Another chance? Concerns about inequality, support for the European Union and further European integration

ABSTRACT Following the 2007–2008 financial crisis, it was expected that the economic downturn and the widening of economic disparities would produce lower support for the European Union (EU) and its continued integration. Using the 2009 European Election Study (EES) data in 27 EU member states, we find that citizens who see greater economic instability and insecurity, regardless of their current economic status, lower their support for the EU as it is but increase their support for continuing integration. Substantively, this suggests that EU citizens may offer the EU another chance to tackle this timely issue and counterbalance market-generated inequality.


Introduction
n the whole, the economic and financial crisis has led to a decline in quality of life [in Europe].' (Eurofound Report 2012) The core aim of the European Union (EU) has been to create widespread and relatively equitable economic growth in Europe for the sake of continental stability and tranquillity. For the most part, this has been achieved and our understanding of support for the EU by its citizens has reflected this such that personal and aggregate economic growth produces higher support for the EU project (see Rohrschneider and Loveless 2010). However, following the 2007-2008 economic crisis, there is a greater percentage of people who may not be objectively 'poor' but feel themselves to be at a heightened risk of economic adversity owing to the rising inequality and economic problems in both their country and the EU (Eurofound Report 2012). What this produces in terms of changes in citizens' support for the EU project remains unclear.
Using the 2009 European Election Study (EES) survey dataset in 27 EU member states, we show that the importance individuals place on addressing inequality is positively correlated with support for further European integration but not for the EU status quo. This is provocative and informs our understanding of popular support of the EU. What is striking about these findings is that there is little evidence that this effect is a direct function of economic 'winning and losing' via individuals' socioeconomic status (Gabel 1998a(Gabel , 1998b. Rather, economic losing and its assumed negative effects on support for the EU may be more nuanced and widespread. Clearly disappointed with the current performance of the EU (in lower support for its ongoing performance) are those citizens who express concern about economic conditions, particularly those who see greater economic instability and insecurity regardless of their current economic status (i.e., 'new losers'). At the same time, these citizens also appear to be more supportive of the EU in the future.
We propose the following understanding. While the EU has long been an economic project coupled with a democratic normative framework, the full range of evidence here suggests that support for the EU moves with a desire for democratic politics to play a stronger, more stabilizing role in the economy. Following the economic crisis, even if the EU is seen to have failed to create adequate economic and social opportunities or has provided these prospects in an unequal manner, EU membership may still represent assurance that both economic and political institutions can and will work effectively. Thus, in addition to traditionally identified groups of 'losers', these 'new losers' appear to be supportive of the EU as a means to buttress democratic power at both national and supranational levels based on a belief that democracy is the key mechanism to combat market-generated inequalities. This potentially suggests that the EU should reflect citizens' preferences for fairness and justice via strong and effective democratic institutions that function to diminish excessive market distortions.

Attitudes about European integration: inequality and support
There is a large literature which seeks to explain individuals' attitudes about European integration (Loveless and Rohrschneider 2008). We turn to the beginning of this literature in which the earliest individual-level models were based on a utilitarian approach, which hypothesized that individuals assess EU membership based on their social position, assessment of their own economic experience and expectations of their nation states' market economy. Under this principle of utilitarianism, there are 'winners' and 'losers' of European integration (Gabel 1998a(Gabel , 1998b. These groups, defined by individuals' social position (economic positions, education, occupational skills and proximity to borders), distinguish between those who would benefit, or indeed lose, from further integration (Gabel and Whitten 1997).
Simply, individuals with high socioeconomic status and low socioeconomic status perceived, and evaluated, the EU differently. The 'winners' regarded this as the expansion of the market and thus opportunities, a perspective particularly pronounced among the more cosmopolitan, mobile and flexible; while the 'losers' viewed the EU in the form of diminishing welfare via a function of declining patterns of national redistribution within expanded markets (Brinegar and Jolly 2005;Gabel 1998a).
