The deserving or undeserving rich? New survey evidence on multimillionaire households in Europe

This article addresses two questions. First, when do people consider the rich to be deserving? A literature survey reveals that it is first and foremost the origin of great wealth that determine public attitudes towards the rich. Rich people “deserve” to be rich if their wealth is perceived as having resulted from competence and hard work rather than from inheritance. Second, who are the rich? Drawing on data from the second wave of the European Household Finance and Consumption Survey (HFCS), it is found that multimillionaire households benefit disproportionately from wealth transfers. Large gifts and bequests alone, however, are not good predictors of rich household status. It is rather the highly educated top heir running a (family) business that best represents the rich in Europe. Such entrepreneurs who benefit from earned and unearned financial resources neither fully correspond to nor contradict the existing public beliefs in the “deservingness” of the rich. It is argued that, while still underresearched, it is the “hybrid rich” that dominate in the twenty-first-century capitalism that is marked by historically high levels of wealth inequality alongside inefficiencies in tax systems.


Introduction
Recent years have seen an increasing focus of scholarship on "the rich," a term that in some studies refers only to the wealthiest individuals in the world and in others may include millions of affluent people (Medeiros and de Souza 2015). 1 The surge in interest was sparked by newly available tax data that deliver reliable information on the rising shares of top income earners and top wealth holders around the world (see Izmirlioglu 2016). The consistent finding across all studies is that the rich are growing richer. Evidence for the U.S. suggests that the top 1% wealth share increased from 25 to 30% in the 1980s to about 40% in 2016 (Zucman 2019). For Europe one can similarly observe a rise in wealth inequality over about the past four decades. The top 10% wealth share, for example, has risen to about 70% in Britain and to about 60-65% in France (Piketty 2014, pp. 340-344). This massive concentration of resources in the hands of the few makes the rich a worthwhile subject of study.
While tax data are the single best variable source for capturing the very top of the wealth distribution, they also have a serious limitation: if not linked to other sources, tax returns contain little information on the processes that lead to top wealth. Thus, only rarely is new light shed on an age-old question of sociology: Who are the rich? (Sorokin 1925;for exceptions, see Keister 2014;Wolff 2000).
The current study contributes to narrowing this research gap by exploring characteristics of rich households. Its main aim is to probe the relative importance of inheritance in the accumulation of great wealth. The study design is motivated by the observation that people judge the rich to be deserving if their fortune can be attributed to dispositional characteristics such as "hard work" rather than to the status of a person's parents (e.g., Rowlingson and Connor 2011;Zitelmann 2021). To empirically evaluate to what extent the "deserving rich" or "undeserving rich" identified in attitudinal research (e.g., Skilling and McLay 2015) match the "real rich," i.e., the rich in the real world, this article leverages surveys on European households that oversample the wealthy and allow us to investigate what differentiates the rich from the non-rich. The intention of the study is thus to connect two bodies of literature-the (mostly) psychological literature on the perceptions of the rich and the sociological literature on the characteristics of the rich-that were previously seen as existing in discrete compartments.
The article proceeds in two steps. First, it reviews what little literature exists on the deservingness of the rich and on the dominant characteristics of the rich identified in empirical research. Second, it explores survey information on rich households in Europe to examine in particular the role of "unearned wealth," i.e., the acquisition of financial resources via family gifts and bequests. In general, the main finding is that rich households fit neither the picture of the "industrious rich" or "working rich" who owe their circumstances to meritocratic work nor that of the leisured "coupon clippers" or "rentiers" who barely work. Instead, survey evidence shows the rich to constitute a complex in-between category of capitalists that tend to build their wealth on inheritances but are as well committed to work (see also Hansen 2014). The main conclusion drawn is that in the twenty-first-century capitalism, top capitalists and top inheritors tend to be the same people.

