and English agency law have their differences, the elements that we see running through the tests, as well as the cases, are not as different as one may think.

B. U.S. Law

Just like the English courts, the U.S. courts’ test for agency is inconsistent. Though far from offering a blueprint on the application of agency principles in a corporate group, U.S. law still offers many more precedents on such an application. In this section, we will look at five different components that we deem relevant to formulate our agency test. They will also address the outstanding issues seen in English law above.

1. Applicability of Agency in a Corporate Group Context

The notable aspect about U.S. law is that agency can be applied in a corporate group to make a parent company liable for its subsidiary. Many states have adopted § 14M of the Restatement (Second) on Agency. This section recognizes that a corporation can act as an agent for someone else, just like an individual. The rationale is simple. As Holmes stated more than a century ago, “it is plain good sense to hold people answerable for wrongs which they have intentionally brought to pass, and to recognize that it is just as possible to bring wrongs to pass through free human agents as through slaves, animals, or natural forces.” The same must be true when the wrong is to pass through a company. The most famous case to actually apply this principle in a corporate context is Berkey v. Third Avenue Railway Co. Justice Cardozo was of the opinion that there could be situations where “dominion [over a subsidiary] may be so complete, interference so obtrusive, that by the general rule of agency the parent will be a principal and the subsidiary an agent.” Empirical research shows that U.S. courts have consistently applied agency to determine the liability of the parent company in many circumstances. For example, in a 1991 article, Robert B. Thompson found that the U.S. courts applied agency as a basis for corporate group liability in fifty-two cases. In a subsequent survey, Peter Oh found 152 such cases. The wide application of agency to parent-subsidiary liability in the U.S. confirms the appropriateness of such application in England.

2. Distinction Between Piercing the Corporate Veil and Agency

Much like in English law, there has been some degree of confusion in terms of whether agency is separate and distinct from the doctrine of piercing the corporate veil in U.S. law. While the authors are of a firm opinion that they are entirely separate bases for finding liability, this has been a debate in U.S. law for a long time. An example of U.S. authorities confusing, or even combining, piercing the corporate veil with agency can be found in Kashfi v. Phibro-Salomon, Inc., where the court held that “the test determining whether a corporation is acting as an agent for a related corporation is the same as the test imposed under the doctrine of piercing the corporate veil.” This would require the plaintiff to show that “the parent dominated the subsidiary, and that such domination was used to commit a wrong or fraud against the plaintiff.” But this cannot be correct, as there must be a difference between imposing liability on the basis of legally identifying a relationship between the parties (agency), and imposing liability just by identifying a policy reason for it (piercing the corporate veil). While it is possible for there to be cases that satisfy both tests, one cannot overlook the reasoning and journey taken in order to reach the destination. In other words, there must be a significant difference in the court’s reasoning when holding that parties are acting under a relationship of agency and when the court holds that the veil of a company must be pierced. Still, the confusion by courts is well noted. Phillip Blumberg observed that the reasoning applied by courts when applying agency have “been stated in terms virtually identical to those of the common identity or ‘alter ego’ alternative for piercing the veil.”

The correct method to approach it seems to be the approach taken by the Southern District Court of New York in Royal Industries Ltd. v. Kraft Foods Inc., where the court decided that the agency rule “will not undermine jealously safeguarded notions of corporate separateness.” The theory behind any such agency claim is that the subsidiary’s acts were, in both form and substance, those of the parent company. Thus, “there is no veil to pierce—the parent is the only party in interest.” As such, it seems that the intention behind the agency test was not to merge it with piercing the corporate veil. Instead, it is just the application that may have, at times, done so unintentionally. The piercing version of “agency” is, therefore, not relevant for our discussion, and any related precedents will not be discussed below. On the other hand, there is another line of authorities in which the U.S. courts apply traditional agency principles in a corporate group context. This is sometimes called “pure agency law.” It will form the focus of our comparison.

3. Consensus as to the Essential Role of Authority in the Establishment of Agency

§ 14M, however, does not expressly state what will constitute such agency in the context of a corporate group. Instead, § 14M cites three cases in this regard, one of them dealing with actual authority and the other two with apparent authority. The necessary presence and essential nature of authority has been recently elaborated upon in Royal Industries Ltd. v. Kraft Foods:

The parent-principal should not be allowed to escape liability simply because it owns stock in the subsidiary-agent. Rather, as in any agency case, the issue should be one of authority: did the subsidiary have authority, actual or apparent, to act on behalf of the parent? As another court in this district has put it, “a parent corporation may become a party to its subsidiary’s contract under an agency theory if the parent’s conduct manifests an intent to be bound by the contract.”

