1.3.1 Group Identities and Cleavage Formation

The importance of social identities is implicitly acknowledged in classical cleavage accounts (Reference Lipset, Rokkan, Lipset and RokkanLipset and Rokkan 1967; Reference Bartolini and MairBartolini and Mair 1990; Reference WeakliemWeakliem 1993; Reference Knutsen, Scarbrough, Jan and ScarbroughKnutsen and Scarbrough 1995; Reference BartoliniBartolini 2005b). Yet those who insist that cleavages continue to matter have not devoted much attention to answering the question of what constitutes the “glue” linking social groups and political parties. While we have learned a lot about the socio-structural groups underlying the universalism–particularism divide, as well as on the discourse of the political actors mobilizing it,Footnote ⁴ the link between structure and consciousness is far from evident. This is of course an idea as old as the social sciences themselves: a “class in itself” is not yet a “class for itself’ (cf. Reference MarxMarx 1937 [1852], 192), and “categories of analysis” (e.g., based on socio-structural conditions) are potentially far from being “categories of practice” (through which people experience group belonging) (Reference BourdieuBourdieu 1985).

There are also specific reasons for focusing on collective identities when it comes to the universalism–particularism cleavage. We argue that a focus on group identities can help us make sense of some particularities in the emergence of this cleavage. In the absence of a clear-cut link between political discourse and the markers of socio-structural position we use to describe these groups, the link between the two still represents something of a black box. This is particularly true for the (counterintuitive) working-class realignment in favor of the Far Right (see, for example, Reference RydgrenRydgren 2007, the contributions in Reference RydgrenRydgren 2013, and Reference Evans and TilleyEvans and Tilley 2011), as well as with respect to recent work that relates subjective structural position, such as status anxiety (e.g., Reference CramerCramer 2016; Reference GestGest 2016; Reference HochschildHochschild 2016; Reference Noam and HallGidron and Hall 2017; Reference FitzgeraldFitzgerald 2018; Reference BoletBolet 2020), relative economic deprivation (e.g., Reference Rooduijn and BurgoonRooduijn and Burgoon 2018; Reference KurerKurer 2020; Reference Kurer and Van StaalduinenKurer and Staalduinen 2022; Reference ZollingerBreyer, Palmtag, and Zollinger 2023), or the perception of economic and social opportunities (Reference ZollingerHäusermann, Kurer, and Zollinger 2023) to Far Right support. Why exactly do such feelings and perceptions of grievance and vulnerability translate into support for the Far Right, rather than for other parties that cater to economic vulnerability more directly? To understand why culturally connoted appeals resonate with economically defined groups, the next section draws centrally on psychological and sociological approaches that highlight the importance of positive group identifications for individuals (Reference Bornschier, Häusermann, Zollinger and ColomboBornschier et al. 2021; Reference ZollingerZollinger 2022). This accounts for the propensity of the “losers” of economic and social change to seek identification based on categories that correspond only loosely to their objective social position.

The construction of a positive self-image is more self-evident for the relative “winners” of the social changes of the past decades. Indeed, in the initial mobilization of the New Social Movements of the 1970s and 1980s, personal and group identity in the quest for the recognition of difference in terms of gender, sexual orientation, as well as the free choice of lifestyles were closely linked. In a process corresponding to what Reference Snow, McAdam, Stryker, Owens and WhiteSnow and McAdam (2000) have called the “general diffusion” of movement identities, solidarity with the drivers of protest then expanded within broader universalistically minded sectors of society. As movement activists flocked into the emerging Green parties, bottom-up and top-down processes of mobilization and identity construction were intimately related.Footnote ⁵ But explaining the inclination of parts of the middle class to vote for the New Left is by no means trivial either. While the literature has identified education and work logic as determinants of universalism (Reference KriesiKriesi 1998; Reference OeschOesch 2006a; Reference Kitschelt and RehmKitschelt and Rehm 2014), an ethnographic approach reveals how these groups’ economic preferences are embedded in broader, culturally defined worldviews of deservingness and fairness, as well (Reference WestheuserWestheuser 2021; Reference Damhuis and WestheuserDamhuis and Westheuser 2023).

In a nutshell, then, the key question is how the winners and losers of the social transformations of the past four to five decades see and describe themselves. We believe such a perspective can go a long way in explaining why certain party appeals resonate with specific social groups, while others fail to do so.

1.3.2 The Role of Agency

Creating and reproducing these nonevident links between social groups and parties obviously assigns a nontrivial role to political agency.Footnote ⁶ We believe that the level of social identities – the intermediate level in Reference Bartolini and MairBartolini and Mair’s (1990) much-noted threefold conception of cleavages – is a good place to study agency.Footnote ⁷ It is here that the self-definitions of social groups intersect with the appeals by political parties to give meaning to grievances. In grasping this link, we can draw on the literature on social movements that highlights how collective action frames point to injustices and combine them with a definition of the group or social category in question (Reference Gamson, Aldon and MuellerGamson 1992; Reference Klandermans and JudithKlandermans 2001).Footnote ⁸ Reference Klandermans and JudithKlandermans (2001) theorizes two processes that translate the “raw material” of a cleavage into collective action. On the one hand, meaning is constructed bottom-up at the interface between networks of personal interaction and media-based public discourse. On the other hand, these interpretations are reinforced during campaigns, where social actors undertake deliberate attempts to persuade voters and where they stake out who the group’s antagonists are. The latter process is crucial because the social movement literature as well as the more classical sociological literature both highlight the group-binding effects of conflict (Reference CoserCoser 1956; Reference StrykerStryker 1980; Reference MarksMarks 1989; Reference Gamson, Aldon and MuellerGamson 1992). Combined, these two processes result in what Reference Snow, Burke Rochford, Worden and BenfordSnow et al. (1986) refer to as “frame alignment,” meaning in our case that individuals’ and parties’ interpretations of grievances come to overlap. Incorporating the idea that party appeals resonate with the way groups would describe themselves also helps us understand how Far Right parties succeed in mobilizing diverse structural groups (as emphasized in the recent literature on different logics of Far Right voting, for example, Reference DamhuisDamhuis 2020; Reference Harteveld, Van Der Brug, De Lange and Van Der MeerHarteveld et al. 2022).

Both the bottom-up processes in which group identities are constructed, as well as the role of political agency in reinforcing and nourishing these identities, lead us to expect fundamental similarities between countries:

1. (a) Similarities in Terms of the Raw Materials for a New Cleavage. We start here from the insight that each individual holds multiple identities with the potential of being politically relevant. Building on Reference Stryker, Stryker, Owens and WhiteStryker (2000), as well as the classical literature on cross-cutting cleavages (e.g., Reference LijphartLijphart 1979; Reference Rokkan, Flora, Kuhnle and UrwinRokkan 1999), the relative salience of these identities should determine which of them will shape political alignments. Identity salience increases, according to Reference StrykerStryker (1980), as individuals interact with members of the same group. Because our personal interactions are patterned by social structural position – chiefly in terms of class, education, and urban-rural residence – identity salience is not entirely voluntaristic. Instead, it is biased toward those identities that are most strongly reinforced at the workplace and in everyday life. The resulting expectation is that the grievances resulting from the transition to a knowledge economy will lead to similar identity potentials across the set of advanced democracies that we study.

2. (b) Convergence of Mobilization Frames. Framing constitutes a creative, collective effort at meaning construction. It “draw[s] on the cultural stock of images of what is an injustice, of what is a violation of what ought to be” (Reference Zald, McAdam, McCarthy and ZaldZald 1996: 266). At the anti-universalistic pole of the cleavage, the Far Right has converged on a particularistic frame that emphasizes the preservation of traditional (national) communities and status hierarchies (e.g., Reference AntonioAntonio 2000; Reference MinkenbergMinkenberg 2000; Reference BetzBetz 2004; Reference BornschierBornschier 2010; Reference Elgenius and RydgrenElgenius and Rydgren 2019). One of the core elements of the Far Right’s ideology is indeed its nostalgic component, as several scholars have highlighted (Reference Betz and JohnsonBetz and Johnson 2004; Reference DuyvendakDuyvendak 2011; Reference Elgenius and RydgrenElgenius and Rydgren 2019, Reference Elgenius and Rydgren2022). This discourse can be expected to resonate strongly with social groups that feel deprived relative to a supposedly better past (e.g., Reference Elchardus and SpruytElchardus and Spruyt 2012; Reference Burgoon, van Noort, Rooduijn and UnderhillBurgoon et al. 2019; Reference Engler and WeisstannerEngler and Weisstanner 2021). Combined with the large literature that has pointed to a fundamental similarity in the competitive spaces in West European party systems (e.g., Reference KitscheltKitschelt 1994; Reference Marks, Hooghe, Nelson and EdwardsMarks et al. 2006; Reference Kriesi, Grande and LachatKriesi et al. 2008; Reference BornschierBornschier 2010; Reference Kriesi, Grande and DolezalKriesi et al. 2012; Reference Hutter and KriesiHutter and Kriesi 2019) this again leads us to expect a fundamental similarity in terms of the group identities underlying the universalism–particularism cleavage.

