The racialization of welfare support as means to further welfare state cutbacks – spillover effects in survey populations and media reports in Austria

ABSTRACT The spillover of racism towards welfare opinions has previously been studied in several countries. This paper analyses how racism and welfare are formed intertwined in Austria. This remains an important topic, as the welfare state not only distributes material resources but is also seen as a site of socialization – at least within the German-speaking literature. Besides identifying a potential spillover effect within survey data for Austria, the article also discusses how the topics of the racialized, especially in the form of the migrant “other” and welfare state and support, are rhetorically linked within newspaper articles. Four rhetorical strategies are distinguished within the empirical material using the notion of the undeserving “other” to argue, among others, for general welfare state cutbacks. These strategies also introduce arguments for extreme cases of social exclusion within the welfare state that affect existing views on, and possibly also, society and social life itself.


Introduction
In 2019, the coalition between the Austrian People's Party (ÖVP) and the Freedom Party of Austria (FPÖ) passed a new family allowance bill, which introduced a mechanism that diminished support per additional child in the family. In a national council debate, the then-acting minister of social affairs justified this by arguing: "We no longer accept that families with ten children or moremost of the time families coming from Arab or African regionsreceive social benefits the common worker can never reach". 1 This was applauded by members of both coalition parties and very much perpetuated the themes that had dominated the preceding election campaign: migration and its cost for the Austrian welfare state.
Although this instance exemplifies "welfare chauvinism" (Hjorth 2016), it even more so demonstrates how the topics of welfare and the racialized "other" are tightly entangled in policy debates. By playing off notions of the "hard-working man" against families from "Arab and African regions" and by promising that the restructuring will mainly hurt these families, the Austrian family allowance model was fundamentally reoriented in a way that both ascribed different values to children but also what family form is seen as socially desired. By promising the exclusion of some, the Austrian welfare system was fundamentally changed for all.
This rhetorical connection between the notion of the racialized "other" and welfare policies is addressed in the literature via the spillover of racism towards the opinions and constitution of the welfare state (Brown 2013). Originally stemming from studies on the welfare state in the USA (Gilens 1995), the spillover has become more prominent in recent years in other countries, including European ones (Larsen and Dejgaard 2013;Harell, Soroka, and Ladner 2014), adding to the otherwise more common topic of welfare chauvinism. However, this rhetorical blending of policies against the racialized "other" and welfare policy also resonates with sociological literature that emphasizes the social function of the welfare state as an important instance of socialization (Lessenich 2008). This enhancement of the welfare state's status from solely distributing material resources further emphasizes the relevance of the spillover of racism as a topic in shaping the welfare state but so far has been rather considered separately.
These observations underlie the two research questions that direct the empirical analysis presented and discussed in this article. First, is there a spillover of racism towards opinions about the welfare state among the Austrian population? This is examined using data from a 2017 telephone survey. As a follow-up question, this paper also investigates how racism and opinions towards the welfare state become intertwined, which this research does not take for granted. For the latter question, the article reflects Hall's (2018Hall's ( , 126ff orig. 1982 argument about the media's importance in shaping and forming worldviews, ideas, and ideologies. Therefore, a qualitative content analysis of newspaper articles published before and at the time of the 2017 telephone survey was also conducted.
The following section provides background information about Austria to contextualize why the country is an interesting case study for this paper's research questions. The connection between the welfare state and racism is illustrated by their shared characteristic of demarcating members of societies into groups with different access to resources. This enables embedding the concept, literature, and studies on the spillover of racism into welfare opinions and on welfare chauvinism into the broader literature on the welfare state. The next section discusses the research questions and methods before demonstrating the empirical results. First, correlation and multivariate regression analysis is used to identify the spillover within the survey population, which is followed by the results from the newspaper article content analysis. Four categories were identified before and during the content analysis, which represent the different ways migration or the "other" are connected to the welfare state that promote demands to constrain and claw back social benefits. The following section discusses the interconnections between the quantitative and qualitative data and their complementary nature, while the article closes with remarks on the significance of this analysis, the limits of the study, and its contribution to the literature.
Why Austria as a case study?
As the first country within the European Union to include a "populist radical right party" (Ennser-Jedenastik 2020) in government (2000)(2001)(2002)(2003)(2004)(2005), Austria demonstrated relatively early that racism is not only dominant at the margins of society, but also within the alleged mainstream or centre of society. Upon the dissolution of this coalition government, right-wing parties and positions became increasingly involved at both the national and local levels. This shift took place within a welfare state characterized by Österle and Heitzmann (2019) as fitting into the conservative, male-breadwinner model that is under constant pressure to economise. The regime is dominated by insurance-based social benefits (e.g. unemployment benefits, pensions) where the accumulation of premiums is closely tied to participation in the labour market. The latter is also inscribed in Austrian social assistance law, e.g. the current iteration explicitly states that it should guarantee the "optimal functionality of the labour market", 2 and omnipresent in public and policy debates. These public debates on the welfare state and social policy take place in a media landscape 3 dominated by the tabloid pressand especially by one particular newspaper, the Kronen Zeitung, with an explicit racist anti-migrant agenda (Schadauer and Schäfer 2019).