Recently, the Delaware Bankruptcy Court in In re Washington Mut., Inc. set out in more detail the pathways to establish agency:

Agency can be established if the subsidiary had actual or apparent authority to act for its parent. Actual agency is created by written or spoken or other conduct of the principal which, reasonably interpreted, causes the agent to believe that the principal desires him so to act on the principal’s account. Actual agency depends upon the actual interactions of the putative agent and principal and not on the perception a third party may have of the relationship. Implied agency, by contrast, depends upon the reasonable conclusion of a third party, based upon the actions of the principal, that the agent has the authority to bind the principal. Agency is the relationship that results from the manifestation of consent by one person to another that the other shall act on his behalf and subject to his control, and consent by the other so to act. The principal’s consent may be manifested by written or spoken words or any other conduct of the principal directed at the agent in the case of actual authority, or at a third person in the case of apparent authority.

It is, therefore, clear that the core of agency relies on there being authority granted from the principal to the agent to act on the principal’s behalf. Without such authority, whether actual or apparent, there cannot be a relationship of agency. Courts have found agency through actual express authority and apparent authority in several cases, but implied authority presents a bigger challenge.

a. Express Authority

Like in England, express authority presents fewer problems for U.S courts. In re Washington Mut., Inc. was a case decided on express authority. In this case, the court relied on the express wording that the company entered into the disputed agreement “for itself and on behalf of” the parent company. This constituted an express agency agreement. In fact, multiple cases have made it clear that, if such an express agreement is present, it does not matter whether there are shareholders or control between the parent and the subsidiary. The U.S. experience here undoubtedly strengthens the belief that there should be a clear application of agency where there is express authority. The problem of express authority is that it is rarely found in cases. This is not surprising, given that a formal express agency agreement can be perceived as unnecessary between a parent company and its subsidiary when the former exercises substantial control over the latter.

b. Apparent Authority

It is mentioned above that English law should find agency in corporate groups through apparent authority, but there is a lack of established precedent. Precedents in U.S. law show how this can be done, even though successful cases are still rare. The best example is again found in In Re Washington Mut., Inc., where the court held that there was apparent authority to establish an alternative ground for establishing agency between the parent and the subsidiary. The court found that it was reasonable for the plaintiff to infer that the subsidiary had apparent authority to enter into the agreement in question because the parent company’s president (who also happened to an officer of the subsidiary) executed the agreement. The agreement also included a specific provision that purported to bind the parent company to use of one of its trademarks. In fact, U.S. courts have found apparent authority between companies even when there was no shareholding between them. For example, in Pullman v. Alpha Media Pub., Inc., using the theory of apparent authority, the court found that a company was liable for the Ponzi scheme set up by another company because the former licensed the use of its trademark. The licensee put the mark in the marketing materials used in the scheme, which were then distributed to the plaintiff. The court concluded that it was reasonable for the plaintiff to rely on these representations. But this does not mean that finding apparent authority is easy. Apparent authority requires the principal to have made a representation to the third party, who reasonably relied on such representation, and agency cannot be established if the parent company has an established relationship with the third party, or when such reliance is absent. In the end, taking relevant circumstances into account, it is a matter for the court to decide whether the representation in question suffices to establish the requisite apparent authority. Moreover, it is a matter of principle that agency between a parent and its subsidiary can be established through apparent authority. This brings us to the most crucial question, which has been troubling the English courts for decades: whether control can constitute implied authority in the absence of an express agency agreement.

4. The Unclear Relationship Between Authority and Control

U.S. courts have not been clear in their answer to this question. What can be confidently said is that control plays a huge part in U.S. agency law. Under most formulations, control is a standalone factor, in addition to authority. In §1.01 of the Restatement (Third) of Agency, the definition of agency requires control. Agency can only be present where “one person (a ‘principal’) manifests assent to another person (an ‘agent’) that the agent shall act on the principal’s behalf and subject to the principal’s control, and the agent manifests assent or otherwise counsels so to act.” Therefore, many courts formulated their test with both authority and control as the essential components. A classic example is the following three-element test adopted by the Eighth Circuit: “Agency is a legal concept that depends upon the existence of certain factual elements: (1) the manifestation by the principal that the agent shall act for him; (2) the agent’s acceptance of the undertaking; and (3) the understanding of the parties that the principal is to be in control of the undertaking.” The first two elements deal with authority, while the third deals with control. But why control is an essential element is never quite explained properly nor do courts explain the connection between authority and control.

a. Why Is Control Important in the First Place?

