1. Introduction
The unequal distribution of wealth is an important part of the increase in economic inequality that has occurred in Western rich democracies during the past decades (Alvaredo et al., Reference Alvaredo, Chancel, Piketty, Saez and Zucman2017; Piketty, Reference Piketty2014). While the abolition of taxes on private wealth in the 1990s and 2000s has contributed significantly to this development (Glennerster, Reference Glennerster2012; Hacker and Pierson, Reference Hacker and Pierson2010; Piketty, Reference Piketty2014; Saez and Zucman, Reference Saez and Zucman2019; Volscho and Kelly, Reference Volscho and Kelly2012), survey evidence suggests that majorities of the population support the taxation of wealth (Dallinger, Reference Dallinger2021; Rowlingson et al., Reference Rowlingson, Sood and Tu2021). Furthermore, issues of wealth taxation have gained momentum in the wake of the financial crisis of 2008ff. and resurged in recent debates about distributing the financial burdens resulting from the COVID-19 pandemic (BBC, 2020). Thus, understanding when and under what conditions citizens in rich democracies support wealth taxation is important to inform public policy choices at the beginning of the 21st century.
The literature on popular support for redistribution through taxation and welfare state policies often sets out from a political economy perspective. Therein, citizens are supposed to support redistributive measures out of rational self-interest, i.e. when they expect to gain from them (Meltzer and Richard, Reference Meltzer and Richard1981; Romer, Reference Romer1975). Yet, recent studies find that people often misperceive the extent of overall economic inequality as well as their own position within the income and wealth distribution (Bellani et al., Reference Bellani, Bledow, Busemeyer and Guido2021; Bublitz, Reference Bublitz2020; Engelhardt and Wagener, Reference Engelhardt and Wagener2018; Gimpelson and Treisman, Reference Gimpelson and Treisman2018), thus biasing the cognitive foundations that underlie the formation of redistributive preferences. In contrast, the deservingness perspective stipulates that support for redistribution depends on whether people regard economic outcomes to be justly deserved or not (Laenen et al., Reference Laenen, Rossetti and van Oorschot2019; Van Oorschot, Reference Van Oorschot2000). While previous studies show that attitudes towards wealth taxation are influenced by indicators of self-interest, (mis)perceptions of the status quo of inequality and deservingness assessments (Alesina and Angeletos, Reference Alesina and Angeletos2005; Groß and Lang, Reference Groß and Lang2018; Limberg, Reference Limberg2020; Ragusa, Reference Ragusa2014; Rowlingson and Connor, Reference Rowlingson and Connor2011; Skilling and McLay, Reference Skilling and McLay2014), extant research has not addressed how self-interest and deservingness relate to each other.
Therefore, this article aims to contribute to the state-of-the-art by investigating how indicators of self-interest, both objective and subjective, as well as popular deservingness opinions about the wealthy affect popular support for wealth taxation in Germany. Specifically, we not only ask how these factors shape popular support for wealth taxation, but also how they interact. Do notions of ‘undeserved’ wealth dampen opposition to wealth taxation among wealthy citizens, thereby off-setting their self-interest? Or does, vice versa, an image of the wealthy as ‘deserving’ decrease support for wealth taxation among the non-wealthy, whose self-interest would otherwise prompt them to favour it? Through investigating how deservingness and self-interest interact, our study contributes to an emerging literature on popular support for progressive taxation (Barnes, Reference Barnes2014; Edlund, Reference Edlund2016; Liebig and Mau, Reference Liebig, Mau, Mau and Veghte2007; Limberg, Reference Limberg2020; Limberg, Reference Limberg2019; Ragusa, Reference Ragusa2014; Sumino, Reference Sumino2015). While these studies are inspired by research on preferences for redistribution and welfare attitudes (Svallfors, Reference Svallfors, Castles, Leibfried, Lewis, Obinger and Pierson2010), attitudes towards taxation are still less well understood (Barnes, Reference Barnes2014). Specifically, the relative contributions of various explanatory factors and their potential interrelations have not yet been fully appreciated.
To study these issues empirically, we use a vignette design which was embedded in a large standardized survey on the perception of inequality in Germany from 2017. In contrast to regular survey items, which are often relatively abstract, this approach accommodates for the fact that issues of taxation are ‘complex and context-dependent’ (Liebig and Mau, Reference Liebig, Mau, Mau and Veghte2007: 100) and allows for more realistic and rich descriptions of fictitious wealthy persons for which we vary different characteristics to examine how these affect support for wealth taxation.
