Sage Journals: Discover world-class research

Abstract

The European Union has progressively communitarised its migration policy. The formation of public support for this integration of a core state power presents an intricate puzzle. On the one hand, immigration is part and parcel of the conflict around the opening and closing of nation states, and thereby mobilises nativist views and Euroscepticism. On the other hand, the European Union may serve as a shield against external threats such as uncontrolled immigration. This article sheds light on this conundrum by examining how refugee arrivals affect public support for a common European migration policy across 28 European Union member states between 1992 and 2021. The results lend support to a post-functionalist logic of an identitarian backlash against integration and a collective action logic of instrumental solidarity in line with national interests.

Keywords

European Union migration policy refugees responsibility-sharing solidarity

Introduction

Immigration has become a major challenge to the development and stability of the European Union (EU). Migration has increasingly become politicised in many EU member states, and leading politicians have argued that the future of the EU hinges on solving this critical issue (Rankin and Oltermann, 2018). While the EU has gained increasing competences in immigration control over time (for a detailed overview see Boswell and Geddes, 2011), it continues to struggle with developing an effective policy response that would establish responsibility-sharing among its member states. An effective migration policy is critical as the EU’s legitimacy rests on the combination of internal freedoms, namely open borders within Europe, and restrictive migration control at the common external border (Lutz and Karstens, 2021). Mass support for a common European migration policy is thus crucial to preclude disintegration dynamics. Indeed, domestic public opinion can act as a powerful constraint on European policymaking because national governments have strong incentives to be responsive to their citizenry when they negotiate policies at the European level (Mauro and Memoli, 2021). The recent years of high immigration into Europe have fostered Europeans’ perceptions that immigration is out of control and that the EU member states are deeply divided on the issue, which has hindered an effective policy response (Connor, 2018; Harteveld et al., 2018; Lutz and Karstens, 2021). Nevertheless, we only have a limited understanding of how these circumstances shape the public’s support for a common European migration policy. When do citizens prefer to have immigration control at the European level rather than at the national level?

The formation of public support for the communitarisation of immigration policies presents us with a theoretical puzzle. On the one hand, immigration issues form part of the political conflict around the opening and closing of nation states, which mobilises nativist views and Euroscepticism (Kriesi et al., 2012; Toshkov and Kortenska, 2015). Some observers believe that the EU’s ineffective response to large-scale immigration and the electoral rise of anti-immigration parties that demand their countries’ return to national border controls have thrown the Union into a profound constitutional crisis (Byrne et al., 2020). On the other hand, other narratives have portrayed the EU as a shield against external threats such as uncontrolled immigration (Conti et al., 2019; Isernia and Cotta, 2016). According to his perspective, irregular immigration from non-European countries should call for more European competences in immigration control. In times of numerous refugee arrivals, we have observed both more expressions of Euroscepticism in domestic politics and calls for greater European solidarity by member states that feel overburdened. Hence, there are competing arguments and empirical observations on whether immigration mobilises or undermines support for a common European migration policy. This article sheds light on this matter by examining how refugee arrivals affect the public support for a common European migration policy.

The extensive literature on public support for (or opposition to) European integration (see Hobolt and De Vries, 2016 for an overview) pays little attention to migration policy despite good reasons for why EU support should vary across policy fields. The few existing studies on public preferences for an EU migration policy seek to explain public support primarily through individual-level factors based on utility and identity (Conti et al., 2020; Erisen et al., 2020), but do not pay particular attention to the way contextual factors, such as migration pressure and refugee arrivals, shape these preferences. Two exceptions are Conti et al. (2019), who find a small positive association between the number of resident refugees in a country and popular support for an EU migration policy, and Heizmann and Ziller (2020), whose comparative study of the 2015 refugee crisis shows that higher numbers of asylum-seekers are associated with higher support for their better (re)distribution among EU member states. Nevertheless, these cross-sectional studies fail to capture the dynamics of public support for EU competences on immigration in a context of asymmetric and fluctuating migration movements. The broader literature on immigration attitudes indicates that citizens’ attitudes and preferences on the matter are indeed affected by the number of immigrants arriving in a country (e.g. Hangartner et al., 2019; Gorodzeisky, 2022). These studies conclude that citizens’ most common response is to feel threatened and to embrace more critical attitudes to immigration and higher support for more restrictive policies. Building on these studies, I ask how member states’ fluctuating immigration numbers shape public support for an EU migration policy. Understanding the underlying dynamics of public opinion allows us to gain important insights into the legitimacy of core European institutions and the prospects for a common European response in times of crisis.

Explaining public support for an EU migration policy

The European migration policy has evolved gradually as a result of previous integration steps. The creation of a common European market that guarantees the free movement of goods, services and capital, as well as people, and the abolition of internal border controls within the Schengen area have created spill-over effects to the field of migration policy (Oltmer, 2022). Such a deepening of European integration in migration policy is a prerequisite for solving the collective action problem of a borderless Europe and establishing effective governance of migration into and within the EU (Eilstrup-Sangiovanni, 2021). In an open Europe where border controls between member states have been abolished and people enjoy freedom of movement, national governments are constrained in their ability to address irregular immigration and fight against transnational threats, such as organised crime or terrorism (Helbling and Meierrieks, 2020; Semyonov et al., 2008). Not only are member states exposed to the risks of immigration, but they are also vulnerable to the migration policy choices of their counterparts. To preserve the integrity of the free movement area and the wider European single market, member states have a strong interest to cooperate and establish a joint European migration policy. For this reason, the EU’s migration policy is mainly concerned with irregular immigration from non-European countries and focuses on strengthening its competences in border control as well as responsibility-sharing in the processing of asylum claims (Boswell and Geddes, 2011). Nevertheless, EU member states are reluctant to surrender national sovereignty in this core area of statehood that is linked to national identity and security. Thereby, public opinion has grown into an increasingly important constraint on integration as the domestic politicisation of transnational policy-issues has replaced the ‘permissive consensus’ for further integration with a ‘constraining dissensus’ (Hooghe and Marks, 2009). In addition to domestic political conflicts, member states also experience significant asymmetries in their migration pressure from outside the EU. Hence, the unsolicited immigration into Europe creates tensions among member states that struggle to establish effective cooperation and responsibility-sharing.

