Even if you have good intentions you might not find altruistic punishment worth the effort, but think about what happens when someone misbehaves in your social group or your workspace.

Corrective actions don’t have to come from one person in isolation

Often multiple people will be affected by their misbehavior, therefore corrective actions don’t have to come from one person in isolation, but can come from a group in coordination. Let me give you a real world example: within the forest of Ethiopia local groups manage their forests as a common good, for example they set their own rules about how individual members of their groups can harvest wood for fuel.

This creates a classic problem of cooperation, each member of the local group might benefit from collecting extra wood for themselves, but everyone harvested would indiscriminately the common resource would disappear. So these groups set up patrols, they go through the forest and they look for people who were taking more with them they should, but that’s quite a lot of work.

Spending time wandering through the forest and looking for those cutting down too much wood reduces the time available for meeting the needs of one’s own family. Researchers ran simple two player laboratory games with the members of 49 different local Ethiopian groups and they used data from those games that characterize each person’s attitude toward cooperation.

The people who most benefit to their communities

About 50% of the people were what they called conditional cooperators – if the other player cooperated they cooperate right back, but if the other player doesn’t cooperate then they behave selfishly. About 10% were free riders – they try to take advantage of the other players generosity. The remaining people were either consistently altruistic or had a more variable set of attitudes.

Then the researchers looked at each of those 49 groups and compared each group’s real-world resource management to how its members play the games. They found that groups whose members were mostly conditional cooperators had the most trees remaining in the forests, they had managed their local force most effectively for the common good. Why? Because the people who were conditional cooperators in the game tended to go out on patrol, they were the ones looking to enforce the social norm of cooperation.

It’s particularly striking the people who most benefit to their communities were the conditional cooperators in the game not the altar us who always cooperated and definitely not the free riders who never cooperated. Conditional cooperation involves cooperating when cooperating is deserved, but punishing when that’s necessary. Real world groups that include a substantial portion of conditional cooperators are better able to manage their shared resources.

Altruistic punishment helps

Let me summarize this key point: if people are allowed to coordinate their punishment of non-cooperators, and even a small group can be effective in shaping the long-term trajectory of their society pushing it toward general cooperation, even if there still a small minority of persistent misbehaviors. Communication is critical for the coordination, the Nobel laureate Elinor Ostrom was a pioneer in the economic modeling of how common resource, like a shared forest in Ethiopia can be managed successfully by local cooperation.

She and her colleagues showed that all touristic punishment works best when people communicate freely. Communication, helps people coordinate their punishment making conditional cooperation more effective and communication can be used as a low-cost social signal. Often, we don’t want to punish violators of social norms immediately that’s too effortful and it carries too much for social cost, so a communicated threat provides an often effective low-cost substitute.

Given how communication can make coordination more effective one might expect that altruistic punishment works better in homogenous groups than in diverse groups, but that’s not true. Field experiments have shown that altruistic punishment helps in diverse groups but has little effect in homogenous groups, perhaps because the threat of punishment carries more weight in diverse groups but less weight in groups that are held together by kinship and social ties.