GENERAL NOTICE

In January 2025, this online platform will be integrated into Boomportaal (www.boomportaal.nl), after which this platform will be discontinued. From that moment on, this URL will automatically redirect to Boomportaal.

DOI: 10.5553/PLC/258999292019001003004

Politics of the Low CountriesAccess_open

Article

Split-Ticket Voting in Belgium

An Analysis of the Presence and Determinants of Differentiated Voting in the Municipal and Provincial Elections of 2018

Keywords split-ticket voting, local elections, voting motives, Belgium, PR-system
Authors
DOI
Show PDF Show fullscreen
Abstract Author's information Statistics Citation
This article has been viewed times.
This article been downloaded 0 times.
Suggested citation
Tony Valcke and Tom Verhelst, "Split-Ticket Voting in Belgium", Politics of the Low Countries, 3, (2019):205-226

    This article tackles the particular issue of split-ticket voting, which has been largely overlooked in Belgian election studies thus far. We contribute to the literature by answering two particular research questions: (1) to what extent and (2) why do voters cast a different vote in the elections for the provincial council as compared to their vote in the elections for the municipal council?
    The article draws on survey data collected via an exit poll in the ‘Belgian Local Elections Study’, a research project conducted by an inter-university team of scholars.
    Our analysis shows that nearly 45% of the total research population cast a split-ticket vote in the local elections of 2018. However, this number drops to one out of four if we only consider a homogenous party landscape at both levels by excluding the numerous votes for ‘local’ lists (which occur mostly at the municipal level). This finding underlines the importance of accounting for the electoral and institutional context of the different electoral arenas in research on split-ticket voting in PR systems. In the Belgian context, split-ticket voting in 2018 also differed between the different parties and regions. Furthermore, it was encouraged by a higher level of education and familiarity with particular candidates. This candidate-centred and strategic voting was matched by party identification and the urban municipal context favouring straight-ticket voting. Other factors such as region, a rural municipal context and preferential voting seemed more relevant to determine voting for local parties than using the instrument of split-ticket votes as such.

Dit artikel wordt geciteerd in

    • 1 Introduction

      In Belgium, local elections for the municipal and provincial council have coincided since 1994.1xThis article uses the term ‘local elections’ to refer to both municipal and provincial elections. Nevertheless, both elections differ substantially in terms of stature, public attention and political relevance. Partly due to the decreasing political relevance and competence of the provincial level, provincial elections are often seen as ‘second order’ and receive far less public attention (Valcke & Reynaert, 2013; Valcke, Reynaert, & Steyvers, 2011). On the other hand, they are more nationalized in terms of the political parties on offer (Marien et al., 2018). Municipal elections, on the other hand, are characterized by the strong presence of local parties and a substantial amount of preference voting, hinting at closer connections between the electorate and the political candidates (Blaise, Faniel, & Istasse, 2018). Considering these differences, the Belgian local elections could be a fruitful arena for studying the presence and determinants of split-ticket voting, the practice when voter i votes for party x in the municipal election and votes for party y in the concurrent provincial election.
      Thus far, the Belgian case has been absent in research on split-ticket voting. Our contribution aims to fill this gap, explaining two basic research questions. The first question asks how (i.e., to what extent) voters have cast a different vote in the elections at the provincial level compared to the elections for the municipal council in 2018. This descriptive analysis will account for differences between the Flemish and the Walloon region, as well as differences in the party political landscape. The second research question seeks to explain why voters split their vote between elections that are taking place at the same time. Drawing on election studies and the specific split-ticket voting literature, this explanatory analysis assesses the impact of a series of independent variables at the ‘individual level’ (gender, age, education, voting behaviour, left-right orientation, link with political satisfaction, party identification) and at the ‘structural level’ (region, type of municipality) on split-ticket voting.
      Studying split-ticket voting in Belgian local elections can be relevant in different ways. Most generally, it can enrich our insights into the core process of democracy. Indeed, although other forms of (more direct) democratic participation have been on the rise for some time, voting can still be regarded as the democratic standard in Western liberal democracies where political participation is primarily based on the competition between political parties and political candidates (Dalton, 2008; Goodin, 2008; Kersting, 2013). Secondly, our article adds to the specific literature on split-ticket voting by studying this aspect of electoral behaviour in the context of local elections in a proportional electoral system and testing specific theoretical determinants such as voters’ social characteristics and their political attitudes and behaviour. As such, the article broadens the strand of split-ticket voting research, which thus far has mostly been developed in the context of mixed-member systems. Finally, studying local government and local elections is important to further our understanding of the Belgian political system as a whole. Although the importance of local government in Belgium should not be overestimated (see below), local elections do play an important role in Belgian politics. Local politics are considered to be the closest to the citizen with a high level of visibility and effectiveness (Dassonneville, Hooghe, & Marien, 2013). Moreover, both the municipal and provincial government are rewarded with higher levels of trust and satisfaction (among citizens as well as companies) than the regional, federal or European governments (VRIND, 2014, 2016, 2017). Politicians operating at higher government levels recognize this and do not hesitate to participate in local elections (dual mandate holding is widespread in Belgium, see Van de Voorde, 2019).2xNot only do MPs value the importance of the local-level elections, but so do members of the regional and federal cabinet; in the 2018 local elections, only two out of nine Flemish ministers did not figure on a candidate list for the municipal or provincial elections, and at the federal level only three out of 18 ministers (including four secretaries of state) could withstand the temptation of the local political arena. The importance of ‘local’ elections also refers to their status as a test case or popular vote for policies (and for politicians) at higher government levels. This is especially the case when the timing of different elections is not too far apart; for example, the federal and regional elections of 26 May 2019 took place only about half a year after the elections for the municipal and provincial councils on 14 October 2018. Moreover, the results of the provincial elections are seen as a good reflection of, or even ‘predictors’ for, the electoral strength of parties at the regional and national level due to similar electoral rules and attitudes (Marien et al., 2018; Valcke & Reynaert, 2013).3xI.a. ‘provincial’ districts for the elections of the federal and Flemish regional parliament, D’Hondt system for the distribution of seats, a more homogenous and ‘nationalized’ party landscape. In recent elections, regional parties have invested quite substantially in their local embedding, thus contributing to the so-called ‘nationalization’ of local politics and local elections (Dassonneville et al., 2013; Marien, Dasonneville, & Hooghe, 2015).
      The article is organized as follows. The next section presents a short overview of the literature on voting motives and split-ticket voting, scrutinizing how this concept is defined and which crucial aspects we need to take into account. Section 3 outlines the specific institutional context of local government and local elections in Belgium. Data and methods are discussed in Section 4. In the fifth section we answer our descriptive research question, while Section 6 tackles the subsequent, explanatory goal of the article. Finally, conclusions and a discussion are presented in Section 7.

    • 2 Theoretical Perspectives on Split-Ticket Voting

      For many years, scholars have been puzzled over the question as to why voters vote differently in elections. This is especially the case when those elections take place at the same time and voters are provided with the opportunity to cast a so-called split-ticket vote, i.e., a vote for different party lists on both (or, in the case of multiple elections, on all or some) concurrent elections. In this section we give an overview of the motives which theoretically underpin voters’ specific choices as mentioned in election research and in the split-ticket voting literature. We will use these motives to operationalize the specific variables included in our explanatory analysis. Yet first, we define and operationalize the article’s central concept of split-ticket voting.

