The Coalition Politics Hypothesis

Game theories let us analyze precise models of social situations. While each model leaves out much that is important, the ability to see how an entire set of payoffs, info, and acts work together can give powerful insights into social behavior. But it does matter a lot which games we think apply best to which real situations.

Today the game most often used as a metaphor for general social instincts is the public goods game, where individuals contribute personal efforts to benefit everyone in a group. This is seen as a variation on the prisoner’s dilemma. With this metaphor in mind, people see most social instincts as there to detect and reward contributions, and to punish free-riders. Many social activities that on the surface appear to have other purposes are said to be really about this. Here, “pro-social” is good for the group, while “anti-social” is bad. Institutions or policies that undercut traditional social instincts are suspect.

While this metaphor does give insight, the game I see as a better metaphor for general social instincts is this:

Divide The Dollar Game … There are three players … 1, 2, 3. The players wish to divide 300 units of money among themselves. Each player can propose a payoff such that no player’s payoff is negative and the sum of all the payoffs does not exceed 300. … Players get 0 unless there is some pair of players {1, 2}, {2, 3}, or {1, 3} who propose the same allocation, in which case they get this allocation. …

It turns out that in any equilibrium of this game, there is always at least one pair of players who would both do strictly better by jointly agreeing to change their strategies together. …

Suppose the negotiated agreements are tentative and non-binding. Thus a player who negotiates in a sequential manner in various coalitions can nullify his earlier agreements and reach a different agreement with a coalition that negotiates later. Here the order in which negotiations are made and nullified will have a bearing on the final outcome. … It is clear that coalitions that get to negotiate later hold the advantage in this scheme. (more)

That is, most social behavior is about shifting coalitions that change how group benefits are divided, and social instincts are mostly about seeing what coalitions to join and how to get others to want you in their coalitions. Such “social” behavior isn’t good for the group as a whole, though it can be good for your coalition. Because coalition politics can be expensive, institutions or policies that undercut it can be good overall.

In this view of social behavior, we expect to see great efforts to infer each person’s threat point – how much they and a coalition would lose if they leave that coalition. We also expect even greater efforts to infer each person’s loyalty – what coalitions they are likely to prefer and help. And we expect great efforts to signal desirable loyalties and threat points. When shifting coalitions are important, we expect lots of efforts to go into seeing and changing the focal points people use to coordinate which new coalitions form, and to seeing who will be pivotal in those changes.

At a meta level, people would also try to infer what other people think about these things. That is, folks will want to know what others think about various loyalties, threat points, and focal points, and in response those others will try to signal their opinions on such things. In other words, people will want to know how well others can track and influence changing fashions on these topics. At a higher meta level, people will want to know what others think that still others think about these things, i.e., they’ll want to know who is seen to be good at tracking fashion. And so on up the meta hierarchy.

When people talk, we expect them to say some things directly and clearly to all, to influence overall focal points. But we expect many other messages to be targeted to particular audiences, like “Let’s dump that guy from our coalition.” When such targeted messages might be overheard, or quoted to others, we expect talking to be indirect, using code words that obscure meanings, or at least give plausible deniability.

A social world dominated by shifting coalitions would spend modest efforts to influence temporary policies, such has how to divide up today’s spoils, and more efforts on rare chances to change longer term policies that more permanently divide spoils. Even more effort would be spent on rare chances to change who is possible as a coalition partner, For our forager ancestors, killing someone, or letting a new person live nearby, could change the whole game. In a firm today, hiring or firing someone can have similar effects.

This view of social behavior as mostly about shifting coalitions raises the obvious question: why doesn’t most social behavior and conversation seem on the surface to be about such things. And the obvious homo hypocritus answer is that we do such things indirectly to avoid admitting that this is what we are doing. Since coalition politics is socially destructive, we have long had social norms to discourage it, such as the usual norms against gossip. So we do these things indirectly, to get plausible deniability.

This can explain why we place such a high premium on spontaneity and apparent randomness in conversation and other leisure behavior. And also why we seem so uninterested in systematic plans to prioritize our efforts in charity and other good causes. And why we drop names so often. When we manage our shifting coalitions, we prefer to stay free to quickly shift our conversations and priorities to adapt to the changing fashions. If you ever wonder why the news, public discourse, and academia seem so uninterested in the topics most everyone would agree are really important, this is plausibly why.

GD Star Rating
loading...
Tagged as: ,
Trackback URL:
  • arch1

    1) I think that people actually talk a lot about coalition politics – much gossip explicitly concerns (or implicitly consists of) precisely that. That said, the fact that ‘gossip’ has such negative connotations tends to support your thesis.

