I Don’t Believe in Ability

I don’t believe in ability

— design notes from a consequentialist.

In my last semester of undergrad, I chose the principal-agent problem as the topic of my capstone essay. The case was Daewoo Shipbuilding. The National Pension Service, which manages the nation’s retirement funds, had lost over 200 billion won on the company’s stock and bonds, and the two institutions were burning legal fees fighting each other. The moment a principal entrusts work to an agent, costs arise. They arise if you monitor, and they arise if you don’t.

My conclusion then was simple. Agency costs are a constant. They cannot be eliminated. And if they cannot be eliminated, a growing total pie must offset them. A firm must grow by more than its agency costs.

Years later, the thesis is unchanged. Only the resolution seems to have sharpened.

I pursue meritocracy.

Yet I do not believe in ability itself.

It sounds like a contradiction, but it is the cleanest position. Ability is not observable. Ability is a label we attach retroactively after seeing results. We look at someone who succeeded and say “no wonder — talented,” at someone who failed and say “incompetent.” Whether Daewoo Shipbuilding’s management was capable or incapable was revealed only after the results blew up. The moment you mistake that label for a substance, proxy indicators like school names and résumés begin to impersonate ability.

So I don’t open up hearts and look inside. I look at results.

That’s too abstract, so let’s cut the problem down small.

There’s a kid at school full of ambition. That kid raises a hand to volunteer. What should you do?

Let them.

There is no need to verify whether altruism exists in that kid’s heart. It isn’t observable anyway. What matters is the result that a hand went up, and the result that the volunteering gets done. What’s wrong with being ambitious? The service happens.

And here is where it gets real.

The moment the hand went up, the kid stepped into a trap.

The kid probably came in to collect something — volunteer hours, a line in the school record. But once aboard this orbit, the volunteering must be done. Whatever the motive was, the structure makes the kid serve. The kid benefits, and the board gains the service. Nobody loses.

A bad trap makes its participants lose, and breeds resentment. When resentment piles up, people dig escape holes. A good trap benefits its participants, so there is no reason to flee. If anything, they walk in deeper on their own. Voluntariness is left intact; only the direction is set.

An action that began as calculation becomes habit through repetition, and once it is habit, at some point the grain changes. People look at their own behavior and adjust their attitudes backward to match.

“Huh — I keep doing this. I must be the kind of person who likes this.”

People don’t volunteer because they’re good. They become good because they volunteer. In a properly designed society, that is.

Surveillance does not change people. Remove the surveillance and they revert. So you must surveil forever. Incentives don’t change people either. Cut the reward and the behavior stops.

A good trap changes the person. Past a certain point, even without the trap, the kid heads in that direction on their own.

The final form of a system is one that keeps running even after the system is gone.

So a person who decides to trust only results may inevitably become a designer. A person who believes in motives has no need for systems — just entrust things to good people. But a person who has decided not to open hearts must make the system render the verdict instead of the heart. At that moment one question remains: with what will you measure results, and where will you make them flow?

The architect’s job is not to fix people. It is to carve the terrain so that ambition lands as good output. A board where people walk in following their own interest, land on a virtuous orbit — and, in time, become genuinely good.