What Democracy Do We Want? The Problematic Focus on the Median Voter

One of the key principles of democracy is responsiveness (May 1978: 1). Democracy is based on institutions, which allow citizens to control their governments by granting them a political mandate based on citizens’ policy-making preferences, and revoking it in subsequent elections if the government does not respond to citizens’ expectations. However, in all democracies the interests and preferences of citizens on political issues vary. As citizens will never have identical opinions on an incumbent government, direct control by the citizens requires a rule that establishes how opinions are aggregated and to whom governments should be responsive. This article argues that the widespread belief that democracy should resort to the rule of the median voter is too shortsighted. Democracy may take many different shapes and forms that go beyond the rule of the majority, but the extent to which democracies make use of these options differs vastly between countries (Bühlmann et al. 2011). Reviewing new empirical research from The National Center of Competence in Research (NCCR) Democracy, we show that the still dominating focus on the median voter risks excluding important groups and legitimate preferences within the population. More inclusive models of democracy not only allow a larger number of citizen preferences to be considered, but they also lead to less conflict and higher degrees of political support (among minorities). Both, political practitioners and academics usually associate democracy with majority rule, which is widely understood as the rule by the median voter. From a theoretical perspective, the importance of the median in politics stems from the fact that “the position of the median voter is the only policy that would be preferred to all others by a majority of voters” (Powell, 2000: 163). A more careful investigation of the effect of institutions associated with the majoritarian model of democracies, such as the plurality rule or multi-round majoritarian elections, shows that they often fail to represent the median voter (see Grofman 2004 for a review). However, the main underlying question of this discussion are less the failures of majoritarian institutions to represent the median voter, but the normative ideal as such. From an empirical perspective, an abundance of studies deal with questions of policy responsiveness and issue or policy congruence (Golder and Stramski 2010; Huber and Powell 1994; Powell 2000; Powell and Vanberg 2000; Wlezien and Soroka 2012), or representation more generally (e.g. Shapiro et al. 2010). Most studies of policy congruence primarily analyse whether policy-making reflects the position of the median voter or how preferences among legislators reflect preferences of the population as a whole. Going beyond this, there is increasing empirical interest in identifying systematic deviations from the median voter model. In some democracies, specific groups of citizens tend to be more influential. In particular, these are wealthy citizens, specific interests, which are more easily organised than others, or particularly engaged citizens (Bartels 2008; Gilens 2012). These studies report a bias of representation towards privileged groups, which is understood as undermining the quality of democratic representation. Instead, we ask whether the particular consideration of groups other than majorities can improve democratic representation, i.e. we discuss the quality of representation beyond the common focus on the median voter. In political theory, the dangerous consequences of the primacy of the median voter have been highlighted in discussions about the possible tyranny of the majority. Unconditional rule by the majority, often discussed as one attribute of populist democracy (see Kübler and Kriesi 2017), places the views of the majority above those of minorities, and risks neglecting the interests of minorities that warrant protection.1 This critique particularly applies to societies with stable identity-based cleavages where the prerequisite of majority rule does not hold; namely the assumption that today's majority may be tomorrow's minority (Guinier 1994, chap. 1). In this context, preferences tend to be linked closely to cleavages that correspond to subgroups in the population, and there, majority rule easily transforms into the rule of the majority over those citizens identifying as the minority. Specifically, in democratic decisions made by majority rule, majorities can overrule minorities, even on issues which are very salient for minorities, but in which the majority has no stake. Once we consider that some issues (e.g. language policies, territorial organisation) are more salient for minorities (e.g. linguistic or regional groups) than for the majority, then majority rule will be revealed as unsatisfactory: it ponders to the preferences of the little affected majority, and neglects the importance of the issue for minorities (Brighouse and Fleurbaey 2010). In the absence of special guarantees for the minority this may even happen with proportional representation. Exclusion on the basis of group identity fundamentally violates the ideal of political equality, which demands that all citizens be treated as politically equal.2 While the literature on varieties of democracy has long acknowledged that democracies come in many shapes and forms – the most important distinction being between consensus and majoritarian democracies (Lijphart 1999) – the literature on policy responsiveness has mostly overlooked this tension and analysed policy responsiveness from a majoritarian perspective of democracy (e.g. Hobolt and Klemmensen 2007; Kang and Powell 2010; Roberts 2010). Even in proportional systems, legislative decisions are eventually taken by simple majority votes. The consideration of the preferences of minority groups should, however, be a key element of all democracies, even, and particularly, if they contradict the position of the majority (Kymlicka 1995; Van Cott 2005). This does not mean that the preferences of minorities should generally overrule the majority. Instead, it implies that the preferences of minorities should be recognised if they concern certain minority-relevant policies and/or if preference differences overlap with identity based groups in society. There are different political institutions, which allow for the inclusion of political and social minorities into