Coming together or growing apart? Globalization,class and redistributive preferences

Abstract

Globalization refers to a growth in international trade, the integration of financial markets and investment flows as well as greater labour mobility. It is also associated with greater market volatility and intensified competition. The general question we address is, does globalization strengthen or weaken public support for governmental redistribution? However, rather than assuming that globalization undermines or buffers welfare state support generally, we suppose that this effect might be class-specific. It might lead to a reduction in welfare state support for better-off classes while those at the lower end of the social strata may call for more governmental intervention. We test this hypothesis on the basis of cross-sectional data from the fourth round of the European Social Survey (ESS) (fielded 2008/2009). The results of multilevel analyses of 31 countries indicate that while welfare state support is generally lower in globalized countries, class cleavages in welfare attitudes are neither intensified nor diluted under conditions of globalization.

Keywords

Cross-country research economic globalization governmental redistribution social classes welfare state support

Introduction

From cross-country research we know that in European countries the welfare state’s major institutional arrangements receive broad support (e.g. Jæger, 2006; Mau, 2003; Svallfors, 2012). Moreover, there is evidence that over the course of the last three decades, there have been few signs of a fundamental welfare backlash or a legitimacy crisis of the welfare state (Mau and Veghte, 2007). Welfare attitude research has also demonstrated that a mixture of motives underpins attitudes and that self-interest as well as other factors such as moral obligations or reciprocity norms play a part (Mau, 2003; Van Oorschot, 2000). However, that does not mean that the welfare state is entirely uncontested in the eye of the public which is unsettled over issues like immigration, deservingness criteria or perceptions of the efficiency of welfare state intervention (Svallfors, 2012).

Scholar and lay wisdom widely accepts that one major challenge welfare states are facing is the process of globalization. Though there is a vast literature dealing with the socio-economic consequences of globalization and the globalization – welfare state’s nexus with regard to the fiscal power of the state (e.g. Brady, 2009; Brady et al., 2007; Schulze and Ursprung, 1999), much less is known about changing attitudes towards welfare state intervention in the course of globalization, in particular with regard to the role of class (for the positioning of political parties see Burgoon, 2012). For this reason, we raise the question: is globalization threatening welfare state support and can different levels of globalization account for different patterns in class attitudes towards welfare state policies?

Basically, there are two opposing (sometimes implicit, sometimes explicit) stories present in the literature. Some authors have argued that welfare states will not only survive globalization, but that a strong welfare state is a precondition to market openness as it makes market integration more acceptable and less costly socially (Garrett, 1995, 1998; Rieger and Leibfried, 2003; Rodrik, 1998). As trade openness and financial liberalization are likely to bring about greater economic volatility and the spread of social risk, they will generate higher demand for welfare protection leading to higher rather than lower welfare expenditure levels (Dreher, 2006a). The second story sees the welfare state at risk under conditions of globalization. It suggests that the redistributive welfare state may lose its public support, particularly if the government’s capacity to tax elites is feeble (Boix, 2003; Soifer, 2009). Also the end of national containment and the rise of insecurities are seen as undermining large-scale solidarity (Bauman, 2001). Thus, under conditions of porous national boundaries, the basis for national solidarity and the willingness to help others may be at risk of withering away (Koster, 2007). In this article, we address the nexus between economic globalization and welfare state support. However, we have reasons to assume that neither of the aforementioned stories holds for the population at large, but that attitudinal patterns differ by social class. In other words, rather than assuming that economic globalization per se affects attitudes towards the welfare state, we suppose that its effects are class-specific with the likely effect of enforcing class differences.

Our article is organized as follows: in the next section, ‘Welfare state support under conditions of globalization’, we discuss the nexus of globalization and attitudes towards the welfare state in more detail. The following section, ‘Data, method and analytical strategy’, is dedicated to our empirical setup. In the penultimate section, ‘Empirical findings’, we present our results. The article concludes with the discussion of our central question leading this study: is there a growing divide of welfare state support between social classes?

Welfare state support under conditions of globalization

Our study makes references to two intensively debated topics: first, globalization and its possible effects on welfare state support (the first sub-section below); second, the role of class as a category for different socio-economic standings in shaping the relation between globalization and attitudes towards redistribution (the second sub-section).

Globalization and welfare state support

The effect of globalization on national welfare systems is a mainstay of topical debates in the social sciences. Though economists have argued that globalization promotes growth and thereby positively affects the living conditions of highly developed countries (e.g. Baldwin and Winters, 2004; Dollar and Kraay, 2001; Dreher, 2006b), it has also been associated with increasing inequality, heightened competition and greater income volatility (e.g. Brady, 2009; Lee, 2006; Lübker, 2004; Standing, 2008; Taylor-Gooby, 2004).

These extensive debates cannot be unfolded here. Instead, we are interested in the link between globalization and public support for the welfare state: how does the former affect the latter? Along the lines of the globalization debate, two major accounts need to be mentioned here. The first one, sometimes called a ‘compensation thesis’ or ‘domestic demand thesis’, is represented by authors such as Garrett (1998), Rodrik (1998) and Rieger and Leibfried (2003). They consider the political standing of the welfare state positively related to globalization since state intervention can cushion and compensate for the harmful effects of globalization and make it politically acceptable. In other words, as world market integration and higher levels of economic openness come in tandem with greater volatility, there will be more demand for welfare state intervention. From this point of view, government protection performs an ‘insulation function’ (Rodrik, 1998: 1011) backed by popular support. Against this background, market insecurities generated by economic globalization may foster welfare state support from citizens and also from political parties which may ‘champion compensation as an answer to international trade’ (Burgoon, 2012: 627). From attitude research, we know that periods of economic strain as well as rising inequality tend to increase public support for the welfare state (Blekesaune, 2007; Dallinger, 2010; Finseraas, 2009; Jæger, 2011). With that said, we assume that greater market volatility and intensified competition associated with economic globalization may boost the demand for welfare state redistribution.

According to the second account, it is doubtful that economic globalization really strengthens commitment to the welfare state; on the contrary, it may put welfare state legitimacy at risk. Since the welfare state does not only distribute benefits, but also imposes costs that have to be borne collectively, globalization may increase opposition to the welfare state, in particular when the opportunities ascribed to globalization seemingly outweigh the potential benefits of a compensating state. Above all, the cost-sensible middle classes might be prone to some degree of welfare scepticism and prefer a leaner state as they trade off expected costs and benefits (Harms and Zink, 2003; Lupu and Pontusson, 2011). One could argue that significant sections of the population may not fear more intense (global) competition and experience globalization primarily as offering opportunities rather than risks. Thus, rather than staying committed to a generous and redistributive tax and transfer system, people may turn towards the free market and shift away from ‘an all-in-the-same-boat philosophy of shared risk towards a go-it-alone vision of personal responsibility’ (Hacker, 2006: 34). Finally, as the welfare state does not only serve the self-interest of citizens, but also organizes solidarity across the society, open borders and the integration into a global market may well affect the solidaristic commitment of citizens negatively and make them less regarding of others (Koster, 2007). This might be the case when social cohesion and the foundations of solidarity, which underpin the redistributive welfare state, are undermined by processes of globalization.