Figure 1 summarizes the discussion so far (also foreshadowing the focus of subsequent sections of this Element). It illustrates how we can think of cleavages becoming consolidated. As party narratives give meaning to existing group boundaries, these politicized identities inform voters’ preferences within specific ideological schemas (also see Reference HuddyHuddy 2001; Reference StubagerStubager 2009). Once formed, they may also shape party politics in lasting ways as mobilization markets are “narrowed” (c.f. Reference MairMair 1997; Reference Rokkan, Flora, Kuhnle and UrwinRokkan 1999) by salient us-versus-them distinctions into which voters are socialized. Put differently, existing interpretations of what conflict is about limit the receptiveness of voters to new political appeals. Hence, even in light of ongoing socio-structural change and party entrepreneurship, collective identity antagonisms generate a certain inertia to party system change, confining the effects of idiosyncratic shocks or individual election campaigns.

Figure 1

Processes of cleavage formation: social closure and political mobilization

An important question concerns the proper level of abstraction to study group identities. Rather than studying broad political identities that are defined by party ideology (e.g., Reference Sartori and StammerSartori 1968; Reference Bartolini and MairBartolini and Mair 1990; Reference Knutsen, Scarbrough, Jan and ScarbroughKnutsen and Scarbrough 1995; Reference MairMair 1997), we suggest focusing on an intermediate level of specificity.Footnote ⁹ These identities – for example, feeling close to “cosmopolitans,” or to “rural people” – are more specific than the broad political allegiances that cleavage theorists have traditionally focused on. At the same time, the identity categories we develop are sufficiently abstract to study the degree to which group identities antagonize New Left and Far Right voters. We elaborate on the choice of these groups in Section 2.

1.3.3 The Transformation of Party Systems and Ideological Party Blocks

Looking at the process of realignment through the lens of ideological party blocks is central to our approach. If transformations in the structure of society create social groups with a common identity, then these potentials inevitably play out quite differently depending on the specific configuration of the party system in different countries. Combined with our focus on group identities, studying party competition in terms of ideological blocks implies a higher level of abstraction that allows us to discern similar patterns underlying varying political entrepreneurship. It allows us to reconcile the structuralist core of our argument with the ever more apparent leeway that actors enjoy in tapping into and giving coherence to raw and often rather diffuse political potentials. Our contention, then, is that politicians and “issue entrepreneurs” for the most part move within the ideological space without fundamentally altering its basic contours or the group divisions underlying competitive dimensions.

This perspective is very much in line with classical cleavage theory (e.g., Reference Rokkan, Flora, Kuhnle and UrwinRokkan 1999). It is also consistent with individual-level accounts developed by cleavage theorists, who distinguish between volatility comprised of voters crossing cleavage lines from within-block movements (Reference Bartolini and MairBartolini and Mair 1990). While the former indicates electoral change signaling either dealignment or realignment, the latter accounts for struggles within ideological blocks over strategies to achieve goals, the rivalries between leaders, and the disappointments that partisans may experience with respect to their government’s policy record. In this way, cleavage theory has always assigned room for voters to make choices based on nuances in issue positioning and emphasis, the profiles and the charisma of specific candidates, or even economic voting.

We revive the analytical perspective of cleavage formation because it allows us to incorporate two aspects of political agency that are central for our purposes. For one thing, we seek to distinguish (a) country variation that reflects fundamental differences in the timing and the strength of the structural manifestation of the universalism–particularism cleavage from (b) more superficial and situational differences caused simply by the fact that the agents of mobilization differ by country. The prime example of the latter is that the mobilization of the universalistic pole of the new cleavage was spearheaded by Green parties in some countries and their established Social Democratic or Socialist counterparts in others (Reference KitscheltKitschelt 1994; Reference Häusermann and KitscheltHäusermann and Kitschelt 2024). The impact of these differences in the division of labor on the political left on competitive policy spaces has been relatively minor compared to that variation triggered by competition between mainstream Right and Far Right parties over the particularistic potential (Reference Kriesi, Grande and LachatKriesi et al. 2008, Reference Kriesi, Grande and Dolezal2012; Reference BornschierBornschier 2010; Reference Bremer, Schulte-Cloos, Hutter and KriesiBremer and Schulte-Cloos 2019; Reference Jasmine, van Ditmars, Hutter and KriesiLorenzini and van Ditmars 2019). This is also mirrored in the far more extensive literature on strategic competition between the established Right and its Far Right challengers (the literature is too large to quote in full, but see, e.g., Reference IgnaziIgnazi 1992; Reference MeguidMeguid 2008; Reference BornschierBornschier 2012; Reference Abou-Chadi and KrauseAbou-Chadi and Krause 2018; Reference van Spanje and de Graafvan Spanje and de Graaf 2018; Reference Spoon and KlüverSpoon and Klüver 2020; Reference Bale and KaltwasserBale and Rovira Kaltwasser 2021; Reference De JongeDe Jonge 2021).

The higher level of abstraction in studying party competition at the level of ideological blocks also helps us to reconcile the two contrasting developments that have spurred diametrically opposing interpretations of contemporary party competition: First, the rise of instability and fragmentation due to an increasing role of agency, and second, the idea of a realignment suggesting that politics remains anchored in fundamental cleavages. While voters remain committed to broader political block ideologies, their choice set – composed of parties that differ with respect to specific issues or issue emphasis, but that align in their basic cleavage positions (Reference Steenbergen, Hangartner and De VriesSteenbergen, Hangartner, and De Vries 2015; Reference Oskarson, Oscarsson and BoijeOskarson, Oscarsson, and Boije 2016) – expands. There is also evidence that affective polarization transcends the partisan level and divides ideological blocks, rather than just parties (Reference BantelBantel 2023).

Theoretically, based on the discussion in Section 1.1, we would expect the mobilization of the New Left and the Far Right to sequentially introduce a division within the Left and the Right. Since the 1980s, Social Democratic parties have competed with the Greens (and sometimes other challengers) over the support of voters holding universalistic group identities. On the political right, as first manifested in France in the early 1980s, established conservative parties increasingly faced competition from the Far Right in rallying voters that more strongly endorse particularistic group identifications. In the meantime, as the last bastions of resistance against the Far Right challenge are falling, we see this competition throughout Western Europe. Theoretically, we would thus expect the existence of up to four blocks along the universalism–particularism cleavage: The traditional Left and the New Left on the one side of the political spectrum, and the traditional Right and the Far Right on the other.

2.1.1 The Relevance of Cleavage Formation in the Knowledge Economy

The political landscapes of Western Europe are widely considered to have been historically shaped by a small number of key conflicts, triggered by the “critical junctures” of the national and industrial revolutions. These changed the fabric of society and became durably articulated by parties and related organizations, such as unions, churches, or associated social clubs. Building on Reference Lipset, Rokkan, Lipset and RokkanLipset and Rokkan’s (1967) seminal work, traditional cleavage theory expanded on the role that collective identities played in this translation of structural disruption into political competition: A sense of shared identity is central to collective action, and it mediates the nonobvious step from objective group belonging to political mobilization. Strong existing identities, cemented by embeddedness in social organizations, also constrain new forms of political mobilization. They partly stabilize existing conflicts even when structural conditions change (e.g., where Catholic workers were historically unavailable for mobilization as members of the working class) (Reference Rokkan, Flora, Kuhnle and UrwinRokkan 1999; Reference BartoliniBartolini 2000; Reference BornschierBornschier 2010). For these reasons, an identity component or “normative element” became an established part of the most influential definition of a “cleavage” as comprised of structural, political, as well as identity divides.

