“These benefits are ours because we were here first”: relating autochthony to welfare chauvinism and welfare ethnocentrism

ABSTRACT

A substantial portion of Europeans opposes granting welfare benefits to immigrants (welfare chauvinism) and to longer-established ethnic minorities (welfare ethnocentrism). We aim to explain welfare chauvinism and welfare ethnocentrism by introducing autochthony as a novel determinant. Autochthony is the general belief in entitlements for firstcomers. Using a representative sample of British (N = 3,516) and Dutch (N = 1,241) natives, we find that autochthony indeed predicts higher welfare chauvinism, even after taking into account a great range of existing explanations. Moreover, an experiment among British natives showed that autochthony not only explains welfare chauvinism towards immigrants but also welfare ethnocentrism towards established ethnic minorities and Muslims. However, autochthony did not explain welfare ethnocentrism towards black Britons. Our findings indicate that the argument “we were here first” can help to shed light on welfare attitudes that are of growing importance in diverse Western societies.

KEYWORDS:

• Autochthony
• welfare chauvinism
• welfare ethnocentrism
• collective psychological ownership

Introduction

The welfare state is a central institution in Western societies that aims to protect and promote the economic and social well-being of citizens.¹ Welfare provisions are scarce, which makes exclusion criteria inevitable, and political discussion concerning those criteria increasingly focuses on issues of migration and ethnicity (Kymlicka and Banting 2006; Newton 2007). A substantial portion of European citizens sees immigrants as less entitled to welfare benefits than the rest of the population (Reeskens and Van Oorschot 2012). This attitude, labelled welfare chauvinism, has received increasing scientific attention, and studies have explained individual differences in welfare chauvinism with ideological beliefs, attitudes towards ethnic relations and migration, and economic insecurity (Ford 2016; Kros and Coenders 2019; Van Oorschot 2006; Reeskens and Van Oorschot 2012). Moreover, people do not only oppose welfare entitlements for migrants but also for longer-established ethnic minority groups whose migration status is less salient (welfare ethnocentrism, Ford 2016). Although empirical work on welfare ethnocentrism is scarce, explaining it is of increasing relevance in European societies where ethnic diversity is rising (Castles and Miller 2009), and large established ethnic minority communities have existed for several generations.

This paper introduces autochthony as a neglected but potentially very relevant underlying ideological belief that may explain welfare chauvinism and welfare ethnocentrism. Autochthony is a belief in entitlements for firstcomers. “We were here first” is often considered a valid argument for claiming ownership of a country and for being entitled to determine what happens within its territory (Geschiere 2009). Based on autochthony, natives may feel entitled to exclusive use of “their” welfare state, and bar newcomers from it, simply because they arrived later and are therefore not regarded as rightful owners (Martinovic and Verkuyten 2013).

Anthropological studies have pointed out the importance of autochthony in Western European populist discourse in justifying welfare chauvinism (Ceuppens 2006). Although systematic quantitative studies have shown that autochthony can explain prejudice towards migrants, opposition towards Muslim expressive rights, and support for collective action against refugees (Hasbun Lopez et al. 2019; Martinovic and Verkuyten 2013; Smeekes, Verkuyten, and Martinovic 2015; Verkuyten and Thijs 2019), no study has yet examined the importance of autochthony as a basis for welfare chauvinism and welfare ethnocentrism among the general public. This is surprising because autochthony and the exclusive right to “our” resources suggest an intuitive link to views about social welfare.

In this study, we test whether autochthony contributes to explaining welfare chauvinism and welfare ethnocentrism, while also taking into account important existing explanations. We relate autochthony to welfare chauvinism using representative samples of British and Dutch natives, which also enables us to examine whether autochthony is a relevant factor in different welfare regimes. Using the British sample, we furthermore examine whether autochthony explains opposition towards welfare entitlements for migrants only (welfare chauvinism), or also for ethnic minorities, Muslims, and black Britons, whose migration status is assumed to be less salient (welfare ethnocentrism). Whereas welfare ethnocentrism has not been studied much, there is extensive literature on welfare chauvinism, and we build our theoretical framework primarily around this literature.

Theoretical framework

Explaining welfare chauvinism

In Western European democracies, citizens consent to contributing parts of their income to a common pool with the aim of assisting fellow citizens who are ill, unemployed, or at an age to be entitled to receive a pension (Ford 2016). This welfare state system is resource intensive and demands a high level of solidarity among citizens. The qualification criteria for assistance have become intensely debated (Kymlicka and Banting 2006). Migration and ethnicity have become important topics in these discussions (Gilens 1999; Kymlicka and Banting 2006; Newton 2007; Reeskens and Van Oorschot 2012), due to increasing cultural diversity in western societies (Schmidt-Catran and Spies 2016). Western welfare states were generally founded in times of greater cultural homogeneity. Mass migration and rising diversity are argued to have introduced new strains on these systems (Reeskens and Van Oorschot 2012). The idea that immigrants are less deserving of welfare support has become a popular and effective political message for populist radical right parties (De Koster, Achterberg, and Van Der Waal 2013). Sympathy with this message is widespread among native majority voters (Van Oorschot 2006). On average, 76 per cent of citizens of twenty-four European countries believe immigrants’ rights to social benefits should be conditional in some way, which is referred to as welfare chauvinism in the soft sense (Reeskens and Van Oorschot 2012). Moreover, 7.5 per cent adhere to welfare chauvinism in the strict sense, meaning that immigrants should never obtain rights to social benefits.

