Support for an EU‐wide social policy? A country‐comparative analysis of unemployment perceptions

Wang, Hequn Article — Published Version Support for an EU‐wide social policy? A country‐ comparative analysis of unemployment perceptions International Journal of Social Welfare Provided in Cooperation with: John Wiley & Sons Suggested Citation: Wang, Hequn (2024) : Support for an EU‐wide social policy? A country‐ comparative analysis of unemployment perceptions, International Journal of Social Welfare, ISSN 1468-2397, Wiley Periodicals, Inc., Hoboken, NJ, Vol. 34, Iss. 1, https://doi.org/10.1111/ijsw.12687 This Version is available at: https://hdl.handle.net/10419/313730 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by-nc-nd/4.0/ ORIGINAL ARTICLE Support for an EU-wide social policy? A countrycomparative analysis of unemployment perceptions Hequn Wang Universität Hamburg, Hamburg, Germany Correspondence Hequn Wang, Universität Hamburg, Hamburg, Germany. Email: [email protected] Abstract This article explored the influence of unemployment perceptions on attitudes towards an EU-wide social policy that guarantees a minimum standard of living for the poor across 18 European countries. The article relied on a theoretical framework that highlights the interaction among economic self-interest, ideology, and perceptions. Using data from Eurostat and the European Social Survey, the results show that Europeans with more negative perceptions of national unemployment or the living conditions of the unemployed were more likely to support an EU minimum income scheme. This association was particularly strong among individuals with non-egalitarian values or right-leaning ideology and remained relatively consistent across different national contexts. Additionally, support was stronger in countries with poor economic and welfare conditions. Overall, the findings reveal a high perceived legitimacy among Europeans for implementing a policy measure that aims to tackle poverty in the EU. KEYWORDS EU social policy, minimum income, perceptions, public opinion, social Europe, unemployment INTRODUCTION The European Union intervenes in the social policy domain to deal with inequality and to combat poverty and social exclusion within the EU, leaning towards a ‘social Europe’(Martinsen & Vollaard, 2014). Still, one in five Europeans was at risk of poverty or social exclusion by 2022, that is, having an income below the national poverty threshold, being socially deprived or excluded from the labour market as defined in the EU (Eurostat, 2023). In addition, the situation varied largely between countries, ranging from 12% in Czechia up to 34% in Romania. A related question is whether there is general support for the EU's intervention in social policy for the sake of the well-being of Europeans and perhaps stronger support in countries with a higher share of a poor population. Although to date social policy remains mainly within the area of national authorities, political debates on fostering solidarity and redistribution beyond national borders are ongoing. These debates have gained increased Abbreviations: ESS, European Social Survey; GDP, gross domestic product; ICC, intraclass correlation coefficient; SD, standard deviation. Received: 13 December 2023 Accepted: 12 June 2024 DOI: 10.1111/ijsw.12687 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2024 The Author(s). International Journal of Social Welfare published by Akademikerförbundet SSR (ASSR) and John Wiley & Sons Ltd. Int J Soc Welf. 2025;34:e12687. wileyonlinelibrary.com/journal/ijsw 1of14 https://doi.org/10.1111/ijsw.12687 importance in recent years due to major global crises such as climate change and the COVID-19 pandemic. This has reinforced the political relevance of studying public opinion on social policy at the EU-level. Against this background, this article analysed EU citizens' attitudes towards the introduction of an EU-wide social benefit scheme that guarantees a minimum standard of living for all poor people in the EU. At the national level, minimum income benefits are already in place in all member states to reduce poverty and social exclusion. However, in most countries, the benefit levels fall far below national poverty thresholds and vary significantly across countries. Therefore, promoting a minimum income benefit scheme at the EU-level would address inequalities within and between member states. Potential alternative policies, such as a framework on adequate minimum wages that the EU has settled on by the end of 2022, aim to enhance minimum income protection for employed persons in the EU. Instead of addressing in-work poverty specifically like that, an EUwide minimum income benefit would provide minimum income protection to all individuals in need, including the unemployed who live in poverty. In addition, such a policy measure would require rich member states to invest more financially, promoting redistribution at the EU-level. A thorough understanding of Europeans' support for the policy measure thus contributes to legitimizing further integration within the EU, and in particular, to achieving the EU priority to tackle poverty in general. While there has been a long-standing tradition of studying attitudes towards