decision-making. Inclusive parliaments not only offer space for the battle between the governing and the main opposition parties, but also allow smaller (political) minorities to be heard. Elections under proportional rule can help minorities to get represented in parliament, although this is usually not sufficient for their inclusion into decision-making. In parliaments, super-majority rules or veto-rights for certain minority groups for constitutional changes and/or legislation imply that reforms need to be agreed between majorities and minorities, and usually lead to coalition governments. De-facto vetoes for minorities are also possible in elections. Several countries in Africa apply special rules on the territorial distribution of votes, which a candidate needs to gain in order to be elected. In countries where social or ethnic groups live in separate parts of the territory, minorities can effectively veto certain candidates from becoming elected through territorial vote distribution requirements (Bogaards 2003) or through election formulas which guarantee that the preferences of majorities and ethno-territorial minorities are equally taken into account (Bochsler 2012). To our knowledge, while such rules are applied in elections, there is less extensive use thereof in direct democratic decision-making, apart from double majority rules (i.e. where an approval of the majority of the citizens and of the sub-states is required to pass a vote). A school of scholars attributes similar effects to majoritarian institutions if they are applied in sufficiently heterogeneous places or in heterogeneous electoral districts. To be elected, candidates need to win a majority of the votes, which is often easier to achieve through alliances with social and political minorities. This is expected to lead to a centripetal effect, where the candidates elected take minority-friendly positions (Horowitz 2003; Reilly 2001). However, in some societies, and for certain issues, majorities living in heterogeneous contexts tend to adopt positions hostile to minorities. In these contexts, the centripetal effect inverts, and the result seems unacceptable for minorities (for the US: Key 1949; for an application to Serbia: Bochsler 2013). Some political theorists demand that decision-making power is distributed proportionally to people's stakes in a particular decision (Brighouse and Fleurbaey 2010: 137), or that voting rules be adapted to ensure that identity-based minorities are not consistently outvoted (Guinier 1991).3 There are different attempts to translate this debate into empirical types of democracies. Arend Lijphart has argued in “Patterns of Democracy” (1999) that democracies can be categorised into a majoritarian and a consensual type. Consensus democracies are characterised primarily by the first two institutions discussed here, proportional representation – leading to large party systems – and enhanced veto points (e.g. super-majority rules or minority vetoes), which require negotiations and compromises between parties. This is associated with more inclusive policy-making, where political minorities are also included in some of the decision-making processes. In Lijphart's tradition, Bochsler and Kriesi (2013) have extended the analysis of the variety of democracies to the 1990s and 2000s, and to more recently democratised countries in Central and Eastern Europe, Latin America, Africa and Asia. Capitalising on the multidimensional measure of democracy, offered by the Democracy Barometer, they find that real democracies can be distinguished according to five dimensions. Here, the distinction between consensus and majoritarian democracies is still among the most important.4 Consensus democracies are open to including political minorities. In Bochsler and Kriesi's analysis, they are particularly characterised by an open electoral system with low thresholds, large party systems, a high degree of representation of women and minorities, and indicators of a corporatist economic model.5 The model also extends to new democracies, with South Africa as the archetype of a proportional democracy. The cases at the opposite, majoritarian pole include both old democracies (e.g. United Kingdom or Canada), and newer (and less stable) democracies, such as Turkey, Indonesia, or Hungary (in the period of investigation). The choice between different models of democracy reflects the normative question on the principles of representation. While majoritarian democracies emphasise the representation of the median voter, proportional democracies allow for the inclusion of citizens who belong to political minorities, and substitute majoritarian decision-making with compromises and negotiations. Empirically, the two principles are hard to reconcile, and the distinction between the two types constitutes a trade-off (Bochsler and Kriesi 2013). The fears of a tyranny of a majority have also been reflected in a vast empirical literature dealing with the effect of direct democratic institutions on minority rights and civil rights. However, there is wide disagreement as to how direct legislation relates to the consensus-majoritarian dimension. It entails elements of both concepts. Hug and Tsebelis (2002) show that direct democratic institutions introduce a new veto player into the political system - the median voter of the population. Relying on this argument, Vatter (2009), in an analysis of 23 democracies, shows that referendums create an incentive for the extension of the governing coalitions,. The more extensive sample of the countries of the Democracy Barometer, also including third-wave democracies in Latin America and in Central and Eastern Europe (with more frequent practice of referendums and initiatives) does not reveal a correlation. The direct-representative dimension appears as a separate dimension in Bochsler and Kriesi's (2013) analysis. Turning to social and political minorities, studies have looked at direct democracy and its effect on manifold minority issues, ranging from; the rights of same-sex couples over abortion rights; “English-only” laws, prescribing English as the only language to be used for state action in US states; and anti-discrimination laws or citizenship rights (e.g. Gamble 1997; Helbling and Kriesi 2004; Vatter and Danaci 2009). Results obtained depend heavily on the issue, the case of study, and the methodology employed. Hug suggests that direct democratic institutions offer the means to correct legislation