The Principal-Agent Problem: Why Your Incentives Are Never Aligned

In 2012, Greg Smith — a vice president at Goldman Sachs — published an op-ed in the New York Times that detonated across Wall Street like a grenade with a twelve-year fuse. He resigned publicly, in print, and explained why.

"It makes me ill how callously people talk about ripping their clients off," he wrote. He described a culture where clients were referred to as "muppets" — puppets whose money could be extracted through complex products they didn't understand. Bankers didn't ask "what's best for the client?" They asked "what has the highest margin for us?"

The outrage was enormous. Goldman's stock dropped $2.2 billion in a single day. Commentators called it a betrayal of trust, a moral failing, a crisis of corporate culture.

It was none of those things. Or rather, it was all of those things — but they weren't the cause. They were the symptom. What Greg Smith described wasn't a broken culture at one bank. It was a structural feature of every relationship where one person hires another person to act on their behalf. Economists have a name for it. They've been writing about it since the 1970s. And it explains why your financial advisor, your doctor, your real estate agent, your mechanic, your lawyer, and your elected representative are all, to varying degrees, optimising for something other than your wellbeing.

It's called the Principal-Agent Problem, and once you see it, you'll never interact with a professional the same way again.

The Structure of Misalignment

The setup is simple. A principal (you) hires an agent (a professional) to act on your behalf, because the agent has expertise or access that you lack. You hire a doctor because you can't diagnose yourself. You hire a financial advisor because you don't understand derivatives. You hire a real estate agent because you don't know how to navigate a housing market.

The problem is equally simple: the agent's interests are not your interests. They overlap — the agent wants you to be satisfied enough to keep paying — but they diverge in specific, predictable, and often invisible ways.

Your doctor wants you to be healthy. Your doctor also wants to avoid malpractice suits, finish appointments on time, and — if they own a practice — maintain revenue. When "keep the patient healthy" and "avoid liability" point in the same direction, the system works. When they diverge — when the safest treatment for the doctor's career isn't the best treatment for the patient's body — the agent's interest wins.

This isn't because doctors are bad people. It's because incentive structures shape behaviour more reliably than character does. As Goodhart's Law teaches us, people optimise for what they're measured on. And agents are measured on their own outcomes, not yours.

Information Asymmetry: The Enabler

The principal-agent problem only bites because of information asymmetry — the agent knows things the principal doesn't. If you could evaluate your doctor's recommendations as well as another doctor could, the problem would dissolve. You'd simply overrule the bad recommendations and follow the good ones.

But you can't. That's why you hired them. The expertise gap that makes the agent valuable is the same gap that makes the agent dangerous. They know more than you, which means they can serve your interests — or their own — and you often can't tell the difference.

The principal-agent problem exists wherever two conditions are met: someone acts on your behalf, and you can't fully verify whether they acted in your interest. This describes virtually every professional relationship you have.

The Agents in Your Life

The principal-agent problem isn't a theoretical curiosity. It's operating in real time across your finances, health, career, and most significant purchases.

Your Financial Advisor

Rachel Torres, a marketing director in Denver, hired a financial advisor in 2018 on the recommendation of a friend. The advisor seemed knowledgeable, attentive, and trustworthy. He put Rachel into a portfolio of actively managed mutual funds. The performance was decent. Rachel felt taken care of.

In 2021, Rachel's nephew — a finance graduate — looked at her portfolio and found something she didn't know. Every fund in her portfolio charged a front-end load of 3-5% and had annual expense ratios above 1.2%. Her advisor received a commission on each fund. Lower-cost index funds — which consistently outperform actively managed funds after fees — would have earned Rachel roughly $47,000 more over the five years she'd been invested.

Her advisor hadn't stolen anything. He hadn't lied. He'd recommended products that were legal, reputable, and genuinely provided investment exposure. He'd also chosen the products that paid him the most, from a universe that included cheaper alternatives that would have been better for Rachel. The advice was shaped by the advisor's compensation structure, not Rachel's financial goals.

This is the principal-agent problem in its purest form. The advisor's interest (maximise commissions) overlapped with Rachel's (invest money) but diverged on the specific execution (which funds). The information asymmetry (Rachel couldn't evaluate fund fee structures) prevented her from detecting the divergence. And the relationship's surface — trust, friendliness, apparent competence — concealed the structural misalignment underneath.

Your Real Estate Agent

You're selling your house. Your agent earns a 3% commission. On a $400,000 sale, that's $12,000. If the agent could get you $420,000 by holding out for another two weeks, the extra $20,000 for you translates to only $600 more for the agent. Two additional weeks of work, open houses, and uncertainty — for $600.

The agent's rational move: encourage you to accept the first reasonable offer and close quickly. More transactions per year, at slightly lower prices, earns the agent more than fewer transactions at slightly higher prices. Your agent's optimal strategy (sell fast) directly contradicts your optimal strategy (sell high).

Economists Steven Levitt and Chad Syverson tested this using data on Chicago home sales. They compared how long agents kept their own homes on the market versus how long they kept clients' homes on the market. The result: agents kept their own homes listed an average of ten days longer and sold them for 3.7% more. When their own money was on the line, they suddenly found the patience they didn't have with yours.

Your Doctor

A surgeon who is paid per procedure has a structural incentive to recommend surgery. A surgeon on a fixed salary does not. The evidence bears this out: a landmark Dartmouth Atlas study found that Medicare spending per patient varied by more than 100% between regions — not because patients were sicker in expensive regions, but because doctors in those regions ordered more tests, more procedures, and more interventions. The variation correlated with physician supply and payment structure, not patient need.

Your doctor isn't trying to harm you. Your doctor exists inside an incentive structure that rewards activity over outcomes. Every unnecessary test generates revenue. Every extra appointment fills a slot. Every defensive referral avoids a lawsuit. The system doesn't select for malice. It selects for whatever the incentive points toward.

