See the introduction to and explanation of this series of posts.

“Agent-Based Models,” Scott de Marchi and Scott E. Page

I remember a time– 8 years ago? 10?– when I was briefly convinced that we were headed for a really huge convergence and transformation in the social sciences: as experimental economics and behavioural psychology made more progress on realistic understandings of decisionmaking; as everyone came to take seriously the balance among self-interest, love of dominance or power, reciprocity, in-group altruism, and altruistic punishment for norm violation that seems to run in our evolved psyches; as the traditional theory of spontaneous order got supplemented by the understanding of emergence and by the generation of work that included Axelrod and Ostrom; and as normative theory absorbed (and engaged with) all of these lessons, as two generations ago it absorbed (and contested) develoments in rational choice theory and economics; we might really start to see a Grand Unified Theory of human social orders in our lifetime. And part of what convinced me of that was exposure to the overlap of complexity theory and agent-based modelling associated with the Santa Fe Institute; that looked to me like the point where all these things would come into contact.

(As an aside: the people who talk about such unification these days are probably more likely to be believers in the power of Big Data. I’m not one. Looking for your keys under a 100,000 watt laser-powered streetlamp doesn’t solve the structural problem with only looking under streetlamps because that’s where the light is.)

I don’t know that I changed my mind about that view so much as I stopped thinking about it. Years later, all of those literatures and approaches seem more entrenched in their separateness. But anyway, this article did a great job of reminding me about the potential power of agent-based modelling. (And, contrary to my ex ante worries, it’s not at all technically dense; it’s very accessible to the nonspecialist.) De Marchi and Page are very modest about agent-based modelling, taking care in particular to talk about it as a complement to game theoretic modelling– less committed to equilibrium analysis or an assumption of optimizing behavior, more able to model social spaces and interactions as well as complicated decision rules, but therefore sometimes lacking in the power game theory has to think about underlying equilibria. It seems to me that game theory is instead a special case of agent-based modelling, but that de Marchi and Page find it impolitic to say so. (But that’s sheer speculation on my part.) Even with modesty and possible understatement, though, they make clear how useful and illuminating agent-based modelling can be, and how much it can add alongside game-theoretic and microeconomic/ rational choice approaches.

As far as I’m concerned this piece comes awfully close to the Platonic ideal of what one of these ARPS should look like and accomplish– and it’s reminded me to buy Page’s book Complexity and Diversity, which has been on my list for a long time.

Next up, another part of the discipline I used to be on top of: indigenous peoples’ politics. “Tribal-State Relations” by Laura E. Evans.