This is an open access article distributed under the terms of the Creative Commons
Attribution License (Before the start of the COVID-19 pandemic, respondents in the global opinion poll “What worries the world?” ranked economic inequality as their greatest worry (Ipsos, 2020). There is ample evidence for the negative impact of inequality on many domains of public and private life including mental and physical health, crime, and social cohesion (Becker et al., 2021; Buttrick & Oishi, 2017; Pickett & Wilkinson, 2015; Wilkinson & Pickett, 2017). Yet despite long-standing acknowledgement of the issue and its urgency, inequality still continues to rise for 70 percent of the global population (United Nations Department of Economic and Social Affairs [UN DESA], 2020). The increasing gap between rich and poor and the concentration of wealth at the top of the economic ladder has drawn concern from economists and international organizations (Organisation for Economic Co-operation and Development [OECD], 2011; Piketty, 2015) as well as politicians and even the CEOs of investment banks (Song, 2014).
Given that trickle-down economics have failed to deliver on their promises (Holt & Greenwood, 2012), it seems that attempts to achieve more equal societies need to actively reverse the decades-long shift of resources toward those at the top. However, neither such redistributive policies themselves nor widespread demands for them have manifested on a larger scale (Bartels, 2005). On the contrary, the median-voter model based assumption that support for redistribution would rise in tandem with inequality (Meltzer & Richard, 1981) was disproved across several countries (Kenworthy & McCall, 2008).
Against this background it seems particularly important to better understand what drives this reluctance toward redistributive policies. The present research sets out to contribute to that aim by investigating three possible influences on individuals’ opposition to redistribution: attitudes toward the goal, the means, and the recipients of redistributive policies. Contrasting an intergroup and an ideological perspective, we argue that while opposition to redistribution can be construed as a form of harm against those at the bottom of the economic ladder, unlike discrimination, it is mainly driven by ideological factors. Below, we discuss these mechanisms and outcomes in more detail.
Inequality and Redistribution
Given that the goal of redistributive policies is to reduce social inequality, the most intuitive and straightforward explanation for why people oppose these policies is that they do not consider a more equal society to be a desirable or important outcome (Bartels, 2005; Samuelson, 2001). Anti-egalitarian and pro-hierarchy views as contained in the construct of Social Dominance Orientation (Pratto et al., 1994) have indeed been linked to a rejection of equality-enhancing measures (Sidanius, 1993). However, the idea that a lack of support for redistribution is merely a result of large-scale indifference to inequality is contradicted by the consistent finding that a majority of people do report great concern about the issue (Ipsos, 2020; Mau & Heuer, 2016). Moreover, research has shown that experimental treatments that successfully increase participants’ perception of inequality as a problem barely affect their support for redistribution (Kuziemko et al., 2015) and that even individuals who are opposed to inequalities such as large income differences often still reject specific plans to reduce them (Page & Jacobs, 2009).
This strongly suggests that there are additional factors besides attitudes toward inequality itself that are associated with people’s stance on redistributive policies. We examine two such factors that we consider particularly relevant from a social psychological perspective: the influences of classism, that is, negative perceptions of those who would most benefit from redistribution, and capitalism, that is, endorsement of the dominant economic ideology which opposes government interventions in the free market. We propose that while classism should be linked to interpersonal discrimination against this target group, it is the ideological rather than the intergroup approach that is the most promising in explaining opposition to redistribution.
The Intergroup Perspective: Classism and Redistribution
The growing gap between rich and poor has brought with it the increased salience (Jetten et al., 2017) of an identity category that had, perhaps prematurely, been declared dead a few decades earlier (Pakulski & Waters, 1996): that of social class. Social class is defined as an individual’s position in the economic hierarchy of society determined by income, education, and occupational prestige, with the term ‘lower social class’ designating those at the bottom of that hierarchy (Adler et al., 1994; Kraus et al., 2017; Oakes & Rossi, 2003). Social class shapes individuals’ thinking and behavior (Fiske & Markus, 2012; Stephens et al., 2014) as well as their cultural expression (Becker et al., 2017) and plays an important role in determining outcomes and opportunities in many areas of life, from health (Lachman & Weaver, 1998), to education (Carnevale & Rose, 2003), to their choice of romantic partners (OECD, 2011). Social class membership is signaled and communicated frequently in everyday interactions and constitutes a crucial element of person perception (Gillath et al., 2012; Kraus et al., 2017).
