Voting by Adaptive Agents in Multi-Candidate Elections

Here we investigate how strategic adaptation changes our understanding and predictions of voting systems. We model populations of agents voting in multiple elections over time that revise their voting strategies based on electoral outcomes. Using this framework we examine the dynamic formation of electoral outcomes. Allowing agents to adapt their strategies refines the predictions of rational choice models and allows for a more realistic basis for understanding elections. We find that Approval and Plurality voting rules generate similar, relatively intuitive behavior, while Borda Count results in less intuitive, and often miscoordinated, outcomes.