The determinants of transnational solidarity in the EU

Abstract

This article studies transnational solidarity in the EU from a perspective that has not yet been applied in the literature. It asks to what extent transnational solidarity depends on territorial cleavages (solidarity based on territorial, i.e. on national communities) or on functional cleavages (solidarity based on social groups cutting across nation-states). It studies this question based on the 2018–2023 waves of a YouGov survey covering up to 17 countries per wave, with a particularly detailed analysis of the 2023 wave. The survey includes data on transnational solidarity in various crises. A longitudinal analysis of the influence of territorial and functional characteristics on transnational solidarity is combined with a cross-sectional analysis of their joint influence. The results indicate that functional criteria are more prominent for transnational solidarity than territorial criteria, with the proviso that the functional criteria also include a territorial component.

Keywords:

• Transnational solidarity
• European Union
• territorial cleavages
• functional cleavages

Solidarity is tied to a community or social group whose members are expected to support each other reciprocally (Lahusen and Grasso 2018). It serves as the ‘glue that binds society and prevents it from disintegrating’ (Banting and Kymlicka 2017: 5). In the European Union, the solidarity that binds its member states together is rather thin compared to the solidarity at the level of its member states (Kuhn 2015). However, there is more European solidarity in the EU than meets the eye at first sight, as is shown by Ferrera and Burelli (2019) and Gerhards et al. (2020): Europe’s citizens exhibit a higher degree of solidarity with other member states and their citizens than many social scientists and politicians, especially from the Eurosceptic camp, have presumed so far. Today, Europe constitutes a space of solidarity distinguishable from national and global solidarity (Gerhards et al. 2020; Oana and Truchlewski 2023). However, European solidarity is context-dependent and also depends on individual characteristics linked to social groups. In terms of context, it varies by country (Katsanidou et al. 2022), by region (Reinl et al. 2023), by type of crisis (Cicchi et al. 2020; Kuhn and Kamm 2019), by macro-economic context (Vasilopoulou and Talving 2020), by the specifics of the solidarity instruments proposed by the policymakers and the way they are framed (Beetsma et al. 2022; Ferrara et al. 2023). In terms of individual characteristics, transnational solidarity in Europe is more strongly associated with cultural than with economic orientations, most notably with cosmopolitanism and altruism (Baute et al. 2019; Bechtel et al. 2017; Diez Medrano et al. 2019; Kleider and Stoeckel 2019; Kuhn et al. 2018), although there are regional differences in this respect, as pointed out by Otjes and Katsanidou (2017) and by Hutter and Kriesi (2019), who show that the combination of an economic and political crisis during the Great Recession made the economic dimension much more conspicuous in Southern Europe.

More specifically, we study ‘transnational solidarity’ in the EU, i.e. the solidarity between citizens from one country (the donor country) with another EU country (Reinl 2022: 1376), from a perspective that has so far not been applied in the respective literature. We ask to what extent this kind of solidarity depends on territorial cleavages (solidarity based on territorial, i.e. on national communities) or on functional cleavages (solidarity based on social groups cutting across nation-states). In the EU, territorial cleavages refer to conflicts between member states. In contrast, functional cleavages refer to conflicts constituted by ideological oppositions between social groups over distributive issues that cut across member states.

According to Caramani (2004), we have assisted in the deterritorialization of cleavages during the formation of nation-states. This deterritorialization of cleavages has occurred due to the reduction of ‘exit’ options through national boundary building and the concurrent rise of voice within the nation-states (i.e. of the politicisation and the channelling of competing demands through representative institutions). The consequent development of voice within the containers of the nation-states implied the development of institutional channels of representation (the electoral, corporate, and territorial channels) and the opposition among individuals and groups along specific functional cleavages (Caramani 2004: 17). For the formation of the nation-states, external boundary-building, and internal political structuring were two sides of the same coin.

As has been pointed out by Ferrera et al. (2024), the construction of the EU polity took place under the least favourable circumstances: mass democracy and the welfare state had greatly enhanced the bonds among their citizens, and those between them and their elected territorial authorities at the level of the nation-states. Nation-states are the political units that have achieved the highest degree of boundary closure, with territorial boundaries resulting from state formation and membership boundaries from nation-building. Against this backdrop, the fundamental question thus becomes how the integration process can overcome the resistance against political structuring and ‘system building’ at the EU level (Bartolini 2005: 386) – resistance directed against EU centre-formation, cultural standardisation across member states, and against the building of enlarged and possibly shallower identities and social solidarity across member states. With the end of the ‘permissive consensus’ on European integration, this kind of resistance has been increasingly politicised, as post-functionalists have pointed out (Hooghe and Marks 2019, 2018). But, importantly, it has been politicised at the domestic level of the member states.

As mentioned, we would like to take a closer look at a specific form of solidaristic attitudes among EU citizens – transnational solidarity of citizens. The solidarity of citizens with other member states is crucial for the ongoing processes of EU polity-building. Such solidarity is necessary for common policymaking at the EU level, particularly for policies that have redistributive implications. In studying transnational solidarity, we focus on the extent to which these attitudes are based on territorial or functional cleavages. Deep territorial divisions, i.e. between member states, would imply more limited possibilities for enacting solidaristic policies at the EU level.

