1 Core concepts

Principal–agent theory examines situations in which one party delegates tasks, authority, or specialized judgment to another. The framework is used to explain why delegation is efficient, and why it can also produce inefficiency when the two parties have different goals, different information, or different tolerance for risk. At its core, the theory asks how to structure incentives so that the delegated party behaves in ways that are broadly consistent with the delegator’s interests.

1.1 Principal and agent roles

The principal is the party that assigns a task or confers decision rights. The agent is the party that carries out the task or exercises that authority. In many settings the principal wants a result, while the agent controls the actions needed to produce it. The distinction is often not rigid: a person can be a principal in one relationship and an agent in another.

1.2 Delegation and decision rights

Delegation allows the principal to take advantage of specialization, expertise, or lower operating costs. The agent may have better technical knowledge or closer contact with the relevant situation, making direct control unnecessary or impractical. Delegation also means that some choices are transferred away from the principal, which creates a need for rules, reporting systems, and incentives that guide the agent’s conduct.

1.3 Information asymmetry

A central feature of principal–agent relationships is that one side usually knows more than the other. The agent may observe effort, opportunities, or private conditions that the principal cannot fully see. This asymmetry affects bargaining, contract design, and the ability to assess performance.

1.3.1 Hidden action

Hidden action refers to behavior that cannot be directly observed or verified by the principal. The agent may choose how hard to work, which methods to use, or whether to take precautions, but those choices are only imperfectly visible through outcomes. This makes it difficult to distinguish good effort from luck.

1.3.2 Hidden information

Hidden information arises when the agent knows facts that the principal does not know at the time of contracting. These facts may concern ability, cost, risk, or quality. Because the principal cannot easily verify them, the agent may have an advantage in negotiation or may be able to choose a contract that exploits that informational gap.

1.4 Incentives and alignment

The main policy problem in principal–agent theory is incentive alignment. The principal tries to reward desirable behavior and discourage actions that are harmful or wasteful. Alignment can be pursued through pay schemes, monitoring, reputation, promotion rules, and contract clauses. Since perfect control is rarely possible, the practical aim is usually to reduce misalignment rather than eliminate it completely.

2 Key problems

Several recurring difficulties make principal–agent relationships hard to manage. These problems often overlap, since the same informational gap can produce both poor effort and poor selection. The theory provides a vocabulary for describing these frictions and comparing alternative institutional responses.

2.1 Moral hazard

Moral hazard occurs when the agent changes behavior after a contract is in place because the principal cannot fully observe or verify the relevant actions. The term is commonly associated with a weakening of discipline once the agent is protected from the full consequences of a choice.

2.1.1 Effort shirking

Effort shirking is the reduction of work intensity, care, or diligence after delegation. Because effort is costly to the agent and not always measurable, the agent may supply less than the principal expects. This is one of the most familiar examples of moral hazard.

2.1.2 Risk shifting

Risk shifting happens when the agent takes on more risk than the principal would prefer, especially if losses are borne partly by others. A contract that insulates the agent from downside losses can encourage actions with high variance even when the principal values steadier outcomes.

2.2 Adverse selection

Adverse selection refers to problems that arise before a contract is chosen, when one side knows more about its own characteristics than the other. The principal may end up offering terms that attract the wrong type of agent or fail to reveal important differences among candidates.

2.2.1 Screening problems

Screening problems involve designing offers that separate more desirable agents from less desirable ones. A principal may use menus of contracts, tests, or qualification standards to learn about hidden characteristics. Poor screening can lead to inefficient matching and weak performance.

2.2.2 Hidden types

Hidden types are unobserved traits such as productivity, reliability, or risk tolerance. Because these characteristics are not immediately visible, they influence the terms under which trade or delegation occurs. The possibility of hidden types often shapes both hiring decisions and contract structure.

2.3 Monitoring costs

Monitoring is costly, time-consuming, and sometimes intrusive. The principal must decide how much oversight is worth the expense. If monitoring is too weak, opportunistic behavior may increase; if it is too intense, it may create bureaucracy, resentment, or unnecessary expense. Optimal monitoring usually balances these trade-offs.

2.4 Contract incompleteness

Contracts are incomplete when they cannot specify every possible future contingency or every relevant duty in enforceable detail. Real-world uncertainty, limited foresight, and legal constraints prevent fully comprehensive agreements. As a result, many relationships rely on broad terms, discretion, and informal understandings.

3 Formal framework

Principal–agent theory is often expressed in formal models that compare the preferences of the two parties under uncertainty. These models show how incentives, risk, and information jointly determine the contract that can be sustained.

3.1 Utility maximization

Both principal and agent are modeled as choosing actions or contracts to maximize expected satisfaction subject to constraints. The principal seeks a desired outcome net of payment and risk, while the agent evaluates compensation against effort cost and uncertainty. The tension between these objectives is the basis of the theory.