Regarding migration, labour market access is markedly restricted for non-EU citizens, aside from professions or fields with a high demand for "foreign" or seasonal workers (e.g. in tourism or agriculture). In addition, the few available correspondence tests also show a strong tendency towards labour market discrimination due to skin colour, ascribed foreign origins, and wearing headscarves regardless of citizenship and education (e.g. Weichselbaumer 2020). Hence, Austria shares many characteristics with several other countries, especially within the European Union (e.g. conservative welfare regimes like Germany and France, the rise of right-wing parties, the dominance of the tabloid press). Likewise, Austria has been at the forefront of normalizing policies within the EU based on racialized ideologies.

Why the welfare state?
Why address the welfare state at all when it comes to the influence of racism? It is important to acknowledge that the following argument may be strongly shaped by a Eurocentric or, rather, an Austrian/German perspective. Although the prominent role of the welfare state is politically contested and while its genesis, meaning, and importance are widely disputed within the research literature, its influential role in shaping contemporary Austrian (and German) society remains relatively unquestioned. Lessenich (2008) sees the welfare state as providing a stabilizing function against changing socioeconomic conditions that have eroded previous social safety nets; it also provides labourready workers and consumers. Additionally, the Austrian/German welfare state is historically connected to industrialization (and colonialism) and hence embedded in given power structures and struggles. It is seen as a means to push through the interests of "capital", but also "wage labour" or the "labour movement", and as a means by the state to claim and solidify power by centralizing, professionalizing, and bureaucratizing social processes. Beyond these historical and structural features, the welfare state is also viewed as a product of and an influence on ideas about freedom, equality, and security, which shape the role, meaning, and importance of social institutions like the labour market, families, and the maintenance of social hierarchies.
Expanding on the effect of the welfare state on society, Kaufmann (2002, 30ff) describes the welfare statespecifically its presence or absence and its structureas an important instance of socialization. It shapes understandings of social interactions, society at large, how people live together, and what they expect of each other and themselves (e.g. what they can/want to do). The welfare state structure, as well as the arguments for and against it, can be considered an important social mechanism that influences how interactions are formed and how actions are entangled (Lessenich 2008, 485). This concerns both the local and global level when considering the welfare state's importance in reifying global social inequality (Weiss 2005).
In terms of the conservative welfare regime that subscribes to a malebreadwinner model, Lessenich (2012) argues that on a structural and cognitive level, there are two lines inscribed in and enacted by the welfare state that demarcate its population: residents who are active in the labour market or not on the one hand, and citizens and non-citizens on the other hand. The rights and prospects for social participation and the opportunities to live or even survive are unequally distributed along these lines of demarcation (2012, 104ff). The first line is legitimized through a glorification of "hard work" and "occupational ambition". By contrast, demarcating between citizens and non-citizens in the second line is not limited to administrative or juridical distinctions, but also draws on another kind of justification for unequal treatment: a form of differentiation based on ideas of a given natural order based on notions of "origin"racism.

What does racism have to do with it?
As an effective demarcation line, racism is seen as one of many practices of exclusion (Hall 2016(Hall , 173 orig. 1989. It leads to the racialization of social conditions and relations, like social inequality or welfare support. This paper draws on Miles' work (2003Miles' work ( orig. 1993) which does not conceptualize group characteristics as per se given, relevant, and influential, but that they are made significant in the process of producing and using the categories to make a difference. In the case of "race", this takes place during the process of racialization, which generates groups who are constructed as homogenized, essentialized, polarized, and hierarchized. While Miles mainly refers to physiological characteristics, many authors including Guillaumin, Balibar, Hall, and Rommelspacher expand upon Miles' concept to include cultural, religious, symbolic, social, and imaginary/phantasmatic characteristics as made relevant through the practice of racialization (see, e.g. Kerner 2009;Terkessidis 2018). When considering these additional features, racism's distinction from other forms of discrimination is not based on characteristics attributed to individuals or groups, but on the mechanisms that enact relations of disadvantage and privilege. One central mechanism is the permanently given threat or practices of excluding the racialized "other" from society, nation states, or life itself for the sake of the "own" group or society's wealth and survival. This includes deportationespecially into uncertain and dangerous environmentswithholding help and support (e.g. by the welfare state), symbolic hostility (e.g. racist campaign posters, destroying memorials), as well as physical violations and assassinations (which very much resonated with Foucault's (2003Foucault's ( , orig. 1976) understanding of racism).
Does this understanding of racialization also apply to the demarcation within the welfare state that is based on citizenship, as described above? Here, I follow Terkessidis (2018) argument that racialization arguably becomes relevant through an accompanying "differentiating power", which is combined with racialization and exclusion as an apparatus of racism. For the welfare state, citizenship, and migration, this differentiating power isamong othersat work between nation states in the form of global inequality, unequally distributed resources, and especially in the unequally distributed power to enable, allow, or restrict rights and opportunities for (trans-)national mobility (Weiss 2005). Therefore, exercizing power based on global inequality to secure resources for the "own" group by excluding "others" and, hence, directing how one group's life is given a higher value than the other can be seen as a mode of racialization. This is true even if solely based on citizenship and the nation state, and especially if this type of demarcation is taken for granted. This racialization of the migrant "other" is also frequently accentuated within nation states by expanding the "migrant" designation far beyond matters of citizenship. For example, by using "foreigner" as a pejorative, by designating, e.g. skin colour, religion, habits, or beliefs as markers of non-belonging and for German speaking countries by normalizing the term Migrationshintergrund (migration background) ) to mark certain citizens as "others" regardless of citizenship. Together, these depict the essentialized migrant "other", as demonstrated in several reports 4 and empirical studies (see, e.g. Terkessidis 2004;Weichselbaumer 2020).