We approach this question from the perspective of Caramani (2023), who has argued that it is possible to conceive of politicised and mass-level integration through the progressive replacement of territorial cleavages with ‘functional’ alignments that cut across territoriality. Caramani defines ‘politicized integration’ as an opposition between social groups rather than between world regions or European member states. Politicised integration implies that citizens and actors of the same side of a cleavage ally across territorial units, thus providing a social base and solidarity bonds for their integration. Traditionally, the transformation of territorial into functional cleavages occurred through the replacement of religious, ethnic, or cultural divisions, i.e. preindustrial divisions with strong territoriality, with divisions based on class. Politicised integration was brought about by industrialisation, modernisation, and class, replacing culture as the basis for political divisions.

Our question is to what extent transnational solidarity among citizens of the EU is based on territorial and functional cleavages. First, we expect territorial cleavages between member states given the strength of the territorial channel of representation within the EU and the importance of intergovernmental coordination as the key decision mode that gives a pivotal role to member state governments. Conflicts and coalitions between member states, which shape cues from national governments and the nature and spatial distribution of crises, are expected to shape transnational solidarity among the respective national publics. Second, we expect transnational solidarity to be shaped by two main functionalist cleavages: a traditional economic or class cleavage, opposing state interventionist positions to pro-market positions, and a new cultural cleavage, opposing nationalists to cosmopolitans. When comparing the role of these cleavages, we argue that to the extent that transnational solidarity is based on territorial cleavages, it will be more precarious for the integration process; to the extent that it is based on functional cleavages, it will imply an integrative potential that suitable political actors could mobilise. Finally, when comparing the two, we also note that while functional conflicts are constituted by ideological oppositions between social groups over distributive issues and have the potential to cut across territorial lines, they may also be territorialised to different extents. That is to say that particular camps of a functional cleavage might dominate in a given member state or that different functional cleavages might be salient to varying degrees in the member states. In our case, a strong territorialisation of functional cleavages lowers their integrative potential.

The article is structured as follows: first, we elaborate on our theoretical perspective. Next, we present the design of our empirical analysis, followed by the presentation of the results. Finally, we conclude by summarising and discussing our results.

Theory: territorial and functional cleavages

As Caramani (2023) points out, integration and politicisation are not incompatible, as is documented by the formation of nation-states: nation-building, i.e. the creation of identities and bonds of solidarity, and democratisation through institutions enhancing mass participation and the politicisation of cleavages went hand in hand. In the EU, conflicts between member states give rise to the quintessential territorial cleavages, while functional conflicts are transnational and cut across territorial lines. Three mechanisms create similarity, convergence, and simultaneous change in the different member states, leading to cross-territorial functional cleavages: (1) a common cause that has similar effects everywhere (e.g. the adoption of a legislative act at the EU level, like the adoption of the Maastricht Treaty, or the creation of transnational party structures); (2) parallel changes in each member state lead to convergence (e.g. technological change or the post-materialist cultural revolution); (3) diffusion (e.g. based on joint membership in the EU and contacts between leaders in successive summits). Replacing territorial with functional cleavages means territorial homogenisation and convergence between parts of the territory undergoing political structuration (Caramani 2015: 30).

In his study of the Europeanization of national party systems to determine the extent of the homogenisation and convergence of these systems across countries, Caramani takes a macro-historical perspective. He finds much convergence and uniformity in developing cleavages across member states from 1848 to 2012 and a close correspondence between EP and national elections from 1974 to 2012. To explain this development, he invokes common factors accounting for uniform swings (e.g. the outcome of WWII, the post-industrial revolution, the outcome of the Cold War, the Maastricht Treaty and globalization), social transformations such as the national revolution, the industrial revolution, urbanisation, or the post-industrial revolution which shaped the national party systems in analogous ways, and events such as the Russian revolution or the fascist reaction to this revolution which account for transnational diffusion effects. His study of the nationalisation at the Europe-wide level finds a sort of ‘Europe-wide nationalization’.

At the end of his analysis, however, Caramani (2015: 291) suggests that the comparison between Europeanization and processes that took place at the level of nation-states should not be taken too far. As defined in his study, Europeanization misses one of the central elements of electoral integration at the national level: a sovereign centre carrying out policies aimed at state formation, administrative-economic integration, and nation-building based on the legitimate use of force. As a result, the integration processes at the mass electorate level took place in the distinct units but simultaneously in units belonging to one system. To characterise the result, he speaks of a situation of ‘segmented similarity’ (p. 292): similar party systems exist alongside each other, but accountability and responsiveness are segmented between them: each party system is accountable and responsive to its national electorate. At the same time, however, the member state governments are also accountable to the EU institutions and other member states’ governments. Indirectly, the segmentation is partly overcome via the territorial representation channel of the EU.

As long as this situation prevails, the similar functional transformation of territorial cleavages in the different member states remains territorially coded and is only potentially conducive to integration processes. Ultimately, the transformation of territorial into functional cleavages is a matter of the possibility of democratic politics beyond the nation-state. As long as the electoral channel of representation at the EU level remains as weak as it currently is, the functional transformation of territorial cleavages remains, at best, a latent potential.

The territorial channel of representation and territorial cleavages

For now, the territorial channel is by far the most important channel of representation at the EU level. Intergovernmental coordination has become the key decision mode in the EU in general and in crisis situations in particular. This decision-mode attributes a pivotal role to member state governments and, in particular, to their top representatives as members of the European Council (Puetter 2015). They provide the critical link between the two levels of the EU polity. As a result of their dual role – that of the head of state or government representing a country in European negotiations and that of a member of the European Council representing Europe back home, the executives of the member states become the key actors in the two-level game that links domestic politics to EU-decision-making. The sequence of multiple crises has increased the need for joint problem-solving and further strengthened the decision-making role of the intergovernmental institutions, with the European Council emerging as the crisis manager in chief through ad-hoc summits focusing on pressing political problems (Fabbrini 2017). Indeed, the European Council has successfully imposed its political views on the legislative process traditionally regulated by the Community Method. The downside of this development is, of course, the aggravation of the ‘democratic deficit’ and the consolidation of an institutional framework that is effective in patching up emergencies but at the cost of amplifying inter-state differences based on a contingent and short-term calculus logic. Intergovernmentalism brings the seeds of polity erosion.