Traditional explanations of welfare chauvinism can be broadly divided into arguments related to ideology, ethnic attitudes, and economic insecurity (Ford 2016; Kros and Coenders 2019; Van Oorschot 2006). The ideology argument approaches welfare chauvinism as a manifestation of general ideas about how society should function. Welfare chauvinism generally coincides with general welfare redistribution preferences. For example, more egalitarian citizens are less prone to endorse welfare chauvinism, as egalitarian ideology calls for economic equality for everyone, without distinguishing between natives and migrants (Frankfurt 1987; Kros and Coenders 2019; Reeskens and Van Oorschot 2012). Furthermore, self-identified right-wing voters are more welfare chauvinist than self-identified left-wing voters (De Koster, Achterberg, and Van Der Waal 2013). Also, authoritarianism relates to more welfare chauvinism as excluding immigrants from the welfare state can be seen as a way to restrict the rights of outgroups who challenge the existing social and political order (Crepaz 2020; Kehrberg 2020).

Other scholars argue welfare chauvinism is rooted in general attitudes towards ethnic relations and migration. White Americans’ attitudes towards welfare policies were found to be better predicted by attitudes towards African Americans than ideology or self-interest (Gilens 1999). In Europe, ethnic threat, which is one of the most important predictors of attitudes towards immigration (Hainmueller and Hopkins 2014), is found to predict welfare chauvinism (Kros and Coenders 2019). In a similar vein, social norms against racism might relate to lower welfare chauvinism, as those who express a strong commitment to anti-racism may consider discriminating against migrant welfare claimants to be a violation of this norm (Ivarsflaten, Blinder, and Ford 2010). Another explanation related to ethnic attitudes concerns conceptions of national citizenship. Wright and Reeskens (2013) proposed that for solidarity to exist, citizens should have a sense of national identity and the extent to which people show solidarity with immigrants depends on their conceptions of the national “we”. Specifically an ethnic conception, with national identity bounded by ethnic ties, is found to relate to welfare chauvinism. Majority members with an ethnic conception of the nation perceive no group ties with immigrants because of their different ethnic descent, which results in a lack of solidarity with immigrants, and a lower willingness to support this group.

Finally, welfare chauvinism is also rooted in economic insecurity. Natives who are, or perceive to be, in more insecure economic positions oppose granting social assistance to immigrants, to prevent the scarce resources to be unavailable when they themselves need them (Ford 2016; Van Oorschot 2006). Although economic risk confounds with ideology and ethnic attitudes, it can also independently predict welfare chauvinism (Kros and Coenders 2019). Both objective indicators (unemployment, welfare dependency) and perceptions of economic insecurity can play a role.

Autochthony and welfare chauvinism

We argue that welfare chauvinism is also related to beliefs about the basis of welfare entitlement Some majority members feel themselves entitled to exclusive use of the welfare state, and the literature on collective psychological ownership provides some distinctive possible explanations for this feeling (Verkuyten and Martinovic 2017). A group can have a sense of possessiveness about an object, place, or idea that they perceived to be “theirs” (Pierce and Jussila 2010). Such feelings of collective ownership are grounded in the psychology of possession which develops early in life (Rochat 2014; Ross, Friedman, and Field 2015; Rossano, Rakoczy, and Tomasello 2011). A collective sense of ownership can similarly be expressed with regard to a country (Brylka, Mähönen, and Jasinskaja-Lahti 2015; Verkuyten and Martinovic 2017). Ownership is typically accompanied by certain rights, such as the right to use what is owned and to prevent others from using it (Nijs et al. 2021; Snare 1972). Therefore, a sense of collective ownership over a nation-state might translate into perceived entitlements to exclusive use of “our” welfare state and the right to exclude others from its benefits, particularly since the welfare state is often perceived to be central to what “our” country entails (Ford 2016).

People have different reasons for perceiving their group as rightful owners, but first arrival (i.e. autochthony) is generally the most important principle for claiming territorial ownership. “We were here first” is used by children, for example to claim ownership of a piece of land where flowers can be picked, or a place on the beach where a sandcastle can be built (Verkuyten, Sierksma, and Martinovic 2015; Verkuyten, Sierksma, and Thijs 2015), and claims of country ownership follow similar logic (Geschiere 2009; Martinovic and Verkuyten 2013; Verkuyten and Martinovic 2017). Anthropological work has shown that autochthony is a strong belief that plays a role in very different intergroup contexts, for example in political struggles between different Cameroonian tribes and in debates about immigration in the Netherlands (Geschiere 2009). Autochthony is expected to specifically underpin welfare chauvinism because it implies a clear boundary between natives, who were “here first”, and migrants, who arrived later (Ceuppens 2006, 2011). Migrants, by definition, are not the first inhabitants, and those who adhere to the principle of autochthony will therefore see migrants as less entitled to a share of resources that are owned by firstcomers (Martinovic and Verkuyten 2013). Moreover, the welfare state is provided via long-established institutions founded by earlier generations, which can establish a sense that the welfare state should be exclusively used by natives. We argue that autochthony is a general belief about entitlements for newcomers, that can provide a distinct underpinning for more specific attitudes, such as welfare chauvinism (Verkuyten and Martinovic 2017). In short, we hypothesize that:

Autochthony is positively related to welfare chauvinism among British and Dutch natives (H1), over and above other ideological, ethnic, and economic determinants of welfare chauvinism.

Autochthony and welfare ethnocentrism

Discussions concerning welfare entitlements focus not only on recently arrived migrants but also on established minority groups. In the United States, race has been a salient feature of welfare debates with negative perceptions of African Americans significantly associated with white opposition towards welfare provision (Gilens 1999). European discussions focus more on migration, with reciprocity being an often-heard argument (“newcomers did not contribute yet”), though there is evidence that white Europeans are also less willing to provide welfare to ethnic minority claimants (Ford 2016; Ford and Kootstra 2017; Kootstra 2016). As European societies become more diverse, and ethnic minorities become more established, race, ethnicity, or religion may become more salient dividing lines, used to define outsiders and exclude them from access to welfare.