In the following paragraphs, I theorise how this context of asymmetric and fluctuating immigration shapes public support for a common European migration policy, instead of an exclusively national one. Public support for an EU migration policy can be understood as a policy-specific preference for the extent to which more (or less) European integration is desirable (e.g. De Wilde and Trenz, 2012). I build on theories of European integration as well as theories of preference formation from the literature on public opinion to derive two theoretical perspectives on the formation of public support for an EU migration policy in the context of migration movements: an identity-driven theory of national closure based on an anti-integrationist backlash and an interest-driven theory of instrumental support for communitarisation based on the potential benefits of European responsibility-sharing.

The first theoretical perspective builds on the idea that immigration causes an identitarian backlash among native citizens, thereby increasing public demands for national closure. The growing presence of immigrants raises the issue’s salience in the public discourse and in citizens’ minds, shaping their political responses (Dennison and Geddes, 2019). The most common of these responses is to feel threatened and to resort to more exclusionary attitudes: immigrants often form a salient out-group that evokes group-centric sentiments of threat among natives who fear that immigrants’ presence would burden them with negative consequences (see Hangartner et al., 2019 for an overview). Studies have identified such anti-immigration sentiments as a key predictor of citizens’ opposition to European integration (De Vreese and Boomgaarden, 2005; Stockemer et al., 2020). Similarly, Harteveld et al. (2018) find that the higher salience of non-European immigration generates Euroscepticism and Lutz (2021) shows that EU citizens withdraw their support for European free movement of people when they associate it with undesired immigration. The 2016 Brexit referendum in the United Kingdom is a case in point for the people who link the EU to increased immigration and demanding more national policy competences to ’take back control’ (Goodwin and Milazzo, 2017).

These findings are in line with the post-functionalist theory of European integration, which argues that national identity and a sense of community are important drivers of integration support (Hooghe and Marks, 2009, 2018). The domestic politicisation of European integration is expected to mainly mobilise Eurosceptic citizens who feel threatened in their national identity. In this identity-based explanation of integration support, citizens are expected to become more sceptical of integration in times of crisis and especially in the areas of core state powers, such as immigration control (Hooghe and Marks, 2005). The issue of immigration constitutes an ideal opportunity for political actors to mobilise voters around nativist sentiments and nationalist identity politics (Van Der Brug and Harteveld, 2021). This is even more likely when immigrants come from non-European countries: such immigration is associated with high levels of perceived symbolic threat which activates in-group favouritism (Czymara and Schmidt-Catran, 2017). The more citizens feel that immigration threatens their national identity, the more defensive they shall become of their national political community and the less appetite they shall have for European solidarity and the transfer of competences in immigration control to the European level. The external threat of immigration should therefore strengthen citizens’ preference for national closure even if European integration can be expected to provide tangible benefits. Consequently, we might expect that a surge in immigrant arrivals shall lead to an identitarian backlash and shall consequently lower support for an EU migration policy.

H1: The higher the number of immigrant arrivals in a member state, the lower the public support for a common European migration policy.

The second explanatory perspective builds on the idea that the public is in favour of a common European migration policy to the extent to which the policy is in the national interest. The collective provision of refugee protection and border controls through European responsibility-sharing offers member states tangible collective benefits, foremost among them the preservation of the integrity of the European area of free movement (see e.g. Bauböck, 2018). In contrast, a failure to establish an effective European migration governance entails the risk of instability, a loss of state control and a deterioration of the human rights situation. These costs fall mainly on the states that face the highest migration pressure. The less exposed member states may thus have few incentives to contribute to a common European migration policy and prefer to free-ride on the efforts of other member states to control migration and host refugees. A common European policy therefore entails solidarity, whereby the more affected member states receive support from their less affected counterparts, who share the costs and responsibilities of hosting refugees and upholding migration control. Therefore, it is not surprising that calls for greater European solidarity in migration governance typically come from the member states that host a disproportionate share of refugees (Trauner, 2016). Similarly, Conti et al. (2020) show that citizens’ support for EU intervention is higher in the policy areas where one’s own country bears disproportionate costs.

Following this collective action logic, we can assume that rational citizens form their policy preferences based on their countries’ national interests and that they should express higher support for a common EU migration policy when they experience higher migration pressure in their country relative to that of others. The more a member state benefits from a common European migration policy relative to exclusively national-level policymaking, the more the public should be supportive of such a communitarisation. This interest-based explanation of EU support is based on sociotropic cost-benefit calculations. We may thus expect that citizens do not simply respond to fluctuating immigrant arrivals in their country, as hypothesis H1 postulates, but rather compare the migration pressure on their country to that on other member states as a bench-marking heuristic to assess the benefits of a common EU policy (cf. Hobolt and De Vries, 2016: 421-423). In the context of asylum governance, Gorodzeisky (2022) has shown that when European citizens form their attitudes on the matter, they consider the potential number of refugee arrivals in their country based on the refugee movements in the wider region. In a similar vein, the number of immigrants in each member state relative to those in the EU in general may serve as a proxy for the extent to which one’s own country might benefit, or lose, from a joint immigration policy at the EU level.

