Introduction

Euroscepticism is on the rise nearly everywhere in Europe. In the United Kingdom (UK), Eurosceptic and populist forces conducted a successful referendum campaign that precipitated Britain’s impending withdrawal from the European Union (EU). In several Eastern European countries, including Hungary and Poland, nationalistic conservative parties with an openly Eurosceptic stance command majorities in parliament and form the government. In Western Europe, Eurosceptic parties on the radical right such as the Front National or the Austrian Freedom Party are winning upwards of 20% of the vote in parliamentary and presidential elections. What most, if not all, of these Eurosceptic forces have in common is that they hold the European Union responsible for uncontrolled immigration, which, according to them, is responsible for the many problems Europe currently faces.

To name a few examples: Nigel Farage, former leader of the UK Independence Party, suggested that the “demand for the rapid implementation of a common EU migration and asylum policy […] would be wholly unacceptable to a United Kingdom that already has levels of immigration that are too high, and as Isis have previously threatened, could lead to half a million Islamic extremists coming to our countries and posing a direct threat to our civilisation” (The Guardian 2015). In a similar vein, the Hungarian Prime Minster Viktor Orbán declared: “For us migration is not a solution but a problem… not a medicine but a poison, we don’t need it and won’t swallow it, every single migrant poses a public security and terror risk. This is why there is no need for a common European migration policy: whoever needs migrants can take them, but don’t force them on us” (The Guardian 2016).

While right-wing and populist leaders frequently blame the EU for uncontrolled immigration and its alleged negative consequences such as crime, economic degradation for natives or a decline of the dominant culture, the scientific literature has (also) increasingly discussed the link between immigration levels and attitudes toward immigration, on the one hand, and Euroscepticism, on the other. They investigate the question of whether negative assessments of immigration trigger higher levels of Eurosceptic attitudes. In this meta-analysis, we offer a systematic review of recent quantitative research evaluating the immigration-Euroscepticism nexus. To that end, we use Google Scholar and search for keywords such as “migration”, “immigration”, “quantitative”, “Euroscepticism“, “Eurosceptic attitudes”. Restricting our search to the period from 2005 to July 2017, we find 21 peer reviewed articles corresponding to our criteria.Footnote 1 They include 84 regression models which directly measure the link between immigration and Euroscepticism. The majority of articles, 16 in total, use attitudes toward immigration or a specific type of immigrants as the independent variable and find strong support (i.e. in above 70% of the cases) for the thesis that negative attitudes toward immigrants lead to increased Euroscepticism. For the five articles that use structural data, or the number of (a certain type of) immigrants, the link with increased Euroscepticism is much weaker and less clear.

This article continues as follows: First, we briefly discuss our research strategy and our study sample. Second, we introduce the main theories—the ethnic composition theory and the contact theory—which link immigration levels and anti-immigrant attitudes to Euroscepticism. Third, we define our dependent and independent variable(s). Fourth, we present and discuss the findings of our meta-analysis. Finally, we draw conclusions from our analysis and provide some avenues for future research.

Research strategy

We aim to provide an overview of previous research evaluating the link between (attitudes toward) immigration and Euroscepticism. To do so, we engage in “an analysis of analyses” (Imbeau et al. 2001:3), that is, we “analyze test results from previous studies through quantitative methods and summarize the findings” (Smets and van Ham 2013: 3). We are working with the findings of published studies, rather than with the primary material that these studies are based upon. Taking every regression model in turn, we identify for each regression weight the influence of any migration proxy present in the model on Euroscepticism. In doing so, we distinguish between attitudes toward immigrants and “hard data” as measured by the percentage or the number of immigrants/or a certain type of immigrants. Using this distinction, we identify 16 articles including 70 regression models which focus on attitudes toward immigration, and 5 articles with 14 regression models, which focus on the influence of structural immigration data on Euroscepticism. Following common practices in meta-analysis, we create three categories: (1) “significant according to theoretical expectations”, (2) “non-significant” and (3) “significant contrary to the literature’s predictions” (Geys 2006, 640). We label these three cases “success”, “no link” and “anomaly”, respectively.