This racialization by national demarcations and of the migrant "other" is also partly addressed in the literature about the welfare state, such as by Béland and Lecours (2008), who analyse the close connection between historical and contemporary social policy and nationalism. Their account of nationalism connects "identity, interests, and political mobilization" with notions of territorial belonging, entitlements, and solidarity. Social policy can be and is used as a powerful tool to produce an idea of national unity, which often appears as exclusive social solidarity. However, the authors state that its impact on prevailing levels of social supportespecially regarding sub-state nationalismis not generalizable.
This ambiguous effect that national demarcations have on welfare support is represented somewhat differently in literature explicitly addressing racism and welfare state. Regarding the USA, Ward (2005, 136) argues that the "racialised welfare discourse escalated" from the 1960s onward and that welfare became increasingly "inextricably tied to race", which led to certain welfare programmes being eliminated. Likewise, Faist (1995) observed a change in German welfare state politics dating to the 1990s, which expanded the previously, mainly class-based welfare regime towards segregation by class and "ethnicity".
This link between the racialized "other" and welfare support has also been addressed in survey studies about "deservingness" that look at opinions towards and the acceptance of welfare support (Harell, Soroka, and Iyengar 2016;Grausgruber 2019). Here, deservingness is seen as a heuristic shortcut to quickly decide "who should get what and why" (van Oorschot 2000). It is a shortcut that simplifies how people form opinions on the legitimacy of welfare supporteven with limited information on the conditions and recipientswhere the migrant "other" is often classified as per se undeserving. In a large experimental vignette study on unemployment compensation, Reeskens and van der Meer (2019) found that even when favourable information is provided on the willingness to work and on the cause of unemployment, this does not necessarily compensate for negative attributions towards the migrant "other". Although not explicitly stated by the authors, this hints towards such demarcations likely being based on homogenization, essentialization, polarization, and hierarchization.

Spillover and welfare chauvinism
The influence of this kind of demarcation constitutes the gateway for racialized views of the "other" within opinion towards the welfare state or certain welfare measures. As Gilens points out in an early survey study on the US welfare system, "beliefs concerning blacks' commitment to the work ethic are the most important dimension of racial attitudes with regard to whites' welfare views" (1995, 994f) and are even better predictors for welfare opinions than self-interest, individualism, and egalitarianism. This was later reproduced by several studies (Kootstra 2016;Alesina, Miano, and Stantcheva 2018).
Following their experimental online survey, Harell, Soroka, and Iyengar (2016) also show a spillover effect among participants from the US, UK, and Canada and point towards media coverage as one possible important source behind the racialization of welfare state opinions. Hall (2016Hall ( , 182 orig. 1989 previously raised similar concerns about the media's role in advancing racialized depictions of belonging to substantiate certain policies (e.g. market individualism, law and order). Hall sees the way topics are selected or ignored, how they are framed, and how prominently (or not) they are promoted as being of social relevance. Their treatment may organize how groups and classes act by influencing how they define "reality" and evaluate contradictory social interests.
To evaluate "why Americans hate welfare", Gilens (1999) conducted a comprehensive analysis of American media between 1950 and 1992 and compared its portrayal of poverty and welfare. He observed an overrepresentation of "blacks" in pictures accompanying "unsympathetic poverty stories" and that they "are unlikely to be found in media stories on the most sympathetic subgroups of the poor, just as they are comparatively absent from media coverage of poverty during times of heightened sympathy for the poor" (1999, 132). Clawson and Trice (2000, 63) later reproduced these results and concluded that newspapers "do not capture the reality of poverty; instead, they provide a stereotypical and inaccurate picture of poverty which results in negative beliefs about the poor, antipathy toward blacks, and a lack of support for welfare programs" (2000,63). However, a comparative study by Larsen and Dejgaard (2013) showed that media coverage differs according to the given welfare regime.
These (publicized) associations of welfare with people of colour (or migration in other national contexts) not only serve to promote the restricted access to welfare support for the alleged "other" (Hjorth 2016), but also to delegitimize overall welfare support. Based on his media analysis, Gilens argues that "the influence of racial attitudes extends beyond explicitly racial issues, and that particular beliefs about blacks can have a striking impact on an ostensibly nonracial issue" (1995,1010). Viewing the racialized "other" as lazy, undeserving, impotent, and not belonging to one's "own" society also negatively influences opinions towards the welfare state and social benefit measures if they are publicly framed as being strongly associated with these "worthless others".
