This is an open access article distributed under the terms of the Creative Commons
Attribution License (The question of responsibility is a central issue in social policy (Dallinger, 2010; Dean, 2004; Jæger, 2006; Wright, 2012). Responsibility for social issues has been understood in different ways in European countries. Social security systems have been created for answering social risks and the needs of citizens. Some countries have emphasized the state’s responsibility for arranging social benefits and services for citizens, while other countries have underlined the responsibility as being that of the citizens themselves.
Scholars have been discussing the question of whether social political solutions follow citizens’ welfare attitudes or whether such solutions are leading those attitudes (e.g., Hills, 2002). On the one hand, it is assumed that social political systems are based on citizens’ social political attitudes (cf. Ervasti, Andersen, & Ringdal, 2012; Kumlin & Stadelmann-Steffen, 2014). On the other hand, it has been found that policy-making institutions influence citizens’ perceptions of social policy (Edlund, 1999). Thus, there are various perspectives on the direction and the mechanisms underlying the association between attitudes and social policy. Nevertheless, there is not much controversy about the existence of that association.
Scholars have also studied the factors that are associated with welfare attitudes. It has been shown that country-level factors, such as the economic situation, are associated with welfare attitudes (Rehm, 2011). Also, individual-level issues, such as demographic factors (Valdimarsdóttir, 2010), people’s living situation (e.g., Pfeifer, 2009) and political ideology (Calzada, Gómez-Garrido, Moreno, & Moreno-Fuentes, 2014) are associated with welfare attitudes. For instance, individuals who position themselves with the left wing political parties are more likely to support the state’s responsibility on social issues (Jæger, 2006). However, the analysis by Valdimarsdóttir (2010) shows that the connection is complex. Furthermore, Blekesaune and Quadagno (2003), for example, approached welfare attitudes as being a result of situational and ideological factors, which they studied at both the individual and country levels. Their results demonstrate that both situational and ideological factors are significantly associated with welfare attitudes.
Individualism versus collectivism can be described as one of the most significant cultural factors. According to Hofstede (1980), individualism is a social framework in which individuals are expected to take care of only themselves and their immediate families. In collectivism, in turn, people are expected to be loyal to the groups to which they belong. Various studies have suggested that individualistic values are associated with political solutions that emphasize citizens’ responsibility (Feldman, 1983, 1988), and collectivistic values, in turn, are associated with political choices that emphasize collective responsibility. On the other hand, the situation is not straightforward. For instance, Sweden is a typical example of an individualistic country (Hofstede Insights, 2018), and Sweden is also an example of a highly developed welfare system with broad universal benefits and services. In turn, Portugal is a country with a collectivistic culture (Hofstede Insights, 2018), but its social political system does not place as much emphasis on the state’s responsibility as with the case of Sweden (Esping-Andersen, 2015; Roosma, Van Oorschot, & Gelissen, 2013; Toikko & Rantanen, 2017). The relationship between country-level cultural values and welfare attitudes is thus a complex issue.
The present study focuses on the association between country-level individualism, political opinions, and citizens’ welfare attitudes. The relationship between individualism and welfare attitudes has been broadly studied, but less work has been conducted with multilevel analysis, in which citizens’ welfare attitudes are explained by the country-level individualism and individual-level factors as political opinions (cf. Feldman, Huddy, Wronski, & Lown, 2019; Feldman & Zaller, 1992; Kosonen, 1987; Shen & Edwards, 2005).
Key Concepts
Attitudes Towards the State’s Welfare Responsibility
Welfare attitudes are widely studied (e.g., Kreidl, 2000; Larsen, 2006; Valdimarsdóttir, 2010; Van Oorschot, 2000), but over the range of different studies, the exact object of a particular attitude is defined in different ways. On the one hand, attitudes can be studied as citizens’ general social political views on issues, such as redistribution and inequality (Cruces, Perez-Truglia, & Tetaz, 2013; Dallinger, 2010). For instance, some scholars have focused on the responsibility of the state (Pfeifer, 2009; Valdimarsdóttir, 2010), attitudes towards welfare policy (Blekesaune & Quadagno, 2003; Svallfors, 1997), and opinions towards social welfare services and social security benefits (Muuri, 2010). On the other hand, research can focus on citizens’ attitudes to welfare for specific target groups (e.g., the poor; Blomberg et al., 2013; Bullock, 2004), or on their opinions as to what kinds of target groups deserve benefits and services (Van Oorschot, 2000; Petersen et al., 2011).
In previous studies, various individual level factors have been found to be associated with welfare attitudes, such as sex/gender, age, educational level and income. As regards sex/gender, women support welfare state more than men (Blekesaune & Quadagno, 2003; Valdimarsdóttir, 2010). Age is a significant factor for welfare attitudes, but results are contradictory. For instance, Blekesaune and Quadagno (2003) found that older people’s attitudes to welfare are more positive than young people’s, but the opposite pattern has also been found (Ervasti, 2001). This may indicate that the relationship between age and welfare state attitudes is curvilinear (u-shaped), such that young and older people support the welfare state more than working age people (cf. Valdimarsdóttir, 2010).