Whereas the theoretical reasoning underlying the link between autochthony and welfare chauvinism is clear given that migrants (by definition) cannot claim to be autochthonous in their new place of residence, the link between autochthony and welfare ethnocentrism is more complex. Although the migration origin of some minority groups (i.e. ethnic minorities, Muslims, black Britons²) is less salient, these groups might not be perceived as autochthonous either. Autochthony might exclude any minority group with a widely perceived migration origin from welfare entitlements because first-occupant status cannot be earned by working hard, trying to integrate, or speaking the language (i.e. ethnic minorities descended from migrants will never become firstcomers) (Feather 2003). Therefore, we expect autochthony to also be related to opposition to welfare entitlements for Muslims, ethnic minorities, and black Britons.

However, autochthony might be a comparatively weaker predictor of welfare ethnocentrism than of welfare chauvinism. Autochthony offers a clear us-them distinction between natives and migrants but the line between natives and longer-established minorities is more blurred. The us-them distinction in welfare ethnocentrism is not primarily based on first occupancy, but more on ethnic, racial, or cultural differences. We therefore expect the relationship between autochthony and opposition to welfare entitlements to be stronger with respect to immigrants than with respect to established minorities.

Additionally, autochthony might be a less relevant explanation of welfare ethnocentrism towards black Britons than towards ethnic minorities and Muslims. Based on self-categorization theory (Turner et al. 1987), we expect that white Britons perceive black Britons as part of a superordinate category of Britons, which leads to more solidarity, a stronger sense that they contributed to society, and therefore, a smaller chance that this group is excluded from welfare entitlements based on autochthony (Gaertner et al. 1999; Hornsey and Hogg 2000; Transue 2007). Black Britons have a long-established status in British society (Perry 2015), suggesting that the group is indeed part of the category Britons, and the use of the label “black Britons” for this group might have a priming effect. To sum up, we hypothesize that, independently of other ideological, ethnic, and economic determinants:

Autochthony is the strongest predictor of welfare chauvinism (towards immigrants), followed by welfare ethnocentrism towards ethnic minorities and Muslims, followed by welfare ethnocentrism towards black Britons. (H2)

Great Britain and The Netherlands

We test our hypotheses among British and Dutch white majority natives. In both countries, welfare chauvinism has become an important message for right-wing populist parties with the UK Independence Party (UKIP) and the Dutch Party for Freedom (PVV) as notable examples (Ennser-Jedenastik 2018). Moreover, both countries have long histories of post-colonial and labour immigration and a persistent influx of new immigrants (Maxwell 2012; Sobolewska, Galandini, and Lessard-Phillips 2017). However, they have very different welfare regimes. Great Britain is a liberal welfare regime, whilst The Netherlands is traditionally categorized as a corporatist regime with socio-democratic characteristics (Esping-Andersen 1990). People in liberal and conservative regimes are generally more welfare chauvinistic than those in social-democratic ones (Van Der Waal, De Koster, and Van Oorschot 2013), which means that British natives are expected to be relatively more welfare chauvinistic. Nevertheless, we have no clear reasons to expect differences in the autochthony-welfare chauvinism association across the two countries. Therefore, we examine the possible country differences in an exploratory fashion.

Methods

Sample and procedure

We used data from the Welfare State Under Strain (WESTUS) survey that consists of five waves collected over a period of 16 months in 2014 and 2015 in Great Britain and the Netherlands (Ford et al. 2015). As questions about autochthony were asked only in wave 4 in the Dutch sample (N = 1.512) and wave 5 in the British sample (N = 4.468), all measures used in our analyses were obtained from these waves, unless mentioned otherwise. The samples were recruited from panels maintained by YouGov in Great Britain and TNS-NIPO in the Netherlands. With weights applied, the samples were representative for the British population as a whole in terms of age, gender, region, social class, party identity and the readership of newspapers, and for the Dutch population as a whole in terms of age, gender, region, and education level. Ethical approval was obtained from the Ethics Committee of the Faculty of Social and Behavioural Sciences of Utrecht University (number 20–175).

We selected only native born participants with two native born parents who indicated “white British” or “Dutch” as their ethnicity (3,714 British and 1,341 Dutch participants). H1 was tested on both country samples and H2 was tested on the British sample only because the experiment to test H2 was only presented to British participants. To test H1, we used a multi-item measure of welfare chauvinism as the dependent variable and included all participants who answered at least one of these items. We also excluded four Dutch participants who did not answer the unemployment item used as a control variable. This left a sample size of 3,516 British and 1,241 Dutch participants to test H1 (total N = 4,757).

To test H2, we used a measure of opposition towards welfare entitlements for specific target groups as the dependent variable. These items were presented later in the questionnaire and were answered by fewer participants than the welfare chauvinism items, which left a total sample of 3,338 British respondents to test H2. All data and code can be found here: https://osf.io/aqc58/.

Measures

Welfare chauvinism

To test H1, we assessed welfare chauvinism with four items. Participants were asked to indicate how long they thought immigrants should work and pay taxes before they were entitled to four different welfare benefits: disability benefits, housing benefits, unemployment support, and income support (Kros and Coenders 2019). Items were measured on 12-point scales ranging from 0 “They should always be entitled to this benefit” to 11 “They should never be entitled to this benefit”. The intermediate options specify the number of years (1–10 years). Higher scores indicated more welfare chauvinism. Welfare chauvinism was treated as a latent factor in a structural equation model.

