Refining vote choice in an ethno-regionalist context: Three-dimensional ideological voting in Catalonia and Quebec

Refining vote choice in an ethno-regionalist context: Three-dimensional ideological voting in Catalonia and Quebec

Electoral Studies 40 (2015) 14e22 Contents lists available at ScienceDirect Electoral Studies journal homepage: www.elsevier.com/locate/electstud R...

842KB Sizes 1 Downloads 32 Views

Electoral Studies 40 (2015) 14e22

Contents lists available at ScienceDirect

Electoral Studies journal homepage: www.elsevier.com/locate/electstud

Refining vote choice in an ethno-regionalist context: Threedimensional ideological voting in Catalonia and Quebec Mike Medeiros a, *, Jean-Philippe Gauvin b, Chris Chhim c a

Centre for the Study of Democratic Citizenship, McGill University, 855 Sherbrooke Street West, Montreal, Quebec, H3A 2T7, Canada Department of Political Science, Universit e de Montr eal, Canada c Department of Political Science, McGill University, Canada b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 December 2014 Received in revised form 8 May 2015 Accepted 15 June 2015 Available online 23 June 2015

Politics in regions or countries with a salient ethno-regionalist cleavage take on a unique dynamic. In addition to ideological positioning on economic and social stances, centre-periphery issues add a third ideological dimension that needs to be acknowledged. Yet, electoral research has rarely given appropriate attention to these three ideological dimensions independently. This paper takes up this challenge by presenting a three-dimensional ideological model of vote choice. The three dimensions are each theoretically portrayed as having a determining and independent influence on voting. Using survey data from the Making Electoral Democracy Work project, we explore determinants of vote choice in recent subnational elections in Catalonia and Quebec. The results show that all three ideological dimensions are significant determinants of vote choice in both regions. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Voting Ideology Ethno-regionalism Catalonia Quebec

1. Introduction On 11 September 2012, up to a million people took to the streets of Barcelona under the slogan of Catalunya, nou estat d'Europa [Catalonia, a new state in Europe] to celebrate Catalonia's national holiday. More than a celebration, this was also a demonstration of popular frustration with both the severe economic crisis and the country's highest court, which two years earlier rejected reforms to Catalonia's status that would have given the region greater autonomy. Meanwhile across the Atlantic Ocean, another region was in social tumult. That same year, Quebec was also the scene of sizeable protests. Unlike the large nationalist demonstrations in Catalonia, these protests focused on tuition fee increases and general discontent toward the government. In the shadow of these events, both regions also held regionallevel elections that year. In the field of electoral studies, it is uncontroversial to say that in order to properly grasp electoral dynamics, economic and social issues cannot be overlooked. These issues are often viewed through a single-dimensional lefteright spectrum. However, the cases of Catalonia and Quebec underscore

* Corresponding author. E-mail addresses: [email protected] (M. Medeiros), jean-philippe. [email protected] (J.-P. Gauvin), [email protected] (C. Chhim). http://dx.doi.org/10.1016/j.electstud.2015.06.004 0261-3794/© 2015 Elsevier Ltd. All rights reserved.

another cleavage very salient to political competition: that of centre-periphery issues (Lipset and Rokkan, 1967). As a result, traditional electoral analyses of ethno-regionalist contexts have explored centre-periphery, economic, and social issues by condensing them into one or two dimensions.1 While this framework has facilitated interpretations of political dynamics, this intertwined treatment of what can be seen as three very distinct facets of political life can leave out important details. As a result, we propose that a three-dimensional analytical perspective, one keeping centre-periphery, economic, and social axes separate, can allow us to get a fuller grasp of electoral competition in regions where ethno-regionalist dynamics come into play.

1 Though the term ‘ethno-regional’ has been described as striving for ‘regional autonomy or secession in the name of ethnic self-determination’ (Hechter and Levi, 1979: 260), the term has also been ascribed to movements without clear ethnic distinctions (see, for example, Falkenhagen, 2013; Müller-Rommel, 2003) as well as those not limited to specific regional territories (see, for example, Jackson, 2013; Lublin, 2014). This paper follows the latter less restrictive interpretation and defines ‘ethno-regional’ as applying to movements not conditional on both precise regional limitations and clear ethnic distinctions (e.g., Roma in Slovakia, Kurds in Turkey, Padania in Italy and Bavaria in Germany). Rather, we employ the term ‘ethno-regional’ to refer to movements motivated by regional autonomy and/or ethnic representation.

M. Medeiros et al. / Electoral Studies 40 (2015) 14e22

This paper presents a three-dimensional ideological model of vote choice. Centre-periphery, economic and social ideological dimensions are portrayed as each having a theoretically important and non-negligible independent influence on a voter's decisionmaking process. More specifically, we use election survey data from Catalonia and Quebec to explore determining factors of vote choice in recent regional elections. The results support our assertion that all three ideological dimensions are significant and separate determinants of vote choice in both regions. The paper proceeds in four parts. First, we review the literature on the role that ideology plays in determining vote choice. We also touch on the uniqueness of partisan competition in ethnoregionalist contexts and why politics in these regions needs to be reconceptualised. Second, we present Catalonia and Quebec as test cases for our proposed theoretical framework. Third, we introduce the survey data used and present the results of our empirical analyses. Finally, we conclude with a discussion of the possibilities brought forth by the results for conceptualizing electoral competition in ethno-regionalist contexts. 2. Ideologies and vote choice For several decades, researchers have worked to uncover the factors underlying an individual's choice to vote for a particular party or a candidate. As a result, a rich body of literature has developed that sheds light on a complicated process. Few theories emphasize the multifaceted nature of voting better than Campbell and colleagues' (1960) ‘funnel of causality’. This theory presents a streamlined explanation of how long-term variables, such as sociodemographic characteristics, lead into more short-term factors, such as current issues and candidates, to ultimately influence a voter's final choice of party or candidate. However, the role of ideology in this process has been somewhat underappreciated. Yet, ideology has lately been presented as one of the key determinants of vote choice (Abramowitz and €l, 2014). This impact might be Saunders, 2006; Medeiros and Noe explained by recent findings, which conclude that ideological tendencies are developed quite early in one's life (Jost, 2006). If this is the case, then early ideological development would lead ideology to be located quite high in the funnel of causality (Dalton, 2008). Ideology can work as a heuristic shortcut that helps individuals navigate what can be overwhelming political landscapes by using their political likes and dislikes (Budge, 1994; Sniderman et al., 1993). In turn, this would allow individuals to develop not only positive feelings toward a party or candidate, but negative ones as €l, 2014). Voters use the ideological well (Medeiros and Noe distinctiveness of political parties to decide for which party to vote (Nie et al., 1979; Rusk, 1987). Ideology thus becomes a crucial element in vote choice. Ideology can hence be explained as a system of beliefs and ideas that help to organize socio-political realities by translating abstract notions into a group of specific policy preferences (Hinich and Munger, 1994). We suggest that these dimensions can be portrayed as policy planes upon which ideological positions represent individuals’ preferences. Most often, ideology is presented through a one-dimensional lefteright axis. As Albright (2010) explains, this is done in order to simplify the complexities of politics and to allow for comparisons. Yet, the utility of a single ideological dimension has long been called into question (Stokes, 1963). In fact, several scholars have echoed the idea that a one-dimensional ideological axis simply cannot account for important political cleavages in societies, thereby leading to a significant loss of political information (Albright, 2010; Castles and Mair, 1984; Stokes, 1963). As a result, researchers have pushed forward and analysed politics through