On the basis of the literature review, there are two narratives of the nexus between economic globalization and welfare state support which we put into two opposing hypotheses:

H1a: In highly globalized countries, support for governmental redistribution should be greater than in less globalized countries.

H1b: In highly globalized countries, support for governmental redistribution should be lesser than in less globalized countries.

Taking class into account: a polarization of welfare state support?

Patterns of class stratification, historically and at present, are driving forces of welfare state development (Korpi, 2006). Following Erikson and Goldthorpe (1992), classes stand for specific ‘experiences of affluence or hardship, of economic security or insecurity, of prospects of continuing material advance or of unyielding material constraints’ (p. 236). Recent research confirms that within and across countries, class is still a central structuring category for different levels of economic vulnerability and subjective economic stress (Mau et al., 2012; Whelan and Maître, 2008).

Welfare attitude research too has highlighted the role of social class for explaining differences in welfare state support. Here, it has been demonstrated that class affiliation matters with the lower classes being more pro-welfare state than the higher classes (Breznau, 2010; Brooks and Svallfors, 2010; Cusack et al., 2006; Svallfors, 2004). As stated by Svallfors (2011; see also Svallfors, 2006), ‘[c]lass is perhaps the most fundamental division with regard to welfare policies’ (p. 806–807) and these class differences are linked to self-interest factors determined by socio-economic positioning as well as certain worldviews and ideologies. Given the role of class for the welfare state, we ask whether conditions of globalization reshape the class–attitude nexus.

Overall, it is contested whether the concept of class will ‘survive’ globalization (Beck, 2013; Curran, 2013). While some authors suggest that globalization cuts across class differences and makes class a less and less suitable concept for understanding the relationship between inequality and politics, others maintain that class differences still matter and are even reinforced in the context of globalization (Goldthorpe, 2002). Accordingly, globalization may add significance to class in the sense that it may intensify differences in risk exposure, material conditions and ideological orientations.

Thus, there might not be a uniform or class-levelling impact of economic globalization. We assume that globalization affects social classes differently and that the link between globalization and welfare attitudes is class-specific. In the literature, authors highlight a split between ‘winners’ and ‘losers’ of globalization along class characteristics of human capital and labour market position (Scheve and Slaughter, 2004; Schulze and Ursprung, 1999; Walter, 2010). The ‘group of winners’ consists most notably of individuals in high class positions filling privileged and well-paid occupations, who are capable of profiting from the extended opportunities in life that come along with processes of globalization. By contrast, the ‘group of losers’ are primarily low-skilled workers who are confronted with higher risks of job loss and pressures on their income (Koster, 2010). Also, life-course research has shown that individual labour market risks associated with class are intensifying under conditions of globalization (Buchholz et al., 2009).

A growing body of literature looks directly at the link between socio-economic status and attitudes towards globalization, in particular aspects such as support for trade openness or state protectionism. Here, it is argued that these attitudes are by and large determined by potential distributional outcomes of globalization with those in higher positions and higher skills being more positive about globalization and low-skilled and unskilled workers being particularly threatened by economic openness and, hence, being more sceptical (Mayda and Rodrik, 2005; O’Rourke and Sinnott, 2001; Scheve and Slaughter, 2001).¹ This effect is likely to play a role in the context of welfare attitudes, too: those with higher class status are more likely to reap the benefits of globalization and thus less in need of a compensating welfare state, whereas those in lower class positions are more vulnerable and therefore demand more state intervention to cushion the negative impact of market integration (for a discussion, see also Dallinger, 2013).

There is also relevant literature dealing with the issue of social policy preferences and individuals risk exposures due to skill specificity, unemployment or their industries’ exposures to international competition. Based on a measure of offshorability, that is, the issue of whether a job can be performed in another country, Walter and Maduz (2009) found that ‘losers’ of globalization are more likely to support the welfare state, while ‘winners’ do not. Rehm (2009) did not find that people in sheltered sectors and in sectors with a competitive advantage are less supportive of the welfare state than those facing stronger competition. However, he found that the occupational level (skill specificity, occupation-specific unemployment rates) matters. In contrast, Finseraas and Ringdal (2012), who include subjective risk perceptions into their analysis, find weak support that firm-specific skills matter and also assert that those in sectors or occupations more open to trade are not more pro-welfare than other groups.

While this research is highly relevant and has nuanced our understanding of welfare state support, we do not, however, follow this track here. This literature suffers from an overly rationalist conception of what determines welfare attitudes leaving aside normative or social aspects which are important drivers of social policy preferences as well (Mau, 2003; Svallfors, 2006). Our focus on class instead takes into account not only that class and attitudes correlate but also that broader social and material conditions structure and determine class attitudes (Svallfors, 2007). We expect that the impact of globalization on welfare state support is moderated by class with lower classes being more inclined to support redistribution and higher classes more likely to defect from the redistributive welfare state under increasing conditions of globalization.

Thus hypothesis H2 reads as follows:

The higher the level of economic globalization of a country, the greater should be the class differences in welfare attitudes – with the lower classes being more pro-welfarist and the higher classes more opposed to welfare state redistribution.

Data, method and analytical strategy

Data

Our empirical analyses are based on data from the fourth round of the European Social Survey (hereafter: ESS4) which was conducted in 2008 and 2009.² The sample used encompasses 31 countries with response rates ranging from 45.7 percent (Croatia) to 78.7 percent (Cyprus) and with a mean response rate of 62.2 percent (see Table 1 for the list of countries). We restrict our analyses to currently employed and unemployed persons aged 18–65, and after list-wise exclusion of missing data, we have 27,869 cases for analyses with our first dependent variable and 28,010 cases for our second dependent variable.

Table 1.

Geographical distribution of preferences for income redistribution (in percent at country-level).

