Potential centrifugal effects of majoritarian features in proportional electoral systems

ABSTRACT

Many scholars argue that proportional representation (PR) has fragmenting and polarizing effects on party systems (centrifugal effects), while the use of single-member districts has moderating effects (centripetal effects). Beyond these two extremes, in order to combat the potential negative consequences of fragmentation and polarization on government, PR systems can incorporate majoritarian features as a moderating influence. We examine this relationship by considering the conditions under which majoritarian features can backfire and contribute to a more extreme government. We present a theoretical example to illustrate this effect across a range of majoritarian rules and highlight the importance of coordination. To stress that our example is not simply a theoretical possibility, we discuss the cases of Poland (2015), Hungary (2014), and Greece (2015), which incorporate majoritarian features through, respectively, a high election threshold, single-member districts, and bonus seats to the plurality winner. We argue that these elements in each case contributed to the legislative power of extreme, populist parties.

KEYWORDS:

• Electoral institutions
• proportional representation
• majoritarian
• coordination

Disclosure of potential conflicts of interest

No potential conflict of interest was reported by the author(s).

Notes

1. Some researchers, however, question whether PR leads to more polarized party systems (c.f. Adams and Rexford 2018; Ezrow 2008).

2. There have been studies that illustrate the limitations of Duverger’s Law, such as Chibber and Kollman’s (1998) analysis of India. See Aldrich and Lee (2016) for a more detailed explanation of the factors leading to two-party systems.

3. Roughly speaking, the parties would be evenly spaced along the left-right dimension – hence the leftmost and rightmost parties are more polarized as the number of parties increases.

4. Our paper is mostly concerned with the representational aspects of PR versus majoritarian systems, i.e. which one leads to policies/governments closer to the median voter. We clarify the conditions under which PR has positive representational effects, whereas Carey and Hix (2011) build on Lijphart (1999) by assuming that more permissive PR systems have a clear representational advantage.

5. The dimension could capture conflict over redistribution, Euroscepticism, or populist, anti-establishment sentiment.

6. We could instead define party position in terms of (inverse) cumulative distribution functions (voter support, given their locations).

7. We do not consider district magnitude beyond the SMD case in our paper. In general, we might expect that a higher district magnitude hurts the ability for a smaller or plurality-favoured party disproportionately to win more seats (e.g. in Figure 1(a), if M = 2, then L and R each win one seat). Comparing such cases, e.g. M = 3 versus M = 4, is outside the scope of our paper, however, since an in-depth consideration arguably adds complicating factors, like the number of parties and their positioning, when analysing its ultimate effects on government (more so than for the features we consider).

8. Note that whereas Cox (1997) focuses on the potential backfire of lower district magnitude, our one example extends that logic to other majoritarian features.

9. For simplicity, we are assuming that a party could govern with 50% of the seats, which is technically not a majority.

10. Future work will consider in more depth the factors that contribute to coordination failure, which may also be more amendable to a larger statistical study that can reassess earlier findings (e.g. Powell and Vanberg 2000; Carey and Hix 2011). In broadening our study, we might also imagine that the effective number of parties and party distributions in the left-right dimension can impact the likelihood of observing centrifugal outcomes. In such a study that takes party distributions as given, the outcome depends on the ability of voters to coordinate (vote strategically) to achieve or avoid particular outcomes. Our analysis of vote choice that we present in this paper is a first step in this direction.

11. For our current study, we are not as interested in differences across the different types of majoritarian features. A statistical analysis would consider each type separately in coding the independent variable of interest, as each variable can only measure one specific majoritarian feature.

12. Although M5S is difficult to classify ideologically, scholars have identified it as populist (Mudde 2018; Mudde and Kaltwasser 2018). Thus, at the very least, M5S is a non-mainstream party and is extreme on an anti-establishment (populist) dimension, which is consistent with our case depicted in Figure 1(a).

13. Having a candidate on the SMD ballot helps ensure party supporters in fact make it to the polls to also cast a party vote for the nationwide seats.

14. We could not find a suitable survey for the 2014 Hungarian elections.

15. Poland poll results: https://en.wikipedia.org/wiki/Opinion_polling_for_the_2015_Polish_parliamentary_electionGreece poll results: https://en.wikipedia.org/wiki/Opinion_polling_for_the_January_2015_Greek_legislative_election

16. If the configuration of parties includes a more extreme rightwing party, r, as in Figure 1(b), then polarization by R decreases votes if we further assume a unimodal (e.g. normal) voter distribution. The gain in votes from gaining supporters for r is smaller than the loss of voters to party L.

17. Of course, ℓ’s location as a spoiler party can also contribute to R’s polarization.

18. If this message of party viability is also conveyed to voters, then voters may also be more willing to support more extreme parties, which reinforces their success.

Additional information

Notes on contributors

Christian B. Jensen

Christian B. Jensen is an Associate Professor of Political Science at the University of Nevada, Las Vegas. His research interests focus on democratic institutions, especially in the European Union. His recent publications include, “Predicting ambiguity: Costs, benefits, and party competition.” with Daniel Lee in Political Research Quarterly.

Daniel J. Lee

Daniel J. Lee is an Assistant Professor of Political Science at the University of Nevada, Las Vegas. His research interests are legislative and electoral politics with a focus on political parties.