Citizens are in favor of social policy if they expect to benefit from it. In this sense, they have a vested interest in supporting social policy. For instance, low level of education is associated with a higher support of the state’s welfare responsibility (cf. Valdimarsdóttir, 2010). Also people whose income does not cover their basic needs are more likely to defend social policies than those people having no economic difficulties (Davids & Gouws, 2013; Niemelä, 2008; Van Oorschot & Halman, 2000). However, the question of interest is complex, because those who contribute to the welfare state as taxpayers are likely to have negative attitudes towards social policy; on the other hand, while people do pay taxes, they also receive benefits from the welfare state.
Comparative studies have found significant national differences in welfare attitudes. For instance, Pfeifer (2009) showed that people have different views on the state’s responsibility for minimum income protection between European countries. The differences between countries are at least partly related to the economic situation at the country-level. Blekesaune (2007) found that lower employment rates are associated with the public opinion that a state should take more responsibility for economic provision and redistribution of income. Increased financial strain was also associated with stronger support for state responsibility for economic provision, but not for redistribution of income. Dallinger (2010) showed that the demand for redistribution decreases with national economic prosperity – even after controlling for social expenditure. Furthermore, Rehm (2011) argues that the country-level economic situation and the level of gross domestic product (GDP), the Gini index, and social expenditures are all important, because they modify the degree of social risk. Hence, welfare attitudes also seem to be associated with the type of welfare state. For instance, Andreß and Heien (2001) found that people within the Liberal Welfare State Model showed a comparatively low level of support for the welfare state, while those within the Conservative Model showed a medium level of support. The highest support was found in the Social-Democratic Model (cf. Larsen, 2008).
However, welfare attitudes have been found to be multidimensional. People have different attitudes towards the different dimensions of welfare state, such as the role of the state in different policy areas, its efficiency, and its intended and unintended outcomes. People are positive about some of these dimensions, and they are more critical about other dimensions. For instance, Roosma, Van Oorschot, and Gelissen (2013, 2014) distinguish between perceptions of welfare state performance (i.e., the ‘is’ or actual performance dimension) and preferences (i.e., the ‘should’ or wished for performance dimension). Furthermore, it has been found that the relationship between welfare attitudes and the dimensions differs between countries: Southern and Eastern Europeans have positive attitudes towards the role of state but a more critical attitude towards the welfare policy outcomes (Roosma, Van Oorschot, & Gelissen, 2013).
Political Ideology and Welfare Attitudes
Political ideology is an important variable to explain attitudes to welfare (e.g., Blekesaune & Quadagno, 2003; Feldman & Zaller, 1992; Kinder & Mendelberg, 2000; Skitka et al., 2002; Zaller, 1992). For instance, Feldman and Steenbergen (2001) emphasize that individuals’ support for social welfare policies can be explained by their general political values. Individuals who position themselves on the left-wing parties are more likely to support the state’s responsibility for social issues (Jæger, 2006). According to Blekesaune and Quadagno (2003), individuals’ egalitarian ideology is associated with their support of welfare policies. However, political ideology is a complex issue, which consists of the right-left dimension and the conservativism-liberalism dimension. Based on these dimensions, Haidt, Graham, and Joseph (2009) distinguish four clusters and name them as: social conservatives, religious left, libertarians, and liberal left.
Furthermore, political ideology is also a collective level issue, which influences welfare attitudes at the individual level. For instance, egalitarianism may be termed as a collective level value, which is associated also with collective responsibility in welfare attitudes. An egalitarian might be one who maintains that people ought to be treated as equals. However, there are different shades of gray in egalitarianism: from fundamental egalitarianism to relational egalitarianism. According to Schwartz (1992), the values are the basis for our attitudinal evaluations. We evaluate people and behaviors, and positively so if they promote or protect the values we share.
Political ideology may be approached as individual and country level issues. For instance, Blekesaune and Quadagno (2003) distinguish between individual level (e.g., political placement or leaning) and country level (e.g., egalitarianism) ideological factors, which affect welfare attitudes. In addition to having a direct effect on attitudes, political ideology also has a moderating effect on attitudes. For instance, Valdimarsdóttir (2010) found that income level correlates negatively with welfare attitudes which emphasize state’s responsibility. However, the relationship between income and attitudes is moderated by political ideology such that there is less attitudinal difference between high- and low income groups that hold a left-wing political ideology than between groups that hold a right-wing ideology. Overall, political ideology seems to be an important factor for attitudes, but its effects are complex and indirect (cf. de Koster, Achterberg, & van der Waal, 2013; Swank & Betz, 2003). In this sense, more research is needed where both individual-and country level situational and ideological factors are included.
Individualism as a Cultural Factor
In general, individualism and collectivism have been the focus of a large discussion in recent decades (e.g., Oyserman, Coon, & Kemmelmeier, 2002; Oyserman, Kemmelmeier, & Coon 2002). The concept of individualism can be defined in different ways. For instance, Schimmack et al. (2005) discuss the two main approaches in their studies of individualism. On the one hand, Hofstede’s (1980) classical analysis explores cultural differences between nations and assumes that individualism and collectivism are opposites at the country-level. On the other hand, individualism can be explored as an individual-level factor. For instance, Triandis and Gelfand (1998) assume that individualism and collectivism are independent factors while distinguishing between two forms of individualism (and collectivism): horizontal and vertical individualism (Singelis et al., 1995). Whereas horizontal individualism combines the concept of an autonomous individual with an emphasis on equality, vertical individualism accepts inequality and competition between individuals.