Theory

In the 21 articles that form the corpus for our research the main theory implicitly or explicitly used to explain the link between immigration, whether real or perceived, and Euroscepticism is the ethnic competition hypothesis or ethnic threat theory (cf. Blalock 1967; Bobo 1988; Olzak 1994). The theory starts with the assumption from “social identity theory” (Tajfel 1981; Tajfel and Turner 1979) that individuals tend to categorise themselves and others into groups and favour their own group (in-group)—via processes of social identification—over others (out-group)—via processes of social contra-identification. The out-group is viewed with a negative bias, either through out-group rejection or in-group favouritism (cf. De Vreese and Boomgaarden 2005; McLaren 2002). Research shows that this is especially true for those that hold anti-immigration attitudes (Sniderman et al. 2002). These people view themselves as “the native majority”, who should have the hegemony over economic, social or cultural resources (Semyonov et al. 2006). If these resources are scarce, they tend to feel threatened by the out-group and may thus seek to defend their privileged position against immigrants, i.e. the intruders and members of the out-group. The perceived or real threats posed by immigrants can be economic: Natives might fear for their jobs, or their social security benefits. The threat can also be linked to challenges at the level of (national) identity and lead to the rejection of foreign lifestyles and cultures (McLaren 2002; De Vreese et al. 2008). Due to such threat perceptions, majority groups will react with exclusionary measures (Olzak 1994: 35).

When it comes to immigration levels (rather than anti-immigration attitudes), the literature also broadly builds on this theoretical strand. Higher immigration levels (under most circumstances) trigger competition among groups and increase the salience (and thus polarisation) of the immigration question in a polity (cf. Azrout et al. 2013b). This in turn fosters perceived out-group threat and social contra-indication (Lubbers and Scheepers 2007; Schlueter and Wagner 2008). Toshkov and Kortenska (2015: 913) suggest that the relationship between immigration levels and anti-immigrant dispositions depends on several factors, including characteristics of the specific out-group and in-group, and the nature of the contact.

Given that the EU—through the Schengen and Dublin agreements as well as important directives e.g. on asylum qualification, procedures and reception standards (Zaun 2017)—is a major stake-holder in immigration policy in Europe, the ethnic threat thesis is seemingly the ideal theoretical construct to explain the effect of either anti-immigration attitudes or structural immigration on Euroscepticism. According to proponents of the theory, both economic and (especially) cultural threats trigger negative sentiments toward European integration because the EU is identified with free migration and cultural diversity. De Vreese and Boomgaarden (2005: 64) argue that EU integration “brings together people from different countries, regions and cultures, and arguably with different religions and ethnicities […]. This fuels negative assessments of these groups, and therefore people holding negative attitudes towards immigrants are more likely to reject the idea of further European integration”.

McLaren (2002: 54) suggests that Euroscepticism is importantly linked to identity politics and that the nation state has become the primary in-group of EU citizens. She argues that “the uniqueness of national cultures and the exclusive control over the resources of the nation-state are […] seen as being under threat by the EU”. Therefore, “the same people who fear […] changes from minority groups living in the country […] are very likely to fear similar changes resulting from the process of European integration”. In other words, to the degree that we assume that in-group bias leads to high appreciation of national traditions, we may deduce that these national sentiments direct resistance to policies leaning toward “integrated nation-states” or further integration at the expense of national sovereignty (De Master and Le Roy 2000; cf. Lubbers and Scheepers 2007: 645).

This applies even more so given that “[r]eligious heterogeneity is a very visible by-product of European integration” (Hobolt et al. 2011: 363). This heterogeneity is most visible in the “religious divide” between Christians and Muslims (Boomgaarden and Freire 2009: 1240), which is all the more relevant considering that Muslim immigrants have made up the largest share of recent immigration waves. In the context of these cultural encounters, European non-Muslims often portray Muslims as “others”, as outsiders who they frequently associate with “terrorism and violence”, “backwardness”, and “inferiority” (Azrout et al. 2013a: 6, 7).