15

two ideological dimensions (see, for example, Inglehart, 1990). Generally, these two dimensions have been presented in terms of a lefteright economic dimension, measuring the desired degree of governmental economic intervention and acceptable income differences between individuals, and a progressiveeconservative social values dimension, measuring the acceptance of social change and individual rights. However, even a two-dimensional approach might not truly capture the nuanced political landscape of a specific country or region. Barnea and Schwartz (1998: 25) state that the ‘unique history, geography, and social circumstances of each country sometimes introduces issues into the political debate that give rise to additional dimensions on which parties adopt different ideological stances'. Consequently, the diversity and complexity of political landscapes leads us to ponder not only going beyond the onedimensional model, but also beyond the two-dimensional model. Transcending the two-dimensional model is particularly relevant for electoral landscapes where ethno-regional divides are salient. As mentioned earlier, the ‘centre-periphery’ dimension can come to the fore in sub-national entities with strong regional identities. On this dimension, preferences on the constitutional status of a region can range from strong centralized power to complete independence. Contemporary movements in regions such as Catalonia, Quebec and Scotland are real-world examples of how relations between a central state and its regions can be a very politically salient issue (Keating, 2001). For voters, this dimension provides extra political information for differentiating parties who may seem similar on the lefteright economic and pros et al., 2006). gressiveeconservative social dimensions (Pallare In many cases, this dimension is indeed incorporated into analyses of voter behaviour. However, its inclusion comes at the expense of the independence of the economic and social dimensions, which are often collapsed into a single-dimensional lefteright axis. Thus, although the centre-periphery dimension is included, many scholars still maintain a two-dimensional framework by condensing two once-separate dimensions into one.2 As mentioned earlier, the analytical utility of collapsing economic and social dimensions into one lefteright axis has been criticized. In an attempt to resolve this issue, we suggest Catalonia and Quebec as test cases for a three-dimensional framework for analysing voter behaviour that views economic, social and centre-periphery as having independent influences on vote choice. 3. Catalonia and Quebec: similar regions on diverging independentist paths In some ways, the political histories of Catalonia and Quebec have often looked similar. However, important differences do exist, especially when it comes to the state of independentism. Catalonia and Quebec are nonetheless similar enough to provide a solid comparative basis, yet diverge sufficiently to allow a thorough examination of the proposed model. 3.1. So far, yet so similar Quebec and Catalonia share important similarities. Firstly, their modern nationalist manifestations surfaced during the same period. Modern Catalan nationalism emerged in the 1960s as a

2 One notable exception is Wheatley and colleagues' (2014) work, whose findings support the importance of exploring these dimensions independently as they show that Scottish voters' policy preferences in the 2011 Scottish Parliament election converge around three distinct dimensions that are analogous to centre-periphery, economic and social axes.

16

M. Medeiros et al. / Electoral Studies 40 (2015) 14e22

struggle against Franco's dictatorship (Guibernau, 2013). This is also when contemporary Quebec nationalism took shape, in the form of popular reaction to the retrograde Duplessis regime (Martel, 1997; McRoberts, 1997). In both regions, this contestation of regressive social models has left a mark on their respective political fabrics. As a result, both Catalonia and Quebec are prone to be more proland and gressive than the rest of their respective countries (Be ~ eira, 2011; Mendelsohn et al., 2007). Lecours, 2005; Lin Both regions are also similar with regard to constitutional matters and secessionism. For the last decade, Catalonia has actively attempted to address its constitutional situation. In 2005, the Catalan government, led by the Partit dels Socialistes de Catalunya (PSC), arrived at a draft agreement with the Spanish gov~ ol (PSOE) and ernment (led by the Partido Socialista Obrero Espan including PSC ministers) on a proposal for a new Statute of Autonomy of Catalonia. The following year, Catalans widely accepted the proposal in a referendum, though with a turnout rate of just under 50%. Although accepted by Catalans, the reform package was opposed by the Partido Popular (PP) d as well as five autonomous regions d who questioned the constitutionality of the document (Martínez-Herrera and Miley, 2010). The issue eventually made its way to the Constitutional Court of Spain, which struck down many of the reform's statutes in 2010. This judgement led to a popular outcry in Catalonia and has positively affected support for independence in a dramatic fashion (Martí, 2013). This story of constitutional dissatisfaction, followed by a failed attempt at reform, resulting in greater discontent that further fans the flames of independence is similar to that of Quebec. The repatriation of the Canadian Constitution from the British Parliament in 1982 created political dissatisfaction in Quebec, leading it to refuse to ratify the Constitution up to the present day. Following this contested constitutional reform, Canada saw a decade of major constitutional debates aimed at reintegrating Quebec into the constitutional fold. This resulted in two separate reform packages: the Meech Lake Accord and the Charlottetown Accord. Both were ultimately rejected and only served to deepen the cleavage between Quebec and the rest of Canada (Russell, 2004). This set off a series of major political events in the province, which included the creation of the independentist federal-level Bloc qu eb ecois (BQ), in 1991, its major victory in the 1993 federal election (in both votes and seats in Quebec), a victory for the similarly independentist Parti qu eb ecois (PQ) in the 1994 provincial election and a referendum on secession in 1995 that the “No” side won by a narrow margin of 1.16%. Perhaps a result of having differing social and constitutional preferences from the rest of their respective countries, both Catalonia and Quebec also place great importance on their regional identities. Subjective identity, the preference between the regional identity and national identity, has been shown to influence Catalan ~ eira, 2011; Montero and Font, 1991). political preferences (see Lin Furthermore, the Spanish identity tends to be weaker in Catalonia ~ oz Mendoza, 2008).3 This same than it is elsewhere in Spain (Mun identity pattern is also present in Quebec (Blais and Nadeau, 1992; Nadeau and Fleury, 1995). Yet, in both cases, attachment to regional identities has not necessarily been at the expense of the centralstate identity; Catalans and Quebecers alike mostly live with nested identities (Martínez-Herrera, 2002; Mendelsohn, 2002). Both cases are also undergoing changes to their party systems. For decades, Catalonia and Quebec had been dominated by two  (CiU) and the PSC in the former main parties: Converg encia i Unio