Studies in the field of social and political sciences have approached the concept of individualism with a narrow meaning, by explaining social issues from an individual perspective instead of a cultural perspective. For example, Bullock (2004) uses the concept of individualism by referring to the individualistic attribution for poverty. According to that author, the individualistic attribution correlates with the attitude of restrictive welfare policy. Also, Shen and Edwards (2005) in their study found that citizens’ attitudes towards strict work requirements are associated with economic individualism. Ervasti (2001) has suggested that both political ideology and individualism influence support for the welfare state. Further, he suggests that political ideology and individualism are associated together. However, Ervasti (2001) does not approach individualism as a country-level cultural factor.
In the present study, we explore the relationships between individualism, political opinions, and welfare attitudes. First, we focus on the association between country-level individualism and citizens’ welfare attitudes. The relationship between individualism and welfare attitudes has been broadly studied but the majority of those studies focus on individual level individualism (cf. Feldman & Zaller, 1992; Shen & Edwards, 2005). We assume that state’s welfare responsibility is less emphasized in individualistic countries (Rantanen, McLaughlin, & Toikko, 2015; Roosma, Van Oorschot, & Gelissen, 2013, 2014). Second, we assume that citizens’ political opinions are linked to welfare attitudes (e.g., Blekesaune & Quadagno, 2003; Calzada et al., 2014; Jæger, 2006). Individuals who lean to left-wing political views are more likely to support the state’s responsibility than individuals who lean to right-wing political views. Third, we study the relationships of political opinions, individualism and welfare attitudes (cf. Ervasti, 2001). We hypothesize that country-level individualism modifies the association between citizens’ political opinions and welfare attitudes. Specifically, we expect that there is a greater attitudinal difference between the supporters of right-wing and left-wing political views in individualistic countries than in collectivist countries.
Method
Hypotheses
In this comparative study, we examined citizens’ attitudes towards the state’s responsibility for social issues, such as unemployment, family services, and care for the elderly in 20 European countries. We ascertained which of the individual-level and country-level factors are associated with citizens’ attitudes towards the state’s welfare responsibility. In particular, we analyzed the relationship between the country-level individualistic culture and the individual-level welfare attitudes. Further, we investigated whether citizens’ political placement is associated with welfare attitudes, and how country-level individualism moderates the effect of political placement on welfare attitudes.
H1: Citizens’ attitudes towards the state’s welfare responsibility are associated with citizens’ political placements. Right-wing people emphasize less the state’s welfare responsibility than the left-wing people.
H2: Citizens’ attitudes towards the state’s welfare responsibility are associated with country-level individualism. In individualistic countries, citizens place a lesser emphasis on the state’s welfare responsibility than in collective countries.
H3: Country-level individualism moderates the effect of political placement on attitudes towards the state’s welfare responsibility. Specifically, we expect the association between political placement and welfare attitudes to be stronger in individualistic countries than in the collectivistic countries.
Data and Variables
The data are based on the 8th Wave of the European Social Survey (ESS, 2016). The data cover 23 countries, and for reasons of data availability, we include only 20 countries (37,743 respondents) in the analysis. The countries are: Austria, Belgium, Czech Republic, Estonia, Finland, France, Germany, Hungary, Iceland, Ireland, Italy, Lithuania, the Netherlands, Norway, Poland, Portugal, Slovenia, Spain, Sweden, and the United Kingdom.
We aimed to explore citizens’ general social political attitudes towards the state’s welfare responsibility. We thus define attitude to welfare in terms of a general meaning without a specific focus on a particular target group. The dependent variable is the sum variable of welfare attitudes, which was constructed from three items. The respondents were asked for their opinions towards the state’s responsibility in ensuring: 1) the standard of living of the aged, 2) the standard of living of the unemployed, and (3) sufficient childcare facilities for working parents. The variable was measured on scales ranging between 0-10, where 0 indicates that it should not be the responsibility of the state at all, and 10 indicates that it should be entirely the responsibility of the state. The reliability (using Cronbach’s Alpha) of the dependent variable is 0.683. We used Templeton’s (2011) two-step approach for testing the normality of the distribution.
In line with Blekesaune and Quadagno (2003), we address situational and ideological predictors at individual- and country-levels. Individual level independent variables are presented in Table 1 and country-level individual variables are presented in Table 2.