Over the past decades, researchers have studied how the “old” cleavage structures have profoundly loosened, focusing on both structural and political change, but largely neglecting the question of how voters perceive their place in this transformed landscape. We place identities center stage in the study of contemporary politics. Taking traditional cleavage theory as a starting point, we conceptualize collective identities as located at an intermediary level between structure and politics (Reference Bartolini and MairBartolini and Mair 1990; Reference BartoliniBartolini 2005b; Reference BornschierBornschier 2010; Reference Bornschier, Häusermann, Zollinger and ColomboBornschier et al. 2021; Reference ZollingerZollinger 2024). Cleavages become consolidated at the identity level as conflict potentials arising from social structure become activated by political actors.

Much existing work points to new potentials for conflict emerging from the structural shift from an industrial to a knowledge society, which we treat as similarly disruptive as the Lipset–Rokkanian “revolutionary” junctures. While the cleavages which characterized West European party systems for decades have lost much of their structuring power, new tensions have arisen from rapid educational expansion, occupational change, feminization of labour markets, concentration of high value-added economic activity in cities, and exposure to the multifaceted process of globalization (Reference KitscheltKitschelt 1994; Reference OeschOesch 2006a; Reference Kriesi, Grande and LachatKriesi et al. 2008; Reference StubagerStubager 2008; Reference Bornschier and RydgrenBornschier 2018; Reference Hooghe and MarksHooghe and Marks 2018; Reference Iversen and SoskiceIversen and Soskice 2019; Reference Garritzmann, Häusermann and PalierGarritzmann, Häusermann, and Palier 2022a; Reference HallHall 2022). These developments change the composition of society, and they benefit some groups more than others (e.g., the higher educated compared to the lower educated; or workers in knowledge-intensive and creative jobs compared to those in routine work). They do so materially, but also more broadly in terms of status, outlook, or the rise and demise of worldviews and ways of life (Reference CramerCramer 2016; Reference GestGest 2016; Reference HochschildHochschild 2016; Reference Noam and HallGidron and Hall 2017; Reference FitzgeraldFitzgerald 2018; Reference KurerKurer 2020; Reference BoletBolet 2021; Reference Kurer and Van StaalduinenKurer and Van Staalduinen 2022; Reference ZollingerHäusermann, Kurer, and Zollinger 2023).

While the timing, strength, and scope of structurally driven societal change has differed across Western Europe, these countries can today broadly be classified as postindustrial, globally integrated, knowledge-based economies. In other words, the “raw material” for a new cleavage (or cleavages) in terms of structural divides has likely emerged across these contexts. In all these countries, structural transformations of social patterns and norms have raised questions related to cultural liberalism and changing gender roles, immigration and multiculturalism, minority rights, or the boundaries of community. These topics have by now been widely taken up by socially liberal parties (especially the Greens) and nativist parties (especially Far Right, but also established conservative parties).

While some have studied the rise of the Far Right and green/left-libertarian parties from the vantage point of political entrepreneurship, the literature on electoral realignment provides a basis for thinking about them as expressions of a new cleavage. This literature links newly arising issue conflicts to the structural changes previously outlined, tracing the emergence of stable new alignments between voters and parties. By now, most observers of party system change in Western Europe will agree on the emergence of a “second dimension” of politics (clearly distinct from the traditional class cleavage), centered around predominantly sociocultural questions of individual liberties, societal organization, and community boundaries. In mobilizing this conflict, parties of the New Left and the Far Right garner disproportionate support, respectively, among a highly educated “new” middle class versus among lower-educated members of the working and “old” middle class (Reference KitscheltKitschelt 1994; Reference Kriesi, Grande and LachatKriesi et al. 2008; Reference BornschierBornschier 2010; Reference Gingrich and HäusermannGingrich and Häusermann 2015; Reference Häusermann, Kriesi, Beramendi, Häusermann, Kitschelt and KriesiHäusermann and Kriesi 2015; Reference Hooghe and MarksHooghe and Marks 2018; Reference Oesch and RennwaldOesch and Rennwald 2018; Reference De Wilde, Koopmans, Merkel, Strijbis and ZürnDe Wilde et al. 2019).

There is cross-country variation regarding exactly how the structural “raw material” has been mobilized, issues bundled and politicized, and how strongly new conflicts have gained expression within party systems. The literature identifies several waves in which the “universalism–particularism” dimension gained political traction, with the New Social Movements of the 1970s and 1980s finding expression in the emergence of the New Left and Green party family, followed by countermobilization driven by the Far Right (e.g., Reference Häusermann, Kriesi, Beramendi, Häusermann, Kitschelt and KriesiHäusermann and Kriesi 2015). These waves of mobilization and countermobilization have played out somewhat differently across countries: especially the conditions for new party entry are well-known to differ (Reference Vries, Catherine and HoboltDe Vries and Hobolt 2020); established parties on the Left did not equally adopt universalist positions early on (Reference KitscheltKitschelt 1994; Reference Rennwald and EvansRennwald and Evans 2014); Far Right countermobilization varied across countries in terms of strength and timing (Reference CarterCarter 2005; Reference Kriesi, Grande and LachatKriesi et al. 2008;); and varying reactions from mainstream right parties to Far Right mobilization have also impacted the salience of second dimension issues (Reference MeguidMeguid 2008; Reference van Spanjevan Spanje 2010; Reference BornschierBornschier 2012; Reference Abou-Chadi and KrauseAbou-Chadi and Krause 2018;).

We contend that bringing in the identity element is crucial to looking past such country variation, and ultimately to assessing whether a new cleavage is emerging. An identity perspective can shed light on the “stickiness” of new group-party alignments, as well as provide important insights into what motivates voters in terms of their perceptions and worldviews.

2.1.2 Identity and Cleavage Formation in Contexts of Realignment

Conceptualizing collective identities as located on an intermediary level between structure and political agency allows us to avoid structural determinism without viewing social identities as entirely socially constructed. It is especially important to make clear that subjective notions of group belonging need not correspond to the objective (educational, class, etc.) categories with which political scientists typically operate. This point also relates directly to a long-standing debate about whether the drivers of political transformations are primarily “economic” or “cultural” (cf. Reference ManowManow 2018; Reference Norris and InglehartNorris and Inglehart 2019): identities can provide the link from conflicts between (objectively defined) “winners” and “losers” of the knowledge economy to what can (mistakenly, we believe) be taken for “mere” culture wars or identity politics conjured up by political elites. To theorize which types of identity contrasts are likely to inform voters’ worldviews and preferences, complementing cleavage theory with insights from sociological and social psychological work on identities is insightful.

Sociologists in Weberian or Bourdieusian traditions have shown how groups monopolize privileges and resources by constructing “symbolic boundaries” of who belongs and who does not (Reference LamontLamont 2000; Reference Lamont and MolnárLamont and Molnar 2002; Reference Savage, Devine and CunninghamSavage et al. 2013; Reference RidgewayRidgeway 2019; Reference DamhuisDamhuis 2020; Reference WestheuserWestheuser 2021; also see Reference BartoliniBartolini 2005b). Material interests can motivate the drawing of (cultural, moral) in-group and out-group boundaries, but so can the quest for maintaining dignity and status. In this vein, especially ethnographic work in the past years provides key insights into the specific group understandings around which a new cleavage may well crystallize: pride in national or rural communities, identification with hard work, or adherence to traditional, conservative, more patriarchal moral standards of success provide a path to positive identity even for objective “losers” of economic and social change (the lower-educated, routine workers, etc.) (Reference CramerCramer 2016; Reference HochschildHochschild 2016; Reference DamhuisDamhuis 2020; Reference WestheuserWestheuser 2021). By contrast, cultural capital is becoming increasingly associated with the cosmopolitan, urban, culturally diverse lifestyles and consumption patterns of the highly educated middle class (Reference FloridaFlorida 2012; Reference Savage, Devine and CunninghamSavage et al. 2013; Reference Flemmen, Jarness, Rosenlund, Blasius, Lebaron, Roux and SchmitzFlemmen, Jarness, and Rosenlund 2019).