      2.1 A Working Definition

      Ticket splitting can be – and has been – studied at various levels: across constituencies, across levels of government, across parties, across electoral rules, across ballot formats and across time. Wang, Lin and Hsiao (2016, p. 194) explain the practice as follows:

      For instance, when elections for the presidency and for the House of Representatives in the United States take place at the same time, voters may choose to vote in the manner of a straight ticket by casting both ballots for candidates from the same party, or to split the ticket by voting for candidates from different parties on different ballots.

      Assessing ticket splitting in a comparative direction necessitates expanding the definition of ticket splitting commonly used in the U.S. literature. We support the idea of Burden and Helmke (2009, p. 2) for a minimalist definition: “a ticket is split if voter i votes for party j in contest r and votes for party –j in some other contest.” This approach allows us to accommodate the diversity in varying institutional contexts in comparative analysis (even cross-nationally). At the same time, we want to make a clear distinction from research on volatility, the practice of switching party preferences of voters between elections (Dassonneville et al., 2013).
      Starting with this general definition, we take into account that ticket splitting may vary along two key dimensions: the first dimension has to do with direction (horizontal versus vertical), and the second dimension refers to timing (concurrent versus non-concurrent elections). Taking institutional structures into account, ticket splitting occurs either horizontally or vertically, whereby horizontal ticket splitting is only possible when multiple, equivalent offices are contested. Vertical ticket splitting occurs where elections are held to fill offices at different levels of government. In the context of our study, which deals with the Belgian local elections at the municipal and provincial level held in 2018, we thus operationalize the concept as a type of vertical ticket splitting in concurrent elections.

      2.2 General Voting Motives

      In order to find theoretical explanations of split-ticket voting, we first look into general theories on voting behaviour. Theories on voting behaviour try to give insight into the wide variety of voting motives, which seem to be evolving over time. Hague, Harrop, and McCormick (2016) structured this wide variety in three clusters of determinants: the social characteristics of voters (such as gender, age, education, profession, religious opinion, membership of societal organizations), their values and attitudes (e.g., individualism, post-materialism, position on cleavages) and their assessment of specific policy themes (i.e., issue voting).
      Two basic streams in the literature on voting behaviour are party identification and issue voting.4xWe will not address the personalization and image of political parties, partly because of pragmatic reasons, and partly because this is considered an outcome rather than a reason for voting (Rahat & Kenig, 2018; Wauters, Thijssen, Van Aelst, & Pilet, 2018). The literature refers to party identification as the result of a socialization process in which citizens (starting from their youth) are influenced by their social environment (i.e., parents, friends and peers) and get used to certain convictions and symbols. Once they are grown up, it is very unlikely that they turn their back on the party they have come to identify with (Kesteloot, 2009). The social class or religious group voters belong to, seem to be of crucial importance in that regard. However, the importance of the social class seems to be diminishing or is at least subject to change. People with a lower educational level, for instance, used to vote for centre or left parties, while nowadays they seem to be more attracted to radical right and conservative parties (Knutsen, 2004, 2018).
      Stemming from rational choice theory, issue voting states that voters will cast an intended and informed vote for parties or candidates that best serve their interests. This has been criticized by scholars referring to rationally ignorant voters: (some) voters are not well informed on politics and are not willing to do so either. However, the supporters of issue ownership theory – albeit focusing on the supply side of elections – stress that voting behaviour can (at least partially) be explained by issue ownership (Bélanger & Meguid, 2008). This implies that voters will vote primarily for parties (or candidates) that stand closest to their own policy preferences.

      2.3 Explanations for Split-Ticket Voting

      We defined split-ticket voting as a type of voting behaviour where voters cast two ballots (or more) for (candidates from) different parties competing in concurrent elections for different political levels or positions. Besides the literature on split-ticket voting focusing on electoral behaviour in the United States – where elections for executive and legislative bodies are held simultaneously (Beck, Baum, Clausen, & Smith, 1992) – there is also substantial interest in split-ticket voting in countries with mixed-member systems for legislative elections, such as Germany, Japan or New Zealand (Karp, Vowles, Banducci, & Donovan, 2002).
      Burden and Helmke (2009) addressed the main explanations for ticket splitting in a number of clusters. Policy balancing is put forward as a first important cluster. In practice, policy is a complex combination of policy inputs from different institutions. Split-ticket voting might then result from the variable importance attributed to policy in the respective electoral arenas under study. In the case of the Belgian local elections, the two government levels (municipal and provincial) are divided along jurisdiction lines and have varying competencies (see below), which could induce split-ticket voting as the result of voters’ balancing of policy between both government levels. The institutional reform implemented in 2018 (the year of the local elections, see below), handing over person-tied competences from the provincial to the regional or municipal level thus could have induced ticket splitting as a manifestation of how voters value particular policy choices (aimed specifically at person-tied or ground-tied competences) of the different parties.
      Second, there is the broad argument of strategic voting. One of the long-standing traditions in the literature is the assessment of whether split-ticket voting is strategic or sincere. Especially studies on mixed-member systems have emphasized ticket splitting as the result of a rational, strategic action on the part of the voter.5xPlescia (2017), however, challenged the general idea that split-ticket voting can be used as a measure of strategic voting. In comparative accounts, split-ticket voting largely emerges as a function of voters voting sincerely in the PR race, but voting strategically in the single-member race to avoid wasting their vote (the wasted-vote hypothesis). Mutatis mutandis this also holds true in a PR system like the Belgian local elections where the provincial elections are much more nationalized and non-national lists play a larger role in municipal elections (see below). An alternative is the threshold-insurance hypothesis which states that in systems with a minimum threshold, voters may be inclined to ticket splitting when they fear a vote for a specific (small) party will be wasted because of the threshold.
      Partisan identification remains an important (if not primary) determinant of citizens’ voting behaviour. Research on Japan and Taiwan (Wang et al., 2016) revealed that the effect of partisan loyalty is mitigated by the institutional design and (new) electoral rules. The empirical evidence supports the mechanical and psychological effects (characterized by Duverger) of citizens’ electoral behaviour. While supporters of major political parties tend to cast sincere straight tickets, identifiers with minor political parties have a higher likelihood of becoming ticket splitters worrying that their votes for trailing candidates may be wasted. To be consistent with partisan loyalty, identifiers with minor parties tend to be sincere in their PR votes but back viable district candidates of major parties. On the other hand, the concept of candidate-centred voting implies that voters’ ideological preferences may be overruled by their preferences for (the characteristics of) specific candidates (from another list).

