Introduction

The left–right dimension of politics is the most dominant ideological dimension in European representative democracies. In Western Europe, the dimension mostly takes on a common pattern, where the Left (compared to the Right) favors state intervention and economic redistribution, as well as socially progressive policies such as positive views toward immigration and feminism (Van Der Brug and Van Spanje 2009). This fits the theory that the left–right dimension is at its core about rejection/acceptance of inequality (MacIver 1947, p. 216; Lipset et al. 1954; Inglehart 1984, p. 293; Bobbio 1996; Lukes 2003; Noël et al. 2021, p. 317; Lindqvist 2023), given that the Left is thus fairly consistently (more than the Right) rejecting inequality along different issue dimensions, regardless of whether the inequality concerns high- and low-income individuals, men and women, or natives and immigrants. However, the left–right dimension in East European (/post-communist European) countries seems to follow a more heterogeneous pattern, where the Left is not necessarily the most vigilant against (different types of) inequality. Research on voters finds that while egalitarianism is associated with left-wing self-placements in Western Europe, the same is not true in Eastern Europe (Thorisdottir et al. 2007; Aspelund et al. 2013). Similarly, Wojcik et al. (2021) find that economic and cultural liberalism are associated with left–right self-placements in Western, but not Eastern, Europe. Their results indicate that terms ‘left’ and ‘right’ have different meanings in different contexts. As does a recent paper by Leykin and Gorodzeisky (2023), who find that right-wing individuals are more favorable to immigration than left-wing individuals in several post-communist countries in Eastern Europe. A seemingly reasonable conclusion then is that while acceptance of inequality seems to separate left- and right-wing individuals in Western Europe, the same is unlikely to be the case in Eastern Europe. This would mean that what is considered left and right in different countries is simply context dependent.

This paper examines whether this conclusion is valid. Contrary to this established view, this paper finds that acceptance of inequality is associated with left–right self-placements in both Eastern and Western Europe. Key to this finding is that previous research on left–right politics in Eastern Europe has overlooked two important aspects that are considered in this paper: (1) citizens’ attitudes toward different inequalities are more likely to be associated with left–right self-placements than an underlying egalitarianism, and (2) attitudes toward an inequality should only be associated with voters left–right self-placements if the specific inequality is important/salient in that political context. Attitudes toward inequalities that are not politicized/salient cannot be expected to correlate with left–right self-placements in the expected direction because it is then unlikely that individuals either (a) copy positions of their chosen parties/side on those inequalities, or (b) choose their left–right self-placement because of such inequalities. This is especially important to consider in Eastern Europe where there is a “structured diversity” of issue salience (Rohrschneider and Whitefield 2009), more so than in Western Europe.

To re-examine the relationship between acceptance of inequality and left–right self-placements in Europe, with the divergence of issue salience in mind, I study voters’ attitudes toward several different inequalities. Drawing on arguments made by MacIver (1947, p. 216), Lipset et al. (1954), Inglehart (1984, p. 293), Bobbio (1996), Lukes (2003), and Joshi (2021), I utilize four different instances of societal inequalities (similarly to Lindqvist 2023) where one group overall has/is doing better than the other—rich/poor, men/women, heterosexual/homosexual individuals, and citizens/immigrants. Since different inequalities are likely to be associated with individuals’ left–right self-placement in different political contexts, there is a clear observable implication: acceptance of each inequality ought to be associated with right-wing orientations, dependent on how politicized the particular inequality is in that context. For example, where immigration is a highly salient issue, we should expect those who are anti-immigration (and thus more accepting of the inequality between immigrants and non-immigrants) to be more likely to place themselves to the right on the left–right spectrum, whereas those who are pro-immigration will be more likely to place themselves to the left. High salience of an inequality should thus make it more predictable which side takes which position on the issue (right-wing individuals should be more accepting of the inequality).