H2: The more immigrants arrive in a member state relative to other member states, the higher the public support for a common European migration policy.

Data and method

I test my hypotheses in a comparative analysis of public opinion in 28 EU member states using data from the Eurobarometer survey (Eurobarometer, 2023), which offers cross-sectional time series data on the political views of the European public. The pooled sample includes 47 survey waves that were fielded between 1992 and 2021 and covers all EU member states, containing a total of $N = 1, 066, 952$ individual-level observations that are aggregated into $N = 998$ country-wave observations.¹ The analysis starts when the EU migration policy was first institutionalised and covers the three decades thereafter. The main outcome of interest is public support for a common European migration policy (rather than an exclusively national one). The Eurobarometer includes a question on whether survey respondents support a joint European policy or prefer an exclusively national policy on immigration. This binary question first appeared in the 1992 wave and has been included continuously in the bi-annual survey for around three decades since then. However, there was a slight modification of the question wording, so I created two separate time series (see Online appendix for the detailed question wording).² The first time series, which runs from 1992 to 2011, asks respondents whether immigration policy should be decided only at the national level or jointly within the EU. The second time series asks respondents whether they are in favour or against a common European migration policy and runs from 2014 to 2021 (including the so-called European refugee crisis of 2015). In both cases, the main dependent variable is the share (in percentage points) of respondents who support a common European migration policy. Therefore, the question measures the preferred level of decision-making but does not ask about the kind of migration policy people want the EU to pursue. While people may have different preferences for how the EU should design its migration policy, but this specific item clearly points to support for a common responsibility in the control and admission of third-country immigrants.

To measure migration pressure as the main independent variable, I combine the survey data from the Eurobarometer with data on the number of refugees from the United Nations High Commissioner for Refugees, which is available from the year 2000 onward (United Nations High Commissioner for Refugees, 2023). Although refugees constitute a subset of all immigrants, there are several reasons why the public should be particularly sensitive to this category of immigrants. First, the European migration policy is mainly concerned with irregular third-country migration – matters of asylum and border control (Boswell and Geddes, 2011). In contrast, the admission of labour and family migrants from non-European countries largely remains in the competence of member states. Second, in Europe, public perceptions of immigrants focus on refugees; when people think of an immigrant, they commonly picture that person as a refugee (Blinder, 2015). Third, refugees are perceived to be particularly threatening, because they are assumed to be poor and culturally distant individuals from the Global South who enter the receiving country irregularly (Marx and Naumann, 2018; Solodoch, 2021). For these reasons, the arrival of refugees is the type of problem pressure most likely to shape the public support for an EU migration policy. For the purposes of my analysis, I measure the number of refugee arrivals in each EU member state both in absolute and in relative terms.³ The absolute number is the total number of asylum applications per 100,000 inhabitants that a member state received in a given year. The relative number is the number of arrivals in a member state as a percentage of the average number of arrivals across all member states. I log-transform both variables of refugee arrivals to account for their skewed distribution (see Online appendix) and lag them by one year in line with the expected causal order. For ease of interpretation, the resulting variables are scaled to a range from 0 to 1.

I include proxies for identity and utility as control variables explaining support for EU migration policy. Identity is measured as the extent to which people feel attached to the EU on a four-point scale (not at all, not very, fairly, very). For each country-wave observation, I calculate the strength of the public’s European identity using the numerical identity scale. As a proxy for utility, the models include the unemployment rate in a given year, a measure that reflects each country’s business cycle and the ease with which immigrants can be integrated into the receiving society and its labour market. Moreover, I also include the share of non-citizens (foreigners) in a country as a contextual control to account for the general migration context beyond refugee arrivals that could result in spurious effects. Finally, to identify heterogeneous effects of refugee arrivals, I rely on respondents’ left-right placement on a scale from 1 (left) to 10 (right). The detailed operationalisation of the variables and their descriptive statistics can be found in the Online appendix.

The empirical analysis proceeds in three steps. First, I conduct a descriptive analysis of the level of public support for an EU migration policy across countries and over time. Second, I estimate linear panel regression models with policy support as an aggregate variable of public support for an EU migration policy depending on the absolute and relative number of refugee arrivals in a country.⁴ I estimate a total of four base models with the two time series that use the two different wordings of the question to operationalise the dependent variable, as well as on whether they include control variables or not.⁵ The estimates are based on fixed effects models, which take into account the hierarchical country-year structure of the data (Moering, 2012). Including country-fixed effects helps avoid omitted variable bias by accounting for all time-invariant country characteristics. This model specification thus identifies the effect of refugee arrivals based on the variation within countries. Thereby, this design is also better suited to the structural occurrence of partial multicollinearity, which stems from the two variables of migration pressure – they are based on the same data and are thus by their nature characterised by a high correlation. As there is hardly any unique variation for the two migration variables in the country-means, the analysis uses the within-country variation instead of the between-country variation.⁶ While partial multicollinearity does not bias the coefficients, it may increase the standard errors and make the estimates unstable (Voss, 2005). To ensure the robustness of the findings, I employ a series of alternative model specifications to demonstrate the stability of the effects.

