Coalition Politics: Simulation of a General Model of Multiparty Competition

经过四十年的理论研究，已经形成了公认的两党选举竞争的基本模式。这一基本模式从根本上是以政党希望赢得选举这一压倒一切的目标为前提的。鉴于模型的形式结构已被理解，通过纳入选民部分的更复杂的行为或候选人的不同程度的能力来检查模型的变体是可能的。大多数关于竞争性选举中选举前行为的理论都集中在“政策盲目”的模式上，即政党为了最大化选票或最大化预期席位而竞争。其他模型关注的是“联盟风险”，即假定政党关心政策，并根据政策协议形成特定的联盟。这项拟议的研究建立在调查人员先前关于多党竞争均衡模型的工作基础上，该模型纳入了政党对选举风险和联盟风险的信念。这项工作将完善一般选举风险-联盟风险模型。将审查两个主要问题：(1)政党是否只是最大限度地增加预期的席位数量，还是其偏好在选举反应中纳入了差异？以及(2)政策考虑的纳入是否为政党声明提供了更好的解释？可以预见，通过将计算机模拟和计量分析相结合，这项研究不仅将增进对民主理论的理解，还将提供一种强有力的方法论工具，为建立两阶段可区分博弈的一般技术提供建议。****

Scholfield Forty years of theoretical research has produced a generally accepted basic model of two-party electoral competition. This basic model is fundamentally premised on the overriding objective that parties wish to win the election. Given that the formal structure of the model is understood, it is possible to examine variants of the model through incorporating more sophisticated behavior on the part of the electorate, or varying degrees of competence on the part of candidate. Most theories of pre- election behavior in competitive elections have focussed on "policy-blind" models where parties compete to maximize votes or to maximize the expected number of seats. Other models focus on "coalition risk" that assume parties care about policy and form specific coalitions based on policy agreements. The proposed research builds on the investigators previous work on an equilibrium model of multiparty competition which incorporates the beliefs of political parties concerning both electoral risk and coalition risk. This work will refine the general electoral risk-coalition risk model. Two major questions will be examined: (1) Do parties simply maximize the expected number of seats or do their preferences incorporate variances in the electoral response? and (2) Does the incorporation of policy concerns provide a superior explanation of party declarations? It is envisioned that this research, through using computer simulation and econometric analysis in combination, will not only advance understanding of democratic theory but also provide a strong methodological tool that will suggest general techniques for modelling two stage differentiable games. ****