Hire someone to act for you — a lawyer, a manager, a senator — and you've just created a problem older than money: their interests and yours are not the same. This tension shapes every institution humans build — the underlying mechanics that kick in the moment you put two people in a room and ask one to act on behalf of the other. This lesson is about one of the most fundamental: the principal-agent problem. It's not a flaw in bad organizations. It's a structural feature of every organization, everywhere, always.

The setup is simple. A principal is whoever has a goal and needs someone else to pursue it. An agent is whoever gets hired, elected, or appointed to do that pursuing. A shareholder is a principal; the CEO they hire is their agent. A voter is a principal; the senator they elect is their agent. A patient is a principal; their doctor is their agent. The principal can't be everywhere at once, so they delegate. And the moment they delegate, a gap opens up. That gap is the problem. The agent knows things the principal doesn't — how hard they're actually working, what tradeoffs they're making, what information they're sitting on. And the agent has their own interests: job security, salary, reputation, comfort, career advancement. Those interests overlap with the principal's, but they never perfectly align. Never.

This is the core of every principal-agent problem: the agent doesn't act against the principal out of malice. They act in their own interest, which the principal accidentally incentivized. The shareholders who designed a bonus tied to quarterly earnings created the misalignment themselves. The voters who only pay attention every four years created the conditions for a senator to serve donor interests the other 1,460 days. The patient who can't read a lab report created the space for a doctor to recommend the more profitable procedure. Milton Friedman (1912–2006, United States) — the economist who argued that a corporation's only social responsibility is to its shareholders — spent decades pointing out that the managers running those corporations were agents with their own agendas, not pure proxies for owner interests. He was right about the problem even when people disagreed about his solution.

So why can't you just fix it? Principals have tried three main approaches for centuries: monitoring (watch the agent closely), incentive alignment (make the agent's rewards track yours), and selection (hire agents whose interests naturally match yours). Each works partially. None works completely. Monitoring is expensive and creates its own problems — agents who perform for the camera and coast everywhere else. Incentive alignment is powerful but always incomplete; you can tie a CEO's pay to stock price, and she'll start buying back shares to inflate it instead of building the business. Selection helps, but people change when they get the job — the loyal deputy who becomes a self-interested executive is a cliché because it happens constantly. The organizations that handle this best aren't the ones that eliminated the gap. They're the ones that designed around it — short feedback loops, transparent reporting, term limits, audits, whistleblower protections. Not solutions. Friction. Enough friction that the agent's self-interest costs them more than it gains them. That's the best any institution has ever managed. Every system of governance you'll encounter in this series — from how votes get counted to how regulators interact with the industries they oversee — is, at some level, a response to this one problem. The gap never closes. The only question is how well it's managed.