My third empirical strategy leverages the so-called called ‘European refugee crisis’ of 2015 as an asymmetric external shock with the massive spike in refugee arrivals in some member states, but not in others (see Lutz and Karstens, 2021 for a similar design). The surge of refugee arrivals in the EU peaked in the summer of 2015 and the issue of immigration reached new heights of public salience at the same time. This context creates a more controlled setting and allows the use of a difference-in-differences (DiD) design to estimate the effect of direct crisis exposure by comparing public support in exposed and not exposed countries before and after the crisis.⁷ The asymmetric crisis affected countries’ absolute and relative numbers of refugee arrivals to different degrees and thus allows us to empirically separate them: the countries that were not directly exposed to the crisis experienced little or no change in their absolute numbers of refugee arrivals but a decline in their relative numbers. Accordingly, hypothesis H1 (absolute refugee arrivals) would lead us to expect that crisis exposure has a negative effect and public support for an EU migration policy to fall primarily in the exposed countries. In contrast, hypothesis H2 (relative refugee arrivals) would lead us to expect a positive effect of crisis exposure, with public support increasing in exposed countries and decreasing in unexposed countries. This design thus allows me to test the two hypotheses without the simultaneous inclusion of the two migration variables and the potential multicollinearity issues associated with it.

Countries’ highly asymmetric exposure to the 2015 refugee crisis, which was a result of their geographic locations, makes it possible to identify a group of strongly exposed countries (Germany, Austria, Hungary, Sweden and Finland) and compare them to the group of countries that were not exposed (Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, France, Ireland, Greece, Italy, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Poland, Portugal, Romania, Slovenia, Slovakia, Spain, and the United Kingdom). The classification is based on a cluster analysis that uses the extent to which the numbers of refugee arrivals during the crisis period exceeded the pre-crisis numbers (see Online appendix for the detailed calculation).⁸ The DiD model compares the two country groups before and after the crisis to estimate the effect of refugee arrivals on public support for an EU migration policy. It does so through an interaction term between a crisis dummy (pre-/post-crisis) and an exposure dummy (exposed/not-exposed member states). Inference in all models is estimated based on standard errors clustered by countries (Bertrand et al., 2004). The crucial identification assumption is that the exposed and the unexposed countries experienced parallel trends in the pre-crisis period. A visual inspection indicates that this is indeed the case, particularly in the survey waves immediately preceding the crisis (see Online appendix). A statistical test shows no significant difference in the time trends of the exposed and the unexposed countries in the pre-crisis time series. Nevertheless, I follow the recommendation by Angrist and Pischke (2014) to include country-specific time trends to make the estimates more robust against the potential violation of the parallel trend assumption by allowing for a degree of non-parallel evolution of the dependent variable in member states in the absence of a treatment effect. Including (time-varying) control variables adjusts the models for potential confounders. Finally, I conduct a series of robustness checks to assess the effect stability using a continuous exposure variable as an alternative operationalisation and identifying effect heterogeneity across and within countries.

Results

In a first step, I describe the general pattern of public support for an EU migration policy. Figure 1 displays the temporal evolution of the percentage of Europeans who express support for a common EU migration policy. With the exception of a short episode in the early 2000s, a majority of all Europeans were in favour of a joint European policy. While support levels declined slightly throughout the 1990s, they have henceforth continuously increased and stabilised at around 75%. This is a significant level of support in light of the fact that immigration control constitutes a core state power and that issues of national sovereignty have become increasingly contested in many member states. What is more, public support for a European migration policy has increased over time despite the further transfer of competences in that policy area to the European level with each treaty revision. Furthermore, major migration crises (in the years 1992, 1999, 2011, and 2015) do not seem to systematically alter the overall support levels. We can thus conclude that despite common perceptions of loss of control and the often contentious nature of migration governance, the possibility of deciding migration policy at the European level enjoys growing majoritarian support among the European public.

Despite this relatively stable majoritarian support, there is significant variation in support across and within member states (see Figure 2). In the 1990s, when the EU had only had very few competences in migration policy, the public supported the communitarisation of this policy field in a majority of member states, most strongly in Italy and Spain. However, in the Northern states of Finland, Sweden, Denmark, the United Kingdom, and Austria only a minority expressed support for a common European policy. This pattern remained largely stable into the 2000s. The Eastern member states that joined in 2004 initially expressed high levels of support. This only changed in 2015 – when a surge of refugee arrivals took place in Europe – when the Visegrád countries became more negative towards a European migration policy. At the same time, public support has increased substantially in other member states, such as Sweden, Germany or Greece, which suggests that the high number of refugees during the crisis lead to more public support for a European approach to international migration, instead of a withdrawal to national closure. Over time, the Southern and Northern countries have converged in their support levels with Southern countries catching up to their Northern counterparts. The new member states in Eastern Europe initially had high levels of support, similar to those of Northern Europe, but have recently diverged from the rest by becoming more sceptic to the idea of a common European policy (see also Online appendix). This descriptive overview reveals significant variation both across countries and over time, as well as region-specific dynamics in policy preferences.

In a next step, I use panel regression models to estimate the effects of refugee arrivals on public support for an EU migration policy (see Table 1). Model 1 on the first time series (2000 to 2012) shows that countries’ absolute number of arrivals has a negative effect, whereas their relative number of arrivals has a positive effect on public support for an EU migration policy. Both coefficients are statistically significant. These results are in line with the theoretical expectations of H1 and H2. The second time series (2014 to 2021) produces the same pattern of results; however, the effects are generally weaker. While all coefficients are of substantial size, in the first period, the absolute number of arrivals has a stronger effect but in the second time period this is the case with the relative number of arrivals. Adding the control variables to the models does not alter the pattern of the results.⁹

Figure 1.

Evolution of public support for an EU migration policy.

Figure 2.

Geographical distribution of support for an EU migration policy.