	Agreement	Disagreement	Mean		Agreement	Disagreement	Mean
Austria	63.7	18.4	2.70	Lithuania	80.3	6.4	2.98
Belgium	70.0	16.2	2.78	Netherlands	51.8	27.6	2.31
Bulgaria	74.8	14.2	3.00	Norway	59.5	19.6	2.51
Croatia	76.7	8.7	2.91	Poland	71.5	15.7	2.75
Cyprus	78.1	10.7	3.00	Portugal	89.7	3.3	3.20
Czech Republ.	51.4	27.0	2.37	Romania	78.4	5.8	3.08
Denmark	40.7	37.8	2.07	Russ. Federat.	68.1	10.5	2.88
Estonia	64.9	15.7	2.69	Slovakia	66.3	13.7	2.78
Finland	75.1	9.6	2.95	Slovenia	85.1	6.2	3.13
France	78.3	10.8	3.05	Spain	80.3	7.9	2.99
Germany	67.6	17.6	2.68	Sweden	62.0	13.1	2.65
Greece	91.7	1.6	3.39	Switzerland	67.1	18.7	2.66
Hungary	84.1	3.7	3.31	Turkey	88.5	5.2	3.29
Ireland	67.6	18.3	2.68	Ukraine	67.8	14.7	2.89
Israel	76.1	8.9	2.99	United Kingdom	57.6	24.4	2.46
Latvia	79.8	6.9	3.10	Average	71.0	13.7	2.84

Source: European Social Survey, 2008; n_i = 27,869; n_j = 31.

Respondents were asked to comment on the statement, ‘The government should take measures to reduce differences in income levels’. The recoded answer categories are: (0) ‘disagree strongly’, (1) ‘disagree’, (2) ‘neither agree nor disagree’, (3) ‘agree’ and (4) ‘agree strongly’. Categories (3) and (4) are collapsed into ‘agreement’ and (0) and (1) into ‘disagreement’. ‘Mean’ is a simple average value for all answers from a given country.

Table 2.

Geographical distribution of preferences for governmental support for the unemployed (means at country-level).

	Mean		Mean
Austria	6.42	Lithuania	8.39
Belgium	6.00	Netherlands	6.19
Bulgaria	7.06	Norway	7.37
Croatia	7.48	Poland	6.15
Cyprus	7.47	Portugal	7.24
Czech Republ.	6.13	Romania	7.72
Denmark	6.68	Russ. Federat.	6.95
Estonia	7.20	Slovakia	5.80
Finland	7.44	Slovenia	6.54
France	6.14	Spain	7.72
Germany	6.37	Sweden	7.38
Greece	8.47	Switzerland	6.25
Hungary	7.12	Turkey	8.01
Ireland	6.66	Ukraine	8.06
Israel	7.49	Unit. Kingdom	5.75
Latvia	8.36	Average	7.02

Source: European Social Survey, 2008; n_i = 28,010; n_j = 31.

Respondents were asked to comment on the statement, ‘It is governments’ responsibility to ensure a reasonable standard of living for the unemployed’. The answer categories are: (0) ‘should not be governments’ responsibility at all’ up to (10) ‘should be entirely governments’ responsibility’.

Both our dependent variables measure preferences for governmental redistribution. The first item captures preferences for income redistribution: ‘The government should take measures to reduce differences in income levels’ (Item B30). In answering this question, respondents could choose among five answers: (1) ‘agree strongly’, (2) ‘agree’, (3) ‘neither agree nor disagree’, (4) ‘disagree’ and (5) ‘disagree strongly’. We recode this item to the value range 0 (‘disagree strongly’) to 4 (‘agree strongly’). The second item measures preferences for governmental support for the unemployed: ‘It is governments’ responsibility to ensure a reasonable standard of living for the unemployed’ (Item D18 from the ESS4-module ‘welfare attitudes’). Here, respondents could choose among 11 answers from (0) ‘should not be governments’ responsibility at all’ to (10) ‘should be entirely governments’ responsibility’. Both variables are treated as metric (continuous) variables.³

Indicators which are relevant to our dependent variables from a theoretical and empirical perspective are incorporated into the statistical models as independent variables. At the country-level, we use the KOF Index of Economic Globalization as a central explanatory variable and Gross Domestic Product (GDP) and Social Expenditure as control variables, which are both likely to be associated with welfare attitudes (Table 5 in Appendix provides an overview of our contextual data). The KOF Index of Economic Globalization measures trade and investment flows as well as the extent to which countries protect themselves from competition via controls on the movement of goods, services and capital (see Dreher et al., 2008). This index is based on data for the year 2008; it deploys a scale from 1 to 100 with higher values indicating more economic globalization.⁴ The GDP serves as a measure of the economic prosperity and general economic condition of a country; it stems from the Penn World Table 7.1 (real GDP per capita; constant prices: Laspeyres; data from 2008). Social expenditure (as percentage of GDP) taps the level of welfare state effort; it comes from the World Social Security Report 2010/2011 edited by the International Labour Organization (data from 2004 to 2007).

At the individual-level, we take as a central explanatory variable (1) class position which captures an individual’s socio-economic position. We utilize the class scheme of Erikson et al. (1979) (in the following: Erikson-Goldthorpe-Portocarero (EGP)) which is linked to occupational skill categories and location within the market and a system of authority. We deploy a six-class version of the EGP scheme (derived from the International Standard Classification of Occupations (ISCO) codes) for all respondents which means that currently unemployed respondents are assigned to the class of their last occupation.⁵ For analytical purposes, we collapse the six-class EGP scheme into three groups: (1) ‘high-skilled employees’ (EGP I and EGP II), (2) ‘medium-skilled employees and workers’ (EGP III and EGP V + VI) and (3) ‘low-skilled workers’ (EGP VII).⁶

Further control variables at the individual-level are (2) gender and (3) age. With regard to gender, we know that women tend to be more pro-welfarist than men (e.g. Alesina and Giuliano, 2009; Cusack et al., 2006; Eagly et al., 2004). Age is also relevant to attitudes towards redistribution because it stands for potential inter- and intra-generational conflicts (e.g. Busemeyer et al., 2009). With regard to unemployment, we know from previous studies that this welfare-dependent group shows stronger preferences for redistribution than other groups (e.g. Alesina and Giuliano, 2009; Amat and Wibbels, 2009; Cusack et al., 2006). As to checking whether not class, but alternatively unemployment matters for the globalization–welfare state nexus, we also run additional models with the unemployed as a separate category (0/1-coded) for cross-level interactions. Tables 6 and 7 in Appendix provide a descriptive overview of our individual-level variables.