As shown in the formulation of the Restatement (Third) of Agency above, in U.S. agency law, control is an essential element generally, not just in the context of a corporate group. As the court in Mouawad Nat. Co. v. Lazare Kaplan Int’l Inc. explained: “the principal’s power to control the agent is an essential element of an agency relationship,” and “the essence of control in an agency sense is in the necessity of the consent of the principal on a given matter.” This seems to suggest that the presence of control implies, or at least suggests, that whatever the subsidiary does must have had some sort of approval or consent from the parent, given the domination exercised by the parent over the subsidiary. In other words, authority could be implied from the control that a parent company exercises over its subsidiary, given that they must have a subsidiary to act on their behalf and within the ambits of a certain scope of authorisation. But, logically speaking, the presence of control does not necessarily mean that there is authority. This is particularly the case in the parent-subsidiary relationship because the parent will always exercise a certain level of control through shareholding. But this does not necessarily mean that it has granted authority to the subsidiary to do all things under the sun.

Another potential reason for control being an essential factor in the U.S. is to distinguish between relationships of agency from other relationships that also involve “one [acting] for the benefit of another.” These could include the likes of “executors, guardians . . . receivers,” to name a few. This is because, in those other examples, “there is no control by the beneficiary.” This second justification seems to bear more weight, given that there can be situations where agency can exist despite the parent company not having total domination over the subsidiary. Whether control is present or not can be completely independent of whether the principal granted such authority, and, as such, the presence of control would serve more purpose if it were to identify whether the relationship between parties was of agency or another relationship where there was someone acting on behalf of a beneficiary, free from the beneficiary’s control. For example, in Patino v. Capital Bonding Corp., the presence of a significant amount of control was sufficient for the Delaware Court of Chancery to hold that there was an agency relationship. It was agreed, among other things, “that [the company] would transfer management authority to [the insurer] . . . that only [the insurer] could modify or terminate the contract, and [the company’s former director] was not allowed on [the company’s] premises without [the insurer’s] permission.” This was sufficient to establish agency despite the fact that the contract specified that the company acting for the insurers would be “independent contractors.”

b. How Do U.S. Courts Handle Control and Implied Authority in Practice?

It is difficult to find any U.S. case that directly addresses the relationship between the two concepts. In practice, however, when there is no actual (express) or apparent authority, many courts just bypass the authority analysis and make a conclusory finding on agency by emphasizing the element of control as the fundamental factor. For example, in Van Maanen v. Youth With a Mission-Bishop, a case that clearly had no express agency, the court avoided the discussion of the relationship between the two concepts by going straight to control, stating that “[t]he central inquiry under the [actual] agency theory of liability is whether the degree of control exerted over the subsidiary by the parent is enough to reasonably deem the subsidiary an agent of the parent under traditional agency principles.” The court went on to deny actual agency based on lack of control. This essentially equated control as implied authority. Similarly, in Coffey v. Fort Wayne Pools, Inc., it was decided that implied actual authority could not be found because the plaintiffs failed to prove that the defendant had a right to control the contractor’s installation of the pools. This not only shows not that control can sometimes be proven simply by looking at one small part of a business or even one particular act, but also shows just how decisive the presence of control can be on whether there was authority, especially implied authority.

There is, however, a lack of consensus on the requisite level of control. Some cases, such as Berkey above, require the highest level of control to justify the finding of agency. On the other hand, some courts, such as the Third Circuit, consider that the control need not reach total domination of the subsidiary by the parent: “Under this [agency] theory, total domination or general alter ego criteria need not be proven.” Some courts simply provide a laundry list of factors that will assist the court. For example, “[t]o determine whether a sufficient degree of control exists, courts review the extent of overlap of officers and directors, methods of financing, the division of responsibility for day-to-day management, and the process by which each corporation obtains its business.”

Others focus instead on the method by which the principal exerts control over the subsidiary. In Kingston Dry Dock Co. v. Lake Champlain Transp. Co., the Second Circuit decided that “the test is . . . rather in the form than in the substance of the control; in whether it is exercised immediately, or by means of a board of directors and officers, left to their own initiative and responsibility in respect of each transaction as it arises.” This would suggest that the level of control is not what matters the most. Instead, it is whether the subsidiary is left to act on its own discretion or accord through its directorship or whether it is the parent who is exercising their control immediately. If the parent was exercising its control immediately, this would indicate that the subsidiary was nothing more than an agent acting on behalf of the parent.