Table 1
Independent Individual Level Variables (Grand Mean Centered, Except Sex and Age)
| Variable | N | Minimum | Maximum | M | SD |
|---|---|---|---|---|---|
| Background factors | |||||
| Sex (0 = female, 1 = male) | 38705 | 0 | 1 | 0.49 | 0.50 |
| Age: Young (0 = working age, 0 = Old, 1 = Young) | 38709 | 0 | 1 | 0.14 | 0.34 |
| Age: Old (0 = working age, 0 = young, 1 = Old) | 38709 | 0 | 1 | 0.22 | 0.42 |
| Situational factors | |||||
| Education | 38184 | -12.86 | 41.14 | 0.00 | 4.05 |
| Income deciles | 30953 | -4.40 | 4.60 | 0.00 | 2.77 |
| Feeling about income | 38230 | -.88 | 2.12 | 0.00 | 0.79 |
| Ideological factors | |||||
| Political placement | 33809 | -4.95 | 5.05 | 0.00 | 2.20 |
Individual level background variables consist of participant sex (0=female, 1=male), and age which is studied with three categories (< 28, 28-64, > 65 years) because, according to previous studies, both young people (Ervasti, 2001) and older people (Blekesaune & Quadagno, 2003) emphasize the state’s welfare responsibility more than working age people (cf. Valdimarsdóttir, 2010). The individual-level situational factors consist of the number of completed years in fulltime education, and a household’s total net income decile in each country. The individual level ideological factor is focused on political placement on an 11-point scale (0 = left; 10 = right). The situational country-level variables consist of GDP, GINI, employment rate, health expenditure, and social expenditure. Country-level ideological factors are based on Hofstede’s dimensions of culture, which are constructed from cross-cultural surveys (Hofstede, 1980, 2011). Hofstede’s cultural framework has been applied in a wide variety of contexts, across the majority of behavioral science disciplines (see for instance, Bochner, 1994; Fernandez, Carlson, Stepina, & Nicholson, 1997; for criticism, see McSweeney, 2002). The latest version of Hofstede’s Cultural Values Model includes a total of six dimensions: power distance, individualism versus collectivism, masculinity versus femininity, uncertainty avoidance, long-term orientation versus short-term normative orientation, and indulgence versus restraint (cf. Hofstede, 2011; de Mooij & Hofstede, 2010). Each of the original factors has been expressed on a scale that runs from 0 to 100 (Table 2). Both individual- and country-level situational and ideological factors are grand mean centered.
Table 2
Country Level Situational and Ideological Factors in 20 Countries (Grand Mean Centered)
| Variable | Content of variable | Min. | Max. |
|---|---|---|---|
| Situational Factors | |||
| GDP | GDP per capita in PPS, Index EU28=100 | -37.43 | 78.57 |
| GINI | Gini coefficient of equivalized disposable income | -5.98 | 6.92 |
| Social Expenditure | Social expenditure on social protection, % of GDP | -12.93 | 5.47 |
| Unemployment | Total unemployment rate, % of active population | -37.43 | 78.57 |
| Health Expenditure | Health care expenditure, % per GDP | -3.18 | 1.63 |
| Ideological Factors (after Hofstede) | |||
| Femininity vs. Masculinity | A higher degree indicates that masculinity (such as achievements) is preferred in society, and a lower degree signifies that femininity (such as care) is valued. | -41.08 | 41.92 |
| Power Distance | A higher degree indicates that hierarchy is clearly established and is executed in society. A lower degree signifies that people attempt to distribute power. | -34.28 | 25.72 |
| Uncertainty Avoidance | A higher degree indicates that regulated behavior is valued. A lower degree in this index shows more acceptance of differing thoughts or ideas. | -39.40 | 36.60 |
| Long-Term vs. Short-Term Orientation | A lower degree (short-term orientation) indicates that traditions are honored. Societies with a high degree (long-term orientation) views pragmatic problem-solving as a necessity. | -27.00 | 32.00 |
| Indulgence vs. Restraint | A lower degree indicates that there is more regulation of people’s conduct and behavior. Societies with a high score allow free gratification of people’s own drives and emotions. | -35.39 | 26.61 |
| Individualism vs. Collectivism | A lower degree indicates that people are expected to be loyal to the group to which they belong (collectivism). A high score indicates a weak interpersonal connection among those who are not part of a core family (individualism). | -42.04 | 19.97 |
Note. Sources: GDP, GINI, Social Expenditure, Unemployment and Health Expenditure were retrieved from the Eurostat (year 2016). Cultural values were retrieved from Hofstede Insights (2018).
Data Analysis
We carried out the analysis as follows. We undertook the descriptive analysis by examining the degree of individualism and the mean of welfare attitudes at the country level. We conducted the variance analysis for confirming the difference in welfare attitudes between the countries. Also, we calculated the Pearson correlation between welfare attitudes and political placement. According to Cohen (1992), the effect size of the correlation is small when the coefficient is at least 0.1, medium when it is at least 0.3, and large when it is at least 0.5.
We performed the main analysis by following the procedure of multilevel modeling. In the first phase, we estimated the null model, and calculated the Intraclass Correlation Coefficient (ICC = 10.7%). In the second phase, we built the individual-level Random Intercept Model, and in the third phase we added the country-level variables to explain the variability in intercepts across countries (the country-level random model). In the fourth phase, we checked whether the slope varied across countries (the Random Slope and Intercept Model), and in the final phase we studied whether country-level individualism moderated the association between individual-level predictors and welfare attitudes within countries.