How do people make sense of the rich?
The notion of "deservingness" was initially developed in the literature on public support for social benefits and services that researches the criteria for a fair distribution of social welfare funds among various groups of poor citizens (Oorschot 2000).
Only recently was it extended to the problem of "the rich" to systematize on which grounds the very affluent are perceived as deserving (Prasad et al. 2009;Rowlingson and Connor 2011). As the literature on attitudes towards different status groups focuses almost exclusively on those close to the lowest level of the socioeconomic spectrum (Horwitz and Dovidio 2017, p. 2), the deservingness of the rich is so far poorly understood. The existing findings suggest that the overall deservingness depends on how the rich are perceived according to the following three deservingness criteria:

Shared identities
This criterion has to do with feelings of similarity and cultural closeness. The rich that flaunt their wealth by, for example, buying sports cars or extravagant mansions are perceived by the majority of people as belonging to an "out-group," while the rich that exhibit (consumption) behaviors mostly attributed to the middle class are rather considered an "in-group" with whom one shares a certain lifestyle and whose wealth is, therefore, more likely to be approved of. That shared (moral) values matter for deservingness opinions is supported by studies that find middle-class individuals to perceive great wealth mostly in a positive light as long as spending is not wasteful or lavish but covers identifiable needs (Sachweh 2012), and by the finding that working-class votes for rich candidates in elections depend on whether the candidate is perceived as being "one of us" or not (Prasad et al. 2009).

Reciprocity
This criterion relates to contributions to general welfare. The greater the perceived contribution of the rich to society (in the past, the present, or the future), the more deserving they are perceived to be. Thus, the rich that are ascribed the role of job creators and contributors to equitable growth and those that spend their money charitably are overwhelmingly judged to be "deserving" (Black and Davidai 2020;McCall 2013).

Causes of wealth
By far the most robust empirical finding in the pertinent literature is that the subjectively perceived source of wealth impacts attitudes towards the rich the most, i.e., whether great wealth is attributed to internal factors (e.g., competence) or to external factors (e.g., family wealth) is of key importance for the rich being judged as either "deserving" or "undeserving" (Davis et al. 2020;Sussman et al. 2014). A large representative European survey conducted by the Allensbach Institute for Public Opinion Research with at least 1000 respondents in Germany, Great Britain, France, Italy, Spain, and Sweden finds that entrepreneurs-prototypes of the hardworking and competent rich-are deemed to most deserve their wealth, while heirs (which stands for effortless accrued wealth) receive the least approval (Zitelmann 2021). 2 The general insight that one can take from this study is that the rich are perceived as deserving in the case of internal attributions, while external attributions of great wealth tend to lead people to consider wealth as being unfairly acquired. Other studies find wealthy entrepreneurs to even be admired (Wu et al. 2018) and rich heirs disdained (Sussman et al. 2014).
Only the last criterium has been thoroughly researched. A comprehensive literature review suggests that internal attributions prevail over external ones (see Table 1). For example, children between 6 and 14 years are reported as becoming able with age to explain richness in terms of having a good job (Sigelman 2012). Tests intended to reveal the subconscious association between mental representations show the rich to be perceived as having high levels of competence and low levels of "warmth," which includes such traits as friendliness, helpfulness, or sincerity (Liu et al. 2017). This combination of high competence and low warmth is universally perceived as being characteristic of the rich (Durante et al. 2017). Not all studies, however, consistently report internal attributions to outweigh external ones. A representative study for Germany finds richness to be attributed predominantly to favorable starting positions rather than to meritorious qualities (Götte 2015), and teenagers in New Zealand rated family background most important in explanations of richness (Stacey and Singer 1985). In general, the use of internal vs. external attributions for explaining richness appears to be ubiquitous.
The fact that individuals who endorse internal explanations for fortunes tend to see the rich in a positive light, while advantages arising from family wealth are much less socially accepted is explained in the literature by the dominance of meritocracy as an ideology that favors earned over unearned advantage (Kluegel and Smith 1986). Meritocracy conveys a number of interrelated ideas that are found to convince members of a society that their social position is deserved and to help maintain their confidence that they live in a "just world" (Dalbert and Donat 2015): people should be rewarded in terms of status-financial or otherwise-according to some deserving attributes. Rewards may be unequal, but that is to be accepted as long as individuals get what they deserve. Finally, the individual is seen as being separable from his or her environment, particularly the family. "He is sui generis, complete within himself" (Brown 2006, p. 168). The meritocratic core belief is that those with the requisite training, experience, talent, and motivation should succeed, while those who fall behind can only blame themselves. There is solid evidence that people in the Western world tend to explain economic success increasingly in terms of meritocratic factors rather than through structural ones (Mijs 2018