Group-specific opposition towards welfare entitlements

In the experiment used to test H2, British participants were randomly assigned to one of six conditions. In each condition, opposition towards welfare entitlements was measured in relation to a specific target group, with the item “I believe [target group] on welfare are deserving of receiving the support they receive from the government” (1 = strongly disagree, 5 = strongly agree). The target group differed per condition and was respectively “migrants”, “ethnic minorities”, “Muslims”, “black Britons”, “white Britons” and “people”. We focused on the comparison between migrants as target group (i.e. welfare chauvinism), and ethnic minorities, Muslims, and black Britons as target groups (i.e. welfare ethnocentrism). The other target groups, “white Britons” and “people”, were not the focus but were kept in the analyses for exploratory comparisons. “White Britons” are the ingroup for our participants and “people” can include all groups. The item was reverse coded so higher scores indicated more opposition.

Autochthony

Autochthony was measured with four items developed by Martinovic and Verkuyten (2013): “The original inhabitants of a country are more entitled than newcomers”, “Every country belongs to its original inhabitants”, “The original inhabitants of a country have the most right to define the rules of the game” and “‘We were here first’ is an important principle for determining who decides on what happens in a country” (1 = strongly disagree, 5 = strongly agree). The construct was validated and found reliable across countries (Hasbun Lopez et al. 2019; Martinovic and Verkuyten 2013), and we treated it as a latent factor.

Political ideology

We controlled for three different indicators of political ideology. Economic egalitarianism was measured with the statement “For a society to be fair, differences in people’s standard of living should be small”, using a 5-point scale (1 = Strongly disagree; 5 = Strongly agree) (Kros and Coenders 2019). Political orientation was measured by asking participants to place themselves on a 7-point scale (1 = Very left-wing; 4 = Centre; 7 = Very right-wing) (Jost 2006; Van Oorschot 2006). Authoritarianism was measured as the mean score on three items, using a 5-point scale (1 = Strongly disagree; 5 = strongly agree) (Reeskens and Van Oorschot 2012). A sample item was “People should do what they’re told. People should follow rules at all times, even when no-one is watching”.

Ethnic attitudes

We also accounted for different indicators of ethnic attitudes. Ethnic threat was measured with two 11-point continuous scale items from the European Social Survey tapping into symbolic and realistic threat. Symbolic threat was measured by asking whether the country’s cultural life is generally undermined (score 0) or enriched (score 10) by people who come to live here from other countries. Realistic threat was measured by asking whether it is generally bad (score 0) or good (score 10) for the country’s economy that people come to live here from other countries. The items correlated strongly (r = .74), and we computed a mean score. Items were reversed with higher scores indicating more threat. Ethnic citizenship conception captures whether people define their national identity in ethnic terms and it was measured as the mean score on two five-point scale items (Wright and Reeskens 2013), indicating how important ethnic characteristics are for being truly British/Dutch (1 = Very unimportant; 5 = Very important). The characteristics were “To have been born in Britain/the Netherlands” and “To have British/Dutch ancestry”.³ Anti-racism norms were measured in wave 3 as the mean score on four statements, using a 5-point continuous scale (1 = Strongly disagree; 5 = strongly agree) (Ivarsflaten, Blinder, and Ford 2010). A sample item was “I don’t want to appear racist, even to myself”. One of the items was reverse coded (“Using stereotypes is OK by my personal values”) so that higher scores indicated stronger anti-racism norms.

Economic insecurity

We also took into account different indicators of economic insecurity. Firstly, a dummy variable indicated whether people were unemployed. Secondly, to measure welfare dependency, respondents were asked whether they claim different types of welfare benefits from the government. That is, disability benefits, housing benefits, unemployment support, and income support. We created a count variable (0 = claims no welfare benefits; 4 = claims all four welfare benefits). Economic insecurity perceptions were measured with the question “how likely or unlikely is it that during the next 12 months, there will be some periods when you don’t have enough money to cover your household necessities?” using a 5-point continuous scale (1 = very unlikely; 5 = very likely).

Background characteristics

We also controlled for basic background characteristics like age and gender. Religiosity was measured by asking “Regardless of whether you belong to a particular religion, how religious would you say you are?”, using an 11-point continuous scale (0 = not at all religious; 10 = very religious). Education was measured as a 7-point continuous variable, harmonized with the ES-ISCED scale, with higher scores indicating higher education. Religiosity and education were both measured in wave 1.

Data analytic strategy

We analysed the data in five steps. First, we performed confirmatory factor analysis using Mplus software (version 8.3; Muthén and Muthén 1998–2017) to test whether the concepts measured with multiple items, i.e. welfare chauvinism, autochthony, authoritarianism, ethnic threat, ethnic citizenship conception, and anti-racism norms captured separate latent constructs. We tested this on the total pooled sample. Second, we tested for the invariance of the main constructs welfare chauvinism and autochthony across both country samples. Third, we examined the descriptive statistics of all variables for the two countries separately. Fourth, to test H1, we regressed the latent construct welfare chauvinism on the latent construct autochthony and all control variables. These analyses were performed using multigroup structural equation modelling to test the invariance of the coefficients between both country samples. Fifth, to test H2, we regressed the one item measuring group-specific opposition towards welfare entitlements on autochthony and all control variables, using the six experimental conditions as groups in multigroup analyses. We tested for significant differences in the regression coefficient of autochthony across conditions (i.e. target groups) by testing whether constraining two coefficients to be equal significantly increased the chi-square, indicating a worse model fit. This way, we could test whether the positive association between autochthony and opposition to welfare entitlement differs across target groups (H2).