3 Though economic reasons have been cited for this Catalan preference (Woolard, ~ oz and Tormos (2015) demonstrate that this is only true for politically 1989), Mun ambivalent Catalans and that identity does truly impact support for secession.

and the Parti Lib eral du Qu ebec (PLQ) and the PQ in the latter. However, vote share for the dominant parties in Catalonia has been ~ eira, 2014). Though mostly attribfalling since 1999 (Rico and Lin utable to new parties, voters have also been turning to alternatives located at either end of the ideological spectrum (Petithomme and ~ eira, 2014). These changes to Garcia, 2013; Rico, 2012; Rico and Lin the partisan landscape came to the fore in the 2012 Catalan election, when Esquerra Republicana de Catalunya (ERC) replaced the PSC as one of the top two parties in terms of seat share. 2012 was also a historic election year in Quebec, as it saw its second minority government in five years (and only the third in the province's history). This recent appearance of minority governments is seen as a direct consequence of weakened PLQ-PQ domination (Bastien et al., 2013). In fact, the PQ has not been able to garner more than 40% of the vote share since 1998.4 As in the Catalan case, new parties such as the Coalition Avenir Qu ebec (CAQ), along with its predecessor the Action D emocratique du Qu ebec (ADQ), and Qu ebec Solidaire (QS) have helped to breathe new life into Quebec's partisan system. While there are considerable similarities between the political situations in Catalonia and Quebec, important differences also exist. Most notably, support for both regions’ respective independentist movements seem to be on different, if not opposing, trajectories. 3.2. The great divide: support for independence To properly understand the variations in the nationalist movements of Catalonia and Quebec, a look at economic factors is necessary. Catalonia has long been an economic powerhouse in Spain (Miller and Miller, 1996). Its economic success has led to it being a net contributor to the Spanish budget, in much greater proportion to its size. This fiscal imbalance, which the 2006 reform attempted to resolve, has been an important grievance fueling land and Lecours, 2014; Guibernau, 2014). Catalan nationalism (Be Therefore it is understandable that the recent economic crisis was ~ eira, an important issue in the 2012 Catalan elections (Rico and Lin 2014). On the other hand, Quebec has traditionally been in a weaker economic position in the Canadian federation and has greatly benefited from equalization payments; a situation believed to have land and historically worked against support for secession (Be Lecours, 2014). However, Quebec weathered the economic crisis quite well and economic concerns did not seem to have a deterlanger, minant role on the 2012 Quebec election (Nadeau and Be 2013). The unsatisfactory economic and constitutional situations helped fuel independentism in Catalonia to new heights (Guibernau, 2013). Led by the CiU, Catalan nationalism had traditionally not been strikingly independentist, but the 2012 Catalan election seems to have marked a turning point. The main electoral issue in 2012 was sovereignty: Catalonia's dret a decidir [right to decide] ~ eira, 2014). The CiU took a more independentist turn, (Rico and Lin which coincided with a sharp rise in support for independence, and campaigned on holding a referendum on self-determination (Ibid.). The governing coalition that resulted from the election, between the CiU and ERC, is the most independentist government Catalonia has ever had (Petithomme and Garcia, 2013). Even amongst the population, Catalans, traditionally perceived as autonomists, have recently become predominantly independentist (Guibernau, 2013). This surge in independentist sentiment among elites and citizens in Catalonia led to a referendum on the region's future being called for 9 November 2014. Though this initiative was blocked by Spain's

4 The PQ had always achieved 40% or more, or very close (39% in 1985), since it first formed government in 1976.

M. Medeiros et al. / Electoral Studies 40 (2015) 14e22

Constitutional Court for being unconstitutional (as only the central government has the right to organize referenda), a non-binding popular consultation was nevertheless organized on the same date as an alternative, resulting in a third of eligible electors voting overwhelmingly for independence.5 The state of independentism is quite different in Quebec. Though the political space had been structured around the sovereignty question, the arrival of the CAQ changed things. Sensing a fatigue among Quebecers, this party explicitly contests the sovereignty cleavage. Its leader has openly stated on multiple occasions that the party was neither sovereignist nor federalist. This electoral calculation seems to have helped the party, as it was able to make the 2012 Quebec election an actual three-horse race. Sovereignty was no longer as determinant to vote choice as it had once been langer, 2013). Even the sovereignist PQ, who was (Nadeau and Be able to cobble together a slim minority government, had to promise during the campaign that it would not hold a referendum on secession if it formed the government. The PQ held on to power for 18 months before calling another election and suffering its worst defeat in terms of vote share since its first election in 1970. These disastrous results followed its federal cousin's fate, as the BQ nearly disappeared in the 2011 federal election after two decades of dominating federal politics in Quebec. Among citizens, support for sovereignty has also seen a steady decline in the last decade (Durand, 2014). We have highlighted the obvious role that the centre-periphery dimension played in politics in ethno-regionalist states. Whether it be on the rise, as in Catalonia, or in a lull, as in Quebec, the constitutional future of a region is a strong determinant of electoral politics. Yet, as we have seen, it is not the only one. The social dimension has also affected the politics of the two regions, as large segments of their populations strived to live in a more progressive society than the rest of their respective countries. Increased electoral support for parties in both regions that are farther from the centre on social issues demonstrates that there is variance on this dimension. As for the economic dimension, the recent crisis has forced economic issues to the forefront, with party positions also varying across the spectrum. Hence, we contend that in order to properly capture vote choice in ethno-regional states that these three dimensions be examined both simultaneously and independently. As a result, we hypothesize that centre-periphery, economic, and social ideological dimensions will have distinct influences on vote choice in Catalonia and Quebec.