When it comes to the lower social class, that perception is often negative. Classism, used here to refer to negative stereotyping and attitudes toward the lower social class (Lott, 2002), is so ubiquitous that even pre-school children already show a preference for wealthier peers, rate lower class individuals as less competent, and hold more negative stereotypes about the lower class than the middle class (Durante & Fiske, 2017). Negative portrayals of the lower social class are common in media formats ranging from scripted reality shows to respectable news magazines (Bullock et al., 2001). Despite this prevalence, classism and social class as a source of identity and exclusion have received comparatively little attention in the social psychological literature (Kraus & Stephens, 2012).
What is well-established is that in general, negative stereotypes often serve to justify and predict harmful behavior against disadvantaged groups (Bodenhausen et al., 1998; Morgan et al., 2013). There are different and evolving ways in which more powerful groups in society can perpetuate harm against disadvantaged groups such as the lower social class. For example, even as instances of more overt or formal discrimination seem to be on the decline, subtler forms, also referred to as interpersonal discrimination, prove highly persistent (Dovidio et al., 2018; Ruggs et al., 2011). This includes behaviors such as social distancing or avoidance of outgroup members that are not illegal or formally reprimandable (Hebl et al., 2002). Here, when we use the term discrimination, these are the types of behavior that we refer to.
While rising inequality has been shown to be detrimental to societies as a whole (Wilkinson & Pickett, 2009), it also disproportionally affects those at the bottom of the economic ladder (Sommet et al., 2018). Therefore, opposing redistributive measures that would reduce inequality and allocate resources to those who lack them, can also be construed as a type of harm against the lower class. Arguably, it can even be considered more damaging than discriminatory behavior because it maintains the material disadvantages that affect outcomes and opportunities (Carnevale & Rose, 2003; Lachman & Weaver, 1998). Thus, if opposition to redistribution does have similar roots to discrimination, classism might be a factor that leads middle and upper class people to not only avoid close contact with the lower class but also to reject measures that would reduce social inequality.
The Ideological Perspective: Capitalism and Redistribution
However, while both are harmful to the lower class, discrimination and opposing redistribution also differ in important ways. Discriminatory behavior usually takes place in the interpersonal realm and while it has been shown that discrimination against lower status groups is more common in political conservatives (von Collani & Grumm, 2009), avoiding contact with such outgroups is not considered a political act per se. Redistribution, on the other hand, takes place in the realm of economic policy and does not only affect those who benefit from these measures but also the broader structures of society and the economy. Thus, opposition to redistributive measures can also be understood as primarily a political preference rooted in the free market ideology of capitalism.
Capitalism in its current neoliberal iteration is the dominant economic ideology of the present era (Harvey, 2005) and as such constitutes the main paradigm through which policies and ideas are evaluated (Palley, 2004). Therefore, it makes sense to assume that endorsement of capitalist ideology plays a role in explaining relatively low levels of support for redistribution (Hing et al., 2019), especially given that, despite galloping inequality, these levels have stagnated since the 1970s when neoliberal laissez-faire capitalism began to replace the more heavily government regulated Keynesian model (Kenworthy & McCall, 2008).
Crucially, one of the central tenets of this ideology is the notion that government interference in the economy, particularly in the form of taxation or welfare, is inefficient and wasteful and an infringement on individual freedom (Friedman, 1962; Hayek, 1960). The role of the state, according to neoliberal capitalism, is to ensure the freedom of the market, but not to try and counterbalance its effects, including rising inequality (Jones, 2014). Thus, if the market rather than democratic government constitutes the best mechanism to efficiently allocate resources (Friedman, 1962; Harcourt, 2010), redistributive policies are unlikely to be met with approval from those who endorse capitalist beliefs. Importantly, this opposition to redistribution is not necessarily tied to perceptions of its recipients. Indeed, firm believers in capitalist ideology should not be swayed in their position on redistribution even if they held entirely positive views of the lower class. Similarly, these beliefs do not have to go hand in hand with a preference for higher inequality, but merely the view that governments should not interfere to reduce it.
