Agent-Based Modelling of Electoral Systems

• In Event: Electoral Systems, Electoral Incentives, and Political Actors

Abstract

In a period of discontent with democracy, the question of reform of electoral rules naturally arises. It is difficult for political scientists to predict the effects of some reforms, however, as many possible combinations of rules are not common enough in the real world to make statistical analysis possible. Formal modelling can help, but becomes intractable or unsolvable if too many rules are included at once. Computational agent-based modelling can help overcome these problems. By representing voters and candidates as agents within a simulation, political scientists can study arbitrary combinations of rules and distributions of preferences in the population. In this paper I offer both a theoretical justification for a computational approach to modelling electoral systems, as well as results from a working version of such a program demonstrating the feasibility of such an approach.

Author

• Alexander Sayer Gard-Murray, Harvard University