I perform several robustness checks to assess the stability of the estimated effects (see Online appendix for detailed outputs). Although it makes theoretical sense to include both migration variables, I estimate a model only with countries’ absolute number of refugee arrivals, confirming the variable’s significant negative effect in the first time series but not in the second one. These results confirm the previous finding that the absolute number of arrivals is the dominant factor in the early period and the relative number of arrivals becomes the dominant factor in the post-2014 period. A re-sampling check excludes countries with a very high correlation between the independent variables that therefore provide limited unique variation. The resulting estimates confirm the results from the base models. Furthermore, to assess the functional form of the model, I draw partial residual plots for the independent variables (see Online appendix) that lend additional support to the hypothesis of a linear relationship between refugee arrivals and public policy support. Finally, a number of treaty changes have expanded the competences of the EU in immigration matters over the time period under investigation. Public support could be responding to these treaty changes in a thermostatic manner. If this is the case, we should observe lower support after major treaty changes. The inclusion of treaty dummies into the first model does not alter the effect of refugee arrivals in any substantive way. The results are thus consistent across different model specifications.

To further corroborate these results in a more controlled setting and to gain a better understanding of their substantive importance, I examine the effect of the so-called ‘European refugee crisis’ of 2015 on public support for an EU migration policy. For this purpose, I estimate DiD estimates (which compare exposed and unexposed countries before and after the crisis event) that treat the refugee crisis as an asymmetric exogenous shock which exposed EU member states to extra-EU immigration to a different degree. The results are presented in Table 2 and indicate that the crisis had a significant effect on public support for an EU migration policy based on countries’ crisis exposure: While the crisis has significantly reduced overall support, direct exposure to the crisis has mitigated this negative effect. These dynamics shape a pattern of relatively stable support in those countries that experienced large numbers of refugee arrivals, while support dropped by almost seven percentage points in countries that did not experience a surge in refugee arrivals during the 2015 refugee crisis. The crisis reversed a stable pre-crisis pattern of European public opinion and established a new pattern of public support that has generally remained stable in the years after the crisis (see Online appendix): Support has hovered around five percentage points higher in exposed countries than in unexposed countries in the aftermath of the crisis, whereas before the crisis the former registered a lower support for an EU migration policy than the latter. These results suggest that not only do Europeans respond to refugee arrivals in their own country, but they also react to arrivals in other countries. The evolution of public support during the 2015 refugee crisis reveals that preference formation follows the relative number of refugee arrivals as support decreased in the countries that experienced an absolute increase in refugee arrivals but less than other member states, and therefore a decrease in their relative share of refugee arrivals. That the crisis resulted in lower support for an EU migration policy overall mainly stems from the unexposed countries – a pattern that is in line with the previous results and confirms H2, which expects an interest-based evolution in public opinion.

I run additional tests and analyses to assess the validity and robustness of these results (see Online appendix for all model outputs). First, I assess the potential heterogeneity of the refugee effect within and across countries. People tend to have different views of immigration and the EU, and they might also respond differently to the migration pressure on their countries. I therefore estimate an individual-level model of support for an EU migration policy which interacts the DiD estimator with respondents’ political ideology, measured by their self-placement on the left-right scale (see Figure 3). My theoretical intuition is confirmed – the results reveal that left-wing individuals only respond to crisis exposure (by increasing their support for an EU migration policy) and right-wing individuals mainly respond to non-exposure (by reducing their support for an EU migration policy). This means that political predispositions moderate people’s crisis responses and that the growing divide among countries is accompanied by an increasing partisan divide within countries: left-wing individuals profess higher levels of support for an EU migration policy than right-wing individuals, a divide that has widened in the aftermath of the 2015 refugee crisis.

This pattern of heterogeneous effects corroborates the previous results. In line with a post-functionalist logic, those on the political right who feel more threatened by immigration tend to have lower support compared to left-wing individuals, and their support is more likely to erode in times of crisis. However, the decline in support does not happen as a uniform identitarian backlash resulting from direct crisis exposure (H1), but only in countries that were spared from a surge in refugee arrivals. In contrast, left-wing individuals tend to increase their support and call for European solidarity in times of crisis, particularly when their countries are directly affected by the crisis. In line with the collective action logic of H2, both ideological camps show a pattern of following instrumental solidarity where their support for a common European migration policy depends on their countries’ crisis (non-)exposure and to the extent to which their countries would benefit from a joint European responsibility on migration matters.

Figure 3.

Heterogeneous effects of the refugee crisis by political orientation.

Table 1.

Panel regression estimates.

	DV: Support EU migration policy (in %)
	(1)	(2)	(3)	(4)
	2000–2012	2014–2021	2000–2012	2014–2021
Independent variables:
Refugee arrivals (absolute)	$-$ 118.14^***	$-$ 23.85^***	$-$ 126.07^***	$-$ 25.14^**
	(13.34)	(7.05)	(16.64)	(9.08)
Refugee arrivals (relative)	96.94^***	29.07^***	109.46^***	31.53^**
	(13.98)	(7.36)	(16.19)	(9.83)
Control variables:
EU-Attachment				7.61^*
				(3.39)
Unemployment rate			0.21^#	0.48^#
			(0.12)	(0.28)
Share of foreigners			$-$ 2.62	$-$ 20.40
			(33.20)	(20.66)
Constant	47.44^***	54.09^***	42.44^***	34.43^**
	(4.77)	(2.41)	(7.07)	(11.46)
Observations	366	389	330	333
Countries	27	28	26	28
$R^{2}$	0.88	0.86	0.89	0.87
Adjusted $R^{2}$	0.87	0.85	0.88	0.86

Note: Linear panel regression models with country-fixed effects and country-clustered standard

errors. Statistical significance indicated as follows: $# p <$ 0.1; $* p <$ 0.05; $* * p <$ 0.01; $* * * p <$ 0.001.

Table 2.

DiD estimates.