In all analyses, we used full information maximum likelihood which allows missing values in endogenous variables. Exogenous variables were therefore endogenized by estimating their variance.⁴ We employed weights for the applied data (i.e. wave 4 in the Netherlands and wave 5 in Great Britain). We conducted ordinary least squares regression analysis with robust maximum likelihood estimation (MLR) to be able to employ these weights and handle non-normally distributed variables.

Results

Measurement model

The expected 6 latent factor model (autochthony, welfare chauvinism, authoritarianism, ethnic threat, ethnic citizenship conception, anti-racism norms) with the 19 items loading on their respective factor fitted the data well according to conventional rules of thumb when British and Dutch participants were pooled together (χ² = 1065.670 (137), p < .001, CFI = .973, RMSEA = .038, SRMR = .048). All items loaded significantly on their respective factor with all standardized loadings above .51. As shown in Appendix B, we estimated different sets of alternative model configurations, but all of these models were significantly worse fits to the data. This shows that our measure of autochthony is empirically distinct from all other latent constructs, and the same holds for our measure of welfare chauvinism. In subsequent models, we treated all multi-item control variables as manifest mean scores, to reduce complexity. Such a model with only welfare chauvinism and autochthony as latent factors also fitted the data well, χ² = 125.510 (19), p < .001, CFI = .993, RMSEA = .034, SRMR = .016.

Measurement invariance

To examine whether British and Dutch participants interpreted the items measuring welfare chauvinism and autochthony in a similar way, we tested for measurement invariance of this two-factor model. A scalar invariant measurement model with equal loadings and equal intercepts fitted the data well (χ² = 353.522 (50), p < .001, CFI = .980, RMSEA = .051, SRMR = .032).⁵ See Appendix C for all model fit statistics. Modification indices suggested that freeing the intercepts of the housing benefits item leads to a significantly better fit (χ² = 246.919 (49), p < .001, CFI = .987, RMSEA = .041, SRMR = .030). British natives scored significantly higher on welfare chauvinism with regard to housing benefits (M = 5.72) than Dutch natives (M = 4.71), t(4,619) = 8.650, p < .001. We used this model with partial scalar invariance as our final measurement model, which allows us to meaningfully compare mean scores and regression coefficients across countries.

Descriptive statistics

Table 1 shows the descriptive statistics by country. British natives tended to agree slightly but significantly more with both the welfare chauvinism (t(4,666) =  4.481, p < .001) and autochthony (t(4,603) = 3.737, p < .001) items than Dutch natives. Only 7.6 per cent of the British and 3.4 per cent of the Dutch sample indicated that immigrants should never be entitled to all four welfare benefits, indicating little support for welfare chauvinism in the strict sense. However, only 2.3 per cent of the British and 2.5 per cent of the Dutch sample indicated that immigrants should always be entitled to all four benefits, indicating substantial support for some degree of welfare chauvinism in the soft sense. Correlations between all variables (see Appendix D) were generally in the expected direction. The correlation between autochthony and welfare chauvinism was positive in both the British (r = .519) and Dutch (r = .453) samples.

Table 1. Descriptive statistics per country.

		Great Britain				The Netherlands
	Range	Valid n	Mean/proportion	SD	α	Valid n	Mean/proportion	SD	α
Welfare chauvinism	0–11	3,516	6.05	3.10	.93	1,241	5.59	2.97	.91
Group-specific opposition towards welfare entitlements	1–5	3,338	3.01	1.07	–	–	–	–	–
Autochthony	1–5	3,471	3.47	1.08	.94	1,232	3.34	.99	.91
Economic egalitarianism	1–5	3,424	3.54	1.01	–	1,224	3.42	.99	–
Political orientation	1–7	3,048	4.01	1.38	–	1,146	4.11	1.26	–
Authoritarianism	1–5	3,505	3.78	.72	.66	1,232	4.02	.57	.63
Ethnic threat	0–10	3,474	4.96	2.68	.77^b	1,234	5.17	2.15	.66^b
Ethnic citizenship conception	1–5	3,480	3.45	1.19	.71^b	1,231	2.77	1.13	.77^b
Anti-racism norms	1–5	3,238	3.56	.77	.79	1,219	3.60	.70	.73
Unemployment	0/1	3,516	.04	–	–	1,241	.03	–	–
Welfare dependency	0–4	3,516	.21	.55	–	1,241	.20	.48	–
Economic insecurity perceptions	1–5	3,395	2.51	1.30	–	1,210	2.51	1.25	–
Age	18–86 / 19–94^a	3,516	49.50	16.25	–	1,241	48.85	17.46	–
Gender (female)	0/1	3,516	.51	–	–	1,241	.49	–	–
Religiosity	0–10	3,487	2.93	2.90	–	1,217	3.55	3.34	–
Education	1–7	3,516	4.91	1.93	–	1,241	3.98	1.74	–

Note: Descriptive statistics were based on manifest mean scores. α is Cronbach’s alpha. All statistics were based on the weighted data. ^aRange before the slash is for the British sample, after the slash for the Dutch sample. ^bCorrelation between the two items in the construct.

Explaining welfare chauvinism

The latent dependent variable welfare chauvinism was regressed on the latent independent variable autochthony and all manifest control variables. A model in which all path coefficients were constrained to be equal across the British and Dutch samples fitted the data well (χ² = 984.201 (267), p < .001, CFI = .970, RMSEA = .034, SRMR = .030), and did not fit significantly worse than a model in which all coefficients were free to vary across countries (TRd = 20.710 (14), p = .109).⁶ This suggests that the coefficients are invariant across the British and Dutch samples and that we can use the structurally constrained model in subsequent analyses.