We focus on Germany because wealth inequality there is comparatively high from a cross-national perspective (Balestran and Tonkin, Reference Balestran and Tonkin2018). Recent research shows that the wealthiest 10 percent in Germany hold over two thirds of overall wealth (Schröder et al., Reference Schröder, Bartels, Göbler, Grabka and König2020), compared to a European average of 52 percent of wealth owned by the richest 10 percent (Balestran and Tonkin, Reference Balestran and Tonkin2018). Also, the increase in income inequality between 2000 and 2005 was driven by gains at the top of the income distribution (Corneo et al., Reference Corneo, Zmerli, Pollak and Nolan2014; Groh-Samberg, Reference Groh-Samberg2016), which further contributed to a concentration of economic advantage. While public support for progressive taxation in Germany is generally strong (Barnes, Reference Barnes2014: 71) and two-thirds of the population advocate for higher taxation of private wealth in political opinion polls (Heinrich et al., Reference Heinrich, Jochem and Siegel2016), the personal net wealth tax has been abandoned since 1997 due to a ruling by the Federal Constitutional Court on the unequal tax treatment of real estate and financial wealth.Footnote 1 Although the fiscal relevance of a reintroduction of this personal wealth tax is subject to debate – estimates range between € 10 to 20 billion Euros annual revenue for a tax rate of 1-1.5 percent (Bach and Thiemann, Reference Bach and Thiemann2016: 83) –, the personal net wealth tax is of great symbolic relevance because it reflects the ‘commitment of the state to redress some of the more egregious inequalities generated by the market economy’ (Bird 1989: 441, cited in Banting, Reference Banting1991: 352).
The article is structured as follows. The next section (2) presents the conceptual framework of the article and discusses political economy and deservingness perspectives on wealth taxation. Thereafter, section 3 discusses the methods and data used, while section 4 presents the results. The concluding section (5) summarizes the findings and discusses limitations as well as promising future research avenues
2. Theoretical perspectives on popular support for wealth taxation
2.1 Political economy perspectives on wealth taxation
Political economy perspectives on attitudes towards public policy and the role of government assume that individuals act rationally according to their economic self-interest. Principally, two arguments are made: first, as income is distributed unequally, the median voter receives less than the mean income and therefore profits from each lump-sum transfer financed through a linear tax on income. Hence, poorer citizens should demand greater redistribution and support progressive taxation, whereas wealthier individuals should be in favor of less redistribution and taxation (Meltzer and Richard, Reference Meltzer and Richard1981; Romer, Reference Romer1975). Second, redistributive expenditures of the state can be seen as a form of insurance supported by risk-averse citizens (Moene and Wallerstein, Reference Moene and Wallerstein2001; Rehm, Reference Rehm2009). Thus, individuals exposed to greater economic risk – such as the unemployed or those at risk of losing their job – should also support taxing the rich more heavily.
Empirically, studies find that popular support for progressive taxation in general is strong among individuals with low incomes, and declines as income rises (Barnes, Reference Barnes2014; Limberg, Reference Limberg2020; Page et al., Reference Page, Bartels and Seawright2013; Sumino, Reference Sumino2015). Indicators of economic risk, such as unemployment or part-time employment, are not significantly associated with greater support for progressive taxation (Limberg, Reference Limberg2020: 183; Sumino, Reference Sumino2015: 1125). Hence, we formulate the following baseline hypothesis:
H1a: Support for wealth taxation is higher among low-income respondents as compared to high-income respondents.
However, recent studies have also found that people tend to misperceive the amount of aggregate inequality in society and their own position within the income distribution (Bublitz, Reference Bublitz2020; Gimpelson and Treisman, Reference Gimpelson and Treisman2018; Engelhardt and Wagener, Reference Engelhardt and Wagener2018; Hauser and Norton, Reference Hauser and Norton2017). While individuals in the upper levels of the hierarchy tend to underestimate their socio-economic standing, individuals in the lower levels tend to overestimate their position (Cansunar, Reference Cansunar2020; Hauser and Norton, Reference Hauser and Norton2017). Thus, the formation of redistributive preferences might be subject to cognitive bias (Kenworthy and McCall, Reference Kenworthy and McCall2007), which makes it important to take into account people’s own subjective assessment of their position in the socio-economic hierarchy and its implications for policy preferences (Bobzien, Reference Bobzien2019; Dallinger, Reference Dallinger2021). Therefore, we formulate the following hypothesis:
H1b: Individuals’ subjective assessment of self-interest, and especially whether they would benefit from redistribution, should increase their support for a wealth tax.