Against this background, it is important to keep in mind that the intergovernmental crisis management by the EU during the more recent crises has led to the emergence of several transnational coalitions between member states (see: Fabbrini 2022; Kriesi et al. 2024; Krotz and Schramm 2022; Truchlewski et al. 2023a):

• the German-French couple, taking a leading role on several occasions (most notably in the case of the RRF (Resilience and Recovery Fund) during the Covid crisis),

• the solidarity coalition, consisting essentially of the southern member states + France, which demanded more solidarity, especially during the Covid crisis,

• the Visegrad 4 coalition, composed of Hungary, Poland, Czech Republic, and Slovakia, which became prominent for the first time during the refugee crisis and, later on, consolidated during the rule-of-law debate during the Covid-crisis,

• the Frugal 4, composed of the Netherlands, Austria, Denmark, Sweden, and, at times, Finland, which made itself known by its opposition to a generous RRF solution in the Covid crisis,

• the UK, which eventually seceded from the EU, but had always been rather sceptical of EU solidarity.

While the southern solidarity coalition was demanding more solidarity among member states, the Frugal 4 opposed such policies during the RRF negotiations in the Covid crisis, and the V4 coalition, in particular, blocked solidaristic policies in the refugee crisis and attempted to block them during the Covid crisis as a result of its opposition to the rule-of-law requirements. During the Ukraine war, however, Hungary’s pursuit of its pro-Russian policy antagonised the other members of the V4 coalition and weakened its joint resolve. The German-French couple attempted to broker compromises during the crisis management, i.e. typically took intermediate positions.

Given the limited availability of information about the EU and the high complexity of the issues debated at the EU level, parties can shape citizens’ attitudes towards the EU (Pannico 2020). However, the importance of intergovernmental decision-making at the EU level, as well as the importance of transnational coalitions, suggests that we need to focus on cues not only or not even in the first place from parties, but in particular from national government representatives who are the pivotal actors in EU policymaking. Several studies suggest that the national governments provide important cues about EU policymaking to the domestic public. Thus, Ares et al. (2017) have shown that spill-over effects from the domestic to the EU-polity support rely on cues from national governments, especially in critical moments, such as the Eurozone crisis, when the role of the national government in the integration process becomes exceptionally salient. Hobolt et al. (2013) provide experimental evidence to show that people are likely to take their cues from the national government rather than from EU officials. Dellmuth and Tallberg (2021) provide experimental evidence that communication by national governments (and civil society organisations) has stronger effects on legitimacy perceptions than communication by international organisations in general.

Given the coalitional line-up of the member states during the crisis management, we expect the attitudes of the national public in the different member states to align with their governments’ coalition membership:

H1: the membership of national governments in transnational coalitions shapes the transnational solidarity among the respective national publics in line with the coalitions’ positions.

More specifically, we expect that voters who support or trust the national government are more likely to follow its cues than voters who do not support or trust it, as people are more likely to accept information from a trustworthy source and to reject information from a non-trustworthy source (Druckman 2001a, 2001b). Whether or not the information received from the government enhances or weakens transnational solidarity depends, of course, on the position of the respective government. In the past, member state governments tended to be pro-European, meaning that individual support or trust in the national government contributed to trust in the EU and transnational solidarity. However, suppose the national government is opposed to EU policies and the EU polity as it stands more generally, as was the case of the Hungarian government during the refugee crisis and the Ukraine war or of the British government under Boris Johnson driving a hard-Brexit line during the Brexit negotiations with the EU. In that case, we expect that individual support or trust in the national government tends to undermine support for the EU and solidarity with other member states:

H2a: trust in the national government reinforces transnational solidarity as long as the national government takes a pro-EU position.

H2b: trust in the national government undermines transnational solidarity if the national government takes an anti-EU position.

Based on the data we shall use in our study, Cicchi et al. (2020: 6) have shown that transnational solidarity varies by crisis. Generally, solidarity is expected to be stronger in crises with exogenous origins (e.g. natural disasters, military attacks, or pandemics) than in crises with endogenous origins (e.g. the Eurozone crisis). If the crisis origin is endogenous, moral hazard narratives get some traction, which reduces solidarity. Moreover, solidarity is expected to be stronger in symmetric than asymmetric crises. In asymmetric crises, like the Eurozone crisis, citizens of the hard-hit countries (who are at the receiving end of solidarity) are likely to be more supportive of transnational solidarity than citizens of countries that are less hit by the crisis (and who are at the contributing end). In such cases, moral hazard considerations again play a role (on the part of the contributors). By contrast, in symmetric crises, like the Covid pandemic, transnational solidarity can be expected to be enhanced by identity-based empathy and interdependence-based sympathy (i.e. ‘rational compassion’, see Genschel and Jachtenfuchs 2021: 9). Alternatively, if transnational solidarity is less driven by motivational than by capacity-related considerations, we might expect less solidarity in symmetric crises: in such crises, like in the Covid pandemic, where all countries are hit similarly, the citizens in each country are likely to consider themselves as victims and to put the priority on overcoming the crisis at home. In each country, the limited resources should first be used to come to terms with one’s own crisis before other countries can be supported (whether for the same or a putative other crisis). This leaves us with two competing expectations:

H3a: empathy effects hypothesis: in an exogenous and/or symmetric crisis, like the Covid pandemic, transnational solidarity is stronger than in an endogenous and/or asymmetric crisis.