Who are the rich really?
Historically, sociological work on the rich has been meager and methodologically troubled by the fact that while some rich people agree-for different reasons (Gilding 2010)-to be interviewed in-depth by social scientists, surveying representative samples of wealthy individuals has been found to be extremely difficult. Given that cooperation and response rate decrease significantly as one moves up to the highest wealth strata (Kennickell and Woodburn 1999), only a single survey has so far allowed us to reliably probe characteristics of the rich. The Survey of Consumer Finances (SCF) is a triennial survey of U.S. households that is collected by the Boards of Governors of the Federal Reserve System, i.e., the central banking system of the United States, and makes systematic use of fiscal information on individual taxable (capital) income to improve the probability of sampling wealthy respondents (Bricker et al. 2016). The applied "oversampling" strategy has proved to lead to a credible (but not perfect) representation of top households (Saez and Zucman 2016). As research by Vermeulen (2018) shows, some SCF sample observations have greater wealth than the poorest Forbes billionaire (even though listed Forbes multimillionaires are excluded from the sample design) and about 15% of the SCF sample holds wealth of over $2 million. Other surveys on the rich, such as the Great British Class Survey (Savage et al. 2013;GBCS) or The Very Wealthy in Germany (Lauterbach et al. 2011;HViD), are informative as well, but they are significantly less efficient in covering the top of the wealth distribution and do not match the methodological sophistication of the SCF, with its complex survey encompassing not only stratified sampling but also item non-response, i.e., participants who refuse or are unable to respond to certain questions. 3 The following literature review is, therefore, mostly confined to SCFbased studies.

Entrepreneurship
Rich lists such as the popular Forbes ranking have become dominated by entrepreneurs (Kaplan and Rauh 2013), and one of the most notable changes in the SCF with regard to the top 1% ranked by household wealth was a huge increase in the share of the self-employed between 1983 and 1992, which almost doubled from 38 to 69%, and stands in stark contrast to the modest gain among all workers, from 15.4 to 17.2% (Wolff 2017, p. 451). Similarly, Keister (2014) makes the observation that members of the top 1% are much more likely than the typical American to be self-employed, which is also confirmed by Ströing et al. (2016) for Germany. While self-employment does not necessarily imply entrepreneurship, it is well established that the self-employed with many employees are clearly overrepresented in the prosperous group (Lindh and Ohlsson 1998).
The reasons why we see entrepreneurs overrepresented among the rich may vary. Some find that entrepreneurs tend not only to earn more but also to save more of their income than workers, which results in higher asset holdings (Meh 2005). Others observe that especially those born into business families tend to choose entrepreneurial occupations (Quadrini 1999). Entrepreneurs with and without inherited advantage tend to base their self-identities on hard work (Kantola and Kuusela 2019) and attribute their prosperity to their innovative ideas and the willingness to work long hours (Schervish 2016).