While the American welfare state dominates analyses of the spillover of racism, welfare chauvinism is a more prominent topic in Europe. Despite a lack of clear and shared definition, as discussed by Ennser-Jedenastik (2020, 2) in an Austrian context, its ability to combine a more "(relatively) leftist social policy position regarding natives with a (relatively) rightist social policy position regarding nonnatives" is one of welfare chauvinism's important features. It introduces a demarcation between potential recipients of social support through differing notions and measures of "natives" and "others", as applied in the research literature (Keskinen, Norocel, and Jørgensen 2016).
There is little overlap in the existing literature on spillover of racism and welfare chauvinism. Conversely, welfare chauvinism tends to shy away from explicitly using racism as explanatory element. In addition, it does not necessarily, conceptually nor empirically lead to a general reduction of welfare support, but rather to a fragmentation of the system. In this regard, using longitudinal survey data Arndt and Thomsen (2019) argue that respondent attitudes towards welfare programmes did not change uniformly over time, but instead appeals for more or less support differed depending on which welfare programme was considered. However, this difference is strongly influenced by how the imagined typical recipients are framed within policy discourses as either "natives" or the "other". This potential influence of the imagined typical welfare recipient could be a nodal point to combine both research strands. The spillover literature may expand welfare chauvinism research towards questions about possible further-reaching effects of excluding the alleged "other" from basic social resources and may emphasize questions about how the "other" is socially constructed. On the other hand, welfare chauvinism and deservingness literature can offer important concepts and explanations on how welfare programmes and regimes function and the varied ways the constructed "other" is excluded.
Although the following empirical analysis primarily focuses on whether a spillover of racism towards welfare opinions can be observed in Austria and how the media frames and links the topics of welfare to the migrant "other", the concluding discussion will reintroduce the possible connection to welfare chauvinism.

Research question and methods
The empirical analyses address two research questions, which the previous arguments suggested are interconnected: Is there a spillover of racism towards opinions on the welfare state within the Austrian population? How are the topics of the migrant or racialized "other" linked to the topic of welfare state within newspaper articles? Two types of empirical material were produced to answer these questions: survey data and a structured inquiry and content analysis of relevant newspaper articles.
The telephone survey was conducted as part of the Solidarity in times of crisis (SOCRIS) project between July and September 2017 and the target population was economically active people in Austria aged 18-65 years, regardless of citizenship. The contact data were randomly selected from public registries and the commissioned polling institute created and maintained the contact database. The random sample was combined with a quota selection based on age, gender, and region that replicated the distribution provided by Statistik Austria, which resulted in 1,004 completed interviews. Unweighted data were used since the analysis focuses on correlations. 5 The opinion towards the welfare state and social benefits was operationalized in two ways (see Table 1): First, as the demand for "less", the "same", or "more public/state support" for refugees, the long-term unemployed, and families with many children. This addresses the question of who should receive support. Second, under which conditions support should be granted, namely whether welfare support should be tied to previous contributions or not.
Racism was surveyed using two questions concerning criminalization and the devaluation of the "other" (see Table 1) and later combined into one indicator. These two questions used the general term "foreigner", allowing flexibility and the potential for racialized othering, as mentioned above. In a first step, bivariate correlations were performed between the welfare support and racism items. Next, multiple regression models were conducted on variables including socioeconomic (age, education, household income 6 ), political alienation, and other ideology variables (authoritarian views, social dominance orientation, performance ethos, nationalism) as controls (see Table 2). The regression models did not include grouping variables like gender and age, because they risk essentializing and homogenizing the results by creating "collateral realities" (Law 2011). Because there were only three options in the level of support question, they were recoded into binary variables for logistic regression (Behnke 2015). For supporting refugees and the long-term unemployed, the "same" and "more support" were merged, while "less support" remained unchanged. By contrast, due to a very low number demanding "less support" for families, "less" and the "same" were merged, while "more support" remained unchanged. Hence, the two variables on refugees and long-term unemployed test if racialized survey responses suggest demands for less, while the one on families tests whether they hinder demand for more support. Finally, an interval scale was used for the variable representing contribution-based social benefits.
The second study encompasses a qualitative content analysis of newspaper articles published between January 2015 and December 2017. This period covers the 2016 Austrian presidential election, the 2017 general election, and continues to the survey administration period. Hence, the analysis considers articles produced under both "normal" and "election" conditions. The relevant newspaper articles were identified using the Austrian Press Agency's APA-OnlineManager 7 database with access provided by the University of Vienna. Several combinations of the phrases "welfare state", "migration", "immigration", "benefit fraud", "need-based minimum benefit", "migrant", "migration background", and "foreigner" were used to identify relevant articles within eighteen Austrian newspapers preselected by the database. These included traditional and tabloid press, and national and regional publications. Ultimately, 106 documents were identified as relevant and comprise many formats including, e.g. reports, interviews, and letters to the editors.
The analysis drew on Mayring's (2000) recommendations for systematic qualitative content analyses combining inductive and deductive approaches. Qualitative content analysis is an interpretative method used to depict and refine categories based on social theory or derive new categories from the empirical material based on the research questions. This is done by successively reading, paraphrasing, and classifying texts to identify relevant passages from the specified and progressively improved coding rules. Two categories were initially selected for the deductive approach. Gilens's (1995) work informed the definition for "equation of welfare support with migration", while "immigration into the welfare state" was based on its prominent role in the 2017 election. However, both categories were further refined using the empirical material. Two additional categories, "desire for controlled migration" and "migration as a threat to the Austrian welfare state", were inductively constructed. Because this article focuses on strategies within the publicized debates, opinions, and reports, no follow-up quantitative analysis was conducted.
