Principal-Agent Problem

Definition

The principal-agent problem arises when a principal (decision-maker) delegates authority or a task to an agent who has private information and objectives that may conflict with the principal's. The fundamental issue is that the principal cannot fully verify the agent's information or actions, making it costly or impossible to align their incentives through direct control. This is not merely a problem of imperfect enforcement, but one where conflicting objectives and decentralized information combine to create genuine constraints on what contracts can achieve.

In the Book

The book identifies this as the core problem of incentive theory: when an agent is "selected for his specialized knowledge and the principal can never hope to completely check the agent's performance" (Arrow 1963, cited in Ch. 1). Laffont and Martimort trace this problem through economic history—from Adam Smith's discussion of sharecropping in agriculture (where tenant-farmers have incentives to hide their use of the landlord's cattle), through Chester Barnard's work on management incentives, to modern regulation and financial contracting.

The problem takes two main forms. With adverse selection (Ch. 3), the agent knows their cost or type before the contract is signed and may misrepresent it. With moral hazard (Ch. 4), the agent takes actions after contracting that the principal cannot observe (like effort level or care). The book's central insight is that under either form of hidden information, the first-best allocation—what would be achieved if the principal knew everything—becomes unachievable. The principal must instead find the best feasible contract, which necessarily involves a trade-off between extracting information rents and maintaining efficiency.

Why It Matters

The principal-agent framework provides a lens for understanding delegation in any context where information is asymmetric: employment relationships, regulation of firms, insurance, lending, governance, and organizational design. It shows that the problem is not moral failure or stupidity, but structural. The framework moves beyond anecdote and allows precise analysis of what constraints bind, how contracts should be designed, and what costs asymmetric information imposes on society. This understanding is foundational to institutional design and policy.