We added the variables to the analysis as follows. We added the individual-level situational variables into the model all at once, and then we removed the weakest (statistically non-significant) variables one at a time. Then, we added the next variables into the groups in the following order: the individual-level ideological variables, the country-level situational factors, and the country-level ideological factors. We used the Restricted Maximum Likelihood (REML) method as a way of estimating the parameters of the models. Heck, Thomas, and Tabata (2014) suggest to use REML in the case of small group samples, because “REML takes into account the loss in degrees of freedom due to the estimation of the P+1 regression coefficients in the model in order to obtain an unbiased estimation of the variance components” (cf. Snijders & Bosker, 1999). In the analysis, we allowed the intercepts to vary, and finally we tested a Random Slope Model. We weighted the data following the recommendations of the ESS.
Results
Country-Level Descriptive Analysis
First, we conducted descriptive analyses on the dependent variable. As can be seen in Table 3, the respondents’ attitudes towards the state’s welfare responsibility vary between the countries (Table 3). On a scale of 0–10, the lowest means of the sum variable were in the Netherlands (6.67), the United Kingdom (6.89), France (7.15), and Hungary (7.20). The highest means of the variable were in Iceland (8.28), Lithuania (8.20), Portugal (8.18), and Spain (8.09). Moreover, the variance analysis indicated a statistically significant difference between the countries, F(19, 38521) = 180.5, p < .001, η2p = .08. The lowest rates of deviation were in Finland (1.22) and the Netherlands (1.24), which indicates that citizens’ welfare attitudes are less polarized than in countries such as Austria (2.03) or the Czech Republic (1.67) where the rates of deviation were the highest.
Table 3
Descriptive Analysis of Welfare Attitudes, Individualism, and Political Placement
| Country | Means of welfare attitudes
|
Index for individualism | Means of political placement
|
Correlation of welfare attitudes and political placement | ||
|---|---|---|---|---|---|---|
| M | SD | M | SD | |||
| Entire data | 7.46 | 1.58 | 69.0 | 4.95 | 2.20 | -.125*** |
| Austria | 7.57 | 2.03 | 55.0 | 4.89 | 2.10 | -.114** |
| Belgium | 7.73 | 1.26 | 75.0 | 4.94 | 1.99 | -.198*** |
| Czech Republic | 7.59 | 1.67 | 57.0 | 5.24 | 2.20 | -.149*** |
| Germany | 7.40 | 1.46 | 67.0 | 4.49 | 1.87 | -.125*** |
| Estonia | 7.80 | 1.49 | 60.0 | 5.38 | 1.96 | -.084 |
| Finland | 7.78 | 1.22 | 63.0 | 5.62 | 2.09 | -.135** |
| France | 7.15 | 1.46 | 71.0 | 5.05 | 2.38 | -.193*** |
| Hungary | 7.22 | 1.67 | 80.0 | 5.90 | 2.50 | -.124** |
| Iceland | 8.28 | 1.28 | 60.0 | 5.18 | 2.22 | -.188 |
| Ireland | 7.20 | 1.51 | 70.0 | 5.22 | 1.91 | -.131* |
| Italy | 8.09 | 1.61 | 76.0 | 5.14 | 2.55 | -.047** |
| Lithuania | 8.20 | 1.60 | 60.0 | 5.39 | 2.02 | -.002 |
| Netherlands | 6.67 | 1.24 | 80.0 | 5.26 | 1.95 | -.151*** |
| Norway | 7.92 | 1.29 | 69.0 | 5.24 | 2.21 | -.181*** |
| Poland | 7.46 | 1.66 | 60.0 | 5.81 | 2.42 | .059** |
| Portugal | 8.18 | 1.51 | 27.0 | 4.64 | 2.48 | -.104** |
| Slovenia | 7.79 | 1.56 | 27.0 | 4.79 | 2.31 | -.021 |
| Spain | 8.09 | 1.66 | 51.0 | 4.46 | 2.17 | -.162*** |
| Sweden | 7.61 | 1.41 | 71.0 | 5.20 | 2.26 | -.176*** |
| United Kingdom | 6.89 | 1.48 | 89.0 | 4.88 | 1.81 | -.212*** |
Table 3 shows the levels of individualism and the mean of welfare attitudes by countries. The greater the degree of individualism, the less likely that citizens would emphasize the state’s welfare responsibility. For instance, the United Kingdom and the Netherlands are countries having high individualism, whereby the state’s responsibility is not emphasized as much as it is in other countries. On the other hand, Portugal and Slovenia are countries high in collectivism whose citizens also emphasize the state’s responsibility.
As far as the Pearson correlations between welfare attitudes and political placement are concerned, over the entire data, welfare attitudes and political placement were negatively correlated. This which means that the left-wing respondents emphasize the state’s welfare responsibility more than the right-wing respondents. However, the correlation differs between the countries, and the negative correlation cannot be found in all countries. The highest correlation coefficient was in the United Kingdom, which is also the most individualistic of the countries. However, even in this case the effect size of the correlation coefficient was small (0.1 ≤ r < 0.3) (cf. Cohen, 1992). On the other hand, in some countries, such as Slovenia, Estonia, and Lithuania, the correlation was not statistically significant, and in Poland the correlation was positive. In a European context, all these countries belong to the group of low individualistic countries. In contrast, the correlation was negative in Portugal, which is a low individualistic country.