Method

To test our hypotheses and to determine the degree to which individual- and country-level characteristics explain the variance in our dependent variables, we conduct hierarchical linear regressions as they are well-suited for the analysis of clustered data. In such a setting, standard ordinary least squares (OLS)-regressions would lead to biased estimates of standard errors, whereas hierarchical modelling brings about separate variances at all levels investigated as well as correctly estimated standard errors (Hox, 2010; Snijders and Bosker, 2012). All multilevel models are estimated with the software MLwiN (version 2.30), using Restricted Iterative Generalized Least Squares (RIGLS). All continuous variables at both levels are grand-mean centred. In all our analyses, the post-stratification weight (including design weight) which is provided with the ESS4-dataset (Edition 4.2) is applied to account for slightly different selection probabilities for individuals (or groups of individuals) in the countries under study. However, countries are weighted equally, which means that the results do not depend on their population or sample size; instead countries are seen as discrete entities.

Analytical strategy

Our analysis begins with a few descriptive findings. Social structural differences, specifically different national compositions of social classes, may cause variations in redistributive preferences. But potential differences in redistributive preferences may also be associated with different degrees of economic globalization, national wealth and welfare state effort at the country-level. For shedding light on this, we employ multilevel analyses (see the preceding sub-section, ‘Method’).

We pursue a strategy of parsimonious modelling. In all multilevel models depicted in Tables 3 and 4, preferences for income redistribution (Item 1) and governmental support for the unemployed (Item 2) serve as our continuous dependent variables. In both cases, we begin with the estimation of an ‘empty’ model (Model 0) without any explanatory variables in order to establish how much variation in our dependent variables is located at the country-level before adding explanatory factors. In the next step, we calculate a random intercept model in which all individual-level explanatory variables are incorporated as fixed effects (Model I). Thereafter, we conduct further analyses by successively adding contextual information: at first, we use the KOF Index of Economic Globalization as the only macro variable (Model II). Adding the control variables GDP per capita (Model III), social expenditure (Model IV), and respectively both country-level variables simultaneously (Model V) should increase the explanatory power of contextual information. In the subsequent eight models, we perform multilevel analyses with cross-level interactions in order to capture the interplay between economic globalization and socio-economic positions; here, we differentiate between cross-level interactions with class affiliations (EGP groups) and with the welfare-dependent group of unemployed (Models VIa to IXb).

Table 3.

Individual-level determinants of preferences for governmental redistribution (linear multilevel models).

	Income redistribution (Item 1)				Support for the unemployed (Item 2)
	Model 0		Model I		Model 0		Model I
	Coefficient (Standard error)	Sig.	Coefficient (Standard error)	Sig.	Coefficient (Standard error)	Sig.	Coefficient (Standard error)	Sig.
Fixed part
Constant	2.846 (0.054)	***	2.862 (0.055)	***	7.032 (0.141)	***	6.952 (0.147)	***
Gender (1 = female)			0.137(0.020)	***			0.141(0.032)	***
Age			0.004(0.001)	***			0.005(0.002)	**
High-skilled			−0.288(0.025)	***			−0.182(0.057)	***
Medium-skilled	Ref.		Ref.		Ref.		Ref.
Low-skilled			0.072(0.016)	***			0.156(0.046)	***
Unemployed			0.138(0.024)	***			0.522(0.057)	***
Random part
σ_u² (variance level 2)	0.091(0.023)	***	0.081(0.021)	***	0.628(0.108)	***	0.595(0.101)	***
σ_e² (variance level 1)	1.030(0.050)	***	0.998(0.047)	***	4.838(0.281)	***	4.787(0.278)	***
Model fit
−2*loglikelihood	83,066.146		82,195.847		126,853.531		126,548.939

Source: European Social Survey, 2008; n_i = 27,869 (Item 1) respectively n_i = 28,010 (Item 2); n_j = 31.

Sig.: ***p < 0.01; **p < 0.05; *p < 0.10.

Table 4.

Country-level determinants of preferences for governmental redistribution (linear multilevel models).

Model	Income redistribution (Item 1)								Support for the unemployed (Item 2)
	Variable	Coefficient (standard error)	Sig.	Variance components and fits					Coefficient (standard error)	Sig.	Variance components and fits
				σ_u²	Sig.	σ_e²	Sig.	−2*loglikelihood			σ_u²	Sig.	σ_e²	Sig.	−2*loglikelihood
II	KOF	−0.007(0.004)	*	0.078	***	0.998	***	82,193.677	−0.018(0.011)		0.581	***	4.787	***	126,547.200
III	KOF	−0.002(0.004)		0.064	***	0.998	***	82,186.581	−0.006(0.012)		0.528	***	4.787	***	126,543.217
	GDP	−0.012(0.003)	***						−0.026(0.013)	*
IV	KOF	−0.006(0.004)		0.079	***	0.998	***	82,192.964	−0.012(0.012)		0.554	***	4.787	***	126,544.717
	Soc. exp.	−0.009(0.008)							−0.045(0.025)	*
V	KOF	−0.002(0.004)		0.066	***	0.998	***	82,186.429	−0.005(0.012)		0.538	***	4.787	***	126,542.830
	GDP	−0.014(0.004)	***						−0.020(0.014)
	Soc. exp.	0.006(0.009)							−0.022(0.026)
VIa	KOF	−0.007(0.004)	*	0.065	***	0.994	***	82,123.177	−0.018(0.013)		0.652	***	4.767	***	126,492.477
	KOF × high-skilled	0.000(0.003)							0.001(0.005)
	KOF × low-skilled	0.000(0.001)							−0.002(0.004)
VIb	KOF	−0.008(0.004)	*	0.081	***	0.997	***	82,180.905	−0.018(0.011)		0.605	***	4.782	***	126,534.477
	KOF × unemployed	0.004(0.002)	*						0.002(0.005)
VIIa	KOF	−0.003(0.003)		0.054	***	0.994	***	82,117.976	−0.007(0.013)		0.562	***	4.767	***	126,489.937
	GDP	−0.009(0.003)	***						−0.024(0.014)	*
	KOF × high-skilled	0.000(0.003)							0.001(0.005)
	KOF × low-skilled	0.000(0.001)							−0.002(0.004)
VIIb	KOF	−0.003(0.004)		0.066	***	0.997	***	82,175.610	−0.009(0.012)		0.548	***	4.782	***	126,531.460
	GDP	−0.011(0.003)	***						−0.020(0.012)
	KOF × unemployed	0.004(0.002)	*						0.002(0.005)
VIIIa	KOF	−0.007(0.004)	*	0.066	***	0.994	***	82,123.093	−0.012(0.014)		0.608	***	4.767	***	126,490.127
	Soc. exp.	−0.003(0.008)							−0.046(0.024)	*
	KOF × high-skilled	0.000(0.003)							0.001(0.005)
	KOF × low-skilled	0.000(0.001)							−0.002(0.004)
VIIIb	KOF	−0.007(0.004)	*	0.082	***	0.997	***	82,180.718	−0.012(0.013)		0.572	***	4.782	***	126,531.127
	Soc. exp.	−0.005(0.008)							−0.045(0.025)	*
	KOF × unemployed	0.004(0.002)	*						0.002(0.005)
IXa	KOF	−0.003(0.003)		0.055	***	0.994	***	82,117.022	−0.006(0.014)		0.573	***	4.767	***	126,489.004
	GDP	−0.011(0.003)	***						−0.017(0.014)
	Soc. exp.	0.009(0.009)							−0.031(0.025)
	KOF × high-skilled	0.000(0.003)							0.001(0.005)
	KOF × low-skilled	0.000(0.001)							−0.002(0.004)
IXb	KOF	−0.003(0.004)		0.068	***	0.997	***	82,175.101	−0.008(0.013)		0.557	***	4.782	***	126,530.202
	GDP	−0.012(0.004)	***						−0.012(0.011)
	Soc. exp.	0.007(0.009)							−0.031(0.024)
	KOF × unemployed	0.004(0.002)	*						0.002(0.005)

