Simon’s (Lost?) Legacy in Agent-Based Computational Economics

The eighth in a series of excerpts from Minds, Models and Milieux: Commemorating the Centennial of the Birth of Herbert Simon.

Marco Castellani and Marco Novarese

In 1960 Herbert Simon proposed the use of computer simulations in order to enrich and improve economic modelling (Clarkson and Simon, 1960). Computer science was among the many fields to which he made contributions. Therefore, his interest in computer simulations should not be surprising. Simon is often quoted in the agent-based simulation literature, where bounded rationality is fully recognised and used as a starting point. And yet Simon’s approach to simulation seems to be quite neglected: there are very few citations to Clarkson and Simon’s paper. And a specialized journal such as JASSS never quotes that paper. Very few economists working in the field seems to know this paper, and no one has pursued his methodological approach. Indeed, Simon (2000) signals to economists to use computers simulations and build realistic theories.

This chapter aims at understanding the reason for this apparent paradox. At the same time we review the variety of approaches and interests Simon pursued. And it will be shown, again, that Simon’s aims and methodology are so different from even heterodox economists. The two points may be linked. Simon was interested in problem solving, often practical problems. For this reason there was no need and indeed no chance to have one and only one approach. According to the problem, he pursued what he thought was the best approach. Simon was a true polymath.