The Agency Problem

When one party (the agent) acts on behalf of another (the principal), their interests inevitably diverge. The principal wants outcomes; the agent wants compensation. This gap—the agency problem—shapes organizations, markets, governance, and anywhere delegation exists.

You hire a real estate agent to sell your house. You want the highest price. They want the deal closed quickly. Your interests overlap but don't align perfectly.

This is the agency problem in miniature. It scales to every principal-agent relationship.

The Structure

Principal: The party who delegates authority. Wants outcomes.

Agent: The party who acts on the principal's behalf. Wants compensation.

The problem arises because:

• Agent has more information than principal (information asymmetry)
• Agent's effort is hard to observe
• Agent's incentives differ from principal's

Where It Appears

Corporate Management

Shareholders (principals) hire executives (agents) to run companies. Shareholders want long-term value. Executives want bonuses, often tied to short-term metrics. Result: short-termism, empire-building, excessive risk-taking with shareholder money.

Politics

Citizens (principals) elect representatives (agents) to govern. Citizens want good policy. Representatives want re-election. Result: policies that look good over policies that work, pandering over truth.

Medicine

Patients (principals) hire doctors (agents) for health. Patients want health. Doctors want income and malpractice protection. Result: over-testing, over-treatment, defensive medicine.

Finance

Investors (principals) give money to fund managers (agents). Investors want returns. Managers want fees. Result: fee structures that benefit managers regardless of performance, excessive trading, closet indexing.

Employment

Employers (principals) hire workers (agents). Employers want output. Workers want wages. Result: monitoring, performance management, the entire apparatus of HR.

Why It's Hard to Solve

Information Asymmetry

Agents know more about their domain. Principals can't evaluate quality directly. The information gap enables exploitation.

Unobservable Effort

You see outcomes, not effort. Was the bad outcome due to low effort or bad luck? Hard to distinguish. Agents exploit this ambiguity.

Multi-Dimensional Tasks

Jobs have many dimensions. Incentivizing one dimension leads to neglect of others. Measure quantity, quality drops. Measure sales, service drops.

Mitigation Strategies

Align Incentives

Make agent rewards depend on principal outcomes. Stock options for executives. Contingent fees for lawyers. But beware: aligned incentives can still be gamed.

Monitor

Oversight reduces information asymmetry. But monitoring is costly and can backfire (agents game the monitors).

Reputation

Long-term relationships change the game. Agents with reputations to protect behave better. One-shot interactions maximize agency problems.

Selection

Choose agents whose values align with yours. Intrinsic motivation beats extrinsic incentives. But selection is hard when you can't observe values.

Reduce Delegation

Do it yourself. Eliminates agency problem but limits scale. Sometimes the best agent is you.

The Decoder View

When you see organizations, markets, or institutions behaving in puzzling ways, ask: who are the agents? Who are the principals? How do their incentives diverge?

The agency problem is structural. It's not about bad people—it's about misaligned incentives. Good people in bad structures produce bad outcomes.

— Decoded by DECODER