Source: European Social Survey, 2008; n_i = 27,869 (Item 1) respectively n_i = 28,010 (Item 2); n_j = 31.

The KOF Index of Economic Globalization measures trade and investment flows as well as the extent to which countries protect themselves from competition via controls on the movement of goods, services and capital. Results of 12 separate multilevel models (RIGLS estimates), each one containing all individual-level covariates that appear in Model I (Table 3), individual-level coefficients and standard errors are almost identical; σ_u² = variance level 2; σ_e² = variance level 1; Models VIa to IXb are with random slopes and include cross-level interactions; in these eight models, the values of −2*loglikelihood are calculated with IGLS estimates of the coefficients.

Sig.: *** p < 0.01; ** p < 0.05; * p < 0.10.

Empirical findings

Descriptive section

As expected, preferences for governmental redistribution differ notably across the 31 countries under study. The average agreement (‘agree strongly’ and ‘agree’) to the statement, ‘The government should take measures to reduce differences in income levels’ (Table 1), amounts to 71 % the average disagreement (‘disagree strongly’ and ‘disagree’) is 14 percent. Agreement is very widespread in Greece (92%), Portugal (90%) and Turkey (89%) whose welfare states are comparatively weak, whereas agreement is lowest in Denmark (41%), the Netherlands (52%) and the Czech Republic (51%). For the latter, an explanation is less quick at hand since, for example, Denmark and the Czech Republic belong to different welfare regimes.

With regard to our second dependent variable measuring preferences for governmental support for the unemployed (Table 2), we see that support for this welfare-dependent group is particularly strong among respondents in Greece (mean value on a scale from 0 to 10: 8.5), Lithuania and Latvia (both 8.4) and rather weak among respondents in the United Kingdom and Slovakia (both 5.8). On average, though, people seem to be comparatively supportive for governmental support for the unemployed (7.0).

Now we turn to the link between class, globalization and preferences for governmental redistribution. A first descriptive account is given in Figures 1 and 2 which show the relationship between economic globalization and welfare state support for the high-skilled, on the one side, and the low-skilled, on the other side.

Figure 1.

Globalization and preferences for income redistribution.

Figure 2.

Globalization and preferences for governmental support for the unemployed.

Figure 1 shows that higher levels of economic globalization are associated with lower levels of support for income redistribution among the high-skilled as well as the low-skilled. Both regression lines run virtually parallel showing that high-skilled are generally less supportive for income redistribution than low-skilled. Here, class differences do not seem to widen with higher levels of globalization. Figure 2 resembles the former one: higher levels of economic globalization are accompanied by lower levels of preferences for governmental support for the unemployed among the high-skilled as well as the low-skilled.

So far, these descriptive results seem neither to provide evidence for our hypothesis, H1a, assuming that high levels of globalization trigger more demand for redistribution, nor for our main hypothesis, H2, claiming different effects of class affiliation across different settings of economic globalization. In the following section, ‘Multilevel analysis’, we scrutinize this relationship by performing multilevel analyses.

Multilevel analysis

We now turn to the findings of our multilevel analyses and start with the so-called ‘empty’ models (Table 3, Models 0) which allow a decomposition of the variances of our dependent variables (Item 1: income redistribution; Item 2: support for the unemployed) into two components: the variance between the countries and the variance within the countries. The intra-class correlation coefficients (calculated according to Snijders and Bosker, 2012: 17–18) indicate that slightly more than 8 percent and more than 11 percent of the variances respectively are attributable solely to differences between the countries. Accordingly, about 92 percent and 89 percent of the variances respectively remain at the individual-level.

Next, we calculate models containing only individual-level variables (Table 3, Models I). Here, both models show similar patterns: as expected, compared to the reference group of medium-skilled employees and workers (EGP III and EGP V + VI), the high-skilled employees (EGP I and EGP II) show a significantly lower demand for governmental redistribution, whereas the low-skilled workers (EGP VII) and the unemployed exhibit a significantly higher demand. Concerning our further controls, we find that women as well as older respondents are more in favour of redistribution; 10.9 percent (Item 1) and 5.2 percent (Item 2) of the total cross-national variance can be explained by the country-specific composition of individual socio-economic characteristics. Thus, the major part of between-country variation remains unexplained.

We turn to the role of country-level determinants and start with the impact of economic globalization on welfare state support. We find that a higher level of economic globalization does not have an effect on support for the unemployed but lowers the level of support for income redistribution (Table 4, Models II). However, this ‘globalization effect’ on peoples’ preferences for income redistribution becomes insignificant when a country’s economic strength is taken into account. Indeed, we find a negative effect of GDP per capita on both measures of governmental redistribution, meaning that in countries with higher GDP, the demand for income redistribution and support for the unemployed is significantly reduced (Table 4, Models III). This finding confirms other research evincing that the demand for income redistribution is lower in economically well-off countries (Dallinger, 2010; Schmidt, 2012).⁷

Another contextual factor which might matter for welfare preferences is the actual level of welfare effort. If we control for this by adding social expenditure (as percentage of GDP) to the KOF Index of Economic Globalization, we see that both context variables show no effect on preferences for income redistribution (Table 4, Item 1, Model IV). However, the level of welfare state effort affects the support for the unemployed: in countries with higher welfare-spending levels, people agree less with the statement that their governments should ensure a reasonable standard of living for the unemployed (Table 4, Item 2, Model IV). If we control for all three macro factors at the same time, only GDP per capita shows a significant negative effect on preferences for income redistribution (but not on preferences for support for the unemployed), again meaning that in countries with better economic performance, preferences for income redistribution are reduced (Table 4, Item 1, Model V).