Such economic orientations to the EU have been historically strong determinants of EU support. Therefore, in the context of the financial crisis of 2007-2008 and continuing economic recession, we might expect a resurgence of the importance of this work (Gabel 1998a;Gabel and Palmer 1995;Gabel and Whitten 1997;Inglehart 1997;Inglehart et al. 1991). However, we suggest that the relationship between individual economic inequality and support for the welfare state at times of economic risk is such that social insurance programmes are attractive not only to those with low incomes, but also to those facing higher risks (this is one basis for inter-socioeconomic group coalitions in support of social protection). This inter-relationship between different socioeconomic groups (i.e., 'winners' and 'losers') of risk and income shapes preferences for redistribution, and thus a renewed role of the state producing the expectation that socioeconomic status as an indicator of economic security would retain or perhaps increase its strength as a predictor of support for the EU project. Note, of course, that the effect of socioeconomic status, in the earliest Gabelian conceptualization, is predicated on the EU being perceived as a guarantor of economic growth (Loveless 2010;Rohrschneider and Whitefield 2006;Tucker et al. 2002) in which individuals would be able to determine what continued integration means to them particularly as 'winners' and 'losers'.
Thus, 'losing' need not necessarily be restricted to (negative) changes in individual income. The distribution of the goods of society may have put more people in a more fragile socioeconomic state or made them feel that they are in a more precarious socioeconomic state. Changes in how individuals benefit from and share in aggregate economic growth or in how those changes are distributed can be reflected in individuals' concerns about inequality. In other words, while many individuals may not be de facto 'poor' or 'losers' in a strict economic definition, they may feel closer to such a state and driving their concern is either being or simply feeling more economically fragile. That is, being 'poor' and the 'fear of becoming poor' can have similar impacts on subsequent life choices (Hammerström and Janlert 1997). These are Europe's 'new losers'.
This concern may be captured by the perception that one is not allocated the 'fair' benefits of society, and is thus likely to affect how people think about the current political and economic status (Kreidl 2000;Osberg and Smeeding 2006;Verwiebe and Wegener 2000;Wegener 2000). If so, the perceived improvement or deterioration of their own social economic wellbeing stands to be a potentially effective determinant of their support for ongoing integration, as well as an evaluative filter through which to assess the EU in its current form. Democratic institutions are the key institutional mechanism available for most citizens to combat excessive and inevitable market distortions (Bollen and Jackman 1985;Reuveny and Li 2003;Szelenyi and Kostello 1996). If the means of democracy are seen to be impartial and fair, citizens having received what they wanted or nottend to accept the outcome, thus producing legitimacy (Rohrschneider 2002). Therefore, democracy can be seen to provide some protection from the inherent inequality, or perceived inequalities, of the market by serving as impartial arbiters of generic social welfare, reducing the effects of market driven inequality.
To bring this in to our understanding of citizens' support for the EU, we link individuals' concerns about inequality to variation in their level of support for the EU through the relationship inequality has to both democratic political institutions and the market. We do not argue that EU citizens want or even prefer an alternative arrangement with political democracy and a free market, but rather that these citizens want the market and democratic institutions to both work effectively (Loveless 2010;Rohrschneider and Whitefield 2006). That is, it is more fruitful to conceive of the market and democracy as mutually reinforcing mechanisms such that markets can produce better economic outcomes for a greater number in conjunction with strong and efficient democratic institutions.
For example, if an economy as a whole provides high living standards and dynamic economic development, individuals often accept relatively high, objective levels of inequality (Bollen and Jackman 1985;Jackman 1975;Jost et al. 2003). Therefore, the balance of market-generated inequalities and effective democratic institutions is plausible as in states with strong democratic political institutions; citizens see these as a bulwark against excessive, inequalities (Bollen and Jackman 1985;Reuveny and Li 2003;Szelenyi and Kostello 1996;Whitefield and Loveless 2013). Thus, when economies fail, democratic political institutions must work. Therefore, we propose that in the wake of the ongoing crisis, the EU may be seen as a potential guarantor of democracy that can, in one of its many functions, combat market-driven inequalities. This expressed desire is manifest in citizens' support for the EU project and its continuation.
Thus, this analysis is not aimed at upending Gabel's work but rather expanding on the definition of 'loser'. We expect both 'traditional' winners and losers of EU integration to reflect long-standing preferences and concern respectively for the EU. However, at the same time, we expect that those concerned about economic conditions to reflect a cross-current of determinants of support for the EU project. Given the state of current research, we suggest that those who have been moved toward greater economic insecurity (i.e., 'new losers') see democracy as the key mechanism to combat market-generated inequalities. Therefore, rather than reflect the anti-EU of the sociodemographic 'losing' profile, our expectation is that those concerned about economic insecurity, the potential 'new losers', are more supportive of the EU as a means to buttress democratic power at both national and supranational levels. We hypothesize that as the level of individuals' preferences for inequality to be addressed increases, they are more likely to support the EU and its continued expansion.