Using survey data from the European Social Survey (ESS) for 27 countries from 2008 to 2016, I show that acceptance of inequality is associated with right-wing self-placements in both Western and Eastern Europe in contrast to previous research. I furthermore demonstrate the importance of considering political salience when examining this relationship for three of the four inequalities (class, immigration, and sexuality), by utilizing multilevel models and interaction effects between the country-level and the individual-level variables. The paper finds that higher salience of an inequality at the country level is associated with a stronger relationship between right-wing self-placements and acceptance of that inequality at the individual level. These findings demonstrate the potential of acceptance of inequality theory to explain left–right politics in various contexts.

Can a single criterion explain the left–right dimension in Europe?

Jost (2006, p. 654) describes the left–right dimension as “the single most useful and parsimonious way to classify political attitudes for more than 200 years.” It is all the more surprising then that there is no consensus as to what the left–right dimension actually entails. Or to be more specific, whether the dimension has any substantive meaning that can travel across time and space. Even though this consensus is lacking, political scientists frequently classify many phenomena and entities as left or right (such as ideologies and policies), and use the left–right dimension to measure the ideological position of parties and voters. Understanding why someone is placed to the left or right in this ideological space would be of utmost importance to political scientists as it would improve measurement validity, as well as for voters since a better understanding of what separates left from right could clarify their vote choices.

Multiple authors have attempted to explain what separates left from right by proposing a criterion (i.e., a guiding principle) that can explain the underlying divide (MacIver 1947; Lipset et al. 1954; Rokeach 1973; Inglehart and Klingemann 1976; Laponce 1981; Silverman 1985; Inglehart 1984; Bobbio 1996; Lukes 2003; Jost et al. 2003; Joshi 2021; Lindqvist 2022). While they are divided on this subject, proposing different explanations, one must conclude that the most popular perspective is that acceptance of inequality separates left from right, as most of the proposed criteria involves acceptance of inequality in some way or another (with Silverman being the only exception of the cited authors). This paper follows literature (Inglehart 1984, p. 293; Joshi 2021; Lindqvist 2023) that posits that inequality can take place along different group divides, such as between more and less affluent individuals, or ethnic majorities and ethnic minorities (where the majority position provides an advantage when deciding on ethnic policy).Footnote 1 Importantly, different social divisions form the basis of the L–R divide at the national level in different countries (Jou and Dalton 2017). The salience of political issues varies between countries, and these variations must be taken into account when examining the correlations between attitudes toward different inequalities and individual left–right self-placements because they change the meaning of the left–right language in that context. For example, when the main dominant divide between political parties in a society is about economic issues, then we would expect that the meaning of the left–right language at the national level should concern economic issues. However, political parties in another society may chiefly be divided on ethnic grounds, and the meaning of the left–right language should instead correspondingly be tied to ethnic politics. As a consequence of this, some countries have left–right landscapes that hardly resemble each other at all.

However, recent research provides reasons for some skepticism of any universal explanation of the left–right dimension. Some researchers question whether the left–right dimension has the same meaning for citizens in different countries in Europe, which would mean that the left–right dimension cannot be used in cross-country comparisons, at least not without taking these differences into account (Zuell and Scholz 2019). While there may be differences in how individuals in various countries understand the left–right terminology, this does not contradict the idea of a criterion and in particular the acceptance of inequality criterion given that it can explain different inequalities depending on their salience. For example, left–right politics might be more about economic inequality in Portugal, while it is more related to the immigration dimension in Austria, but it can nevertheless be explained by the equality–inequality criterion in both instances (the Right is more accepting of the inequality in each case). The main difference between Austria and Portugal lies in the difference in salience of political issue dimensions.

