Electoral Rules and Electoral Behaviour

The Scope of Effects

Ruth Dassonneville, Marc Hooghe, Michael S. Lewis-Beck

• Publisher: Taylor & Francis
• Year: 2018
• eBook ISBN: 9781351273503
• Edition: 1
• Topic: Politics & International Relations
• Subtopic: Politics
• Index: Politics & International Relations

Across representative democracies, there is a strong variation in the rules that govern the electoral process. A classic insight in political science is that these rules, e.g., the presence of a majoritarian or a proportional system have a profound effect on the way a democracy functions. We know less however, about the way voters actually respond to these electoral rules. This kind of effect presupposes that voters not only are aware of the electoral system, but also that they adapt to the incentives offered by the system. In this volume, a group of international scholars investigate whether this is indeed the case. The various chapters in this volume deal with the effect of proportionality, mixed-member systems, compulsory voting and preferential voting. The chapters are based on recent data and state-of-the-art methods. The introduction confronts the findings of the various chapters with the allegedly universal validity of vote choice models in the literature. The research presented in this volume mainly deals with elections in Europe, but the findings speak to the broader community of electoral scholars. The chapters originally published as a special issue in West European Politics.

Do electoral rules have an effect on electoral behaviour? An impact assessment

Ruth Dassonneville, Marc Hooghe and Michael S. Lewis-Beck

ABSTRACT

Electoral democracies worldwide are all organised around elections but the rules under which the elections are organised differ greatly from one country to another. These electoral rules, such as whether voting is compulsory or what electoral system is used, are thought of as strongly affecting voters’ behaviour and the choices they make. If electoral rules indeed shape citizens’ electoral behaviour, the implication is that theories of what explains voters’ choices are country-specific as well. This is in sharp contrast to the idea that theories of electoral behaviour are generalisable. This special issue tackles this question and offers an assessment of the impact of electoral rules on voters’ behaviour, on the one hand, and the generalisability of individual-level theories of voting behaviour, on the other. The collection of papers furthermore offers an important contribution in terms of the kind of electoral rules that are scrutinised, with several papers focusing on the little-investigated phenomenon of preferential voting.

Scholars of electoral behaviour commonly assume voting to be strongly affected by the institutional context in which citizens are embedded. Electoral rules, and efforts to change them, are highly salient topics in the public and political debate as well, because of the assumption that voters’ behaviour is affected by these rules and because of expectations that citizens would behave differently if the rules were changed. The general expectation, therefore, seems to be that electoral rules matter, because they will have an effect on voters, and therefore also on electoral outcomes. With the term ‘electoral rules’ we refer here to the rules under which elections are organised, such as whether voting is compulsory, or whether a majoritarian or proportional electoral system operates. In general, electoral rules are expected to influence whether and how citizens cast a vote. Moreover, we observe a very strong variation with regard to the electoral systems functioning in various countries. Further, these systems tend to be rather resistant to any effort to change them (Farrell 2011). With a few exceptions, like Italy or New Zealand, electoral rules mostly can be regarded as stable country characteristics (Gallagher 1998). Therefore it makes sense to assume that these marked differences could have an impact on the way voters interact with the political system, most notably in the context of elections.

More than half a century ago Maurice Duverger (1951) claimed that electoral systems fundamentally shape voting behaviour and its determinants. Since then, numerous scholars have investigated and reported on the effect of electoral rules on voters (Blais and Carty 1991). First, comparative work that describes electoral behaviour across a number of countries, where electoral rules vary in a significant manner, has given indications of fundamental differences. Research on electoral participation, for example, consistently finds higher turnout levels in countries with a proportional electoral system as compared to majoritarian electoral systems (Geys 2006). Compulsory voting rules, too, clearly have a positive effect on turnout levels, suggesting that these rules matter in the individual decision to vote (Birch 2013). Second, causality claims have been strengthened by experimental work on the impact of electoral systems. Van der Straeten and her colleagues (2010), for example, have run a series of laboratory experiments which conclude that the extent to which citizens vote strategically depends on electoral rules. Third, the low external validity of such lab experiments has been addressed by means of quasi- and natural-experimental studies. These studies also give indications of a non-negligible impact of electoral rules, on electoral participation as well as on voters’ choices (Fowler 2013; Karp and Banducci 1999). A recent study by Sanz (forthcoming) is exemplary in this regard. This paper exploits the natural variation in the electoral rules of local elections in Spain, showing that turnout increases when local representatives are elected on open lists, as compared to closed-list elections. From this brief overview of previous work and different research strategies, it becomes evident that electoral rules seem to matter (Norris 2004). The decision to turn out to vote, not to mention what candidates or parties a voter chooses, are influenced by the rules of the game. Following Duverger, electoral systems apparently have strong and significant psychological effects.