Social psychology further provides rich insights into individual-level psychological motivations of identity formation. Work building on social identity theory documents people’s innate tendency to simplify the social world by categorizing and stereotyping into “us” and “them.” A psychological desire to positively distinguish one’s own group fosters in-group bias and the derogation of out-groups (Reference Henri, Turner, William and WorchelTajfel and Turner 1979; Reference TajfelTajfel 1981; Reference HuddyHuddy 2001; Reference Roccas and BrewerRoccas and Brewer 2002; Reference StubagerStubager 2009; Reference MasonMason 2018; Reference Mason and WronskiMason and Wronski 2018). This aligns, in many ways, with sociological work concerned with the status concept (Reference LamontLamont 2000; Reference ShayoShayo 2009; Reference ZollingerZollinger 2022). Similarly, recent social psychological work that documents identity sorting and reduced “identity complexity” chimes with a notion from cleavage theory (Reference Roccas and BrewerRoccas and Brewer 2002; Reference MasonMason 2018): a lack of cross-cutting identities (e.g., today, if educational, class, and geographical divides align) intensifies identity conflict (Reference Bornschier, Häusermann, Zollinger and ColomboBornschier et al. 2021). This body of work especially highlights the affective component of identity antagonisms, which has recently entered issue-based/spatial accounts of political conflict via the concept of “affective polarization” (Reference Iyengar, Sood and LelkesIyengar, Sood, and Lelkes 2012; Reference Gidron, Adams and HorneGidron, Adams, and Horne 2020; Reference ReiljanReiljan 2020; Reference WagnerWagner 2021; Reference Hobolt, Leeper and TilleyHobolt, Leeper, and Tilley 2021; Reference Hegewald and SchraffHegewald and Schraff 2022).

Integrating these various strands of work results in a model of voting behavior in which cleavage identities mediate between individuals’ socio-structural position and their political behavior. Both in-groups and out-groups are relevant here, as are cognitive and affective aspects of group identification. These subjective self-perceptions correspond imperfectly to objective group belonging, because they serve the social and psychological goals of orientation and a positive self-understanding in a changing world. Identities thus conceptualized provide an entry for studying the “glue” of cleavage formation in the twenty-first-century knowledge economy environment. Comparing against the most pillarized, encapsulated, formally organized manifestations of traditional cleavages, stable and recurring patterns of electoral realignment in today’s fast-paced, individualized, politically volatile societies are somewhat puzzling at first glance. Indeed, the decline in actual partisan identity (not to mention party membership) is a key factor that complicates the study of contemporary cleavages. Our expectation is that studying the intermediary level of identities can reveal regularities and cross-national patterns that are hidden when looking at voting and partisanship alone. We expect voters to have a clearer understanding of the broader political blocks or camps where “people like them” belong (or do not belong) politically. Group identities that clearly link structural divides in the knowledge economy to political options that represent universalism versus particularism would be strongly indicative of a new, fully-fledged cleavage taking shape.

2.2.1 Survey Design

To study identities associated with a universalism–particularism cleavage, we fielded a bespoke online survey in France, Germany, Switzerland, and the UK. The surveys were implemented by the survey company Bilendi between November 2020 and January 2021. We recruited 2,000 participants each in France and Germany and 3,000 participants each in Switzerland and the UK. In Switzerland, we only recruited participants from the German and French speaking parts. In the UK, we only recruited participants from England, to avoid capturing conflicts over national identity (Scottish, Welsh, etc.) more specific to the UK context. The samples are population-representative in terms of education, age, and gender.

We selected these four countries because they are part of the same historical area of political cleavage formation but are at different stages of electoral realignment. While France and Switzerland are representative of countries that experienced early and strong realignment, Germany and the UK are cases of late and less consolidated realignment (Reference Kriesi, Grande and LachatKriesi et al. 2008; Reference BornschierBornschier 2012). The former two countries saw the early establishment of a strong “particularist” Far Right. Switzerland’s major left parties further jointly represent a particularly extreme articulation of the “universalist” New Left side of a new divide (Reference Rennwald and EvansRennwald and Evans 2014; Reference Bornschier, Häusermann, Zollinger and ColomboBornschier et al. 2021). Meanwhile, the Far Right in Germany made a much later breakthrough (including for institutional and historical reasons), and in the UK, the party system constrained the articulation of new divides in the political arena (although the Brexit referendum seems to have been a substitute and catalyst in this respect, cf. Reference Hobolt, Leeper and TilleyHobolt, Leeper, and Tilley 2021).

Importantly, despite this variation in party system change, the structural transformations expected to generate new identity potentials have occurred in all four countries. The fact that the last bastions of resistance against the Far Right have fallen in most Western European party systems suggests that the space for political agency is limited, certainly in fully preventing the emergence of this divide. The strategic action of established political parties helps explain why Far Right parties were able to break into party systems earlier in places like France, Switzerland, Flanders, and the Netherlands, and much later in Germany, Britain, as well as a parts of Southern Europe. As a result, the electoral realignment of educational groups or classes along the second dimension, shifts of dimension salience, and the association of specific bundles of issues with New Left and Far Right parties is more entrenched in some countries than others (differences we address in Section 5). However, while their political expression may vary somewhat, we expect to see similar identity divides anchored in structural divisions across all four countries.

In the survey, we ask respondents about their sociodemographic characteristics, political attitudes, and party preferences. However, the center piece of the survey are novel questions on group identity (building on Reference Bornschier, Häusermann, Zollinger and ColomboBornschier et al. 2021). We work mainly with a series of closed-ended questions in which we ask respondents about perceived closeness to different social groups (“Of the following groups, how close do you feel towards them? By ‘close’ we mean people who are most like you in terms of their ideas, interests, and feelings.”), on a ten-point scale ranging from “not at all close” to “very close.” We consider this survey item a good starting point to extend research in cleavage theory to encompass group identities. The validation of this measure is discussed extensively in the appendix to Section 2.

2.2.2 Selection and Measurement of Identities

The survey asked about belonging to seventeen specific groups, in randomized order. Our aim was to tap into group categories that are closer to the structural foundations of a new divide according to the literature (education, class, place of residence) and others that are closer to the sociocultural distinctions through which these divides become manifest (e.g., conationals versus cosmopolitans). We also aimed to include both identities theorized to be newly emerging and others that were already associated with traditional cleavages. In selecting group categories and developing the wordings, we drew inspiration from qualitative work (e.g., Reference LamontLamont 2000; Reference SavageSavage 2015; Reference CramerCramer 2016), built on a previous Swiss study (Reference Bornschier, Häusermann, Zollinger and ColomboBornschier et al. 2021), and on results from open-ended survey questions (Reference ZollingerZollinger 2024).

Concretely, the survey asked about the following groups (the distribution of responses is shown in appendix Figure A2.2):

• Education: We use three categories, namely, people with a higher education degree, people with medium-level education (vocational training in Germany and Switzerland), and people with lower-level education. Already associated with traditional class divides (Reference BourdieuBourdieu 1984), education has become recognized as the primary structural divide underpinning a new universalism–particularism conflict, especially in knowledge-based economies (Reference StubagerStubager 2008; Reference Kriesi, Grande and LachatKriesi et al. 2008; Reference Iversen and SoskiceIversen and Soskice 2019).

• Class: To tap into the traditional vertical class dimension, we asked respondents how close (or distant) they felt to wealthy people and people with humble financial means. Targeting more horizontal class divisions (Reference OeschOesch 2006b; Reference Savage, Devine and CunninghamSavage et al. 2013) as well as work more versus less associated with the knowledge economy, we further asked about people who do hard, tiring work, people who do creative work, and people who work in the social and education sector.

• Residence: We asked about closeness to urban and rural people, given the geographical dimension of emerging divides in the knowledge society (Reference CramerCramer 2016; Reference FitzgeraldFitzgerald 2018; Reference MaxwellMaxwell 2019; Reference Iversen and SoskiceIversen and Soskice 2019; Reference PatanaPatana 2022).

• National (nativist) identity: The survey asked about closeness to German/Swiss/French/British people and about closeness to people with a migration background. Asking about people with a migration background also allows us to partly tap the ethnic and racial component of divides over diversity in a European context.

• Parochialism/communitarianism versus universalism/cosmopolitanism: Here, we asked about cosmopolitans and people who are down-to-earth and rooted to home. We further asked about self-perceived belonging to culturally interested people, given milieu studies that indicate “cultural capital” becoming increasingly linked to cosmopolitan (urban) lifestyles (Reference SavageSavage 2015; Reference Flemmen, Jarness, Rosenlund, Blasius, Lebaron, Roux and SchmitzFlemmen, Jarness, and Rosenlund 2019).

