Which of the following would support a Primordialist claim related to Christianity and democracy?

1 Introduction*

Political theory has held national identity as important to democracy, primarily because democratic beliefs need a source of community. When a person thinks of herself as belonging to a particular kind of people, the national identity can help develop and strengthen the bonds of community (Calhoun 2007; Canovan 1996; Mill 1861). On the other hand, too intense and restrictive national identity can potentially undermine support for democratic ideals (Kymlicka 2001, 258). This long-standing theoretical debate, whether national identity and democracy are complemental or competing has recently been complemented by empirical evidence. On the individual level, Erhardt Wamsler and Freitag (2020) conclude that the normative dimension of national identity cuts both ways in regard to support for democracy. Support for self-governance is implicit in voluntary civic notions of national identity (Habermas 1992; Verba 1965), and is positively correlated with democratic beliefs. The non-voluntary national identity, where national membership depends on the accident of origin, is related to lower support for the idea of democracy. On the country level, Gabrielsson (2021) shows that the concerns for homogeneous non-voluntary identity, classically understood as nationalism, correlates with lower levels of actual democracy. Building on these insights, this article explores the interplay between national identity and democracy by taking both individual support for democracy and the society’s level of democracy in to account. Support for the idea of democracy concerns beliefs, which capture moral judgements, or core values, that need to be present for actual democracy to exist. The latter, democracy as an actuality, refers to the empirical facts showing how opportunities to participate in free elections and constitutionally protected civil liberties vary between countries (Dahl 1998). Accordingly, how does the level of actual democracy affect individual relationships between non-voluntary national identity and support for the ideal democracy?

The literature suggests that high level of actual democracy contributes to the internalization of the core values of democracy (Muller and Seligson 1994), which should reduce the differences between those who emphasize non-voluntary features of nationhood and others. On the other hand, actual democracy’s emphasis of demos over ethos might threaten the congruence between non-voluntary identity and nation. If that is the case, interactions with higher levels of actual democracy would increase the differences between those who consider national membership as non-voluntary compared to others. Thus, the moderating effect of actual democracy is an open question that needs to be investigated empirically.

This article makes both theoretical and empirical contributions to the literature. First, I advance arguments explaining why and how non-voluntary national identity is negatively correlated with support for democracy and theorize about the role of actual democracy in the link between national identity and support for democracy in the abstract. Second, combining all 30 countries from the integrated dataset of European Value Survey (EVS 2017) with country-level data from Varieties of Democracy (V-Dem 2017), I show that the non-voluntary national identity correlates negatively with support for democracy. In particular, random intercept models reveal that the relationship between holding a non-voluntary identity and support for democracy depends on the actual level of democracy in a country.

2 Support for Democracy

It is often said that although democracy is imperfect, it is the best form of government available. When we compare non-democratic regimes with democratic regimes one might claim that democracies tend to be more prosperous. Such claims are partly generated outside the political domain, where deep-seated beliefs about people are imbedded in core values linked to early-life socialization. In this regard, people think of democracy as something more than procedural fairness and functioning institutions. We articulate support for democratic ideals (Dahl 1998, 9−11, 22−27, 44−46; Dalton 2004). Democracy as an ideal refer to the value that democracy is assumed to provide. Some believe that the value of democracy can be recognized in benign consequences such as social order, justice and benevolent citizens. Others emphasize the importance of democratic decision-making as such, that majority decision-making contains an intrinsic justice, a collective decision-making in which all participants remain equal (Christiano 2009, 288; Estlund 2008; Rawls 1971, 3; Rawls, 1999, 3).

Research on support for democracy sometimes conflates democratic beliefs with preferences for actual conditions such as level of actual democracy (Maseland and van Hoorn 2011), governmental trustworthiness (Newton and Zmerli 2008; Kazimierz and Krystyna 2009), and sometimes lacks clarity due to notions brought from everyday language (Dahl 1998, 26–32). To avoid these pitfalls, I first stipulate the core values of democracy, and then I review previous predictors of support for democracy.