We note, however, if concerns with inequality are driven by the desire to see the economic costs of inequality mitigated through democratic means, support for EU performance will only benefit if the EU is perceived to have performed well. If the EU is perceived to have not performed well, we expect to see a loss of support for current performance. Similarly, this suggests that concerns about inequality could also impact attitudes about the ongoing EU project, inasmuch as the EU is seen to be able to continue (its good) or improve (its bad) performance and support for a continuation of the EU project is likely.

Methodology
We use the European Election Study of 2009 to examine support for the EU (see the Appendix for all data and variables). There are a variety of indicators for support (Loveless and Rohrschneider 2008). Therefore, in order to test the robustness of our approach, we include several of them here: (1) EU membership is good or bad; (2) satisfaction with the EU; (3) EU enlargement is good or bad; and (4) support for more or less European unification. In doing so, we further group the first two ('EU good or bad' and 'satisfaction with the EU') in order capture respondents' orientation to the EU as it is in its current form. Simultaneously, we group the latter two ('EU enlargement good or bad' and 'support for more or less unification') to represent respondents' feelings of the deepening or intensifying of EU membership; in other words, support for the EU's progress toward its goal of continued integration. 1 Table 1 shows the covariation of these dependent variables. While we can see that they vary from one another, none of these variables are substantively correlated with the other and we can conceive of two conceptual groups: the EU as status quo and EU enlargement. We do not formally impose this conceptualization on the data or theory other than to point out the possible distinctiveness of the variables between the EU in its current form and 'deepening' the EU.
There are numerous approaches to the understanding of EU support (see Appendix). The standard model includes communication (social communication, watching mass media and interest in politics); identity (feelings about being described as European and fear of immigrants); ideological congruence and institutional performance (including retrospective and prospective sociotropic economic evaluation, as well as normative preferences for the market and satisfaction with democracy); sociodemographic variables (including self-reported social class, subjective standard of living, age, gender, ideology and education) (Loveless and Rohrschneider 2008). While cross-sectional data run the risk of endogeneity, here we have adhered as tightly as we can to the existing framework developed around determinants of support for the EU to make our findings applicable to a wide number of studies. For our central independent variable of inequality, respondents were asked how they consider the importance of addressing inequality using the straightforward question: 'income and wealth should be redistributed towards ordinary people' (recoded so that agreeing is the higher score). 2 The models are first run as pooled regressions controlling for cross-national differences using country dummies. For the dependent variables 'EU membership: good or bad'; 'satisfaction with democracy in EU'; and 'EU enlargement: good or bad', we used ordered Logit. For 'attitude to European unification', we use ordinary least squares (OLS) regression.
Across all four dependent variables (see Table 2), we see no effect of social communication, and the use of mass media is statistically significant for both 'EU: good or bad' and the enlargement variables. Individuals' interest in politics is statistically significant and positively correlated with the EU as good or bad and for more unification. In congruence with some of the most longstanding theories of support for the EU, both prospective sociotropic economic evaluations and satisfaction with (national) democracy are consistently, positively and statistically significantly predictive of support for both the EU and enlargement. The same can almost be said for retrospective sociotropic economic evaluations, which fails to reach significance for 'EU good or bad'. Similarly, for preferences for market economies, it fails to reach significance for support for EU enlargement. Thus, the ideological congruence and performance variables are nearly uniformly positive and as expected. Overall, the 'identity' variables are the most consistent predictor of support for the EU and EU expansion. They are in the directions expected, with 'feeling  European' positively associated with support for the EU and 'fear' negatively associated across all models.
For the sociodemographic variables, we note that the reliance of Gabel's 'winners and losers' hypothesis on static demographic variables may be breaking down. The richer, younger and more educated no longer appear to see the EU and further integration as a net positive. For one, there is only a slight gender effect for 'EU: good or bad'; education is only positively correlated with the attitude that the EU is good (rather than bad); and only in the case of the 'EU: good or bad' are younger respondents more supportive of the EU. For both 'satisfaction' and 'EU expansion', older respondents are more supportive. Social class shows up in both 'EU' and EU expansion' although self-reported 'standard of living' is positively and consistently correlated with three of the support for the EU variables (excluding 'EU enlargement: good or bad').