4. Data In order to test the premise that we put forth, we make use of voter surveys from the Making Electoral Democracy Work (MEDW) project, which contains data for the September 2012 Quebec and November 2012 Catalan elections. Consisting of a pre- and postelectoral survey, the sample includes 597 voters in Catalonia and 561 voters in Quebec. We try to capture the three ideological dimensions by creating summative rating scales based on survey questions, a technique that has been shown to be preferable to selfreported questions on ideological positioning (Treier and Hillygus, 2009). While vote intentions are taken from post-electoral data, controls and main predictors come from the pre-electoral wave of the survey.

5 The referendum asked two separate questions: 1) Do you want Catalonia to become a State? 2) If so: Do you want this State to be independent? The Catalan authorities decided to keep the same format for the alternative consultation. The result was that 80.76% of the 2.3 million Catalans who had cast a ballot voted Yes on both questions (Generalitat de Catalunya, 2014).

17

First, we create the centre-periphery scale by combining two questions from the surveys. The first refers to independence and asks how respondents would vote if such a referendum were to be held. The second question is an identity question, often known as the Moreno question, which captures subjective identity. Before creating the scale, we test for both validity and reliability. In both Quebec and Catalonia, principal component analysis leads to the variables being part of a single factor, thus confirming that they relate to a common dimension. In terms of reliability, the Cronbach's a scores are 0.73 for Quebec and 0.76 for Catalonia. For the economy dimension, we use two questions. The first is a measure of redistribution that asks ‘Where would you place yourself on a 0 to 10 scale where 0 means very favourable to redistribution of wealth from the richest to the poor, and 10 means you oppose redistribution’. However, the second question is only available in Quebec. It asks participants if they think postsecondary education tuition should be raised or lowered. In Quebec, the variables load on a single factor and reliability is moderate with a Cronbach's a score of 0.52. Since no similar question was asked in Catalonia, we rely only on the redistribution of wealth question. Finally, the social dimension is captured through two questions related to social tolerance. The first question measures to what degree respondents favour tougher prison sentences over rehabilitation programs. The second question captures whether respondents favour welcoming more or fewer immigrants. Yet again, variables load on one single factor and reliability is moderate, as both Catalonia and Quebec have a Cronbach's a score of 0.54 for this scale.6 Robustness checks were carried out by conducting factor analyses on all items (Tables A1 and A2 in Appendix). Three separate factors are identified, thus providing further justification to the use of separate scales for the ideological dimensions. All dimensions are scaled from 0 to 10, where 0 refers to centralisation, the economic left, or social progressivism, while 10 refers to decentralisation, the economic right, or social conservatism. Finally, in order to capture the appropriate influence of our ideological dimensions, a series of control variables were added to the analyses: age, gender, education, income, interest and mother tongue.7 The first five are traditional control variables commonly used in electoral analyses. Mother tongue is especially important to ascertain ethno-regional linguistic political dynamics. It has previously been shown to influence vote choice in past Catalan elec~ oz Mendoza, 2008). This phenomenon is tions (Miley, 2013; Mun also present in Quebec and has even been demonstrated to be a determinant of voting in the 2012 election (Bilodeau, 2013; Nadeau langer, 2013). and Be

5. Results It was noted in Section 2 that ideology in ethno-regional contexts could not be understood in its entirety without taking the centre-periphery, economic and social axes separately into account.

6 Though Cronbach's a scores between 0.50 and 0.70 can be considered moderately low, they are within acceptable limits reported in political survey research (see, for example, Blais et al., 2002; Escandell and Ceobanu, 2010; Gidengil et al., 2012). 7 Education is an 11-point scale ranging from ‘No schooling’ to ‘Doctorate’, while income is a 12-point scale ranging from ‘Less than $20,000’ to ‘$120,000 or more’. Interest is a summative rating scale created from two 11-points scales asking how interested the respondent was in either the ‘current election’ or ‘in politics generally’. The Cronbach's a score for this scale was 0.86. Finally, the mother tongue variable was coded 1 if the respondent's mother tongue was either Catalan in Catalonia or French in Quebec, and 0 if not.

18

M. Medeiros et al. / Electoral Studies 40 (2015) 14e22

Figs. 1 and 2 illustrate this analytical necessity with regard to vote choice by presenting positional variation of voters across parties. Firstly, three-dimensional visualization highlights how voters are clearly separated along the centre-periphery axis. To be more precise, on a scale from 0 to 10, PLQ voters had a mean of 2.1 and CAQ voters a mean of 3.7, as opposed to means of around 6.5 and 8 for QS and PQ voters. This underscores the determinant nature of this axis for vote choice, as has been established in the literature. Thus, we can say that the three-dimensional visualization has good face validity. In substantive terms, the visualization shows in the Quebec case how PLQ and PQ electors differ on this cleavage. While the CAQ is more dispersed, it is mostly clustered at the mid-point of the axis. Also, some QS supporters do seem to be more centralist, but most of them are closer to the ‘periphery’ end of this dimension. On the Catalan front, Fig. 2 shows that the PSC and the Partit Popular de Catalunya (PPC) are clustered around the lower end of the axis with pro-centre scores of 3.6 and 2 respectively, while CiU and ERC voters are clearly positioned on the ‘periphery’ side with mean scores of 8.8 and 9.2. As for the economic and social axes, the figures demonstrate that they capture different issue dimensions for both regions. Since three-dimensional visualization hinders readability by distorting perspective, comparisons of group means were conducted using ttests. These statistical tests (statistically significant to the p < 0.001 level) show that in Quebec, with the exception of PLQ supporters,