Table 2 shows all unstandardized regression coefficients of three models. In the first model, welfare chauvinism was regressed on background characteristics only. Respondents with fewer educational qualifications were more welfare chauvinistic. Age, gender and religiosity were not significant predictors. In the second model, control variables concerning political ideology, ethnic attitudes, and economic insecurity were added. People who were less economically egalitarian, more right-wing, and more authoritarian, were more welfare chauvinistic. Ethnic threat and ethnic citizenship conception were associated with more welfare chauvinism. Social norms against racism were related to less welfare chauvinism. Finally, unemployment and economic insecurity perceptions were not significantly related to welfare chauvinism, whilst more welfare dependency was associated with less welfare chauvinism, which is unexpected based on the self-interest argument. We return to this issue in our discussion.

Table 2. Unstandardized regression coefficients for the model with welfare chauvinism as the dependent variable.

	Model 1	Model 2	Model 3
	Welfare chauvinism	Welfare chauvinism	Welfare chauvinism
Autochthony			.565 (.073)***
Economic egalitarianism		–.202 (.049)***	–.219 (.048)***
Right-wing political orientation		.217 (.043)***	.174 (.043)***
Authoritarianism		.387 (.067)***	.278 (.067)***
Ethnic threat		.453 (.024)***	.377 (.026)***
Ethnic citizenship conception		.336 (.045)***	.194 (.048)***
Anti-racism norms		–.366 (.078)***	–.264 (.079)***
Unemployment		.016 (.218)	–.047 (.225)
Welfare dependency		–.250 (.096)**	–.270 (.097)**
Economic insecurity perceptions		.054 (.039)	.051 (.039)
Age	.007 (.004)	–.002 (.003)	–.004 (.003)
Gender (female)	.171 (.103)	.273 (.087)**	.270 (.087)**
Religiosity	–.024 (.017)	–.049 (.015)**	–.047 (.014)**
Education	–.345 (.029)***	–.082 (.026)**	–.060 (.027)*
R^2 Great Britain	.051	.383	.397
R^2 The Netherlands	.047	.296	.313
N Great Britain	3,516	3,516	3,516
N The Netherlands	1,241	1,241	1,241

*p < .05; **p < .01; ***p < .001.

In the third model, autochthony was added as a predictor to test our first hypothesis. In line with our expectation, a stronger autochthony belief was associated with more welfare chauvinism (B = .565, SE = .073, p < .001), even after controlling for a wide range of other predictors of welfare chauvinism.

Looking at the relative effect size, the standardized coefficients reported in Appendix F show that autochthony was the second strongest predictor of welfare chauvinism (β = .187, SE = .024, p < .001 in Great Britain and β = .177, SE = .024, p < .001 in the Netherlands⁷), after ethnic threat (respectively β = .323, SE = .022, p < .001 and β = .272, SE = .020, p < .001). These results suggest that autochthony is a relevant predictor of welfare chauvinism over and above existing explanations of welfare chauvinism. The general patterns were invariant between British and Dutch majority members, showing the robustness of our results.

Explaining welfare ethnocentrism

First, we examined the mean opposition towards welfare entitlements across the six experimental conditions that referred to different target groups. In each condition, people were asked about the welfare entitlements for one of six target groups (i.e. migrants, ethnic minorities, Muslims, black Britons, white Britons, people). As the superscripts in Table 3 indicate, participants were significantly more opposed to welfare entitlements for migrants than for all other groups. Opposition towards welfare entitlements for ethnic minorities and Muslims was higher than towards black Britons, white Britons, and people in general. Opposition towards welfare entitlements for ethnic minorities and Muslims did not significantly differ and opposition towards welfare entitlements for black Britons, white Britons, and people did not significantly differ either.

Table 3. Descriptive statistics of group-specific opposition towards welfare entitlement, separated by condition.

		Range	Valid n	Mean	SD
Group-specific opposition towards welfare entitlements		1–5
	Migrants		598	3.48^a	1.10
	Ethnic minorities		559	3.14^b	1.13
	Muslims		531	3.05^b	1.11
	Black Britons		535	2.74^c	.92
	White Britons		564	2.80^c	.98
	People		551	2.79^c	.96

*p < .05; **p < .01; ***p < .001. Means with the same superscript are not significantly different from each other at the p < .05 level, based on post-hoc tests in ANOVA’s.

Next, we performed a multigroup structural equation model and performed chi-square difference tests to test for significant differences in the relationship between autochthony and opposition towards welfare entitlements across the six conditions. All coefficients were free to vary across conditions.⁸ Figure 1 shows the unstandardized regression coefficients of autochthony, separated by condition. The coefficients of the control variables were not included in the figure but can be found in Appendix G. Autochthony was significantly related to more opposition towards welfare entitlements for migrants after controlling for all other predictors (B = .241, SE = .061, p < .001), which once again confirms the autochthony-welfare chauvinism relationship. Autochthony was also positively related to opposition towards welfare entitlements for ethnic minorities and Muslims (respectively B = .186, SE = .059, p = .002 and B = .168, SE = .064, p = .008), but not for black Britons (B = .091, SE = .059, p = .124).

Figure 1. Unstandardized regression coefficients for the multigroup model with group-specific opposition towards welfare entitlements as the dependent variable, separated by condition. control variables are included but not reported.

Note: Coefficients with the same superscript are not significantly different from each other at the p < .05 level. The value of low [high] autochthony is the mean of the dependent variable minus [plus] the standard deviation of the independent variable multiplied by the regression coefficient.

The coefficients at first glance seem to be in line with H2, as they suggest that autochthony was able to explain opposition towards welfare entitlements for migrants best, followed by ethnic minorities and Muslims, followed by black Britons. However, most of these coefficients did not significantly differ from each other. The only significant difference shows that autochthony was more related to opposition towards welfare entitlements for migrants than for black Britons (TRd = 4.251 (1), p = .039). H2 was therefore partially accepted.