EU integration in general (e.g., Anderson & Reichert, 1995; Gabel, 1998; Gabel & Whitten, 1997), attitudes towards social-policy making at the EU-level have seldom been discussed explicitly in previous literature until the last few years (e.g., Baute & Meuleman, 2020; Beaudonnet, 2013; Burgoon, 2009; Gerhards et al., 2016; Roosma & van Oorschot, 2021). The recent studies demonstrated that support for EU social-policy making is generally stronger in member states with worse economic and welfare performances (e.g., lower welfare generosity or efficiency). While revealing a contextual impact on Europeans' attitudes, most studies left out how Europeans may perceive their country's conditions differently. Less is known about the specific role that individual perceptions of the national context may play in shaping attitudes. Perceptions at the individual level should be considered in addition to objective contextual factors, as people often exhibit significant misperceptions regarding national indicators, such as the level of unemployment, poverty or inequality (e.g., Cardoso et al., 2016; Choi, 2019; Kuhn, 2019). As a result, scholars found that support for redistribution at the national level is more strongly associated with individuals' perceptions of national inequality than with contextual factors (e.g., Bussolo et al., 2021; Gimpelson & Treisman, 2018). This prompts the question of whether perceptions also play a crucial role in explaining attitudes at the EU-level. The present study contributes to the existing literature by addressing the role of both objective contextual factors and individual perceptions of them in shaping Europeans' attitudes towards an EU-wide social policy in terms of a minimum income benefit scheme. This study centred on the perceptions of the unemployment rate and the living conditions of unemployed individuals within the country (also referred to as ‘unemployment perceptions’or ‘perceptions regarding unemployment’ hereafter). In this study, I adapted a theoretical framework that highlights the interaction among economic self-interest, ideology, and the perceptions. Empirically, I combined macro data on country-level economic and welfare performances with individual-level data from the European Social Survey (ESS) in 2016, which includes 18 EU member states. This article aimed to address the following research questions: First, how are attitudes towards an EU-wide minimum income scheme associated with unemployment perceptions differently in comparison to objective contextual factors regarding the economic and welfare performances of the EU member states? Second, to what extent can individuals' ideological motivations, such as egalitarianism or political ideology, influence the role of the perceptions in shaping the attitudes? Finally, do contextual factors also influence how unemployment perceptions relate to attitudes towards the policy measure? COUNTRY PERFORMANCES AND SUPPORT FOR THE EU'S INTERVENTION IN SOCIAL POLICY Since the Maastricht Treaty, the integration process within the EU has largely extended to the social dimension. The EU has begun to intervene even more in the social policy domain, since the economic crisis in 2007/2008 has increased economic disparities between European countries. For instance, different European Structural and Investment Funds have been carried out, such as the European Social Fund, supporting initiatives that aim to reduce unemployment and promote economic and social cohesion (European Parliament, 2022). However, from the perspective of European citizens, the Europeanisation of social policy resulted in a sudden loss of support for the EU back in the 1990s. There was growing public concern about the EU's intervention in the social policy domain having a negative impact on 2of14 WANG national social security systems (Ferrera, 2005), also described as the ‘Post-Maastricht blues’(Eichenberg & Dalton, 2007). Scholars argued that this reflects an economic self-interest among Europeans, as their attitudes towards integration within the EU depend on how it could influence their country and themselves (see also Gabel, 1998; Gabel & Whitten, 1997). Accordingly, opposition to the intervention by the EU would be particularly strong in member states with favourable economic and welfare conditions. In contrast, support would be strong in member states with unfavourable conditions to start with because their citizens may expect the EU's intervention in social policy to improve the level of social benefits in their country, from which they could possibly benefit (e.g., Burgoon, 2009). In line with the economic self-interest explanation, recent comparative literature found that support for the EU's intervention in social policy is overall negatively associated with national economic and welfare performances. Instead of referring to the EU's intervention as a whole directly, empirical studies mostly narrowed it down to one specific aspect where the EU could possibly intervene. For instance, focusing on three European countries, that is, Germany, Spain and Poland, Gerhards et al. (2016) showed that support for a specialised European social policy was strongest in Poland and weakest in Germany. In Mau (2005), support for EU welfare