Is there a spillover effect within the survey population?
Is there a connection between opinions towards the migrant "other" and welfare policy? This will be first addressed by discussing the survey data. The survey population expressed notably different demands for "more", the "same", or "less" public help, depending on which of the three groups of beneficiaries they were asked about (see Table 3). While few (6 per cent) argued for less and a clear majority (61 per cent) for more support for families with many children, that was not the case for the long-term unemployed and for refugees. By contrast, 28 per cent called for less and 24 per cent for more support for the long-term unemployed, and 30 per cent and 21 per cent for refugees, respectively. The majority of respondents expressed the desire to maintain 2017 support levels, which reflects previous survey studies on public support for these groups by Grausgruber (2019) and van Oorschot (2000).
Additionally, more respondents were in favour of a contribution-based welfare model (around 18 per cent strongly and 30 per cent agree) than opposed to (6 per cent strongly and 24 per cent disagree). Twenty two per cent of the survey population neither agreed nor disagreed that only those paying taxes and making contributions should receive social benefits.
For the racialization variables, approximately 24 per cent agreed and 9 per cent agreed strongly with the statement that "foreigners" increase crime rates in Austria (see Table 4), while 19 per cent neither agreed nor disagreed, 33 per cent disagreed, and 15 per cent strongly disagreed. Hence, roughly one third of the survey population associated "foreigners" with crime. This is much lower compared to the 2016 Social Survey Austria, which covered the entire Austrian population and 65 per cent (rather) agreed with this statement (Hofmann 2016). Regarding the second item, around 42 per cent agreed or strongly agreed that "foreigners" contribute to the welfare of Austria. This Table 3. Frequencies for dependent variablespublic support for families with many children, long-term unemployed and refugees and approval of a contribution-based welfare model. constituted a much higher approval rating than for a similar question in the Social Survey Austria, where only 27 per cent agreed that immigration is beneficial for the Austrian economy. Hence, our survey population displayed fewer negative opinions towards "foreigners" than participants in the Social Survey Austria. However, like the Social Survey Austria, both variables were significantly (indirectly) correlated (Pearson = −.39). Table 5 displays bivariate correlations on how racism may shape opinions towards the welfare state. There were significant correlations between all racialized and welfare variables, although they were weakest concerning support for families with many children. Associating "foreigners" with crime and rejecting their contribution to the welfare of Austria was connected to demands for less support for families, although weakly (tau-b = −.08 and .10). This relationship became more distinct when it came to support for the long-term unemployed (tau-b = −.23 and .28) and was especially strong for refugees (tau-b = −.42 and .40). These two variables also showed an effect on the opinion towards contribution-based social benefits. Associating "foreigners" with crime yielded an effect (r = .43) that was as strong as the refugee support variable. Surprisingly, the question about whether foreigners contribute to the welfare of Austria, which is more directly connected to work, achievements, and performance, demonstrated a smaller impact (r = −.30) on demands for contribution-based welfare than questions about the association between "foreigners" and crime.
These correlations point towards a potential spillover of racism onto opinions towards the welfare state. The linear and logistic multiple regression analyses elaborate upon the effect of racism, including the previously described control variables (see Table 6). Additionally, Table 7 addresses potential multicollinearities between the independent variables, including the correlation coefficient for all independent variables, and shows several significant interrelations, with nationalism and authoritarianism being the highest (0.55). However, in consideration of Allison's (1999) recommendation about critical values, the collinearity statistics in Table 7 do not suggest any effect from extreme and problematic multicollinearity in the regression. The tolerance for each variable is well above 0.4 and the variance inflation factor (VIF) below 2.5. All four regression models showed a significant correlation arising from the racism item, even after controlling for socioeconomic status, education, past and expected changes to household income, alienation from the political system, and other ideologies. In terms of supporting refugees, the racism item was the only significantly influential variable besides the authoritarianism item. A more pronounced, racialized response increased the likelihood of demanding less public or state help for refugees, regardless of socioeconomic status. Hence, rejecting state help for refugees does not seem to arise from the competition over scarce welfare resources or the fear of not receiving future support if needed, but more from negative views of "foreigners" (see also Hofmann 2016).
The results also show that racism influenced opinions about public assistance for the long-term unemployed. Besides the variables stating that only hard work can lead to success (performance ethos) and on authoritarianism, which both showed significant influence, negative views towards "foreigners" increased the likelihood of demanding less public help for the unemployedalthough to a lesser extent than for refugees. Negative views of "foreigners" also significantly reduced the likelihood of demanding more help for families with many children, although to an even lesser degree. In this case, income was the only other significant variable, wherein higher income reduced the likelihood of demanding more support for families.
Preference for contribution-based welfare was affected by several attitude and ideology variables. Disappointment with and alienation from the political system, idealizing "hard work", orientation towards social dominance, and authoritarianism were all positively correlated to preferring a contributionbased welfare approachwith racism, however, showing the strongest effect. A significant segment of the survey population demonstrated that adverse attitude towards "foreigners" accompanied the belief that only people who pay taxes and contributions should receive social benefits.