• Gender/Feminism: An item about closeness to feminists is designed to capture the gender component of new conflicts, with especially the women’s movement having contributed to the emergence of the New Left.

• Religion: Lastly, we ask about closeness to people with Christian-Western values. Traditionally associated with the religious cleavage and Christian Democracy, transformed aspects of this divide could also feed into a newer universalism–particularism divide.

Throughout this Element, the following groups will turn out as differentiating most effectively between the two extremes of the universalism–particularism dimension: nationals, people with a migration background, cosmopolitans, and people who are down-to-earth and rooted to home. We will therefore zoom in on these groups in many of the analyses. Further items designed around the related concepts of identity, social closure, and political mobilization inquired into respondents’ networks or social interactions, perceived overlaps between different types of group boundaries, or voters’ associations of groups with specific parties. The survey also included a conjoint experiment that asked respondents to choose identity profiles to which they felt closer. The exact wording and operationalization of identity-related concepts based on these items will be detailed in the subsequent sections, along with the presentation of results.

2.2.3 Identification and Operationalization of Party Blocks

As already explained, our aim in studying collective identities is partly to look beyond country specificities in party competition to common patterns of cleavage formation. We are not primarily interested, for instance, in the role that social democratic (versus Green/left-libertarian parties) play in mobilizing the universalist side of a cleavage; or in whether a specific mainstream conservative party dabbles in the particularist Far Right field. Our interest lies in more overarching identity divides that provide voters with the cues of where they belong in the political landscape.Footnote ¹⁰

Following this logic, we investigate the political articulation of a new cleavage from a perspective of party blocks. We derive these blocks empirically and validate them theoretically. Starting from an empirical classification makes sense here because, as mentioned, parties’ ideological positions matter more than their belonging to historically grown party families. Our party classification is based on mixture models fitted to the 2019 Chapel Hill Expert Survey (CHES) data on all Western and Southern European countries, excluding Eastern Europe. The mixture model allows for a clear identification of clusters and their structures based on the Bayesian Information Criterion (BIC). Moreover, it probabilistically assigns parties to clusters, which makes it easier to handle parties with amorphous profiles.

We proceed in two steps: First, based on the economic left-right and GAL-TAN dimensions of the CHES, we start with an initial clustering of parties. As per the BIC, this results in three clusters: one that is economically and socially left-wing, one that is economically right-wing and socially progressive, and another that is economically right-wing and socially conservative (roughly conforming to a tripolar model of the political space, but with the third cluster reaching beyond what is typically discussed as the Far Right party family). In a second step, we refine each first-stage cluster based on the issues on which it is most heterogeneous (considering sixteen issues available in the CHES). Note that the BIC does not indicate sub-clustering to be necessary for all first-stage clusters, but we consider within-cluster differences to gain further insights. This second step produces six narrower clusters:

1. 1. Green left, consisting of pro-environmental, economically left, and culturally progressive parties.

2. 2. Traditional left, consisting of economically left and culturally progressive parties that take more moderate environmental positions.

3. 3. Left liberals, consisting of parties that take pro-environmental stances, favor civil rights over law and order, and tend to cater more to urban interests.

4. 4. Right liberals, who are more moderate on the environment, tend more to law and order positions, and have a more rural focus than the left-liberals.

5. 5. Traditional right, consisting of economically right and culturally conservative parties. Those parties, however, favor open societies and are more moderate on moral issues than the radical right.

6. 6. Far Right, who are economically conservative, favor a closed society, and traditional mores.

In subsequent analyses, we focus on three party blocks based on this differentiated perspective: one based on the Far Right cluster, one on the traditional right cluster, and a New Left block that combines the green left and traditional left clusters (both of which are broadly left-wing economically as well as culturally, in line with the tripolar model of political competition, Reference Oesch and RennwaldOesch and Rennwald 2018). We opt to disregard the left liberal cluster (which only concerns the Swiss Green Liberals in the countries we study), and we go on to consider the right liberal cluster specifically for France, where we would otherwise exclude Macron’s La République En Marche as well as the Mouvement Démocrate (MoDem). Table 1 shows the final classifications of parties in our sample.Footnote ¹¹

An important question concerns electoral volatility. Here, we followed Reference Steenbergen and WilliSteenbergen and Willi (2019) and derived consideration sets from our respondents’ stated propensities to vote for different parties (full details are in the online appendix). For each respondent, we identified the list of parties they consider voting for. We then determined whether those parties are mostly situated within one of the blocks we identified, or whether there is a great deal of cross-block consideration. Our methodology suggests that the average consideration set size was greater than one party, suggesting a potential for volatility. However, this volatility appears to have been limited. The percentage of consideration sets that included only parties from one of the four blocks in Table 1 were 66.0 in England, 73.6 in France, 71.9 in Germany, and 65.3 in Switzerland. We conclude that there is constrained electoral volatility in the four countries. In addition to measuring propensities to vote, we also asked respondents for their party choice, that is, which party they would vote for if national elections were held next Sunday. We use the latter measure for most analyses conducted in the following sections.

Having introduced our survey design and established our definition of party blocks, we are now in a position to evaluate our theoretical claims empirically. In the following section, we first study the link between social structure and social identities, before we bring in party choice in Section 4.

3.1.1 Social Structure in Knowledge-Based Economies: Education, Class, Residence

We focus our analyses of the sociocultural rootedness of collective identities in education, occupational class, and urban-rural residence. These characteristics have been shown to deeply structure both the distribution of material and cultural opportunities in today’s societies, as well as political attitudes and voting behavior. At the same time, their link to universalistic-particularistic collective identities is far from trivial, as these sociological categories do not directly materialize in the political organizations and mobilization strategies we observe in Western Europe.

Education is the posterchild of social stratification in the knowledge economy. To quote Reference Hooghe, Marks and KamphorstHooghe, Marks, and Kamphorst (2022): “The university is to the information revolution what the factory is to the industrial revolution.” Not only have we witnessed a tremendous expansion of higher education across Western Europe over the past decades – with today on average 60–70 percent of young people enrolled in some form of tertiary education in Western economies compared to ca. 30 percent in the 1970s (Reference Garritzmann, Häusermann, Kurer, Palier, Pinggera, Garritzmann, Häusermann and PalierGarritzmann et al. 2022b) – but we also see education premiums rise across all developed economies (Reference Weisstanner and ArmingeonWeisstanner and Armingeon 2020).

In line with the literature, we distinguish lower (primary and lower secondary education), middle (upper- and post-secondary education and professional training), and higher levels of education degrees (tertiary education). Its close link to the knowledge economy brings us to prioritize education over narrower aspects of social stratification such as income or wealth.

Beyond socioeconomic status, the unequal occupational dynamic as a consequence of deindustrialisation and service sector growth (Reference OeschOesch 2013), as well as task-biased technological change (Reference Kurer and PalierKurer and Palier 2019) have brought to the forefront conceptualizations of “types” of tasks and education – that is, how the everyday experience of different work logic shapes how people evaluate opportunities, and how they think about universalism and particularism.

We conceptualize this idea of milieu, work logic, or educational socialization by means of two indicators, educational field on the one hand and occupational class on the other hand. In terms of education field, we follow Reference Hooghe, Marks and KamphorstHooghe, Marks, and Kamphorst (2022) and focus on an indicator of the ratio of communicative and cultural skills in relation to all types of skills in a particular education field (cultural, economic, communicative, and technical skills, CECT). This measure is based on four types of resources characterizing different fields of education (Reference de Werfhorst, Herman and Kraaykampvan de Werfhorst and Kraaykamp 2001): cultural, communicative, economic, and technical. In line with Reference Hooghe, Marks and KamphorstHooghe, Marks, and Kamphorst (2022) we allocate the CECT value (between 0 and 1) to fourteen education fields.Footnote ¹²

The distinction between occupational classes along vertical and horizontal lines (Reference OeschOesch 2006b) taps into a very similar idea, that is, the everyday socialization of people through experiences and tasks. Building on extensive research highlighting the relevance of occupational classes for economic resources and risks (Reference OeschOesch 2006b, Reference Häusermann, Kurer and SchwanderHäusermann, Kurer, and Schwander 2014), as well as for preferences and electoral behavior (e.g., Reference Kitschelt and RehmKitschelt and Rehm 2014), we distinguish business owners, technicians, production workers, managers, clerks, sociocultural professionals, and service workers.