Key here is thus to separate between attitudes toward different inequalities. This can reconcile acceptance of inequality theory and the contemporary debate on multidimensional voter preferences. Research shows that voters regularly mix left-wing attitudes on some dimensions with right-wing attitudes on others (Malka et al. 2019). We therefore must consider that an underlying (in)egalitarianism is not necessarily creating consistent left- or right-wing issue attitudes (and accompanying left–right self-placements) for many individuals, at least at the voter level. This is not to say that there could not be an underlying egalitarianism at work (at least for some voters, and political elites), only that (1) many individuals have different attitudes toward different inequalities (i.e., are not consistently egalitarian). Furthermore, (2) those individuals who choose to vote for the Left or the Right because of their political attitudes are more likely to do so in regard to how they view specific inequalities, rather than how they view inequality as a more abstract concept.Footnote 2 When testing whether acceptance of inequality is associated with left–right self-placements at the voter level, it is therefore more appropriate to examine attitudes toward different inequalities rather than individuals’ level of egalitarianism (see also Lindqvist 2023; Lindqvist and Dornschneider-Elkink 2023).Footnote 3

Dividing the more general concept acceptance of inequality into attitudes toward different inequalities is especially important when analyzing left–right self-placements in Eastern and Western Europe simultaneously. There are clear differences in issue salience between countries in Eastern Europe (Rohrschneider and Whitefield 2009) as well as between the two regions, and studies highlight that left–right competition manifests itself very differently in Eastern Europe compared to Western Europe. The Right in Poland and Hungary for example have been found to increase government spending more than the Left (Tavits and Letki 2009). This relationship is commonly assumed to be the opposite in Western representative democracies. Furthermore, researchers find that social conservatism (which is often connected to resistance to egalitarian progressive change) is associated with the West European Right, but with the Left in many countries in Eastern Europe (Rovny and Edwards 2012, p. 54; Hadarics 2017). These patterns suggest that individuals’ and parties’ attitudes toward different inequalities are not consistently bundled together (i.e., consistently accepting or rejecting inequalities) as often in Eastern Europe as in Western Europe.

There is some previous research testing whether acceptance of inequality is associated with right-wing self-placements. Evans et al. find that egalitarian attitudes have a relationship with left-wing self-placements in the British electorate (Evans et al. 1996), and Wiesehomeier and Doyle (2012) find similar results in Latin America. Weber (2012) finds that general egalitarianism is related to left-wing self-placements in Europe, but that this effect is conditioned on party polarization. Jost et al. conduct a meta-analysis of the political psychology literature, and find that acceptance of inequality (as well as resistance to change) is associated with right-wing/conservative political orientation (Jost et al. 2003). Lindqvist (2023) uses World Values Survey data to examine countries in North and South America, Europe, Asia, and Oceania, and finds that acceptance of different inequalities (such as income and gender inequality) is generally associated with right-wing self-placements (whereas attitudes toward government ownership of industry are less associated with political orientation). Cochrane also examines both European and non-European contexts (but studying party manifestos instead of voters) and finds that the Left is clearly more associated with equality than the Right (Cochrane 2015, pp. 73–75).

In contrast, research on Eastern Europe finds that acceptance of inequality is not (consistently) correlated with right-wing self-placements there (Thorisdottir et al. 2007; Aspelund et al. 2013). Yet, it is not clear whether the equality–inequality explanation really fails to hold in this region. Thorisdottir et al. (2007) and Aspelund et al. (2013) measure attitudes toward inequality (with data from the European Social Survey in 2002, 2006, and 2008) with the variable: “He[/she] thinks that it is important that every person in the world should be treated equally. He[/she] believes everyone should have equal opportunities in life.” This variable measures equality of opportunity, which is a related, but importantly, different concept compared to acceptance of inequality as discussed in this paper. The measurement of attitudes toward inequality in this paper instead concerns support for change in an egalitarian direction in regard to salient inequalities between groups, where one group has it (/does) better overall compared to the other group. The difference between these two approaches means that a re-examination might deliver different results.

Findings from two more recent studies also challenge the idea that acceptance of inequality is associated with right-wing self-placements in Eastern Europe. Wojcik et al. (2021) find that cultural liberalism is not consistently related to left-wing self-placements in Eastern Europe. However, they measure cultural liberalism with attitudes toward abortion, gay rights, divorce, and euthanasia, where only gay rights is clearly related to an inequality (that between sexual minorities and the heterosexual majority). From their results, it is not clear that attitudes toward different inequalities are unrelated to left–right self-placements, in particular depending on the salience of each issue. Similarly, although Leykin and Gorodzeisky (2023) find that positive views toward immigration are not correlated with left-wing self-placements in Eastern Europe (even though they are in Western Europe), their analysis does not take into account salience and attitudes toward other inequalities that may explain left–right self-placements.