    • 3 Belgian Local Elections Context

      The overview of the literature has provided us with several interesting findings and theoretical determinants of split-ticket voting. Nevertheless, we must be aware of the effects of different electoral systems. While much of the theoretical insights stem from mixed-member systems, Belgian local elections can be classified as PR systems with one single vote (Deschouwer, 2012).6xThe principle of proportional representation was introduced in 1899, which makes Belgium the very first country to use a list PR system. In this system, voters may choose to cast a list vote and/or a preferential vote for one or more candidate(s) on the same list. The latter is crucial for the interpretation of our data. Furthermore, the electoral arena of the Belgian local elections holds some specific features that could bear upon the theoretical insights developed above. Before we turn to our analysis of split-ticket voting and its determinants in the Belgian local elections of 2018, this section sets out the specific institutional context of local government and local elections in Belgium.
      First, it is important to understand the differences between the municipal and provincial level and their respective local elections. In Belgium, the importance of local government should not be overestimated. In 2013, the local sector represented only 7% of the country’s GDP and 13.2% of the total public sector expenditure (Belfius, 2017, p. 1; CEMR, 2013, p. 7).7xThe strength of Belgian local government at the municipal level is exemplified by the ‘medium low’ rating in the Local Autonomy Index by Ladner et al. (2019). These authors created a new index of local autonomy covering 39 countries between 1990 and 2014. Given the multidimensionality of local autonomy, they used 11 standardized variables that could be combined into seven dimensions: legal autonomy, political discretion, policy scope, financial autonomy, organizational autonomy, access and non-interference. These seven dimensions, however, are not of equal importance, and a weighing procedure put the emphasis on political discretion and financial autonomy. Traditionally, Belgian local government has been labelled as an example of the so-called Franco type, which stands for a type of local government holding few competences and little discretion but ready access to the central government (mainly through party political contacts) (Wayenberg, De Rynck, Steyvers, & Pilet, 2011). Within this local government setting, the second tier (provinces) plays a more modest role than the first tier (municipalities) in terms of competences. Municipalities have general competences and their policy scope increased over the last decades in the domains of public health, social affairs and public safety (De Rynck & Wayenberg, 2010). Belgian provinces (in Flanders and in Wallonia), on the other hand, have been subject to debate and reforms which even questioned their very existence (Valcke et al., 2011; Valcke & Verhelst, 2017).
      Apart from its general outlook and specific competences, both Belgian local government levels differ in terms of their electoral rules and voter attitudes as well. One of the general electoral features in Belgium is that compulsory voting is used for elections at every political level. This means that the voter, once he/she has reached the age of 18, must show up at the polling station. Still, it remains possible to cast a blank or invalid vote (Deschouwer, 2012). Although Belgian municipal and provincial elections have been concurrent since 1994, the electoral rules in use are not entirely the same (Veny, Warnez, & Callens, 2017). The differences concern the system for the distribution of seats (Imperiali in municipal elections and D’Hondt in provincial elections8xThe D’Hondt method is known to be favourable to the smaller political parties.) and the voting right for EU citizens without Belgian nationality (which is not applied in provincial elections; Valcke et al., 2011).
      What is more important for our research on ticket splitting is the omnipresence of non-national (‘local’) lists in municipal elections (Steyvers & De Ceuninck, 2017). In the previous elections, non-national lists presented themselves to the voter in 70% of the municipalities while they were almost absent in provincial elections (Heyerick & Steyvers, 2013; Steyvers, & Heyerick, 2017; Valcke & Reynaert, 2013). This trend continued in 2018. Typical local lists such as ‘municipal interests’ (‘gemeentebelangen’/‘intérêt communal’) and ‘list of the mayor’ (lijst van de burgemeester’/‘liste du bourgmestre’) figured on many ballots again, while the participation of ‘citizen lists’ (‘burgerbelangen’/‘intérêts citoyen’) was particularly striking and more novel. At the provincial level, on the other hand, the electoral landscape is more homogenous and dominated by the national parties. In terms of the electoral competition between those parties, we also have to bear in mind that Belgian local government is characterized as a party government in which parties dominate political life to a strong extent (Steyvers, Reynaert, & Block, 2010). Party loyalty, however, has been increasingly challenged in a post-modern context marked by high voter volatility and the de-pillarization9xThis concept refers to the evolution in which traditional parties (Christian-democrats, liberals, socialists) started to lose their all-embracing grip on society (which was partly reflected by a loyal party vote as well). of Belgian politics (Dassonneville et al., 2013; Deschouwer, 2012).
      Additionally, the municipal elections are characterized by the stronger use of preference votes as compared to their provincial counterparts. Preferential voting in municipal elections amounts to 70% while in provincial elections this figure is around 55% (Valcke et al., 2011; Wauters, Thijssen, Van Aelst, & Pilet, 2018; Wauters, Verlet, & Ackaert, 2010).10xIn Flanders, for instance, 4,511,166 voters took part in the 2018 elections resulting in 4,442,414 preferential votes for the provincial elections, while 11,954,465 preferential votes were given in the municipal elections (data retrieved from the official website of the Flemish government, www.vlaanderenkiest.be). The higher share of preference votes in municipal elections is probably strengthened by the incumbency effect which tends to reward popular mayors in the municipal elections with additional preference votes (Audrey, Pilet, Depauw, & Van Aelst, 2013; Reynaert, Steyvers, & Van de Voorde, 2019).
      A second important theoretical distinction is to be made between the different regions of the country. In Belgium, the regions (Flanders, Wallonia, Brussels Capital) hold the formal authoritative competence over the ‘local sector’ which includes municipalities, provinces,11xSince the provinces only report under the Flemish and Walloon regions respectively, our analysis does not include the Brussels Capital region. public centres of social welfare, police zones and (non-commercial) inter-municipal cooperations. The regional divide often cross-cuts the different factors outlined above. The institutional debate on the future of the provinces, for instance, was more outspoken in the Flemish region. This also remained the case during several consecutive legislative eras (De Ceuninck, Steyvers, & Valcke, 2016). The most recent reform (implemented in 2018, the year of the local elections) concerns the downsized competence framework limiting provincial competencies to the so-called ground-tied competencies and handing over their person-tied competences to the regional or municipal level (Valcke, 2017, 2019).12xIn this instance the number of provincial councillors – which was already downsized earlier – was further diminished to half of the total number prior to the 2018 elections. Although the Walloon provinces have been the subject of debate as well, this was not the case when the 2018 local elections took place. As such, ideological or policy-oriented voting at the provincial level could have been more likely in Flanders, increasing the theoretical chances of split-ticket voting.
      With regard to the omnipresence of local lists in the municipal elections, it is important to stress that the number of communes in which the parties have used their ‘national’ brand name has mainly dropped significantly in Wallonia (Blaise, Demertzis, Faniel, Istasse, & Pitseys, 2018).13xCDH for example presented a list under its own brand name in only 24 municipalities (as compared to 69 in the 2012 elections), PS used the national brand name in 99 municipalities (128 in 2012), MR in 75 municipalities (89 in 2012) and Ecolo in 129 municipalities (157 in 2012) (Blaise et al., 2018, p. 35). Also, differences in the political party landscape refer to the regional divergence of the Belgian government. Since the 1970s, Belgian ‘national’ parties have split up into ‘regional’ but independent branches of the ideological formation (Deschouwer, 2012). To reflect on the election results for the different parties, one has to bear in mind that all parties in Belgium are organized as independent parties in the two regions. In other words, there is no such thing as a ‘Belgian’ socialist or liberal party. Against this background it is interesting to note that the election results for the different parties in the two regions differ. The relative importance of the political ideologies and parties in the two regions is quite different (Deschouwer, 2012), as is the case with the results in the different provinces and electoral districts (Valcke & Reynaert, 2013). Another difference with regard to the regional divergence is the varying effect of the list vote, i.e., a vote for a party, as the effect of the list vote has been reduced more in Flanders than in Wallonia (Valcke, 2006, pp. 58-71). One last regional difference we need to mention is the specific procedure for the ‘election’ of the mayor. In Wallonia, the candidate with the highest number of preference votes from the largest party in the coalition is automatically appointed as mayor (Pilet & Van Haute, 2007). As opposed to this semi-direct form of election, Flemish mayors are still appointed by the Flemish government after having been nominated by the municipal council. Nevertheless, the question remains if (and to what extent) voters can really include this information when casting their vote as the coalition is only formed after the elections. The possible effect on split-ticket voting is thus unclear.
      We conclude this section by referring to two other characteristics of Belgian local elections that might have an effect on ticket splitting: the level of political participation and the electoral competition. In past elections there was a clear difference in political participation between the Flemish and Walloon provinces. It is remarkable that at each election the average number of voters who were absent or who cast a blank or invalid vote was much higher in the Walloon provinces. For instance, in 2012, the absenteeism amounted to 8.2% in the Flemish provincial elections while it was 11.3% in the Walloon provincial elections (Valcke & Reynaert, 2013). Besides, the electoral competition also used to be much stronger in Wallonia than in Flanders in the provincial elections. In the 2012 elections, a maximum of 12 lists figured on the ballot for the Flemish provincial elections (in Vlaams-Brabant, Oost- and West-Vlaanderen) while on the Walloon side the voter could choose between a maximum of 17 lists (in the provinces Liège and Hainaut).

