Concept
Agency Roots of the Relationship
Formation of client-lawyer relationship
Forming the Client-Lawyer Relationship
Agency Roots of the Relationship
Before any rule can say what a lawyer owes a client, something must make a person a client. The law answers by borrowing from agency. The Restatement (Second) of Agency § 1(1) defines agency as the fiduciary relation that results when one person manifests consent that another act on that person's behalf and subject to that person's control, and the other person consents to act in that way.1 Read in the legal setting, the definition makes the client the principal and the lawyer the agent, so a dispute over whether a client-lawyer relationship ever formed is a dispute over whether this definition has been met.
Agency is the law of delegation. A person who needs something done can hand the task to another, and agency supplies the rules that make the handoff work. Three players are always present: the principal, the agent, and the third party with whom the agent deals.2 For a lawyer, the third party might be an opposing party or the counterparty to a deal. The arrangement works because the agent can act for the principal in a way that binds the principal to the third party, and the same feature makes it dangerous when an agent misbehaves. We return to that danger when we ask who decides what in a representation and who bears the loss if a lawyer turns out to be dishonest.
The definition has three working parts, and each of the next three subsections takes one: consent running from both sides, the principal's control, and the fiduciary character of the relationship.
Mutual Consent
The definition asks for consent from both sides. One person must manifest consent that another act on that person's behalf and subject to that person's control, and the other must consent to act in that way.3 A manifestation is something a person says or does that shows consent, so the inquiry turns on what each side expressed rather than on what either privately intended. In a representation, the would-be client manifests consent by asking the lawyer to take up a matter, and the lawyer manifests consent by agreeing. Until both halves are present, the definition is unmet and no relationship exists.
The Restatement then names the two roles. Section 1(2) calls the one for whom action is to be taken the principal, and § 1(3) calls the one who is to act the agent.4 A client is the person for whom the lawyer acts, so the client is the principal and the lawyer is the agent. From here on, a statement about what a principal may demand is a statement about what a client may demand, and a statement about an agent's duties describes the lawyer's.
Both consents are required because of what each side takes on. Agency works because the agent can bind the principal to a third party. A person who asks a travel agent to book an Amtrak ticket from Los Angeles to San Diego must pay for it even though that person never dealt with Amtrak, because the agent had authority to bind the principal to the purchase.5 A lawyer likewise speaks for a client to counterparties and tribunals, and no stranger should become another person's voice without that person's agreement. The agent's side carries its own weight, because the relationship is fiduciary and a person should not take on those duties without agreeing to act. Requiring consent from both sides keeps either status from being imposed on someone who never chose it.
A fact pattern must show two manifestations before it shows a client-lawyer relationship, and nothing in the definition requires that either be written.
Because the law looks to manifestations rather than paperwork, a phone call can carry the whole exchange. A signed engagement letter is excellent evidence that both consents were given, and lawyers prefer one for that reason, but it proves the two manifestations and does not replace them.
The client's request is only half of the test, and an answer that stops there misses the point whenever the lawyer's side is absent.
The owner's wish, however clearly expressed, cannot draft the lawyer into service, because the second half of § 1(1) calls for the other person's consent as well. Later sections show how a lawyer's conduct can sometimes supply that second half; here the lawyer's words refuse it outright.
The consent that creates the relationship also keeps it going, which is why agency is called a consensual relationship. It ends whenever either side withdraws its consent that the agent act on the principal's behalf and subject to the principal's control.6 Either party has the power to end the agent's actual authority by revoking or renouncing it, even if doing so breaches a contract between them and exposes that party to a damages claim; the authority has ended all the same.7 Whether a discharge also breaches a fee agreement is therefore a separate question from whether the relationship has ended. Some fiduciary obligations outlast the consent that created them, a point taken up under the fiduciary character of the relationship.8 Termination as a topic of its own is taught elsewhere.
Acting Under the Client's Control
The second element sits in two phrases of the definition: the agent must act on the principal's behalf, and must do so subject to the principal's control.10 Acting on someone's behalf means doing the task for that person's benefit, so a lawyer's work belongs to the client's matter and serves the client's ends. Control is the phrase that does the sorting, because it separates an agent from a person who merely happens to be doing something useful for another. Control here means the principal is in charge; the principal need not control every last detail of how the job gets done.11
The definition insists on control because of what agency does to the principal. An agent who deals with a third party can bind the principal to that dealing, as the travel agent's booking showed earlier in this chapter. A person who can be held to another's choices needs some way to shape them, and agency law supplies it as a right to direct the agent and to end the agent's authority. Without that right, the agent would make binding choices for someone who could not review them, and the principal would absorb results that no one could be made to justify. The element asks only for a right of control, a lower bar than control of every detail, and the lower bar makes sense because agents are usually engaged for skill the principal lacks. A client who hires a litigator wants the lawyer to choose the research path and draft the briefs, and a rule requiring the client to supervise those details would exclude nearly every professional engagement from agency.