In summary, the general correlation hypothesized by acceptance of inequality theory (H1) and the related interaction effect (H2) are therefore examined in this paper, succinctly formulated, as follows:

H1

Acceptance of inequality is associated with right-wing self-placements.

H2

The relationship between acceptance of an inequality and right-wing self-placement is moderated by the salience of that inequality.

Acceptance of inequality theory stands against empty vessels theory (Sartori 1976, p. 335), which emphasizes that citizens understand the terms in varying ways depending on their context and social background and that there consequently is no one criterion that can separate between left and right in different contexts. This explanation may be understood as a null hypothesis, i.e., that there is no dividing principle that remains the same between contexts (Lindqvist 2023, p. 6). In addition, there are multiple other competing explanations/theories for what separates left from right in politics that can be found in the literature: secularism–religiosity (Laponce 1981), universalism–particularism (Silverman 1985), for–against government intervention in the economy,Footnote 4 change–resistance to change (Jost et al. 2003), and diffusion–concentration of power (Joshi 2021).Footnote 5 This paper is not focused on testing these alternative explanations/theories but the analysis in the main paper, as well as additional analyses in the Online Appendix, control for the most popular alternative explanations when testing H1. Future research will benefit from testing these competing frameworks further.

Data and variables

To test the hypotheses, this paper examines whether respondents’ left–right self-placements in different countries are associated with attitudes toward salient inequalities, using data from the European Social Survey from 2008 and 2016 (ESS 2016, 2008).Footnote 6 The ESS is a cross-national survey that enables this paper to control for many intervening variables, and the two survey years are selected due to the availability of survey items measuring the independent variables. The individual survey answers are analyzed using multilevel models with observations hierarchically structured, i.e., survey respondents (level 1) are located in different countries (level 2). Random intercepts for each country are employed, as well as random slopes for certain variables. All countries were included in the sample as long as they had been democratic for at least ten years at the time of the survey (using scores from the Polity IV index, see Marshall et al. 2014), because democratic experience is often assumed to help individuals develop consistent left–right self-placements and associations (Thorisdottir et al. 2007, p. 183; Noël et al. 2021).

The dependent variable is Left–Right Self-Placement measured on an eleven-point scale, and there are four independent variables that measure different attitudes regarding change in an egalitarian direction. Given the theory, it is logical to treat an individual’s left–right self-placement as being explained by their attitudes toward specific group inequalities. However, left–right self-placements can also in some cases conversely cause issue attitudes and this paper therefore mostly discusses these relationships as correlations.

Attitudes toward inequality are measured with four societal inequalities that can be characterized as cleavages or group divides, but these four inequalities are not exhaustive of the entire universe of different potential inequalities. These dimensions are however useful for this examination because contemporary research demonstrates that they are important/salient issue dimensions in Europe (Caughey et al. 2019). In each case, we expect that any support for the lower group indicates support for less inequality on that dimension. The first group conflict is the class cleavage. This is operationalized with the item Economic Inequality, measured in the ESS data with three statements, where respondents are asked to state their level of agreement: “Large differences in income acceptable to reward talents and efforts,”Footnote 7 “For fair society, differences in standard of living should be small” and “Government should reduce differences in income levels.” The four independent variables are all operationalized by using the available and suitable items. This means that if more survey items are available, measuring the same concept, then they are added to construct the whole variable in that data set. In that sense, the independent variables are not always measuring the same underlying attitude, but function as measurements of individuals’ general level of hostility to/acceptance of inequality regarding each group conflict/cleavage.