Two-Level Regression Analysis
A two-level regression analysis was conducted for examining the factors which are associated with welfare attitudes (Table 4). According to the Null Model (M0), the proportion of variance in welfare attitudes that lies between countries is 10.7%, [ICC = 0.107], which suggests that welfare attitudes vary across countries. Then, the first model was constructed for examining the variability in intercepts across countries. The individual-level factors were added in two waves; the first wave was situational variables (Model 1) (sex, age, education, income) and the second wave was ideological variables (political placement) (Model 2). The full individual-level model (Model 2) explains 7.6% of individual-level variance and 23.2% of country-level variance. It can be seen that men have lower rates in welfare attitudes. Also, people with longer-term education and with a higher income have lower rates in welfare attitudes. Working age people compared to young and older people have lower rates in welfare attitudes. Also, right-wing respondents have lower rates in welfare attitudes which supports Hypothesis 1 and is consistent with the descriptive analysis (see Table 4).
Table 4
Effect of the Individual-Level and Country-Level Situational and Ideological Factors on Welfare Attitudes
| Factor | M0
|
M1
|
M2
|
M3
|
M4
|
M5
|
M6
|
|||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| b | SE | b | SE | b | SE | b | SE | b | SE | b | SE | b | SE | |
| Intercept | -.014 | .073 | -.179 | .071 | -.160 | .068 | -.141 | .082 | -.195** | .065 | -.195** | .056 | -.194** | .055 |
| Individual-level | ||||||||||||||
| Sex: Femalea | .068*** | .090 | .058*** | .011 | .058*** | .011 | .058*** | .011 | .060*** | .011 | .061*** | .011 | ||
| Age: Youngb | .087*** | .018 | .090*** | .018 | .090*** | .018 | .089*** | .018 | .087*** | .018 | .088*** | .018 | ||
| Age: Oldc | .071*** | .014 | .071*** | .014 | .071*** | .014 | .071*** | .014 | .070*** | .014 | .076*** | .014 | ||
| Education | -.010*** | .002 | -.012*** | .002 | -.012*** | .002 | -.012*** | .002 | -.015* | .005 | -.015* | .005 | ||
| Income | -.027*** | .002 | -.025*** | .002 | -.025*** | .002 | -.025*** | .002 | -.024*** | .002 | -.025*** | .002 | ||
| Political placement | -.053*** | .002 | -.053*** | .003 | -.053*** | .003 | -.054*** | .003 | -.053*** | .003 | ||||
| Country-level | ||||||||||||||
| GDP | -.001 | .003 | ||||||||||||
| GINI | -.017 | .030 | ||||||||||||
| Social expenditure | .003 | .030 | ||||||||||||
| Unemployment rate | .041 | .021 | ||||||||||||
| Health expenditure | -.036 | .156 | ||||||||||||
| Cultural dimension: Masculinity | -.002 | .004 | ||||||||||||
| Cultural dimension: Power Distance | -.006 | .007 | ||||||||||||
| Cultural dimension: Uncertainty | -.001 | .008 | ||||||||||||
| Cultural dimension: Long Term Orient. | .002 | .005 | ||||||||||||
| Cultural dimension: Indulgence | -.008 | .006 | ||||||||||||
| Cultural dimension: Individualism | -.010* | .004 | -.011** | .003 | -.011** | .003 | ||||||||
| Cross-level | ||||||||||||||
| Placement * Individualism | -.001*** | .001 | ||||||||||||
| Variance components | ||||||||||||||
| Individual level variance | .852*** | .006 | .799*** | .007 | .787*** | .007 | .787*** | .018 | .787*** | .007 | .780*** | .007 | .783*** | .018 |
| Between-country variance | .082** | .029 | .067** | .024 | .063** | .022 | .061 | .013 | .034* | .011 | .034* | .011 | .033* | .011 |
| Random Slope: Placement | .001* | .001 | ||||||||||||
| Random Slope: Education | .001* | .001 | .001* | .001 | ||||||||||
| Model fitting | ||||||||||||||
| -2LL (REML) | 95973 | 67531 | 67130 | 67154 | 67174 | 66958 | 67050 | |||||||
Note. N = 37664 in 20 countries. Source: European Social Survey, Round 8.
aRef. Male. bRef. Working age + old. cRef. Working age + young.
*p < .05. **p < .01. ***p < .001.
In the next phase, the country-level variables were added to explain the variability in intercepts across countries. In this case, the thesis was that the country-level situational factors (GDP, Gini, social expenditure, employment rate, health expenditure) and ideological factors (Hofstede) would impact the remaining variability in welfare attitudes between countries. It can be noted that controlling for the other predictors in Model 3, the country-level situational factors did not statistically significantly affect welfare attitudes. According to Model 4, individualism was statistically significantly associated with welfare attitude, as is assumed in Hypothesis 2. In individualistic countries, citizens’ attitudes towards the state’s responsibility were lower than in collective countries. Other cultural values were not statistically significantly associated with welfare attitudes, neither were country-level situational factors. Model 4 explains as much as 58.2% of country-level variance.