H3b: resource effects hypothesis: in a symmetric crisis, like the Covid pandemic, transnational solidarity is weaker than in an asymmetric crisis.

Kyriazi et al. (2023) compared pre-Covid (2019) transnational fiscal and economic solidarity with such solidarity during the first wave of the pandemic (June 2020) in six countries – France, Germany, Italy, Spain, Sweden, and the Netherlands. They found that levels of support did not substantially differ in the two periods compared, suggesting that both mechanisms might have been at work.¹

The two main functionalist cleavages

Meanwhile, at the national level, we assist in the return of a cultural cleavage, i.e. of a type of cleavage that Caramani qualified as ‘premodern’: a new cleavage has been taking shape that is explicitly related to the EU integration process and the opening up of the nation-state in economic, cultural and political terms more generally. This cleavage opposes nationalists and cosmopolitans and has come to replace the traditionally dominant religious cleavage in cultural terms. Various authors have used different names to refer to it: ‘GAL-TAN’ (Hooghe et al. 2002), ‘independence-integration’ (Bartolini 2005), ‘integration-demarcation’ (Kriesi et al. 2008), ‘universalism-communitarianism’ (Bornschier 2010), ‘cosmopolitanism-communitarianism’ (Zürn and de Wilde 2016), ‘cosmopolitanism-parochialism’ (De Vries 2017), the ‘transnational cleavage’ (Hooghe and Marks 2018) and the cleavage between the principle of sovereignty and Europeanism (Fabbrini 2019: 62f.). The drivers of this new cleavage have been the radical right and national-conservative political forces which defend the position of those social groups who are opposed to the opening up of the national community to strangers from the outside and are equally opposed to the European integration process and the concomitant loss of national sovereignty and economic independence. On the cosmopolitan side of this new cleavage, the Green parties have been the main protagonists, defending not only ecologist points of view but also universal human rights, peace, and international solidarity. As a result of the rise of this new cleavage, politics in European countries has become increasingly structured into a two-dimensional space, created by the new cultural cleavage, on the one hand, and the traditional economic or class cleavage, driven by the centre-left and centre-right mainstream parties and opposing state interventionist positions to pro-market positions, on the other hand.

These two functionalist cleavages have arguably become established to a greater or lesser extent in all European countries. As Häusermann and Kriesi (2015) showed, despite stark differences in the average levels of voter preferences regarding economic and cultural conflict, their configuration and socio-structural determinants are very similar across Europe, with education as the main determinant of cultural preferences and economic preferences being structured by income and a new class conflict dividing the middle class.

The new cultural cleavage is particularly important for solidarity across EU member states. The proponents of the nationalist position oppose system-building at the EU level. They are expected to oppose policies aimed at redistributing resources across member states. This expectation is supported by previous studies of transnational solidarity in Europe (Baute et al. 2019; Bechtel et al. 2014; Diez Medrano et al. 2019; Kleider and Stoeckel 2019; Kuhn et al. 2018). In addition, we expect those who are opposed to redistributive measures at the domestic level to oppose redistributive measures across member states as well. However, we expect the effect of nationalist opposition against solidaristic measures across member states to be even stronger than the effect of opposition to redistributive measures at the domestic level, given that redistribution across member states is not in the first place related to class but to the defense of the national community against the demands from other such communities. To summarise:

H4a: the new cultural cleavage, indicated by individual positions related to immigration and EU integration, has a powerful effect on transnational solidarity: cosmopolitans favor it, and nationalists oppose it.

H4b: the class cleavage, indicated by individual positions about redistribution within the national community and self-placement on the left-right scale, also impacts transnational solidarity, but to a lesser extent: individuals with left-wing and pro-redistribution attitudes favor it, and vice versa for individuals with right-wing attitudes and who oppose redistribution.

The territoriality of functional cleavages

Caramani (2009: 29) describes the territoriality of political cleavages as the ‘degree to which (linguistic, religious, economic) groups of individuals are opposed along territorial lines’. His prime example is the centre-periphery cleavage. This is sharply contrasted with more functional cleavages, such as the class cleavage. We argue that beyond the continuum from purely territorial to purely functional cleavages, it is possible to observe a territorialisation of functional cleavages. This can occur when functional cleavages overlap with territory. Imagine a simple case of a polity with two territories. One territory is dominated by conservative individuals, while liberals dominate the other. The main cleavage in the polity is left-right, that is, functional. However, in this example, the functional cleavage completely overlaps with territory. Although nominally functional, we argue that such a situation has implications similar to a territorial cleavage. The higher the degree of territorialisation of a functional cleavage, the more difficult it becomes to integrate the polity.

In our case, functional cleavages might be territorialised if, in particular member states, different functional cleavages are salient to different degrees. There is some reason to expect this. We know, for example, that migration is not equally salient among the populations of member states, partly based on their experience of the 2015 refugee crisis (Kriesi et al. 2024). Importantly for our analysis, the degree to which the class or the new cultural cleavage determine solidary attitudes may vary by country. We do not form firm hypotheses on this issue but rather explore the territorial dimension of functional cleavages by looking at heterogenous country effects.