The second inequality concerns ethnic minorities/poor immigrants compared to native citizens. The variable Less Immigration is measured in the ESS data with two items: “Allow many/few immigrants of different race/ethnic group from majority” and “Allow many/few immigrants from poorer countries outside Europe.” Both policies are changes in an egalitarian direction—these immigrants have lesser outcomes in general than the current citizens and are seeking to improve their conditions. The last two inequalities are between women and men, and homosexual compared to heterosexual individuals. Like the class cleavage, each of these group conflicts has two groups that have differences in overall outcomes. Having better outcomes is defined by what people value, for example, money, power, and status. One group has worse overall outcomes (e.g., women earn less than men) and/or less political power (e.g., homosexual individuals are a minority, and women usually occupy less seats in national parliaments).

The variable Intolerance of Homosexuality is measured with the items “Gays and lesbians free to live life as they wish,” “Gay and lesbian couples right to adopt children,” and “Ashamed if close family member gay or lesbian” (the latter two are not in the 2008 data). The former two items are policy prescriptions. If these policies are not in place, then supporting them constitutes favoring change in an egalitarian direction. This is true even if the policy has no implication for heterosexual individuals since (in)equality is a relative concept. If these policies are in place (dependent on the specific country), then opposing them would constitute favoring change in an inegalitarian direction. “Ashamed if close family member gay or lesbian” is not measuring any specific policy, but rather is aimed at measuring homophobia and prejudice. It is possible to be homophobic on a personal level yet favor policies in a more egalitarian direction, such as gay rights (which would result in measurement error). Therefore, as a robustness check, the analysis in a subsequent model uses only “Gays and lesbians free to live life as they wish” (and using only “Men should have more right to job than women when jobs are scarce” for the Anti-Feminist Attitude variable), resulting in no changes to the conclusion of this paper (see Fig. SM6 and SM7 in the Online Appendix).

The last independent variable is Anti-Feminist Attitude, which is measured with one item in the ESS (2016) data: “Men should have more right to job than women when jobs are scarce.” This is a policy description which indicates a change in an inegalitarian direction (for countries where this is not the case). For the ESS (2008) data, the survey item “Women should be prepared to cut down on paid work for sake of family” is also added to the variable. This is not a policy prescription and is therefore (similarly to the attitudinal item measuring homophobia) used as a proxy to measure how much the respondent favors change in an egalitarian or inegalitarian direction between men and women. As stated in the previous paragraph, an additional analysis only using “Men should have more right to job than women when jobs are scarce” produces only minor changes to the results (see Online Appendix).

When explaining left–right self-placements of individuals, Inglehart and Klingemann (1976) suggest three predictors: partisanship, social characteristics (sociodemographic variables), and ideology. Thus, left–right self-placements ought to be a function of these three (coupled with other less important variables). While it is possible to control for sociodemographic variables, it is very difficult to separate an individual’s party choice and an individual’s ideology as these are closely linked (Inglehart and Klingemann 1976). Party choice is therefore not controlled for. However, a separate analysis is conducted where partisanship is controlled for in the Online Appendix, and the results are very similar to those presented in the main paper. Different sociodemographic control variables are however included in the analyses, mimicking previous literature on the topic: Female, Age, Household Income, as well as other standard control variables such as Education Level, Union Member, Religiosity, Rural, Ethnic Minority and dichotomous variables for different religious denominations. The ESS data also allow the control of four psychological variables: Traditionalism, Rule-Following, Need for Security, and Openness to Experience. These are psychological factors that may affect an individual’s left–right position, as well as correlate with an individual’s acceptance of inequality (see also Jost et al. 2003; Thorisdottir et al. 2007). One last control variable is added: More EU Integration, which reflects attitudes toward the EU.

The data from the ESS do not contain some important additional control variables. Specifically, there is no variable measuring attitudes toward government intervention in the economy, freedom, and resistance to change, which have been linked to both ideological orientation and acceptance of inequality in previous literature (Downs 1957; Rokeach 1973; Jost et al. 2003). As a robustness check, separate analyses are conducted (see Online Appendix) where these variables are included as control variables, using data from the World Value Survey (WVS) as well as the European Values Study (EVS). These analyses are overall similar to the main analysis, with the only exception that they do not include the psychological control variables Traditionalism, Rule-Following, Need for Security, and Openness to Experience. The overall results for the independent variables (attitudes toward inequality) are also similar (see Online Appendix for further details).