Another implication is that correcting for biased perceptions – for instance, by providing information about the actual extent of wealth inequality – might change individuals’ redistributive preferences. Indeed, research shows that providing information about actual inequality can increase redistributive preferences (Zimmermann et al., Reference Zimmermann, Heuer and Mau2018), especially if these inequalities indicate a disadvantage for the individual (Engelhardt and Wagener, Reference Engelhardt and Wagener2018). We thus formulate the following hypotheses:
H1c: Information on the unequal distribution of wealth in Germany should increase support for a wealth tax.
2.2 Deservingness opinions
Besides self-interest, several studies have shown that redistributive preferences and attitudes towards public policy are also influenced by opinions regarding the deservingness of specific social groups (Laenen et al., Reference Laenen, Rossetti and van Oorschot2019; Van Oorschot, Reference Van Oorschot2000). Initially, the deservingness perspective originated in studies that sought to explain weak popular support for social policies aimed at the poor (Feagin, Reference Feagin1975; Van Oorschot, Reference Van Oorschot2000; Laenen, Reference Laenen2020). A prominent conception developed by Van Oorschot (Reference Van Oorschot2000) suggests that people apply five criteria to determine the deservingness of the beneficiaries of social policies: control (beneficiaries are not personally responsible for their situation); attitude (beneficiaries are thankful for the benefits received); reciprocity (beneficiaries have contributed previously to social security schemes, or can be expected to do so in the future); identity (beneficiaries are similar to mainstream society); and need (beneficiaries are genuinely needy).
While the CARIN conception has mostly been applied to socio-economically deprived groups, such as the poor, welfare recipients or immigrants (Laenen et al., Reference Laenen, Rossetti and van Oorschot2019; Reeskens and van der Meer, Reference Reeskens and van der Meer2018; Van Oorschot, Reference Van Oorschot2000), researchers have recently also started to investigate how deservingness opinions about privileged groups, such as the rich, impact preferences for redistribution and taxation (Limberg, Reference Limberg2020; McCall, Reference McCall2013; Rowlingson and Connor, Reference Rowlingson and Connor2011; Skilling and McLay, Reference Skilling and McLay2014). The key idea is that if economic outcomes – such a high income or wealth ownership – are regarded as deserved, support for redistribution through taxation should be lower than when they are viewed as undeserved. A fundamental criterion according to which the rich are regarded as deserving is when they have acquired their wealth consistent with meritocratic principles. Therefore, we follow Limberg’s (2020: 174) differentiation between procedural and behavioral deservingness as two analytically distinct sub-dimensions of ‘deservingness valuations’Footnote 2 grounded in meritocracy beliefs.Footnote 3 This distinction differs somewhat from the CARIN conception, though there are overlaps.
Procedural deservingness refers to the degree to which economic success is regarded as independent from family background (cf. Limberg, Reference Limberg2020: 174), therein resembling the control and to some extent also the identity criterion of the CARIN conception (Van Oorschot, Reference Van Oorschot2000). Although modern societies regard themselves as meritocracies, in which individual economic outcomes should not depend on one’s family of origin (Parsons, Reference Parsons and Laumann1970), the accumulation of wealth is particularly ambivalent in this regard. While meritocratic wealth acquisition is principally possible, e.g. through savings from earned income (Skopek et al., Reference Skopek, Buchholz and Blossfeld2014), non-negligible sums of wealth are transferred through inheritance, which are distributed unequally (Beckert, Reference Beckert2008; Tiefensee and Grabka, Reference Tiefensee and Grabka2017). Thus, initial advantages in individuals’ starting conditions, as reflected by a privileged family background, should increase support for the taxation of wealth. In line with Scheve and Stasavage (Reference Scheve and Stasavage2016), taxation can thus be seen as a compensation for benefits that the rich and wealthy enjoy elsewhere, for instance, ‘when it is clear that effort and talent have played less of a role than luck in accumulation of their earnings’ (Scheve and Stasavage, Reference Scheve and Stasavage2016: 47).Footnote 4 Thus, we formulate the following hypothesis:
H2a: If a wealthy person comes from a poor family, support for wealth taxation should be lower, whereas it should be higher if a wealthy person comes from a wealthy family.
