Party System Closure. Party alliances, Government Alternatives, and Democracy in Europe (Oxford University Press, 2021) written by Fernando Casal-Bertoa and Zslot Enyedi is a unique, informative and insightful longitudinal study of how European countries were governed in the last century. Before assessing the methodological advancements that the book proposes in the study of party system institutionalization, it is worth noting that scholars and students who are not familiar with this branch of political science will find in the book a massive amount of information about governments in small and big countries in Europe. The book includes data for much-analysed countries in Western, Eastern and Southern Europe, but also for micro-states (such as San Marino, Andorra and Liechtenstein) or other under-analysed countries (Albania, Moldova, Montenegro, North Macedonia). The third chapter in this regard provides a very informative analysis of the functioning of current party systems in Europe, while chapter four adds to this an historical account of defunct party systems (24 in total). The geographical scope allows the authors to both drawing inferences from the whole universe of elections in Europe (and not just a sample of them). This is per se an ambitious achievement; yet, Casal-Bertoa and Enyedi go beyond this, adding to a cross-time dataset a new analytical perspective which constitutes a significant step forward in the study of European party systems.

In fact, the explicit aim of the book is to present a new measurement, based on the patterns of parties' interactions in government, for the often-elusive concept of party system institutionalization. The authors, drawing from the work of Peter Mair, use the concept of ‘closure’ to measure these patterns. Party system closure is a composite concept, which entails three main dimensions, namely the alternation in government, the novelty of the formula of government composition and the degree of access of new parties to government. According to the authors, a system is considered as closed in two opposite-yet-complementary cases, that is regular alternation in power (the parties in government are replaced at each election as it happens e.g. in a two-party system) and no alternation, meaning that parties in government remain the same over time. Also, systems are closed when the government composition is stable over time and governments are composed of a ‘narrow circle of continuously functioning parties’ (p.19). Contrary to other measurements of party institutionalization, such as party fragmentation or electoral volatility, party closure emphasizes the agency of parties in creating stable patterns of cooperation or rather their inability to do so (open systems). Indeed, this is another important merit of the book: while other measurements focus on voters' electoral behaviour and the consequences of their choices, the concept of closure innovatively shifts its focus on parties and on the government arena. In this respect, the authors provide a compelling argument about the fact that ‘European systems spend the large majority of their time under governments which are, in terms of coalition patterns, copies of earlier governments’ (p. 126). To be sure, party fragmentation and electoral volatility remain fundamental concepts to grasp what happens in an election X at a time t and both can provide a concise measurement to analyse voters' aggregate behaviour. The concept of closure, however, synthesizes long-term trends that the other two concepts admittedly overlook: for example, in volatile environment where preferences for political parties change significantly, there might exist consolidated patterns of cooperation between parties that hinder other parties to access the government. For example, looking at Figure 6.1 (p. 132), which describes the evolutions of measures of volatility and closure per decade during the 20th and 21st centuries, we can observe a divergent pattern emerging between the two measurements at the beginning of the century and since the 80s. From 1900 to 1920, volatility is high and tends to decrease approaching the 20s; on the other hand, European systems are closed, albeit showing a tendency to become more open (especially in Southern Europe, pp. 108–112). Volatility and closure co-vary following similar patterns in the following decades (the ‘Ice Age’ of European party politics), at least until 1980, when they diverge again, in the same direction as before: European systems remain ostensibly closed, while volatility grows constantly. A similar trend can be found also in chapter 7 when the authors explore the relation between party institutionalization and closure: as expected, the two concepts are clearly correlated (with a correlation coefficient of 0.52, p. 162), since the more party are institutionalized in the system, the higher the likelihood to establish long-lasting patterns of cooperation in the governmental arena. However, Figure 7.10 shows that, similarly to what was observed with volatility at the beginning of the 20th century, despite the tendency of European systems to become more open approaching the 20s, party institutionalization increases. Since the 80s, on the other hand, party institutionalization decreases, while ‘the logic of government formation continues to be relatively predictable’ (p. 163). Party fragmentation is less correlated with closure (−0.44) and Casal-Bertoa and Enyedi find a non-deterministic relationship between the two measurements. What is crucial in this case is that closure might persist when systems become more fragmented, yet only up to a certain point. Extreme atomization as in the case of Ukraine at the beginning of the 90s meant a very open system and it is difficult to conceive predictable patterns of cooperation when parties are no longer capable to structure the interests in society. Another important finding, which the authors could have probably emphasized more is that the party system types in each country matter only to a certain extent. Only tripolar and multipolar systems diverge in terms of openness (something that one would expect by looking at coalition possibilities between parties), while grand-coalition, two-party, two-bloc, two-plus-one and centre-based system show remarkably similar levels of closure. What matters, on the other hand, is the age of the political system (pp. 134–137 and 224–225). Older systems are more stable in many respects as the literature has already shown; closure is not an exception to this trend. To this, the authors interestingly point out that closed systems are more likely to survive than open ones, but this relation is mediated by other variables, such as economic development.