In a second step of the analysis of the ESS data, I test whether the political salience of an issue moderates the relationship between attitudes toward an inequality and left–right self-placements (an interaction effect), as stated by H2. To this end, I employ data (Volkens et al. 2020) from the Comparative Manifesto Project (CMP) to measure the average salience of economic issues and multiculturalism (which is closely related to immigration) in available manifestos ten yearsFootnote 8 leading up to the ESS survey. The specific selection of manifesto items can be found in the Online Appendix. Country-level salience is measured as the average percentage of coded manifestos in a country that contained the different dimensions (economic and multiculturalism respectively). Party manifesto salience scores are weighted by the party’s vote share in the corresponding parliamentary election. Unfortunately, neither gender nor same-sex political salience can be measured using the CMP. The Chapel Hill Expert Survey (CHES) likewise has no measurement of the salience of gender politics, but contains a variable where experts were asked to assess the salience of “social lifestyle, (e.g. homosexuality)” for each party (Polk et al. 2017), and I use this item in the analysis. Stecker and Tausendpfund (2016) use the same CHES item in their analysis, arguing that it approximately measures the same dimension at the party level as the ESS survey item “Gays and lesbians free to live life as they wish” measures at the individual level. The closest previous CHES data available for each ESS survey are utilized to measure the salience of same-sex politics at the country level (see more detailed information on this measurement in the Online Appendix).

Analysis and results

The variables are analyzed in multilevel regression models with different data sets separated. The independent variables are standardized in all models (in the paper and the Online Appendix) to facilitate comparison (Table SM2 and SM3 in the Online Appendix display descriptive statistics). The ESS provides data for 27 countries (11 in Eastern Europe, 16 in Western Europe). The four independent variables have random slopes in the multilevel models as they are expected to differ by country. More EU Integration also has a random slope since its correlation with Left–Right Self-Placement substantively varies in different countries (Van Elsas et al. 2016). The multilevel regression model coefficients are found in Table SM4 in the Online Appendix.

The expectation of acceptance of inequality theory is that different inequality dimensions will be associated with left–right self-placements in different contexts. Figures 1 and 2 display the coefficients and their 95% confidence intervals in the multilevel models for the four independent variables Economic Inequality, Anti-Feminist Attitude, Intolerance of Homosexuality, and Less Immigration, for each of the countries in the ESS data. The overall results demonstrate strong support for H1. Acceptance of inequality in at least one of the four areas is associated with right-wing orientation in all countries. For some countries, all four variables have significant positive coefficients (such as the Netherlands and Croatia).

Fig. 1
figure 1

Multilevel regression model explaining left–right self-placements. The figure displays coefficients and 95% confidence intervals (calculated with 400 bootstrap samples) by country

Full size image
Fig. 2
figure 2

Multilevel regression model explaining left–right self-placements. The figure displays coefficients and 95% confidence intervals (calculated with 400 bootstrap samples) by country

Full size image

Overall, the independent variables are in almost every country either positively significant or non-significant. However, Anti-Feminist Attitude and Intolerance of Homosexuality have rather weak coefficients (and many times non-significant). Instead, attitudes toward economic inequality and immigration are seemingly more closely related to left–right self-placements in most countries, which is expected due to the higher levels of salience of these issues. This seems to be the case especially in later years, as the coefficients for attitudes toward immigration are stronger in 2016 compared to 2008. This pattern corresponds well with the trend of increasing salience of immigration as a political issue in recent decades. It is also possible that the 2016 results may (in part) be due to the refugee crisis in Europe in 2015, which increased the salience of immigration.