While other, disregarded endogenous variables may influence the overall model, the influence of the racialization variable can nevertheless be assumed as stable, given similar results from previous survey data that also stress its significance with different variable combinations (e.g. Gilens 1995;Grdešić 2019). Therefore, the results from this study's regression analysis  ETHNIC AND RACIAL STUDIES support the notion that a spillover effect of racialized views exists within the survey population. Nonetheless, the data alone cannot fully clarify how racism and welfare opinions become connected in preferences for reduced welfare support when negative views of the racialized "other" are present.
Connecting the topics of welfare support and the racialized migrant "other" in media reports How is the notion of a racialized "other", especially represented as a migrant "other", entwined with welfare support? The newspaper article analysis addressed this question covering the years 2015-2017. It examined the manifold ways that the welfare state is linked to "migration" or the migrant "other", both in general terms and regarding specific measures, e.g. social assistance. The qualitative content analysis applied a framework based on four thematic categories that reflect the research question: The desire for controlled migration, migration as a threat to the welfare state, immigration into the welfare state, and equating welfare support with migration. Most of these categories are tightly interconnected and overlapping, meaning that their separation is primarily for analytical purposes. The following discussion uses phrases and arguments translated by the author to illustrate the results, while the references and quotes are consecutively numbered and information about the newspaper articles and the original German-language quotes are included in a separate online document. 8

Desire for controlled migration
The first category, desire for controlled migration, addresses controlled migration and entails feeling entitled to decide who is and is not allowed to migrate (into Austria). Therefore, the migrant "other" is separated into groups who are allowed or asked to immigrate, and groups not welcome to come or stay, which connects to a demarcation dominated by the alleged needs of the Austrian labour market. As the article "Does Austria need Immigration? Yes but … " stated: "This country with its low birth rate needs young, working people who keep Austria and the economy running. But it is not always the skilled workers needed, who are immigrating" (1). Or formulated more bluntly in a comment by an economist: "(Welfare) states are like magnets for people fleeing from poverty and attract more of them than advisable from an economic point of view" (2).
These ideas accompanied the view that migration is not for the sake of the migrants, but for that of the host country. As Faßmann, who was Austria's then-"migration expert", and later its minister of education, argued in 2016: "Without a doubt migration has to be oriented toward qualification and oriented toward the interests of Austria" (3). This notion of migration stems from a colonial worldview that reduces migrants to resources for the sole sake of a country's own betterment. This opinion was also present in several newspaper articles and published opinions when expanded towards the welfare state, such as: "Migrationespecially of young and well-educated personsis particularly important for countries with a comprehensive welfare state" (4), as prominently argued on one newspaper's front page.
However, concerning the especially relevant matter of the spillover of racism, the Austrian welfare state itself appeared under scrutiny over whether it promotes migration that benefits or harms the country. On the one hand, the welfare state was depicted as attracting the "wrong", namely poor, migrants. On the other hand, it was suspected of not only pushing away qualified migrants, but also of producing the outgoing migration of hardworking and well-qualified Austrian citizens. Both arguments were raised in an article by an author who regularly publishes in several Austrian newspapers. He saw the "voluminously dimensioned welfare state" (5) as an important factor in attracting less-qualified workers who are predestined to remain in the lower segment of the labour market, while none of the "thousands of Austrian high-performance migrants" would pass on the opportunity to leave the republic for a "portulent welfare state" (6). By contrast, the author argued that there can only be one welfare state: It can either be a comprehensive welfare state that includes undesired migration (in-and outgoing) or a restricted one with desired, controlled migration. The author showed a clear bias towards the latter, which was also shared by the Ministry of the Interior's Migration Council in arguing for an adapted social system, as "qualified migrants do not ask for social benefits but for low taxes" (7).

Migration as a threat to the Austrian welfare state
The second category, migration as a threat to the Austrian welfare state, points in a similar direction, but takes on a more dramatic tone and sees migration as a threat to the Austrian welfare state. Several articles pointed out the (possible) high costs of migration to the welfare state, especially to the hard-working taxpayers. One opinion piece raised the question of who in Austria will pay for the costs of forced migration in its title, with the author later blaming refugees for levelling down income in the low-wage segment, higher taxes, increased government debt, a reduced per-capita income, and endangering the social system in Austria (8). This was accompanied by a prediction of the end of the welfare state "as we know it". Several articles, comments, and letters to the editors expressed certainty that migration and the welfare state were necessarily mutually exclusive; for example, an editorial framed as a confession commented: Here, one also has to be true to oneself: Do we want to preserve the present level of our welfare state to some extent? Or do we want to share and refrain from certain comforts? Both will be difficult to fulfil. (9) The argument's subtext was stated more bluntly in a reader's letter: "What do we want: open borders or a functioning welfare state? Both are not possible", which extended the argument to specify that Austria's current (migration-friendly) policy hides behind a "phony moral adverse to the interests of our own people" (10), hence also indicating a strong ideology that distinguishes between an agonistic "we" and the "other".