voters had on average a higher score on the social axis than on the economic axis. This suggests that voters tended to be more rightleaning on social issues than on economic ones. This is particularly evident when examining the changes in party clusters for the PQ and the CAQ. PQ voters are more to the left in economic terms, with a mean of 3.6, but spread along the conservative side in social terms with a mean of 5.8. A different pattern is noticeable for the CAQ, whose voters are spread out along the economic axis, with a mean of 5.4, while being much more concentrated on the conservative side of the social axis; being the more socially right-wing party in Quebec with a mean 6.5. In a three-dimensional environment such as Fig. 1, this means that CAQ supporters are mostly positioned at the opposite end of the origin, or the ‘back corner’ of the chart from the given perspective. While the number of QS voter is too small to perceive a clear trend in the figure, its voters were found to be the most leftist on both the economic and social fronts (means of 2.8 and 4.2 respectively). Fig. 2 shows the three-dimensional visualization for Catalonia. CiU voters follow the pattern of CAQ voters in that they are more on the conservative side on the social axis (mean of 7) while being more clustered around the mid-point on the economic axis (mean of 3.83). As for supporters of the PSC, they are grouped on the conservative side of the social dimension with a mean of 5.9, while leaning much more to the left on the economy with an average score of 2.5. Comparisons of group means were also conducted,

Fig. 1. Positioning of Quebec voters by party.

M. Medeiros et al. / Electoral Studies 40 (2015) 14e22

19

Fig. 2. Positioning of Catalan voters by party.

using t-tests, for the Catalan parties. The statistically significant results indicate that the voters for all parties were more socially conservative than they were economically conservative. As these data show, economic and social dimensions are not necessarily congruent and must be taken into account separately. It does not necessarily follow that voters who are more socially conservative are also economically conservative. Collapsing these dimensions into a single lefteright dimension would dilute this nuance of the Quebec and Catalan party systems. With the distinctiveness of each axis established, the next step is to determine if each of the three axes actually has an independent impact on vote choice. Multinomial logistic regressions are used to measure the dimensions’ specific influence on the vote for the four major parties in each region. The results for Quebec, in Table 1, further support our threedimensional model. The multinomial logistic regression presents log-odds with the incumbent PLQ as the reference category. The centre-periphery dimension has a statistically significant influence on vote choice for the CAQ, QS and PQ. The effects suggest that the more sovereignist a voter is, the less likely they are to vote for the PLQ, as opposed to the other three parties. Likewise, more socially or economically leftist voters are more likely to prefer the PQ or QS over the PLQ. The coefficients for the CAQ suggest that the biggest ideological distinction between CAQ and PLQ voters lies in the centre-periphery axis, rather than social or economic lefteright. In terms of the control variables, three had a significant effect. Voters whose mother tongue is French were more likely to vote QS

Table 1 Multinomial logistic regression of vote intentions in Quebec.

Social Economy Centre-periphery Age Male Education Income Interest French Intercept N pseudo R2

PQ/PLQ

QS/PLQ

CAQ/PLQ

0.25* (0.10) 0.35**(0.12) 0.87*** (0.11) 0.02 (0.02) 0.52 (0.49) 0.13 (0.13) 0.10 (0.07) 0.16 (0.10) 1.59 (0.81) 1.50 (1.85) 532 0.35

0.45***(0.13) 0.47***(0.13) 0.64*** (0.12) 0.02 (0.02) 0.4 (0.53) 0.04 (0.14) 0.08 (0.08) 0.01 (0.11) 1.99* (1.01) 0.68 (1.94)

0.08 (0.08) 0.07 (0.08) 0.27** (0.10) 0.01 (0.01) 0.42 (0.33) 0.22* (0.10) 0.13*(0.05) 0.08 (0.07) 1.39** (0.44) 0.92 (1.19)

Numbers reported are log-odds. Standard errors in parentheses. *p < 0.05, **p < 0.01, *** p < 0.001.

or CAQ than PLQ. This is expected, as the PLQ is usually the most popular party among Anglophone and Allophone voters in Quebec. Wealthier voters are also more likely to vote CAQ over PLQ, which is echoed through the CAQ's rightist economic policies. Finally, less educated voters tend to vote CAQ rather than PLQ. The results for Catalonia, in Table 2, also support our choice of analytical framework. The regression model uses the incumbent CiU as the reference category. The social dimension had a significant effect for both the PSC and ERC. This means that more socially

20

M. Medeiros et al. / Electoral Studies 40 (2015) 14e22

Table 2 Multinomial logistic regression of vote intentions in Catalonia.

Social Economy Centre-periphery Age Male Education Income Interest Catalan Intercept N pseudo R2

PSC/CiU

PP/CiU

ERC/CiU

0.49** (0.17) 0.25* (0.10) 0.77*** (0.12) 0.01 (0.02) 0.41 (0.53) 0.18 (0.23) 0.21 (0.13) 0.05 (0.14) 1.53** (0.51) 0.49** (0.17) 337 0.50

0.11 (0.19) 0.03 (0.13) 1.33*** (0.27) 0.01 (0.02) 0.10 (0.69) 0.07 (0.29) 0.17 (0.14) 0.01 (0.16) 0.68 (0.77) 0.11 (0.19)

0.25***(0.07) 0.15* (0.06) 0.24 (0.13) 0.02 (0.01) 0.10 (0.27) 0.09 (0.13) 0.20* (0.08) 0.02 (0.06) 0.73* (0.31) 0.25*** (0.07)

Numbers reported are log-odds. Standard errors in parentheses. *p < 0.05, **p < 0.01, *** p < 0.001.