    • 4 Data and Research Methods

      Our analysis uses data that have been collected in the context of the inter-university research project ‘Belgian Local Elections Study.’14xThe Centre for Local Politics of Ghent University coordinated a team of researchers from Antwerp University, Hasselt University, Vrije Universiteit Brussel, Université Libre de Bruxelles, Université Catholique de Louvain and Université de Namur. This project consisted of an exit poll carried out in a representative sample of 45 municipalities in Belgium on the municipal and provincial election day of 14 October 2018. In each municipality involved, two interviewers approached a random selection of voters (i.e., every fifth voter leaving the polling station) and asked them to complete a survey and a mock ballot.15xThe mock ballot replicated citizens’ specific vote (i.e., the party and/or candidates the citizens had voted for in their municipality). The response rate of this procedure was 43.6% and rendered a total dataset of 4,589 cases. After data cleaning we ended up with a research population of 3,330 unique voters.16xThis process involved deleting structural non-response and citizens who did not complete the question on municipal and/or provincial voting behaviour, as well as the respondents from the Brussels region (since provincial elections are not organized in the 19 municipalities from this region).
      The survey included a range of questions dealing with socio-demographic characteristics of the voters, voting motives and behaviour, political attitudes, preference voting and other political attitudes. With regard to our research focus, both provincial and municipal voting behaviour were surveyed. The main focus of the additional questions of the survey was on the municipal level. Yet given the context of Belgian local politics and elections, variables concerning the municipal level can be used as general predictors (e.g., satisfaction with democracy) but also as specific elements to differentiate this political arena from its provincial counterpart, hence to explain the possible use of split-ticket votes (e.g., party identification, preference voting and familiarity with a candidate on the municipal ballot). The dataset was completed by adding the region and the type of municipality of the respondents.
      First, we answer our descriptive research question by presenting a general overview of the degree of split-ticket voting in the local elections of 2018 and the way these split votes are distributed. Considering the specificity of the Belgian local political context, we refine the general image by also considering split-ticket voting in a uniform party landscape (i.e., the major parties that ran for election at both levels, hence excluding local lists17xIn the research project, local lists are specified as all lists with a non-registered/protected number.) and in a context without blank votes. Furthermore, our analysis grasps the difference in split-ticket voting between the two regions and the different political parties, as both elements have proven to be determining features of the local electoral arena in Belgium.
      Second, an explanatory analysis seeks to uncover the determinants of split-ticket voting as categorized in the three research populations outlined above (i.e., total population, uniform party landscape, uniform party landscape excluding blank votes). These determinants have been selected on the basis of the broad theoretical clusters introduced in Section 2. We organize them in three groups.
      The first group consists of the social voter characteristics such as gender, age (in categories: young, 18-34 years; old, 65+; the reference group, 35-64) and education (i.e., having obtained a university or college degree). While these variables are often treated as general determinants of voting behaviour, education might also be a further clue for issue voting as well as split-ticket voting (and its underlying mechanisms of policy balancing and strategic voting) in particular.
      The second group includes the voters’ values, attitudes and voting behaviour. Two variables reflect the particular importance of preference (candidate-centred) voting and grassroots politics in Belgian municipal elections, which could consequently increase the odds of split-ticket voting – whether or not the voters cast a preference vote in the municipal election and whether this vote was inspired by the familiarity with a particular candidate.18xGiven the main focus of the research project on the municipal elections, these questions have not been asked for the provincial elections. Party identification (i.e., whether or not voters feel close to one particular party) is a third attitudinal variable which is expected to decrease the odds of split-ticket voting. As general determinants of voting behaviour, we also include voters’ ideology (i.e., their self-reported position on a 0-10 left-right scale) and their satisfaction with democracy (i.e., rating the satisfaction with the way democracy is working in the voter’s municipality on a scale from 0-10). As a proxy for the socialization process behind party identification, our model includes the variable of whether the voter was a primary voter since the last local elections of 2012. Novices would then be expected to display a lower degree of party identification, hence having a better chance of ticket splitting.
      Finally, our model takes account of the regional differences in the institutional setting of Belgian local elections (for instance, which could impact policy balancing) and also includes the socio-economic setting of the municipal electoral context (i.e., differentiating between urban municipalities, rural municipalities and semi-urban municipalities as a reference group).