The Present Research
The aim of the present research is to contribute to our understanding of why individuals oppose redistributive policies. We include three explanations in our model based on the goal, means, and recipients of redistribution. We also include discrimination against the lower social class as a second outcome which allows us to investigate how opposing redistribution differs from interpersonal harmful behavior toward the lower class in its antecedents (Figure 1).
Figure 1
Proposed Model
To assess rejection of the goal of redistribution, a reduction of social inequality, we use Social Dominance Orientation (SDO; Sidanius & Pratto, 1999). SDO reflects a preference for inequality and hierarchically structured intergroup relations. It is usually more common in members of dominant social groups and is associated with prejudice and discrimination against low status groups (Sibley & Duckitt, 2013). Individuals high in SDO tend to oppose equality-enhancing measures and discriminate against disadvantaged groups (Sidanius, 1993). Therefore, we propose that SDO will predict both discrimination against the lower social class and opposition to redistribution.
To assess negative perceptions of the recipients of redistributive policies, we employ different measures of classism. Investigations of media depictions of the lower social class reveal that a wide range of classist stereotypes exist (Bullock et al., 2001; O. Jones, 2012). It has been argued that, more so than other forms of prejudice, classism is common among liberals as well as conservatives (Baron, 2016; Eribon, 2018; O. Jones, 2012) and that both conservative and liberal news sources reinforce class-based distinctions and pass negative judgement on the lower class (de Goede, 1996). From conservative sources, this includes depictions of the lower class as disregarding societal norms and traditions or linking lower class membership to criminality and substance abuse (Sidel, 1996). In liberal narratives, the lower class is often associated with ethnocentric and nationalist political causes like Brexit (Eatwell & Goodwin, 2018; Telford & Wistow, 2020) and thought to be particularly prone to prejudice against other marginalized groups (Metzgar, 2020). More directly blaming this group for their own economic status, the lower class is also stereotyped as lacking in work ethic, self-reliance, and ambition (Cozzarelli et al., 2001; Durante & Fiske, 2017; Lindqvist et al., 2017), attributes that are held up as desirable across the political spectrum (Coates, 2018; Peck, 2001). In order to reflect this diversity of stereotype content, we designed measures of liberal and conservative classism, based on the respective representations, for our first study and a measure of classism focusing on a perceived lack of hard work and drive for the second study.
Finally, we assess opposition to government intervention in the economy by measuring individuals’ preferences for capitalism and free markets. We predict that capitalist ideology will be associated with redistribution but not discrimination against the lower class, whereas classism should show the opposite pattern of associations. Because identification with one’s own middle or upper class ingroup may be related to more discriminatory behavior against the lower class outgroup (Hudson et al., 2003) and higher socio-economic status may increase distance to the lower class and self-interested opposition to redistributive policies, these two constructs as well as age and gender are added to the model as controls.
Study 1
Method
Participants
Power analysis for linear regression analysis with four independent variables (f2 = .15, β-1 = .95, α = .05) suggested a sample size of at least 129. Research assistants recruited 183 participants in a medium-sized Western German town and via social media. Seven participants were excluded for not completing all central measures and four were identified as multivariate outliers, leaving a final sample of 172 (44% female, Mage = 39, SD = 17.8). More than half (57%) were university educated, 41% currently employed, 40% in education or training, and 12% retired. A combined 33% indicated they would vote for conservative, right-wing or libertarian parties while 50% indicated they would vote for green, left or social-democratic parties and 17% were undecided, did not intend to vote or intended to vote for other parties. Of our participants, 17% identified as lower middle class, 51% identified as middle class, 31% as upper middle class and 1% as upper class. Participants identifying as lower class were screened out at the beginning of the survey.