	DV: Support for an EU migration policy
Crisis	$-$ 6.76 $* * *$
	(1.16)
Exposure	25.59
	(18.47)
Crisis*Exposure	5.93 $*$
	(2.29)
Constant	28.17^*
	(12.81)
Observations	253
Countries	23
Survey waves	11
R²	0.93
Adjusted R²	0.91

Note: DiD model with country-specific time-trends (not shown) and country-clustered standard errors.

The statistical significance is indicated as follows:

$* p <$ 0.05; $* * p <$ 0.01; $* * * p <$ 0.001.

A second type of potential heterogeneity has to do with how the effect of refugee arrivals varies across different country contexts. The regional trends in the dependent variable show that support for an EU migration policy dropped significantly in Eastern European countries in the aftermath of the 2015 refugee crisis, but did not do so in the Northern and Southern member states (see Online appendix). One could therefore come to believe that the effect of crisis exposure might be driven by Eastern member states and be a regional phenomenon rather than the result of refugee arrivals per se. When I drop Eastern European countries from the model, the effect remains positive but it is smaller and loses statistical significance. Therefore, the results are largely due to the comparison between the crisis-exposed countries in the West and the unexposed countries in the East. Looking at the country-specific trends of public support for an EU migration policy, we indeed see that Eastern European countries are those that have most strongly reduced their support during the crisis (see Online appendix). The exception to this pattern is Hungary, which experienced a large surge in refugee arrivals and increased its public support during the crisis, in line with my theoretical expectations. The trends for the Western European countries are more diverse. The strongly-exposed Germany and Sweden increased their support despite their already very high support levels (around 80%), and most countries that were not directly affected by the crisis tended to have lower levels of support, yet to a significantly smaller degree than their counterparts in Eastern Europe. It thus can be concluded that the aforementioned result patterns also occur within regions, but that the strongest effects result from a polarisation between Western destination countries and Eastern bystander countries.

This result of cross-country variation is also indicative of the fact that member states’ crisis exposure was not entirely binary but varied somewhat, with countries becoming exposed to different degrees. To account for this variation, I re-estimate the DiD-model with a continuous treatment variable of gradual crisis exposure using two different kinds of operationalisation (see Online appendix for the operationalisation and the model output). These alternative models confirm the previous results of a positive effect of direct crisis exposure. Nevertheless, the effect is only statistically significant when operationalisation uses data from 2015 and not the data combining the years 2015 and 2016. This pattern suggests that the evolution of public support mainly resulted from the peak of the crisis in 2015.

Conclusions

Migration has become a crucial policy field for the EU, with significant pressures from functional integration needs and domestic opposition to immigration in many member states. Therefore, public support for the transfer of the sovereign competence of migration control – a core state power that typically is the exclusive domain of the nation state – to the EU is hence crucial to understand the EU’s democratic legitimacy in times of contested globalisation. This study has analysed the extent to which Europeans support a common European migration policy and the role that the problem pressure of refugee arrivals plays in shaping public preferences. It suggests that the European public has largely expressed stable majoritarian support for a common European migration policy. Despite the strong interlinkage of anti-immigration sentiments and Euroscepticism in public opinion as well as the common perception of a control loss regarding immigration into the EU, the European public is clearly supportive of communitarising competences on immigration policy on the European level.

At the same time, strong fluctuations over time and across countries can be observed. To explain this pattern of public preferences, this study assesses the influence of member states’ different positions within the European migration regime. How do the very asymmetric numbers of refugee arrivals shape public preferences in EU member states? Has the often unsolicited arrival of refugees led to an identitarian backlash and a popular demand for a renationalisation of migration control? Or do people assess the individual positions of their countries and the extent to which the latter benefit from an EU migration policy?

My analysis of longitudinal survey data shows that a higher number of refugee arrivals is associated with lower support for a common EU migration policy, which confirms the hypothesis that increasing problem pressure makes people more likely to favour national responses to European ones. At the same time, countries’ relative migration pressures increase public support for an EU migration policy. This is thus supporting evidence for two simultaneous dynamics of how Europeans respond to refugee arrivals. On the one hand, refugee arrivals seem to contribute to a nativist backlash in favour of strengthening national sovereignty. The respondents with negative views on immigration tend to show lower support for an EU migration policy. On the other hand, public preferences seem to also respond to the broader European context. The larger the share of refugees in one’s own country relative to other member states, the more one is inclined to support a common European migration policy. This demonstrates that national interests do play an important role and that preference formation follows a collective action dynamic where support is a function of the extent to which one’s own country benefits from Europeanisation. This effect was strongest during the 2015 refugee crisis where member states’ highly asymmetric exposure and the salience of the issue facilitated benchmarking in the formation of policy preferences. These dynamics were most visible in Eastern European bystander countries which reduced their public support, whereas Western destination countries which experienced high numbers of refugee arrivals increased their support for a common European policy. Overall, contextual migration pressures in the form of refugee arrivals are a powerful predictor of public support for an EU migration policy; they are not confined to a short-term crisis effect but constitute an important driver of preference formation.

These findings have implications for the prospects and limits of European migration governance and for EU support more broadly. First, the pattern and evolution of public preferences are largely in line with the national governments’ discourses, with the most exposed countries most actively calling for European solidarity on immigration. Hence, public support is likely to pose an additional constraint on the EU’s efforts to overcome the collective action failure in refugee protection. The 2015 refugee crisis increased the pressure for European coordination on migration matters but it also deepened the divide between member states, as people are concerned with their countries’ national interest. This means that the collective action logic and free-rider dynamics that have been theorised to understand government actions in asylum governance (e.g. Thielemann and Armstrong, 2013) can also be found in the formation of public policy preferences.