    • 5 Descriptive Analysis: Split-Ticket Voting in the Local and Provincial Elections of 2018

      Our first basic research question assesses the extent to which voters have cast a split-ticket vote (or, on the contrary, a straight-ticket vote) in the 2018 municipal and provincial elections. First, we offer a general overview of the degree to which split-ticket voting was used in the elections (Table 1). As the table shows, the overall number of split-ticket votes in the local elections of 2018 is quite high. Only slightly more voters have voted for the same party (55.3%) instead of splitting their votes between parties across the municipal and provincial level (44.7%). Additionally, we can see that split-ticket voting is more common in Wallonia (50.3%) than in Flanders (41.7%).
      Yet, the high percentage of split-ticket votes is obviously strongly determined by the different stature of both local elections. A quick look at Tables 2 and 3 demonstrates the different party political landscape at the municipal and the provincial level. Whereas local and/or alliance lists attracted a large share of the municipal votes (nearly 20% in Flanders, and almost 34% in Wallonia), this is far less the case at the provincial level (namely 2% in Flanders and 2.6% in Wallonia; these lists are presented under the header ‘other’ in the table). At the more nationalized provincial level, these local lists did not participate in a systematic and comprehensive way in the elections. Obviously, such lists are also often formally and personally distinct from the local lists or alliance lists at the municipal level. This implies that a vote for a local/alliance list at the municipal level almost inevitably leads to a split-ticket vote between both levels (an effect which is even stronger in Wallonia due to its larger share of national parties competing under a local label in the municipal elections). As such, the typical features of both electoral arenas have a considerable impact upon the total number of split-ticket voting in Belgian local elections. To refine our analysis, we therefore also take a look at split-ticket voting excluding this variegated composition of the ballot at both levels stemming from the large proportion of local lists and alliance lists at the municipal level. If we only consider a uniform party landscape comprised of votes for the non-local (‘traditional’) lists at both levels, including blank votes (N = 2,491), the number of straight-ticket votes substantially increases to 73.9%. Given the larger share of local votes in the municipal elections in Wallonia, the difference between both regions consequently decreases as well. Excluding blank votes from this overview (N = 2,331) further augments the proportion of straight votes to 75.5%.
      In summary, a cautious conclusion would be that in the ‘uniform’ part of the electoral party landscape, nearly one-fourth of the Belgian voters cast a split-ticket vote in the local elections of 2018. Taking the overall composition of the ballots (and the different electoral arenas they imply) into account augments the number of split-ticket votes to nearly 45%.

      Table 1 General overview of split-ticket voting in the 2018 local elections in Belgium
      Straight ticketSplit ticketN
      General population55.344.73,330
      Flanders 58.3 41.7 2,173
      Wallonia 49.7 50.3 1,157
      Uniform population (Votes for non-local parties)73.926.12,491
      Flanders 73.1 26.9 1,732
      Wallonia 75.8 24.2 759
      Uniform population excluding blank votes75.524.52,332
      Flanders 74.1 25.9 1,643
      Wallonia 79.0 21.0 689

      Table 2 Split-ticket voting in the 2018 local elections in Flanders
      Flanders
      Provincial 2018 (%)
      Municipal 2018 (%)CD&VGroenN-VAOpen VLDPVDASp.aVlaams BelangBlankOtherN (municipal votes)% (municipal votes)
      CD&V 65.8 8.68.86.11.54.72.21.70.5 407 100
      Groen 5.6 75.1 5.63.66.42.00.40.80.4 249 100
      N-VA 2.82.3 80.5 6.30.50.75.31.20.5 431 100
      Open VLD 4.06.312.1 68.4 1.11.15.21.10.6 174 100
      PVDA 4.310.60.02.1 72.3 8.50.02.10.0 47 100
      Sp.a 5.27.44.13.07.0 66.8 3.72.60.4 271 100
      Vlaams Belang 1.11.15.30.00.04.3 87.2 1.10.0 94 100
      Blank 3.04.56.03.01.51.56.0 73.1 1.5 67 100
      Alliance 3.733.325.97.43.79.33.71.911.1 54 100
      Local list 14.816.610.628.52.49.55.05.07.7 379 100
      N (province votes) 378 353 492 303 91 260 159 94 43 2,173 100
      % (of total province votes 17.4 16.2 22.6 13.9 4.2 12.0 7.3 4.3 2.0

      Second, we refine our analysis by examining split-ticket voting across the different political parties. The results are presented per region in Table 2 and Table 3 respectively. Each table displays the percentage of voters per party combination at both levels. The underlined numbers represent straight-ticket votes (votes for the same party at both levels), the other numbers display the split-ticket votes, i.e., the proportion of party voters in the municipal elections (rows in the table) that voted for a different party in the provincial elections (columns in the table). In addition to the different parties that ran for both elections, the table also includes the blank and invalid votes (labelled ‘blank’).
      In the Flemish region, where split-ticket voting is generally lower compared to Wallonia,19xYet we find a strong and significant association between the municipal vote and the provincial vote in Flanders (Cramer’s V = 0.577, p = 0.000) and in Wallonia (Cramer’s V = 0.564, p = 0.000). it is surprising to note that the three traditional parties (CD&V, Open VLD and Sp.a) appear to have the least loyal voters across both electoral arenas. From among their municipal electorate, around 35% cast a vote for a different party in the provincial elections. This could be considered a further proof of the so-called de-pillarization of Belgian politics, according to which the traditional parties have lost their ancient, loyal electorate. Parties with a more explicit political profile, on the other hand, seem to better succeed in attracting or retaining a loyal electorate. The largest share of straight voting is to be found both left (Groen and PVDA) and right (N-VA and Vlaams Belang) of the political centre. Another general pattern involves splitting votes between left-wing parties (PVDA, Sp.a and Groen) and between right-wing parties (N-VA, Open VLD, Vlaams Belang) respectively, with the Green party being an exception to this trend and CD&V confirming its position as the party in the centre of the political spectrum. The ‘local’ votes from the municipal level are cast in different directions in the provincial elections. Votes from alliance parties predominantly go to the Green (Groen) and regionalist (N-VA) party in the province, while voters from local lists at the municipal level tend to vote quite often for the liberal party (Open VLD) in the province. The two most extreme parties (PVDA on the left and Vlaams Belang on the right) clearly profit less in the provincial elections from split-ticket votes coming from local lists in the municipal elections.

      Table 3 Split-ticket voting in the 2018 local elections in Wallonia
      Wallonia
      Provincial 2018 (%)
      Municipal 2018 (%)cdHDéFiEcoloMRPSPTBPPBlankOtherN (municipal votes)% (municipal votes)
      cdH 64.6 0.018.84.210.40.00.02.10.0 48 100
      DéFi 0.0 67.4 4.72.316.39.30.00.00.0 43 100
      Ecolo 4.31.1 68.8 6.511.82.20.01.14.3 93 100
      MR 4.20.83.4 81.5 5.91.70.01.70.8 119 100
      PS 3.71.15.62.2 79.8 5.20.02.20.0 267 100
      PTB 1.70.05.20.010.3 80.2 0.00.91.7 116 100
      PP 4.20.08.30.00.08.3 70.8 4.24.2 24 100
      Blank 0.01.713.311.710.010.01.7 51.7 0.0 60 100
      Local list 15.25.723.119.821.16.20.33.15.7 389 100
      N (province votes) 112 57 200 196 343 147 19 55 30 1,159 100
      % (of total province votes) 9.7 4.9 17.3 16.9 29.6 12.7 1.6 4.7 2.6

      The voting pattern in Wallonia is remarkably different. Apart from the higher share of overall split-ticket votes, which is partly caused by the fact that more national parties ran under a local header in the municipal elections, the two traditional government parties (the liberal MR and socialist PS) have the highest number of straight-ticket votes together with the far-left PTB. The votes of the remaining parties are more often split across both elections. Another remarkable figure is the low percentage of blank straight votes (51.7%). More than in Flanders, citizens who cast a blank vote in the municipal elections tend to vote for one of the big parties (PS, MR, PTB, Ecolo) at the provincial level in Wallonia. Municipal votes for a local list are distributed quite evenly over the traditional parties (cdH, MR, PS plus Ecolo) in the provincial elections, which also relates to the absence of a strong regionalist and far-right party in this region. The remaining party patterns within the group of split-ticket votes are less pronounced. A trade-off from cdH-votes to Ecolo can be observed, yet it is above all the PS which attracted municipal votes from different parties (the left PTB and Ecolo, but also cdH and DéFi) in the provincial election.