These assessments of the alleged problem that migration poses to the continuity of the welfare state were often followed by demands to implement welfare state reforms that reduce support for migrants. However, these also served to undermine general social benefitsand especially social assistance, as was extensively described in one published comment. By invoking the spike in the number of people migrating towards Austria in 2015, the author asked whether the welfare state provides enough incentives for labour market participation, especially compared to the minimum wages in many industries. To him, the current level of social assistance has subverted the work ethos, which he saw as precious and worth protecting. Although the author started by discussing migration, he transitioned to the broad statement that "a reduction of social assistance is not just necessary so that we will recognize our welfare state in many years to come but especially to come back to a point where hard work is rewarded again" (11). The shift from demanding welfare reform in light of the migrant "other" to a universal reduction for all Austrian residents appeared less subtly in a newspaper's letter to the editor: "For foreign families, where one has never ever worked or who live from social benefits alone, cut all support, also for all refusing to work" (12).

Migration into the social system
Although it was present well beforehand, immigration into the Austrian welfare system became more salient during the 2017 election (13). This was addressed in interviews (14) and speeches (15) by the leading ÖVP candidate, which elevated its prominence in newspapers and media who were already being courted by the "populist radical right" (Ennser-Jedenastik 2020) FPÖ (16). The ÖVP continued the pull argument described above, but also claimed that migration had already and will further damage the welfare state in Austria. This was accompanied by demands to reduce welfare support below the already-low minimum rate for "foreigners", which had been implemented in several federal states (Bundesländer) only to be later abolished by the constitutional court. 9 According to a report in the newspaper DerStandard, the regulation for reduced minimum benefit aimed at the "others" in the state of Lower Austria, however, also showed negative consequences for residents not (necessarily) deemed as " foreigners" (17).

Equation of welfare support with migration
All of the texts and articles described so far may lead to welfare beneficiaries being associated with "foreigners"; however, this also occurs directly. During the 2017 election, the ÖVP leader's statement that "every other social assistance beneficiary in Vienna is a foreign citizen" was heavily repeated by many newspapers without any context (18).
For one tabloid press, the number of "foreigners" receiving minimum benefits alone was newsworthy enough to publish the title, "Social Benefits for Around 45,000 Foreigners". This was followed by "just 60 per cent of the millions of taxes goes to Austrians, the rest to other citizensjust a few of them from the EU" (19). The tabloid press also regularly publishes articles on cases of what they consider inappropriate amounts of received welfare support, usually indicating the recipients' alleged "origins" or "foreign roots" (20).
Reports, comments, and published rumours about benefits fraud further underlined the notion that mainly "migrants" benefit from welfare support. During the 2016 Austrian presidential election, an article reproduced the rumour spread by one candidate that: […] when representatives of a Catholic charity recommend a family of refugees to relocate to Vienna, although they already found a flat and work in Salzburg, just to receive more from social assistance than from working, then we know that benefit fraud has become endemic. (21) Another article stating the false claim that "bigamist Muslims" would receive additional welfare support for their "second wives" (22) was criticized as false and discriminatory by the Austrian Press Council, a self-regulated and voluntary controlling body for good and ethical standards in the press. The article, however, is still online.
Alternative narratives that combine migration and the welfare state differently were comparatively rare. Some argued that migration has a positive effect on financing and organizing the welfare state, and without a colonial undertone or agenda (23). Likewise, some noted that the welfare state's main function is to support people who do not have enough to survive and taking this away would be contradictory (24). Additionally, two articles also discussed the spillover effect of racism and how Austrian political parties use it to undermine the welfare state (25). Nevertheless, most articles casted a damning light on the relationship between migration and the welfare state in (at least) the four ways described in this study and summarized in Table 8. Furthermore, all four strategies were accompanied by demands for economization, cutbacks, and tightened controls.

Discussion
The results from both the survey and content analysis suggest that in Austria, racialized demarcations are tied to opinions about welfare and social support not just for the alleged "other", but also the form and extent of welfare support itself. The survey results align with similar studies conducted in other countries (e.g. Harell, Soroka, and Ladner 2014;Alesina, Miano, and Stantcheva 2018). Negative opinions towards "foreigners" are both entangled with increased demands for less refugee support, but also the long-term unemployed. These are also tied to demands for a contribution-based welfare system and, to a lesser extent, to denying additional support for families with many children. However, the figures alone cannot fully infer causality from correlation. Rather, the quantitative analysis shows that racialized ideologies mainly hold the potential to inform opinions towards the welfare state. At the same time, the newspaper article analysis illustrates how the notion of the migrant "other" is made relevant to the welfare state itself. All four identified narratives connected the welfare state with migration that made them seem naturally co-dependent in a harmful way. In this regard, the welfare state is described as limiting the potential for Table 8. Overview of content analysis-categories and specifics.

Category Specifics
Desire for controlled migration Distinction between right (workers) and wrong (welfare beneficiaries) migrants Migration for the (economic) sake of one's own country (Extensive) Welfare state driving away the highly skilled and hardworking Welfare state as magnet for (unwanted) migration Migration as a threat to the Austrian Welfare State High costs of migration for the welfare state Either migration or welfare state Demands for welfare reformsreduction of benefits for migrants but also in general Migration into the social system Migrant's access to welfare support (EU and general) Lowering social support below the minimum rate Equation of welfare support with migration Scandalization: Needs-based minimum benefits and other social support mainly for migrants Benefit fraud by migrants controlled migration and attracting the "wrong" migrants on the one hand. This unwanted migration, on the other hand, is seen as endangering the welfare state itself and hindering access to welfare resources for the "own" group. Both accumulate into demands for welfare cutbacks as solution for both alleged problems.