Yet, despite these differences, one thing is clear: if implied actual authority can be based purely on control, this would suggest that control alone suffices to create an agency relationship, even when there is no express authority or apparent authority. This control-centric approach in agency has been criticized. Blumberg thought that the practice of courts finding agency based entirely on control was caused by an “oversimplification” by Justice Cardozo and relegates agency to nothing more than “single-factored piercing.” Instead, he argued that the focus should be on the consensual relationship between the parent and the subsidiary, which is the hallmark of authority for agency. But, despite providing plenty of examples of cases relying on the “misunderstood” control test, he did not cite any case that actually adopted an authority-based test. The authors managed to identify several such cases. In Mobil Oil Corp. v. Linear Films, Inc., the court clearly distinguished between control and authority as separate elements. It further stated that “[a] vital prerequisite to imposing liability based upon customary agency principles is finding a close connection between the relationship of the two corporations and the cause of action . . . . [T]he relevant question is whether the alleged principal . . . directed the specific actions of the alleged agent.” But this substituted test still sounds like a control test, with its reference to the principal directing the subsidiary. It is rare to find courts that elaborate on this standard’s relationship with control, which was not done in Mobil Oil. In that case, the court simply concluded that agency could not be found because the principal in the case was not incorporated at the time of the transaction. The only case that actually applied the standard ended up citing factors very similar to control. In Fuller v. Midland Credit Management Inc., after citing the Mobil Oil test, the court went on to find the requisite “close relationship” between the parent and the subsidiary based on the facts that (1) they are parent and subsidiary; (2) the parent devised collection policies and practices for the subsidiary; (3) they shared common officers and directors; (4) the subsidiary had no employees; (5) they shared offices; (6) the subsidiary was formed to insulate the parent from liability; and (7) representations by the parent appeared to have made the subsidiary act on its behalf. The first six factors constantly appear in piercing cases when it comes to assessment of control, while the last factor may be more suited for apparent authority.

Stephen Bainbridge uses the example of Walkovszky, an early case on parent-subsidiary liability, to further illustrate the shortcoming of a control-centric test:

Walkovszky involved a close corporation with a single dominant shareholder. Such a shareholder’s ability to elect the board of directors by definition gives him control of the corporation. In exercising that control, such a shareholder’s decisions naturally are based on his own best interests. The corporation thus has no interests other than the interests of the dominant shareholder. Hence, it begs the question to describe a corporation, as the Walkovszky court did, as a “dummy” for its individual shareholders who are in reality carrying on the business in their personal capacities for purely personal rather than corporate ends. The close corporation with a single dominant shareholder has no “corporate ends” separate from those of its owner. In apparent recognition of this fact, courts generally require plaintiff to show something more than mere control.

This is probably why many U.S. courts ended up requiring the additional presence of fraud or injustice to establish agency. However, as mentioned above, this brand of agency is made essentially the same as piercing. Judge Dore even thought that the use of agency in the parent-subsidiary context is not appropriate beyond express agency.

In the end, it seems that the relationship between control and authority is running in circles. The truth seems to be what Professor Powell found almost more than ninety years ago:

But it cannot be doubted that as a matter of common-sense deduction from the facts and under ordinary principles of agency, the subsidiary is the agent of the parent corporation. A similar degree of control as between two individuals would inevitably create a “factual agency” in law and fact. If, without mention of the corporate immunity of stockholders, it was left to a jury to determine whether or not the subsidiary were the servant of the parent corporation, can there be any doubt as to what their verdict would be?

Effectively the U.S. courts’ approach boils down to two methods of finding agency. As Justice Cardozo’s judgment has indicated, they require either a high standard of control or an additional element of fraud. The former is an agency principle based on control (barring express or apparent authority), as discussed above, and the latter is piercing the corporate veil.

5. Additional Emphasis on Scope of Agency

One may question whether there is any real significance in distinguishing agency and piercing. The answer is yes. This is mainly illustrated by the emphasis on the scope of the agency, which can serve as an additional element for the agency. Many courts, in addition to control and authority, highlight that the liability in question must fall into the scope of the agency.