Design

As Reinl (2022) has pointed out, transnational solidarity is a multidimensional concept with at least two underlying dimensions: risk-sharing provided for fellow EU member states in crisis and redistributive policies striving for more strongly embedded solidarity in the overall EU integration process. We study transnational solidarity in terms of risk-sharing in crisis situations based on a set of surveys conducted by YouGov. This survey has been conducted yearly since 2018 for many countries. The dataset covers 11 countries for six years, 13 for four years, and 17 for the last two years. All survey waves include data on our dependent variable – solidarity in various crisis situations. However, data on the independent variables included in the yearly surveys vary from one survey to the other. In particular, adequate measures for the two dimensions of the political space have only been included in the 2023 survey. We combine a longitudinal analysis of the influence of individual political attitudes and the territorial and crisis context on solidarity with a cross-sectional analysis of the joint influence of the two types of determinants.

The dependent variable is a solidarity index, composed of attitudes about five different crises in another EU country – natural disasters, such as a major earthquake or catastrophic flooding, a major debt crisis, a major refugee crisis, a major unemployment crisis, and a military attack by a country outside of the EU. The respondents were asked whether their country should or should not provide major help to an EU country suffering such a crisis, with dichotomous response categories. The responses to the five crises form a strong factor (EV = 3.0), which will be used as the main dependent variable in the analyses.² Unfortunately, in the 2018 and 2019 surveys, the formulation of these questions varied somewhat: the respondents were not asked whether ‘your country’ should or should not provide help but whether ‘EU countries’ should do so. This alternative stimulus will likely elicit more solidarity than the later, more specific formulation. And it, indeed, does so. Therefore, we can only use the last four years for most of our longitudinal analyses. We shall use all six years only for the test of H2a/b, the country-specific effect of government trust, which focuses on the interaction effects between trust and country-year.

When asked explicitly how the respondents’ own country should provide help, they prefer joint help. Thus, the 2023 survey asked an additional question about how the respondents’ own country, once it had decided to help another EU member state in a crisis, should provide the help: should it act on its own or as part of a joint effort managed by the European Union? In response to this question, 70.1% prefer joint effort over bilateral help (preferred by 15.8%), while 14.1% do not know how to answer.³ However, we do not know whether the respondents considered bilateral or joint EU solidarity in responding to our solidarity questions. Moreover, to relate our results to the results of other studies, it is important to remember that we are asking about transnational solidarity, not EU support or EU identity.

We use individual country dummies for the territorial effects or dummies for the country coalitions (German-French couple, southern Europe, Frugal 4, V4, other central and eastern European, UK). For the analyses based on the longitudinal file, we distinguish between three periods, which correspond to three different crisis situations: the pre-crisis period (2018–2019), the Covid crisis period (2020–2021), and the period of the Ukraine war (2022–2023).

For the functional effects, we use five variables and the usual socio-demographic controls (age, gender, education, and an assessment of one’s economic situation). As mentioned, the two dimensions of the political space can only be fully operationalised for the 2023 survey. In this case, we operationalise immigration-related attitudes with a battery of four items (inspired by the ESS items), which form a strong factor (EV = 2.4): To what extent do you think your country should allow the following types of people to come and live here: (1) people from the same ethnic group as most people in your country, (2) people from a different ethnic group from most people in your country, (3) people from poorer countries outside Europe, (4) people from poorer countries inside Europe. We operationalise European integration-related attitudes with two classic indicators: the classic territorial identity question (which we recoded into three categories – one has an exclusively national identity, one has a mixed (national-EU) identity/an exclusively EU identity, or one has none of these identities/does not know); and a question about trust in the EU. A single question operationalises the redistributive attitude: Do you think the government should redistribute income from those who are better off to those who are worse off? The responses to this question were recorded on an 11-point scale. In addition, we have the classic self-placement on a 7-point left-right scale (with don’t know recoded to the mid-point). Finally, we can rely on two indicators for trust: trust in the national government and trust in the EU, measured on a five-point scale.

For the longitudinal file, we only have corresponding items for left-right orientation and the two trust variables. However, they are not identical regarding trust in the national government.⁴ In addition, we use the question about past vote. We recode this information into party families which are comparable across the 17 countries included in the survey: radical left, greens, social-democrats, liberals (social and conservative ones), conservatives (including the Christian democrats and national-conservatives such as PiS and Fidesz), radical right, others (e.g. regional parties) and non-voters. We recoded the non-dummy variables to the 0–1 range to render the effects across independent variables comparable.

Results

We first present the territorial effects and then the effects of the individual-level attitudes, i.e. the functional effects.

Territorial effects: differences between member states and crises in terms of solidarity

Figure 1 presents the predicted margins for the country effects on the solidarity index, controlling for individual-level characteristics, based on the 2023 data.⁵ The horizontal dashed line in the graph represents the sample average. The countries are ordered according to their coalition affiliations. Southern European countries stand out as most favourable to solidarity across member states, aligning with expectations (H1). In contrast, also aligning with expectations, the Finns and the Dutch, two Frugal coalition members, are highly critical of EU solidarity, followed by the Lithuanians. However, note that by 2023, the Danes and the Swedes, the other two members of the Frugal coalition, are aligned with the European average. The citizens of one of the three V4 countries, which are part of the sample – Hungary, are rather critical of transnational solidarity, again in line with expectations. In addition, we find that the transnational solidarity of German and French citizens is below average, too, a result that was not necessarily expected. Croatians and Romanians, by contrast, are rather more solidaristic than average.

Figure 1. Solidarity index by country: 2023, predicted margins.