responsibility was much stronger in Mediterranean than in Continental countries, and was weakest in Scandinavian countries. Moreover, studies revealed that a higher level of national social protection relates to lower support for the welfare responsibility of the EU (Beaudonnet, 2013; Burgoon, 2009) and for a specific EUwide income benefit scheme (Baute & Meuleman, 2020). Similarly, the support is negatively associated with national economic conditions (Beaudonnet, 2013; Burgoon, 2009; Roosma & van Oorschot, 2021). Another growing branch of literature linked support for redistribution to the perceived, not the objective situation of a country. The main reasoning is that individuals tend to largely misperceive the actual state of their society with respect to national indicators such as the extent of unemployment, poverty, and inequality (e.g., Arin et al., 2021;Cardoso et al., 2016;Choi,2019; Kuhn, 2019;Kunovich,2012). Given these misperceptions, previous studies found that individual perceptions of, for example, national inequality, rather than actual contextual factors of a country, strongly associate with the demand for national redistribution (e.g., Bussolo et al., 2021; Gimpelson & Treisman, 2018). This raises the question of whether individuals' perceptions also have strongerexplanatorypowerthanobjective contextual factors in analysing attitudes towards redistribution at a supra-national level, that is, the EU-level. To date, only two empirical studies (Baute & Meuleman, 2020; Roosma & van Oorschot, 2021) drew attention to the impact of perceptions about national context on attitudes towards redistribution at the EU-level. Using the same individual-level survey data from the ESS in 2016, the results of both studies reveal that perceptions play a similarly important role as objective national economic and welfare conditions in explaining attitudes towards an EU-wide minimum income benefit scheme. First, the more negative the standard of living for unemployed or retired individuals in the country is perceived, the stronger the support for the policy measure. Second, individuals from countries with worse welfare performances tend to exhibit stronger support. However, neither of the studies discussed the gap between the objective national context and the individual perceptions or explicitly explained why perceptions should matter in addition to national context. 1 This article aimed to address this research gap, with a focus on individual perceptions regarding unemployment. In addition to the perceived standard of living for the unemployed addressed in the two previous studies, the present study included a second measure of unemployment perceptions, that is, the perceived national unemployment rate. Unemployed Europeans, who usually stand at the bottom of national income distribution (e.g., de Graaf-Zijl & Nolan, 2011; Gallie et al., 2003), should be among the primary target groups of an EUwide social policy that aims to improve the well-being of those in poverty. Therefore, perceptions related to both the proportion of unemployed individuals in the country and their living conditions are particularly relevant for analysing attitudes towards such a policy measure. This article accounted for both these individual unemployment perceptions as well as objective national economic and welfare performances, and investigated different mechanisms of how they affect attitudes towards an EUwide minimum income benefit scheme. 1 Instead, both studies shifted their focus to additionally analyse how individuals' expectations about the EU's impact on national social benefits shaped their attitudes. Baute and Meuleman (2020) demonstrated that these expectations fully accounted for the contextual impact on the attitudes, leaving aside perceptions of the national context. Roosma and van Oorschot (2021) investigated whether the role of the perceptions could be attributed to a ‘hope-or-fear reasoning’at the individual level: Individuals with negative (positive) perceptions of the living conditions of unemployed persons would expect an EU-wide social policy to improve (worsen) the level of social benefits in their country, and consequently, support the policy measure more (less) strongly. However, they found no empirical evidence to support this reasoning. SUPPORT FOR AN EU-WIDE SOCIAL POLICY? 3of14 EXPLAINING ATTITUDES TOWARDS AN EU-WIDE SOCIAL POLICY The economic self-interest explanation is a common approach for analysing country variations in attitudes towards redistribution and social policy at the EU-level, as previously discussed. At the individual level, empirical evidence indicates that it is not individuals' socioeconomic status that primarily shape these attitudes, but their ideological motivations (e.g., Baute & Meuleman, 2020; Gerhards et al., 2016). For instance, politically left-leaning individuals and individuals with egalitarian values are more in favour of an EU-wide social policy. Both economic self-interest and ideological motivations are traditionally considered as important theoretical explanations for the demand for national redistribution (see Alesina & Giuliano, 2011), before being applied to the supra-national level. A more recent study by Bussolo et al. (2021) proposed