Ideology produces little consistency other than those who subscribe to the most left ideological positions are less likely to support the EU as it is and more likely to support enlargement (although not more unification). At the same time, we see that those who self-identify at the farthest right positions are also supportive of further unification and, in contrast to those on the left, of the EU as it is. These are the least clear of the findings, although one might posit that the negative support of the left and support from the right for the EU, status quo, is indicative of a clear market position (recall the strong positive effects of individuals' market preference and prospective sociotropic economic evaluations). Thus, we might consider that both the left and right might support expansion such that the left would prefer to see more democracy and the right a continuation of the EU's apparent market profile.
This conclusion is not unreasonable given the individual-level findings for the inequality variable. As individuals agree with the notion that income and wealth should be redistributed towards ordinary people, support for the EU as it is drops; yet, support for further integration increases. Therefore, in accordance with our theoretical expectation that individuals' attitudes towards addressing inequality should increase support for the EU the findings also support the theoretical notion that individuals regard the EU as a means to reinforce substantive democratic governance (at both the national level and within the EU itself). Finally, although not presented in Table 2 for space considerations, almost all of the countries' dummy variables were statistically significant, suggesting possible cross-national variation in support for the EU and its enlargement. As a result, we propose to investigate further with a multilevel analysis.
We first examine the intra-class correlation coefficient (ICC) to find whether there is a higher degree of resemblance between lower-level units belonging to the same level-two unit (i.e., random effects) suggesting country-level effects. However, for each of the dependent variables, none of the ICCs in the empty models reach higher than 10 per cent; although the proportional reduction of variance in models which include all of the individual-level variables for both 'EU membership' and 'satisfaction with democracy in the EU' are 17.1 per cent and 30.4 per cent respectively, suggesting possible significant cross-national differences. 3 This is not surprising. The standard model of support for the EU has increasingly required controlling for national-level variables that play a role in shaping crossnational differences in popular perceptions of the EU (Kritzinger 2003). Popular perceptions of the EU are accustomed by national institutional factors in that, as countries move from lower levels of both economic and political performance, citizens move from economic to political criteria (Anderson 1998;Norris 2000;Rohrschneider 2002;Rohrschneider and Loveless 2010;Sanchez-Cuenca 2000).
Yet, what the EU citizenry may regard as excessive inequality may have little to do with inequality per se but depend on whether the economy as a whole provides high living standards and dynamic economic development (Bollen and Jackman 1985;Jackman 1975). As seen above, the extent that EU member states have relatively wealthy economies, or economies that have relatively high rates of growth, demonstrates that inequalities will be perceived as legitimate and acceptable, autonomous of objective levels of inequality, particularly in conjunction with robust democratic political institutions. If this is the case, then we would expect individuals' concerns about inequality to also vary by national characteristics.
Therefore, we include macro-level predictors that are theoretically linked to individual-level variation in support for the EU with macro-economic and political performance. We include gross domestic product (GDP) per capita (2009), the World Bank's Government Effectiveness measure (Kaufmann et al. 2009), and, given the focus of this analysis here, the Gini Index of income inequality (2009). Doing so, this not only tests the role of nationallevel effects on support for the EU and integration, but more importantly serves as a test of the robustness of the above individual-level findings.