We have hypothesised that trust in the national government increases solidarity with other EU countries, except if the national government explicitly takes anti-EU positions (H2a-b). Figure 2 presents the effects of trust in government on the solidarity index, controlling for past vote, for each country during the three periods we distinguish for testing this particular hypothesis – the ‘pre-crisis’ (2018–2019), ‘Covid-crisis’ (2020–21) and ‘Ukraine war’ period (2022–23).⁶ The results support the hypotheses. In general, trust in the national government enhances, as expected by H2a, EU solidarity. Suppose one trusts the most important representative of national interests at the table of EU decision-making, i.e. the national government. In that case, one is ready to be more solidaristic with the EU member states, independently of other relevant factors. However, there are three exceptions to this general result that are in line with H2b. The first exception concerns Hungary and Poland, two V4 countries. In Hungary, the government’s consistent anti-EU policymaking reverses the effect of government trust in the two later periods (no data for the pre-crisis period in this case). In Poland, the effect of government trust was also negative or inexistent in the two earlier periods but became positive during the Ukraine war. This development of the effect of government trust is in line with the more accommodating EU-positions adopted by the Polish government under the impact of the Ukraine war. The second exception concerns the UK under Boris Johnson’s Conservative government. As the Johnson government steered the UK into a hard Brexit, trust in this government ended up undermining solidarity with EU countries suffering a major crisis. The third exception refers to the populist governments in southern Europe. Thus, in the pre-crisis period, trust in the Tsipras government (2015-2019) in Greece reversed the effect of government trust on transnational solidarity. During the same period, trust in the Conte I government in Italy (June 2018-September 2019), a coalition of the populist radical right (Lega) and the populist left (M5S) also considerably reduced the effect of trust in government on solidarity. Finally, note that under the new right-wing government in Sweden (October 2022-) – a minority government supported by the radical right, the effect of government trust on solidarity has also been considerably reduced, even if it remains positive. These results support H2a-b and the importance of government cueing for transnational solidarity.

Figure 2. Solidarity effect of trust in government by country and crisis period: average marginal effects.

Next, we turn to the variation of transnational solidarity by crisis. Table 1 presents the period effects for different models: model 1 includes just the period effects, model 2 adds the coalition dummies, model 3 the individual trust variables, plus socio-demographics, model 4 also includes the past vote, and model 5 the interactions between government trust and coalition dummies. With the introduction of the coalition dummies, the effects shown correspond to the effects for the reference category – the French-German couple, which has roughly average period effects. As it turns out, the differences between the two crisis periods are quite limited. Still, the COVID crisis is associated with somewhat less solidarity, contrary to the sympathy/empathy hypothesis H3a but in line with the resource hypothesis H3b. Its negative effect is partially suppressed by the individual-level characteristics, i.e. it increases once we control for the latter.

Table 1. Period effects, by model: OLS-effects and t-values^a.

	Model 1	Model 2	Model 3	Model 4	Model 5
	b/t	b/t	b/t	b/t	b/t
Covid crisis, ref
Ukraine war	0.020***	0.020***	0.037***	0.037***	0.038***
	(7.607)	(7.836)	(3.559)	(3.541)	(3.670)
Controlling for	–	Territorial	Territorial + trust	Territorial + trust + past vote	Territorial + trust + past vote + interaction territorial-trust

^a Slovakia, Croatia, and Bulgaria are excluded from this analysis because they were only part of the 2022 and 2023 samples.

Figure 3 allows us to assess the differences between the crisis periods concerning solidarity by country, controlling for individual characteristics.⁷ We mainly observe stability in individual countries, except that in the Frugal 4 coalition, solidarity increases somewhat in the Ukraine war situation. These are the countries that were highly reluctant to contribute to the Resilience and Recovery Fund during the Covid crisis. This seems to support the resource hypothesis. However, the differences between the crisis situations are limited even in their case.

Figure 3. Country effects by crisis period: predictive margins.

The level of solidarity varies from natural disasters (76% solidarity) via military attacks (60% solidarity) and refugees (50%) to unemployment (37%) and public debt (33%), which supports the empathy hypothesis H3a. The period differences also vary by crisis, which clarifies the country differences we have just observed in Figure 3. Figure 4 presents the development of crisis-specific solidarity for the various countries. First, solidarity in the case of natural disasters remains very strong across countries from one crisis to the other. It is mostly stable but increases somewhat among the Frugals and decreases somewhat in Romania. Second, solidarity concerning debt and unemployment crises is also rather stable; the Frugals are the least solidaristic when it comes to such crises, while the southern Europeans are more solidaristic than the others. Changes from one crisis period to another are very limited in this respect, undermining the resource hypothesis. Thus, regarding economic crises, the Frugals refuse to be solidary during a symmetric crisis (as the Covid crisis) and more generally at any moment, which neither supports the empathy nor the resource hypothesis. By contrast, solidarity in the case of a major refugee crisis has generally increased during the Ukraine war: it did so most clearly in Lithuania, the three Nordic countries, and the UK, all countries where it was very low before the war. However, this increase is hardly related to resource considerations. It can rather be explained by the inflow of a new type of refugees into Europe due to the Russian attack on Ukraine and by the much more accommodating policy of the EU that applied to them (Moise et al. 2024).

Figure 4. Coalition and country effects by crisis period and type of solidarity: predictive margins.