a concise framework that particularly addresses the interplay between the economic selfinterest, ideological motivations, and perceptions of national context in explaining demand for redistribution at the national level. Focusing on perceptions of national inequality, the basic theoretical assumption of the authors is that a stronger demand for national redistribution is associated with higher perceived national inequality, lower socioeconomic status, and a left-leaning ideology or similar views on social justice. Moreover, both self-interest and ideology play a crucial role in shaping individuals' perceptions. In particular, the authors argued that citizens of one country are exposed to the same national socioeconomic conditions, such as the extent of unemployment, poverty, and inequality within the country. Individuals' socioeconomic characteristics and ideology, including political orientation, determine how they may interpret such national context differently. Accordingly, better national economic and welfare performances lead to more positive perceptions among individuals regarding the national context, such as lower perceived national inequality. Conditional on the same national context, individuals with a lower socioeconomic status or a left-leaning ideology tend to exhibit more negative perceptions and regard a high level of unemployment, poverty or inequality as more problematic. In this article, I adapted the theoretical approach of Bussolo et al. (2021) to explain attitudes towards redistribution at the EU-level. As opposed to their focus on inequality perceptions, this article explored the role of perceptions regarding unemployment in shaping attitudes towards an EU-wide social policy. Overall, I assumed a similar mechanism with regard to how socioeconomic status, ideological motivations, and perceptions regarding unemployment within a country relate to attitudes at the EU-level. Individuals with a lower socioeconomic status are likely to support social assistance at the EU-level as an ‘imperfect substitute’(Burgoon, 2009, p. 433) to assure their living conditions. In other words, driven by the self-interest motivation, individuals will support social policy at the EU-level if they believe their country and themselves could benefit from it. Following this logic, individuals who perceive a high unemployment rate or a low standard of living for the unemployed in their country should be more in favour an EU-wide social policy. Therefore, I hypothesised that the more negative the perceptions regarding unemployment in the country, the more likely the support for an EU-wide social policy (H1). Considering both the economic self-interest and ideological motivations, a few previous studies (e.g., Armingeon & Weisstanner, 2022; Margalit, 2013) addressed a conflict of interest by analysing the interplay between individuals' socioeconomic status, such as income, and political ideology for shaping demand for national redistribution. They found that demand for national redistribution is overall stronger among left-leaning individuals and poor individuals. In addition, poor right-leaning individuals and rich leftleaning individuals also exhibit relatively strong demand, both of which are stronger compared to rich right-leaning individuals, though less strong compared to poor left-leaning individuals. With regard to the role of economic self-interest and how it could be influenced by ideology, their findings reveal that left-leaning ideology mitigates the influence of socioeconomic status on the attitudes. Accordingly, individuals with a leftleaning ideology or egalitarian values may face a similar conflict of interest when they hold positive perceptions regarding unemployment, that is, perceptions of a low unemployment rate or a high standard of living for the unemployed in the country. Although not addressed in their theoretical framework directly, Bussolo et al. (2021) provided empirical evidence that the left-leaning ideology diminished the role of perceptions in shaping attitudes towards national redistribution. I expected ideology has a similar damping effect at the EU-level, leadingtomysecondhypothesis,namelythe relationship between unemployment perceptions and support for an EU-wide social policy is stronger among individuals with a right-leaning ideology or non-egalitarian values (H2). As elaborated, national economic and welfare performances play a role in the formation of individuals' perceptions, determining their between-country variation. Accordingly, the influence of unemployment perceptions on support for an EU-wide social policy could be partly 4of14 WANG explained by objective contextual factors. Moreover, previous studies empirically confirmed that attitudes towards social policy at the EU-level are negatively associated with objective country performances. However, these studies did not analyse the role of the perceptions or discuss them in comparison with the contextual factors, as discussed in the previous section. Instead, they pointed out that contextual factors' impact on attitudes at the EU-level results from public concern about the EU having a negative impact on national social security systems (e.g., Baute & Meuleman, 2020; Roosma & van Oorschot, 