The core values of democracy extend beyond the actual functioning of governments to the degree to which we support the notion of democracy itself. Support for democracy is not limited to how it is. It encompasses also how society could be organized. Building on the idea that how we want to govern ourselves is based on values, we need to describe these values. In an uncontroversial sense, the core values of democracy are defined as intrinsic equality and personal freedom, which means that “we ought to regard the good of every human being as intrinsically equal to that of any other” (Dahl 1998, 9−11; see also Christiano 2008). A person’s life, liberty and happiness are not intrinsically superior or inferior to the life of, liberty and happiness of any other. Intrinsic equality is the basic guiding principle, allowing for a variety of belief and practice in diverse society. The reciprocal principle, in which the parties respect one another, is motivated on moral grounds (Rawls 1972, 356−9; Weale 1985). Thus, intrinsic equality is the least common denominator for the variety of democratic ideals.

A person supportive of the idea of democracy thinks about citizens as moral-political equals, meaning that institutions and citizens should not act on fundamental intrinsic inequality that favors some communities and oppress other communities (Forst 2002, ch. 2). Linked to democratic practice: when democratic beliefs are socially institutionalized, a majority will not oppress a minority or remove democracy even if the majority have these opportunities, and citizens will not use their democratic right to vote for non-democratic candidates. The presences of the core values of democracy thus help to avoid controversial decisions and election results.

According to previous research, authoritarianism and age are the main two predictors of support for democracy in the abstract. Following Adorno and colleagues (1950), authoritarian-minded citizens, who think about people as types rather than constantly shifting and dissolving groups, have proven to score low on support for democracy. The state of mind depicted by belief in obedience, or submission to someone else’s authority, as well as acting of dispensing of that belief through the oppression of one’s subordinates, is not consistent with intrinsic equality (Allport 1954, 395–408; Dahl 1998, 66–68). In this regard, extreme views both on the left and right (Greenbergh and Jones, 2003), religiosity (Lipset 1981), and non-voluntary national identity (Erhardt, Wamsler and Freitag 2020, 59, 62), have been identified as being indicative of authoritarianism and thus correlate negatively with support for democracy.

In the contemporary discussion on “democracy’s fading allure” (Plattner 2015), age is the main predictor. The evidence is not conclusive. For instance, Foa and Mounk (2017a, 2017b) argue that the youngest generations in consolidated democracies exhibit regression away from robust support for democracy. Critics of these studies have pointed out that the findings are built on biased country selections and mislabeling of newer democracies as “consolidated.” In contrary, modernization scholars assert how the core values of democracy are more prevalent among younger generations who grew up in secure material environments. The young generation may be critical of political authorities, yet, committed in their democratic belief (Norris and Inglehart 2019; Kirsch and Welzel 2019; Voeten 2017). Wuttke and colleges (2020) have analyzed survey data from eighteen European countries employing cohort, life-cycle and period effects and found strong evidences supporting that citizens have not given up their support for democracy. This result is compelling in view of the fact that there is no conflict between age and intrinsic equality. Empirical research has mostly overlooked cultural accounts such as national identity when estimating support for democracy. Therefore, exploring the interplay between national identity and democracy contributes to the literature.

3 Defining Non-voluntary National Identity

Nation generally designates a group of people with common language, common cultural features, and a history of living together normally on the territory they regard as their home. As identity includes both self-identification and perceptions of others, national identity too expresses both angles. In respect to self-classification, a person has a national identity insofar as the person thinks of herself, that is, classifies herself, as belonging to a particular kind of people who is separated from others by virtue of belonging to a particular nation. Thus, a national identity is a “way of thinking”, a representation or a concept, linked to a specific nation (Miller 2003, 114−5). In contrast to the nation, national identity is not a group of people because the individuals who make up a nation usually represent and conceptualize the nation in different ways.

Regarding how a person represents the nation and thinks about other people includes normative stances distinguishing those who do belong to the nation from those who do not. The normative approach has been labeled using different distinctions: civic v. ethnic (Kohn, 2017 [1944]), primordialism v. instrumentalism (Bačová, 1998), ascriptive/objectivist v. civic/-voluntarist (Jones and Smith 2001), inclusive v. exclusive (Tudor 2018). What these distinctions have in common is that they raise the question of whether an individual’s membership in a nation should be regarded as voluntary or non-voluntary.