The issue of the scope of the agent’s authority is closely related to the issue of consent to the agency. The question of whether an agent’s conduct was ratified by his principal is closely intertwined with the question of whether the agent’s conduct was within the scope of the agency.

Similarly, the Delaware Court of Chancery held that: “For liability to attach under customary agency, ‘an arrangement [must] exist between the two corporations so that one acts on behalf of the other and within usual agency principles, [and] the arrangement must be relevant to the plaintiff’s claim of wrongdoing.’”

The significance of scope is that the parent only grants authority to the subsidiary for a limited period or a particular task. Proving that the act of the subsidiary fell within the permitted scope would go a long way to show that they are acting within the authorized parameters set by the principal, that is, that they had authority to do so. It is also notable that, in terms of scope, agency can be differentiated from piercing by way of the time in which an evaluation of the state of affairs is made. While piercing is an ex-post evaluation of whether the transaction involved injustice sufficient to justify it, agency is an ex-ante evaluation of the existing relationship between the parties when the transaction occurred. In other words, agency merely reveals existing relationships as they are, instead of trying to find something that is not already there.

C. The Test

The U.S. cases confirm that authority is an essential component of any agency test for a corporate group. This is consistent with general agency law in both England and the United States. The U.S. experience of expressing actual authority and apparent authority confirms our understanding of their role in agency law in a corporate group context. Therefore, English courts should not hesitate to find agency if the evidence is clear that these bases of authority are present in a case.

It is also clear from the cases that agency can arise out of individual instances, and, in all cases, it is a defined relationship which already exists before an act is carried out by an agent. This is in stark contrast to piercing the corporate veil, where the courts carry out an ex-post analysis to see if justice would deem it necessary to “alter” the way a transaction had originally occurred. As such, another essential component is ensuring that the subsidiary’s acts as an agent arise from the very scope that the parent has so authorised. Only then would liability be attributed, based on an agency relationship that already existed at the time the act was carried out, instead of morphing the nature of a previous act into an agency relationship. It also guarantees that the subsidiary had authority from the principal on the particular matter. Accordingly, English courts should learn from the U.S. experience and pay attention to the scope of agency as well.

Last, but definitely not least, control raises the most issues. As U.S. agency law stands, control is an essential factor that must be present before agency can be found, even though there does not seem to be one agreed-upon justification in the cases or the literature as to why that is the case. England has never required control to be present in agency law generally, but control has somehow always found a way into the discussion. In any event, from the analysis above, it should be clear that U.S. law effectively allows the presence of excessive control to establish implied authority and agency. Cases with express agency relationships and apparent authority are far more straightforward. Hence, most cases that require a step-by-step analysis through this agency test would be ones where the courts need to infer a subsidiary’s authority out of a parent-subsidiary relationship. Looking at it in this way, it makes sense that excessive control would be highly persuasive, as this is in line with the requisite mindset that the principal is willing to provide much-needed authority to the agent to carry out a task for them, under parameters laid down by the principal.

But it must be noted that, whether one finds them pervasive or not, nothing in the U.S. cases enable us to overcome the hurdle we previously encountered in Smith, Stone & Knight. Control, or even excessive control, on its own, cannot establish agency as authoritatively dictated by the House of Lords in Salomon and subsequently confirmed in J. H. Rayner. That is probably why the editors of Bowstead take a rather vague position on control: “[t]he idea of control may also be relevant where it is contended that a company is an agent of its parent,” without specifying its exact role. If one agrees with Powell’s view that general agency law will regard control exercised by the parent over the subsidiary as constituting agency, one way to reconcile the conflicting views of general agency law and Salomon is that company law has changed agency law as far as separate legal personality is concerned. After all, a company is a statutory creature, and the House of Lords simply interpreted the statute such that control, no matter how dominating, cannot establish agency under English law. While control may still be highly relevant when coupled with some other circumstances such as previous dealings with third parties in order to find implied authority, it will never suffice on its own. It is not just a matter of the precedential weight of Salomon, but also a matter of general judicial view on the strictness of separate legal personality. Even pre-Prest, there were only a handful of cases that acknowledged the possibility of a parent being made liable for the acts of its subsidiary—whether through agency or piercing. Recent empirical research identified 213 such cases between 1885 and 2014. This contrasts starkly to the liberal use of piercing or agency in U.S. law. Thompson has called piercing the most litigated issue in U.S. corporate law, and the relative ease of piercing the corporate veil successfully under U.S. law is well documented. In a separate research conducted by one of us, there were 1,044 piercing cases between 2012 and 2014 alone. The determination the U.K. Supreme Court displayed in limiting the applicability of piercing makes it even more unlikely that it would fundamentally alter the course of more than a century old English company law by overruling Salomon in this regard. Lord Justice Kerr stated in the Court of Appeal decision in J. H. Rayner:

The crucial point on which the House of Lords overruled the Court of Appeal in [Salomon] was precisely the rejection of the doctrine that agency between a corporation and its members in relation to the corporation’s contracts can be inferred from the control exercisable by the members over the corporation or from the fact that the sole objective of the corporation’s contracts was to benefit the members. That rejection of the doctrine of agency to impugn the non-liability of the members for the acts of the corporation is the foundation of our modern company law.

With that, and considering the general agency test set out by Bowstead and Blumberg, the test we propose is as follows:

The agency relationship of principal (parent company) and agent (subsidiary company) may be constituted through:

1. Mutual consent to the existence of the agency relationship by both principal (parent company) and agent (subsidiary company). This can be displayed through:

(a) Express Authority:

a. By the conferring of authority by the parent company on the subsidiary, which may be express, or implied from the conduct or situation of the parties (control by the parent company alone may not imply an agency relationship), and may or may not involve a contract between them;

b. Retrospectively, by subsequent ratification by the parent company of acts done on the parent’s behalf; or

(b) Apparent Authority:

A parent company may be liable under the doctrine of apparent authority in respect of subsidiary who is not that person’s agent at all; or may be estopped as against a third party from denying the existence of an agency relationship; and

2. The subsidiary’s conduct on behalf of the principal falls within the scope of authority which was granted to them by the principal.

V. How Agency Works With Other Principles

A. Filling the Void Left by Piercing?

Given the long-standing confusion between agency and piercing, the focus of this section is the relationship between our agency test and the doctrine of piercing post-Prest. Our test has substantially limited agency to express actual authority and apparent authority. A powerful counterargument is how a more liberal use of agency can fill the loophole left by a narrowing piercing principle in Prest to combat fraud or other form of corporate abuse beyond evasion of existing liability. In Prest, there would be a gap in the law, as there would be situations of fraud that do not reach the threshold of piercing the corporate veil but would fall outside the scope of agency due to their fraudulent nature. Restatement Second on Agency § 14M states:

[I]n some situations, a court may find that the subsidiary has no real existence or assets, that its formal existence is to cloak a fraud or other illegal conduct . . . . It may be found that the parent company is the real party to a transaction conducted by the illusory subsidiary and responsible for its transactions as a principal.

This is the scenario found in Trustor AB and Gilford Motor, which Lord Sumption and Lord Neuberger found to be covered and explained by the agency instead. Fraud, therefore, is what one may call a justification to imply an agency relationship unto the subsidiary company.

It is, however, submitted that, to establish an agency relationship through “fraud or illegal conduct,” is dangerously close to making agency law another “metaphor,” risking a re-entry into the mists of piercing the corporate veil. In Lord Neuberger’s judgment, His Lordship was of the opinion that Gilford illustrates this danger. In Gilford, Mr. Horne, the managing director of a company, undertook not to solicit customers of the company during and after termination of his employment. Termination eventually did occur, and a company was set up on behalf of Mr. Horne, with Mr. Horne’s wife and one of his employees acting as the sole directors and shareholders of the company. The court found that:

[The] company was formed as a device, a stratagem, in order to mask the effect carrying on of a business of Mr. E.B. Horne. The purpose of it was to try to enable him, under what is a cloak or sham, to engage in business which, on consideration of the agreement which had been sent to him just about seven days before the company was incorporated, was a business in respect of which he had a fear that the plaintiffs might intervene and object.