3.2 Expected utility and risk aversion

Expected utility analysis allows the parties to care not only about average returns but also about the distribution of outcomes. Risk aversion matters because an agent who dislikes risk may require a premium to accept variable pay. This creates a trade-off between stronger incentives and insurance.

3.3 Incentive compatibility

An allocation is incentive compatible when the agent finds it optimal to follow the intended action or report truthful information. Incentive compatibility is central because contracts only work if they are attractive from the agent’s point of view, not merely from the principal’s.

3.3.1 Participation constraints

Participation constraints require that the agent prefer accepting the contract to opting out. If compensation is too low or too uncertain, the agent may refuse the agreement. The principal must therefore provide enough expected value to secure cooperation.

3.3.2 Incentive constraints

Incentive constraints require that the agent prefer the promised action over alternative, self-serving actions. These constraints formalize the need for effort, honesty, or compliance to be individually rational once the contract is in force. They are often the main restriction on what the principal can implement.

3.4 First-best and second-best outcomes

A first-best outcome is one that would be achieved if all actions and information were fully observable and enforceable. In practice, this ideal is rarely attainable. A second-best outcome is the best feasible arrangement given the constraints created by hidden information, hidden action, and limited enforcement.

4 Contract design

Contract design aims to shape behavior through payment rules, monitoring, and the allocation of authority. Different contract forms solve different parts of the incentive problem, and each has its own limitations.

4.1 Fixed-wage contracts

Fixed wages pay the agent a set amount independent of observed performance. They provide income stability and reduce exposure to risk, but they offer weak incentives if effort is hard to observe. Such contracts are common when tasks are routine or when performance measurement is unreliable.

4.2 Performance-based pay

Performance-based pay links compensation to measurable results. It can strengthen incentives by rewarding success and penalizing failure. However, it may also encourage manipulation of metrics or excessive focus on what is measured at the expense of unmeasured aspects of performance.

4.3 Bonuses and penalties

Bonuses and penalties add positive or negative adjustments to a base contract. They are often used to motivate short-term goals, quality standards, or compliance. Their effectiveness depends on whether the underlying performance measure is credible and whether the agent believes the rule will be applied consistently.

4.4 Commission contracts

Commission contracts pay a share of generated revenue, output, or sales. They are useful when individual effort is closely related to measurable outcomes. At the same time, commissions can push agents toward aggressive selling, selective client targeting, or short-term gains that do not always match the principal’s broader objectives.

4.5 Relative performance evaluation

Relative performance evaluation compares one agent’s results with those of peers or benchmarks. This can reduce the effect of common shocks, such as weather or market-wide changes, and help isolate individual contribution. It is especially useful where absolute performance is noisy.

4.6 Relational contracts

Relational contracts are based on ongoing relationships, reputation, and the expectation of future cooperation rather than on detailed legal enforcement. They are common when formal enforcement is weak or when flexibility is valuable. Trust and repeat interaction help sustain behavior that is hard to specify in advance.

5 Analytical models

Formal agency models provide structured ways to analyze different sources of conflict between the parties. They are used to compare contract forms, identify limiting constraints, and explain observed organizational patterns.

5.1 Principal–agent model under risk

In the standard risk model, the principal offers a contract to an agent who chooses effort before an uncertain outcome is realized. The contract must balance incentive power against the agent’s dislike of income volatility. Stronger incentive pay can improve effort but may require greater risk-bearing by the agent.

5.2 Hidden action models

Hidden action models focus on actions that cannot be directly observed. The principal infers effort from outcomes, but outcomes are also affected by randomness. These models clarify why full reimbursement or pure salary may be inefficient in some settings, even if they provide insurance.

5.3 Hidden information models

Hidden information models study situations in which the agent has private knowledge before the contract is finalized. The principal may offer different contract options to induce self-selection. These models are useful for understanding screening, truthful reporting, and market segmentation.

5.4 Multi-task agency

Multi-task agency examines cases in which the agent performs several tasks, some measurable and others not. If incentives are tied mainly to measurable tasks, the agent may neglect activities that are harder to quantify but still important. This creates a common reason for mixed compensation systems.

5.5 Dynamic principal–agent models

Dynamic models extend the analysis across time. They allow for learning, reputation, career concerns, and repeated contracting. Over time, the principal may observe more about the agent, while the agent may respond to future rewards or penalties, making long-term relationships especially important.

6 Applications

Principal–agent theory is applied in many fields because delegation is common in economic and organizational life. The details differ across settings, but the underlying structure of asymmetric information and incentive conflict remains similar.

6.1 Employment relationships

In employment, workers often choose their effort level after being hired, while employers cannot observe all aspects of performance. Incentive systems, promotion ladders, supervision, and workplace norms are used to reduce shirking and improve coordination.

6.2 Executive compensation

Corporate boards and shareholders face a classic agency problem with managers. Executive compensation may include salary, bonuses, stock options, and long-term awards intended to link management behavior to firm performance. The challenge is to reward value creation without encouraging short-term manipulation.