Second, the results of this study show that transnational problem pressures may undermine public support for European integration and therefore make addressing such challenges more difficult. A functionalist necessity of collective European action does not inherently raise public support in member states. Rather, there is a risk that failing to provide effective solutions to transnational issues such as immigration may not only deepen the political divide in Europe, but also undermine the political legitimacy of the EU as the adequate actor to address those issues. That is, Europe could see itself entrapped in a context of a continuous migration crisis that simultaneously intensifies the need for collective action and reduces the political support for a common European response.

While this study provides new insights into the public support for integrating a core state power at the European level, two sorts of limitations should be mentioned. First, the extent to which domestic political discourses shape how public preferences respond to contextual migration pressures remains an open question. Second, the lack of panel data and aggregate-level analysis does not allow for the identification of the individuals that have changed their preferences in response to the arrivals of refugees, nor what individual-level mechanisms are at play. Future research should therefore further investigate the domestic conditions and individual characteristics that mediate and moderate people’s tendency to see the EU as a saviour or as a liability in the face of external threats. Moreover, it is yet to be determined to what extent the findings of this study are applicable to other transnational policy issues that require collective action at the European level.

Supplemental Material

sj-pdf-1-eup-10.1177_14651165241226721 - Supplemental material for Between common responsibility and national interest: When do Europeans support a common European migration policy?

Supplemental material, sj-pdf-1-eup-10.1177_14651165241226721 for Between common responsibility and national interest: When do Europeans support a common European migration policy? by Philipp Lutz in European Union Politics

Supplemental Material

sj-dta-3-eup-10.1177_14651165241226721 - Supplemental material for Between common responsibility and national interest: When do Europeans support a common European migration policy?

Supplemental material, sj-dta-3-eup-10.1177_14651165241226721 for Between common responsibility and national interest: When do Europeans support a common European migration policy? by Philipp Lutz in European Union Politics

Supplemental Material

sj-dta-4-eup-10.1177_14651165241226721 - Supplemental material for Between common responsibility and national interest: When do Europeans support a common European migration policy?

Supplemental material, sj-dta-4-eup-10.1177_14651165241226721 for Between common responsibility and national interest: When do Europeans support a common European migration policy? by Philipp Lutz in European Union Politics

Supplemental Material

sj-r-5-eup-10.1177_14651165241226721 - Supplemental material for Between common responsibility and national interest: When do Europeans support a common European migration policy?

Supplemental material, sj-r-5-eup-10.1177_14651165241226721 for Between common responsibility and national interest: When do Europeans support a common European migration policy? by Philipp Lutz in European Union Politics

Supplemental Material

sj-dta-6-eup-10.1177_14651165241226721 - Supplemental material for Between common responsibility and national interest: When do Europeans support a common European migration policy?

Supplemental material, sj-dta-6-eup-10.1177_14651165241226721 for Between common responsibility and national interest: When do Europeans support a common European migration policy? by Philipp Lutz in European Union Politics

Footnotes

Acknowledgements

Previous version of this article have benefited from valuable comments by Ben Crum, Thibaud Deruelle, Lukas Lauener, and the participants of the research seminars at the University of Geneva (Switzerland) and the Vrije Universiteit Amsterdam (Netherlands), the 2022 ecsasuisse conference in St.Gallen (Switzerland) and the 2023 conference by the Council of European Studies (CES) in Reykjavik (Iceland).

Declaration of conflicting interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung (208858).

ORCID iD

Philipp Lutz

Data availability statement

All data generated or analyzed during this study are included in this published article [and its supplementary information files].

Supplemental materials

The supplemental materials are available online.

Notes

References

Angrist

Pischke

(2014) Mastering ’metrics: The path from cause to effect. Princeton, NJ: Princeton University Press. DOI:https://doi.org/10.5860/choice.189854.

Bauböck

(2018) Refugee protection and burden-sharing in the european union. Journal of Common Market Studies 56: 1–16. DOI:https://doi.org/10.1111/jcms.12638.

Bertrand

Duflo

Mullainathan

(2004) How much should we trust differences-in-differences estimates? Quarterly Journal of Economics 119: 249–275. DOI:https://doi.org/10.1162/003355304772839588.

Blinder

(2015) Imagined immigration: The impact of different meanings of ‘immigrants’ in public opinion and policy debates in britain. Political Studies 63: 80–100. DOI:https://doi.org/10.1111/1467-9248.12053.

Boswell

Geddes

(2011) Migration and Mobility in the European Union. London: Palgrave Macmillan. ISBN 9780230007475.

Byrne

Noll

Vedsted-Hansen

(2020) Understanding the crisis of refugee law: Legal scholarship and the eu asylum system. Leiden Journal of International Law 33: 871–892. DOI:https://doi.org/10.1017/S0922156520000382.

Connor

(2018) A majority of europeans favor taking in refugees, but most disapprove of eu’s handling of the issue. Pew Research Center, available at: https://pewrsr.ch/2D7Cjmj (accessed 14 November 2022).

Conti

Mauro

Memoli

(2019) Citizens, immigration and the eu as a shield. European Union Politics 20: 492–510. DOI:https://doi.org/10.1177/1465116519834648.

Conti

Mauro

Memoli

(2020) Immigration, security and the economy: Who should bear the burden of global crises? burden-sharing and citizens’ support for eu integration in italy. Contemporary Italian Politics 12: 77–97. DOI:https://doi.org/10.1080/23248823.2020.1712015.

10.

Czymara

Schmidt-Catran

(2017) Refugees unwelcome? changes in the public acceptance of immigrants and refugees in germany in the course of europe’s ‘immigration crisis’. European Sociological Review 33: 735–751. DOI:https://doi.org/10.1093/esr/jcx071.