Since individual level factors are significantly related to welfare attitudes, we explored whether the slope of individual-level predictors varies randomly across countries. In Model 5, the slope variance of education (Wald Z = 1.97, p < .05) and political placement (Wald Z = 2.09, p < .05) were significant, which suggests that the slopes vary across countries in the sample. In this sense, it is necessary to hold reservations on the results of the full individual-level model (M2) which suggests that higher education and right-wing leaning political views were associated with lower rates of welfare attitudes.
In simple models, the effects of predictors on the outcome do not depend on the values of other predictors in the model. In the next step (M6), we examined the cross-level interaction: whether individualism as a country-level ideological factor moderates (i.e., enhances or diminishes) the association between individual-level predictors and welfare attitudes within a country. According to Model 6, the interaction between individualism and political placement is statistically significant. The interaction indicates that the relationship between political placement and welfare attitudes depends on the value of country level individualism. This means that overall individual-level political placement is negatively associated with attitudes towards the state’s welfare responsibility, but the higher degree of individualism slightly enhances the effect of political placement on welfare attitudes. Hence, the relationship between political placement and welfare attitudes is moderated by country level individualism such that there is less attitudinal difference between the left-wing and right-wing of political views in the collectivist countries than in the individualistic countries (Figure 1).
Figure 1
The association of political placement and the state’s welfare responsibility on the low and high levels of individualism.
The simple slopes for the association of political placement and state’s welfare responsibility on the low and high levels of individualism are -.07 and -.11, respectively (both ps < .001). After including the interaction term in the model, there was still significant variability to be explained, both within countries (Wald Z = 113.52, p < .001) and between countries (Wald Z = 2.38, p < .017). The full model (M6) explains 8.1% of individual-level variance and 59.8% of country-level variance.
Discussion
With the current study, we aimed to examine the association between individualistic values and citizens´ attitudes towards the state’s welfare responsibility (welfare attitudes). The study shows that individualism, as a national value, is associated with attitudes towards the state’s welfare responsibility. We found an association between individualism and welfare attitudes both between countries and at the individual level. Hence, individualism seems to be a factor that needs to be taken into account when considering citizens' attitudes towards the state's welfare responsibility. Further, individualism moderates the influence of political placement on welfare attitudes, although the effect is relatively weak. Overall, the association of political placement and welfare attitudes is complex and their connection is not visible in all countries.
The state's welfare responsibility is strongly supported in European countries, but there are also variations related to various individual-level factors. The state’s welfare responsibility is more supported by women than by men. Also, people with lower levels of education and with lower income offer more support for the state’s welfare responsibility than their counterparts, which is consistent with Blekesaune and Quadagno (2003) and Valdimarsdóttir (2010). However, it is necessary to be cautious about the effect of educational level because, according to the present study, it varies from country to country. Further, overall, left-wing people support the state’s welfare responsibility more than right-wing people (e.g., Jæger, 2006). The state’s high welfare responsibility has been a traditional aim for left-wing parties (Esping-Andersen, 1990), which can also be found in citizens' current welfare attitudes. Further, the state’s welfare responsibility is also more supported by individuals who are out of the workforce. Consistent with this, we found that young and older individuals support state’s responsibility for welfare to a higher extent than working age individuals (cf. Blekesaune & Quadagno, 2003; Ervasti, 2001). This is understandable because young and older people might need more social benefits and services than the group of working age people (Davids & Gouws, 2013).
Welfare attitudes are also associated with country-level individualism, which seems to have a significant role in explaining welfare attitudes between countries. The countries with a high degree of individualism have lower support for the state’s welfare responsibility than the countries with a low degree of individualism. The United Kingdom and the Netherlands are examples of individualistic countries, where citizens do not emphasize the state’s role in welfare as much as in less individualistic countries, such as Portugal and Slovenia. The countries which belong to the extremes of the degree of individualism are also far from each other in terms of the welfare state performance. For instance, in the Netherlands welfare state performance (actual level of social benefits and services) is much higher than in Portugal. In this sense, welfare attitudes might be linked with the actual level of social policy and its benefits and services. Hence, it is essential to take into consideration what Roosma, Van Oorschot, and Gelissen (2014) found that in the countries where the welfare state's performance is higher, the citizens’ support of social policy is lower than in the countries where the welfare state's performance is lower (cf. Toikko & Rantanen, 2017). However, the degree of individualism cannot be directly related to the high or low welfare state performance in social policy because, for instance, the Nordic countries seem to have a high welfare state performance but also relative high support (i.e., preferences) for the state’s welfare responsibility. In this sense, the question concerning the relationships between the degree of individualism, citizens’ opinions on welfare state performance and citizens’ preferences is a complex issue, which needs a great amount of further research.