Finally, we observe very interesting country differences in the development of solidarity in the case of a major military attack, none of which, however, support H3a or H3b. In most countries, the level of solidarity has hardly changed over the past years. However, there are three countries (Hungary, Romania, and Greece) where the readiness to be solidary with another country that suffers from a military attack has decreased. In comparison, this readiness has increased in three other countries (Finland, Sweden, and the UK) – a contrast that is in line with our cueing hypotheses. In the case of Hungary, the Orbán government has adopted a pro-Russian policy, which is likely to account for its public’s clear-cut deviation from the overall European development. Romania’s government is unambiguously in line with the EU and NATO but avoids the strident tone of Poland or the Baltic states. One reason may be its closeness to Moldova, the region’s most vulnerable country in terms of Russian aggression. Interestingly, the Romanian media ‘comprehensively supports the “smokescreen” strategy about military deliveries and other support of Ukraine’. Moreover, some popular narratives in Romania may explain the lack of public support for military assistance to Ukraine; thus, it is widely accepted in Romania that Russia is an impossible opponent to defeat and that it is better to avoid confrontation. Peace between the warring parties should be achieved quickly; otherwise, the war would threaten Romania’s interests.⁸ The Greek government has also clearly come out in favour of Ukraine. However, the Greek left-wing opposition has taken a more ambiguous position,⁹ and, more recently, the Greek government clashed with Ukraine over a list of ‘International war sponsors’ published by Ukraine, which included some Greek private companies.¹⁰

In the case of Finland and Sweden, by contrast, both governments have reacted by taking sides against the aggressor: they both applied to become members of NATO. Next to the US, the UK has become one of Ukraine’s greatest military donors, and it is also hosting a training program supported by several allies. These governments’ pro-Ukrainian positions are reflected in the increasingly solidaristic attitudes of the respective publics.

Functional effects: effects of individual political attitudes

For our research question, it is crucial to compare the territorial effects referring to differences between countries with the functional effects referring to differences between individuals. As can be seen from Figure 5, which is based exclusively on the 2023 data, the individual functional effects tend to be much larger than the country effects.¹¹ The largest country effects, the effects for the southern coalition countries, are roughly equivalent to the weaker individual-level effects – the effects related to redistributive attitudes at the national level, national vs. EU-related identities (with people with both national and mixed identities being more solidary than people with no explicit identity), and left-right self-placement (with citizens from the left being more solidary than citizens on the right).¹² Trust in government has an even weaker effect. The strongest effects by far are exerted by immigration attitudes and trust in the EU, i.e. by the indicators of the new cultural cleavage. As we discussed in the theory section, immigration attitudes and trust in the EU also have a territorial aspect, which means that they have mixed effects on solidarity in terms of the distinction between territorial and functional characteristics. However, from the point of view of Caramani’s theory, these individual-level characteristics must deploy their effects, on average, across all countries. Territorial and functional characteristics jointly explain roughly one-third (30%) of the variance in EU solidarity at the individual level.

Figure 5. Attitudinal effects vs country effects, 2023: average marginal effects.

We have further pursued the variation of the impact of individual character characteristics from one country to another. Figure 6 presents the corresponding results for immigration and redistributive attitudes.¹³ The dashed horizontal lines in the two parts of the figure indicate the average effects and the solid lines indicate the null effects. On the left-hand side, we find the variation of the effect of immigration attitudes across coalitions. Most importantly, the effect is very strong in all coalitions. However, it also varies significantly between coalitions, strongest in the southern member states and weakest in the central-eastern European member states. A more detailed analysis by country reveals that the effect is particularly strong in Italy and particularly weak in Bulgaria, Lithuania, and Romania. In all the other coalitions, it is situated close to the average. On the right-hand side of the figure, we present the variation of the effect of redistributive attitudes. It is far weaker than the effect of immigration attitudes in all coalitions. It is significantly above average in the UK and below average in the V4 countries. Despite significant variation across coalitions, the overall similarity of the effects for a specific attitude suggests a large potential for politicising the new cultural and the conventional left-right cleavage across EU member states. In other words, in principle, the ground for a more integrated politicisation of the cleavages in the EU is prepared. What is missing are the respective channels of mobilisation.

Figure 6. Attitudinal effects by coalition, 2023: average marginal effects.

Figure 7 presents the variation of attitudinal and country effects across the Covid and the Ukraine crises, based on the longitudinal data, controlling for past vote. We already know from Figure 2 that the country effects hardly vary from one crisis to the other. Figure 7 now shows that the same applies to the indicators of the functional cleavages – trust in the EU (for the new cultural cleavage) and left-right orientation (for the class cleavage), even if the effect of left-right orientation becomes somewhat less pronounced in the Ukraine crisis. The reduced effect of left-right orientation means that the people on the right have become relatively more solidaristic during the Ukraine crisis. This result aligns with the right’s comparatively greater solidarity with immigrants from Ukraine, which Moise et al. (2024) have observed before. Additional analyses show that, independently of the overall left-right orientation, the radical right is less solidaristic than the party families of the moderate right in both periods. The same applies to the non-voters. Importantly, the proxy indicator for the new cultural cleavage, trust in the EU, has by far the strongest effect on transnational solidarity in both crises.

Figure 7. Attitudinal effects vs country effects, by crisis: AMEs.

In sum, the results concerning individual attitudes confirm the key importance of the new cultural cleavage for transnational solidarity – cosmopolitans are much more supportive of such solidarity than nationalists (H4a). The traditional class cleavage is also of some importance, with the left being more solidary than the right, but to a lesser extent than the new cultural cleavage (H4b).