The counterpart of the ethnic threat hypothesis, the contact theory, does not really provide any strong theoretical counterweight to this dominant paradigm. In its simplest form, the contact theory postulates that increased contact between different ethnic groups could undermine mutual stereotypes and thus reduce cultural tensions (Allport 1954). Under favourable circumstances increased levels of immigration could increase native residents’ chances of coming into contact with people from different ethnic backgrounds, which, in turn, could reduce prejudices and enhance understanding for immigrants (van Klingeren et al. 2013). The link between increased understanding for immigrants and support for European integration is not (adequately) specified in the literature. Some authors seem to reverse the logic from the above literature between anti-immigration attitudes and (decreasing) public support for the EU and simply claim the opposite for greater acceptance of immigrants (cf. van Klingeren et al. 2013). In addition, it has been suggested that more transnational interconnectedness is conducive to a cosmopolitan worldview, which tends to diminish Euroscepticism (cf. Kuhn 2011).

Contact theorists themselves affirm that the contact hypothesis can only work under certain circumstances that is, e.g. “prejudice … may be reduced by equal status contact between majority and minority groups in the pursuit of common goals. The effect is greatly enhanced if this contact is sanctioned by institutional support … that leads to the perception of common interests … between members of the two groups” (Allport 1954: 281). In Europe, these preconditions largely do not apply. The political situation is characterised in many countries/regions by strong anti-immigrant parties, a latent hostility toward Muslims, a fear of Islamic terror and an upsurge of nationalism. In such an environment, there are few or no shared common goals between (Muslim) immigrants and the majority native population. There is little intergroup cooperation, no equal status between the groups, and little action by the authorities to support positive interactions between natives and (Muslim) newcomers (Azrout et al. 2013b: 482). Amir (1976) as well as Nagel (1995) highlight that contact under such unfavourable conditions may increase prejudice and intergroup tension. According to Paolini et al. (2010), negative contact makes individuals more aware of their respective group memberships (i.e. it causes high category salience), which, in turn, tends to make them more hostile toward immigrants. With negative attitudes toward immigrants we are back at the stand of literature and the logic that links this with Euroscepticism (cf. first part of this section).

Thus, from a theoretical perspective, there is rather unanimity that both anti-immigrant attitudes and high immigration levels are likely to trigger higher levels of negative attitudes toward European integration. Our “analysis of analyses” will show to what degree these theoretical expectations apply. Before discussing the empirical findings of previous published research however, we will briefly present how immigration attitudes and immigration levels as well as Euroscepticism are operationalised.

Definition of variables

Independent variable: immigration/immigration attitudes

The major distinction in the literature is between attitudinal factors toward immigration and “hard data” on the number of immigrants. The first operationalisation focusing on attitudes makes up the majority of our sample of studies; 16 out of the 21 articles and 70 models. Within this broad category there is wide variation. The most common indicator is general attitudes toward immigration (normally measured on a 0 to 10 scale ranging from very negative to very positive) (see Kuhn 2011; Tillman 2013; Yavçan 2013). This indicator is used in more than 60% of cases. Aside from this general measure, some articles focus on attitudes toward a certain type of immigrants, such as Muslim immigrants (e.g. Lubbers 2008), immigrants from EU member states (Lubbers and Scheepers 2007) and immigrants from non-EU member states (Vasilopoulou 2016). Another set of articles tries to specifically capture the type of threat immigrants might pose, focusing on ethnic motives (e.g. Barbulescu and Beaudonnet 2014) or economic motives (e.g. Garry and Tilley 2009). A third array of studies particularly focuses on assessments of EU migration policy as a trigger/driver for Eurosceptic attitudes (Balestrini 2011; Domanov 2013) (see also Table 2).