The control that general agency requires is also thinner than the control that marks an employment relationship. An employer has the right to control the day-to-day details of the job, such as the uniform a server wears and the way dishes are served and cleared, while general agency requires only that it be clear who is the principal and who is the agent.12 A lawyer who exercises independent professional judgment about research and tactics therefore still acts under the client's control in the sense the definition uses, because the client remains the person in charge.
A stem that stresses how little the client did day to day is inviting the wrong question. The inquiry is whether the client holds the right to direct the lawyer and the power to end the engagement, and a client who delegates nearly everything remains a principal for that reason.
Courts applying the control element outside the law office show what its absence looks like. The two decisions below are not client-lawyer cases; they appear here because each treats control as the essential feature. In Hollingsworth v. Perry, the proponents of a California ballot initiative defended the measure and claimed to act for the people of California. The Court rejected the claim because the most basic feature of agency was missing, the principal's right to control the agent's actions: the proponents answered to no one, decided for themselves and without review what arguments to make and how to make them, were not elected, and could not be removed under any provision.13 With no one able to direct or remove them, there was no principal at all.
Taylor v. Sturgell presents the control question from the other side. There the issue was whether a later litigant could be treated as the agent of a party bound by an earlier judgment, so that preclusion would apply. The Court cautioned that courts should be wary of finding preclusion on this basis, that a mere whiff of tactical maneuvering will not suffice, and that agency principles indicate preclusion is appropriate only if the putative agent's conduct of the suit is subject to the control of the party bound by the prior adjudication.14 The two cases differ in where the control is missing. In Hollingsworth, no principal could direct or remove the proponents, so the relationship had no one at the top. In Taylor, a possible principal was identifiable, and the open question was whether that party controlled the later suit. Applied to a law office, the same logic means that a person the client cannot direct or discharge is not the client's agent, however closely their interests align.
Fiduciary Character
The definition calls agency "the fiduciary relation" that results from the two manifestations of consent, and that label has consequences of its own.16 A fiduciary holds a position of trust and must act for the other person's benefit rather than the fiduciary's own. The earlier elements told you when a client-lawyer relationship exists; this one tells you what the lawyer owes once it does. Because the relationship is fiduciary from the moment both consents are present, the duty of loyalty attaches then, and the later rules on conflicts and confidentiality elaborate that one idea for lawyers.
Loyalty is the core of the fiduciary obligation. The law presumes that a principal expects, and is entitled to expect, that the agent will use the skill and experience for which the agent was retained to further the principal's interests. An agent who instead promotes his own interest at the expense of the principal's violates the agency relationship and is liable for any loss that results.17 The presumption makes sense given what delegation involves: the principal hands over decisions and information that the principal cannot easily check, so a disloyal choice can be costly before anyone notices.
A stem may stress that the seller, not the client, paid the lawyer, hoping to make the payment look harmless. The question remains whose interest the lawyer served, and here the lawyer served the lawyer's own at the client's expense.
Loyalty is not the agent's only obligation. An agent must also act in good faith, confine his acts within the scope of his actual authority, obey the principal's instructions, and use due care and reasonable diligence in the business entrusted to him.18 Those duties return when we reach client control and malpractice exposure.
Hollingsworth also shows what a missing fiduciary relation looks like. The proponents had taken no oath of office and were bound only by the same ethical constraints that apply to every other party in a legal proceeding, so they owed the people of California nothing of the sort an agent owes a principal.19 They were free to pursue a purely ideological commitment to the law's constitutionality without regard to resource constraints, changes in public opinion, or effects on other state priorities, and that freedom is exactly what a fiduciary lacks. Control and loyalty fit together. Control gives the principal the means to direct and remove the agent, and the fiduciary duty assures the principal that the agent, while serving, serves the principal. A person who answers to no one and owes loyalty to no one is not an agent.
Fiduciary duties can also outlast the relationship that created them. Consent creates an agency, and withdrawal of consent ends the agent's actual authority, yet certain obligations persist, chiefly the obligation not to use confidential information learned during the relationship.20 The duty of loyalty is strong enough that aspects of it continue to bind the agent after the relationship has ended.21 Information keeps its value after the engagement is over, and if the duty stopped at termination, a principal could be left at the agent's mercy simply by ending the relationship. At the same time, as a matter of agency law a former agent is free to compete with the former principal.22 The line therefore falls between using the principal's confidences, which stays forbidden, and competing as such, which does not, and drawing that line in practice is harder than stating it.
The rules governing lawyers and former clients are more detailed, and later readings take them up. The structural point is enough here: duties that arise when the relationship forms can bind the lawyer long after the engagement ends, which is why the formation question deserves care.