Voter Knowledge, Competitive Media and Party-Competition: An Agent-Based Model

Traditional models of spatial voting assume that voters choose the candidate closest to them in the policy space. This requires that, during a campaign, these voters have acquired enough information so that they can preferentially rank-order all candidates and vote correctly. However, many voters depend for their political information on mass media outlets like newspapers (or newscasts) which may have incentives to tailor their reporting, for example, if this helps them increase sales or in case they are motivated by policy themselves. This then affects the information available to voters and possibly their vote choice. The current paper presents a computational spatial model that investigates the dynamic interrelationship between voters, candidates and the media in a 2-party political system. The model addresses questions about the quality of information available to voters in a political system and how characteristics of a media system (e.g., the number of independent media outlets) may affect political outcomes like how representative political competition is to the wishes of the median voter. The contributions of this paper are twofold. To begin with, it aims at improving our understanding of the “information environment” from which many individuals receive their political information. Instead of treating the media as an exogenous “conveyor belt” of information, the project investigates how decision rules available to media outlets affect what information is available to the electorate. In addition, the paper adds to the (theoretical) literature of dynamic policy-based party competition, by investigating what role media may play in such models.