    • 6 Explanatory Analysis: Determinants of Split-Ticket Voting in Belgian Local Elections

      Having outlined the general patterns of split-ticket and straight-ticket voting in the Belgian local elections of 2018, the final part of the analysis scrutinizes its determinants. Building on the insights from the descriptive analysis, we run three separate multivariate logistic regression analyses that determine the odds of casting a split-ticket vote or not. In the first, the dependent variable is the total population of split-ticket votes. Secondly, we assess the split-ticket votes for a uniform party landscape, i.e., only including voters who have cast a vote for one of the non-local party lists in both elections (including blank votes). The third and final regression analysis studies split-ticket voting in the uniform party landscape excluding blank votes as well, since the latter forms a particular sub-group of voting which might affect split-ticket voting to a different extent (for instance, those votes might be protest votes or uninformed/uninterested votes, which could relate differently to theoretical determinants such as education or political satisfaction and rational voting).
      The independent variables are clustered in three groups: individual characteristics of the voters (gender, age, education), political attitudes and behaviour (ideological self-placement, preferential voting in the municipal elections, familiarity with a candidate for the municipal elections, satisfaction with democracy, being a novice voter) and structural features of the electoral arena (region, type of municipality). Having tested our explanatory model for multicollinearity, the results of the analyses are presented in Table 4.

      Table 4 Determinants of split-ticket voting: A multivariate logistic regression analysis
      Total population (split-ticket vote)Uniform population (split-ticket vote)Uniform population excluding blank votes (split-ticket vote)
      Personal characteristics
      Gender (female)0.9900.9821.091
      Age (ref: 35-64 yrs.)
      Young 1.0081.0961.128
      Old 1.0851.1131.156
      Education (higher) 1.554*** 1.356*** 1.356**
      Political attitudes & behaviour
      Preference vote (yes) 1.336*** 1.1051.196
      Familiarity (yes) 1.343*** 1.249 1.322**
      Party identification (yes) 0.603*** 0.473*** 0.441***
      Ideology (0-10)0.9990.9650.966
      Satisfaction democracy (0-10)1.0220.9981.006
      Primary voter since 2012 (yes) 1.358* 1.2321.215
      Structural characteristics
      Region (Wallonia) 1.349*** 0.9130.797
      Municipality type (ref: semi-urban)
      Urban 0.573*** 0.551*** 0.549***
      Rural 2.443*** 1.2531.270
      Model
      N2.3061.7021.589
      Nagelkerke R2 0.111*** 0.065*** 0.076***
      Constant 0.278*** 0.6330.596

      * p < 0.10, ** p < 0.05, *** p < 0.01

      With regard to the voters’ personal characteristics, the Belgian context shows that age or gender do not determine split-ticket voting in a significant way. The effect of education, on the other hand, is significant and systematic. The fact that highly educated voters are more likely to cast a split-ticket vote in concurrent elections seems to confirm the argument derived from rational choice theory stating that split-ticket voting (partly) results from a strategic choice at the ballot. A second important determinant is party identification. Much in line with our theoretical expectations, the analysis shows that voters who strongly identify with a particular party tend to support this party irrespective of the electoral arena at stake. More general political attitudes such as ideological orientation and satisfaction with democracy do not have a significant impact upon split-ticket voting.
      Other patterns in our data could be connected to the grassroots nature of local politics and elections in Belgium. For instance, familiarity with a particular candidate not only determines preferential voting, it also increases the odds of split-ticket voting in local elections. In other words, close connections between voters and candidates and the resulting candidate-centred voting might thus compensate for party alignment and its corresponding straight-ticket voting. The latter is also significantly higher in an urban context in which close and personal connections between voters and candidates are less likely.
      The remaining significant results seem evocative of the particular electoral context in Belgian local politics. Indeed, the positive effects of preferential voting in municipal elections, residing in a rural municipality and residing in the Walloon region only applies to the total research population of the elections. Conversely, the chance of split-ticket voting as such is, thus, not significantly higher in Wallonia, in rural municipalities and in the case of preferential voting for municipal elections in a uniform party landscape which does not consider specific local lists. This suggests that region, rural context and preferential voting could be seen as determinants of local list voting in the Belgian 2018 elections – which almost inevitably leads to split-ticket voting in the Belgian context – rather than as determinants of the pure system of split-ticket voting as such. Finally, we find a (less significant) impact of primary voters, who appear to be more inclined to cast a split-ticket vote as the socialization theory would expect.

    • 7 Conclusion and Discussion

      Although (more) direct forms of participation have been on the rise for some time, in modern democracy voting still remains the primary means by which citizens participate in politics. It is not surprising then, that voting and elections keep garnering scholarly attention. But while much is known about voters in stable, two-party and presidential systems, there is still a lot to uncover to understand how citizens express their preferences through the ballot box across a variety of institutional contexts. Elections frequently call upon voters to make multiple decisions. Often voters do simply vote for the same party in different elections, but the volatility of elections in Belgium and around the world suggests that this is not always the case. Research on how often and why voters engage in split-ticket voting in other parts of the world challenges the theoretical underpinnings of the voting literature, while at the same time widening our knowledge of the enormous institutional diversity.
      This article has shed light on the particular context of PR systems by studying split-ticket voting in Belgian local elections. The analysis showed a surprisingly high proportion of split-ticket votes between the municipal and provincial elections of 2018. This practice is higher among the highly educated part of the electorate and among voters familiar with particular candidates. Our study therefore confirms candidate-centred and strategic voting motives as important determinants of split-ticket voting. On the other hand, split-ticket voting is less likely among the group of voters displaying strong party identification as well as in an urban context where the distance between voters and candidates will be larger. Regional differences seem to confirm waning party loyalty from voters of the traditional parties in Flanders, yet less in Wallonia. The diverging party streams between both regions is remarkable as well, although we should keep in mind that Walloon national parties more often participated in the municipal elections under the label of a local list.
      The most important contribution of this article to the literature probably stems from the clear impact of the electoral and institutional landscape of the different political levels under study. Indeed, scrutinizing split-ticket voting in a uniform party landscape changes the picture of split-ticket voting substantially. In a divergent electoral context marked by the strong presence of local lists and local list votes at the municipal level, the effect of some determinants (preferential voting, a rural municipal context, region) seems indirect. Rather than promoting the odds of split-ticket voting as such, those elements increase the chance of local list voting at one particular level, which subsequently increases the possibility of a split-ticket vote as those particular local lists tend to compete on this electoral level solely. In this line of reasoning, one could expect a similar effect from the electoral bonus of incumbent mayors in municipal elections. In our view, distinguishing between pure and indirect determinants of split-ticket voting provides a more realistic picture of the electoral process which takes into account the fundamental importance of the electoral stature of the concurrent elections. More fine-grained analyses, including interaction effects between different determinants, could therefore take our understanding of this phenomenon another step forward.

