Tuesday, December 31, 2013

A Game of Institutions Book 3: The Sanctioners' Dilemma

Recall from last chapter the proposition I encouraged you to consider: though the reasoning may be fallacious or rooted in vice, people may be motivated by payoff asymmetries and previously-accumulated earnings. Let's consult our simplified table of game archetypes to see where we might look next.

Cooperative Non-Cooperative

In S&T, recall that euvoluntary constitutions require no ex post enforcement to maintain cohesion. I used the example of low-exit-cost partnerships to illustrate these. I think that some of the Ostroms' examples would qualify too. But once we start shuffling around the box above, we move away from the euvoluntary conditions toward... something else. I do hope you're willing to forgive me for staying in the "symmetrical" row rather than the "cooperative" column, but since I just got done posting about the Amish, I trust you'll be willing to bear with me.

Exclaves fascinate me. Pseudo-enclaves even more so. In the case of the Amish, we've got an extra layer of governance embedded (loosely) in the ordinary federal structure. But here's the fascinating bit. Unlike a Olsonian stationary bandit, the Amish lil' leviathan generates inalienable quasi-rents. After all, there's really no technology that allows a taxing authority to expropriate the warm sense of belonging to a community or the fellowship of hearth and kin. 

But, ha ha, there's quite obviously a private incentive to shirk. Amish individuals can obtain higher private utility by flaunting the rules of their society. Of course, since the SPNE of this is for everybody to shirk, it is in the constitutional interest to have an enforcement mechanism. And what could be better than yanking the club card? Credibly threaten to revoke the quasi-rents of family and community, and hey-yo, the decision calculus changes. Elementary economics.

Now, I'm not 100% sure that this sits squarely in the corner of the PD box, but at least in terms of expected utility, it's my understanding that Amish communities are marginally more egalitarian than, say, old world Europe. I'm sure that some families are more influential or wealthier than others, but the Amish Church itself lacks the hierarchy that characterizes the Catholic or Orthodox Churches. The local bishop is the community leader, but it's an unpaid, elected post. I'm not enough of an expert to know whether or not hereditary elites are common, but at least nominally, the Amish value the virtues of thrift, temperance, piety, prudence, and industry. Good work makes good people. An expression of Nicomachean ethics. Again, from an outsider's perspective, it seems as if envy within the community isn't as big a component of private motivation as it is among the English.

So yes. Shunning is coercive. But it has to be in a game that generates inalienable quasi-rents. No other punishment option is either appropriate or in the choice set. Besides, if it was rough enough for Socrates to guzzle hemlock, it's severe enough to keep wayward flock members in line.

A parting note: I had originally planned to calculate and share with you the comparative statics for all these games, but soon I discovered that LaTeX editors are both a pain in the rump to use and don't display for butt in my reader, so if you were looking forward to more "math", I'm afraid I'll have to disappoint you. Sorry.

Edit: As AG notes, bishops are (can be?) selected by lot (sources I've found in a quick Google search vary, so maybe it varies by community) and more importantly, the decision to shun cannot be made lightly. There's a risk of schism, and even if there weren't, there are increasing quality-adjusted returns to network size. If the strength of your organization depends on good members, you'd want to be dang sure that truncations are well and truly in the group interest. Over-pruning a tree is as fatal as drought.


  1. Am I missing something, or just behind the times? I thought "cooperative" games implied enforceable contracts, whereas "non-cooperative" games were the ones with self-enforcing agreements. Thus implying that coordination games are non-cooperative, as they require no external enforcement mechanism for players to arrive and stick at equilibria.

    Mike's definition of euvoluntary exchange assumes "conventional" ownership (hence no external enforcement per se), but appears to apply only to items that can be exchanged synchronously. Or do I have a bunch of reading to catch up on?

    1. See Fundenberg&Tirole "Game Theory" Chapter 10 (page 397 in the 1991 edition) for a discussion of cooperative and non-cooperative games. My table is consistent with their definition.

      And "conventional ownership" seems to invoke at least a Hayekian rule of law. Looking at it this way, I think it's reasonable to say there's some external enforcement by custom if not by any particular, specific agent (Vol 1 of Law, Legislation, and Liberty).

      But this whole series is drawn from the Shughart/Thomas piece, which borrows the term "euvoluntary" without the same specific definition Munger uses. You can find the paper in the "S&T" link above.


Do you have suggestions on where we could find more examples of this phenomenon?