Education
Empirically, Wolff does not find clear evidence that more education paid off in terms of entry into the ranks of the top 1% of wealth holders over the period 1983 to 1992. In general, however, the very rich are clearly a well-educated group: in 1983, 76% were at least college graduates, compared to 21% of all households (Wolff 1998, p. 145). In a similar vein, an exploration of the National Longitudinal Survey of Youth (NLSY) shows that nearly 40% of millionaires had advanced degrees and 30% had at least a college degree (Keister 2005, p. 73). The existing evidence for Germany also reveals that millionaires are more highly educated than the average population (Schröder et al. 2020a).
The bulk of the literature remains silent on whether the relationship between a high level of education and great wealth is causal or not (Keister and Lee 2014, p. 20). A recent study investigates what drives differences in wealth across education groups over the life cycle on the basis of a rich longitudinal dataset covering the entire Swedish population and allowing, inter alia, controls for parental characteristics. Girshina (2019) finds that education has a positive, profound, and long-lasting effect on net worth.

Inheritance
There is solid evidence based on population registers that heirs with the highest income and wealth receive by far the largest bequests (Bastani and Waldenström 2021). Even if surveyed households tend to understate the value of inherited wealth (to give the impression that they "earned" their wealth), Wolff and Gittleman (2014) find that wealth transfers increase monotonically, with a big jump for households holding at least $1 million in assets. Keister and Lee (2014) find that 42.5 percent of those in the top 1% of the wealth distribution and 44% of those in the next 9% inherited at least some amount. Finally, evidence from the German HViD survey suggests that receiving gifts and bequests is a key factor for about two-thirds of the very wealthy (Ströing et al. 2016).
If put together, the cited studies allow for a cautious portrait of the rich in the United States. Generalizations of the few insights gained on other societies, however, appear difficult, as, among other things, overall inequalities in income and wealth differ hugely between the United States and Europe (Piketty 2014). The cited work can nevertheless provide guidance as to relevant features that research should consider. I therefore place entrepreneurship, education, and especially inheritance at the center of the subsequent analysis. While comparative accounts of the rich in Europe do not exist in the literature, cross-country research on the determinants of household wealth has either detected no systematic association between household net worth and country-level characteristics (Semyonov and Lewin-Epstein 2013) or has identified varying substitution effects between welfare state provisions and private wealth that are pronounced in the middle and at the lower end of the wealth distribution but not at the upper end (Fessler and Schürz 2018). I therefore start from the assumption that great wealth is similarly associated with entrepreneurship, education, and inheritance across European countries.

Data
The study uses data from the second wave of the European Household Finance and Consumption Survey (HFCS), which was coordinated by the European Central Bank (ECB). All data were gathered over the years 2014 and 2015. The HFCS collects representative household-level data on households' finances and consumption.
Oversampling strategies for wealthy households to remedy potential non-observation bias vary widely across all 18 participating euro area countries. Cyprus based the oversampling on household (HH) information from electricity bills, for example, while Germany used geographical income information, and Greece real estate information. Considered here are only six countries with the highest "effective oversampling rate" (for the top 10% households), which demonstrates the degree to which the share of wealthy households in the sample exceeds their share in the population (see Table 2). All other countries are neglected because of comparatively poor coverage or even non-coverage of the rich. Clearly, the national banks of Spain and France were most effective in targeting wealthy households. For instance, the category above €2 million of household wealth already in the first HFCS wave contained 544 sample observations for Spain, representing 139,539 households (Vermeulen 2018, p. 9). While there remains, of course, a very large gap between the richest household in the HFCS and the poorest person on national rich rankings (Bach et al. 2019), the HFCS nevertheless offers the first unique possibility to gain insights into the rich in Europe.
In the social sciences, there is no unanimity in the definition of the rich, and social scientists have used a number of different practical approaches to measuring richness (Medeiros 2006). Most authors have so far taken a positional approach by equating the rich with the top 1%, top 5%, or top 10% of the wealth distribution. Only very recent research has started to investigate whether people agree on a "rich line," defined as a threshold at which material resources become considered excessive. Using vignettes standing for families with different levels of wealth that make it possible to measure the level of agreement with a series of statements, Robeyns   2021) found that practically no one considers wealth levels below €1 million rich, while for a majority the "rich line" has been crossed if households own over €2 million. To identify the rich, I therefore use as a lower limit household net wealth of €2 million. I further top-coded the variable net wealth at €100 million to exclude seven outliers (see Table 2). This setting of the upper and lower limit allows the focus to be entirely on households that are commonly perceived by respondents as "rich" (e.g., "millionaires next door") and not "super-rich." 4 As shown in Fig. 1, rich households are concentrated in each country at levels of net household wealth between €2 and €5 million. The dispersion in wealth among the rich is the largest for Spain and France. To make sure that results are robust to different definitions of the rich, analyses were also conducted for the "top 10%" in the wealth distribution (see Online Appendix). Another methodological issue involves the handling of missing values through item non-response. The ECB suggests a multiple imputation approach and makes five implicates available that are constructed following the guidelines of Rubin (1987). I use all five implicates as well as survey weights in all descriptive statistics. Within the regression framework I estimated two models with five imputations, one with unweighted data and one using survey weights. As the parameters Fig. 1 Distribution of the rich. Reduced dataset (N = 27,737 households); calculations over all five implicates. Each dot represents a household were substantively similar, I report regression results based on unweighted data only (Winship and Radbill 1994).