Although not the focus of our study, we did find – as one might expect – that autochthony was not significantly related to opposition towards welfare entitlements for white Britons. Furthermore, autochthony was significantly more related to opposition towards welfare entitlements for migrants (TRd = 10.053 (1), p = .002) and ethnic minorities (TRd = 4.241 (1), p = .039), compared to white Britons. However, Figure 1 shows no significant difference in how autochthony was associated with opposition towards welfare entitlements for black Britons and white Britons (TRd = .620 (1), p = .431). Surprisingly, autochthony was also positively related to opposition towards welfare entitlements for people as a general category (B = .152, SE = .074, p = .039).

Discussion

As a lack of solidarity among co-citizens puts a strain on the viability of the welfare state, it is important to understand why people oppose welfare entitlements for newcomers and established ethnic minorities (Kymlicka and Banting 2006). Our research is the first to show that autochthony can help to explain welfare chauvinism, and to some extent, welfare ethnocentrism. People who believe in the entitlements of firstcomers are more welfare chauvinistic towards migrants and more welfare ethnocentric towards ethnic minorities and Muslims. These relationships exist even when taking into account a great range of existing explanations. Moreover, the relationship between autochthony and welfare chauvinism is found to be robust across British and Dutch natives.

One could argue that welfare chauvinism and welfare ethnocentrism are rather hard to tackle when driven by autochthony. The principle of autochthony might be rather engrained in people’s minds as it is based on the intuitive psychology of possession (Rochat 2014) and is used to establish ownership in a range of contexts. This might make it very hard to debunk autochthony as a meaningful principle. Also, autochthony can be argued to be a strict principle as first-occupant status cannot be earned by later comers in any way. Migrants will never become firstcomers.

However, there are three reasons why our study shows a more nuanced picture. First, we find that only a very small portion of Dutch and British natives support welfare chauvinism in the strict sense that immigrants should never obtain rights to social benefits (Reeskens and Van Oorschot 2012). Even those who strongly believe in autochthony are still generally in favour of granting conditional social benefits to migrants.

Second, autochthony is not the only predictor of welfare chauvinism. All of our indicators of ideology and ethnic attitudes were also related to welfare chauvinism. In contrast, unemployment and perceived economic insecurity were not related to welfare chauvinism. Contrary to what we expected, people who are dependent on welfare were less welfare chauvinistic, which suggests that empathy feelings with fellow welfare recipients plays a more pronounced role in explaining welfare chauvinism than self-interest considerations (Van Oorschot 2008).

Third, our experimental findings showed that native Britons who endorse autochthony do not bluntly exclude every minority group with a migration origin from welfare entitlements. Although autochthony was associated with more welfare ethnocentrism towards ethnic minorities and Muslims, it was unrelated to welfare ethnocentrism towards black Britons. The latter finding is in line with our argument deduced from self-categorization theory (Turner et al. 1987), that white Britons could perceive black Britons as part of a superordinate category of Britons, which would lead to more solidarity and a lower inclination to exclude this group from welfare entitlements based on claims of autochthony. Minority groups might at a certain point be seen as sufficiently established to be included in a superordinate category of the national population and therefore, to be granted welfare entitlements. We note that our results might be partly due to the priming effect of the label “black Britons” used in our experiment. Using the label “blacks” for this group or otherwise emphasizing a common identity for other groups (e.g. “British Muslims”) might yield different results (see Verkuyten and Thijs 2010).

Related to this, we found that native Britons were surprisingly open to granting welfare entitlements to black Britons. They opposed welfare entitlements for black Britons as much as they opposed welfare entitlements for white Britons and people in general, and less than they opposed welfare entitlements for migrants, ethnic minorities and Muslims. Next to a superordinate social categorization, there are at least two additional possible explanations. First, most black Britons descend from the Caribbean and their closeness to British culture, partly resulting from Britain’s Caribbean colonial history and their predominantly Christian denomination, can increase solidarity (Van Oorschot 2008). Second, native British might refrain from speaking out explicitly against black Britons to prevent feeling overtly racist.⁹

Autochthony was significantly more strongly related to opposition towards welfare entitlements for migrants than for black Britons, which suggests that those who adhere to autochthony distinguish between minority groups. However, we found no other significant differences in the relationship between autochthony and opposition towards welfare entitlements for the four different target groups (migrants, ethnic minorities, Muslims, and black Britons). This could be partly due to the fact that the labels used in our experiment are open for interpretation. Different minority groups can overlap in people’s minds (see Braun et al. 2019). Still, the strengths of the relationships were in the expected direction.

Although not the focus of our study, another notable finding is that native Britons who endorsed autochthony were more opposed to welfare entitlements for people in general (and not for white Britons). This suggests that when native Britons think about granting welfare entitlements for people in general, they might overall not have native welfare recipients in mind but rather think of people with a migration status in particular, as they perceive welfare policies as mainly benefiting this group. A similar pattern is found in the United States where discussions about welfare benefits have focused on race and white Americans’ attitudes towards welfare policies strongly depend on their attitudes towards African Americans (Gilens 1999). This is an indication that ideologies such as autochthony can not only help to explain possible eroding support for group-targeted welfare policies, but for the welfare state in general. Public support for an extensive welfare system might erode the more migrants and minorities are overrepresented among welfare dependants or are strongly perceived as such.