The second set of models are run as multilevel, fixed-effect models that allow for random variation on the interceptas the mean level of support controlling for the individual-level effectsby macro-indicator (i.e., country). Note that in Table 3 the substantive results of using OLS regression were the same as ordered Logit (as in Table 2); thus we present the OLS regression output to ease interpretation. 4 In Table 3, self-reported class, standard of living, ideology, age and education are the same as in the previous models (see Table 2). The remnant gender effect from above is now completely missing. The performance variables are the same and remain strong, as do both identity variables and the communication variables. For ideology, we acknowledge that the variable used for concerns about inequality potentially includes respondents' support for redistribution, and thus there is the potential for an endogenous relationship with respondents' left-right ideological self-identification. To address this, we have added an interaction of respondents' ideological identification and countries' levels of income inequality. In doing so, we see that despite the left's apparent disappointment at the EU as it is, these same identifiers are more supportive in countries with higher inequality. However, we fail to find any effect for right identifiers for the EU as it is, as well as for both ideological identifiers for the EU as an ongoing project. However, this is less worrisome, as expectations about future performance (in the case of redistribution) are not a theoretical determinant of EU support. Controlling for cross-national effects, we find that, in every case, as government effectiveness increases across countries, the mean level of support for the EU and its expansion decreases. For EU support, this corresponds to some extent to existing work in which the quality of country-level democracy inversely affects support for the EU (although we see no corresponding effect from GDP per capita; see Rohrschneider and Loveless [2010]). For further integration, this confronts the competencies argument that pits the EU against national governments' abilities to manage emergent economic challenges. We are unable to completely unravel this here; however, we examine this further in Table 4. At the same time, changes in the level of aggregate  income inequality (i.e., Gini index) appears to reduce somewhat support for EU enlargement. At the individual-level, individuals' concern about addressing inequality is nearly exactly the same as before (see Table 2), only the negative coefficient for 'satisfaction with democracy in the EU' has disappeared (it is only three one-hundredths of a percentage point away from statistical significance, i.e. p < 0.06). In other words, controlling for cross-national differences in economic performance, democratic institutional performance and national levels of income inequality, the findings here are nearly unchanged, indicating an empirically robust finding at the individual-level offering clear generalizability. What this suggests theoretically is that individuals may prefer a stronger EU presence (i.e., 'unification', 'enlargement') that does not exist in the status quo. This is suggestive of a stronger state role and one that corresponds to a stronger affinity for the EU: an overall assessment of the EU and more unification/enlargement.

Discussion
The post-crisis period had effects on the economic welfare of EU citizens almost immediately (Eurofound Report 2012). If we understand the EU as primarily a market promoter via integration of national economies, it is reasonable to expect that those pushedor who perceived themselves and others as pushedtoward a more fragile personal economic condition might be more critical of the EU and ongoing integration. Kriesi et al (2008) argue that economic competition led by changes in the American economy, cultural diversity and the competition between national governments and perceived encroachment of supranational politics have driven European societies in the theorized directions of Gabel's initial contribution over the past decades of EU expansion. However, it seems that 'losers' are not only losers in continued integration but also in the reduction of states' public sector capacity and political willingness to continue the welfare state. The findings here comport with this highly plausible understanding in three ways. First, those that want inequality to be addressed appear to be open to further integration, while being at the same time dissatisfied with the current EU performance in this context. Individuals' concerns with inequality depress support for the EU, suggesting that the perceived 'democratic deficit' continues (Rohrschneider 2002). However, popular dissatisfaction with the functioning of democracy tends to produce a desire for more, rather than less, democracy (Dalton 2004;Norris 2000). We find that support for EU integration via individuals' concerns about addressing inequality suggests a strong connection between the strong democratic enforcement that the EU could potentially offer. 5 While not tested directly here, this suggests that the EU's response to the crisis has been disapproved, but that the EU has a positive possible role to play. Whether this is one that supersedes the nation state or whether it is one of backing the collectivity of the European project is not clear. One might posit that the EU is being called upon to address inequality substantively in addition to national action or in the presence of national inaction.
We have no direct way to assess the answer to whether the EU or national governments are perceived as primarily responsible for the stabilization of markets and domestic/international economics. However, in the most recent Eurobarometer (No. 81, Spring 2014 [European Commission 2015]), respondents were asked, 'In your opinion, which of the following is best able to take effective actions against the effects of the financial and economic crisis?' 6 The responses are set out in Table 4.
By including previous Eurobarometer data (EB78, EB79, and EB80 [European Commission 2015Commission , 2016European Commission and European Parliament 2015]), we demonstrate that there is little variation between the EU and national governments, and therefore, in the minds of EU citizens, this question may not have a clear-cut answer. This suggests that the EU and the national governments may have 'another chance' to prove to EU citizens their willingness and ability to deal with this. Doing so would provide legitimacy to the responsible institutions. However, the lack of change and no evident lead contender in the minds of citizens underscores an ambivalence, as both national governments and the EU are followed closely by the G20 and the IMF (and the aggregated other, none and don't know category).