Behavioural deservingness refers to the role that individual effort and merit play in the accumulation of wealth (Limberg, Reference Limberg2020: 174), which also mirrors the control dimension in the CARIN scheme (Van Oorschot, Reference Van Oorschot2000). Accordingly, Rowlingson and Connor (Reference Rowlingson and Connor2011: 440) note ‘that rich individuals may be seen as deserving if they are considered to be ‘responsible’ for their situation by working hard and taking advantage of the opportunity to do well.’ Conversely, wealth is regarded as ‘undeserved’ when it is acquired non-meritocratically, e.g. through inheritance or stock-market speculation. Studies find that whether wealth was acquired meritocratically or not indeed matters for citizens’ support for taxation (Limberg, Reference Limberg2020). For instance, Liebig and Mau (Reference Liebig, Mau, Mau and Veghte2007: 110-111) show that respondents would support higher taxes on income derived mainly from inheritance while supporting lower tax rates on income derived mainly from work.Footnote 5 Similarly, Ragusa (Reference Ragusa2014: 348) demonstrates that the approval of greater taxation of the rich is particularly influenced by perceptions of whether these individuals have earned their wealth and/or whether it stems from inheritance. Overall, questions of how wealth is acquired should play an important role in justifying taxation. Thus, we formulate the following hypothesis:
H2b: If a wealthy person has acquired his/her wealth through his/her own work, support for a wealth tax should be lower. On the other hand, if a wealthy person has acquired his/her fortune through luck (inheritance, marriage) or coincidence (speculation), support for a wealth tax should be higher.
2.3 Interactions between self-interest and deservingness
In most of the literature, self-interest and deservingness opinions are treated as separate – and at times competing – explanatory factors (Cook and Barrett, Reference Cook and Barrett1992; Van Oorschot, Reference Van Oorschot2000; Sachweh et al., Reference Sachweh, Ullrich, Christoph, Mau and Veghte2007). While previous research has investigated how the importance of deservingness criteria varies in accordance with social values – e.g. cultural orientations, humanitarianism or political orientations (Gielens et al., Reference Gielens, Roosma and Achterberg2019; Hansen, Reference Hansen2018; Petersen et al., Reference Petersen, Slothuus, Stubager and Togeby2011) –, the interaction between deservingness and self-interest has received scant attention. However, it is not implausible to assume that, for some social groups, deservingness might matter less than for others, and that therefore both factors interact. For instance, given their socio-economic position and the potential losses incurred by redistributive policy, high-income groups have a strong self-interest to oppose wealth taxation, regardless of whether they regard wealth as deserved or undeserved. Therefore, one might expect that among privileged groups, notions of ‘undeserved wealth’ might be less important than among disadvantaged groups. In turn, research has also shown that meritocracy beliefs form an important part of the ‘moral economy’ of affluent groups (Sachweh, Reference Sachweh2017). Hence, notions of ‘deserved wealth’ might increase the opposition to wealth taxation among the affluent.
In turn, disadvantaged groups are not at risk of experiencing economic losses through wealth taxation. Therefore, their self-interest generally prompts them to support a wealth tax. This support might increase if they regard wealth as undeserved, or decrease if wealth is viewed as deserved. Altogether, then, the relative importance of deservingness valuations for different social groups can be hypothesized to depend on whether or not these deservingness valuations align with the respective group’s self-interest. Thus, we formulate the following hypotheses:
H3a: Among privileged groups, notions of ‘deserved wealth’ further decrease support for wealth taxation, while ‘undeserved wealth’ should be off-set by their self-interest.
H3b: Among disadvantaged groups, notions of ‘undeserved wealth’ increase support for wealth taxation, while ‘deserved wealth’ should be off-set by their self-interest.
3. Data and methodological approach
Our analysis is based on data from a primary survey on inequality and justice perceptions in Germany.Footnote 6 A total of 2,089 German-speaking individuals in private households aged 18 and over were surveyed in a multi-stage three-tier random sampling framework and interviewed personally between April and June 2017.Footnote 7 In addition to the vignette analysed in this article, this survey included items that measured citizens’ perceptions of the extent, reasons and consequences of social inequality as well as extensive sociodemographic information (Thümmel and Geiss, Reference Thümmel and Geiss2017).
The vignette on the wealth tax was presented to the respondents at the end of the first quarter of the survey. Vignette designs have various advantages over item-based studies, which we now discuss.
