Corporations Defect Where Individuals Cooperate

Robert Axelrod's work on the iterated prisoner's dilemma is one of the more genuinely useful results in social science: cooperation emerges without altruism, provided the game repeats and the players expect to meet again.

What is less often noticed is that the conditions producing that result are conditions human individuals satisfy and corporations do not. The same game theory that explains why people cooperate explains why firms defect.

Why Individuals Cooperate

We are reasonably well equipped for the iterated prisoner's dilemma and the tragedy of the commons, and the equipment is evolutionary rather than moral.

We evolved in small groups with repeated interactions among the same people. In that environment, defection is a bad strategy: you will meet this person again, they will remember, and reputation propagates through a group small enough for everyone to know everyone. Under those conditions the game rewards something like tit-for-tat, and we came pre-loaded with the instincts that implement it — a sense of fairness, a taste for reciprocity, discomfort at cheating, and considerable appetite for punishing cheaters at cost to ourselves.

Commons problems yield to the same equipment. Small groups sharing a resource can and routinely do sit down, agree limits, and enforce them informally. Ostrom documented this happening across many cultures without any external authority.

None of this required anyone to be good. It required repetition, memory, and a small enough group.

Why Corporations Do Not

Now ask what a corporation brings to the same game.

It is not that firms never cooperate. Two companies supplying each other under repeated contracts cooperate reliably, because that is a genuinely iterated game with reputation attached. Where the structure supports cooperation, corporations cooperate.

But when a firm sees an opportunity to harm its environment for gain — to externalise a cost onto people who are not party to any transaction with it — it takes the opportunity far more readily than an individual would.

The reason is not that corporate executives are worse people. It is that the instinct which restrains individuals has no corporate equivalent.

Our reluctance to harm the surroundings was built in an environment where the surroundings were our own small tribe. Anything that hurt other people hurt us, so the instinct never needed to distinguish self-interest from group-interest — they coincided. That coincidence is what our moral equipment encodes.

A corporation is not governed by that psychology. It is governed by interest, and strictly speaking harming others can benefit it. There is no evolved reluctance because there was no evolution — no ancestral corporations that failed because they poisoned their neighbours.

The Structural Point

This is one instance of a more general claim: organisations are steered by interests rather than by the individuals nominally heading them.

A chief executive with an ordinary human conscience presides over an entity whose decision procedures do not have one. He can override them occasionally, at some personal cost. He cannot make the entity conscientious, because the entity is a set of incentives and reporting lines, and those will still be there when he leaves.

Which is why appeals to corporate responsibility so reliably underperform. They address the psychology of the people inside, and it is not the psychology of the people inside that is producing the behaviour.

What Would Actually Work

If the diagnosis is right, the remedy is not moral exhortation but restoring the conditions that make defection unprofitable — the conditions individuals get for free from small-group life.

That means making the game genuinely iterated for firms: liability that persists, reputational effects that attach and stay attached, and counterparties with the standing to retaliate. It means making the harmed parties into players rather than externalities, since a cost imposed on someone with no move in the game is exactly what the corporate structure is optimised to find.

Ordinary self-interested firms in such a configuration behave far better than well-intentioned firms in the current one. That is the same conclusion the prisoner's dilemma reaches about people, and there is no reason it should stop applying at the corporate boundary.