2021). In general, rich countries and countries with generous welfareprovisionaremorelikelytobenegatively affected by the intervention of the EU, as competences aretobeshiftedfromthenationaltotheEUlevel (Baute et al., 2018; Ray, 2004). Consequently, driven by their economic self-interest, opposition from the citizens to the EU's intervention in social policy could be strong. Against this background, I assumed that individuals from countries with better economic and welfare performances are less likely to support an EU-wide social policy (H3). Considering both the literature that discussed the role of welfare state performances and the theoretical approach of Bussolo et al. (2021), objective contextual factors and subjective perceptions regarding unemployment have an impact on attitudes towards an EU-wide social policy that is partly independent from each other. Furthermore, there should be between-country variation in the role of the perceptions, depending on a country's actual economic and welfare performances. Given that the contextual factors could affect attitudes directly, support for an EU-wide social policy in a country may already reach a sufficiently high level due to the poor economic and welfare performances of the country. Consequently, perceptions regarding unemployment may no longer matter in that country. In contrast, attitudes towards an EU-wide social policy may vary strongly in rich countries or countries with generous welfare provision, depending on how individuals perceive unemployment within their own country. Following this logic, I expected that the relationship between unemployment perceptions and support for an EU-wide social policy is stronger in countries with better economic and welfare performances (H4). Figure 1depicts a full picture of the mechanisms with regard to the formation of attitudes towards an EU-wide social policy that I applied for studying the role of perceptions regarding unemployment. Arrows with black straight lines in the figure respectively denote the discussed hypotheses (H1–H4). Arrows with black dashed lines illustrate a more general interplay among socioeconomic status, ideology, and unemployment perceptions that I did not explicitly address in terms of hypotheses. DATA AND METHODOLOGY Data To examine the hypotheses empirically, the present study used individual-level data from the 8th round of the European Social Survey (ESS8) that was held in 2016 in 23 countries, including five non-EU countries (European Social Survey Round 8 Data, 2016). This study focused on EU member states to analyse attitudes towards an EU-wide social policy. The following 18 countries remained in the sample: Austria (AT), Belgium (BE), Czechia (CZ), Estonia (EE), Finland (FI), France (FR), Germany (DE), Hungary (HU), Ireland (IE),Italy(IT),Lithuania(LT), the Netherlands (NL), Poland (PL), Portugal (PT), Slovenia (SI), Spain (ES), Sweden (SE), and the United Kingdom (GB). After data cleaning, the sample consisted of 29,286 individuals in FIGURE 1 Formation of attitudes towards EU-wide social policy. SUPPORT FOR AN EU-WIDE SOCIAL POLICY? 5of14 total, with sample sizes ranging from 1114 in Slovenia to 2623 in Germany. 2 For the first time, ESS8 included a specific survey question about attitudes towards a potential EU-wide minimum income benefit scheme. The introduced proposal comprised the following three features: (1) ‘The purpose is to guarantee a minimum standard of living for all poor people in the European Union’; (2) ‘The level of social benefit people receive will be adjusted to reflect the cost of living in their country’; (3) ‘The scheme would require richer European Union countries to pay more into such a scheme than poorer European Union countries’. On a scale of 1 (‘strongly against’)to4(‘strongly in favour’), respondents were asked if they would be against or in favour of such a scheme. Notably, over 80% of the respondents fell into the middle two categories (‘against’or ‘in favour’). In 12 out of the 18 countries, below 5% (30 respondents or less in several countries) selected ‘strongly against’and ‘strongly in favour’, respectively. Therefore, I combined the first two and the last two of the original categories to generate a dichotomous dependent variable, denoting whether respondents would support an EU-wide minimum income benefit scheme or not. Eventually, 67% of all respondents were supportive of the policy measure. This article presented the results using the dichotomous variable as the dependent variable, but the results remained robust when using the original variable with four categories. I used two items in ESS8 on perceptions of the unemployment rate and the standard of living for the unemployed within the country as measures of unemployment perceptions. First, ESS8 asked respondents to estimate how many in every 100 working age people in their country are unemployed and looking for work. Respondents were provided with answering categories ranging from 1 (‘0–4’)to11(‘50 or more’) with an interval of 5 percentage points for their responses. They were then asked to assess the standard of living of people who are unemployed in their country on a scale of 0(‘extremely bad’)to10(‘extremely good’). Following the operationalisation