Finally, we study where people live as a key dimension of social structuration in the knowledge economy. Given the creative, interpersonal, and communicative aspect of knowledge economy jobs, it is argued that the structuring impact of urban-rural – or centrist-remote – residence is particularly strong in the era of knowledge-based growth. Indeed, opportunities in terms of employment, access to education and jobs, as well as socially and culturally diverse activities tend to concentrate in urban centers (Reference Iversen and SoskiceIversen and Soskice 2019). Hence, we also explore how urban-rural residence relates to group identities. We measure objective residence in terms of population size of respondents’ place of residence, calibrated to the national context. We distinguish bigger cities (more than 100,000 inhabitants; in Switzerland more than 50,000 inhabitants), (sub)urban areas, and small towns and rural areas (less than 5,000 inhabitants; in Switzerland, less than 2,000).

3.1.2 How Socio-structural Categories Translate into Sociocultural Group Identities

In this section, we show how group identities – and in particular culturally connoted group identities – relate to socioeconomic categories of social structuration. A different question, of course, is why ascriptive social categories relate to subjective group perceptions that do not correspond straightforwardly to the related socioeconomic categories. The step from objective group belonging to subjective identity (and eventually political mobilization) goes to the heart of political sociology and is nontrivial in any case. The link between education, class, and residence to cosmopolitanism, feminism, national identity, feeling down-to-earth and rooted to home, feeling close to people with migration background, and so on seems even more complex to theorize. However, as developed in this section, the rise of the knowledge economy has transformed life conditions, prospects, and risks of individuals not only in material-economic terms but also in broader sociocultural terms.

As the previous section argued, both cultural sociology and social psychology provide us with tools to understand why people identify with social groups that provide them with self-value, status, and a sense of recognition. Because of such motivational drivers, a new cleavage may crystallize around pride in national or rural communities, identification with hard work, or adherence to traditional, conservative, more patriarchal moral principles – in demarcation from the increasingly cosmopolitan, urban, “alternative” habitus of the educated middle class. Translating this idea to our study of changing group identities, it explains why “losers of the knowledge economy” may rather draw group boundaries in terms of cultural identities based on notions of “hard work,” being “down-to-earth,” or nationality. Similarly, the “winners of the knowledge economy” may tend to draw recognition and identity from values and habits linked to universalistic orientations – for example, cosmopolitanism, openness, and support for social and political minorities – rather than from their education or occupational class.

Research on the material determinants of behavior has started to integrate some of the insights on the need for status recognition and self-worth through group boundaries. One key aspect is the increasingly strong focus on relative measures of grievances, acknowledging that people evaluate their social situation in comparison to reference groups (Reference Kurer and Van StaalduinenKurer and van Staalduinen 2022; Reference ZollingerBreyer, Palmtag, and Zollinger 2023; Reference ZollingerHäusermann, Kurer, and Zollinger 2023). Experiencing relative economic decline – or the threat of it – is indeed positively correlated with support for radical right and radical left parties (Reference Burgoon, van Noort, Rooduijn and UnderhillBurgoon et al. 2019; Reference KurerKurer 2020; Reference Engler and WeisstannerEngler and Weisstanner 2021; Reference Kurer and Van StaalduinenKurer and van Staalduinen 2022).

It is important to highlight the implication of taking these mechanisms seriously. The long-standing debate about whether the drivers of political-electoral behavior are primarily “economic” or “cultural” (cf. Reference ManowManow 2018; Reference Norris and InglehartNorris and Inglehart 2019) has been misleading. The question is not whether people radicalize politically because they are either “poor” or “racist.” Rather, people experience material transformations of their life circumstances in a broad array of forms: in terms of opportunities for mobility, whether they are part of an upward social–economic–cultural dynamic or a precarious one, whether the world looks increasingly similar to what they are familiar with or the opposite, and so on. These very real transformations lead individuals to (re-)imagine themselves as members of social groups that provide order, reassurance, and status. Collective identities provide the link from conflicts between (objectively defined) “winners” and “losers” of the knowledge economy to what can mistakenly be taken for “mere” culture wars or identity politics. Conflict over group identities and perceptions of deservingness, fairness, or “common-sensical” evaluations entail economic as well as cultural evaluations (Reference Damhuis and WestheuserDamhuis and Westheuser 2023). This is likely why material compensation seems to have very limited effect on preventing radicalization among declining groups (Reference GingrichGingrich 2019). Identity divides develop a powerful ideological–political map through with individuals interpret the world, a map that solidifies through mobilization and representation – in short, processes of cleavage formation.

3.2.1 Validating the Structural Rootedness of Identities

We start with Figures 2–4, which provide the findings of linear regression models, pooled across countries (with data weighted by education, gender, and age), predicting closeness to key identity groups by education level, education field, occupational class, residence, age, and gender. Since the “objective” sociodemographic attributes of respondents relate in a plausible and straightforward manner to the corresponding “subjective” group identities (Figure 2), we validate our measure of group identity (closeness) and substantiate that identity conflicts are tightly linked to people’s material life conditions.

Figure 2 shows this validation link between objective and subjective groups for three exemplary identities.Footnote ¹³ The level of education clearly and significantly predicts closeness to people with high education. Education also has some explanatory leverage for class identities (work in the social/education sector) and urban identities, but decidedly less so than for education itself. Rather, closeness to class identities is better explained via both field of education (CECT) and occupational class. People with a high share of communication and cultural skills in their education profile, as well as sociocultural professionals, identify particularly strongly with people working in the social and education sector. Conversely, production workers, who epitomize traditional, routine manual labor, feel most distant to people working in the education and social sectors. Finally, objective residency predicts closeness or distance toward urban identity most strongly.

Figure 2

Education level and field, class, and place as determinants of closeness toward groups defined by education, work logic and residence

Figures 3 and 4 go beyond the validation of subjective group identities. Rather, they show that even key sociocultural group identities are firmly rooted in education, educational field, occupational class, and/or residency. The pervasive and consistent predictive power of education level is particularly striking: People with high education feel close to cosmopolitans, to people with migration background, to feminists, culturally interested people, and to people with Christian-Western values. At the same time, they feel relatively distant from people who describe themselves as down-to-earth and rooted to home, and from people who identify via their nationality. These patterns are important because of their consistency, and because they contradict objective patterns to some extent. For instance, people with migration background “objectively” tend to have rather below-average education levels, yet patterns of identification are reversed.

Figure 3

Education level and field, class and residence as determinants of closeness toward culturally connoted identity groups (part I).

Figure 4

Education level and field, class and residence as determinants of closeness toward culturally connoted identity groups (part II).

The fact that CECT and occupational class show strong links to some of the culturally connoted group identities despite controlling for education level in the models shows that the immediate “task experience” (in training or work) matters beyond the mere level of human capital. CECT shows strong negative links with national identity and positive ones when it comes to closeness to people with migration background, feminists, and culturally interested people. Occupational class as a determinant of closeness shows more variation. On average, production workers and sociocultural professionals appear as the most antagonistic classes, especially with regard to closeness to people who are down-to-earth and rooted to home, as well as people with migration background, feminists and culturally interested people, from whom production workers demarcate themselves. Finally, people living in big cities feel relatively closer to all the universalistic identities and more distant to all the particularistic ones. On the other hand, people living in small towns and rural areas are not clearly distinct from people in small cities or agglomerations. It seems that the identity-dividing line in terms of residency is clearly between urban residents in big cities and everyone else.

A final word on the category of people with Christian-Western values. This group identity is less clearly attributable to the universalistic or particularistic pole. Closeness to these people relates more clearly to elderly male, highly educated managers. As we will see in later analyses, this identity is indeed not part of the straightforward universalism–particularism divide but relates more closely to a traditional right-wing conservative identity.

Through the analysis of predicted values of closeness (see appendix Figure A3.1) we see that education levels predict about the same substantive effect in terms of closeness to the direct correlate (closeness to people with a high education), as to key culturally connoted groups – cosmopolitans and people with migration background.