Measures
Capitalism
Capitalist ideology was measured with two items (r = .66) asking participants to indicate their economic preferences on scales ranging from 1 (‘socialist’ / ‘the state should control the economy’) to 7 (‘capitalist’ / ‘control of the economy should be left to the market’).
SDO
Social Dominance Orientation was measured with the SDO-7 short scale (Ho et al., 2015) consisting of eight items (e.g., ‘Group equality should not be our primary goal’; α = .82).
Classism
Based on analyses of the media coverage of the lower social class, we designed two measures of classism. Liberal classism was measured with four items (e.g., ‘Most lower class people don’t support equal rights for minorities’, ‘Many people from the lower class hold ethnocentric views.’; α = .71). Conservative classism was also measured with four items (e.g., ‘Most lower class people don’t show enough respect for authorities’, ‘People from the lower class don’t appreciate the importance of law and order for a functioning society’, α = .74).
Discrimination
Interpersonal discrimination against the lower social class was measured with four items (α = .65) focusing on avoidance and social distancing (e.g., ‘I would rather not send my child to a school attended by lower class children’, ‘I would rather not be friends with someone from a lower class background’).
Redistribution
Two items (r = .41) measured support for redistributive policies that would benefit the lower social class (‘I’m in favor of abolishing child benefits for families that don’t need them so poor families can receive higher child benefits’, ‘I would be willing to pay a little more for public transport if that means low income people can use it for free’). We operationalized this outcome with relatively small measures that would direct resources toward the lower social class rather than mentioning contentious concepts like taxation in order not to artificially increase the effect of capitalist ideology.
Control Variables
The MacArthur ladder was used to assess subjective socioeconomic status (SES) on a 10-point scale with higher scores indicating higher status. Participants’ identification with their own social class was measured with four items (e.g., ‘I identify with other people from my social class’; α = .84). Unless otherwise indicated, items were rated on scales from 1 (‘strongly disagree’) to 7 (‘strongly agree’).
Results and Discussion
Data, code and materials are available as Supplementary Materials. Analyses were conducted in R Version 4.1.0 (R Core Team, 2021). Means, standard deviations, and intercorrelations between the variables are displayed in Table 1.
Table 1
Descriptive Statistics and Correlations With Confidence Intervals
| Variable | M | SD | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
|---|---|---|---|---|---|---|---|---|---|---|
| 1. Age | 38.74 | 17.76 | ||||||||
| 2. SES | 5.84 | 1.35 | .19* | |||||||
| [.04, .33] | ||||||||||
| 3. Identification | 3.86 | 1.45 | .13 | .32*** | ||||||
| [-.02, .28] | [.18, .45] | |||||||||
| 4. Capitalism | 3.62 | 1.35 | .12 | .24** | .14 | |||||
| [-.03, .26] | [.10, .38] | [-.01, .29] | ||||||||
| 5. SDO | 2.65 | 1.06 | .20* | .27*** | .24** | .60*** | ||||
| [.05, .34] | [.12, .41] | [.08, .38] | [.49, .69] | |||||||
| 6. Classism (conservative) | 3.35 | 1.10 | -.13 | .06 | .15 | .12 | .35*** | |||
| [-.28, .02] | [-.09, .21] | [-.01, .29] | [-.03, .27] | [.21, .48] | ||||||
| 7. Classism (liberal) | 4.12 | 1.11 | -.07 | .20** | .13 | .08 | .16* | .50*** | ||
| [-.22, .08] | [.05, .34] | [-.02, .28] | [-.07, .22] | [.01, .31] | [.37, .60] | |||||
| 8. Discrimination | 2.62 | 1.10 | .10 | .24** | .30*** | .24** | .38*** | .43*** | .37*** | |
| [-.06, .25] | [.09, .38] | [.15, .43] | [.09, .37] | [.24, .50] | [.29, .54] | [.23, .50] | ||||
| 9. Redistribution | 4.13 | 1.67 | .08 | -.14 | -.14 | -.54*** | -.47*** | -.18* | -.04 | -.24** |
| [-.07, .23] | [-.28, .01] | [-.29, .01] | [-.64, -.43] | [-.58, -.34] | [-.32, -.03] | [-.19, .11] | [-.38, -.09] |
*p < .05. **p < .01. ***p < .001.
