Humans have long used norms to great advantage to coordinate behavior. Each norm requires or prohibits certain behavior in certain situations, and the norm system requires that others who notice norm violations call attention to those violations and coordinate to discourage or punish them.

This system is powerful, but not infinitely so. If a small enough group of people notice a minor enough norm violation, and are friendly enough with each other and with the violator, they often coordinate instead to not enforce the norm, and yet pretend that they did so. That is, they let cheaters get away with it.

To encourage norm enforcement, our social systems make many choices of how many people typically see each behavior or its signs. We pair up police in squad cars, and decide how far away in the police organizational structure sits internal affairs. Many kinds of work is double-checked by others, sometimes from independent agencies. Schools declare honor-codes that justify light checking. At times, we “measure twice and cut once.”

These choices of how much to check are naturally tied to our estimates of how strongly people tend to enforce norms. If even small groups who observe violations will typically enforce them, we don’t need to check as much or as carefully, or to punish as much when we catch cheaters. But if large diverse groups commonly manage to coordinate to evade norm enforcement, then we need frequent checks by diverse people who are widely separated organizationally, and we need to punish cheaters more when we catch them.

I’ve been reading the book Moral Mazes for the last few months; it is excellent, but also depressing, which is why it takes so long to read. It makes a strong case, through many detailed examples, that in typical business organizations, norms are actually enforced far less than members pretend. The typical level of checking is in fact far too little to effectively enforce common norms, such as against self-dealing, bribery, accounting lies, fair evaluation of employees, and treating similar customers differently. Combining this data with other things I know, I’m convinced that this applies not only in business, but in human behavior more generally.

We often argue about this key parameter of how hard or necessary it is to enforce norms. Cynics tend to say that it is hard and necessary, while idealists tend to say that it is easy and unnecessary. This data suggests that cynics tend more to be right, even as idealists tend to win our social arguments.

One reason idealists tend to win arguments is that they impugn the character and motives of cynics. They suggest that cynics can more easily see opportunities for cheating because cynics in fact intend to and do cheat more, or that cynics are losers who seek to make excuses for their failures, by blaming the cheating of others. Idealists also tend to say what while other groups may have norm enforcement problems, our group is better, which suggests that cynics are disloyal to our group.

Norm enforcement is expensive, but worth it if we have good social norms, that discourage harmful behaviors. Yet if we under-estimate how hard norms are to enforce, we won’t check enough, and cheaters will get away with cheating, canceling much of the benefit of the norm. People who privately know this fact will gain by cheating often, as they know they can get away with it. Conversely, people who trust norm enforcement to work will be cheated on, and lose.

When confronted with data, idealists often argue, successfully, that it is good if people tend to overestimate the effectiveness of norm enforcement, as this will make them obey norms more, to everyone’s benefit. They give this as a reason to teach this overestimate in schools and in our standard public speeches. And so that is what societies tend to do. Which benefits those who, even if they give lip service to this claim in public, are privately selfish enough to know it is a lie, and are willing to cheat on the larger pool of gullible victims that this policy creates.

That is, idealists aid cheaters.

Added 26Aug: In this post, I intended to define the words “idealist” and “cynic” in terms of how hard or necessary it is to enforce norms. The use of those words has distracted many. Not sure what are better words though.