As national identity concerns both self-classification and normative stances about other people, the non-voluntary national identity represents two different kinds of entities. The first referent is a certain kind of national identity, a person who emphasizes the importance of origin and cultural markers learned by early socialization when she classifies herself as a member of the nation. The other is the requirements set for other people in order to be qualified as a member of the nation. That is, a particular way of being a member in which the membership is conditioned by non-voluntary features. In practice, these two entities generally coincide. For instance, a Norwegian who claims that a true member of the nation should be white and Christian, is normally white and Christian and self-classified as Norwegian. By generalization to all nations, my stipulative definition of the non-voluntary national identity is as follows:

Person A has a non-voluntary national identity if and only if A holds that the true member of the nation (i) classifies herself as [nationality] and (ii) appeals to national oriented properties that lie out of reach of ones’ will.

The requirement of being born of country ancestors and phenotypical characteristics is non-voluntary in all contexts. Cultural markers as language and religion requirement are not fully non-voluntary. New language learning and religious conversion are negotiable and something that we can choose to do, meaning that appropriate adjustments in cultural practices affect opportunities to be classified as a true member of a nation. However, this possibility does not annul the non-voluntary aspects. In general, native language and imposition of religious category are closer to non-voluntarism through early socialization. In respect to personal integrity, the pressure to convert and the pressure to learn the national language differ. New language learning is far less controversial, while the requirement to convert is controversial. Furthermore, historically, partly as a result of the theories of race, religion and descent became closely associated within the concept of nation (Hobsbawm 1992). Thus, religion requirement through early socialization generally counts as non-voluntary. To be more concrete, the non-voluntary national identity varies in restrictiveness. For instance, requirement on being born in the nation is less restrictive than requirements of being born of country ancestors with specific phenotypical characteristics and religious belonging.

4 Non-voluntary National Identity and Support for Democracy

In order to scrutinize the negative relationship between the non-voluntary national identity and support for democracy I have formulated two arguments. The first one concerns the relations between non-voluntarisms and non-universalizable stances, and the second one highlights the prioritization of national agency over individual freedom. Regarding the former, the core values of democracy and the non-voluntary national identity appear to be incompatible. The core values of the ideal democracy hold that no person is naturally superior to another and any relations of authority between them stand in need of justification. The justification of inequality on the basis of non-voluntary features does not emerge from or lend support for the assertion of intrinsic equality.

My second argument asserts that the general negative relationship remains even if a country is ethnically and racially homogeneous. This is because the non-voluntarist claim can weigh against the force of other claims; a prioritization of national agency over individual freedom. More precisely, the emphasis on non-voluntary features of national identity generally entails that the individual is not perceived as a free agent. Agency is rather ascribed to the nation. The duties towards one’s nation can therefore supersede individuals’ right to autonomy (cf. Aristotle, Politics Book II, section 1261a; Greenfeld and Eastwood 2005, 259). For instance, contemporary scholars have observed that non-voluntarist representations claim that women have a moral obligation to give birth to new members of the nation and to nurture them for the sake of the nation (Yuval-Davis 1997); metaphysical notions such as “the will of the nation” should not be contested by individual citizens or foreigners (Sata and Karolewski 2019); security and order of the nation are prioritized over person freedom (Mudde 2017, 3−5; cf., Allport 1954, 395−424); the world is perceived as a member of the nation, not as an autonomous human being (Wodak 2019, 65; aspiration to become an identical part of a national collective (Fromm 1994/1941; Kitschelt 2017, 360−1); national culture is not understood as individual lifestyles, but as a community in conflict with other communities (Biskamp 2019, 103). That is not to say that non-voluntary national identity automatically holds a certain combination of political views. Rather, it tends to prioritize national agency over individual agency. Thus, the incompatibility between non-voluntarism and the core values of democracy support and the prioritization of national agency over individual freedom support the following hypothesis:

H1: People who articulate higher levels of non-voluntary features of national identify are less supportive of the ideal democracy.

5 National Identity and Support for Democracy in Context

In hypothesizing how actual democracy moderates individual relationships between national identity and support for democracy, I start by illustrating the broad picture of the interplay between national identity and societal conditions. Then, I narrow it down to conditions making the non-voluntary national identity salient. This is key to the better understanding of the dynamics between national identity and democracy.