6.3 Insurance contracts

Insurance illustrates moral hazard because coverage can reduce the insured party’s incentive to avoid losses. Deductibles, co-payments, and coverage limits are common devices for preserving some responsibility while still providing protection against major risks.

6.4 Financial intermediation

Banks, lenders, and investors often rely on intermediaries who have specialized knowledge about borrowers or assets. Agency issues arise when intermediaries make lending, trading, or screening decisions on behalf of others. Monitoring, covenants, and capital requirements are used to reduce these tensions.

6.5 Regulation and public policy

Regulators act on behalf of the public, but they may have different information, priorities, or responsiveness than the people affected by their decisions. Principal–agent analysis helps explain oversight structures, reporting requirements, and the design of public agencies.

6.6 Supply chain relationships

In supply chains, firms depend on distributors, retailers, contractors, or suppliers whose actions influence quality, timing, and cost. Contracts often combine quantity targets, quality standards, audits, and relationship-based cooperation to align incentives across the chain.

The basic framework can be extended to settings with many participants, collective action, and more complex organizational structures. These extensions show that agency problems are not limited to simple one-to-one relationships.

7.1 Multiple principals and multiple agents

Many real arrangements involve several principals or several agents at once. Competing objectives can create coordination problems, while multiple agents may free-ride on one another’s effort. The resulting structure is often more difficult to contract than the simple model suggests.

7.2 Team incentives

Team incentives reward collective output rather than individual contribution. They are useful when performance is joint or when individual effort is hard to isolate. However, they can weaken personal accountability unless paired with peer monitoring or internal discipline.

7.3 Common agency

Common agency occurs when one agent serves several principals simultaneously. The agent may receive conflicting instructions or exploit the principals’ lack of coordination. This situation is common in lobbying, finance, and supplier relationships.

7.4 Delegation and organizational design

Organizational design concerns how authority is distributed within firms and institutions. Decisions about hierarchy, decentralization, reporting lines, and oversight can be understood as responses to agency problems. The goal is to place decision rights where information is best while maintaining accountability.

7.5 Behavioral principal–agent models

Behavioral models incorporate psychological factors such as fairness concerns, reciprocity, overconfidence, and limited attention. These approaches suggest that real agents may not respond purely to monetary incentives. As a result, social norms and perceived legitimacy can matter alongside formal contracts.

8 Empirical and practical considerations

Although principal–agent theory is highly influential, applying it in practice requires careful measurement and interpretation. Real contracts are shaped by legal institutions, industry norms, and imperfect data, all of which affect what can be observed and enforced.

8.1 Measuring incentives

Researchers and practitioners must determine how strongly pay depends on performance and how sensitive the agent is to those rewards. Measuring incentives is difficult because compensation packages often contain multiple components and because behavior can respond in indirect ways.

8.2 Observability and verification

Not every relevant action can be observed, and not every observed event can be verified in a formal setting. Distinguishing measurable output from true effort is a persistent challenge. The more difficult verification becomes, the more likely parties are to rely on proxy measures or informal judgment.

8.3 Contract enforcement

Even well-designed contracts require enforcement to be credible. Enforcement may depend on courts, arbitration, internal discipline, or reputation. Weak enforcement reduces the value of elaborate clauses and increases reliance on simple, robust arrangements.

8.4 Empirical testing of agency models

Empirical research tests agency models by examining how compensation, monitoring, and organizational structure affect performance. Because many factors move together, causal inference can be difficult. Researchers often use natural experiments, panel data, or institutional variation to isolate agency effects.

8.5 Limitations of the theory

Principal–agent theory is powerful but not exhaustive. It may simplify motivations, understate cooperation, or assume a level of rational calculation that does not always exist. It also depends on measurable outcomes and stable preferences, which are not present in every setting. Even so, it remains one of the most important tools for analyzing delegation and incentives.

</INTERNAL_LINK_CANDIDATES> Moral hazard (behavioral change after contract formation that harms the principal) Adverse selection (pre-contract hidden characteristics affecting matching) Information asymmetry (unequal access to relevant information between parties) Hidden action (unobservable effort or behavior) Hidden information (private knowledge known before contracting) Incentive compatibility (contract conditions making desired behavior optimal) Participation constraint (condition that acceptance is preferable to opting out) First-best outcome (ideal efficient result under full observability) Second-best outcome (best feasible result under constraints) Fixed-wage contract (compensation independent of performance) Performance-based pay (compensation linked to measurable results) Commission contract (pay tied to sales or output) Relative performance evaluation (comparing performance against peers or benchmarks) Relational contract (informal agreement sustained by trust and repetition) Multi-task agency (one agent handling multiple tasks with different measurability) Dynamic principal–agent model (agency model extended over time) Executive compensation (pay arrangements for corporate managers) Insurance contract (policy balancing coverage and incentive effects) Common agency (one agent serving multiple principals) Organizational design (allocation of authority and structure within institutions) </INTERNAL_LINK_CANDIDATES>