Dependent variable
The explanandum in this study is membership in the group of rich households, i.e., households with net worth over €2 million and below €100 million. It is important to note that as the HFCS uses the household as the unit of analysis, information on financial resources refers not to individuals, but to households. Net wealth is total household assets less total liabilities. Assets include real assets 5 (e.g., household main residence, other residential real estate, vehicles, shares in self-employed businesses) and financial assets (e.g., deposits, shares, bonds, mutual funds); liabilities include mortgages, loans, and other uncollateralized debt (e.g., credit card debt). Entitlements to public and occupational plans and social security funds are excluded from the HFCS wealth concept. I follow Arrondel et al. (2014) in distinguishing safe (deposits, life insurance) from risky financial assets (investments in mutual funds, bonds, and publicly traded shares). The business assets category includes publicly traded and non-traded businesses as well as businesses with and without self-employment.

Independent variables
Entrepreneurs are households that own "business in which at least one member of the household works as self-employed or has an active role in running the business" (ECB 2013, p. 22).
Inheritances and (inter vivos) gifts are captured by two survey questions. The first asks whether the household inherited the household main residence or received it as a gift. The second asks whether any member of the household received other substantial wealth transfers, including money, real estate, or any other valuable asset as a lifetime gift or a bequest. For up to three past inheritances or gifts, households are also asked to report the values for each transfer separately. Particularly large wealth transfers are identified by distinguishing top 10% gifts and bequests from the rest (for a similiar approach, see Schneebaum et al. 2018).
Education captures the highest level of completed education of the household's reference person. 6 The three levels distinguished based on the International Standard Classification of Education (ISCED) are: primary (ISCED 0 and ISCED 1), secondary (ISCED 2-4), and tertiary (ISCED 5 and ISCED 6). 5 Households that own "real estate for business activity" are all households that own property which is not privately used, rented or vacant but used, for example, for production purposes. 6 The reference person in the household is defined in accordance with the "Canberra definition," i.e., applying the following rule in the order given until one person is identified: "One of the partners in a registered or de facto marriage, with dependent children; one of the partners in a registered or de facto marriage, without dependent children; a lone parent with dependent children; the person with the highest income; the eldest person" (UNECE 2011, pp. 65-66).

Control variables
The control variables age, gender, marital status, and household structure are all known to correlate heavily with household wealth (Bover 2010;Killewald et al. 2017). The composition of households is considered by differentiating between single households, couple households with or without children, and all other households.