While the case never explicitly used the language of “piercing the corporate veil,” it has been seen as the case that laid the groundwork for this principle. While it is widely agreed that the decision was correct in granting an injunction against the company and Mr. Horne, their lordships had different interpretations as to why one should be granted. Lord Neuberger was of the opinion that agency could have dealt with this case without having to resort to the creation of piercing the corporate veil. His Lordship believed that the “cloak or sham” referred to in the original judgment merely referred to the fact that the company was actually acting on behalf of Mr. Horne and that the business was going to be “carried on for [him].” The legendary Lord Denning also took a similar position when he decided that the companies in Wallersteiner v Moir “were just the puppets of Dr. Wallersteiner . . . . they were his agents to do as he commanded. He was the principal behind them.” Both Lord Neuberger and Lord Denning seemed to focus on the control that the principal had over the subsidiary, to the extent that it was merely carrying out the principal’s wishes. But Lord Neuberger’s view clearly did not represent that of the U.K. Supreme Court. In Lord Sumption’s judgment, despite thinking that the injunction in Gilford Motor could be covered by the general injunctive power of the court instead of triggering piercing, he clearly stated that Gilford Motor was decided on piercing. In fact, he placed emphasis on the fact that “the case is authority for what it decided, not for what it might have decided.” In addition, drawing fraud as control was previously rejected in Yukong, where Justice Toulson was clear in holding that it should not mean anything for a subsidiary to act on the direction of the principal and for its benefit. While the benefit of the subsidiary assisting the principal to commit fraud is clearly directed towards the principal, this should not impact the relationship of the parties, as focusing on the benefit would lead to diminishing the significance and importance of the limited liability regime. Many forms of corporate vehicles are created for the benefit of the principal in carrying on its business. But, if mere benefit can amount to agency, the floodgates would be thrust wide open.

Perhaps more importantly, general agency under English law is not constituted by fraud, and this article contends for a return to general agency laws and principles. In our discussion of U.S. law above, fraud is an element of piercing the corporate veil, not agency. This should be the same under English law. In Prest, Lord Sumption spent much effort in identifying a principle under English law that “no court in this land will allow a person to keep an advantage which he has obtained by fraud.” According to him, piercing the corporate veil is part of that principle, but he never links fraud directly with agency. It is also true that he regarded Trustor AB, which involved finding shareholder liability in a fraudulent scheme, as an agency case. But to use Lord Sumption’s own words, “the case is authority for what it decided, not for what it might have decided.” Reading the judgment of Trustor AB, the court based its reasoning inequitably on piercing instead of agency. While the court made some reference to agency, it did not spend any time explaining how agency arose in the case. With all due respect, Lord Sumption was wrong in interpreting the case as being based on agency.

In our view, there may be legitimate arguments that frauds other than evasion of existing liability should be addressed. What can clearly be seen through the liberal U.S. piercing the corporate veil principle is that there is an undeniable need for a way to deal with injustice. But we contend that the correct way to do so is to revisit the scope of piercing. In fact, other Law Lords in Prest have expressed reservations that the piercing doctrine should only be limited to evasion of liability. Exactly how to redefine that boundary of piercing is something beyond the scope of this article, but it is clear that agency should not be the rug to sweep the uncertainties of piercing under. It has never been the role of agency law, nor should it be agency law’s future role. To do otherwise would be to inject a level of uncertainty into agency law generally, which is highly undesirable.

In short, we are of the opinion that each principle should be used to serve its unique and particular purpose. Agency and piercing the corporate veil should play distinct parts in the landscape of parent-subsidiary liabilities. Strictly speaking, agency should not be used to deal with fraud. Agency is not a tool to combat fraud, whereas piercing is a principle that was created for the very purpose and aim of combating unjust situations. Otherwise, agency will be seen as the new piercing the corporate veil. Should fraud be accepted as something which, when coupled with control, can imply an agency relationship, then it is foreseeable that many cases which used to plead piercing the corporate veil would now plead agency instead. In that case, English courts would have shut the gate on the uncertainty and the possibility of piercing the corporate veil becoming the go-to relief for injustice, but open one called agency, which would effectively serve the exact same purpose and be used in the exact same way. We, therefore, suggest that agency is not a viable substitute for the courts to deploy as a form of relief against fraud and injustice.

B. Why Does the Distinction Matter?

One might ask more fundamentally why there should be such a strict distinction between agency and fraud. This is indeed the view in Prest where Lord Walker considers that “‘piercing the corporate veil’ is not a doctrine at all . . . . It is simply a label to describe the disparate occasions on which some rule of law produces apparent exceptions to the principle of the separate juristic personality.” Agency, as another exception to Salomon, can then simply be covered by the “piercing” label. But confusing the two doctrines is unacceptable. That is certainly the view under U.S. law. Agency preserves corporate separateness, while piercing effectively merges the companies. But does this theoretical difference matter in practice? After all, one may argue that whatever the pathways, “[t]he ultimate consequence—the attribution of legal rights or responsibilities—is crucial, not the doctrinal route leading to that result.” It has been discussed above that agency should only be established when the subsidiary acts within the scope of agency. Agency may only exist in certain particular transactions that fall within the authorization of the parent company, while piercing is more generally applied to all dealings of the subsidiary in determining whether circumstances justify merging the parent with the subsidiary. The distinction, therefore, would have an implication on identifying the extent and scope of the parent’s responsibility.

