The Institutional Determinants of Government Formation in Parliamentary Democracies: Assessing the Influence of the Confidence Relations on the Type of Coalition

The defining characteristic of parliamentary democracy is the dependence of the executive on the legislature, making parliaments central in shaping how governments are formed. This dependent relationship between the executive and the legislature is most clearly expressed in three institutional procedures, mainly parliamentary votes: the vote of investiture, the vote of confidence, and the vote of no confidence. Together, we refer to these as the confidence relations, and we argue that they capture a core characteristic of parliamentary regimes that shapes the balance of power between executives and legislatures. The institutional design of parliamentary confidence relations shapes the rules influencing how political actors bargain over government formation. However, the research on coalition formation rarely treats investiture, confidence, and no confidence rules as a single, holistic institutional configuration. This paper examines whether and how the confidence relations affect the probability of forming different types of coalition governments. Building on a parliament centered measure of “confidence power” resulting from our conceptualization of the confidence relations, we define four types of confidence relations—simple majorities, increased majorities, negative confidence, and constructive confidence. Empirically, we analyze the institutional confidence relations and resulting typology of parliamentary confidence power in 20 parliamentary democracies. We extend this typology to a time-varying measure by coding institutional reforms in the confidence relations over time. We then link country-time observations of confidence relations types to a comprehensive dataset of coalition governments and model coalition formation outcomes as the probability to form different coalition types. We fit a panel multinomial model with country fixed effects and translate the estimates into a staged probability tree model to predict probabilities of coalition type by confidence relations type. Our identification strategy leverages within-country institutional change to distinguish institutional effects from time-invariant national traits, estimating how shifts between confidence relations types change coalition types probabilities while accounting for clustered observations and systemic conditions. That is, we assume that the types of coalitions formed change according to the type of confidence relations in a country, and when countries change their confidence relations, they shift from one type of coalition to another. The paper’s main contribution is to bring an important but neglected aspect of institutional design of the confidence relations in general, and their change in particular, into the study of coalition formation. We will show that coalition types are not only a function of systemic and political preferences, but also of the institutional rules that govern how governments are formed, sustained, and removed through the institutional procedures that structure executive responsibility to the legislature. By integrating formation, survival, and termination procedures into a single longitudinal framework, the study provides a new basis for comparative claims about how parliamentary institutions structure coalition bargaining and the types of coalitions that emerge.