In a broader sense, there is a correspondence between representations of a nation (national identities) and existing nations (objects). We need to consider how specific societal conditions correspond to the national identity. The context in which national identities are formed is a complex of basic aspects such as the international order based on nations (Billig 1995, 7) and their locations (Kohn 1944), and diversity (Hjerm and Schnabel 2010); social institutions such as national lived culture (Calhoun 1997; Bonikowski 2016), constitutions (Schnabel, Behrens and Grötsch (2017), and economic wealth (Gellner 1983, 1987: see also Erhardt Wamsler and Freitag 20201); and social currents of national feelings (Durkheim 1895/1964, 4). The correspondence goes in two directions. On the hand, the sheer factuality of these conditions influence ones’ national identity. On the other hand, the individual national identity captures how people think of themselves, by interaction with other people and how they choose to act in the light of their thoughts. In this regard, the individual conceptualization of the nation refers to, and can even change, such societal conditions as exemplified above. Taken together, individuals and their thoughts represent different features of the nation, and, a particular nation affects how people think of themselves. This recursive relationship is inherent to the dynamic at any given point and to the evolution and formation of democracy.

Following this line of reasoning national identity is both a basic reference point and a specific way in which the individual represents the nation in her mind. However, although the interplay between individual and nation has evolved for thousands of years, the non-voluntary aspect of her national identity only becomes explicit in response to a perceived threat. Contemporary examples are found in the violent clashes in Rwanda (Bhavnani and Backer 2000), the former Yugoslavia and Soviet republics by the cleansing of non-nationals (Storey 2002). An individual’s non-voluntary conceptualization of the nation may also become salient in reaction to immigration, multiculturalism, globalization, a racialized hegemony, perceived threat by equality policies, the nativist manifestations among political elites, and so on. As long as the interplay between national identity and nation goes on without restraint, there is no need to appeal for the national oriented properties that lie out of reach of ones’ will (such as being born of country ancestors and phenotypical characteristics). The emphasis on a non-voluntary class arises when the congruence between the non-voluntary identity and nation is violated. The interaction with such entities is likely to be significant when we predict beliefs and behavior in regard to the non-voluntary national identity. The reasons are two-fold. First, the sense of social position provides the basis of violated congruence between non-voluntary identity and nation. The accentuation of the in-group identity preserves the integrity and the position of the dominant group, which usually emerges from historical hierarchical racial (non-voluntary) structures (Blumer 1958). Second, according to Allport, “national involvement” is the most important single factor in institutionalizing the dichotomized worldview. To conceptualize the nation in non-voluntary features becomes a positive anchorage providing definiteness and protection of the in-group, which entails a demand to control the perceived threat (Allport 1954, 395−424).

Against this backdrop, let us call this “the interplay between nation and non-voluntary identity”, I expect that the relationship between non-voluntary national identity and support for democracy will depend on the level of actual democracy. I propose two causal social mechanisms, which point in two different directions in respect to the moderating role of actual democracy. Regarding the amplifying capacity, the threat towards the social position of the non-voluntary identity will amplify the negative correlation between non-voluntary identity and support for democracy. Specifically, contemporary liberal democracies have generally emerged from ethnically homogeneous entities. At the same time, higher levels of actual democracy delineate citizens in voluntarist features, which widen the boundaries of national membership (Habermas 1992; for empirical findings see Gabrielsson 2021). The widening of group-boundaries is the result of hierarchy-attenuating legitimizing myths; a political doctrine that empathizes universal rights of man. The hierarchy-attenuating myths are tied to the patterns of beliefs and practices that form the civic culture and behavior devoted to equality (Pratto, Sidanius, and Levin 2006). This pretension, that no person is naturally superior to another, challenges the non-voluntary conceptualization of “the people”, which ought to coincide with one nation within a distinct geographical territory. As liberal democracies generally reprehend legitimate national membership as an extension of non-voluntary heritage, it creates a tension between non-voluntarist stances and the nation. Nevertheless, identification with non-voluntary properties still persist in these conditions, and a higher level of actual democracy is likely to affect the take on democratic beliefs. Thus, the hypothesized amplifying impact derives from the intensified conflict about intrinsic equality. The capacity of actual democracy to widen national group boundaries challenges the perpetual link between the non-voluntary identity and nation, while the moral justification for inequality in regard to favoritism about phenotypical characteristics and other non-voluntary properties is rejected. Accordingly, hierarchy-attenuating legitimizing myths and the defining of citizens in voluntarist features, and the conceptualization of national boundaries established with non-voluntary features, cannot be combined without a contradiction in principle. If this is true, the interaction with institutionalized democracy reinforces the inconsistency between the non-voluntarism and intrinsic equality, and supports the following hypothesis:

H2: Higher levels of actual democracy amplify the negative correlation between the non-voluntary identity and support for democracy.