C. A Quick Word on Direct Liability

Direct liability is a relatively new development in English law. The principle is based on tort law and, more particularly, on the assumption of responsibility by the parent over the subsidiary’s acts. It can serve as another exception to Salomon in practice. In theory, however, English courts have made it clear that direct liability is not piercing the corporate veil. The Chandler court simply holds the parent company liable for its own conduct. Thus far, all the U.K. Supreme Court cases applying this principle deal with jurisdictional issues, and the principle is very much in constant evolution. But it is still proper to make some brief observations on the relationship between agency and direct liability. First, since direct liability has its basis in tort law, it does not overlap with agency in non-tort cases, such as contract cases. Second, agency and direct liability overlap in tort cases. Both agency and direct liability could be imposed on a parent company when it makes representations regarding the assumption of certain safety measures on behalf of the subsidiary. The overlap will certainly be more significant if the courts end up applying a control-based test for agency. For example, in Fuller v. Midland Credit Management Inc. above, the court considered the fact that the parent devised collection policies and practices for the subsidiary. This is no doubt a factor for direct liability as well. Thus, adopting an agency test not based on control would further serve as a way to distinguish between these two distinct principles.

VI. Conclusion

The authors set out to find a suitable test for agency law in the English corporate context. This took place at a time when the principle of piercing the corporate veil had its role significantly limited in Prest. Naturally, agency may be seen as a fan-favourite to fill the void that is left behind, but this cannot, in our view, be allowed. While piercing still has a limited role to play to deal with injustice in a corporate context, the proposed test by the U.K. Supreme Court is very much restricted to those who use separate legal entities to evade existing liabilities. As such, many other instances of injustice may slip through the grasp of this now defunct principle.

But agency is not an effective substitute. It must not fall into the “mists” that piercing the corporate veil did. In the words of Blumberg, if we allow for agency to be applied in a way that does not really reflect agency principles, then we “simply add another metaphor to the arsenal of jurisprudence by metaphor.” As such, the authors have attempted to distinguish the agency principle from piercing, in particular barring any consideration of fraud in its determination. In our respectful opinion, this falls outside the scope and role of agency.

One must not forget that, even with fraud out of the picture, agency still has plenty to offer to the issue of parent liability in a corporate group. This is reflected mostly in the clear and simple way in which one can find agency through express actual authority and apparent authority. Admittedly, however, they are unlikely to be broadly applied, as stated above. In any event, broad application would be contrary to the traditionally strict safeguard of the principle of separate legal personality since Salomon. As Judge Dorn remarked in Lowendahl v. Baltimore & Ohio R.R. Co., “any severely logical application of agency rules would destroy the protection afforded stockholders by incorporation.” Agency is a narrow exception for sure, but we submit that there is nothing wrong with a principle being narrow, insofar as it is properly applied. We do not wish to push the envelope of agency law beyond what we believe is an accurate reflection of traditional agency principles simply because there is a gap in the law that needs be filled.

We streamline agency’s application by establishing a two-step test. While express actual authority and apparent authority are straightforward and do not have the potential for much expansion, implied actual authority seems to be an area that could potentially expand into a tool for combating injustice. After considering the difference between U.K. and U.S. laws on the general view of separate legal personality, it is concluded that control alone is not a proper basis to find implied authority. We believe that the added layer of scope helps further distinguish agency as a separate principle from piercing. In any event, we find this test to be a good starting point for the courts to adopt moving forward.

In the end, one thing is clear. U.S. courts use piercing the corporate veil to deal with injustice. The U.K. initially did so as well, but their lordships decided that the principle was going too far. Now that we have taken out the overarching principle used to combat these kinds of situations, English law needs, and indeed deserves, a new hero. In our opinion, however, agency law is not that hero. This is confirmed by the American experience. Nonetheless, we have planted the seeds for a principle that can greatly assist the courts in dealing with parent-subsidiary liabilities in a corporate context. How this plant grows and develops in the future is left in the capable hands of the English judges, whom we urge to use and develop it further. In our humble opinion, the time has come for agency to fall in line and finally play its proper role within a corporate group context.