Finally, we repeated all regressions for each of the four countries, to analyze cross-sectional consistency. Table 2 summarizes – for the culturally connoted group identities – the links between education levels, CECT, occupational class and residency, and closeness to group identities. The main finding is that we find strikingly consistent patterns across countries. In particular for the “cosmopolitan-rooted to home” pair of group identities, we find identical structural foundations of closeness in most countries. The same goes for closeness to people with migration background, feminists, and culturally interested people. For the other groups, the socio-structural foundations are less consistent. For national identities, in particular, the baseline level of identification is high, and the clearest patterns are a relatively strong demarcation by sociocultural professionals and people with communicative-cultural education.

In light of the analyses of this section, we are now able to distinguish a set of identity groups that stand out by their socio-structural foundation and consistency, and which epitomize the universalistic–particularistic identity divide: cosmopolitans and people with migration background as groups to whom individuals at the universalistic pole feel distinctively closer; people who are down-to-earth and rooted to home as groups to whom individuals at the particularistic pole feel distinctively closer. In the further analyses, we will focus on these groups to exemplify our analyses.

3.2.2 The Fundamental Role of Education

So far, we have shown that group identities are clearly and substantively rooted in socio-structural categories, and that these links are consistent across a range of identities and across countries. In this section, we further substantiate the claim that education is one of, if not the key, structural basis of the new cleavage. To do so, we leverage two original elements of our survey: the extent to which individuals perceive education groups to differ along other dimensions than human capital, and a conjoint survey allowing us to show which attributes of a hypothetical person drive a feeling of closeness to this person.

First, the perception of overlaps between different divides. We asked respondents the following question: “Please imagine people with different levels of education (for instance, people with and without a university degree). Do you think these educational groups also differ with regard to the following characteristics?” The response items concerned social class, place of residence, hobbies, and values. If education is a key structural element of cleavage formation, then respondents should perceive education as having a structuring effect on these different dimensions. Moreover, this perception should be particularly strong among individuals at the extremes of the universalism–particularism divide. Figures 5 and 6 confirm these expectations. In Figure 5, we see that across all countries, large majorities of respondents think that people with different levels of human capital also differ in terms of class, residency, hobbies, and values. Hence, education emerges as a heuristic associated with both clear structural and sociocultural correlates for most people.

Figure 5

The perceived structuring effect of education: distribution of evaluations whether people who differ in education also differ in other structural and cultural characteristics

Figure 6

The intensity of cosmopolitan identity as a determinant of the perceived structuring effect of education

If this association of education with other aspects of social distinction is indeed the foundation of the new cleavage studied in this Element, then it should be particularly strong among respondents with clear identities relative to this cleavage. To test this, Figure 6 regresses responses to the aforementioned question on the intensity of identifying with cosmopolitans. We measure intensity as the absolute deviation from the respective country mean. As we would expect, we see that respondents with more intense identities (both positive and negative toward cosmopolitan people) perceive the structuring effect of education more strongly.

The conjoint analysis we conducted adds further evidence to the claim that education is the key structural basis of the new cleavage. In the pairwise conjoint comparisons, we asked respondents to indicate which person they feel closer to. The hypothetical individuals were characterized by randomized values on four attributes: cosmopolitanism/openness versus national/parochial orientations; work and work ethic, reflecting different work logics; residence (city versus countryside); and the person’s affinity to and perspectives on education.

Contrary to the linear regressions whose findings we show in Figures 2–4 and in Table 1, the conjoints allow us to observe the relative importance respondents attribute to different aspects of identity. In Figure 7, we show the marginal means for respondents with high and low education level. Findings are largely consistent across countries. Three findings are particularly relevant: first, high- and low-educated respondents differ clearly and significantly regarding the extent to which specific value orientations, work logics and residence yield closeness to a hypothetical person; second, despite defining the subgroups by education, values, work logic and residency yield stronger effects than characteristics more closely linked to the attribute of education itself. Hence, education is the structural basis, but it translates into different group identities. Third, when it comes to value orientations (cosmopolitanism, values, culture), the effects are not just different in strength, but even diametrically opposed. These effects highlight the strongly divisive potential of these values.

Figure 7

Choice (closeness) conditional on education: marginal means

Importantly, when replicating Figure 7 for sociocultural professionals and production workers, the findings look somewhat weaker but very similar, including the diametrically opposed effects when it comes to value orientations. Hence, we are not arguing that education level is the only relevant basis for cleavage formation. Rather, task/field socialization clearly adds to the formation of relevant cleavage foundations.

3.2.3 Evidence for Cleavage Formation

In a last step of this analysis of the structural foundation of the new cleavage, we report findings on social network formation and symmetry in the identity hierarchies of people at the extremes of the universalism–particularism divide.

A first piece of evidence for actual cleavage formation results from our analysis of social networks. Given the socio-structural underpinnings of the universalism–particularism divide in terms of overlapping and concentrating educational experiences, workplace, or geographical contexts, we expect the formation of social “milieux.” Voters are socialized into social networks, develop notions of “people (un)like them,” and these notions themselves may further limit social interactions across group boundaries. Figure 8 provides evidence for the formation of social milieux along similar identities. It is based on survey questions asking all respondents about personal interactions with members of the different groups we focus on. More specifically, we asked how often “they have personal conversations or spend their time” with people who belong to these groups. We see that closeness to typical universalistic and particularistic identities predicts contact with people from the same category of identities. The closer respondents feel to cosmopolitans, the more they report direct interaction and time spent with people with migration background. The more people identify as down-to-earth and rooted to home, the more frequent they describe their interactions with “conationals.” We chose conationals and people with migration background as network groups, as they are objectively extremely prevalent in society. It is highly unlikely that people only interact several times a year or less with conationals, or that they hardly ever have a conversation with someone with a migration background. Yet, the subjective closeness to these ideal-typical universalistic and particularistic identities predicts respondents’ reporting and perception of their social network quite strongly.

Figure 8

Social network formation along identity lines: frequency of having personal conversations and spending free time together

Another indication of such milieu and cleavage formation is that individuals at the extremes of the cleavage recognize the other side as antagonist. This idea implies a mutual awareness and recognition of divisive values and attributes. Figure 9 presents the same conjoint findings as shown in Figure 7 for individuals with strong universalist–particularist identities. We see that the same attributes tend to elicit opposite reactions from respondents with strong cosmopolitan identity as opposed to respondents who identify strongly with people who are down-to-earth and rooted to home. Hence, the two sides of the cleavage indeed seem to read and interpret the attributes through similar – yet opposite – lenses.

Figure 9

Choice (closeness) conditional on “cosmopolitan” and “down-to-earth and rooted to home” identities: marginal means

5.2.1 Ideological Blocks

Our first aim is to understand how accurate voters’ mental map of their national party systems is: Do voters perceive the same links between identities and party blocks that emerged from our analysis in Section 4? We start with the universalistic identities: cosmopolitans and people with migration background. As Figure 13 shows, both identities are strongly associated with parties on the left, in line with our previous findings.Footnote ¹⁸ This holds both in Germany and in Switzerland, and reflects the fact that both countries saw an early transformation of the Left, most evident in the emergence of Green parties.

Figure 13

Perception of group–party associations, universalistic identities

Next, we move to the key particularistic identities, conationals, and people who are “down-to-earth and rooted to home.” Here, Figure 14 shows a somewhat more nuanced picture. In Germany, both identities are clearly associated with the mainstream Right. Down-to-earth people are even somewhat associated with the Left. In Switzerland, by contrast, both identities are associated with the Far Right. Going back to Figure 10 in the previous section, we see that the perception of Swiss voters closely corresponds to actual voting behavior. In Germany, both identities are almost equally strongly associated with voting for the mainstream Right and for the Far Right. In a direct comparison with Switzerland, it thus seems accurate that these identities are also more strongly associated with the mainstream Right by voters themselves.

Figure 14

Perception of group–party associations, particularistic identities

Finally, in Figure 15, we look at the more traditional identities, “people who do hard, tiring work” and rural people, for which we expect political associations to have shifted compared to a Lipset–Rokkanian world. For “rural,” we find the same pattern as for the particularistic identities: It is associated with the mainstream Right (and partly with the Left) in Germany and with the Far Right in Switzerland.Footnote ¹⁹ Interestingly, “Hard, tiring work” is associated with the Left in Germany but with the Far Right in Switzerland. A close look at the data used in Section 4 shows that these perceptions are again very accurate for Switzerland, in the sense that they match how identification relates to vote choice. In Germany, by contrast, both identities are only weakly correlated with political behavior. Interestingly, these perceived voter–block links are almost identical for those who strongly identify and those who do not identify at all with a group themselves (see appendix Figures A5.1–A5.6).