This line of reasoning has indeed a solid theoretical basis. The presence of institutions, one assumes, shapes human behaviour and upholds this behaviour, since these norms are ultimately interiorised by the citizens. This kind of loyalty leads to social stability, as Parsons (1951) famously assumed in his classical theory. Electoral rules can be considered as part of the institutional framework of a society because they determine the allocation of power and because their stability plays a constitutive role in politics (Bowler and Donovan 2003). The agency of individual voters and citizens, therefore, should be bound by these electoral rules. This implies that norms of rational behaviour by citizens can also vary, depending, for example, on whether the institutions they live in are based on a majoritarian or a representative logic.

While electoral systems are generally found to matter for the vote choice, a more fundamental theoretical implication of Duverger’s work tends to be unappreciated. That is, the claim that electoral systems have a systematic and strong impact on voters calls into question the generalisability of vote choice models, across countries and electoral systems. Scholars examining the determinants of the vote choice have devoted quite some attention and effort to detecting general patterns. Theories, survey methods, as well as question wording for crucial concepts have travelled from one country to another. Originally, similarities in the question wording of survey items of national election surveys in different countries were the result of rather ad hoc efforts by ‘“missionaries” from Michigan’, the birthplace of the highly influential American National Election Studies (Knight and Marsh 2002: 173). Recently, more systematic efforts at harmonisation have been undertaken – for example, the True European Voter project has set out to harmonise the data from existing election studies across Europe, so allowing the testing of fundamental theories of voting behaviour. In addition, the Comparative Study of Electoral Systems (CSES) project offers a leading example of coordination in data gathering, as it revolves around the administration of identical questions in election studies from over 50 democracies worldwide. These large data efforts are in part inspired by the notion that comparative analyses will allow us to shed more light on the general characteristics of the vote choice process.

Such efforts permit a more systematic assessment of the generalisability of some fundamental theories. These theories of the voting behaviour originate mostly, though not exclusively, from scholars investigating electoral behaviour in the United States. Take, for example, the focus on partisanship as a key variable for explaining voters’ behaviour and choices. While the concept of partisanship emanates from the ground-breaking work of Campbell and his colleagues (1960), many other scholars have sought to measure and verify its impact on the vote choice in countries worldwide, assuming that the concept of partisanship could be present in other political systems as well (Bartle and Bellucci 2014; Lupu 2015). Similarly, theories of economic voting have travelled from the United States, where they were originally proposed and tested, to a large and varied set of electoral contexts (Lewis-Beck and Stegmaier 2000). Many more examples could be referred to, such as work on strategic voting – which has been examined in majoritarian as well as in proportional electoral systems (Gschwend 2007) – or work on the impact of party leaders on the vote choice (Aarts et al. 2013). These studies share a fairly obvious conviction that some fundamental theories of voting behaviour are generalisable across a large and varied set of contexts. Finding that patterns are generalisable can be important from a normative point of view as well. Citizens using their vote to hold incumbents and politicians accountable for their performance stands as a key aspect of elections, one allowing for the realisation of representative democracy (Przeworski et al. 1999). Holding politicians accountable is, indeed, an essential component of democracy. As such, finding economic voting in a large and varied set of democracies bears witness to their well-functioning, representative nature.

The essential challenge of Duverger’s work, however, lies in the impact of institutional and systemic differences between countries. If he is right, and electoral systems fundamentally shape voters’ behaviour and the choices they make, differences in electoral rules severely limit the potential to generalise research results on the individual-level factors that explain the vote. The impact of electoral rules could be such that the effect of individual-level determinants on the vote depends on specific characteristics of the electoral system. To express it differently, maybe in some countries one specific set of vote choice determinants is important, while in other countries, with different electoral rules, a totally different set of determinants might affect citizens’ vote choices. That could mean that the search for generally applicable vote choice models is in vain, because the vote choice process is highly context-specific. In that case, the obvious conclusion should be that it is impossible to determine universal vote choice determinants, since voters tend to respond to a specific set of electoral rules. In summary, the effect of the electoral system on the vote choice process can be seen as a major challenge for electoral research and the generalisability of theories of voting behaviour, a challenge not often seriously addressed. The goal of this special issue is to offer such an assessment by investigating in a systematic and comparative manner what impact electoral rules have on specific vote choice processes and determinants.