Now we turn to our main focus, the interaction between socio-economic class, economic globalization and welfare attitudes. Remarkably and contrary to our main hypothesis, H2, we do not find any significant cross-level interaction effects between economic globalization and class affiliation, neither for our first item measuring preferences for income redistribution nor for our second item measuring preferences for governmental support for the unemployed. This finding remains stable if we add GDP per capita or social expenditure (as percentage of GDP) to the models (Table 4, Models VIa, VIIa, VIIIa, IXa). In other words, class matters, but class differences in welfare attitudes neither seem to widen nor to diminish with higher levels of economic globalization.

After having found that economic globalization does not enforce class differences in attitudes, the question arises whether a more direct measure of the labour market status like unemployment yields different results as suggested by parts of the literature (Dallinger, 2014; Rehm, 2009). While class stands for a broader concept referring to a more general market status, unemployment refers to a critical life event. Looking at the first cross-level interaction between economic globalization and unemployment status, we see indeed a significant result for preferences for income redistribution (Table 4, Item 1, Model VIb). It signals a polarization in the sense that with higher levels of economic globalization, the unemployed stay more supportive of welfare state redistribution than the employed groups. Adding either GDP per capita or social expenditure (as percentage of GDP) or both additional macro determinants to the models does not change this ‘polarization effect’: in highly globalized countries, preferences for income redistribution are generally lower; however, the unemployed still differ from the general pattern by staying more pro-welfarist (Table 4, Item 1, Models VIIb, VIIIb, IXb).

Table 4 also shows that adding explanatory information at the country-level increases the quality of model fit (as indicated by changes in −2*loglikelihood).⁸ Bearing in mind the sequence of nested models, we first look at models without cross-level interactions: here, we observe that including the KOF Index of Economic Globalization does not improve the quality of model fits significantly. However, we achieve significantly better model fits after adding GDP per capita. Adding social expenditure (as percentage of GDP) though does not further improve the model fits. However, we receive significantly better model fits with models with cross-level interactions (as compared with those without them). Overall, for both of our dependent variables Models IXa turn out to be the best-fitted models. These are the models that incorporate all three macro determinants and test for cross-level interaction effects between economic globalization and class affiliation.

What is the bottom-line of our findings? First, there is a significant effect of economic globalization on preferences for income redistribution (our first item) as long as the KOF Index of Globalization is the only country-level variable taken into account. However, a country’s economic strength outweighs the impact of its level of globalization. Second, contrary to our main hypothesis, H2, we found no significant interactions between economic globalization and class affiliation regarding preferences for income redistribution. However, we found significant interactions between economic globalization and unemployment: unemployed respondents display less defection from the redistributive welfare state – even under conditions of high levels of globalization.⁹ Third, preferences for governmental support for the unemployed (our second item) react to the level of welfare state effort meaning that welfare generosity weakens redistributive preferences with respect to those in need, most possibly a ceiling effect.

Conclusion

The impact of globalization on national welfare states is widely discussed in the literature. While one part of it focusses on the role of tax competition and the issue of high levels of capital mobility and the spread of risk and insecurity, the other part highlights the critical issue of welfare state support and legitimacy. Here, one camp underlines the protective function of the welfare state making greater economic openness more acceptable, while the other one sees welfare state legitimacy at risk. Although this literature makes the explicit argument that there is a link between risks associated with globalization and domestic preferences for governmental redistribution, empirical studies addressing this issue are still rare. The article approaches this topic by addressing the question of whether preferences for income redistribution and support for unemployment protection vary under different levels of economic globalization. The first question we raised is: does globalization reinforce or weaken welfare state support? Against the background of the literature on globalization on the one hand and on class and welfare state support on the other hand, we suggested that this link might be class-specific. Thus, the second and more specific question we raised is whether class cleavages in welfare attitudes deepen under conditions of globalization.

To answer these questions, we analysed comparative data for 31 countries from the fourth round of the European Social Survey. We found that in highly globalized countries (measured by the KOF Index of Economic Globalization), preferences for income redistribution tend to be lower (confirming our hypothesis H1b assuming that high levels of globalization trigger less demand for redistribution), whereas economic globalization does not seem to affect peoples’ preferences for support for the unemployed. However, once we include GDP per capita in our models, the significance of this ‘globalization effect’ vanishes suggesting that the economic condition of a country matters more than its degree of economic globalization.

As our main interest was directed at class-specific differences of preferences for governmental redistribution under different levels of economic globalization, we furthermore performed multilevel analyses with cross-level interactions. Disconfirming our main hypothesis H2 claiming different effects of class across different settings of economic globalization, we did not find any significant cross-level interactions implying that class differences in welfare attitudes neither diminish nor intensify with higher levels of globalization. Against the background of the debate on the role of class in the era of globalization, we therefore suggest that class still matters, but also that globalization neither undermines nor exacerbates class differences in attitudes. This is not to say that market status is of no relevance at all for differences in preferences under different levels of globalization, but class, often understood as the best proxy for material as well as ideological group-positioning, does not seem to exert a heavy influence here. We rather found an, albeit modest, influence of unemployment status on attitudes towards redistribution, a finding which confirms other research (Dallinger, 2014). Hence, the experience of a critical life event matters more than belonging to a socio-economic class category.

Finally, we want to address this study’s limitations. Besides the fact that our findings rely on cross-sectional data (making it impossible to draw conclusions about causal effects rather than correlations), we focussed on three macro determinants – economic globalization, economic strength and welfare generosity – and two of them (economic globalization and economic strength) are undeniably closely intertwined. We chose them for conceptual reasons but future research regarding preferences for governmental redistribution should enhance the explanatory power of level 2-units (i.e. countries) by adding further suitable country-level determinants. Additionally, it would be worthwhile to include more appropriate individual-level determinants because it can be expected that further covariates will increase the proportions of explained variances in all models. With this in mind, our study should be considered a starting point for tackling the problem of explaining the dynamics of welfare state support under conditions of economic globalization.

Footnotes

Appendix

Table 7.

Descriptives for models with item 2 (individual-level variables).

	Minimum	Maximum	Mean	Standard deviation
It is governments’ responsibility to ensure a reasonable standard of living for the unemployed	0	10	7.019	2.323
Gender (1 = female)	0	1	0.464	0.499
Age (years)	18	65	40.436	11.664
High-skilled (1 = yes)	0	1	0.365	0.481
Medium-skilled (1 = yes)	0	1	0.427	0.495
Low-skilled (1 = yes)	0	1	0.208	0.406
Unemployed (1 = yes)	0	1	0.101	0.301

Source: European Social Survey, 2008; n_i = 28,010; n_j = 31.

