Agent-Based Computational Models - A Formal Heuristic for Institutionalist Pattern Modelling?

Institutionalist economists have always been criticizing the neoclassical way of studying the economy, especially because of its obsession to a very strict and flawed formalism. This formalism receives critique also from advocates of agent-based computational economic (ACE) models. The criticism seems to be similar to that of institutional economists. Although some authors consider ACE models to belong to a completely new way of thinking about economics, many concepts of ACE have been anticipated by institutionalists: Although using a different vocabulary, ACE proponents speak about cumulative causation, realistic agents, explanatory models, dynamic relations among individuals and the necessity to see the economy as an systemic whole rather than from an atomistic perspective. Consequently, the emergence of the ACE framework may not be left unconsidered by institutionalist economists. This paper investigates the consistency of ACE models with the institutionalist research program as defined by Myrdal, Wilber and Harrison and other original institutionalists and discusses whether ACE models can be a useful heuristic for institutionalist "pattern modelling". I study the ability of ACE models to provide a holistic, systemic and evolutionary picture of the economy, the conception of agents in ACE models, and ask whether they can help to understand the social stratification of a society with its power relations. I also compare ACE models with earlier attempts to formalize institutionalist analysis, e.g. by Bush and Elsner (Theory of Institutional Change), Hayden (Social-­Fabric-Matrix) or Radzicki (System Dynamics).