The pattern of correlations between the constructs suggests a distinction between intergroup and ideological elements. While there was a strong correlation between the two types of classism and both were associated with SDO and discrimination, no links to capitalism emerged and only conservative classism was weakly correlated with redistribution. Moreover, discrimination, but not redistribution, was associated with higher SES and stronger class identification.
Figure 2
Path Analysis Model of Associations Between Classism, SDO, Capitalism, and Discrimination and Redistribution
Note. Coefficients presented are standardized regression coefficients.
†p < .08. *p < .05. **p < .01. ***p < .001.
We conducted path analysis using the lavaan package (Rosseel, 2012) to test our regression model. Standardized coefficients are displayed in Figure 2. The results largely confirmed our hypotheses. As expected, capitalism, B = -.50, SE = .10, p < .001, 95% CI [-.69, -.30], and SDO, B = -.41, SE = .14, p = .004, 95% CI [-.69, -.13], negatively predicted redistribution while the effects of both liberal, B = .11, SE = .12, p = .354, 95% CI [-.12, .33], and conservative classism, B = -.06, SE = .12, p = .614, 95% CI [-.30, .18], were not significant. Of the control variables, only age had a significant effect with older participants showing more support for redistributive policies, B = .02, SE = .01, p = .011, 95% CI [.004, .031].1 For discrimination, conservative, B = .25, SE = .08, p = .002, 95% CI [.09, .41], and liberal classism, B = .19, SE = .08, p = .014, 95% CI [.04, .34], as well as social class identification, B = .16, SE = .06, p = .004, 95% CI [.05, .27], were significant predictors while capitalism was not, B = .03, SE = .07, p = .606, 95% CI [-.10, .17]. Against our expectations, the effect of SDO did not reach significance either, B = .17, SE = .09, p = .080, 95% CI [-.02, .35].
These results offer support for the idea that opposition to redistributive policies is best understood as an ideological position rooted in a preference for an unequal society and a rejection of government interference in the economy. Construing it as primarily another type of harmful action against the lower social class on the other hand did not emerge as a satisfactory explanation. Redistribution differed from discrimination, which was not predicted by capitalist ideology but only by classism. Importantly, both liberal and conservative forms of classism predicted discrimination, indicating that concerns about the negative effects of classist sentiments in liberal circles may be well founded (Baron, 2016; O. Jones, 2012; Tschulanov, 2020) even if these stereotypes do not seem to directly affect policy attitudes. The lack of a correlation between capitalism and classism also lends support to the thesis that a rejection of government interference in the free market is not necessarily linked to or dependent on negative perceptions of those who might benefit from such interference.
One caveat to our conclusions is that we measured classism with stereotypes that did not contain attributions about poverty. However, research has shown that stereotypes implicitly blaming the lower social class for their socio-economic status by depicting them as lazy or lacking ambition are common as well (Cozzarelli et al., 2001; Durante & Fiske, 2017; Lindqvist et al., 2017). These stereotypes are likely derived from widely endorsed meritocratic beliefs about status being based on hard work and deservingness (McCoy & Major, 2007). Meritocratic beliefs also serve to legitimize inequality (Bettache et al., 2020) and are negatively related to support for redistribution (García-Sánchez et al., 2020; Rodriguez-Bailon et al., 2017). Therefore, this form of classism cannot be separated from the ideological perspective on redistribution as clearly as the ones used in Study 1. However, this also suggests that it can provide interesting additional insights into the role of classism in redistribution support and constitute a strong test of our hypotheses. Consequently, in Study 2, we proceed to test the same model from Study 1 but with a measure of classism that includes stereotypes based on meritocratic beliefs.