Unlike many other studies (e.g., Dallinger, 2010; Rehm, 2011), in the present study we did not find an association between welfare attitudes and situational country-level factors such as gross domestic product (GDP), the Gini index, health expenditures, social expenditures or the rate of employment. In our study, we thus found that welfare attitudes are associated with national-level cultural values, not with practical social political solutions. Individualism as a cultural value refers to an individual's right, freedom and self-realization. If these values are followed in social policy, it leads to social and political solutions, where the state's welfare responsibility is aimed at being as limited as possible. However, although this kind of result has not been broadly reported before, it is somehow related with the results of Saguy, Gruys, and Gong (2010) who showed that social-structural explanations and political solutions are not emphasized in the United States, a highly individualistic country, as much as they are in France, which is less individualistic.
Furthermore, the present study indicates that citizens’ political placement plays a key role in explaining welfare attitudes. Overall, individual-level political placement is negatively associated with the state’s welfare responsibility. The more right-wing leaning the political placement, the less support for the state’s measure on welfare issues (e.g., Feldman & Steenbergen, 2001; Jæger, 2006). Also, in the present study we found that the relationship between political placement and welfare attitudes depends on the degree of individualism. In this sense, country-level individualism reduces attitudes towards the state’s welfare responsibility but, on the other hand, it emphasizes the significance of political placement or leaning. Compared to less individualistic countries, in more individualistic countries, there is a greater attitudinal difference between the supporters of right- and left-wing political parties concerning the state's welfare responsibility. However, people who support left-wing parties also emphasize the state’s welfare responsibility in individualistic countries.
The results of this study have some limitations. First, although the present study explains the variance between the countries quite well, it does not explain the variance within the countries so well. Although the results are statistically significant, there still remains some significant variability that is unaccounted for. In other words, although welfare attitudes in individualistic and collectivist countries are different, there are also significant variations in attitudes towards the state’s welfare responsibility in each country. Second, in this study individualism is considered as a one-dimensional concept. This premise, which is based on how Hofstede (1980) conceptualized individualism, does not take into account the distinction between horizontal and vertical individualism (Triandis & Gelfand, 1998), However, this distinction is very important socio-politically. Indeed, as aforementioned, Nordic welfare states are typical examples of horizontal individualism, while the United States is an example of vertical individualism. In the USA support for welfare is much weaker than in the Nordic countries. Horizontal and vertical individualism also affect the welfare attitudes in different ways (Rantanen, McLaughlin, & Toikko, 2015). Unfortunately, the scale for horizontal and vertical dimensions of individualism was not available in the ESS data. Third, the present study approaches attitudes at the general level, without making a distinction between the different groups of recipients and their deservedness with respect to benefits and services. It is obvious that citizens’ attitudes vary according to the target groups, such as the elderly, the unemployed, and families with children (e.g., Van Oorschot, 2000; Petersen et al., 2011). Fourth, in terms of political ideology standing, this study focused on the left-/right-wing dimension, without taking into consideration the current discourse on right-wing populism or immigration. Yet, they are both at the center of current societal discussions on attitudes and values.
Conclusion
The question of welfare responsibility is a central issue of social policy. The European discussion on welfare responsibility has focused on individual and state responsibility. The social and political debates themselves might be polarized but national social political systems are based on combinations from both extremes of the responsibility dimension. Indeed, the Nordic welfare state model, for instance, is based on the responsibility of state, the system also includes elements which emphasize the duties of individuals. Similarly, the Liberal welfare state emphasizes the responsibility of individuals, but the state also has a significant role in securing the welfare of citizens. In this sense, there is not a pure social policy system in Europe which is based solely on the responsibility of individual or state. Social political systems rest on the acceptance of citizens to whom the systems are expected to serve and, on the other hand, citizens' attitudes reflect the existing social policy systems. There are between-country differences (e.g., the Portuguese versus the British) as regards the attitudes towards the state's welfare responsibility. However, in each country there are those people who support the welfare responsibility of the state and there are other people who support the welfare responsibility of the individuals.
According to the present study, the degree of national level individualism is a significant factor for understanding citizens’ attitudes towards the state’s welfare responsibility. The higher the level of individualism, the lower the probability that the state's welfare responsibility is supported. The results suggest that social policy is based on national cultural values more than one may assume. Individualism is a more significant factor than other country-level situational factors, such as national wealth or investment in social and political systems.
National solutions for social policy are deeply political choices. In the more global picture, citizens who support the left-wing political parties emphasize more the state's welfare responsibility than citizens who support the right-wing political parties, although the connection is weak. According to the present study, the degree of individualism modifies the association of political placement and welfare attitudes. In individualistic countries, the attitudinal difference between the right-wing and the left-wing views is greater than in less individualistic countries. The degree of individualism is a kind of lens which effectively portrays the division of European attitudes toward the state's welfare responsibility. In this sense, polarization towards the welfare state is more likely in individualistic countries than in collectivistic countries (cf. Brito Vieira, Carreira da Silva, & Pereira, 2017).
The present study emphasizes the importance of national cultural values in social policy making. In recent decades, the development of individualism has been a strong point in Western countries. In this sense, it is understandable that Liberal social policies have received attention in Europe. However, in the future, collective values may become more pronounced, for instance, due to immigration. If collective values gain more attention, this will very likely be reflected in future social policy making.