ABSTRACT
Primaries are becoming the preferred mechanism to select parliamentary elites in an increasing number of parties across Europe. Recent literature has shown that certain party-level characteristics favour the implementation of this inclusive method of selection. However, despite its great relevance in party life, the impact of the electoral system has not attracted enough attention. In this article we wonder if the type of electoral list (blocked or not blocked) has an effect on how participative the method of candidate selection is. We expect primaries to be more popular where electoral lists are blocked, as an intermediate mechanism for the electorate to influence an otherwise leadership-driven process of candidate selection. However, the results of a survey of 2,561 candidates from 37 parties show that this expectation is not met, as primaries seem to be more common precisely in non-blocked lists systems, with this effect being especially visible among right-wing parties. These findings suggest that the type of list might play a key role in this important aspect of internal party life.
Acknowledgments
We are indebted to the CIUPARCRI team based at the Universidad Pablo Olavide for allowing us to use the preliminary data from their project in Spain, which served as a playground for this article. The CIUPARCRI database was part of the final CCS database. We would also like to thank Sorina Soare for her assistance in party classification of Romanian parties.
Disclosure statement
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Data availability statement
Publicly available datasets were analysed in this study. This data can be found here: https://www.comparativecandidates.org
Supplemental data
Supplemental data for this article can be accessed online at https://doi.org/10.1080/14782804.2023.2195619.
Notes
1. Two other dimensions are used by the Hazan and Rahat (2001), including candidacy (who can be selected) and the voting/appointment systems (the procedure to determine the selection). Additionally, they also refer to functional (de)centralisation, which refers to functional representation for candidates that represent certain social groups or sectors such as trade unions, women, and minorities.
2. See the books edited by Cross and Katz (2013) and by Boatright (2018) for a comprehensive analysis of the effects of primaries on voter turnout, polarisation, and internal dynamics of power, among others.
3. It should also be noted that parties might use different selection strategies depending on the list position at stake, e.g. primary for the top position to increase participation, delegate appointments for mid-tier positions to and direct appointment for bottom positions.
4. By worse positions we refer to lower chances of being elected, not necessarily (or not only) the actual position in the list. In some countries, the law requires that the candidates must be ordered alphabetically.
5. The question in the core questionnaire in English asked ‘Who was most influential in deciding on your candidacy nomination’.
6. 2013 and 2017 are not present in the sample as there was not any election with valid cases during these two years.
7. The CCS also includes other relevant countries for this analysis, such as Austria, Denmark, and The Netherlands. However, our dependent variable had no valid cases for Denmark, while Austria and The Netherlands had insufficient cases for the independent variables.
8. We acknowledge that a widely accepted classification of the type of lists is missing in the comparative literature. Blocked lists are sometimes defined as ‘closed lists’, while non-blocked lists may be referred to using the broad term of ‘open lists’ or ‘preferential voting’ (Passarelli 2020), which would include both blocked and unlocked lists, as well as free lists. It is precisely to avoid considering in the same group free lists (such as true panachage in countries like Switzerland and Luxembourg) together with list systems allowing for preferential voting in the same party lists (such as Sweden, Denmark or Belgium) that we only considered the latter and labelled these as non-blocked lists.
9. Values from 0 to 4.99 were coded as left, while values from 5 to 10 were coded as right. The Workers’ Party of Belgium, the Communist Party of Finland, the People’s Party (Dan Diaconescu) in Romania, and We Can (Podemos) in Spain were the leftmost parties with an original value of 1.2, while the Flemish Block scores the higher value to the right (9.66). Our models were run using the original 0–10 variable, with similar results.
10. We used this threshold to approximately account for a full legislature.
11. For a detailed discussion on the topic of personal representation and its consequences, see Colomer (2011).
12. The visualisation of the predictive margins of Model 2 are available in in the Annex.
13. The three-level models, using countries as an additional cluster, also show significance in an additive model for the same variables.
14. The data can be found in . For the Unified Democratic Coalition and the Social Democratic Party, the data refer to two elections (2011 and 2015).
15. It should be noted that although the additive model is significant in both specifications (two and three levels) at 0.05 level, the interactive model is only significant at 0.05 level in the two-level model.