Vignettes are short texts that describe a hypothetical situation or person and contain certain characteristics or stimuli that vary either experimentally or randomly between the vignettes (Auspurg et al., Reference Auspurg, Hinz, Liebig, Sauer, Engel, Jann, Lynn, Scherpenzeel and Sturgis2015). As there is currently no wealth tax levied in Germany and the concept may represent a ‘fictitious issue’ for some (e.g. younger) respondents (Sturgis and Smith, Reference Sturgis and Smith2010), this approach is better suited for measuring attitudes towards wealth taxation than, for example, an item-based survey, in which the risk of statistical artefacts due to social desirability cannot be ruled out. In contrast to a purely item-based query, vignette scenarios are closer to real-world situations due to their greater complexity (Rossi and Anderson, Reference Rossi, Anderson, Rossi and Nock1982). This characteristic prevents respondents from focusing their attention on a single characteristic, which reduces the tendency towards socially desirable response behaviour (Auspurg et al., Reference Auspurg, Hinz, Liebig, Sauer, Engel, Jann, Lynn, Scherpenzeel and Sturgis2015).
In our survey, we followed Jann (Reference Jann2008: 108-112) in relying on a so-called between-subjects design in which each respondent evaluated only one vignette. This kind of design avoids socially desirable response behaviour as well as learning- and fatigue effects, which can result from the evaluation of several vignettes (Auspurg et al., Reference Auspurg, Hinz and Liebig2009; Jann, Reference Jann2008: 110-111). Thus, each of our respondents was randomly assigned a description of the financial situation of a fictitious Mr. Müller and asked whether or not he/she agreed that Mr. Müller’s wealth should be taxed (see Figure 1). The agreement with this question is the dependent variable of our research design. Coding was reversed so that high values indicate support for a wealth tax.
Figure 1. Vignette design and variations.
Source: own depiction, varying characteristics underlined.
In the description, three characteristics varied according to our hypotheses (underlined in the Figure): information on the distribution of wealth in Germany (information/no information), information on the social origin of Mr. Müller (poor family/wealthy family), and information on the ways in which wealth was accumulated (work/earnings/marriage/stock-market speculation).Footnote 8 Considering all combinations of these three characteristics resulted in a comparatively small vignette universe of 16 vignettes (2x2x4). In the survey, no combination of characteristics was excluded, and each vignette was evaluated several times (130 times on average). Correlations between the factors were thus minimized (orthogonal design).Footnote 9
The characteristics varied in the vignettes represent the central independent variables in our analyses. By randomly varying information about wealth inequality as well as our fictitious Mr. Müllers social origin and mode of wealth accumulation, the effect of these features on respondents’ support for wealth taxation can be estimated by regression analysis and interpreted causally because of the orthogonality of the design (cf. Jann, Reference Jann2008: 112).
Even though the dependent variable was measured on a 7-point ordinal scale, we analyse our data using OLS regression because a Brant test indicated the violation of the parallel regression assumption for some model specifications (Long, Reference Long1997). The three vignette features were coded nominally. As reference categories, we chose ‘no information on wealth distribution’ and ‘poor family’ in the case of the two dichotomous characteristics (‘information on wealth distribution’ and ‘social origin’). For the modes of wealth accumulation, we chose ‘work’ as the reference category because this was the only meritocratic characteristic. In addition to the vignette characteristics, two indicators of respondents’ characteristics self-interest were used to operationalize self-interest: income and their subjective gain from a more egalitarian income distribution. Control variables include education, age, gender, and region (East/West).
Income was included in the analysis as a categorical variable so that possible non-linear effects can be detected. This variable was based on respondents’ monthly net household equivalent income. The categorisation was based on median income and divided the respondents into three groups: low incomes (less than 70 percent of median net equivalent household income); medium incomes (between 70 and 150 percent of the median); and high incomes (above 150 percent of the median). The group of low incomes was chosen as the reference category. Subjective self-interest was measured with the following question:
‘If incomes were to become more equal in Germany, some people would receive a higher income and some would receive a lower income. Do you think that your personal income would then (1) definitely increase, (2) probably increase, (3) stay the same, (4) probably decrease, or (5) would definitely decrease?’
This item was recoded into three categories: gain from redistribution (answer categories 1 and 2), no change (answer category 3), and loss from redistribution (answer categories 4 and 5).
Age was measured in years and divided by ten such that one-unit increases represented a ten-year increase in age. Gender and region were dummy coded (1 = female, 1 = East Germany). Education was grouped into three categories: low (Certificate of Secondary Education), medium (General Certificate of Secondary Education), and high (General Qualification for University Access).