Second, it is not simply those that find themselves in more precarious economic positions whose concern about inequality affects their support of the EU project. It appears that evaluations are not only economic but also sociotropic (Rohrschneider and Loveless 2010). That is, the system can be seen as too unfair, thus making inequality representative of this as individuals' assess societal differences in access and opportunity to the EU, rather than their own access. It is possible that a citizen could support the notion that income and wealth should be redistributed and at the same time be satisfied with the level of inequality in one's country. However, it is likely to be a small portion of the population compared to those who are worried about inequality and support redistribution, as we are aware that concerns about inequality are strongly related to support for redistribution (Corneo and Gruner 2002;Finseraas 2012;Kenworthy and McCall 2008;Rehm 2009). The question here is a generic, normative preference for the project itself rather than a preference for specific policy outcomes. 7 Thus, it appears that the recent economic crisis has affected support for the EU and its continuation. There is a more widespread concern about inequality and the role of the EU (lower support for the EU as it is), as well as optimism for the project (support for the ongoing project) following the economic crisis.
This more general concern about addressing inequality leads us to the third and final way our findings point towards a better understanding of support for the EU. By linking higher levels of concern for addressing inequality with lower support for the EU as it is and higher support for further EU integration, it is not unreasonable to draw a preliminary conclusion that the EU is perceived as a meaningful enforcer of democratic principles of the EU and the member states to deal with excessive market distortions (Rohrschneider and Whitefield 2006). In conjunction with the social justice literature that points to the underpinning values of fairness and justice via strong and effective democratic institutions and processes that drive perceptions of inequality in societies, EU membership may be more than mere economic integration in the minds of many citizens. Membership may instead represent assurance that both economic and political institutions work effectively.
An alternative to our theory might suggest that citizens concerned with inequality may believe that the market basis of the EU is part of the reason why inequalities exist. This is intuitive, as support for the EU has ebbed and flowed over the past several decades, with overall support declining (Roth et al. 2013). There has been, however, no clear drop in support since the onset of the economic crisis. Despite the EU's original role as a facilitator of an expanding market, its current and expanded role includes the prospect as a guarantor of democracy for many member states. In the context of the competing effects of EU-level democracy and liberal market economy, there is work to suggest that individuals can, and indeed do, distinguish among EU-level competencies (Caldeira and Gibson 1995;Rohrschneider 2005), suggesting that other EU-level entities such as the European Central Bank, rather than the entire EU project, may be implicated for undesired economic outcomes. While we cannot address this directly here, this question is nonetheless an important one for future research.

Conclusion
We find that individuals' support for the EU to address inequality is independently and strongly correlated with negative support for the EU as it is and positive support toward a deepening of EU integration. This finding is both seldom contingent on individuals' socioeconomic location, making it a common explanation of support for the EU, as well as normatively supportive of stronger democratic institutional performance. This in turn allows us to examine the changing nature and role of the EU in the eyes of EU citizens given new economic realities.
Citizens' support for the EU has long relied on the perception of the EU's political and economic performance (Anderson and Reichert 1996;Carrubba 1997;Eichenberg and Dalton 1993;Gabel 1998a;Gabel and Whitten 1997;Rohrschneider and Loveless 2010). The EU's institutional response to inequality, regardless of whether it is economic or political, directly shapes support for the EU. We know that perceptions of inequality and redistribution are strongly driven by desires for democratic institutions to 'do more' (Bollen and Jackman 1985;Reuveny and Li 2003;Szelenyi and Kostello 1996;Whitefield and Loveless 2013). Thus, the theory that 'new losers' are worried about their economic status and about the future prospects of institutional responses fit neatly into the larger framework of economic and political performance evaluations. If the EU is seen as currently under-responsive, there is the hope that it can 'do more' in the future. This would appear as a loss of support for the current performance of the EU, but hopeful of positive support in the futureas our findings have demonstrated.
This suggests that, following the recent economic crisis, citizens' attitudes and orientations to the EU may incorporate salient and timely issues related to the dramatic and trying economic changes, such as inequality. Yet inequality not only heightens individual level concerns about economic stability, it has also been demonstrated in other contexts to force demands on democratic politics (Castillo and Zmerli 2012;Whitefield and Loveless 2013). Citizens' concerns with inequality appear to implicate concerns with the effective function of the EU as a stabilizer of democratic governance. This is not necessarily new, in that both comparative democratic and economic performance has long been a source of citizens' orientations to the EU. However, it does appear to suggest that citizens want democratic political institutions to be available and effective in addressing inequality. The findings further suggest that some of the effectiveness should originate from the EU. Whether this comes as a function of the EU itself in terms of policy responses, or the reinforcement of substantive democratic national governance, remains unclear and is worthy of further examination.