Importantly, the same pattern is found for East European countries as for West European countries, in stark contrast to previous studies on the same topic that found no link between acceptance of inequality and left–right self-placements in Eastern Europe (Thorisdottir et al. 2007; Aspelund et al. 2013). However, there are two important differences between the two regions. One, countries in Western Europe regularly exhibit stronger relationships between attitudes toward different inequalities and left–right orientations, compared to East European countries. This is partially expected given that these countries generally have shorter democratic experiences, and both parties and voters there are therefore likely to have less stable ideological foundations. See for example the prevalence of centrist anti-establishment parties in this region, who unlike radical left and right parties in Western Europe do not attack elites based on ideological grounds (Engler 2023). Another possible explanation for the weaker relationships could be that other inequalities are important in Eastern Europe, but this paper is unable to evaluate this.

Two, in Eastern Europe, there is a clearer distinction between different countries compared to Western Europe, mainly because more coefficients do not reach statistical significance. The Czech Republic, Estonia, Slovakia, and Romania have rather strong coefficients on the economic dimension, but not as strong for the other issues. Latvia and Hungary instead have stronger coefficients on the immigration dimension, compared to their coefficients for Economic Inequality. Conversely, the strongest significant coefficients for Poland, Croatia, and Slovenia are the ones for attitudes toward gender and sexuality.Footnote 9 This variation in which attitudes correlate with left–right positions in different East European countries is an indication of that different political conflicts are important in different contexts there. H2 states that issue salience can help explain these different patterns. This claim is important for acceptance of inequality theory because it creates an observable prediction: the link between acceptance of a specific inequality and right-wing self-placements should be stronger in contexts where that inequality is salient. This paper tests this prediction by including the country-level salience of each issue in the multilevel models. The salience measures are mean centered (non-mean-centered initial salience data is available in the Online Appendix) because each interaction effect is then estimated at the mean-level of salience for the other issues. All interaction effects are statistically significant and positive (see tables for both years in the Online Appendix). Figure 3 shows the interaction effects between country-level issue salience and attitudes toward that specific inequality in 2008 and 2016 (the multilevel models can be found in Table SM15 in the Online Appendix). The two-way interactions are presented in Fig. 3 by plotting the estimated coefficients of acceptance of inequality, conditional on issue salience in the interactions, where the shaded areas indicate the 95% confidence intervals (Solt and Hu 2021).Footnote 10 The country-level salience observations are displayed at the bottom of each graph in the form of histograms (Solt and Hu 2021; Berry et al. 2012, p. 668). The results demonstrate that the positive correlation between acceptance of a specific inequality and right-wing self-placements is stronger when the issue itself is more salient. In most cases, the estimated coefficients are not significant at very low levels of issue salience, in line with our expectations. Altogether, this provides strong evidence for H2 because high levels of issue salience are associated with issue attitudes more clearly conforming to the expectations of acceptance of inequality theory. This is in line with previous research that has shown that salience of immigration correlates strongly with polling for the far-right (Dennison and Geddes 2019, p. 114).

Fig. 3
figure 3

Multilevel regression models explaining left–right self-placements using ESS data (2008 and 2016), with interaction effects between survey respondents’ acceptance of different inequalities (level 1) and country-level salience (level 2). Shaded areas indicate 95% confidence intervals

Full size image

Conversely, lack of salience may help explain some of the weaker cases for the hypotheses. In particular, the Czech Republic and Lithuania produce the weakest results in Figs. 1 and 2, with a few significant negative coefficients. These cases provide important insights, best exemplified by the Czech Republic. The economic dimension is strongly positively correlated (coefficient around 0.5), while attitudes toward homosexual individuals in 2008 and 2016, as well as women in 2008, are negatively correlated with left–right self-placements. These coefficients suggest that support for equality on the socio-cultural dimension in the Czech Republic is correlated (albeit weakly) with right-wing self-placements. It should be noted, however, that the only consistent effect in the Czech Republic in the EVS (see analysis in the Online Appendix, Table A11) and ESS data (as well as when controlling for partisanship) is the positive coefficient for Economic Inequality (the same is true for Lithuania). Economic Inequality also has a much stronger relationship with left–right politics than the socio-cultural variables. In fact, the Czech Republic has the strongest coefficient for economic inequality of all post-communist countries in both 2008 and 2016. The strength of the class cleavage is also in line with the literature on politics in the Czech Republic (Hloušek and Kopeček 2008). The same can be said for Lithuania (Jurkynas 2004), where the economic cleavage (albeit weaker than in the Czech Republic) has also been more salient compared to other issues. The Czech Republic and Lithuania thus partially conform to the theoretical framework, albeit as weaker cases since the coefficients for social (/immigration) attitudes are expected to be non-significant and not slightly negative when the economic dimension is dominant (at the expense of social issues). Conversely, the fact that Economic Inequality for Hungary in 2008 (see Fig. 1) is not significant does not necessarily contradict H1. Class has had “little bearing on political divisions” and “the relevant dimension of substantive political conflict in Hungary is cultural” (Vegetti 2019, p. 78), in line with the positive coefficients for the non-economic independent variables in this paper.