Our study has several limitations. We are unable to make claims about causality and cannot rule out the possibility of bidirectional associations, as autochthony was not repeatedly measured in different waves. Although we argued that autochthony is a general underlying belief that translates into more specific attitudes (Verkuyten and Martinovic 2017), it is possible that autochthony is also used to justify pre-existing welfare chauvinism (Crandall and Eshleman 2003). Autochthony can make it seem like welfare chauvinist ideas are not unjust or discriminatory but are self-evident consequences of the intuitive principle of ownership for firstcomers. However, our assumption that autochthony is an underlying ideological belief that helps to explain welfare chauvinism and welfare ethnocentrism is theoretically derived and based on empirical research. We examined autochthony as a general ideological belief (which did not refer to any context) to predict welfare chauvinist attitudes in a specific context. We believe that a reverse causal order from welfare chauvinism in a specific context to a general ideological belief is less likely. Moreover, experimental evidence shows that autochthony precedes ownership beliefs (Friedman and Neary 2008; Verkuyten, Sierksma, and Martinovic 2015) that in turn cause different reactions (Nijs, Martinovic, and Verkuyten 2022). However, none of these studies has examined welfare chauvinism or welfare ethnocentrism and although the assumed causal structure is likely to underlie our correlational findings, more experimental work is needed to unpack causality.

We measured welfare chauvinism by asking participants how long immigrants should work and pay taxes before they are entitled to welfare benefits. Part of our argument was based on the idea that people who endorse autochthony can oppose immigrants’ welfare entitlements based on first occupancy and independent of immigrants’ contribution. Hence, one might expect an even stronger autochthony-welfare chauvinism relationship if welfare chauvinism is measured in terms of opposing welfare entitlements for migrants in general. An advantage of our applied measure was that we could show that welfare chauvinism in the soft sense is very widespread among the general public: most people view migrants’ right to social benefits as conditional (Reeskens and Van Oorschot 2012).

There are several possible directions for future research. First, autochthony is not the only principle to determine collective ownership. “We made the country as it is today” or “the country made us who we are”, referred to as the investment and formative principle (Verkuyten and Martinovic 2017), can also be used as arguments for country ownership, and therefore, as arguments to determine who is entitled to welfare support. The interplay between autochthony and other principles can shed more light on the importance of ownership claims for welfare chauvinism. Secondly, future research could also examine the interplay between autochthony and other predictors of welfare chauvinism in more detail, to further position the role of autochthony in relation to other explanations. For example, ethnic threat can function as an independent predictor as in the current study, but it can potentially also moderate the relationship between autochthony and welfare chauvinism (Hasbun Lopez et al. 2019; Martinovic and Verkuyten 2013). Similarly, ethnic citizenship conception and autochthony can be independently related to welfare chauvinism, as ethnic citizenship conception is about who is a true citizen, while autochthony is about who is a rightful owner. But one might also argue that ethnic citizenship mediates the effect of autochthony or vice versa, as tested by Verkuyten and Martinovic (2015) using correlational data. Longitudinal designs can help infer how autochthony interplays with other predictors of welfare chauvinism. Thirdly, testing the robustness of our results in other contexts is another possible future direction. For example, autochthony might have a different effect in settler societies, in which the majority group is not the autochthonous population (Nooitgedagt et al. 2021). Finally, to understand the challenges of the welfare state in multicultural societies, it is crucial to understand welfare state attitudes and welfare chauvinism among minority groups (Galle 2019). It is worth investigating how a minority status influences autochthony and its relationship to welfare chauvinism.

In closing, our study showed that autochthony can help explain welfare chauvinism and welfare ethnocentrism among Western Europeans. The argument “we were here first” is not only relevant in people’s day-to-day lives, but also sheds light on public support for group-targeted welfare policies and the perceived legitimacy and viability of the welfare state in modern Western societies.

Acknowledgements

The authors thank Maykel Verkuyten and reviewers at the ICS Forum Day for their very helpful feedback and suggestions on drafts of this article.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This research has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no 715842) awarded to the second author. The data collection was supported by the Open Research Area in Europe for the Social Sciences (ORA) joint funding scheme and financed by the Netherlands Organisation for Scientific Research (NWO) [464-11-060] and the Economic and Social Research Council (ESRC) [ES/K006126/1].

Notes

1 This article is based on the first author’s doctoral dissertation (Nijs 2022).

2 The data allow us to explore this on the British sample only.

3 We included ethnic citizenship conception as previous research argued that the kind of national identification people hold is a more important predictor of welfare chauvinism than the extent to which they identify with the national group (Wright and Reeskens 2013). However, as a robustness check, we also included two measures of the extent of national identification (chauvinism and national attachment) as control variables. This did not substantially alter our results, as shown in Appendix A.

4 We could not endogenize the dichotomous control variables unemployment and gender when testing H1, as this would pose numerical problems for Mplus. Unemployment and gender were therefore uncorrelated with all other predictors. The two variables were endogenized when testing H2.

5 A metric model with free intercepts fitted significantly better than a scalar model (TRd = 187.457 (6), p < .001). However, as the Chi-square difference test is sensitive to sample size (Putnick and Bornstein 2016) and all other fit indices suggest a good fit of the scalar model, we did not continue with the full metric invariant model.

6 See Appendix E for the results of the model in which all coefficients were free to vary across countries.

7 Although the coefficients were constrained to be equal across country samples, the standardized coefficients slightly differed across country samples, because they were standardized in the group.

8 Here we take into account that predictors may be differently related to opposition towards welfare entitlements, depending on the specific target group.

9 Indeed, as displayed in Appendix G, anti-racism norms were strongly related to opposition towards welfare entitlements for black Britons but not related to opposition towards welfare entitlements for migrants. Anti-racism norms were also, but with a lower significance level, related to opposition towards welfare entitlements for ethnic minorities, Muslims, and white Britons.