in Baute and Meuleman (2020), egalitarianism was measured as a latent variable constructed from three items. Respondents were asked to report on a scale of 1 (‘agree strongly’)to5(‘disagree strongly’), how much they agreed or disagreed with each of the following three statements: (1) ‘Large differences in people's incomes are acceptable to properly reward differences in talents and efforts’;(2)‘For a society to be fair, differences in people's standard of living should be small’;(3)‘The government should take measures to reduce differences in income levels’. I recoded the responses for the construction of the latent variable so that the higher the score, the stronger the egalitarianism. Political ideology was measured by the item in which respondents were asked to place themselves on the scale of political left–right, ranging from 0 (‘left’)to10(‘right’). I reversed the coding of the item for the analysis. Due to a high number of missing values (13% of all respondents), I included political ideology only for additional analysis. According to the theoretical framework elaborated in the previous section, individuals' socioeconomic characteristics also have an impact on their attitudes towards an EUwide social policy, as depicted in Figure 1.Therefore,inthe regression analysis, I included as controls gender, age, and socioeconomic status. The latter was measured by four different indicators: household income, welfare dependency, employment status, and education. Since this study focused on attitudes towards an EU-wide social policy, I additionally controlled for respondents' emotional attachment to Europe. Summary statistics of all individual-level variables are provided in Table A1 in the Online Appendix. At the country level, I used Eurostat data in 2016 for measuring objective national economic and welfare performances. I calculated welfare efficiency as the ratio between the percentage of the population at risk of poverty before and after social transfers. 3 The efficiency scores can vary between 0 and 100. The higher the scores, the higher the welfare efficiency. The indicator shows the extent to which national social transfers protect citizens from poverty, which better measures the welfare performance of one country than other measurements such as the social welfare generosity scores (see Scruggs, 2014). For robustness checks, I used net expenditure on social protection benefits (as the percentage of gross domestic product [GDP] per capita) and at-risk-of-poverty rate (after social transfers) as alternative contextual factors that reflect the national welfare situation more generally. In addition, I used unemployment rate for measuring national economic performance and included GDP only for robustness checks because it correlated statistically highly with welfare efficiency (Pearson's correlation: 0.64). 4 Descriptive statistics of all contextual factors used in the main analysis and the robustness checks are provided in Table A2 in the Online Appendix. Table A2 also shows 2 For descriptive analysis, post-stratification weights and population size weights were used in combination to reduce sampling error and potential non-response bias and to adjust the sample size for each country in proportion to its population size. 3 I used the Eurostat data in which pensions were excluded from social transfers. The at-risk-of-poverty threshold was set at 60% of the national median equivalised household income. The Eurostat codes are ilc_li10 (before social transfers) and ilc_li03 (after social transfers). 4 Eurostat codes for net social expenditure, at-risk-of-poverty rate, unemployment rate, and GDP are spr_net_ben,ilc_li03,une_rt_a_h and nama_10_pc, respectively. 6of14 WANG statistics on the perceptions of the unemployment rate and the living standard of the unemployed at aggregate level (i.e., mean values for each country), which I additionally calculated for the descriptive analysis. Methodology I applied multilevel modelling to account for countrylevel variations in attitudes towards an EU-wide social policy according to national economic and welfare performances. Since the dependent variable is dichotomous, I estimated linear probability random intercept models to test the first three hypotheses. For cross-level interactions, random slope models are more adequate than random intercept models, since the former allows for variations at the cluster level with respect to the coefficients of lower-level covariates (Heisig & Schaeffer, 2019). However, my sample contained only 18 countries and a relatively large number of individuals per country. In the case of a small sample size at the cluster level, estimates of parameters referring to context effects are likely to be biased, and may cause additional reliability problems for including random slopes to test cross-level interactions (Bryan & Jenkins, 2016). To tackle this specific issue, one may consider two-step modelling (Achen, 2005;Heisigetal.,2017), which estimates crosslevel interactions in a more straightforward manner: in a first step, regressions are estimated separately for each cluster, and in a second step, coefficient estimates of variables of interest obtained from the first-step regressions are regressed on the