on the left a voter is, the more likely they are to vote PSC or ERC over CiU. The same can be said of the economic dimension, as economically left voters tend to choose these same parties over CiU. However, the centre-periphery axis shows statistically significant differences for all parties when compared to CiU, except for ERC. This leads us to believe that being more pro-centre increases the chances of voting PSC and PPC over CiU, whereas the difference between CiU and ERC on that same front is too limited to have an effect on voters. As for the control variables, two have an effect on vote choice. Income seems to positively influence the CiU vote, as wealthier voters are less likely to vote PSC or PPC than CiU. Finally, having Catalan as a mother tongue is also shown to have a positive impact on vote choice, as these voters are less likely to vote for the PSC and ERC than CiU. Our results clearly show that the three dimensions were distinct determinants of vote choice for the 2012 sub-national elections in Catalonia and Quebec. These findings support our assertion that the centre-periphery, economic and social dimensions influence in an independent manner vote choice in ethno-regional contexts. Thus, the results point to the importance of separately examining these three dimensions in order to properly understand vote choice in ethno-regionalist contexts. 6. Conclusion The 2012 Catalan Parliament election campaign was arguably highlighted by the debate around centre-periphery issues. However, this factor did not eliminate discussions around economic and social policies. This contrasts with the 2012 Quebec election campaign, where sovereignty did not play as large a role as it had in past elections and took more of a back seat to economic and social matters. Yet, centre-periphery issues were nonetheless present. The importance of centre-periphery, social and economic issues lead us to believe that the standard two-dimensional view used in regions with a strong centre-periphery cleavage (i.e. that of centre-periphery versus lefteright) is insufficient. While previous studies of electoral behaviour in these contexts have included the centre-periphery axis, we propose that including this factor should not come at the expense of considering the importance of the ‘traditional’ economic and social dimensions in structuring political outcomes. We have instead proposed a three-dimensional analytical framework for understanding election dynamics in these regions, with Catalonia and Quebec as test cases. As our results show, exchanging a general lefteright axis for two separate economic and social axes allows for slightly more nuanced views of political competition in Catalonia and Quebec. Voters can have economically left and socially right opinions and vice versa. These differences would perhaps be glossed over using a unidimensional lefteright axis.

Though it is evident from the political events that unfolded prior to the 2012 Catalan election why the centre-periphery dimension was a determining factor in vote choice, the Quebec case needs a bit of an explanation. Seeing that early on in the campaign, the PQ had ruled out any chance of holding a referendum on secession if they were to form government, it would be fair to presume that this langer dimension would become less salient. As Nadeau and Be (2013) bring forth, the centre-periphery dimension was indeed less determinant in the 2012 election than in previous contests, but it nevertheless remained a decisive factor in vote choice. We believe that the perpetual influence of the centre-periphery dimension on Quebec politics is due to the fact that the Quebec party system is structured along this cleavage. Thus, the PQ could not escape questions on how it planned to further independence and its nonindependentist opponents refused to let voters forget its deepseeded desire to hold a referendum on secession. The centreperiphery issue hence remained salient and determinant. In Catalonia, this political dimension has long been a constant of regional politics before the events leading to the surge in support for independence. This is perhaps due to the parties being also partly divided along the centre-periphery axis. As a result, we believe that when the centre-periphery axis structures the party system, it will be a salient and determining factor in electoral politics. Catalonia and Quebec are obviously not the only two polities to have their party systems divided along the centre-periphery dimension. Scotland and Flanders are, for example, recurrent cases in the literature on ethno-regional politics. Electoral surveys administered in these two regions often explicitly explore the influence of the centre-periphery axis on vote choice. Yet, electoral research overlooks this dimension in other regions. Bavaria also has a party system that is partly divided along the centre-periphery axis (Hepburn, 2010), but, to the best of our knowledge, electoral surveys in Bavaria have not taken into consideration the centreperiphery axis, thereby leaving a void in the understanding of electoral dynamics. The centre-periphery dimension can also be present in political contexts that are not structured along this axis. Australia and the United States have been shown to have regions in which the centre-periphery dimension plays a more prominent role, even though political parties, even at state-level, are not divided along this axis (Brown and Deem, 2014; Kincaid and Cole, 2011). Seeing that the centre-periphery divide is not intimately tied to the partisan landscape, we presume that the centre-periphery axis plays more of a contextual influence on vote choice in such cases. However, to our knowledge, there is a lack of electoral research into the influence of the centre-periphery dimension on electoral dynamics in American and Australian states where this axis could be salient. Our findings point to the greater information that results from exploring the social and economic axes separately while at the same time accounting for the centre-periphery axis. Therefore, this paper contributes to a more refined picture of electoral dynamics specifically in Catalonia and Quebec, while aiming to speak to ethno-regional contexts in general. Though more research is necessary to uncover what can be gained through viewing centre-periphery, economic, and social issues as independent and individual elements of political behaviour in distinctive settings, our results justify applying the threedimensional analytical framework in other elections in ethnoregional contexts. Acknowledgements  Blais and Damien Bol for sharing the We are grateful to Andre Making Electoral Democracy Work (MEDW) data with us and for

M. Medeiros et al. / Electoral Studies 40 (2015) 14e22

inviting us to present an earlier version of this article at a panel they organised at the 2014 European Consortium for Political Research General Conference in Glasgow. We would also like to thank Pedro Riera for his insightful comments and this journal's anonymous referees for their help and suggestions. Appendix These tables present the resulting factor analyses used to determine the dimensions. We only use loadings of 0.45 (20% variance overlap between variable and factor), as shown in the shaded areas of Tables A1 and A2 (Tabachnick and Fidell, 2007: 668).

21

Table A4 Multinomial logistic regression of vote intentions in Catalonia using factors

Social factor Economy factor Centre-periphery f. Age Male Education Income Interest Catalan Intercept N pseudo R2

PSC/CiU

PPC/CiU

ERC/CiU

0.89* (0.44) 0.85** (0.30) 3.47*** (0.59) 0.01 (0.02) 1.06 (0.56) 0.26 (0.25) 0.37** (0.13) 0.01 (0.16) 1.01 (0.62) 2.13 (1.76) 368 0.47

0.51 (0.49) 0.24 (0.39) 5.46***(0.88) 0.01 (0.03) 1.02 (0.78) 0.13 (0.32) 0.28* (0.14) 0.01 (0.17) 0.09 (0.93) 1.75 (2.05)

0.58*** (0.16) 0.49** (0.17) 0.96* (0.46) 0.02 (0.01) 0.05 (0.28) 0.04 (0.13) 0.17 (0.09) 0.02 (0.07) 0.75* (0.34) 0.80 (0.82)

Numbers reported are log-odds. Standard errors in parentheses. *p < 0.05, **p < 0.01, p < 0.001.

***

Table A1 Factor analysis of dimensions in Quebec using principal-component analysis N ¼ 814. Component

Factor 1

Economy e redistribution Economy e tuition Social e sentences Social e immigration Centre-periphery e referendum Centre-periphery Moreno Eigenvalues Proportion

Factor 2

0.4407

Factor 3

Uniqueness

0.9206 0.6451

0.1481 0.3495 0.3167 0.2700 0.1909 0.1782

0.7972 0.8498 0.8797 0.9049 1.84 0.31

1.40 0.23

1.30 0.21

Note: Loadings smaller than 0.20 were omitted.