However, a few alternative narratives and arguments show that different potential approaches exist but are not used. This is arguably deliberate, as previously suggested by Hall. Most of the published claims made by the authors are demonstratedly unproven as fact; for example, there is no clear and final value that represents the financial cost of migration versus its benefits for Austria (see, e.g. Prettenthaler, Neger, and Apostle 2019). Instead, published figures were usually incomplete and (intentionally) misleading; for example, regarding the share of the unspecified "foreigners" receiving social benefits while ignoring their potential previous contributions to the social system, whether they receive full support or a supplement to an existing occupation, that the alleged costs amount to a miniscule fraction of Austrian welfare expenditures (Atzmüller, Knecht, and Bodenstein 2020), and especiallyhow "foreigners" would survive without support. The way the authors frame the relationship between migration and the welfare state in their articles is deliberate and reflects previous media studies discussed above (Gilens 1999;Clawson and Trice 2000;Larsen and Dejgaard 2013).
The strategies used in these publications also show how migration itself is racialized via the topic of welfare support. They introduce citizenship, nation states, and "nativeness" as self-evident demarcations in access to welfare without addressing the roots of this differentiating power. Furthermore, the right to make migration decisions and determine who is allowed to benefit is taken for granted by using the survival of the welfare state as a rhetorical device to differentiate between whose needs, life, and survival are worth more and whose are worth less.
Nevertheless, welfare chauvinism and, hence, excluding the "other" from access to welfare support, is a policy stance taken by many of the analysed articles, but many go even further. Most of the identified articles also use migration as a means to challenge the welfare state and social support itselfnot just for the migrant "other", but for society in general. However, like in the welfare chauvinism literature, only select welfare programmes, especially social assistance, have a strong, negative association with the "other"either by allegedly pulling in the "wrong migrants" or by being exploited by the undeserving "others". If there is a causal effect, meaning a spillover of racism onto the opinions towards welfare programmes, the narratives identified within the media reports show (some of) the ways this causality is produced, enacted, and cultivated.

Conclusion
The article offers a particular approach to addressing the spillover of racism towards welfare state opinions. First, it uses the research literature to embed existing studies on the spillover within the broader (German-speaking) welfare state literature as an instance of socialization. Next, the results from the empirical material and analysis both address whether a spillover exists within the Austrian population and how media discourses helped entangle these two issues. So far, most publications have only discussed these topics separately.
At the same time, this study has some limitations. First, it required assuming that the newspaper articles and survey opinions towards the welfare state and the alleged "others" are connected based on the research literature, but this could not be tested. Rather, they are solely linked topically. Second, the special population used for the survey sample may foster or hide certain correlations present in the general population. Furthermore, limited resources meant that racism and opinions towards the welfare state within the survey were operationalized in a rudimentary manner. Finally, the qualitative content analysis of the media reports is a first approach on this topic in Austria that must be expanded upon and refined in future iterations.
Some additional topics could only be briefly discussed. Content analyses of newspaper articles can stimulate debates on the media's responsibility, influence, and role within society. Here, Hall's (2018Hall's ( , 148f orig. 1982) position that the media always acts within the sphere of ideologieswhether journalists like it or notmust be reiterated. However, determining how to adequately address this within the national and international media landscape goes beyond the scope of this article. Nevertheless, such future discussions are advisable, because the racialization of welfare state opinions exists within ongoing debates about the general social functions ascribed to the welfare state: its role permitted vis-a-vis the labour market as well as the general demarcation between labour-market and non-labour market citizens (Lessenich 2012) and between the deserving and the non-deserving (Faist 1995). Literature on welfare chauvinism and deservingness may help accomplish this and sharpen the empirical analysis, which will be incorporated into future studies.
This article advanced the research on the public opinion towards the welfare state, welfare chauvinism and deservingness by stressing the potential wide(r)-reaching consequences of associating welfare programmes with the racialized "other". This especially concerns the form of the migrant "other", which remains speculative in other conclusions reached by welfare chauvinism literature (e.g. Atzmüller, Knecht, and Bodenstein 2020). Focusing on the spillover highlights how the racialization of welfare measures also impacts fundamental qualities of the welfare state. This includes whether it should provide a living for those outside of the labour market, if need should play a more important role than previous contributions, and if everyone should be treated equally or not. These questions are magnified when combined with the topic of the racialized "other". The racialization of social benefits debates enablesor at least more easily enableslegitimizing benefits well below the defined minimum standards necessary to survive (in Austria), or to outright refuse them, even if only (currently) argumentatively. Following Kaufmann's (2002) notion of the welfare state being an instance of socialization, this paper introduced the idea that using the figure of the racialized "other" survival is no longer guaranteed or envisaged by (Austrian) society and the state if someone is not deemed useful to society, and being initially pitted against the racialized "other" is not necessarily restricted to them.