![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
This study highlights the role of ordinary voters in facilitating electoral reform, particularly in the context of electoral system referendums. We investigate six factors that may shape voters’ preferences: partisanship, ideology, political efficacy, system support, government performance, and value orientations. Analysis of longitudinal data from the New Zealand Election Study reveals that small party support, left-leaning ideology, and satisfaction with democracy are the most consistent and substantive predictors of preference for proportional over plurality rules, while authoritarian values and low efficacy often exert the opposite effect. Results are discussed with reference to both the 2011 referendum in New Zealand and applicability to debates over electoral reform in other countries.
Notes
1. Voters can also influence electoral system choice by preventing change. For example, constitutional amendments to replace the existing STV system with plurality were twice rejected by Irish voters (1959, 1968).
2. For details on the series of events culminating in the adoption of MMP, see Jackson and McRobie (Citation1998) and Vowles (Citation1995).
3. Further confirmation of an informed and engaged public can be found in deliberations over electoral reform in a citizens' assembly in British Columbia, where Carty (Citation2004, p. 182) found participants ‘willing to put the time and energy into doing the hard work’ of examining details of alternative electoral systems.
4. Note that support for electoral change may be orthogonal to partisanship. For example, interest in electoral reform in Canada ‘does not have a partisan base’ (Cross, Citation2005, p. 77).
5. It is often the case that parties disadvantaged by past election results no longer deem the system unjust once they return to power (Lundberg, Citation2007, p. 489), and Labour may be no exception.
6. ‘Confidence and supply’ agreements refer to parties' promises to support the government on votes of confidence while remaining outside of the cabinet. Coding only parties in the cabinet as winners does not yield substantively different results.
7. While electoral change may be the consequence of ‘system failure’ (Renwick, Citation2010, p. 67; Shugart & Wattenberg, Citation2001, p. 572), this is a sufficient but not a necessary condition.
8. One can interpret rejection of plurality as a stronger form of protest voting than opting for an opposition party (Banducci & Karp, Citation1999, p. 231).
9. However, plurality supporters criticised MMP for empowering party leaders, since party headquarters determine the ranking of candidates (and hence their likelihood of election) on the proportional component of the system.
10. In the 1993 referendum in New Zealand, 81% who rated government performance as very good voted to keep plurality, while 78% of those who rated the government as very poor opted for MMP (Levine & Roberts, Citation1994, p. 249).
11. A parliamentary committee examining this issue in 2001 recommended that a referendum should not be held since more time was needed to assess how MMP functions. All parties took part except New Zealand First, which refused to participate.
12. The 1996 election resulted in an unexpected National-New Zealand First coalition government, which was generally regarded as unsuccessful. Dissatisfaction with both the coalition negotiation process and the government's performance led to declining support for MMP (Karp & Bowler, Citation2001).
13. All analyses were replicated by substituting the MMP vs. plurality dependent variable with a dichotomous choice between systems that yield a proportional outcome (preference for MMP, STV, PR in general) versus those based on competition in single-member districts (plurality, alternative vote, preferential vote). Doing so does not produce substantially different findings from those reported below. Data available from author upon request.
14. Class is not included, because in some surveys more than 40% of respondents declined to identify themselves as belonging to any class.
15. Bivariate correlations between this item and the dependent variable range between 0.75 and 0.82 in the five surveys where this question was asked.
16. Bivariate correlations between government type and electoral system preference (both coded dichotomously) in the four surveys where the former item was included in the questionnaire are consistently in the 0.65–0.66 range.
17. United Future (including its predecessor) elected seven list MPs in 2002, and two in 2005, but none in 1996, 1999, 2008, and 2011.
18. One possible explanation is that, in contrast to the Maori party, there is no party aimed specifically at mobilising women. Media discourse that rarely highlights greater women's representation under MMP may be another reason.
19. The bivariate correlation between age and support for MMP reached 0.197 in 2005, but declined in subsequent years (0.148 in 2008 and 0.135 in 2011).
20. For example, support for the new MMP system in Scotland declined when more citizens find parliament not meeting their expectations (Curtice, Citation2004, p. 333).