Now let us examine the potentially reducing moderating effects. In accordance with this hypothesis, high levels of actual democracy are acknowledged as a condition suitable for the internalization of support for the ideal democracy (Muller and Seligson 1994). The concrete aspect of the hierarchy-attenuating legitimizing myth is the actual political institutions of modern European representative democratic government. When a society provides frequent and fairly conducted elections and freedom of expression, and grants access to alternative sources of information and the right to form independent association, and include all persons subject to the laws of that state, the society does something with the individual. The actual democracy invokes thoughts and actions conditioned by theses constituents, which emphasis its primary role: socialization. When these conditions exist, they are likely to internalize support for the ideal democracy more than any feasible alternative (Dahl 1998, 147; see also Estlund 2008, 42−45, 98). Following this idea, higher levels of actual democracy will make it possible to emphasize non-voluntary features of national identity and still being supportive to democracy in the abstract. In this regard, interactions with consolidated liberal democracy decrease the difference in support for the ideal democracy between the non-voluntary take on national identity and others. In a nutshell, the differences in support for democracy as a function of non-voluntarism become smaller as actual democracy increases. I posit as follows:

H3: Higher levels of actual democracy reduce the negative correlation between the non-voluntary identity and support for democracy.

6 Data, Measurements and Methods

The source of individual-level data used for the empirical part of the study is the EVS 2017. It covers many European countries with sufficient country-specific sample sizes to enable robust statistical analysis. Here, we analyze data from surveys of populations in 30 countries, including 56,368 respondents. The respondents are selected via representative multi-stage or stratified random sample of the adult population of the country, 18 years old and older.

6.1 Independent Variables

A large share of previous empirical research about normative conceptions of national identity relies on the civic v. ethnic distinction2 (Kohn 1944; for empirical findings see e.g., Berg and Hjerm 2010; Hjerm 2000; Larsen 2017; Jones and Smith 2001). As a number of commentators have pointed out, this approach also has disadvantages. First, the properties that create these indices are arbitrary as they can belong to “ethnic” in one part of the world and “civic” in another part of the world. Second, the method shows neither whether or not, nor how, the civic items coexist with ethnic items on an individual level and on an aggregate level. Third, the “civic/Western/political vs. ethnic/Eastern/cultural-argument” concerns European and Asian countries in particular, not all countries in the world (e.g., Özkirimli 2005; Smith 1983).

Some scholars have used an exploratory method called Latent Class Analysis (LCA) to reveal the different types of national identities that exist in a country (Bonikowski and DiMaggio 2016b; Soehla and Karim 2021). However, recent research casts doubt on the efficacy of this strategy (Eger and Hjerm 2021). A third approach distils three dimensions of national identity: civic (voluntary), cultural (in between voluntary and non-voluntary) and ethnic (non-voluntary) definitions (e.g., Pehrson, Vignoles and Brown 2009, 29−30; Shulman 2002).

Considering the limitations of the civic v. ethnic distinction and the exploratory strategy, I suggest a shift from the civic v. ethnic distinction to the voluntary v. non-voluntary distinction. In doing so, I use three items to create the additive index capturing non-voluntary national identity: “ancestry” and “born”, which necessarily define non-voluntarist membership; “religion”, which capture requirements about culture learned by early socialization (Cronbach’s alpha .76). The index is normally distributed and takes values from 1 to 10. Table 1 reports the number of respondents and how the questions about national identity were answered.

In general, religion requirements are unimportant, and the other criteria of national membership are considered to be important.

6.2 Dependent Variables

When considering survey data on support for democracy, it is important to note that the data measure citizens’ perceptions of the idea of democracy rather than actual government performance (Herbert 2013, 2). The “regime abstraction”, occurs when participants answer in relation to an ideal form of democratic governance, rather than in relation to how democracy is actually performing (Kiewiet de Jonge 2016) as a present democratic system (Zaller and Feldman 1992).