In contrast to the common operationalisation of immigration by attitudes, the second operationalisation using structural data or the percentage of immigrants is much less frequent, employed by only 5 articles and 14 models. These operationalisations include the proportion of immigrants from non-EU member states, the percentage of immigrants from CEE countries (e.g. Toshkov and Kortenska 2015) and the number/percentage of new immigrants (Goodwin and Heath 2016) (see Table 4).

Dependent variable: euroscepticism

The term “Euroscepticism” is often considered “a generic, catch-all term encapsulating a disparate bundle of attitudes opposed to European integration in general and opposition to the EU in particular” (Szczerbiak and Taggart 2017: 12). Accordingly, Hooghe and Marks (2007: 120) suggest “the term expresses doubt or disbelief in Europe and European integration in general”. This broad usage of Euroscepticism is also common in the articles we review.

Most of the studies investigate general attitudes toward the EU (e.g. Balestrini 2011; Kuhn 2011; Van Klingeren et al. 2013), based on the question: “Generally speaking, do you think that (your country’s) membership in the European Community (Common Market) is a good thing (0), neither good nor bad (1), or a bad thing (2)”, or a variant of this question.

A smaller group of researchers adopt a slightly different perspective. While still investigating evaluations of the EU, they use questions which focus explicitly on (further) EU integration (De Vreese and Boomgaarden 2005; Boomgaarden 2009; Yavçan 2013; Toshkov and Kortenska 2015; van Elsas et al. 2016). Finally, a few scholars use specific types of Euroscepticism. For example, Lubbers and Scheepers (2007) use “political euro-scepticism” as their dependent variable and investigate it through questions regarding the appropriate level of decision-making (regional, national, EU, international) on policies in different areas (e.g. immigration, crime, social welfare). Luedtke (2005) investigates attitudes toward EU control over just one policy area (immigration policy). Despite these nuances, we consider all these operationalisations of Euroscepticism to be sufficiently similar to cluster them together. Considering that all articles capture Euroscepticism either generally (in 80 of the cases) or in a specific policy field (in 20%), the more specific operationalisations should capture part of the general anti-EU attitudes and should thus be highly correlated with these general measures.

Findings and discussion

Our analysis of analyses highlights that there is strong support in favour of the relationship between increased anti-immigration attitudes and more Euroscepticism. Nearly three out of four regression models confirm this relationship, which appears solid across various operationalisations of anti-immigrant attitudes. Whether researchers use questions measuring general attitudes toward immigration, questions gauging the economic or cultural threat posed by immigration or more specific questions that tap into individuals’ opinions about EU migration policy, all studies find that individuals who are more critical of immigration or a specific aspect of immigration are more likely to be critical of European integration. The exception is attitudes toward EU immigrants. In the one study that uses this indicator (i.e. Lubbers and Scheepers 2007), critical attitudes toward migrants from other EU countries do not trigger increased Euroscepticism in three out of four cases (see Tables 1 and 2).

Table 1 The effect of (anti-)immigration attitudes on Euroscepticism
Full size table
Table 2 The effect of different types of anti-immigration attitudes on Euroscepticism
Full size table

When it comes to the link between structural data on immigration and Euroscepticism the relationship is less clear. In fact, in nearly two-thirds of the cases, a high or increased percentage of immigrants does not lead to a rejection of European integration. Rather in most cases, there is no relationship between the number of immigrants in an area and evaluations of the EU (Tables 3 and 4).

Table 3 The effect of structural immigration data on Euroscepticism
Full size table
Table 4 The effect of different operationalisation of immigration levels on Euroscepticism
Full size table

This meta-analysis’ main finding is that negative attitudes toward immigration trigger increased opposition to European integration. In contrast, high percentages of immigrants are unrelated to individuals’ assessment of the European Union. This finding warrants some further discussion. The literature discussing the link between immigration and anti-immigrant attitudes might help us make some sense of this result. This literature increasingly comes to the conclusion that there is no relationship between real levels of immigrants in a geographical area and perceived levels (Hainmueller and Hopkins 2014; Pottie-Sherman and Wilkes 2017, Stockemer 2016). In other words, there is now some solid scholarly evidence that respondents do not know about the real number of immigrants in their municipality, region or country. In other words, the percentage of immigrants who they think lives in their neighbourhood, city, region or country is uncorrelated with the actual number, and in many instances overestimated (Stupi et al. 2016). What counts is individuals’ assessment of immigration. If European citizens think that there are (too) many immigrants and/or if they assess immigration’s general influence, and/or its economic or social impact negatively, they are more likely to have anti-immigrant attitudes.