Table A2 Factor analysis of dimensions in Catalonia using principal-component analysis N ¼ 797. Component

Factor 1

Economy e redistribution Social e sentences Social e immigration Centre-periphery e referendum Centre-periphery Moreno Eigenvalues Proportion

Factor 2

Factor 3

Uniqueness

0.9936

0.0109 0.3357 0.2845 0.1411 0.1408

0.7888 0.8435 0.9261 0.9207 1.74 0.35

1.34 0.27

1.01 0.20

Note: Loadings smaller than 0.20 were omitted.

As a robustness check, we present the regression models using the predicted factors rather than the summative-rating scales. Inferences are essentially the same and pseudo-R2s are slightly higher. Small differences could be due to the number of survey participants that varies between the different ways of creating the scales.

Table A3 Multinomial logistic regression of vote intentions in Quebec using factors.

Social factor Economy factor Centre-periphery f. Age Male Education Income Interest French Intercept N pseudo R2

PQ/PLQ

QS/PLQ

CAQ/PLQ

0.82** (0.25) 1.03*** (0.29) 3.74*** (0.40) 0.03 (0.02) 0.01 (0.53) 0.06 (0.14) 0.10 (0.07) 0.27 (0.14) 0.75 (0.77) 0.43 (2.07) 472 0.37

1.17*** (0.32) 1.45***(0.37) 2.99***(0.48) 0.01 (0.02) 0.5 (0.54) 0.08 (0.15) 0.10 (0.09) 0.11 (0.13) 1.60 (1.09) 2.43 (2.31)

0.23 (0.20) 0.17 (0.18) 1.03**(0.34) 0.01 (0.01) 0.54 (0.34) 0.23* (0.11) 0.13* (0.05) 0.01 (0.08) 1.04* (0.44) 0.03 (1.11)

Numbers reported are log-odds. Standard errors in parentheses. *p < 0.05, **p < 0.01, *** p < 0.001.

References Abramowitz, A.I., Saunders, K.L., 2006. Exploring the bases of partisanship in the American electorate: social identity vs. Ideol. Polit. Res. Q. 59 (2), 175e187. Albright, J.J., 2010. The multidimensional nature of party competition. Party Polit. 16 (6), 699e719. Barnea, M.F., Schwartz, S.H., 1998. Values and voting. Polit. Psychol. 19 (1), 17e40.  Ge langer, E., lineau, F., 2013. Une e lection extraordinaire. In: Bastien, F., Be  Ge langer, E., lineau, F. (Eds.), Les Que  be cois aux urnes: les partis, Bastien, F., Be dias et les citoyens en campagne. Les Presses de l'Universite  de Montre al, les me Montreal. land, D., Lecours, A., 2005. The politics of territorial solidarity nationalism and Be social policy reform in Canada, the United Kingdom, and Belgium. Comp. Polit. Stud. 38 (6), 676e703. land, D., Lecours, A., 2014. Accommodation and the politics of fiscal equalization Be in multinational states: the case of Canada. Nations Natl. 20 (2), 337e354. ral : le vote des minorite s visibles a  l'e lection Bilodeau, A., 2013. L'avantage libe  Ge  be coise de 2012. In: Bastien, F., Be langer, E., lineau, F. (Eds.), Les que  be cois aux urnes: les partis, les me dias et les citoyens en campagne. Les Que  de Montre al, Montreal. presses de l'Universite Blais, A., Nadeau, R., 1992. To be or not to be Sovereignist: Quebeckers' Perennial Dilemma. Can. Public Policy 18 (1), 89e103. Blais, A., Gidengil, E., Nadeau, R., Nevitte, N., 2002. Anatomy of a Liberal Victory: Making Sense of the Vote in the 2000 Canadian Election. Broadview Press, Peterborough. Brown, A.J., Deem, J., 2014. A Tale of Two Regionalisms: Improving the Measurement of Regionalism in Australia and beyond. Regional Studies (Available on CJO: http://www.tandfonline.com/doi/pdf/10.1080/00343404.2014.980800. Budge, I., 1994. A new spatial theory of party competition. Br. J. Polit. Sci. 24 (4), 443e467. Campbell, A., Converse, P., Miller, W., Stokes, D., 1960. The American Voter. Wiley, New York. Castles, F., Mair, P., 1984. LefteRight political scales: some ‘Expert’ judgments. Eur. J. Polit. Res. 12 (1), 73e88. Dalton, R., 2008. Citizen Politics: Public Opinion and Political Parties in Advanced Industrial Democracies, Fifth ed. CQ Press, Washington, DC.  ne rations, Souverainete , Parti Que  be cois. Retrieved 8 August, Durand, C., 2014. Ge 2014, from. http://ahlessondages.blogspot.ca/2014/06/generationssouverainete-parti-quebecois.html. Escandell, X., Ceobanu, A.M., 2010. Nationalisms and anti-immigrant sentiment in Spain. South Eur. Soc. Polit. 15 (2), 157e179. Falkenhagen, F., 2013. The CSU as an ethno-regional party. Ger. Polit. 22 (4), 396e420. s participatiu. Generalitat de Generalitat, de Catalunya, 2014. Resultats del proce Catalunya, Barcelona. Gidengil, E., Nevitte, N., Blais, A., Everitt, J., Fournier, P., 2012. Dominance and Decline: Making Sense of Recent Canadian Elections. University of Toronto Press, Toronto. Guibernau, M., 2013. Secessionism in Catalonia: after democracy. Ethnopolitics 12 (4), 368e393. Guibernau, M., 2014. Prospects for an independent Catalonia. Int. J. Polit. Cult. Soc. 27 (1), 5e23. Hechter, M., Levi, M., 1979. The comparative analysis of ethnoregional movements. Ethn. Racial Stud. 2 (3), 260e274. Hepburn, E., 2010. Small worlds in Canada and Europe: a comparison of regional party systems in Quebec, Bavaria and Scotland. Regional Fed. Stud. 20 (4e5), 527e544. Hinich, M.J., Munger, M.C., 1994. Ideology and the Theory of Political Choice. University of Michigan Press, Ann Arbor. Inglehart, R., 1990. Culture Shift in Advanced Industrial Society. Princeton University Press, Princeton. Jackson, R.J., 2013. Global Politics in the 21st Century. Cambridge University Press, Cambridge.

