principal-agent problem

You hire a contractor to renovate your kitchen and then go to work all day. You want it done well and cheaply; the contractor would rather cut corners and pad the bill. You can't watch every nail. Or: shareholders own a company but hire a CEO to run it; the CEO might chase a private jet and empire-building over the owners' profits. Whenever one person (the principal) relies on another (the agent) to act on their behalf, but the two want different things and the principal can't fully monitor the agent, you have the principal-agent problem.

The problem has two ingredients: a conflict of interest (the agent's goals differ from the principal's) and asymmetric information (the agent knows more about their own effort and choices than the principal can observe). Because the agent acts in their own interest while spending the principal's money or risking the principal's outcomes, the agent may shirk, take hidden risks, or self-deal. Note this is moral hazard wearing organizational clothes: it is the same hidden-action logic applied to delegated relationships.

The principal-agent problem is the silent engine behind much of how organizations are designed: performance-based pay, stock options for executives, sales commissions, auditing, boards of directors, professional codes, and contracts that tie reward to results all exist to align the agent's incentives with the principal's. The honest catch is that no contract can perfectly close the gap — measurable targets can be gamed (teachers teaching to the test, bankers chasing short-term bonuses at long-term cost), so designing incentives is an ongoing, imperfect art, not a solved problem.