4. Results: Support for wealth tax among the German population
In the following, we first examine how characteristics of the wealthy (as measured in our vignette) influence respondents’ support for the taxation of wealth. Figure 2 displays the effects of the vignette characteristics and the respondent’s characteristics indicating objective and subjective self-interest in graphical form (see also Model 1 in Table A.1 in the appendix).
Figure 2. Support for wealth taxation by vignette characteristics and self-interest, multivariate results.
Source: ‘Inequality and Justice Orientations in Germany” survey 2017, weighted data. R2 = 0.06; N = 1.824.
We find that providing information on the distribution of wealth has no influence on the approval of the wealth tax, whereas both coming from a wealthy family and non-meritocratic wealth accumulation via inheritance, marriage, or stock-market speculation significantly increase support. Hypotheses H2a and H2b are thus supported. The effect of social origin points in the expected direction, thereby confirming the assumption formulated in H2a that wealth taxes are justified as a compensation for the advantages of a privileged origin. Regarding H2b, the regression analyses display a clear ranking according to which stock-market speculation has the strongest impact on the support for wealth taxation, followed by marriage and inheritance. These results suggest that work, as an indicator of meritocratic wealth accumulation, is rewarded in the sense of H2b, whereas indicators of wealth accumulation via chance and luck – such as inheritance, marriage and stock market speculation – have the opposite effect. Accordingly, H2b is supported. On the one hand, respondents seem to consider procedural deservingness in their evaluations as the taxation of wealth is regarded as a compensation for the initial advantage of a privileged origin. On the other hand, they also take into account behavioural deservingness, as the ‘effortless’ acquisition of wealth in the form of inheritance, marriage, and stock-market speculation increases support for wealth taxation, whereas meritocratic wealth accumulation through work lowers support.
With regard to the indicators of self-interest represented in the vignette characteristics, Hypothesis H1c, according to which knowledge of the unequal distribution of wealth in Germany should increase support for the taxation of wealth, must be rejected in this form. While the possibility that the information contained in the vignettes was not new to respondents and therefore had no effect cannot be ruled out, another plausible interpretation might be that the way in which the information was presented was not optimal. For instance, a graphical visualization of the wealth distribution might have been more powerful, as some previous studies suggest (Cansunar, Reference Cansunar2020; Fernández-Albertos and Kuo, Reference Fernández-Albertos and Kuo2015; Franko et al., Reference Franko, Tolbert and Witko2013).
Figure 2 also displays how the respondents’ characteristics serving as indicators of self-interest – income and subjective self-interest – are related to support for wealth taxation. While there is no significant association between income and wealth tax support, which does not support Hypothesis H1a, the expectation of not benefiting or even losing from a more egalitarian income distribution is associated with lower support for wealth taxation, which supports Hypothesis H1b. Thus, subjective assessments of one’s own economic position and how it might be affected by a more egalitarian distribution matter more for wealth tax support than objective indicators of self-interest (Bellani et al., Reference Bellani, Bledow, Busemeyer and Guido2021; Bobzien, Reference Bobzien2019; Cansunar, Reference Cansunar2020; Dallinger, Reference Dallinger2021). Interestingly, this is not only the case for those expecting to lose under more egalitarian conditions – e.g., by having to pay higher taxes – but also for those who expect no change. This suggests that opposition to wealth taxation also extends into groups who do not expect economic losses if incomes were distributed more equally.
Importantly, studies on perceived inequality and the effects of information indicate that the impact of information about inequality matters most if it relates to individuals’ own socio-economic positioning (Engelhardt and Wagener, Reference Engelhardt and Wagener2018; Fernández-Albertos and Kuo, Reference Fernández-Albertos and Kuo2015; Franko et al., Reference Franko, Tolbert and Witko2013). Therefore, we interacted our indicators for objective and subjective self-interest with the information dimension of the vignette. The results are displayed in Figure 3, which shows how the predicted level of wealth tax support differs between those with and without information on aggregate wealth inequality across objective and subjective self-interest (cf. Models 2 and 3 in Table A.1). While there is no interaction with income, information on wealth concentration in Germany significantly interacts with respondents’ subjective self-interest.Footnote 10 Providing information increases support for wealth taxation among respondents who expect no change from a more egalitarian income distribution, while it does not affect attitudes of those expecting to lose under a more egalitarian distribution. Thus, providing information on wealth concentration apparently can counteract the otherwise negative attitudes towards a wealth tax among citizens who themselves do not expect to be affected by greater economic equality.
Figure 3. Interaction: Self-Interest and Information on Wealth Concentration.