    • References
    • Audrey, A., Pilet, J.-B., Depauw, S., & Van Aelst, P. (2013). De partij, de leider of een ‘gewone’ kandidaat? Het gebruik van de voorkeurstem bij de gemeenteraadsverkiezingen van 2012. In R. Dassonneville, M. Hooghe, S. Marien & J. B. Pilet (Eds.), De lokale kiezer. Het kiesgedrag bij de Belgische gemeenteraadsverkiezingen van oktober 2012 (pp. 117-140). Brussel: ASP.

    • Beck, P. A., Baum, L., Clausen, A. R., & Smith, C. E. (1992). Patterns and sources of ticket splitting in subpresidential voting. American Political Science Review, 86(4), 916-28.

    • Bélanger, E., & Meguid, B. M. (2008). Issue salience, issue ownership and issue based vote choice. Electoral Studies, 27, 477-491.

    • Belfius (2017). De financiële situatie van de Vlaamse lokale besturen. Budget 2017. Brussel: Belfius.

    • Blaise, P., Demertzis, V., Faniel., J., Istasse, C., & Pitseys, J. (2018). La préparation des élections communales et provinciales du 14 octobre 2018. CRISP, 2381-2382, 165p.

    • Blaise, P., Faniel, J., & Istasse, C. (2018). Les résultats des élections provinciales du 14 octobre 2018. CRISP, 2386-2387, 115p.

    • Burden, B. C., & Helmke, G. (2009). The comparative study of split-ticket voting. Electoral Studies, 28, 1-7.

    • CEMR (2013). Local and regional Europe. Factsheets. A figure-based portrait of local and regional Europe. Brussels: Council of European Municipalities and Regions.

    • Dalton, R. J. (2008). Citizenship norms and the expansion of political participation. Political Studies, 56(1), 76-98.

    • Dassonneville, R., Hooghe, M., & Marien, S. (2013). Partijtrouw en volatiliteit in lokale verkiezingen. In R. Dassonneville, M. Hooghe, S. Marien & J. B. Pilet (Eds.), De lokale kiezer. Het kiesgedrag bij de Belgische gemeenteraadsverkiezingen van oktober 2012 (pp. 35-66). Brussel: ASP.

    • De Ceuninck, K., Steyvers, K., & Valcke, T. (2016). As you like it or much ado about nothing? Structural reform in local government in Belgium. In K. Dede, & U. Sadioglu (Eds.), Theoretical foundations and discussions on the reformation process in local governments (pp. 237-264). Hershey: IGI Global.

    • De Rynck, F., & Wayenberg, E. (2010). Belgium. In M. Goldsmith, & E. Page (Eds.), Changing government relations in Europe. From localism to intergovernmentalism. New York: Routledge.

    • Deschouwer, K. (2012). The politics of Belgium. Governing a divided society. Basingstoke: Palgrave Macmillan.

    • Goodin, R. E. (2008). Innovating democracy: Democratic theory and practice after the deliberative turn. Oxford: Oxford University Press.

    • Hague, R., Harrop, M., & McCormick, J. (2016). Comparative government and politics. An introduction (10th ed.) Basingstoke: Palgrave Macmillan Education.

    • Heyerick, A., & Steyvers, K. (2013). Lokaal of nationaal? Het beste van twee werelden combineren. In H. Reynaert & K. Steyvers (Eds.), De verkiezingen van 14 oktober 2012: de kracht van verankering? (pp. 99-121). Brugge: Vanden Broele.

    • Karp, J. A., Vowles, J., Banducci, S., & Donovan, T. (2002). Strategic voting, party activity, and candidate effects. Testing explanations for split voting in New Zealand’s new mixed system. Electoral Studies, 21(1), 1-22.

    • Kersting, N. (2013). Online participation: From ‘invited’ to ‘invented’ spaces. International Journal of Electronic Governance, VI(4), 270-280.

    • Kesteloot, S. (2009). Evolutie van de campagnevoering in Vlaanderen tussen 1946 en 2007. Hoe gaan de Vlaamse traditionele partijen om met veranderende kiezers? Gent: Universiteit Gent, Faculteit Politieke en Sociale Wetenschappen (PhD Thesis).

    • Knutsen, O. (2004). Social structure and party choice in Western Europe: A comparative longitudinal study. Basingstoke: Palgrave Macmillan.

    • Knutsen, O. (2018). Social structure, value orientations and party choice in Western Europe. Basingstoke: Palgrave Macmillan.

    • Ladner, A., Keuffer, N., Baldersheim, H., Hlepas, N., Swianiewicz, P., Steyvers, K., & Navarro, C. (2019). Patterns of local autonomy in Europe. Cham: Palgrave Macmillan.

    • Marien, S., Dassonneville, R., & Hooghe, M. (2015). How second order are local elections? Voting motives and party preferences in Belgian municipal elections. Local Government Studies, 41(6), 898-916.

    • Marien, S., Put, G.-J., Smulders, J., Maddens, B., Deschouwer, K., & Govaerts, I. (2018). Van dorpsstraat naar Wetstraat. Samenleving & Politiek, 25(9), 55-60.

    • Pilet, J.-B., & Van Haute, E. (2007). Eerste lessen uit de automatische benoeming van burgemeesters in Wallonië. Res Publica, 49(1), 34-45.

    • Plescia, C. (2017). On the mismeasurement of sincere and strategic voting in mixed-member electoral systems. Electoral Studies, 48, 19-29.

    • Rahat, G., & Kenig, O. (2018). From party politics to personalized politics? Party change and political personalization in democracies. New York: Oxford University Press.

    • Reynaert, H., Steyvers, K., & Van de Voorde, N. (2019). Burgemeesters in Vlaanderen: lokale voorkeurstemmen en nationale mandaten? In H. Reynaert (Ed.), Verrekijkers voor lokale besturen: een lange(re)termijnvisie?! (pp. 133-160). Brugge: Vanden Broele.

    • Steyvers, K., & De Ceuninck, K. (2013). De kracht van verandering of plus ça change? De gemeenteraadsverkiezingen van 2012 in het licht van de trends sinds de fusies van 1976. In H. Reynaert, & K. Steyvers (Eds.), De kracht van verankering? (pp.15-40). Brugge: Vanden Broele.

    • Steyvers, K., & Heyerick, A. (2017). Fifty shades of Rokkan? Reconceiving local party system nationalisation in Belgium. Croatian And Comparative Public Administration, 17(4), 509-538.

    • Steyvers, K., Reynaert, H., & Block, T. (2010). Team work or territorial war? Assessing the introduction of a management team in local government: The Flemish region of Belgium. Public Management Review, 12(1), 11-31.

    • Valcke, T. (2017). De rol van de provincie als intermediair bestuursniveau: tussen concurrentie en complementariteit (pp. 15-39). In T. Valcke (Ed.), Het beleid van de Provincie West-Vlaanderen: 1995-2017. Brugge: Provinciebestuur West-Vlaanderen.