where parliaments are not following the preferences of the median voter, and thus lead to policy outcomes which are more strictly oriented towards the median voter (Hug 2011). As a result, this can also undermine the protection of social and political minorities from majoritarian decision-making. Investigating five fields of minority rights, Bochsler and Hug (2015) show that the effect of direct democracy on minorities depends on voter preferences. Direct democratic institutions reinforce the median voter tendency of legislation. While they tend to offer higher protection towards minorities with a strong political support among the population than representative democracy institutions, they also can undermine the protection of minority groups, which are unpopular in their countries. The question of who is represented in politics is not only a normatively relevant exercise, but also has concrete political implications, as multiple contributions from NCCR Democracy have shown. For instance, studying ethnically distinct population subgroups, Cederman et al. (2010a) have repeatedly shown that excluded ethnic groups are more likely to be involved in civil conflict with the government, especially during democratization periods (Cederman et al. 2010b) and in times of elections (Cederman et al. 2013). This not only applies to countries in transition, but also to democracies. Studying the countries covered by the Democracy Barometer, Hänni (2017a) shows that ethnic minority groups who lack inclusion, and whose preferences are not represented in politics, are more likely to protest against the government – both violently and non-violently. But even when dissatisfaction and exclusion do not translate into actions against the state, they have potentially harmful consequences for democracy, as excluded or underrepresented groups exhibit less support for the democratic regime (Hänni 2017b), participate less in politics (Burns et al. 2001; Phillips 1998), and are generally less politically involved (Bühlmann and Schädel 2012). This applies not only to ethnically distinct groups (Banducci et al. 2004; Pantoja and Segura 2003), but also to young citizens (Bousbah 2016), women (Bühlmann and Schädel 2012), and to socio-economically poorer groups (Schäfer 2013). Data from the Democracy Barometer suggests that poorer people are systematically (self-)excluded from the political process, whereas research on Latin America provides evidence that poorer groups are less well represented in politics due to their numerical underrepresentation in parliament (Carnes and Lupu 2015)6. Even if political exclusion in these cases does not necessarily involve anti-government actions, it is problematic as it may endanger the long-term stability of democracies. Both, participation and enduring support for the regime are necessary pillars of a functioning democracy. When voters abstain from politics and lose support for the political system, democracy risks not only normative, but also empirical failure. The importance of control by the citizens (vertical control) in democracies is uncontested. Every democratic government allows its citizens to regularly renew or abolish its mandate through free and fair elections. While the regular holding of elections is undisputed, there is less agreement on how to aggregate public opinion and how to choose representatives. In this contribution we critically evaluate the primacy of the median voter in politics and discuss the theoretical and empirical limitations related to this model. From a theoretical perspective important advances in the literature depart from the idea of pure median voter representation and attempt to include as many perspectives as possible into politics (Lijphart 1984: 4). Bochsler and Kriesi (2013) show that these models do not remain theoretical but are applied by real world democracies, resulting in democracies of many different shapes and forms. Departing from representing (only) the median voter is crucial for ensuring the stability of today's democracies. Heterogeneous democracies have better chances to succeed and survive if important subgroups within a state are represented in politics in one way or another (Juon and Bochsler 2017; Linder and Bächtiger 2005). While this has long been acknowledged and is increasingly well achieved with regard to identity-based ethnic groups, there remains room for improvement with respect to other identity-based groups. In particular, the focus of political representation on privileged groups can be a challenge to democracy: it risks opening the door for (‘populist’) newcomers in the political institutions (Bornschier 2017) and for contestation outside the representational arena (Hänni 2017a). In the 21st century, responsiveness has been challenged from other directions too. Technocratic governments without democratic legitimation have been on the rise in Southern Europe as are populist movements who claim to be responsive to ‘the people’, but exclude those who do not belong to the political community according to their definition (see Kübler and Kriesi 2017). At the same time, international organizations with questionable democratic legitimacy have become more influential (see Freyburg et al. 2017). Questions of responsiveness and representation will stay salient in the debate about the quality of democracy. Future research should ask more explicitly to whom democracies are responsive. Daniel Bochsler is an Associate Professor of Political Science at the University of Copenhagen. He has been an Assistant Professor of Comparative Politics (Democratization) at the University of Zurich and at the Centre for Democracy Studies Aarau from 2011 to 2017, and is a co-project leader of the Democracy Barometer. His research deals with political institutions, especially in culturally diverse countries and in the context of democratization. Address for Correspondence: Daniel Bochsler, University of Copenhagen, Dept. of Political Science, Øster Farimagsgade 5, DK-1353 København K, [email protected]. Miriam Hänni is a postdoctoral researcher at the University of Konstanz. She holds a PhD from the University of Zurich. Her research interests include representation and responsiveness, heterogeneous societies, and questions of electoral competition. Her research has been published in Legislative Studies Quarterly, Political Studies, and the Journal of Ethnic and Migration Studies.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.