Source: ‘Inequality and Justice Orientations in Germany” survey 2017, weighted data. R2 = 0.06; N = 1.824.
Next, we explore how self-interest interacts with the deservingness valuations described in the vignette (see Figures 4 and 5 and Table A.2 in the appendix). First, in Figure 4 we look at interactions between the procedural deservingness dimension with income and subjective self-interest. Our findings show that deservingness valuations matter particularly among low-income groups or those subjectively expecting to benefit from redistribution, respectively. Here, the predicted level of support for wealth taxation is significantly higher if the wealthy come from a wealthy family rather than from a poor family. Interestingly, low levels of wealth tax support can be found among low-income respondents when the wealthy are supposed to come from a poor family. This indicates that procedural deservingness valuations reflecting meritocratic ideals affect tax preferences particularly among low status-groups (both objectively and subjectively defined), which partly contradicts Hypothesis H3b. By contrast, procedural deservingness valuations do not significantly affect the predicted level of wealth tax support among middle- or high-income groups, nor among those who subjectively expect no change or income losses under a more egalitarian distribution. Hence, self-interest apparently counterbalances the otherwise positive effect of procedural deservingness on wealth tax support for these groups. Thus, even though the affluent often display strong meritocratic orientations in prior studies (Sachweh, Reference Sachweh2017), these orientations do not translate into corresponding redistributive preferences but appear to be offset by potential threats to their socio-economic standing. This finding is in line with Hypothesis 3a.
Figure 4. Interaction: Self-interest and procedural deservingness.
Source: ‘Inequality and Justice Orientations in Germany” survey 2017, weighted data. R2 = 0.06; N = 1.824.
Looking at behavioural deservingness valuations in Figure 5, we see that across the indicators for objective and subjective self-interest, different modes of wealth acquisition matter for respondents’ tax preferences. Low-income groups, and those expecting to benefit from redistribution, respectively, show significantly lower levels of wealth tax support when wealth has been accumulated through work (rather than through inheritance, marriage, or speculation), thus reflecting their meritocratic orientations noted above. Middle-income groups, and those who expect no change from greater equality, are least supportive of taxation when wealth has been accumulated trough work and inheritance, which may reflect meritocratic orientations as well as the significance of inheritance as a mode of wealth acquisition in the middle-class (Korom, Reference Korom2017). Finally, high-income groups and those who expect to lose under more egalitarian conditions are particularly supportive of wealth taxation when wealth has been accumulated through luck (stock market speculation), which may indicate a distancing from paths to richness considered as “illegitimate”. While formal significance tests indicate that adding the interaction terms between deservingness valuations and objective and subjective self-interest is no significant improvement compared to model 1, some contrasts are statistically significant in the graphical display. However, taking the number of cases on which these interactions are based into account, these findings should be interpreted cautiously.Footnote 11
Figure 5. Interaction: Self-interest and behavioral deservingness.
Source: ‘Inequality and Justice Orientations in Germany” survey 2017, weighted data. R2 = 0.06; N = 1.824.
5. Discussion
In this article, we have investigated how objective and subjective self-interest, information on wealth concentration and varying conceptions of deservingness, as they are reflected in perceived characteristics of the wealthy, impact popular support for wealth taxation in Germany. We paid particular attention to the interrelation between these factors, asking to what extent notions of ‘deserved’ or ‘undeserved’ wealth may offset or amplify the impact of self-interest. Empirically, we have used a vignette design embedded in a representative standardized survey from 2017. Therein, we described a fictitious wealthy person for whom we systematically varied social background characteristics (poor vs. rich family) as well as the mode of wealth accumulation (work, inheritance, marriage, speculation). Furthermore, we also varied whether information on the extent of wealth inequality in Germany was provided or not in order to test whether correcting potential misperceptions of inequality would shift preferences in favour of taxation.
Our results show that respondents’ subjectively perceived self-interest – here operationalized through the assessment of whether one would expect to benefit or lose under a more egalitarian income distribution – had a greater impact on respondents’ tax attitudes than respondents’ actual income. This is in line with research arguing that rather than individuals’ objective location, it is the subjective assessment of their socio-economic position that affects redistributive preferences (Bellani et al., Reference Bellani, Bledow, Busemeyer and Guido2021; Bobzien, Reference Bobzien2019; Dallinger, Reference Dallinger2021; Gimpelson and Treisman, Reference Gimpelson and Treisman2018). Importantly, the absence of a significant impact of income is not due to a mediation effect of subjective self-interest (cf. Weisstanner and Armingeon, Reference Weisstanner and Armingeon2021). Since subjectively defined self-interest and income are only moderately correlated in our data (Cramer’s V = .2), (mis)perceptions of one’s own socio-economic status are an important factor to consider next to (mis)perceptions of overall societal inequality (as emphasized by Hauser and Norton, Reference Hauser and Norton2017). Future research might therefore pay greater attention to how status perceptions and (mis)perceptions of societal inequality interact in affecting redistributive preferences.
