A Game Theoretical Approach to Institutions for Collective Action

ICYMI, the Shughart and Thomas piece (here) really is quite good indeed. As an eagle-eye view of the literature on institutional choice, they cover quite a bit of ground, from Leeson's œuvre on pirate compacts to Buchanan, Tollison, Shughart & al.'s club goods theories to the Ostroms' contributions on commons governance to the formal high constitutional theories of Tiebout, Buchanan, and Hirschman. If you wanted a good place to find the peaks in the constitutional formation literature, this is it.

But that's not what piqued my interest. What they've done is propose a fascinating taxonomy that runs very closely to an idea that's been rattling around my attic for at least a year. You see, underpinning competing theories of constitutional formation are typically game theoretic models. Usually what you'll see are claims that constitutions help people solve coordination problems. But occasionally, someone will forward the proposition that what's being solved is a repeated PD.

Well, which is it?

By proposing a euvoluntary-voluntary-coercive continuum, S&T tell the discerning reader "it depends". And that furthermore, the degree to which the game departs from balanced cooperation determines its euvoluntarity. Why is this important? A few reasons.

1. Euvoluntary things, be they exchange or institutional choice, tend not to offend. Both characteristics of fairness and justice are satisfied when arrangements are euvoluntary, so parties external to the arrangement are unlikely to interfere with the terms of the constitution, at least on behalf of the signatories. 
2. Euvoluntary institutional arrangements should be stable. No need for ex post enforcement suggests that everyone is happy, there's no incentive to shirk, and therefore there should be no reason to dissolve the arrangement. (Hint to Adam G: this is where you come in)
3. Voluntary arrangements are better than none at all, but they could signal some durable payoff asymmetries. Moving from a coordination game to a Battle of the Sexes game means that the coordinating mechanism necessarily privileges some parties more than others. Hence the ex post enforcement as the constituents might want to jockey to obtain the greater payout. This is your Hume. This is your McCloskey. Oh man, I'm going to get in trouble for that, aren't I?
4. To the extent that voluntary arrangements lead to the generation of local elites, dissolution (or reform) pressures may mount over time from within and paternalistic expropriation pressures may mount from without. This is where I think Tiebout has the most bite, but it has to be a Tiebout tempered with Coase: departing or adjusting a voluntary constitutional arrangement is unlikely to be cheap.
5. Coercive constitutions are where your Hobbes, your North, your Olson work. This is pretty familiar territory, and I think it's probably fair to say that the assumption of non-cooperative games underpins what most folks think of the organizing principles of nation-states. And I think that's probably a fair model. A Hawk-Dove game is pretty much the proper way to model crime.
6. More to the point, this is the proper way to model revolution, or at least to answer the question why some organizations are stable while others topple almost as soon as they are established. Good comparative institutional analysis can be illustrated clearly and concisely with basic game theory.

So, that's why I really like the paper. It puffs some good wind into the sails hoisted by Buchanan in his '63 address at the Southerns. Public Choice scholars should be able to generate several handsful of hypotheses by skimming those points above. As for me, I'd be tickled to just expand this post into a piece. I think there's some meat there.