EVS provides data capturing confidence in democratic institutions; characteristics and function of democracy; citizen’s evaluation of institutional performance; importance of democracy: preferences about political system. Among these aspects of I use the two best suited indicators capturing support for the ideal democracy: “Having a democratic political system” is 1 = very god, 2 = fairly good, 3 = fairly bad, 4 = very bad. “Democracy may have problems but it’s better than any other form of government.” Responses vary from 1 to 4, where 1 = agree strongly, 2 = agree, 3 = disagree, 4 = disagree strongly. “How important is it for you to live in a country that is governed democratically?”, where 1 means is “not at all important” and 10 means “absolutely important.” The two items (r = .41) are collapsed into an index, “support for democracy”, and take values between 1 and 10 (n=50,548). Some studies include reverse coding of items capturing authoritarianism when measuring support for democracy (democracy–autocracy preference (DAP) scale, e.g., Magalhães, 2014), while other studies reports support for democracy and authoritarianism separately (e.g., Inglehart 2003). In this sample, including “having a strong leader who does not have to bother with parliament and elections” gives a Cronbach’s alpha of .38. Therefore, I only use the two variables capturing support for the ideal democracy explicitly. The sample distribution of the dependent variables is displayed in Table 2.

6.3 Country-level Variables

Varieties of Democracy (V-Dem) provides a multidimensional and disaggregated dataset that reflects this complexity, and specifically measures these five high-level principles of democracy (see Table 3). The V-Dem project is one of the largest, reliable social science data collections, and describes the level to which democracy is established for independent states. These measurements are well established and often referred to when the level of development of democracy is used in empirical research (Lindberg et al. 2014).

V-Dem uses two of these five principles to measure liberal democracy: fair elections (including freedom of association and freedom of expression, alternative sources of information, universal suffrage and the degree to which government policy is vested in elected political officials), and liberal principles (protection of individual and minority rights, the strength of rule of law and constitutionally protected civil liberties, independence of judiciary).

I also include several controls to factor out the possible confounding influence of citizen characteristics: political preferences, political trust, income, GDP per capita, and age. For specifications, see Appendix 1.

6.4 Statistical Methods

Due to the hierarchical structure of the data, multilevel methods are required to account for the interdependence of respondents nested within countries. Random intercept models, with national identity and support for democracy constituting level 1 and country-level of actual democracy constituting level 2, account for this dependency between individual observations and national contexts (cf., Hox 2010). Multilevel models also provide opportunities to investigate cross-level interactions. In this case, how the level of actual democracy moderates the individual-level relationship between national identity and support for democracy.

7 Results

The empty model reports a high mean (8.50) and low SE (0.02). The variance component between countries is 0.64 and 1.65 between individuals within countries. It gives us an interclass correlation coefficient at 13.2 %, meaning that 13.2 % of the variance exist between countries. In other words, properties at country-level as level of actual democracy can help explain support for democracy on individual level. The continual decrease of log likelihood provides an accurate model-fit.

Model 1 shows that non-voluntary identity is negatively correlated to support for democracy. Sex, left or right political views, and religiosity are insignificant and excluded from remaining models. Other individual-level control variables exhibit positive effects. My first hypotheses state that people who articulate higher levels of non-voluntary features of national identify are less supportive of democracy. The size of the fixed effect (−.07) is comparable with political trust (.08), which is considered to closely related to support for democracy. Considering also the high mean, the evidence to H1 must be seen as strong and compelling. The variance between countries has decreased with 12.2 % and indicates that means of non-voluntary features of national identity varies between countries. The decrease of individual-level variance reaches 6.6 %. It makes sense due to the fact that H1 is supported.

Model 2 adds fixed effects on country-level and interactions effects. The fixed effects on country-level are not hypothesized. However, the hypothesized interaction affect is easier to interpret if country-level fixed effects are reported. As we ca see, there is a strong positive correlation between level of actual democracy and support for democracy even after controlling for GDP per capita. I stated two hypothesis, which point in two different directions in respect to the moderating role of actual democracy. Regarding the amplifying capacity, H3 suggests that high level of actual democracy contributes to the internalization of support for the ideal democracy, which should reduce the differences in support for democracy between non-voluntary national identity and others. H2 states that actual democracy’s emphasis of demos over ethos threatening the congruence between non-voluntary identity and nation resulting in an amplified negative correlation between non-voluntary national identity and support for democracy. This means that individuals scoring high on non-voluntary features of national identity should be less supportive to democracy in countries with institutionalized protection of individual and minority rights, and constitutionally protected civil liberties. Cross-level interaction effects between individual level (non-voluntary national identity) and the country-level variables (level of liberal democracy) are estimated in order to test these hypotheses. As model 2 shows, higher levels of actual democracy affect support for democracy in accordance to H2. The cross-level interaction effect is negative (−.03), meaning that higher levels of actual democracy amplify the negative correlation between non-voluntary national identity and support for democracy. H3 is false. For illustrative purposes, predicative margins are generated using model 2.