Analytical strategy
First, I descriptively analyze the wealth composition of rich and non-rich households in order to identify the unique financial traits of each group. Second, I investigate the systematic relationship between membership in the group of rich households and several key variables using logit regression models. Logit regression makes it possible to measure the effects of independent variables on a dichotomized variable (i.e., non-rich vs. rich) and provides, inter alia, tests of significance of the relations. Regarding probabilities, logit models describe non-linear effects that cannot be expressed in a single coefficient (Best and Wolf 2015). Therefore, regression results are reported using average marginal effects (AME) that are insensitive to non-linearity and provide as well readily interpretable effect estimates (Breen et al. 2018). AMEs (estimated by STATA's margins post-estimation command) provide a single summary measure of how a one-unit change in an independent variable changes the probability of belonging to the group of rich households, averaged across the whole sample.

Descriptive results
The first descriptive impression supports our expectation: the rich are clearly a distinctive group. Among other things, the findings suggest a strong relationship of inheritance, entrepreneurship, and education with top wealth status. Table 3 demonstrates that sociodemographic characteristics also vary between the rich and nonrich. While heads of rich households are especially likely to be married and male, single households are more common in the group of the non-rich. Further, the rich are clearly better educated and older by average age.
To explore investments, Fig. 2 shows the composition of gross assets. It is notable that the household main residence (HMR) accounts for more than half of the portfolio of the non-rich. In contrast, businesses account for the largest proportion of the portfolio in the case of the rich. Other real estate is the second largest share in the portfolio of both, the non-rich and the rich households.
To probe whether entrepreneurs figure high within the group of rich households warrants a more fine-grained analysis of asset holdings. Figure A1 in the Online Appendix considers two specific types of real estate and business assets: ownership of self-employed businesses and of real estate for business activity. The results are clear-cut: over 60% of all rich households report self-owned business. Thus, unlike investors, the majority of rich households work for companies that they own. Slightly more than 10% of these rich entrepreneurial households also own real estate for business purposes. Both types of ownership can only very rarely be observed in the case of the non-rich.
Marked differences between the rich and non-rich emerge also with regard to the number of hours normally worked per week in the main job. In all countries except France, reference persons in rich households report working more than forty hours. In the case of Belgium, the number even amounts to sixty hours. For non-rich households, in contrast, the average number of hours comes close to the standard full-time working week. The rich in the HFCS are, thus, not only predominantly entrepreneurs, they also work long hours. Figure 3 further shows that there is a clear gradient in the chances of receiving gifts and bequests, with 65% of all rich households and only 35% of all nonrich households having received at least one wealth transfer in the past. Interestingly, it is not the household residence that "trickles down" from one generation to the other but substantially more often family wealth transfers that include money, dwellings, land, stocks, or business assets. This descriptive evidence suggests that the rich may be a group not only whose merit is highly rewarded but also whose members have a privileged family background in common.  While important, the identified correlates of rich households tell us little about which factor predicts great household wealth the most. I therefore apply a regression framework to better assess the determinants of richness in a multivariate setting.