Conclusion

We have analysed the impact of territorial and functional cleavages on transnational solidarity in the EU. We have found support for the hypothesis that transnational coalitions between member states influence the transnational solidarity among the citizens of the corresponding member states. We have also found that the governmental composition moderates the effect of governmental trust on solidarity. Both of these results point to the impact of cueing by member state governments – the key actors in the territorial channel of representation in the EU. We found mixed support for our two competing hypotheses about the crisis-specific impact on transnational solidarity: overall, solidarity varies by crisis in line with the empathy hypothesis. However, some drivers of country differences either refer to highly specific aspects of the problem pressure during a given crisis (e.g. the type of refugees flowing into the EU) or they are linked to country-specific policies in some member states (e.g. the EU’s refugee policy, the pro-Russian policy of Hungary, and the pro-Ukrainian policies of the UK, Finland and Sweden during the Ukraine war), which are also providing specific cues for the national publics. The resource hypothesis did not get any support: the greater overall solidarity of the Frugals during the Ukraine crisis is unrelated to their solidarity in economic terms and entirely associated with their greater refugee solidarity during this crisis.

Overall, we found individual characteristics to be more important for transnational solidarity, especially the attitudes related to the new cultural cleavage, which is in line with the results of several previous studies to which we referred. However, even if individual-level characteristics prevail and their impact is similar across countries, this does not necessarily mean that functional criteria predominate as determinants of transnational solidarity. The key individual determinants associated with the new cultural cleavage – immigration attitudes and trust in the EU – have a territorial component. One could argue that the new cultural cleavage resembles the centre-periphery cleavage and reproduces it at the supranational level. Or, in other words, like the premodern cleavages, the new cultural cleavage has clear territorial roots.

Moreover, individual-level attitudes are consequential for the most important channel of representation in the EU’s current decision-making procedures, the territorial channel of representation. This is a result of the fact that the key political attitudes co-determine the composition of the government in a given member state and that the cueing by the national governments, in turn, influences the transnational solidarity of the attentive public in the member states. This means that although the individual-level attitudes have similar and strong effects on transnational solidarity across member states, this similarity remains segmented. In other words, the new cultural cleavage may have effects that differ territorially from one member state to the other depending, among other things, on the composition of the national government and its policies, on the incidence of the crises on the national population, and the distribution of the relevant individual attitudes in the various member states. As long as the functional cleavages do not find adequate expression at the EU level, their homogenising effect remains hypothetical, and EU politics remains dominated by territorial considerations, which undercut the transnational solidarity in the system.

Acknowledgements

We want to thank our colleagues from the SOLID team, Ann-Kathrin Reinl, and Daniele Caramani for their helpful comments on earlier versions of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

Our research was supported by the European Research Council under the Synergy Grant number ERC_SYG_2018, grant number 810356, for the project Policy Crisis and Crisis Politics: Sovereignty, Solidarity, and Identity in the EU post-2008 – SOLID.

Notes on contributors

Hanspeter Kriesi

Hanspeter Kriesi is a Part-time Professor at the European University Institute, Florence, Italy, where he was the Stein Rokkan Chair holder at the Department of Political and Social Sciences from 2012 to 2020. He is one of the three Principal investigators of the SOLID-ERC project [[email protected]]

Alexandru D. Moise

Alexandru D. Moise is a Research Fellow at the European University Institute, in the SOLID-ERC project. His research interests include political behaviour, European politics, health policy, and populism. He has published in the British Journal of Political Science, the Journal of European Public Policy, West European Politics, and the Journal of Health Politics Policy and Law, among others [[email protected]]

Ioana-Elena Oana

Ioana-Elena Oana is a Research Fellow at the European University Institute, in the SOLID-ERC project. Her research agenda focuses on political behaviour, public opinion formation, and European integration and solidarity. Her work has been published in the European Journal of Political Research, West European Politics, and Journal of European Public Policy, among others. She has also co-authored two monographs with Cambridge University Press [[email protected]]

Notes

1 Other studies document increasing EU support during the Covid crisis – as a result of a rally-around-the-flag effect (Truchlewski et al. 2023b) or greater identification with the EU (Truchlewski et al. 2023b; Chapter 11). However, these results on EU support do not allow us to draw any conclusions ­regarding the outcome of interest here – transnational solidarity.

2 Given the dichotomous response categories of the five variables, we performed a polychoric factor analysis. The factor loadings of the crisis indicators vary from .65 (military attacks) to .82 (debt and unemployment).

3 The results of Oana and Truchlewski (2023) align with this finding. They found that bilateral solidarity received less support than EU solidarity.

4 In 2022 and 2023, trust in the national government and the EU was measured on a five-point scale. In previous years, the same applies to trust in the EU. Trust in the national government, however, was operationalized with a set of items asking about trust in the national government to make things better concerning the economy, climate, military defense, protection from terrorism, protection from crime, food standards, employment opportunity, one’s financial situation, health provisions, and immigration. We have created a single factor based on these ten items for these years.

5 For details, see column 2, Table A1 in the Online Appendix.

6 For details, see column 2, Table A2 in the Online Appendix.

7 For details, see column 3, Table A2 in the Online Appendix.

8 Ibid.

9 https://www.euractiv.com/section/all/short_news/greeces-tsipras-blasts-pms-ukraine-policy-favours-diplomacy/.

10 https://www.politico.eu/article/new-sanctions-against-russia-stuck-limbo-greece-hungary-protest-ukraine-war/.

11 For details, see column 3, Table A1 in the Online Appendix.

12 These country effects are maximum residual effects that result from controlling the individual-level effects we include in our model (Germany is the reference category). The country effects might be weaker if we were able to control for additional, relevant individual-level characteristics, which we have, however, not measured (such as threat perceptions or satisfaction with government performance).

13 For details, see Table A4 in the Online Appendix.