    • Valcke, T. (2019). Provincies in Vlaanderen: toekomst van het intermediair bestuursniveau. In H. Reynaert (Ed.), Verrekijkers voor lokale besturen: een lange(re)termijnvisie?! (pp. 45-66). Brugge: Vanden Broele.

    • Valcke, T., & Reynaert, H. (2013). Analyse van een electoraal tweestromenland. De provincieraadsverkiezingen van 14 oktober 2012 in Vlaanderen en Wallonië. In H. Reynaert, & K. Steyvers (Eds.), De verkiezingen van 14 oktober 2012. De kracht van verankering? (pp.41-80). Brugge: Vanden Broele.

    • Valcke, T., & Verhelst, T. (2017), Doorstart of doodlopende straat voor de provincies? Een comparatieve analyse van de steun voor functionele en territoriale hervormingen in veertien Europese landen. In H. Reynaert (Ed.), Veranderende lokale besturen voor en door de mensen?! (pp. 113-128). Brugge: Vanden Broele.

    • Valcke, T., Reynaert, H., & Steyvers, K. (2011). Belgium. In H. Heinelt & X. Bertrana (Eds.), The second tier of local government in Europe. Provinces, counties, départements and landkreise in comparison (pp.27-55). London: Routledge.

    • Van de Voorde, N. (2019). Municipal councillors in parliament, a handicap for legislative activism? Parliamentary productivity of dual mandate-holders in the Belgian federal assembly between 1995 and 2014. Parliamentary Affairs, 1-21. doi:10.1093/pa/gsz011

    • Veny, L., Warnez, B., & Carlens, I. (2017). De regelgeving voor de lokale verkiezingen van 14 oktober 2018. Tijdschrift voor Gemeenterecht, 4, 239-256.

    • VRIND (2014). Vlaamse Regionale Indicatoren 2014. Brussel: Studiedienst van de Vlaamse Regering.

    • VRIND (2016). Vlaamse Regionale Indicatoren 2016. Brussel: Studiedienst Vlaamse Regering.

    • VRIND (2017). Vlaamse Regionale Indicatoren 2017. Brussel: Studiedienst Vlaamse Regering.

    • Wang, T. Y., Lin, C., & Hsiao Y. (2016). Split-ticket voting under MMM. In N. F. Batto, C. Huang, A. C. Tan, & G. W. Cox (Eds.), Mixed-member electoral systems in constitutional context. Taiwan, Japan, and Beyond. University of Michigan Press.

    • Wauters, B., Verlet, D., & Ackaert, J. (2010). Meer gewicht voor voorkeurstemmen op lokaal vlak: broodnodige of overbodige hervorming? Burger, Bestuur en Beleid, 6(2), 103-114.

    • Wauters, B., Thijssen, P., Van Aelst, P., & Pilet, J.-B. (2018). Centralized personalization at the expense of decentralized personalization: The decline of preferential voting in Belgium (2003–2014). Party Politics, 24(5), 511-523.

    • Wayenberg, E., De Rynck, F., Steyvers, K., & Pilet, J.-B. (2011). Belgium: A tale of regional divergence? In: J. Loughlin, F. Hendriks, & A. Lidström (Eds.), The Oxford handbook of local and regional democracy in Europe (pp. 71-95). Oxford: Oxford University Press.

    Noten

    • 1 This article uses the term ‘local elections’ to refer to both municipal and provincial elections.

    • 2 Not only do MPs value the importance of the local-level elections, but so do members of the regional and federal cabinet; in the 2018 local elections, only two out of nine Flemish ministers did not figure on a candidate list for the municipal or provincial elections, and at the federal level only three out of 18 ministers (including four secretaries of state) could withstand the temptation of the local political arena.

    • 3 I.a. ‘provincial’ districts for the elections of the federal and Flemish regional parliament, D’Hondt system for the distribution of seats, a more homogenous and ‘nationalized’ party landscape.

    • 4 We will not address the personalization and image of political parties, partly because of pragmatic reasons, and partly because this is considered an outcome rather than a reason for voting (Rahat & Kenig, 2018; Wauters, Thijssen, Van Aelst, & Pilet, 2018).

    • 5 Plescia (2017), however, challenged the general idea that split-ticket voting can be used as a measure of strategic voting.

    • 6 The principle of proportional representation was introduced in 1899, which makes Belgium the very first country to use a list PR system.

    • 7 The strength of Belgian local government at the municipal level is exemplified by the ‘medium low’ rating in the Local Autonomy Index by Ladner et al. (2019). These authors created a new index of local autonomy covering 39 countries between 1990 and 2014. Given the multidimensionality of local autonomy, they used 11 standardized variables that could be combined into seven dimensions: legal autonomy, political discretion, policy scope, financial autonomy, organizational autonomy, access and non-interference. These seven dimensions, however, are not of equal importance, and a weighing procedure put the emphasis on political discretion and financial autonomy.

    • 8 The D’Hondt method is known to be favourable to the smaller political parties.

    • 9 This concept refers to the evolution in which traditional parties (Christian-democrats, liberals, socialists) started to lose their all-embracing grip on society (which was partly reflected by a loyal party vote as well).

    • 10 In Flanders, for instance, 4,511,166 voters took part in the 2018 elections resulting in 4,442,414 preferential votes for the provincial elections, while 11,954,465 preferential votes were given in the municipal elections (data retrieved from the official website of the Flemish government, www.vlaanderenkiest.be).

    • 11 Since the provinces only report under the Flemish and Walloon regions respectively, our analysis does not include the Brussels Capital region.

    • 12 In this instance the number of provincial councillors – which was already downsized earlier – was further diminished to half of the total number prior to the 2018 elections.

    • 13 CDH for example presented a list under its own brand name in only 24 municipalities (as compared to 69 in the 2012 elections), PS used the national brand name in 99 municipalities (128 in 2012), MR in 75 municipalities (89 in 2012) and Ecolo in 129 municipalities (157 in 2012) (Blaise et al., 2018, p. 35).

    • 14 The Centre for Local Politics of Ghent University coordinated a team of researchers from Antwerp University, Hasselt University, Vrije Universiteit Brussel, Université Libre de Bruxelles, Université Catholique de Louvain and Université de Namur.

    • 15 The mock ballot replicated citizens’ specific vote (i.e., the party and/or candidates the citizens had voted for in their municipality).

    • 16 This process involved deleting structural non-response and citizens who did not complete the question on municipal and/or provincial voting behaviour, as well as the respondents from the Brussels region (since provincial elections are not organized in the 19 municipalities from this region).

    • 17 In the research project, local lists are specified as all lists with a non-registered/protected number.

    • 18 Given the main focus of the research project on the municipal elections, these questions have not been asked for the provincial elections.

    • 19 Yet we find a strong and significant association between the municipal vote and the provincial vote in Flanders (Cramer’s V = 0.577, p = 0.000) and in Wallonia (Cramer’s V = 0.564, p = 0.000).


Print this article

Politics of the Low Countries will be published by Radboud University Press. New submissions can be be submitted on our new website: https://www.plc-journal.eu/