Multivariate results
In Table 4, logit regression is used to assess separately the importance of different key variables for rich household status (Model 1 to 3). All regression models control for the same sets of individual-and household-level variables and include country fixed effects. I find all three key variables to exhibit a quite similar positive and significant association with rich household status. The marginal effect of entrepreneurship is the largest, at 13 percentage points (p.p.), followed by top gifts and bequests at 12 p.p., and tertiary education at 10 p.p. One of the main results is, thus, that having received a large wealth transfer in the past increases the probability of being rich by 12 p.p..
Model 4 simultaneously includes all three key variables. The marginal effect of entrepreneurship decreases to 11 p.p while that of top wealth transfer and education drops to 9 p.p and 7 p.p., respectively, which suggests that, for example, a small part of the education effect is due to a correlation with the two other key predictors. The identified differences in the predictive power of the three key variables are quite small. Tertiary education matters the least, which suggests that Fig. 2 Shares of assets relative to total household gross wealth (in %). Reduced dataset (N = 27,737 households); weighted data, calculations over all five implicates the formally best educated households are not necessarily the most knowledgeable of economic opportunities. The overall insight gained is that a single household characteristic is not sufficient to predict rich status. The determinant process that drives the accumulation of great wealth appears rather to be a complex one that is simultaneously impacted by different factors. I therefore continue by probing the interactive effect of all key predictors.
The robustness check conducted to predict households that belong to the top 10% of the wealth distribution reveals partly different results especially because marginal effects for top inheritance are larger than for entrepreneurship. Thus, the model used in Table 4 is sensitive to the definition of the rich. The effects of the predictors reported in Table 5 can only be interpreted in relation to the reference category, which encompasses lower-educated, non-inheriting and non-entrepreneurial households. Once again, we can see that entrepreneurship and top gifts/bequests make richness similarly more probable by between 9 and 13 p.p. The joint effect is, however, clearly greater: top wealth transfers combined with entrepreneurship are associated with an increase of 34 p.p. The probability of more highly educated, entrepreneurial, and inheriting households belonging to the rich is even an astonishing 49 p.p. higher if compared to the reference group. Thus, the main takeaway is that a combination of household characteristics best predicts whether a household is likely to surpass the richness line or not; entrepreneurship alone or top inheritance alone turn out to be comparatively poor predictors.
It is important to note that the results for interacting determinants stay similar if the analysis is conducted for the top 10% in the wealth distribution. What differs is only the size of marginal effects (see Appendix).
Country-specific analyses (see Fig. 4) reveal quite similar distinctive patterns of predictors to those identified in the pooled dataset. What one can observe is partly a difference in the magnitude of marginal effects. For Spain, entrepreneurs with top inheritance and tertiary education have a probability of belonging to the top wealth group of about 60 p.p, while for Belgium the same effect is significant but much lower. Still, entrepreneurship combined with top inheritance is associated with high probability in the case of Belgium, which suggests that education can mediate differently the relation of the other main predictors with richness. Cross-country differences, however, should be interpreted with great caution, as the country comparison of the pre-harmonized HFCS data is impaired due to remaining differences in the methodology (Tiefensee and Grabka 2016) such as the varying sampling frame used to oversample the wealthy (see Fig. 2 and Table 1). Figure 4 shows, however, consistently for every country studied that the probability of becoming rich is highly dependent on the interplay of entrepreneurship and inheritance, which confirms our central findings from pooled data.

