Full Text

Analyzing coalition governments in Western parliamentary democracies has followed more or less the same design throughout the years. The objective has been to “predict” or “explain” the formation of governments in general. Although the explanatory power of models has improved, even the most comprehensive are able to predict government formation in less than half of the cases. 1 This can mean either that the process of coalition formation is to a significant extent idiosyncratic or that current models have not been fully specified. It is the latter approach that is taken by the book Puzzles of Government Formation. The research design follows recent suggestions about the implementation of mixed methods and nested analysis2 and sets out to first determine the cases that are not explained by coalition theories. Classical office and policy based theories are used to define such “puzzles”. Institutional theories are mentioned, but only to state that they will not be included in the initial analysis (p. 10). Although the vast majority of coalitions are thus left unexplained, the authors decide to focus on cases that are relatively least likely (pp. 14-15). The specific selection of cases is not fully explained and many other even more unlikely cases could have been chosen. Nevertheless, instances of government formation from Sweden, Norway, Austria, Spain, Finland, the Netherlands and Belgium were chosen as in-depth case studies. These cases can be considered puzzling because they were either minority governments, oroversized coalitions, which according to office based theories should not form, or non-connected coalitions and coalitions excluding the median legislator, which should not form according to policy based coalition theories. Thinking about solutions to these “puzzles”, the case studies indeed suggest variables, which in principle could be used in improved models. For example, a variable for multiple interconnected party systems could be operationalized and included, as well as electoral volatility and the fortunes of individual parties (p. 199). Furthermore, one of the lessons of the book is that more attention should be paid to defining coalitions (pp. 195-196). Many of the cases suggest that parliamentary or policy coalitions that do not include membership in government can serve the same 1 Lanny W. Martin and Randolph T. Stevenson, “Government Formation in Parliamentary Democracies,”American Journal of Political Science 45, no. 1 (2001): 33–50. 2 Hanna Bäck and Patrick Dumont, “Combining Large-n and Small-n Strategies: The Way Forward in Coalition Research,”West European Politics 30, no. 3 (2007): 467–501.CEU Political Science Journal. Vol. 8, No. 2 263 function as government coalitions. A recurrent theme is the skill of political leaders (p. 196) or specific commitments made prior to coalition bargaining. These, however, are factors that cannot be included in general models. We can thus say that this book succeeds in the objectives it sets for itself in that it is able to propose variables that could indeed improve general models, but at the same time it shows the limitations of such modelling. Although the book purports to solve puzzles, it also raises some questions. What is “deviant” is determined by what we imagine to be the factors that should influence the formation of coalitions. The book begins by taking into account policy and office theories and disregards institutions in determining what is puzzling. By now, however, it is beyond evident that institutions help to explain a lot of variance in coalition formation and office and policy alone are quite poor predictors. Perhaps the cases that were selected would not have been so puzzling, if institutional factors that we already know have a very significant impact were taken into account. Indeed, the case studies make a lot of reference to different institutional factors in explaining the “puzzling” cases. The other puzzle implicitly raised relates to the role that assumptions should play in coalition research. It is noted both by the editors in the introduction (p. 5) and by Ilja van Beest in his account of social psychological studies of coalitions (p. 37) that such research in politics should be more critical and self-aware about its assumptions and not so reluctant to reconsider the role of rational choice theory, the source of the assumptions. Yet the way that the puzzles in this book are defined follows exactly the prescriptions of coalition formation theory soaked in classical rational choice assumptions about the behaviour of political parties. In this sense the book still starts from theories and assumptions that are known to be very poor approximations of reality and that it itself says should not be made light-heartedly and should perhaps be reconsidered. By and large, office and policy theories tell us that minimal winning governments and connected governments are the norm and everything else is puzzling. Reality has been telling us that in fact they are not the predominant norm. In this context what perhaps is truly puzzling is that we still take them to be the norm, even if we say that we should not and even if we know that they are not. But is it also perhaps in this last characteristic of the book that its greatest merit for research in coalitions lies. For it shows again that what is puzzling for office and policy theories alone, has in fact quite obvious explanations in the calculations and behaviour of parties, suggesting yet again that research in coalitions should start looking far over and beyond the former. It indicates important elements that have been missing in models, but also what remain beyond the reach of general models of coalition formation. It is thus essential reading for anybody interested in coalition research and some of the more recent developments therein.Book Reviews 264 Bibliography Bäck, Hanna, and Patrick Dumont. “Combining Large-n and Small-n Strategies: The Way Forward in Coalition Research.” West European Politics 30, no. 3 (2007): 467–501. Martin, Lanny W., and Randolph T. Stevenson. “Government Formation in Parliamentary Democracies.” American Journal of Political Science 45, no. 1 (2001): 33–50.