ABSTRACT
Oversized coalitions are not unusual in Central and Eastern Europe (CEE), but there is less attention to the puzzle of why political parties seek to form them. Empirical studies of CEE often lack a systematic comparative perspective of the different mechanisms. Theoretically informed studies face a difficulty of the application to coalition games in CEE. To fill the gap, this article investigates pathways of oversized coalitions in CEE with Qualitative Comparative Analysis (QCA). This study shows that multiple causal paths can be constructed by combined conditions relating to main hypotheses from coalition theory. The major findings are: (1) bargaining power concentration would have a contextual effect on oversized coalitions; (2) median in policy dimension(s) could be seen as a key concept to provide an explanation of the coalition pattern in CEE; and (3) bicameral and semi-presidential structure cannot be rejected as a combined factor for broader coalitions in CEE.
Acknowledgements
An earlier version of this article was presented at the 24th World Congress of Political Science at Poznań, at the 10th ECPR General Conference in Prague. I would like to especially thank Aron Buzogány for invaluable research assistance.
Disclosure Statement
No potential conflict of interest was reported by the author.
Notes on contributor
Sho Niikawa is a postdoctoral researcher at Sophia University, Japan. His research covers comparative political institutions, party politics, coalition study, qualitative methods.
Notes
1 Although there are other relevant indices, e.g. the Shapley-Shubik power index, I apply the Banzhaf index as this can be understood as a standard measure of the bargaining power in comparative coalition studies (e.g. Bergman, Ersson, and Hellström Citation2015).
2 The fragmentation score of 3.5 does not seem extreme. The tendency for fragmented party systems in CEE countries can be measured by the average score, which is 3.71 for the 48 CEE cases between 1991 and 2011.
3 I can propose a higher threshold of 300 based on empirical data for Western European countries. Even if we use this method to set the threshold that can distinguish comparatively small parliaments, e.g. Finland (200), Belgium (197) from other larger countries, e.g. Spain (350) and Sweden (362), the analytical robustness of this article is confirmed.
4 Cases over this threshold are rare in CEE countries. The average score is 9.22 across the 48 cases between 1991 to 2011. Therefore, in observed cases scoring over 20.00, I expect centrifugal party competition.
5 The prime minister’s institutional power may also be hypothesised (Bergman, Ersson, and Hellström Citation2015). However, this hypothesis still seems to be needed to theorise what mechanism of prime ministerial power can be expected for oversized coalition formations.
6 Including those two countries does not change the results.
7 Analyses centred around Y-combinations give rise to other problems, such as unrealistic combinations in the possible coalitions being included, or to probabilities of possible patterns being estimated with an assumption of independence from irrelevant alternatives (IIA).
8 I use a simple multi-value QCA, because there are still empirical and methodological problems with applying the advanced version, namely “fuzzified” multi-value QCA (Thiem Citation2014).
9 Because QCA is a tool for analysing sufficient condition(s), necessary condition(s) is checked in advance (Schneider and Wagemann Citation2012, 139–144). Consequently, no single necessary condition for oversized coalitions is identified (see Appendix ).