The micro- and the macro-level, direct and indirect effects

When assessing the impact of electoral rules on the vote choice process, it is important to distinguish between different levels of analysis. Electoral rules are macro-variables specific to a particular context and usually highly stable – as is evident from some recent failed attempts to reform the electoral system in Britain and several Canadian provinces (Fournier et al. 2011; Whiteley et al. 2013). The vote, in contrast, is a micro-level variable, as it is individual citizens who decide whether to turn out to vote and whom to vote for in the polling booth. Any analysis of the impact of electoral rules has to be attentive to this difference in levels, because the problem of ecological fallacy implies that we cannot straightforwardly make inferences about individuals’ behaviour from an aggregate-level analysis.

The literature on electoral rules does include a large number of studies that focus exclusively on the macro-level. Work in this tradition has offered important insights and a number of aggregate-level findings on the consequences of electoral rules prove to be quite robust. To mention only one example, levels of turnout are consistently found to be higher under proportional representation (PR) rules. Aggregate-level studies have a number of important advantages, such as the fact that they rely on precise estimates of the vote share or turnout, and not on respondents’ reported behaviour, which is more error-prone. Aggregate-level studies, however, do not allow strong conclusions to be drawn about individuals’ political behaviour. Substantial variation in citizens’ characteristics and attitudes can simply not be accounted for by an analysis of aggregate turnout patterns. Observed differences between countries that are ascribed to different electoral rules might therefore well be spurious.

It is such concerns that have motivated scholars to develop datasets that would allow the impact of electoral systems to be examined by means of individual-level analyses. The data from the CSES project, for example, consist of individual-level measures of citizens’ behaviour as well as their attitudes in a large set of democracies. By means of these data, or by pooling other survey data from different countries, research designs can be developed that allow the consequences of differences in electoral rules at the individual level to be assessed.

Assessing the impact of electoral rules at the individual level is also theoretically important, because aggregate-level differences should be grounded in individual-level mechanisms. While some macro-level findings on electoral rules are strong and robust, our understanding of the individual-level mechanisms explaining these macro-observations is still limited (Blais and Aarts 2006). A number of scholars have used the available comparative individual-level survey data to investigate such questions, and to examine the impact of contextual factors by means of individual-level data. Surprisingly, however, these studies do not always corroborate the strong macro-level evidence. The detailed analysis by Brockington (2004) of individual-level sources of turnout on CSES data, for example, did not allow him to unveil what individual-level mechanisms explain why turnout is higher under PR rules. This only adds to the theoretical puzzle: empirically we know there are strong differences between countries. If, however, there is no direct impact of country-level variables on the behaviour of citizens, it remains a challenge to explain the occurrence of these differences, since other causal mechanisms must be at play. Electoral rules might, for example, have an effect because they affect political actors’ behaviour. Hooghe et al. (2006), for example, suggest that the effects of electoral reform (specifically the introduction of an electoral threshold in Belgium) cannot be explained by the effect it had on individual voters, but solely by the effect it had on the strategic and anticipatory behaviour of political parties. While on an aggregate level one might draw the conclusion that changing the rules would have had an effect on voters’ behaviour, this clearly was not the case.

As an additional advantage, individual-level comparative research allows for looking beyond the direct effects of electoral rules on citizens’ political behaviour. Individual-level theories of the vote choice usually depart from quite general models and assume that some traditional vote choice determinants, such as partisanship or the economy, affect voters’ decisions in the same way across democracies. By combining individual-level data from different countries and settings, one can relax this assumption, and verify whether citizens’ vote choice calculus is effectively the same regardless of the electoral rules. Doing so, we gain insights into the indirect effects of electoral systems, as well as in how electoral rules moderate individual-level mechanisms. Previous work along these lines that examined the indirect and moderating impact of electoral rules, however, has not given indications of a strong impact of electoral rules. This observation has led Thomassen (2014b: 19) to conclude that ‘Political institutions are less relevant for people’s attitudes and behaviour than often presumed’. If this observation is correct, it would imply that (changing) electoral systems should not have a major impact on the way voters ma...

Table of contents

• Cover
• Half Title
• Series
• Title Page
• Copyright
• Contents
• Citation Information
• Notes on Contributors
• 1. Do electoral rules have an effect on electoral behaviour? An impact assessment
• 2. Candidate-centred electoral systems and voter turnout
• 3. Rules, institutions and the economic vote: clarifying clarity of responsibility
• 4. The electoral consequences of austerity: economic policy voting in Europe in times of crisis
• 5. Ticket-splitting in mixed-member systems: on the importance of seat linkage between electoral tiers
• 6. Too much choice, too little impact: a multilevel analysis of the contextual determinants of preference voting
• 7. The impact of compulsory voting on inequality and the quality of the vote
• Index