‘High-skilled’ = EGP I and EGP II; ‘medium-skilled’ = EGP III and EGP V + VI; ‘low-skilled’ = EGP VII.

Acknowledgements

We wish to thank Traute Meyer and the two anonymous reviewers for their valuable suggestions and comments on an earlier version of this paper. Furthermore we thank Nate Breznau for proof-reading this paper as well as Christoph Burkhardt and Jan Mewes for helpful comments.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

Notes

References

Alesina

Giuliano

(2009) ‘Preferences for Redistribution’, IZA Discussion Papers No. 4056, March, available at http://ftp.iza.org/dp4056.pdf.

Amat

Wibbels

(2009) ‘Electoral Incentives, Group Identity and Preferences for Redistribution’, Estudio/Working Paper 2009/246. Madrid: Juan March Institute.

Baldwin

R.E.

Winters

L.A.

(eds) (2004) Challenges to Globalization: Analyzing the Economics. Chicago, IL: University of Chicago Press.

Bauman

(2001) Community: Seeking Safety in an Insecure World. Cambridge: Polity Press.

Beck

(2013) ‘Why “Class” Is Too Soft a Category to Capture the Explosiveness of Social Inequality at the Beginning of the Twenty-First Century’, The British Journal of Sociology 64: 63–74.

Belsley

Kuh

Welsch

R.E.

(1980) Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. New York: John Wiley.

Blekesaune

(2007) ‘Economic Conditions and Public Attitudes to Welfare Policies’, European Sociological Review 23: 393–403.

Boix

(2003) Democracy and Redistribution. Cambridge: Cambridge University Press.

Brady

(2009) ‘Economic Globalization and Increasing Earnings Inequality in Affluent Democracies’, Research in Sociology of Work: Economic Sociology of Work 18: 149–81.

10.

Brady

Beckfield

Zhao

(2007) ‘The Consequences of Economic Globalization for Affluent Democracies’, Annual Review of Sociology 33: 313–34.

11.

Breznau

(2010) ‘Economic Equality and Social Welfare: Policy Preferences in Five Nations’, International Journal of Public Opinion Research 22: 458–84.

12.

Breznau

Lykes

V.A.

Kelley

Evans

M.D.R.

(2011) ‘A Clash of Civilizations? Preferences for Religious Political Leaders in 86 Nations’, Journal for the Scientific Study of Religion 50: 671–91.

13.

Brooks

Svallfors

(2010) ‘Why Does Class Matter? Policy Attitudes, Mechanisms, and the Case of the Nordic Countries’, Research in Social Stratification and Mobility 28: 199–213.

14.

Buchholz

Hofäcker

Blossfeld

H.P.

Mills

Kurz

Hofmeister

(2009) ‘Life Courses in the Globalization Process: The Development of Social Inequalities in Modern Societies’, European Sociological Review 25: 53–71.

15.

Burgoon

(2012) ‘Partisan Embedding of Liberalism: How Trade, Investment, and Immigration Affect Party Support for the Welfare State’, Comparative Political Studies 45: 606–35.

16.

Busemeyer

M.R.

Goerres

Weschle

(2009) ‘Attitudes towards Redistributive Spending in an Era of Demographic Ageing: The Rival Pressures from Age and Income in 14 OECD Countries’, Journal of European Social Policy 19: 195–212.

17.

Curran

(2013) ‘Risk Society and the Distribution of Bads: Theorizing Class in the Risk Society’, The British Journal of Sociology 64: 44–62.

18.

Cusack

Iversen

Rehm

(2006) ‘Risk at Work: The Demand and Supply Sides of Government Redistribution’, Oxford Review of Economic Policy 22: 365–89.

19.

Dallinger

(2010) ‘Public Support for Redistribution: What Explains Cross-National Differences?’, Journal of European Social Policy 20: 333–49.

20.

Dallinger

(2013) ‘Economic Openness and Domestic Demand for Social Protection: A Multi-Level Analysis of Social Security Preferences between 1990 and 2006’, Comparative Sociology 12: 585–616.

21.

Dallinger

(2014) ‘Globalisierung und die Nachfrage nach sozialer Sicherheit. Eine kritische Analyse des “domestic-demand”-Ansatzes’, Berliner Journal für Soziologie 24: 59–88.

22.

Dollar

Kraay

(2001) ‘Trade, Growth, and Poverty’, World Bank Discussion Paper No. 2615, June. Washington, DC: World Bank.

23.

Dreher

(2006a) ‘The Influence of Globalization on Taxes and Social Policy: An Empirical Analysis for OECD Countries’, European Journal of Political Economy 22: 179–201.

24.

Dreher

(2006b) ‘Does Globalization Affect Growth? Evidence from a new Index of Globalization’, Applied Economics 38: 1091–110.

25.

Dreher

Gaston

Martens

(2008) Measuring Globalization: Gauging Its Consequences. New York: Springer.

26.

Eagly

A.H.

Diekman

A.B.

Johannesen-Schmidt

M.C.

Koenig

A.M.

(2004) ‘Gender Gaps in Sociopolitical Attitudes: A Social Psychological Analysis’, Journal of Personality and Social Psychology 87: 796–816.

27.

Erikson

Goldthorpe

J.H.

(1992) The Constant Flux: A Study of Class Mobility in Industrial Societies. Oxford: Clarendon Press.

28.

Erikson

Goldthorpe

J.H.

Portocarero

(1979) ‘Intergenerational Class Mobility in Three Western European Societies: England, France and Sweden’, The British Journal of Sociology 30: 415–41.

29.

Finseraas

(2009) ‘Income Inequality and Demand for Redistribution: A Multilevel Analysis of European Public Opinion’, Scandinavian Political Studies 32: 94–119.

30.

Finseraas

Ringdal

(2012) ‘Economic Globalization, Personal Risks and the Demand for a Comprehensive Welfare State’, in Ervasti

Andersen

J.G.

Fridberg

Ringdal

(eds) The Future of the Welfare State: Social Policy Attitudes and Social Capital in Europe, pp. 68–87. Cheltenham: Edward Elgar.

31.

Garrett

(1995) ‘Capital Mobility, Trade, and the Domestic Politics of Economy Policy’, International Organization 49: 657–87.

32.

Garrett

(1998) ‘Global Markets and National Politics: Collision Course or Virtuous Circle?’, International Organization 52: 787–824.

33.

Goldthorpe

(2002) ‘Globalisation and Social Class’, West European Politics 25: 1–28.

34.

Hacker

J.S.

(2006) The Great Risk Shift: The Assault on American Jobs, Families, Health Care and Retirement – And How You Can Fight Back. Oxford: Oxford University Press.