Conclusions
Previous research has shown that we tend to instantaneously construct plausible stories to explain why someone has become rich. A pervasive habit of mind is to attribute richness either predominantly to one's own efforts or to external causes. While more research is certainly needed to firmly establish which of the two explanatory modes prevails in different societies, the current state of knowledge is that internal attribution of richness occurs more often, i.e., the rich are Fig. 4 Interacting determinants of richness. Results from country-specific logit regressions (average marginal effects). Reduced dataset; multiply imputed data; identical controls as in Table 5. Whiskers indicate 95% confidence intervals held to be deserving because they have "earned it with their own hands" and "through their own efforts." In theory, the real rich might well match the imagined rich, i.e., the accumulation of great wealth might be driven first and foremost by "earned income." This article aimed at unpacking the black box of richness by probing empirically the effects of inheritance, entrepreneurship, and tertiary education on rich household status in Europe. Achieving this goal has become possible with the ECB launching the HFCS survey, which oversamples the wealthy in different countries and thus allows insights into how the rich differ from the rest of the population in Europe.
This study finds the rich to be hybrids that elude dichotomous causal thinking. For, one the one hand, multimillionaires in six different European countries were found to be likely to have received large family transfers ("unearned wealth"), and, on the other, they derive most of their wealth from self-employed businesses ("earned wealth). What is more, tertiary education was found to mediate the relationship between entrepreneurship and top wealth accumulation. It is the combination of these three key factors that predicts best whether a household will join the group of the rich.
One can best theorize the rich analyzed as capitalists in the classical sense: owners of (family) business who employ wage earners to produce goods and services for sale, on the condition that they produce enough to not only cover their wages (and other costs) but also to provide the owner with profit. As owner-entrepreneurs they fit the image of "progressive" entrepreneurship as a field of economic opportunityseeking that rewards talents and risk-taking. As those who either enter into existing family enterprises or build businesses partly based on large family transfers, they exhibit "regressive" features associated with "patrimonial capitalism" (Piketty 2014, p. 237), in which dynastic forms of wealth play a preeminent role.
The survey does not reveal how entrepreneurially orientated firms owned and managed by the rich are. The reported working hours suggest only that the rich describe themselves as being guided by a strong work ethic (see also Kantola and Kuusela 2019). It is similarly impossible to know whether the "hybrid rich" represent the "deserving" or the "undeserving rich." Attitudinal research has so far relied too heavily on one-dimensional depictions of the rich ("entrepreneur" vs. "heir"), pre-structured questionnaires, and Likert-like scales that anchor study participants' responses to questions posed. These measurement approaches fail to account for the hybrid nature of the rich identified in this study. Narrative-style or "story" vignettes that accurately portray the real rich, as identified in the article, would do a much better job of capturing attitudes. Such research is especially warranted as we know that internal or external attributions of great wealth do not only systematically impact perceptions of deservingness of the rich but lead as well to low or high levels of support for redistributing policies in general and taxes on the rich specifically (Ragusa 2015;Sainz et al. 2019). That survey respondents may hold differentiated perceptions of "family capitalists" is only indirectly suggested by studies examining attitudes towards inheritance taxation. While respondents often tend to favor progressive taxation of large inheritances, they are overwhelmingly in favor of tax relief for the inheritance of company assets (Groß and Lang 2018). Broad sections of the population may thus hold a positive image of the rich portrayed in this article.
It is empirically observed that while wealth inequality has risen in recent decades, advanced democracies have not turned to new or higher taxes on the rich (Emmenegger and Lierse 2022). It is therefore crucial to know whether and why there is public opposition to taxes on top wealth owners (Rowlingson et al. 2021). Most recent research has shown that providing facts about the role of inheritances in today's society changes attitudes towards inheritance taxation (Bastani and Waldenström 2021). An interesting topic for future research would therefore be how the propagation of facts on the "hybrid rich" alters opinions on the deservingness of those whose wealth is perceived to be excessive. Such studies might provide fuller insights into how malleable attitudes towards the rich are and how (changing) internal and external attributions of great wealth relate to tax preferences.

Limitations
Despite its contributions, this article, like most analyses, builds on data that have clear limitations. As the study uses cross-sectional data, all the analyses can only establish an association, and not a causal relationship, between the key drivers of the wealth identified and top household status. What it does answer, therefore, is the question of who the rich are, rather than how they came to be that way. An important conundrum that remains to be solved is especially whether entrepreneurs accumulate more wealth, or whether people with wealthy family backgrounds are more likely to become entrepreneurs.
Further, the study disregarded the "super-rich," such as the German manager Stefan Quandt, who owns almost half of BMW shares, as the survey data analyses only capture "the rich" and not the "super-rich" in a representative way. It is only very recently that progress in narrowing the data gap for net assets located above three million up to a quarter of a billion has been made in survey research (Schröder et al. 2020a, b). Once these data become available, the research design proposed in this article could be extended to the investigation of the "super-rich." Another limitation is missing information on what was inherited or received as a gift. Given the wealth portfolio composition of the rich investigated, it may be that large family wealth transfers go hand in hand with the transition of businesses from one generation to the next. It is only with complete information on gifts and bequests that one can distinguish unambiguously between entrepreneurs that benefited, for example, from large money gifts and those who became (at least part-) owners of family businesses. It is, however, well known that respondents are particularly likely to refuse to give detailed answers to survey questions on the origin of their inherited wealth. Thus, unfortunately, the quality of survey information on family wealth transfers is unlikely to improve in the future.
Funding Open access funding provided by University of Graz.

Data availability
The HFCS dataset analyzed in this article is available for research purposes. You can