Figure 1 shows that actual democracy effects the relationship between non-voluntary national identity and for support for democracy. The slopes indicate that the respondents emphasizing non-voluntary features of national identity exabit considerably lower support for democracy when actual democracy is at the highest than when it is at the lowest. The slopes are significant at the 0,01 percent probability level. Actual democracy has differentializing effect for democracy support outcomes. A person scoring high on non-voluntary national identity has a considerably higher probability of scoring lower on support for democracy when actual democracy is at the highest than when it is at the lowest. Another way to put this is that the difference between the non-voluntary national identity and others is larger in consolidated liberal democracy compared to non-democracies.

Finally, to further challenge my findings, I conducted a series of robust analyses. The theoretical strands are also tested by using alternative measurements of actual democracy at county-level: deliberation (DDI), egalitarianism (EGDI), electoral processes (EDI) and political participation (PDI) (V-dem 2017). The robustness checks confirm H1 and H2 (Appendix 2).

8 Conclusions

This article describes the interplay between actual democracy and how the non-voluntary national identity relates to support for democracy. The initial expectations on the individual relationships are verified. People who articulate higher levels of non-voluntary features of national identify are less supportive of ideal democracy. Most significantly, I also maintained that contextual characteristics in the countries under study have an impact on the individual relationship. I theorized on the interplay between nation and non-voluntary national identity of the non-voluntary national identity, and performed statistical tests using multilevel models. The results show that high level of actual democracy amplifies the likelihood that people scoring high on non-voluntary national identity score low on support for democracy.

The findings suggest that the non-voluntary national identity poses an obstacle to the development of support for democratic ideals. The non-voluntary national identity democracy stands in an inconsistent relation to intrinsic equality, and, the prioritization of national agency over individual agency does not fit the moral autonomy making up the democratic mind-set.

The interaction with higher levels of actual democracy goes in two directions. Actual democracy is believed to be a condition suitable for the internalization of intrinsic equality and a positive take on personal freedom, which relates to lower aggregate scores of non-voluntary national identities (Gabrielsson 2021). However, the actual liberal democracy also violates the congruence between nation and the non-voluntary national identity. Given that, the interplay between nation and the non-voluntary identity synthetizes the following explanation: as actual democracy increases, aggerate scores of non-voluntary national identity reduces, and, the diminishing proportion of the non-voluntary national identity will decrease in support for democracy in the abstract.

Two shortcomings should be noted. First, the statistical tests do not provide a causal explanation at the individual level. Other reasonable interpretations of the estimates are at hand. First and forest most, we might consider the reversed causal relationship on individual level, that is, high level of actual democracy amplifies the likelihood that a person scoring low on support for democracy scores high on non-voluntarism. However, this interpretation can be explained by the interplay between the non-voluntary national identity as well. Fixed unequal identity positions based on non-voluntary features paired with an emphasis on collectivism, in which the agency is attributed to nation over individuals, are incompatible to intrinsic equality and personal freedom. Therefore, the interaction with hierarchal-attenuating individualism of the kind that liberal democracy manifests will reinforce the covariation of classical nationalist stances with negative evaluation on the idea of democracy. Still, it would be appropriate with an argument that spells out the causal direction on the individual level. First, as notes earlier, previously research shows that aggregated scores of voluntarist civic nationhood together with lower scores of non-voluntarisms are a precondition to higher levels of liberal democracy (Gabrielsson 2021). Second, given that these features of national identity serve as a sufficient condition to achieve higher levels of liberal democracy, these findings are hard to tie in with the reversed causal direction (that is, the less support for democracy, the more non-voluntarism). In doing so, we have to present a theory that challenges the interplay between nation and national identity, and explain how and why lower levels of liberal democracy amplify the likelihood that the true democrat is a non-nationalist. It does not seem plausible. In other words, the causal direction proposed by the first hypotheses is reasonable.

