When working on the roadmap on ICT for governance and policy modelling, I started considering these two domains as a bit juxtaposed, coming out of complicated committee negotiation rather than sharing a well defined identity. They seemed, at first sight, as simply two domains that shared the fact of “not being dealt with by other program priorities.
This is quite typical of EU research negotiation, where internal discussions sometimes obscure the genuine importance of a theme. As one interviewee once told me, we’re working on a solution in search of a problem.
Yet I now realize how much these two domains share, and how much the actual culture of the people working on these themes is similar. There is a kind of anthropological similarity between the two communities.
Basically, both agent-based modellers and gov20 people share a refusal of traditional IT and economic thinking. On IT: they are against “IT automation” , where IT subsitutes humans through online services, decision support systems, knowledge management tools. Instead, they share the view of IT as a human centric tools, in the sense of Engelbart and Ciborra.
In economics, they are against general equilibrium theory. which assumes that all individuals are rational and that market tools are the most efficient tool to reflect preferences, and that future systems behavior can be induced by the past in a linear way.
So, the reality can be described by a 2X2 matrix


Traditional / Reductionist / Modern

Non-traditional / Complex / Post-modern


IT automating

The more users, the lees performant the application is

User needs are predictable ex-ante

IT augmenting

Network effects Applications get better
the more people use it

User-needs are dynamically and iterative integrated into development


General equilibrium theories

Humans considered as rational and average

Linear models can predict future

Agent-based modelling

Human considered as diverse and
accounting for social influence

Non-linear effects are generated from agent interaction and system dynamics

Instead, both these two communities share two principles: a) that human are diverse, non reduceable, and not susbtitutable by machines and b) that complex, non linear results can be generated by very simple network interaction.
In my next post, I will dig deeper into how these two principles are adopted by the two communities.

UPDATE: another principle they share is the notion of “emergence”. Both gov20 and agent-based modeling focus on using ICT to identify and capture emergent behaviour ex post, rather than codifying it ex-ante