    2) If the desire to keep coalition options open were the primary reason stayed away from ‘serious’ topics in conversation, wouldn’t media and content providers still see a big demand for information on serious topics which could be consumed in relative privacy?

    • http://overcomingbias.com RobinHanson

      On #2, the theory that people are hypocritically pretending to care about what they talk about, but really use that talk to coordinate coalition politics, doesn’t at all predict they will try in secret to really care about those topics.

  • Tom

    If you ever wonder why the news, public discourse, and academia seem so uninterested in the topics most everyone would agree are really important, this is plausibly why.

    Yes, and news, public discourse, and academia are also significantly controlled by elite groups with their own agendas.

  • IMASBA

    “This view of social behavior as mostly about shifting coalitions raises the obvious question: why doesn’t most social behavior and conversation seem on the surface to be about such things. And the obvious homo hypocritus answer is that we do such things indirectly to avoid admitting that this is what we are doing.”

    Couldn’t it be that coalition politicking does happen but is given fairly low priority (that most of our efforts do go towards temporary policies of the entire group) because a) there has been selection in the past against people who were too busy with the cloak and dagger stuff instead of making preparations and allies against natural calamities and b) people, to some extent, choose to ignore coalition politicking impulses, not just because we fear punishment or condemnation from our culture but because that culture has been around so long and actually makes some good points (coalition politicking does indeed come with a price to the overall group and there’s no telling when you’ll be in the part of the group that pays that price) that we really do feel sorry, not just sorry for getting caught, if we break its rules, making abstaining from coalition politicking similar to abstaining from crime?

    • http://juridicalcoherence.blogspot.com/ Stephen Diamond

      Couldn’t it be that coalition politicking does happen but is given fairly low priority…

      You really think that coalition politicking has low priority? (The importance of coalition politics is Hanson’s assumption–which to me seems painfully obvious–not his conclusion.)

      • IMASBA

        Only halfway through does he start speculating about a society dominated by coalition politcking. I challenge the assumption that just because it exists and is important it dominates society, to me that seems more logical than assuming everyone but Robin and myself are really good at playing hidden agendas (unless we’re both autists shouldn’t we notice this stuff in our own lives if we looked closely enough if it really is what the vast majority of humans are doing?)

      • http://juridicalcoherence.blogspot.com/ Stephen Diamond

        (unless we’re both autists shouldn’t we notice this stuff in our own lives if we looked closely enough if it really is what the vast majority of humans are doing?)

        I take it Robin has noticed the same, as I think I have. But how could you have looked closely enough, when you just now learned of the theory?

      • IMASBA

        I can look back in hindsight can’t I? In any case I think it’s healthy to stay skeptical about claims that supposedly “everyone” does something all the time that you’ve never done, while you’re a normal person with a normal life as far as you know.

  • IMASBA

    “If you ever wonder why the news, public discourse, and academia seem so uninterested in the topics most everyone would agree are really important, this is plausibly why.”

    Seriously? This sounds like one of those “90% of all men/women/rich people, etc…” articles in some bad magazine where they try to convince you that 90% of the members of the group you belong to do or think something even though you knwo you, and other members of the group you know don’t do or think that.

    No, the reason why AMERICAN news and public discourse seem uninterested in “important” topics is because there are better ratings to be gotten with other material and for academia it’s a matter of a privileged elite not being able to connect with people from other groups because of the third world level of inequality between and separation of different groups of society (how many economists grew up in the ghetto?)

    • http://juridicalcoherence.blogspot.com/ Stephen Diamond

      No, the reason why AMERICAN news and public discourse seem uninterested in “important” topics is because there are better ratings to be gotten with other material…

      Hanson’s theory concerns why there are better ratings to be gotten with other material.

      • IMASBA

        It was my interpretation that he suggested the media actually are actively manipulating things very secretively to further the goals of their own coalition. What you seem to be saying is that the public select for media that give them information they can use for their own personal coalition building, they seem to be doing a rather bad job in that case, though of course it’s possible people are hardwired to subconsciously believe that even gossip about strangers they’ll never meet is useful for information for their own personal coalition building.

  • Yadal

    This is a fairly obvious nitpick, but to give the model greater accuracy you should take into account that those genetically related have a desire to keep family members in their coalition independent of the spoils- not generally a strong enough bias to overcome selfish preferences, but strong enough to be a noticable factor.

    • http://overcomingbias.com RobinHanson

      Of course.