Second, abstraction regarding support for democracy may be a problem in opinion survey research when the term refers to different things to different people. Beyond “intrinsic equality” some respondents might picture the core values of democracy in metaphysical extensions as “the will of the people”; democracy as a way to tracking the truth (epistemic democracy, e.g., Estlund 2008); or, deliberative aspects (Habermas 1986). In order to better capture democratic beliefs, or support for democratic ideals, I see two possible routes. One is to operationalize the moral judgments capturing intrinsic equality. The other one is to take the psychological pre-dispositions of “democratic personality” in to account (Allport 1954, 429). In this regard, mental flexibility is indicative. The mentally flexible can handle ambiguity and have less need to categorize quickly or to persist in categories once made. The democratic personality has an inward-looking curiosity about imaginative processes, fantasies, and theoretical reflections, and a generalized state of friendliness and goodwill and holds a positive view of human nature, as well as empathic abilities. The democratic personality understands another person’s state of mind, and adapts to it (ibid., 1954, 429, 431, 435–6; see also Hjerm et al. 2020).

I suggest two directions for future research: monitoring the extent to which national identity affect support for democracy in non-European countries; exploring the impact of other country-level variables such as constitutions, nativist manifestations among political elites, perceived national threat, globalization and immigration.

Given the vast literature on the topic of how national identity relates to democracy, and the comeback of nationalist stances in political domain, national identity ought to be incorporated into the study of democracy. Empirical research on this topic helps to evaluate the democracy-oriented consequences of peoples’ views and beliefs about their nations and bring clarity whether national identity and democracy are coimplemental or competing.

We cannot choose our national origin, yet, we can choose our concept of the true people. This choice affects prospective conditions of democratic beliefs and democratic governance. The more restrictive national identity, regarding non-voluntary features of national identity, the less democracy.

Appendix 1

See Table A1. All control variables are mean-centralized. Data regarding GDP per capita comes from World Bank (2017). The remaining variables/index are standardized (take values between 1 and 10).

Political trust is an index based on four variables: “Please indicate how much confidence you have in each of the items presented in the next questions. Is it a great deal, quite a lot, not very much or none at all? Please indicate how much confidence you have in Parliament/ The justice system/Political parties/Government?” 1 = a great deal, 2 = quite a lot, 3 = not very much, and 4 = none at all.

I have used a harmonized variable to capture education (Q81): “What is the highest educational level that you have attained?”0 = Less than primary, 1 = Primary, 2 = Lower secondary, 3 = Upper secondary, 4 = Post-secondary non-tertiary, 5 = Short-cycle tertiary, 6 = Bachelor or equivalent, 7 = Master or equivalent, 8 = Doctoral or equivalent.

Income: “Here is a list of incomes and we would like to know in what group your household is, counting all wages, salaries, pensions and other incomes that come in. Just give the letter of the group your household falls into, after taxes and other deductions.” The variable takes values between 1 and 10, were 1= 1th decile and 10 = 10th decile.

Age of the respondents is recoded in 7 intervals, were 1 = 18–24 years, 2 = 25–34 years, 3 = 35–44 years, 4 = 45–54 years, 5 = 55–64 years, 6 = 65–74 years, 7 = 75 and more years.

Political views: “In political matters, people talk of ‘the left’ and ‘the right’. How would you place your views on this scale, generally speaking?” The variable takes values between 1 and 10, were 1 = left and 10 = right.

One variable is used to capture religiosity: “Independently of whether you go to church or not, would you say you are …” 1= religious person, 2 = not a religious person, 3 = a convinced atheist.

Appendix 2

The theoretical strands are tested by using alternative measurements of actual democracy at county-level: Deliberative Democracy Index (DDI, Table A2), Egalitarian Democracy Index (EGDI, Table A3), Electoral Democracy Index (EDI, Table A4), and Participation Democracy Index (PDI, Table A5). The robustness checks confirm H2 (Higher levels of actual democracy amplify the negative correlation between the non-voluntary identity and support for democracy).

Why would the argument made by Samuel Huntington about future global conflicts be characterized as Primordialist?

In what way does a Primordialist explanation of the politicization of identity differ from a constructivist explanation?

Which of the following best characterizes regimes in the study of comparative politics quizlet?

Which of the following statements best characterizes Madison's dilemma?