Figure 15

Perception of group–party associations, contested identities

Ultimately, these three graphs support our main contention that voters have a relatively accurate mental map of group–party links on the level of ideological blocks. They consistently identify the block whose voters most strongly identify with a social group as the main representative of that social group. Moreover, telling differences between Germany and Switzerland emerge. At the universalistic pole, associations between groups and parties are very similar in both countries. At the particularistic pole, however, down-to-earth, Swiss, and rural people are all predominantly associated with the Far Right SVP in Switzerland. In Germany, by contrast, these identities are (still) clearly associated with the mainstream Right. Finally, people “who do hard, tiring work” are still associated with the traditional class cleavage (and thus with the Left) in Germany. At almost 50 percent, this is the strongest association we see with any German party block across all six identities. In Switzerland, the picture is very different. Here, the same group is associated with the new cleavage and the Far Right in particular. The realignment of the traditional working class seems largely complete in the eyes of Swiss voters. This suggests that the timing and intensity of realignment also affects voters’ perceptions of the party system.

In Figures A5.7–A5.9 in the appendix, we show the same graphs for France and England. Here, the character of a two-party system comes out very clearly in the UK. Not only cosmopolitanism and a migration background are clearly associated with the left but also down-to-earth and hard work. This points to the weakness of the Far Right, as well as to the persisting upper class connotation of the mainstream right. Whereas rural is associated to an equal extent with the left and the mainstream right, conationals is the only category that is associated to a substantial degree with the Far Right (although less than with the mainstream right). Other research indicates that given the constraints of the British party system, Brexit might be more tightly linked to the political articulation of a new cleavage than partisanship, which cross-cuts the Leave/Remain camps (Reference Hobolt, Leeper and TilleyHobolt, Leeper, and Tilley 2021).

In France, we see stronger evidence of an association of the particularistic pole with the Far Right. Here, nationals are most strongly associated with the Far Right, which also scores relatively highly on rural, down-to-earth, and hard work. However, the association of hard work with the left is even stronger, thanks to the reinvigoration of the state–market cleavage by Mélenchon’s France Insoumise (the more detailed results show that it is this party, rather than the other components of the left, that voters associate with the hard working). This suggests that the relative strength of cleavage dimensions matters for the realigning potential of the new cleavage: Where the economic dimension remains polarized, the left and the Far Right may compete for some of the same voters. In this sense, France is more similar to Germany, where the Social Democrats were able to retain voters harboring these identities due to their reluctance to fully embrace a New Left position, than to Switzerland. Consistent with the three other cases, cosmopolitanism and migration background are clearly associated with the New Left block in France.Footnote ²⁰

5.2.2 Cohort Differences

In the next step of the analysis, we look at cohort differences in the perception of parties. Here, our hypothesis is that the mental map of younger voters, who have been politically socialized after the transition to the knowledge economy and the politicization of new issues, should reflect the new cleavage in more crystallized ways than the mental map of older voters. This difference should most strongly concern those identities which always mattered to politics but in which there has been a shift in who represents them, such as hard-working or conational. By contrast, we would expect weaker differences for identities such as “cosmopolitan,” since older people probably did not develop a strong political image of these identities in their youth, that is, before these groups were strongly politicized.

Indeed, when looking at the assignment of “cosmopolitan people,” we hardly see any differences between respondents under the age of forty and respondents over the age of sixty (Figure 16). This is quite different, however, for people “doing hard, tiring work” (Figure 17). Here, the oldest respondents in both countries still seem to see the party system through the traditional class-cleavage structure. This is much less the case for younger respondents. In Germany, a plurality still associates people doing hard, tiring work with the Left (i.e., with Social Democrats), but the share is much lower for younger than for older voters. In Switzerland, a plurality of younger respondents even associates this group with the Far Right. Similarly, there is a substantial age difference for conationals. Among the elderly in both countries, this identity is more strongly associated with the mainstream right than among younger voters (Figure 18).

Figure 16

Perception of group–party associations for cosmopolitans, two age groups

Figure 17

Perception of group–party associations for people who do hard work, two age groups

Figure 18

Perception of group–party associations for conationals, two age groups

Of the other three identities of interest, the pattern for people with “migration background” is very similar to the pattern for “cosmopolitans.” There is a somewhat higher share of “don’t knows” among the old, but the relative order of the parties is the same across age groups. Similarly, the pattern for “down-to-earth” is very similar to the pattern for conationals (see appendix, Figures A5.10 and A5.11). The only identity where age plays a role in one country but not in the other is “rural” (Figure 19). Here, Germany shows the familiar pattern of a stronger alignment with the new cleavage among the young than among the old. This is not the case in Switzerland, where majorities of all age groups associate “rural” with the SVP. This reflects the fact that the SVP used to be a farmer’s party before it developed into a Far Right party.

Figure 19

Perception of group–party associations for rural people, two age groups

5.2.3 Differences within the Left Block

We now move on from looking at blocks to the level of individual parties. Since there is only one Far Right party in both countries, and since the demarcation between FDP and CVP in Switzerland is of less theoretical interest in the context of the new cleavage, we focus on the demarcation within the left block regarding the universalistic identities. For Switzerland, this means that we compare the Social Democrats and the Greens. In Germany, we add Die Linke as a third party within the left bloc.

As Figures 20 and 21 show, there are considerable differences between both countries. In Switzerland, cosmopolitans and people with a migration background are more strongly associated with the Social Democrats than with the Greens. In Germany, both identities are much more strongly associated with the Greens.

Figure 20

Perception of group–party associations on the left, universalistic identities

Figure 21

Perception of group–party associations on the left, contested identities

Regarding people “who do hard, tiring work,” we had already seen that this identity is clearly associated with the Far Right in Switzerland and more with the Left in Germany. Figure 21 now reveals that this pattern is mainly driven by the Social Democrats, and, to some extent, by Die Linke. This confirms that we really observe a more traditional perception of cleavages here: This group is not at all associated with the New Left (the Greens) but with the more traditional Social Democratic or even post-Communist Left. Rural, finally, is an identity where we don’t see strong differences between the two countries.

To some extent, these differences are likely due to differences in the baseline of party sizes. The Swiss Social Democrats are roughly twice as strong as the Greens, while the German Greens and Social Democrats roughly performed on the same level in many recent (state-level and European) elections. This makes it much more likely a priori that a “typical” universalist voter in Switzerland votes for the Social Democrats. Nevertheless, these effects are much stronger than what could be explained by pure baseline effects.

To further explore this within-block variation, we now connect perceptions of the party system with individual vote choices. In Section 4, we demonstrated that identities are systematically associated with the choice of different party blocks. Here, we now further differentiate this relationship within the left block.

5.2.4 Relationship with Vote Choice

Figures 22 demonstrates the association between closeness to cosmopolitans/hard-working people and support for Social Democrats/Greens. Both figures further demonstrate striking differences between Germany and Switzerland. In Switzerland, identifying as a cosmopolitan is almost equally strongly correlated with voting for the Social Democrats and with voting for the Greens. In Germany, by contrast, this identity only correlates with support for the Greens, while there is no correlation at all with voting Social Democrat.

Figure 22

Closeness to cosmopolitans/hard-working people and support for left parties

The right-hand part of the figure shows the reverse picture for people doing hard, tiring work. In Switzerland, this identity is negatively correlated with voting for both left parties. In Germany, by contrast, it is not correlated with the Social Democrats but negatively with the Greens.

Thus, we again see cross-country differences that are in line with our argument about the role of early and late realignment and the Swiss Social Democrats’ comparatively early adoption of a New Left platform. In Switzerland, universalism is significantly associated with support for the Social Democrats. In Germany, by contrast, the prototypical universalistic identities, cosmopolitanism, and migration background are not significantly related with voting for the Social Democrats. The profile of the Greens is much more similar in both countries. While the prototypical universalistic identities are strongly linked to the Greens, the more particularistic identities are correlated with reduced support for the Greens.