  • http://bur.sk/en Viliam Búr

    Coalition shifting encourages pro-social behavior. Imagine a group of people who are unable to leave each other; for example prisonners or children in school. We would expect a strong pecking order, with highest-status members being cruel to everyone else to preserve the order.

    Now imagine a group of people where anyone is free to leave; for example an open-source software community. Suddenly it makes sense for the highest-status people to act nice; otherwise the remaining members could leave and create their own alternative community.

    It’s the threat of “voting by feet” that motivates higher-status people to act pro-socially. Removing the threat gives the higher-status people more power. So it makes strategic sense for them to create rules that make “voting by feet” more difficult: to encourage group loyalty and hostility towards outsiders.

    • http://overcomingbias.com RobinHanson

      Status hierarchies usually aren’t arbitrary – they are based on actual abilities to win fights, etc. Coalition politics is quite consistent with status hierarchies – see the world of chimps for example. With coalition politics, status depends more on your abilities in such politics, such as your abilities to read fashion, pick good allies, etc.

  • efalken

    Perhaps this is why people vote as if they are ‘groupish’, not ‘selfish’. Further, why coalitions have to be subtle, coalitions have to be plausibly protecting victimization and promoting egalitarianism as opposed to simply being selfishly advancing their parochial interests.

    • Douglas Knight

      Do coalitions have to be subtle? Do they have to be about victimization and egalitarianism?

      To me these seem like very recent developments, not eternal truths.

      • efalken

        Well, I should qualify that, because while historically disadvantaged groups can claim to be victims, teachers can’t. Unions, sugar producers, veterans, etc. are all pretty self-serving groups that talk about bigger virtues: the children, defending the country, etc. In the end mainly a pretext for self-serving rents.

        Coalitions work best when small, because it’s easier to identify ‘cheaters’. As coalitions are basically conspiracies, if the masses know that group X is in a conspiracy, they will squash them like a bug. In the old days the cant wasn’t today’s ephemeral virtues, but rather, things related to honor. They believed there was a natural aristocracy which had special privileges because they had special virtues granted by God. The Holy Roman Empire was really just a dynasty of the Habsburg family among others.

        Of course, you can subjugate a populace through force, as in colonialism, but then you need some sort of apartheid and a big advantage in technology. I don’t think those were ‘conspiracies’ because they were open.

      • http://juridicalcoherence.blogspot.com/ Stephen Diamond

        I don’t think coalitions must be conspiratorial; for one thing, a coalition might constitute a majority.

        But to form ruling coalitions, elites must represent their interests as the interests of the masses.

        This seems likely to be an “eternal truth.”

  • ParanoidAltoid

    For prisoner’s dilemmas, the outcome is that both players defect. But the best-for-everyone outcome is for both to cooperate. One “solution” to the prisoners’ dilemma is to step back and work out a deal such as “If you defect, you have to pay a fine.”

    So what is the solution to divide the dollar games?

    If we assume linear utility (as opposed to diminishing marginal utility), then it doesn’t really matter what the outcome is. If the shifting of coalitions is costly, then we’d want norms that determine how goods are distributed so no wasteful competition is needed. The norms could be totally arbitrary. Anyone who goes against distribution-norms would need to be punished.

    If we assume that the goods have diminishing utility, as a lot of things do, then perhaps a norm of equality would be useful, as it ensures both that everyone gets what they need, and sets a norm that prevents costly coalition-forming as mentioned above.

    • http://juridicalcoherence.blogspot.com/ Stephen Diamond

      Yes (to the last paragraph) but we don’t have the absolute psychological freedom to pick norms: equality norms work effectively when there’s something physical to match. (See Status inflation and deflation: Prestige, the essence of status, permits broad allianceshttp://tinyurl.com/lxokdf7 )

    • IMASBA

      Coalition forming and switching comes with a price to the overall group, so if you want to limit it you have to make it so that no one can ever get a really great reward (at least not without a really great contribution towards the overall group) and that damage to the overall group is borne by all members. Basically what most people would consider a fair and safe society.

  • Pingback: Recomendaciones | intelib

  • Pingback: Overcoming Bias : Careers Need Allies

  • Zvi Mowshowitz

    The NYC rationalist group tried out a version of this game, and we had a very fun and interesting night. It culminated in this new game design I’m now working on: http://thezvi.livejournal.com/147624.html

  • Pingback: Overcoming Bias : Next Step, Exogamy?

  • Pingback: Overcoming Bias : Who Wants Standards?

  • Pingback: Overcoming Bias : Trustworthy Telepresence