Study 2
Method
Participants
We recruited 263 participants from local groups, student organizations, and via social media. Twenty-three participants were excluded because they did not complete the central measures and eight were identified as multivariate outliers, leaving a final sample of 232 (31% female, Mage = 52, SD = 18.2). Two thirds (68%) were university educated, the majority (57%) currently employed, 12% in education or training, and 26% retired. A combined 24% of participants indicated they would vote for conservative, right-wing or libertarian parties while 72% indicated they would vote for green, left or social-democratic parties and 3% were undecided or did not intend to vote. Of our participants, 24% identified as lower middle class, 46% identified as middle class, 28% as upper middle class and 2% as upper class. Participants who identified as lower class were screened out at the beginning of the survey.
Measures
We used the same measures of capitalism (r = .69), SDO (α = .82), interpersonal discrimination (α = .70), and redistribution (r = .30), as well as socioeconomic status and class identification (α = .84) as in Study 1. Classism was measured with six items (e.g., ‘People from the lower class often lack ambition’, ‘Most lower class individuals don’t manage their money responsibly’; α = .91) based on the classism subscale of the Intolerant Schema measure (Aosved et al., 2009).
Results and Discussion
Descriptive statistics and correlations between variables are displayed in Table 2. Unlike the ones in Study 1, our new measure of classism was associated with capitalism which supports the idea that the stereotypes contained in it stem from meritocratic beliefs. However, consistent with Study 1 results, classism was strongly linked to discrimination but only weakly linked to support for redistribution.
Table 2
Descriptive Statistics and Correlations With Confidence Intervals
| Variable | M | SD | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
|---|---|---|---|---|---|---|---|---|---|
| 1. Age | 51.66 | 18.15 | |||||||
| 2. SES | 5.91 | 1.54 | -.03 | ||||||
| [-.16, .10] | |||||||||
| 3. Identification | 3.95 | 1.55 | .07 | .13* | |||||
| [-.06, .20] | [.00, .26] | ||||||||
| 4. Capitalism | 3.17 | 1.43 | .02 | .27*** | .09 | ||||
| [-.11, .15] | [.14, .38] | [-.04, .21] | |||||||
| 5. SDO | 2.43 | 1.15 | .09 | .21** | .10 | .61*** | |||
| [-.04, .22] | [.08, .33] | [-.03, .23] | [.52, .68] | ||||||
| 6. Classism | 3.05 | 1.34 | .02 | .12 | .21** | .36*** | .41*** | ||
| [-.11, .15] | [-.01, .25] | [.08, .33] | [.24, .47] | [.30, .52] | |||||
| 7. Discrimination | 2.16 | 1.08 | -.05 | .12 | .16* | .40*** | .48*** | .60*** | |
| [-.18, .08] | [-.01, .25] | [.03, .28] | [.28, .50] | [.37, .58] | [.51, .68] | ||||
| 8. Redistribution | 4.62 | 1.73 | .09 | -.11 | -.06 | -.39*** | -.42*** | -.26*** | -.27*** |
| [-.04, .21] | [-.23, .02] | [-.18, .07] | [-.49, -.27] | [-.52, -.30] | [-.38, -.14] | [-.39, -.15] |
*p < .05. **p < .01. ***p < .001.
Path analysis results also mirrored our findings from Study 1. Standardized coefficients are displayed in Figure 3. Again, capitalism, B = -.31, SE = .09, p = .001, 95% CI [-.49, -.13], and SDO, B = -.37, SE = .12, p = .002, 95% CI [-.60, -.14], but not classism, B = -.09, SE = .09, p = .300, 95% CI [-.26, .08], negatively predicted redistribution while age had a positive effect, B = .01, SE = .01, p = .025, 95% CI [.002, .024].2 Discrimination was predicted by classism, B = .38, SE = .05, p < .001, 95% CI [.29, .47], and this time, the effect of SDO was significant as well, B = .23, SE = .06, p < .001, 95% CI [.11, .36]. Capitalism did not significantly predict discrimination against the lower social class, B = -.05, SE = .05, p = .267, 95% CI [-.04, .15].