Conclusions

This short meta-analysis provides a timely addition to the literature examining the links between structural data on immigration and immigration attitudes on Euroscepticism. Studying recent research, we find that there is rather strong support for the proposition that negative opinions about immigrants trigger negative opinions about Europe. This evidence is solid across operationalisations of immigration attitudes. However, at the same time we also find that higher percentages of immigrants do not tend to make individuals more Eurosceptic.

This finding, which is based on pre-refugee crisis data, becomes all the more interesting in the (post-)refugee crisis setting. Once the survey data become available, future research could look at how the so-called European refugee crisis from 2014 to 2016 shaped Europeans’ attitudes toward immigration and the EU. Did the refugee crisis trigger more negative assessments of migration, which then could have led to increased Euroscepticism? In addition, does this meta-finding concerning the non-relationship between structural data on immigration and Euroscepticism also hold in the crisis context? Most likely so: it is unlikely that the 2.4 million migrants that entered the European Union in 2015, which “only” make up around 0.5% of the EU population that have turned Europeans more hostile to immigration (Eurostat 2016). Rather it is more likely that the pictures of refugee boat after refugee boat arriving at the Italian coasts, the pictures of segregated neighbourhoods, the pictures of terrorist attacks, some of which have been conducted by recent immigrants from the Middle East and North Africa, along with a securitisation of migration in public discourses, that make individuals wary not only of immigration, but also of the European integration project, which has been held responsible. This is exacerbated by the fact that the European Union seems unable to comprehensively resolve the refugee and asylum crisis (Niemann and Zaun 2018), which is a minor crisis, if we talk in real numbers, but a major crisis in terms of perceptions.

An additional question that would merit (further) research is the following: is there any connection in the assessment of immigrants and citizens’ appreciation of the EU between countries that did not take any migrants such as Hungary, but whose government employed a very critical and anti-immigrant rhetoric from the beginning to the end of the crisis, and countries such as Germany and Sweden, which, at least in 2014 and 2015, were accommodating of foreigners both rhetorically and on the ground?

One weakness of the current research agenda on the immigration-Euroscepticism nexus is that findings largely remain at the level of correlation, while the causal mechanisms at work remain to some extent subject to speculation. This points to the need for greater efforts at theory-building (and refinement). The current theories that are principally relied upon—such as the competition hypothesis or ethnic threat theory—tend to be routinely invoked without sufficiently engaging with specific causal tenets/mechanisms. Moreover, this theoretical strand seems under-specified. The causal link between immigration (attitudes) and Euroscepticism is rather weak. Competition hypothesis/ethnic threat theory does not sufficiently connect immigration (attitudes) with Euroscepticism at a causal level, largely because this paradigm makes no assumptions about European integration as such. Closely related, additional/complementary explanations have been (largely) ignored, both in the relevant discussions and analyses of this topic.

For example, insights from the securitisation literature (Waever 1995; Buzan et al. 1998)—postulating that migration is increasingly framed as a security threat (Bigo 2002), which has also been found in the EU context where migration has been more and more linked to terrorism and crime (Huysmans 2005)—could shed additional light at the immigration-Euroscepticism nexus: securitisation closely connects the majority in-group with the minority out-group in the direction suggested by research findings so far. Framing immigration as a security threat may thus provide a complementary (and deeper) explanation for out-group rejection on the part of the native population and also further connects such developments to Euroscepticism, as the EU tends to be an important part of the (negative) framing.

While future research should tackle these issues, we provided a much-needed summary of the current stage of literature measuring the link between immigration (attitudes) and Euroscepticism.