The dynamic battle for pieces of pie-Modeling party support in multi-party nations

When teams of rival politicians compete for public support, they are essentially playing a zero sum game where one party's gains tend to come from the losses of one or more of their opponents. Despite this, most analyses of party support across time model the dynamics associated with a single party's support. In nations where only two parties are competing for votes, this approach is fine. But in nations with more than two parties, much of the substance of what is going on in party competition is lost. In this paper we illustrate the usefulness of a modeling strategy proposed by Philips et al. (2015) for estimating and interpreting the causal relationships that shape trade-offs in party support as they evolve over time. We extend their work by modeling public support for four parties instead of three and by developing the ability to model dynamic changes in party characteristics. We estimate our models on monthly data from the United Kingdom and Germany.