Figure 3
Path Analysis Model of Associations Between Classism, SDO, Capitalism, and Discrimination and Redistribution
Note. Coefficients presented are standardized regression coefficients.
*p < .05. **p < .01. ***p < .001.
The results of Study 2 provide further support for an interpretation of opposition to redistribution as a result of interrelated beliefs about the desirability of inequality as well as the role of the state in interfering in the economy to reduce it. Comparing the effects of classism in both studies also suggests that while negative perceptions of the lower class are not a main driver of opposition to redistributive policies that would benefit this group, they likely do serve as a justification for such a position when they contain elements that blame the lower class for their disadvantages and paint inequality as fair. While they still did not significantly predict redistribution when entered into the model with SDO and capitalism, these classist stereotypes did show stronger associations with both capitalism and redistribution, which might be due to the fact that they simultaneously contain an expression of meritocratic beliefs which are a part of neoliberal capitalist ideology (Bettache et al., 2020).
General Discussion
This research set out to better understand opposition to redistributive policies, particularly in light of the apparent contradiction between high levels of concern about rising inequality across many countries without a corresponding rise in efforts to combat it through redistribution (Bartels, 2005; Ipsos, 2020; Mau & Heuer, 2016; OECD, 2011). We put forward two possible mechanisms that may contribute to individuals opposing redistribution beyond a preference for an unequal society: an intergroup perspective that conceptualizes opposition to redistribution as a form of harmful behavior against its beneficiaries, the lower social class, and an ideological perspective that views it as an endorsement of the capitalist belief that governments should not interfere in the free market. We proposed that capitalist ideology would be the most promising predictor of redistributive preferences because it is, theoretically at least, independent of both preferences for inequality and perceptions of the lower social class. That is, even individuals who are concerned about the level of inequality and hold favorable views of those most affected by it may still oppose redistributive policies if they subscribe to the idea that the market is the best mechanism for allocating resources. On the other hand, individuals might hold negative stereotypes about the recipients of redistributive policies and still support redistribution when it is in line with their ideological positions.
The two studies we conducted provided support for the distinction between ideological and intergroup mechanisms and the importance of the former over the latter in predicting support for redistribution. Comparing redistribution with interpersonal discrimination revealed distinct patterns of associations whereby capitalism and SDO were strongly linked to redistribution, but classism was not. On the other hand, discrimination was predicted by classism and, in Study 1, identification with participants’ own social class, but not by capitalism. By investigating different clusters of classist stereotypes we were also able to show that stereotype content affects these associations to some extent. Stereotypes derived from liberal and conservative media portrayals that depict the lower social class as, for example, holding ethnocentric views or lacking respect for authorities were not associated with either endorsement of capitalist ideology or support for redistribution. Stereotypes painting the lower social class as lazy and lacking in ambition and self-reliance, on the other hand, reflect widely endorsed meritocratic beliefs that are also a tenet of neoliberal capitalism (Bettache et al., 2020). As such, we predicted that they could not be wholly separated from the ideological factors and indeed, this form of classism was associated with capitalist ideology and redistribution. Importantly, however, it still did not emerge as a significant predictor in our model. One possible explanation is that, similar to how meritocratic beliefs can justify existing inequality (Hadarics et al., 2021), these stereotypes serve primarily as a justification for opposing redistribution rather than an antecedent.
Our results show similarities to other lines of research that compare the effects of intergroup or interpersonal relations and ideological factors on attitudes toward social change, particularly where economic or material issues are concerned. Studies on intergroup contact have found that while positive encounters improved advantaged group members’ attitudes toward disadvantaged outgroups, this did not automatically translate to increased support for a more equal distribution of resources (Saguy et al., 2009). However, advantaged group members were more likely to engage in collective action on behalf of a disadvantaged group’s equality when intergroup contact was not only positive, but also entailed explicit political communication about the unequal group status (Becker & Wright, 2022). Similarly, vicariously experiencing financial hardship via one’s friends only affected support for government efforts to address inequality if political discussions occurred within the friendship network (Newman, 2014). Finally, interactive and personalized treatments to increase empathy with families on the poverty line lead to increased perception of poverty and inequality as serious issues, but affected redistributive attitudes only when information about the effectiveness of such policies was provided, explicitly challenging the capitalist belief that government spending is wasteful (Kuziemko et al., 2015).
