The behavioral economist Ernst Fehr and his colleagues have shown that incorporating inequality aversion into their models allows them to predict whether people cooperate or be selfish in economic situations. They defined two distinct forms of inequality aversion: envy and guilt.

In some situations people will be willing to sacrifice

People strongly dislike having less than what others have – that is envy, and they less strongly dislike having more than what others have – that’s guilt. Both of these are examples of reference dependence, people care not just about absolute outcomes but outcomes relative to some reference point. Their inequity aversion model implies that in some situations people will be willing to sacrifice some of their own money in order to reach a more equitable outcome.

Inequity aversion can explain some biases and people’s choices like rejection of unfair offers in the ultimatum game, but inequity aversion by itself can’t explain the prevalence of cooperation. To get an intuition about why imagine that there are group of people interacting in a market. Some are very inequity verse but some art they’re just selfish. The inequity averse people will be at a competitive disadvantage compared to the selfish people, those who act to maximize their own interest without regard to cooperation.

Selfish people are willing to accept unfair divisions of resources as long as the unfairness works to their benefit. So, a selfish person would do better than an inequity of person when the deck stacked in their favor. without some mechanism for enforcement inequity aversion can’t explain the persistence of cooperation in the equilibrium state of the markets.

We want to be nice to other nice people

Another explanation is that of reciprocity. I’ll define reciprocation here in terms of intentions. When we interact with others we make judgments about their motives and then we act in the same way, we reciprocate good intentions leading to cooperation. In one such model reciprocity and economic games, different choices are characterized as either involving nice or mean intentions.

Players gain utility not just from their own outcomes but from the product of their own intentions and the other player’s intentions. If the other player is nice and it’s good to be nice if the other players mean that’s good to be mean. This sort of model implies that our behavior in a social interaction depends on our inferences about someone else’s character. We want to be nice to other nice people.

Niceness can be mutually reinforcing leading to cooperation, but meanness can also be mutually reinforcing. As such reciprocation, like inequity aversion, provides a necessary precondition for cooperation, but it is insufficient. We still need some enforcement mechanism so how can cooperation be enforced? In humans altruistic punishment takes a different form.

Let’s suppose you’re watching two people play a trusting game. The investor offers everything to the trustee, but the trustee keeps everything, the trustee breaks the social norm of cooperation. let’s also suppose that you have the opportunity to spend five dollars of your own money to punish the uncooperative trustee. If you spend $5 the trustee will lose $20 from their ill-gotten gains.

Altruistic punishment is powerful, but expensive

People are willing to engage in such altruistic punishments in both laboratory and real-world situations.  They want to enforce social norms, they get angry at the cheaters and that emotion pushes people to take actions against their own self-interest. When people engage in altruistic punishment there are two primary effects:

  • the first is straightforward it corrects deviant behavior when a partner fails to cooperate – for example the punished fish starts cleaning correctly or the punished human cheater cooperates in the next game;
  • the second effect is more subtle altruistic punishment increases one’s confidence that an unknown partner will cooperate in the future – if people can’t be punished for failing to cooperate in social cooperation unravels with even a few bad apples.

People see antisocial behavior that goes unpunished and they lose confidence in others that leads to a cycle of diminishing cooperation over time. Punishment is possible a few good apples can enforce cooperation even in very large groups.

Altruistic punishment is powerful, but expensive. It requires time and energy to confront someone who is disturbing their social group or is shrinking the duties in the workspace. Even if you have good intentions you might not find altruistic punishment worth the effort.