35.

Hainmueller

Hiscox

M.J.

(2006) ‘Learning to Love Globalization: Education and Individual Attitudes toward International Trade’, International Organization 60: 469–98.

36.

Harms

Zink

(2003) ‘Limits to Redistribution in a Democracy: A Survey’, European Journal of Political Economy 19: 651–68.

37.

Hox

J.J.

(2010) Multilevel Analysis: Techniques and Applications, 2nd edn. New York: Routledge.

38.

Jæger

M.M.

(2006) ‘What Makes People Support Public Responsibility for Welfare Provision: Self-Interest or Political Ideology? A Longitudinal Approach’, Acta Sociologica 49: 321–38.

39.

Jæger

M.M.

(2011) ‘Macroeconomic and Social Change and Popular Demand for Redistribution’, available at: http://ssrn.com/abstract=1799951.

40.

Korpi

(2006) ‘Power Resources and Employer-centered Approaches in Explanations of Welfare States and Varieties of Capitalism: Protagonists, Consenters, and Antagonists’, World Politics 58: 167–206.

41.

Koster

(2007) ‘Globalization, Social Structure, and the Willingness to Help Others: A Multilevel Analysis Across 26 Countries’, European Sociological Review 23: 537–51.

42.

Koster

(2010) ‘Welfare State Attitudes and Economic Integration in the European Union, 1992–2002: A Multilevel Investigation across 24 Countries’, Policy and Politics 38: 179–95.

43.

Lee

J.-E.

(2006) ‘Inequality and Globalization in Europe’, Journal of Policy Modeling 28: 791–6.

44.

Leiulfsrud

Bison

Solheim

(2010) Social Class in Europe II: The European Social Survey 2002–2008. Trondheim: Norwegian University of Science and Technology.

45.

Lübker

(2004) ‘Globalization and Perceptions of Social Inequality’, International Labour Review 143: 91–128.

46.

Lupu

Pontusson

(2011) ‘The Structure of Inequality and the Politics of Redistribution’, American Political Science Review 105: 316–36.

47.

Mau

(2003) The Moral Economy of Welfare States: Britain and Germany Compared. London; New York: Routledge.

48.

Mau

Veghte

(eds) (2007) Social Justice, Legitimacy, and the Welfare State. London: Ashgate.

49.

Mau

Mewes

Schöneck

N.M.

(2012) ‘What Determines Subjective Socio-economic Insecurity? Context and Class in Comparative Perspective’, Socio-Economic Review 10: 655–82.

50.

Mayda

A.M.

Rodrik

(2005) ‘Why Are Some People (and Countries) More Protectionist Than Others?’, European Economic Review 49: 1393–430.

51.

O’Rourke

Sinnott

(2001) ‘The Determinants of Individual Trade Policy Preferences: International Survey Evidence [with Comments and Discussion]’, in Collins

S.M.

Rodrik

(eds) Brookings Trade Forum, pp. 157–206. Washington, DC: Brookings Institution.

52.

Rehm

(2009) ‘Risks and Redistribution: An Individual-Level Analysis’, Comparative Political Studies 42: 855–81.

53.

Rieger

Leibfried

(2003) Limits to Globalization: Welfare States and the World Economy. Cambridge: Polity Press.

54.

Rodrik

(1998) ‘Why Do More Open Economies Have Bigger Governments?’, Journal of Political Economy 106: 997–1032.

55.

Scheve

K.F.

Slaughter

M.J.

(2001) ‘What Determines Individual Trade-policy Preferences?’, Journal of International Economics 54: 267–92.

56.

Scheve

K.F.

Slaughter

M.J.

(2004) ‘Economic Insecurity and the Globalization of Production’, American Journal of Political Science 48: 662–74.

57.

Schmidt

(2012) ‘The Development of Public Demand for Redistribution: A Pseudo-Panel Model for Decomposing within- and between Effects’, GK SOCLIFE Working Paper Series 10, May. Cologne: University of Cologne.

58.

Schulze

G.G.

Ursprung

H.W.

(1999) ‘Globalisation of the Economy and the Nation State’, The World Economy 22: 295–452.

59.

Snijders

T.A.B.

Bosker

R.J.

(2012) Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling, 2nd edn. London: Sage.

60.

Soifer

H. D.

(2009) ‘The Redistributive Threat: State Power and the Effect of Inequality on Democracy’, BWPI Working Paper 93, May. Manchester: Brooks World Poverty Institute.

61.

Standing

(2008) ‘Economic Insecurity and Global Casualisation: Threat or Promise?’, Social Indicators Research 88: 15–30.

62.

Svallfors

(2004) ‘Class, Attitudes and the Welfare State: Sweden in Comparative Perspective’, Social Policy and Administration 38: 119–38.

63.

Svallfors

(2006) The Moral Economy of Class: Class and Attitudes in Comparative Perspective. Stanford, CA: Stanford University Press.

64.

Svallfors

(2007) ‘Introduction’, in Svallfors

(ed.) The Political Sociology of the Welfare State, pp. 1–29. Stanford, CA: Stanford University Press.

65.

Svallfors

(2011) ‘A Bedrock of Support? Trends in Welfare State Attitudes in Sweden, 1981–2010’, Social Policy and Administration 45: 806–25.

66.

Svallfors

(ed.) (2012) Contested Welfare States: Welfare Attitudes in Europe and Beyond. Stanford, CA: Stanford University Press.

67.

Taylor-Gooby

(2004) ‘Open Markets and Welfare Values: Welfare Values, Inequality and Social Change in the Silver Age of the Welfare State’, European Societies 6: 29–48.

68.

Van der Meer

Te Grotenhuis

Pelzer

. (2010) ‘Influential Cases in Multilevel Modeling: A Methodological Comment’, American Sociological Review 75: 173–8.

69.

Van Oorschot

. (2000) ‘Why Pay for Welfare? A Sociological Analysis of Reasons for Welfare Solidarity’, The Netherlands’ Journal of Social Sciences 36: 15–36.

70.

Walter

(2010) ‘Globalization and the Welfare State: Testing the Microfoundations of the Compensation Hypothesis’, International Studies Quarterly 54: 403–26.

71.

Walter

Maduz

(2009) ‘How Globalization Shapes Individual Risk Perceptions and Policy Preferences: A Cross-national Analysis of Differences between Globalization Winners and Losers’, WCFIA Working Paper 09-0015, August. Cambridge, MA: Harvard University.

72.

Whelan

C.T.

Maître

(2008) ‘Social Class Variation in Risk: A Comparative Analysis of the Dynamics of Economic Vulnerability’, The British Journal of Sociology 59: 637–59.