Patterns of party cooperation : the role of past experience in the formation of pre-electoral coalitions, party mergers and government coalitions

This work analyses the three major forms of party cooperation –pre-electoral coalitions, party mergers and government coalitions –adopting a dynamic perspective, in order to show the crucial role played by past experience in explaining the recurrence of these phenomena. In doing so it complements the existing literature, mostly centred around the role played by office seeking and policy-seeking motivations in explaining parties’ behaviour, by adding a new focus on political time, whose meaning and importance are discussed in chapter one. Chapter two focuses on pre-electoral coalitions (PEC). Data show how more than 60% of the PECs formed at election t-1 are repeated in the current election. This suggests two questions. Does history of past cooperation matter in PEC formation? Is the process that brings parties to form a PEC in the current election the same whether they have already cooperated in the previous one or not? Using a random-effect dynamic probit model with transition probabilities, we find that two parties that formed a PEC in the previous election are almost 33% more likely to form the same PEC in the current one. Second, in PEC formation, parties seem to be only concerned about the maximization of their vote share, while, in PEC continuation, their decision to form the same PEC again is exclusively policy driven. Chapter three analyses party mergers, focusing on the importance of incentives for rebranding –defined as the will to move into the political space toward the median legislator– and on the influence of past experience of cooperation –defined as the existence of a previous pre-electoral pact– in explaining why parties agree to join and loose their autonomous identity. The analysis is performed on the Italian case, which, with its seven cases of merger in 60 years, represents an interesting and statistically useful unit of observation of a generally rare phenomenon. We run a rare event logistic (Relogit) regression on an original dataset on party policy preferences, derived from parliamentary debates on the vote of investiture of Italian governments from 1946 to 2008. Our results confirm that, other things being equal, the probability of having a merger increases when the incentive of re-branding increases and for those parties that have previously been involved in a pre-electoral coalition. In chapter four we build on the existing literature on government coalitions, confronting “classic” explanations with the political hearth model. We try to answer the following questions: does the largest party have more chances to form a minority government when it holds the core position? Do the classical office- and policy-seeking theories still hold true when a core party exists? Do behavioural factors, such as familiarity and inertia, still have a significant impact on government formation in the presence of a party that has a so strong bargaining power? We limit our analysis to the Italian case, by building an original dataset which comprises all the potential coalitions between all the parties represented in the low chamber in each formation opportunity in Italy from 1946 to 2008. Our results show that when the largest party constitutes a core, the potential governments that are more likely to form are either single party cabinets held by it or coalitions including them. When the core is absent, potential governments are less likely to form if they are minority governments, and the more parties they include. Moreover, when the core is absent incumbent parties are more likely to form the government again. This is no more significant when a core party exist.