Notes
1. Exploratory factor analysis shows a weak but two variable latency. Available from authors. 2. All models were tested for multi-collinearity. 'Addressing inequality is not a proxy for other values and/or positions. We point out that it is not specifically correlated with ideological positions (r = -0.13; p ≤ 0.001, N = 23,647); the role of the state in the market (r = 0.07; p ≤ 0.001, N = 27,069); market preferences (r = -0.05; p ≤ 0.001 N = 25,130); or individuals' self-reported standard of living (r = -0.13, p ≤ 0.001, N = 26,567). Thus, 'addressing inequality' here is largely independent of many commonly assumed predictors. 3. ICC in empty models: 'EU membership: good or bad': 9.9 per cent; 'Satisfaction with democracy in EU': 4.3 per cent; 'EU enlargement: good or bad': 8.5 per cent; and 'Attitude to European unification': 6.0 per cent. For 'EU enlargement: good or bad' and 'Attitude to European unification', the proportional reductions of variance in full models are 6.9 per cent and 9.1 per cent respectively. 4. The multilevel ordered Logit results are available from the authors. 5. Whether this will transpire remains unanswered, as we lack the capacity to replicate our model with more recent data. For example, using the 2014 Election Study (EES) data is problematic and analytically hazardous owing to changes in the nature and availability of many of the necessary variables. Only half of the dependent variables are included, more than half (56 per cent) of the independent variables are operationally different, while 19 per cent are missing entirely. This introduces substantial opportunity for both measurement error and omitted variable bias. Therefore, the model was not conducted with EES 2014 data. 6. Question QC3a in the survey, see page 27 of 'First Results' source: http://ec. europa.eu/public_opinion/archives/eb/eb79/eb79_first_en.pdf. 7. Note that the question here asks, 'Income and wealth should be redistributed towards ordinary people' making no reference to country, party, or specific policy.
or neither good nor bad? RC: Good thing, Neither, Bad thing. Reverse coded to make findings intuitive, DK to missing. Satisfaction with democracy in EU (q. 85): How satisfied are you, on the whole, with the way democracy works in the European Union? Very satisfied, Fairly satisfied, Not very satisfied, or Not at all satisfied. Refused and DK to missing.
Attitude to European Unification (q. 80): Some say European unification should be pushed further. Others say it already has gone too far. What is your opinion? Please indicate your views using a 10-point-scale. On this scale, 1 means unification 'has already gone too far' and 10 means it 'should be pushed further'. What number on this scale best describes your position?
EU enlargement is good or bad (q. 83): In general, do you think that enlargement of the European Union would be a good thing, a bad thing, neither good nor bad. RC: Good thing, Neither, Bad thing. Reverse coded to make findings intuitive, DK to missing.
Satisfaction with democracy in [country] (q. 85): How satisfied are you, on the whole, with the way democracy works in [respondent's country]? Very satisfied, Fairly satisfied, Not very satisfied, or Not at all satisfied. Refused and DK to missing.

Inequality
Address inequality (q. 63): Income and wealth should be redistributed towards ordinary people. RC: Strongly Agree, Agree, Neither, Disagree, Strongly Disagree. Reverse coded to make findings intuitive, DK/NA recoded to Neither.

Communication
Social communication (q. 18): How often talk to friends/family about election? Often, Sometimes, Never. Reverse coded to make findings intuitive, DK to missing.
Mass media (q. 16, q. 17, q. 20): How often watch program about election on TV (q. 16)/read about election in newspaper (q. 17)/ look into website concerned with election (q. 20)? Often, Sometimes, Never. Reverse coded to make findings intuitive, DK to missing. Simple arithmetic sum of the three.

Political interest
Interest in politics (q. 78): To what extent would you say you are interested in politics? RC: Very, Somewhat, a little, not at all. Reverse coded to make findings intuitive, DK to missing. Government effectiveness: Reflects perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government's commitment to such policies (Kaufmann et al. 2009).