Skip to content

Glossary · Game Theory

Principal-Agent Problem

also: agency problem · agency theory · principal–agent problem · moral hazard · multitasking problem · incentive design

Definition

The principal-agent problem arises when one party delegates work to another whose actions are imperfectly observable and whose interests differ. Jensen and Meckling (1976) and Holmström (1979) formalized the resulting trade-off between risk sharing and incentives; sales compensation, agency fee structures, and attribution disputes are its everyday forms inside a growth organization.

Because effort is hidden, the principal can only contract on outcomes, and outcomes are noisy, so strong incentives load risk onto the agent and weak incentives invite shirking. Holmström's informativeness principle says every signal that carries information about effort should enter the contract, which is the theoretical case for relative performance evaluation. Holmström and Milgrom (1991) added the multitasking problem: when only some tasks are measurable, paying for those tasks pulls effort away from the unmeasured ones, which is why high-powered commission plans corrode customer quality and why agencies paid on media spend overspend. Baker (1992) showed that gaming follows whenever the performance measure diverges from the principal's real objective, a pattern later popularized as Goodhart's law.

Essays on this concept