Why Character Doesn't Fix It

The tempting response is: just find an honest agent. A good doctor. An ethical advisor. A trustworthy lawyer. Surely character is the variable.

Character helps at the margins. It doesn't solve the structural problem. Because the principal-agent problem isn't about bad people. It's about systems that make good people act against their clients' interests without realising it.

The Moral Licensing Effect

Psychologist Anna Merritt found that people who have recently demonstrated their moral character are subsequently more likely to act in self-serving ways — a phenomenon called moral licensing. The financial advisor who genuinely believes he's a good person — who volunteers on weekends, who donates to charity, who coaches his kid's soccer team — is slightly more likely, not less, to recommend the high-commission fund. His moral self-image gives him unconscious permission to serve himself in domains where the moral stakes feel lower.

He doesn't think "I'm screwing my client." He thinks "this fund is perfectly fine, and I happen to benefit from recommending it." The self-serving behaviour is disguised, even from himself, by a moral identity that says "I'm not the kind of person who would do that." Cognitive dissonance smooths the conflict. The agent genuinely believes they're acting in your interest — which makes them more convincing, not less.

The most dangerous agents aren't the dishonest ones. They're the honest ones whose incentive structure has quietly corrupted their judgement without triggering their moral alarm system. You can protect yourself against a liar. You can't protect yourself against someone who believes their own conflicted advice.

Character is a variable. Incentive structure is a constant. When they conflict, incentive structure wins — slowly, invisibly, with a smile and a handshake and a genuine belief that the recommendation is sound.

Reading the Incentive Map

The defence against the principal-agent problem isn't suspicion — if you distrust every professional relationship, you'll never get medical care, legal help, or financial advice. The defence is incentive literacy: the ability to read how your agent is compensated and predict where that compensation will distort their recommendations.

The Three Questions

Before accepting significant advice from any agent, ask:

"How do you get paid?" This is the most important question in any professional relationship, and the one most people never ask. Fee-only financial advisors (paid by the hour or by percentage of assets) have different incentives than commission-based advisors (paid per product sold). Salaried doctors have different incentives than fee-for-service doctors. Understanding the payment structure tells you where the misalignment lives.

"What would you recommend if you weren't being compensated for it?" Most honest agents will give a straight answer to this question — and the gap between their actual recommendation and their hypothetical recommendation reveals the incentive distortion. If the answer is "I'd recommend the same thing," you have some evidence of alignment. If the answer shifts, you've found the divergence.

"What would you do if this were your money / your body / your house?" Levitt's study showed that agents behave differently with their own assets. Asking this question explicitly invites the agent to switch from their professional frame (optimised for their incentives) to their personal frame (optimised for outcomes). It doesn't guarantee honesty. It changes the lens.

Align the Incentives, Don't Just Trust the Person

The structural fix is better than the interpersonal fix. Whenever possible, choose agents whose compensation aligns with your outcome:

Fee-only financial advisors over commission-based ones. They earn the same whether you buy an index fund or a loaded mutual fund.
Flat-fee lawyers over hourly lawyers for defined tasks. An hourly lawyer benefits from complexity. A flat-fee lawyer benefits from efficiency.
Outcome-based compensation wherever available. A real estate agent who earns a bonus above a target price has an incentive to hold out. An agent on a flat percentage has an incentive to close fast.

You can't eliminate the principal-agent problem. But you can choose the version of it that points your agent's self-interest in the same direction as your wellbeing.

The Agent You Don't See

The most consequential principal-agent problem in your life isn't with your doctor or your financial advisor. It's with your employer.

You are the agent. Your employer is the principal. And the misalignment runs in both directions.

Your employer hired you to produce value for the organisation. You accepted the job to produce value for yourself — income, experience, meaning, status. These overlap substantially. But they diverge in specific, predictable places.

Your employer wants you to work on the highest-priority project. You want to work on the project that develops the most transferable skills. Your employer wants you to stay late for the launch. You want to go home to your family. Your employer wants you to hire the candidate who's best for the role. You want to hire the candidate who won't outshine you.

Most office politics is the principal-agent problem playing out between employer and employee, each optimising for their own outcome while claiming to optimise for the shared one. The employee who says "I'm doing what's best for the team" while padding their resume. The manager who says "this is best for the company" while protecting their department's headcount. The executive who says "this is about shareholder value" while designing a compensation package that rewards them regardless of shareholder value.

None of these people are lying. They've just confused their map for the territory — mistaken their own incentives for objective reality, because the incentives are constant and the objectivity is effortful.

Greg Smith left Goldman Sachs and wrote a book. Goldman issued a statement calling his account misleading. The news cycle moved on. Nothing structurally changed at Goldman, or at any other investment bank, because the principal-agent problem isn't a culture issue that a viral op-ed can fix. It's an architecture issue that requires redesigning how agents are compensated, monitored, and held accountable.

Smith was right about what he saw. He was wrong about why it existed. It wasn't because Goldman had uniquely bad people. It was because Goldman had the same incentive structure as every other institution where agents act on principals' behalf — an incentive structure that systematically rewards agents for serving themselves while maintaining the appearance of serving you.

The uncomfortable truth isn't that your agents are dishonest. It's that honesty isn't sufficient protection against a structural force that operates below the level of conscious intention. Your financial advisor can be a wonderful person and still put you in the wrong fund. Your doctor can be deeply caring and still order the unnecessary test. Your real estate agent can be a friend and still rush the sale.

They're not failing you. The system is working exactly as designed — for them. The only thing that changes the outcome is changing the design. And the only person who will ever change the design of your relationships is you, because you're the only one in the room whose incentive is to get it right.