22

M. Medeiros et al. / Electoral Studies 40 (2015) 14e22

Jost, J., 2006. The end of the end of ideology. Am. Psychol. 61 (7), 651e670. Keating, M., 2001. Nations against the State : the New Politics of Nationalism in Quebec, Catalonia, and Scotland. Palgrave Macmillan, Basingstroke. Kincaid, J., Cole, R.L., 2011. Citizen attitudes toward issues of Federalism in Canada, Mexico, and the United States. Publius: J. Fed. 41 (1), 53e75. ~ eira, R., 2011. ‘Less at stake’or a different game? regional elections in catalonia Lin and Scotland. Regional Fed. Stud. 21 (3), 283e303. Lipset, S.M., Rokkan, S., 1967. Cleavage structures, party systems, and voter alignments. In: Lipset, S.M., Rokkan, S. (Eds.), Party Systems and Voter Alignments: Cross-national Perspectives. The Free Press, New York. Lublin, D., 2014. Minority Rules: Electoral Systems, Decentralization, and Ethnoregional Party Success. Oxford University Press, Oxford. : re ^ves, luttes et de route du Canada Martel, M., 1997. Le deuil d'un pays imagine bec et la francophonie canadienne, français: Les rapports entre le Que 1867e1975. University of Ottawa Press, Ottawa. Martí, D., 2013. The 2012 catalan election: the first step towards Independence? Regional Fed. Stud. 23 (4), 507e516. Martínez-Herrera, E., 2002. From Nation-Building to building identification with political communities: consequences of political decentralisation in Spain, the Basque country, Catalonia and Galicia, 1978e2001. Eur. J. Polit. Res. 41 (4), 421e453. Martínez-Herrera, E., Miley, T.J., 2010. The constitution and the politics of national identity in Spain. Nations Natl. 16 (1), 6e30. McRoberts, K., 1997. Misconceiving Canada: the Struggle for National Unity. Oxford University Press, Oxford. €l, A., 2014. The forgotten side of Partisanship negative party Medeiros, M., Noe identification in four Anglo-American democracies. Comp. Polit. Stud. 47 (7), 1022e1046. Mendelsohn, M., 2002. Measuring national identity and patterns of attachment: quebec and nationalist mobilization. Natl. Ethn. Polit. 8 (3), 72e94. Mendelsohn, M., Parkin, A., Pinard, M., 2007. A new chapter on the same old Story? public opinion in quebec from 1996e2003. In: Murphy, M. (Ed.), Canada: the State of the Federation 2005. Quebec and Canada in the New Century; New Dynamics, New Opportunities. McGill-Queen's University Press, Montreal, pp. 25e52. Miley, T.J., 2013. Blocked articulation and nationalist hegemony in catalonia. Regional Fed. Stud. 23 (1), 7e26. Miller, H., Miller, K., 1996. Language policy and identity: the case of catalonia. Int. Stud. Sociol. Educ. 6 (1), 113e128. ~ a: lealtad y transferencia de votos Montero, J.R., Font, J., 1991. El voto dual en Catalun micas. Rev. Estud. Políticos 73, 7e34. en las elecciones autono

Müller-Rommel, F., 2003. Ethnoregionalist parties in Western europe. In: De Winter, L., Tursan, H. (Eds.), Regionalist Parties in Western Europe. Routledge, London. ~ oz, J., Tormos, R., 2015. Economic expectations and support for secession in Mun catalonia: between causality and rationalization. Eur. Polit. Sci. Rev. 7 (2), 315e341. ~ oz Mendoza, J., 2008. From National Catholicism to Democratic Patriotism? an Mun Empirical Analysis of Contemporany Spanish National Identity. Department of Social and Political Sciences. Universitat Pompeu Fabra, Barcelona.  2013. Un mode langer, E., le ge ne ral d’explication du vote des Nadeau, R., Be  Ge be cois. In: Bastien, F., Be langer, E., lineau, F. (Eds.), Les Que  be cois aux Que dias et les citoyens en campagne. Les presses de urnes: Les partis, les me  de Montre al, Montreal. l'Universite s et appui a  la souverainete  Nadeau, R., Fleury, C.J., 1995. Gains linguistiques anticipe bec. Can. J. Polit. Sci. 28 (1), 35e50. du Que Nie, N., Verba, S., Petrocik, J., 1979. The Changing American Voter. Harvard University Press Cambridge. s, F., Lago, I., Mun ~ oz, J., 2006. The 2005 autonomous election in Galicia: the Pallare end of conservative hegemony 1. Regional Fed. Stud. 16 (4), 481e494. Petithomme, M., Garcia, A.F., 2013. Catalonian nationalism in Spain's time of crisis: from asymmetrical Federalism to Independence? Fed. Idea(March) 1e18. Rico, G., 2012. The 2010 regional election in catalonia: a multilevel account in an age of economic crisis. South Eur. Soc. Polit. 17 (2), 217e238. ~ eira, R., 2014. Bringing secessionism into the mainstream: the 2012 Rico, G., Lin regional election in catalonia. South Eur. Soc. Polit. 19 (2), 257e280. Rusk, J., 1987. Issues and voting. In: Long, S. (Ed.), Research in Micropolitics. JAI Press, Greenwich Conn. Russell, P.H., 2004. Constitutional Odyssey: Can Canadians Become a Sovereign People? University of Toronto Press, Toronto. Sniderman, P., Brody, R., Tetlock, P., 1993. Reasoning and Choice: Explorations in Political Psychology. Cambridge University Press, Cambridge. Stokes, D., 1963. Spatial models of party competition. Am. Polit. Sci. Rev. 57 (2), 368e377. Tabachnick, B.G., Fidell, L.S., 2007. Using Multivariate Statistics, fifth ed. Pearson Education, Boston. Treier, S., Hillygus, D.S., 2009. The nature of political ideology in the contemporary electorate. Public Opin. Q. 73 (4), 679e703. Wheatley, J., Carman, C., Mendez, F., Mitchell, J., 2014. The dimensionality of the scottish political space: results from an Experiment on the 2011 Holyrood elections. Party Polit. 20 (6), 864e878. Woolard, K.A., 1989. Double Talk: Bilingualism and the Politics of Ethnicity in Catalonia. Stanford University Press, Stanford.