cluster-level indicators. Given this context, I employed all three possible models, that is, random intercept models, random slope models, and two-step modelling, to test my last hypothesis regarding cross-level interactions. I aimed to draw robust conclusions by comparing the results of these different models. Country-level indicators were additionally standardised for better interpretation of the results. RESULTS Support for an EU-wide minimum income scheme: Descriptive statistics As mentioned, a majority (67%) of all respondents were in favour of an EU-wide minimum income benefit scheme. However, there were distinct country differences. Figure 2displays that the share of the supporters ranged from 48% in the Netherlands to 92% in Portugal. Support was stronger in Southern and Eastern European countries compared to countries in the North and West. Figure 2suggests that the richer the country, or the better the welfare performance of the country, the lower the possibility of supporting the policy measure. This assumption was backed up when looking at the correlation between the support and country characteristics regarding the economic and welfare performances. Figure 3a shows that the higher the unemployment rate, or the lower thewelfareefficiencyofonecountry,thehighertheshareof respondents in that country that were (strongly) in favour of an EU-wide minimum income benefit scheme. This suggests a negative relationship between national economic and welfare performances and support for an EU social policy. Consistent with the results regarding the objective country characteristics, Figure 3b illustrates a positive correlation between unemployment perceptions and support for the policy measure: the higher the unemployment rate was perceived at aggregate level, the higher the share of respondents of one country in favour of an EUwide minimum income scheme. The share of supporters was negatively correlated with aggregate-level perceptions on the standard of living of the unemployed in one country. In addition, Figure 3shows a clearer and stronger linear correlation between the latter two (Pearson's correlation: 0.82) compared to the correlations for the other three presented indicators. It seems that the average perception regarding the standard of living of the unemployed within one country could better explain country differences in the attitudes. To sum up, descriptive statistics suggest that national economic and welfare performances as well as perceptions regarding unemployment matter for analysing attitudes towards an EU-wide social policy like a minimum income benefit scheme. Further analyses are needed to investigate their underlying mechanisms. FIGURE 2 Support for EU-wide minimum income benefit scheme in 18 European countries. Using ESS8 data, 18 European countries, N=29,286; Weighted statistics. SUPPORT FOR AN EU-WIDE SOCIAL POLICY? 7of14 Impact of unemployment perceptions and country performances This section shifted the focus from aggregate-level perceptions to perceptions at the individual level. Following the theoretical model, I tested in this section the extent to which unemployment perceptions relate to attitudes towards an EU-wide minimum income benefit scheme, and how the role of the perceptions differs alongside individual ideology. In addition, I investigated how the attitudes are associated with national economic and welfare performances. The results are depicted in Table 1. Column (1) shows that, at the individual level, the probability of supporting an EU-wide minimum income benefit scheme increased significantly by 0.5 percentage points as the perceived national unemployment rate increased by one unit, that is, by 5 percentage points. The probability of supporting the policy measure decreased significantly by 1.6 percentage points as the perceived standard of living of the unemployed in one's country increased by one unit. Accordingly, the higher the unemployment rate or the worse the standard of living as unemployed was perceived by individuals, the more likely their support for an EU-wide minimum income scheme. This confirmed my first hypothesis that support for such a social policy is positively associated with more negative perceptions regarding unemployment (H1). The results of Column (1) additionally indicate a positive association between support for the policy measure and egalitarianism. Column (2) then shows that for respondents with non-egalitarian values, the probability of supporting an EU-wide minimum income benefit scheme increased significantly by 3 percentage points as the perceived national unemployment rate increased by one unit (5 percentage points). However, the more FIGURE 3 Support for EU-wide minimum income benefit scheme. Pearson's correlation between the share of supporters in one country and national performances in (a) is 0.54 for unemployment rate (left) and 0.65 for welfare efficiency (right). The correlation between the share of supporters and aggregate-level unemployment perceptions in (b) is 0.52 for the perceived unemployment rate (left) and 0.82 for the perceived living standard of the unemployed (right). All correlations are highly significant (p< 0.001). Weights applied for ESS data (N=29,286, 18 countries). 8of14 WANG