With regard to misperceptions of societal inequality, informing respondents about the degree of wealth inequality in Germany did not affect their tax preferences in our vignette design. The provision of information as such seems insufficient to shift preferences in favour of taxation. This partly contradicts previous studies that did find significant information effects (Cansunar, Reference Cansunar2020; Zimmermann et al., Reference Zimmermann, Heuer and Mau2018).Footnote 12 We do find, however, that providing information on aggregate inequality significantly interacts with subjective self-interest. Information on wealth concentration in Germany increases support for wealth taxation among respondents who expect no change from a more egalitarian income distribution, and who might therefore be indifferent to wealth taxation. As this group amounts to 44 percent of respondents in our sample, the policy implications of this interaction are significant.
With regard to deservingness valuations, our findings suggest that the legitimacy of wealth taxation strongly depends on the characteristics of the wealthy. In particular, behavioural deservingness perceptions reflecting how a potential taxpayer has accumulated his/her wealth play an important role. While ‘unearned’ wealth accumulated by chance and luck has a positive influence, wealth accumulation through work – as an indicator of merit – significantly reduces support for wealth taxation. Similarly, if a wealthy person comes from a poor family, support for taxing wealth is lower than when a wealthy person comes from a rich family. These results suggest that respondents consciously or unconsciously assess the deservingness of the wealthy when assessing the legitimacy of taxation. Wealth and the wealthy are therefore not viewed as a homogeneous category. When wealth is acquired in a meritocratic way, support for taxation is lower, whereas non-meritocratic wealth acquisition is associated with a higher approval of its taxation (cf. also Alesina and Angeletos, Reference Alesina and Angeletos2005; Limberg, Reference Limberg2020; Ragusa, Reference Ragusa2014).
The importance of meritocratic belief is also reflected in the way in which deservingness valuations interact with indicators of objective and subjective self-interest. Interestingly, especially respondents with low incomes and those who see themselves as potentially benefiting from a more egalitarian income distribution show low levels of wealth tax support when the wealthy come from a poor family and when wealth was accumulated through work. This indicates that meritocratic beliefs, which are commonly found to be particularly salient for high-status groups (Sachweh, Reference Sachweh2017), may off-set the impact of self-interest in low-status groups. In middle- and high-status groups, by contrast, we find that some non-meritocratic modes of wealth acquisition – such as inheritance (in the middle class) or marriage (in high income groups) – may also decrease tax preferences. Since inheritance or marriage might to some extent reflect typical pathways to wealth in middle- or high-status groups, meritocratic beliefs appear to be off-set by their objective and subjective self-interest.
Along these lines, a promising avenue for future research would be to explore in greater detail different aspects of the behavioural deservingness dimension. While our study has focused primarily on the way in which wealth was accumulated, Ragusa (Reference Ragusa2014) suggests that the ways in which wealth is used – e.g., for one’s own good or the common good – also affect tax preferences. Hence, it would be interesting to investigate whether philanthropic actions, donations or charity can serve as a legitimation of socio-economic advantage that keeps demands for greater taxation of wealth in their place (McGoey, Reference McGoey2015). Similarly, it would be interesting to study how the rich have been portrayed in popular discourse (Waitkus and Wallaschek, Reference Waitkus and Wallaschek2022) and how this contributes to particular images of the ‘deserving’ or ‘undeserving’ rich which shape popular debates about inequality, redistribution and taxation.
Finally, our study is not without limitations. It refers to a single country – Germany – in which issues of deservingness surrounding wealth are rarely publicly debated. Despite occasional media reports, the wealthy remain rather secluded and are not publicly visible as a group.Footnote 13 This may differ from Anglo-Saxon liberal nations, in which wealth is more readily interpreted as a sign of economic success. Thus, in order to determine to what extent our results are context-specific or generalizable across nations, replications of our vignette design in other countries would be highly welcome.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S004727942200099X
Competing interests
The author(s) declare none.
Open access