Therefore, it would seem that combating negative stereotypes and improving intergroup relations alone is unlikely to achieve significant progress in the fight against rising economic inequality as long as the underlying ideology remains. In fact, it has been shown that rising inequality leads to increased meritocratic perceptions of the rich as competent and the poor as incompetent (Connor et al., 2021) rather than the other way around. This supports the idea that at least some forms of classism may be consequences of and justifications for a lack of redistribution more than a cause of it. And while exposure to more unequal societies can give rise to populist resentment against the elite (Hartwich & Becker, 2019), it also lowers support for redistribution (Roth & Wohlfart, 2018), again suggesting that the key to understanding the self-perpetuating nature of economic inequality lies less in intergroup perceptions and primarily in the widespread acceptance of the key tenets of neoliberal capitalism that present inequality as inevitable and redistribution as an infringement on human freedom (Hayek, 1960; Monbiot, 2016).
Limitations and Directions for Future Research
The two studies presented here provided initial support for the importance of ideological over intergroup factors in explaining support for redistribution. However, despite the largely consistent results across both studies, some caution is warranted when it comes to the generalization and interpretation of these findings. Because both studies used correlational datasets, any inferences about causation can only be tentative and the mechanisms explored here need to be tested further in experimental or longitudinal studies.
Moreover, our research focused on the context of social class and classism which has received comparatively little attention in the intergroup literature. While we believe that the insights into the relationship between different forms of classism and discrimination as well as capitalist ideology and redistribution constitute an important contribution, they are also specific to this context. It might be of interest for future research to examine whether different mechanisms emerge when another, intersecting intergroup dimension is added, such as redistribution in the form of reparations for Black Americans. Given the link between opposition to welfare spending in the U.S. and negative stereotypes about Black people (Cooley et al., 2019; Gilens, 1999), it might well be the case that intergroup factors play a more central role in predicting attitudes toward redistribution when the focus is on social categories other than class. Similarly, it might be of interest to investigate the associations we found in countries that differ from Germany with regard to their implementation of neoliberal principles of economic freedom (Bettache et al., 2020).
Finally, even though the patterns of results were consistent, correlations between our two redistribution items were lower than desirable which limits the reliability of this measure. This may be because individuals’ support for specific policies is influenced by factors beyond their support for redistribution in general. Because we explored different stereotype content in our classism scales, we kept the redistribution measure the same across both studies to improve comparability. However, our conclusions would be strengthened by future research corroborating these findings using a broader range of redistributive policies or a more abstract measure of redistribution attitudes. We also made an effort to choose the measures in these studies in such a way that capitalist ideology and opposition to redistribution did not measure identical attitudes, but it is still possible that they tap into the same latent construct.
Conclusion
Social inequality is widely unpopular and can have detrimental effects even on those who are not materially deprived by it (Wilkinson & Pickett, 2009). Despite widespread consensus that it constitutes one of the most pressing global issues of our time (Ipsos, 2020), its rise is continuing and even accelerating (Piketty & Saez, 2014; UN DESA, 2020) without a corresponding increase in support for redistributive policies (Kenworthy & McCall, 2008). The research presented here lends support to explanations stressing the role of the dominant economic system (Hing et al., 2019) by showing that ideological, more so than intergroup mechanisms appear to be central to opposition to redistribution. Thus, we conclude that while challenging negative perceptions of those suffering most from rising inequality is important in reducing interpersonal discrimination, it is unlikely to lead to much progress in tackling the root problem unless dominant narratives of neoliberal capitalism are also challenged.