While this paper finds support for acceptance of inequality theory, there remain alternative explanations that can explain the results. It is for example possible that the Right is primarily traditional and not interested in what might be argued as recklessly fast change, and therefore does not support change in an egalitarian direction (concerning the four dimensions measured in this paper). The same is true for government intervention in the economy—the Right might be skeptical of government intervention and therefore oppose egalitarian change (although this would not explain non-economic change toward more equality). However, when controlling for Resistance to Change and Privatization using World Value Survey and European Values Study data, the coefficients for acceptance of inequality are still significant and positive (these analyses can be found in the Online Appendix). Nevertheless, Resistance to Change and Privatization also demonstrate isolated correlations with Left–Right Self-Placement in the expected directions (see Fig. SM2 and SM3 in the Online Appendix). Thus, it is possible that there are other dividing lines between left and right, which coexist with the acceptance of inequality criterion in Europe. Further research is needed (especially in contexts outside of Europe) to test these and other competing explanations, as well as the acceptance of inequality criterion.

Discussion and conclusion

This paper provides new empirical evidence demonstrating that acceptance of inequality is associated with right-wing self-placements in both Eastern and Western Europe, in contradiction to earlier research. Importantly, the relationship between acceptance of an inequality and right-wing self-placements is stronger in contexts where the inequality is a salient issue, as the theory predicts. This finding may extend to contexts outside of Europe, where other issues/inequalities than the ones examined in this paper are salient (Jou and Dalton 2017).

In Europe, this relationship is seemingly especially important for two dimensions: the economic and the immigration/ethnic dimension. The broader theme in Europe in the twenty-first century is thus reflected in the results of this paper. The economic dimension has had an important role for the left–right dimension but has been increasingly challenged by the ethnic/immigration dimension, especially in Western Europe. Populist radical right parties have grown with refugee and immigration issues at the heart of their campaigns. This paper demonstrates that this pattern is seemingly very predictable. For example, if refugee immigration and multiculturalism become more important in East European countries where these issues are not very salient yet, then it likely follows that pro-immigrant attitudes will become better correlated with left-wing self-placements. We can be more certain of this if the issue becomes the main ideological issue.

The findings of this paper are important for the comparative study of the left–right dimension, but in particular for research on the left–right dimension in Eastern Europe. As previous research has pointed out, there is much more diversity in these countries’ left–right dimensions, compared to how the left–right dimension manifests itself in Western Europe. Nevertheless, this paper showcases that it is possible to find context-independent themes, even in this region. Thus, left–right politics in Eastern and Western Europe may share more similarities than previously thought, even if only one (i.e., the role of acceptance of inequality).

This study has some limitations, but the paper nevertheless has many benefits for future studies in political science as it adds to our knowledge of the left–right dimension. It is arguably the most popular and important single dimension of politics in representative democracies. Understanding whether there is a context-independent theory that can explain the difference between left and right in politics is therefore of utmost importance for political science. Such a distinction could potentially be an important explanatory variable. Understanding this dimension could furthermore improve citizens’ understanding of politics, informing the citizenry of what tends to structure their politicians’ ideological stances. All of this highlights the need for more research examining whether there is an underlying core disagreement of the left–right dimension.