11.

Dennison

Geddes

(2019) A rising tide? the salience of immigration and the rise of anti-immigration political parties in western europe. Political Quarterly 90: 107–116. DOI:https://doi.org/10.1111/1467-923X.12620.

12.

De Vreese

Boomgaarden

(2005) Projecting eu referendums: Fear of immigration and support for european integration. European Union Politics 6: 59–82. DOI:https://doi.org/10.1177/1465116505049608.

13.

De Wilde

Trenz

(2012) Denouncing european integration: Euroscepticism as polity contestation. European Journal of Social Theory 15: 537–554. DOI:https://doi.org/10.1177/1368431011432968.

14.

Eilstrup-Sangiovanni

(2021) Re-bordering europe? collective action barriers to ‘fortress europe’. Journal of European Public Policy 28: 447–467.

15.

Erisen

Vasilopoulou

Kentmen-Cin

(2020) Emotional reactions to immigration and support for eu cooperation on immigration and terrorism. Journal of European Public Policy 27: 795–813. DOI:https://doi.org/10.1080/13501763.2019.1630470.

16.

Eurobarometer (2023) Eurobarometer. European Commission. Available at: https://europa.eu/eurobarometer/.

17.

Goodwin

Milazzo

(2017) Taking back control? investigating the role of immigration in the 2016 vote for brexit. British Journal of Politics and International Relations 19: 450–464. DOI:https://doi.org/10.1177/1369148117710799.

18.

Gorodzeisky

(2022) The influx of refugees and temporal change in attitudes towards asylum seekers: A cross-national perspective. European Sociological Review 38: 648–662. DOI:https://doi.org/10.1093/ESR/JCAB066.

19.

Hangartner

Dinas

Marbach

et al. (2019) Does exposure to the refugee crisis make natives more hostile? American Political Science Review 113: 442–455. DOI:https://doi.org/10.1017/S0003055418000813.

20.

Harteveld

Schaper

de Lange

et al. (2018) Blaming brussels? the impact of (news about) the refugee crisis on attitudes towards the eu and national politics. Journal of Common Market Studies 56: 157–177. DOI:https://doi.org/10.1111/jcms.12664.

21.

Heizmann

Ziller

(2020) Who is willing to share the burden? attitudes towards the allocation of asylum seekers in comparative perspective. Social Forces 98: 1026–1051. DOI:https://doi.org/10.1093/sf/soz030.

22.

Helbling

Meierrieks

(2020) Transnational terrorism and restrictive immigration policies. Journal of Peace Research 57: 564–580. DOI:https://doi.org/10.1177/0022343319897105.

23.

Hobolt

De Vries

(2016) Public support for european integration. Annual Review of Political Science 19: 413–432. DOI:https://doi.org/10.1146/annurev-polisci-042214-044157.

24.

Hooghe

Marks

(2005) Calculation, community and cues: Public opinion on european integration. European Union Politics 6: 419–443. DOI:https://doi.org/10.1177/1465116505057816.

25.

Hooghe

Marks

(2009) A postfunctionalist theory of european integration: From permissive consensus to constraining dissensus. British Journal of Political Science 39: 1–23. DOI:https://doi.org/10.1017/S0007123408000409.

26.

Hooghe

Marks

(2018) Cleavage theory meets europe’s crises: Lipset, rokkan, and the transnational cleavage. Journal of European Public Policy 25: 109–135. DOI:https://doi.org/10.1080/13501763.2017.1310279.

27.

Isernia

Cotta

(2016) A sword or a shield? Globalization, European integration, and public opinion. In: Westle B and Segatti P (eds) European Identity in the Context of National Identity: Questions of Identity in Sixteen European Countries in the Wake of the Financial Crisis. Oxford: Oxford University Press, 211–235. DOI:10.1093/acprof:oso/9780198732907.003.0008.

28.

Kriesi

Grande

Dolezal

et al. (2012) Political Conflict in Western Europe. Cambridge: Cambridge University Press.

29.

Lutz

(2021) Loved and feared: Citizens’ attitudinal ambivalence towards free movement in the european union. Journal of European Public Policy 28: 268–288. DOI:https://doi.org/10.1080/13501763.2020.1720782.

30.

Lutz

Karstens

(2021) External borders and internal freedoms: How the refugee crisis shaped the bordering preferences of european citizens. Journal of European Public Policy 28: 370–388.

31.

Marx

Naumann

(2018) Do right-wing parties foster welfare chauvinistic attitudes? a longitudinal study of the 2015 ‘refugee crisis’ in germany. Electoral Studies 52: 111–116. DOI:https://doi.org/10.1016/j.electstud.2018.01.011.

32.

Mauro

Memoli

(2021) The role of public opinion in eu integration: Assessing the relationship between elites and the public during the refugee crisis. Journal of Common Market Studies 59: 1303–1321. DOI:https://doi.org/10.1111/jcms.13183.

33.

Moering

(2012) The fixed effects approach as alternative to multilevel models for cross–national analyses. The GK Soclife working paper series.

34.

Oltmer

(2022) The Borders of the EU. Europena Integration, “Schengen” and the Control of Migration. Springer. ISBN 9783658391997.

35.

Rankin

Oltermann

(2018) Future of eu hinges on solving migration issue, says merkel. The Guardian. Available at: https://www.theguardian.com/world/2018/jun/28/future-of-eu-hinges-on-solving-migration-issue-says-merkel.

36.

Semyonov

Raijman

Gorodzeisky

(2008) Foreigners’ impact on european societies: Public views and perceptions in a cross-national comparative perspective. International Journal of Comparative Sociology 49: 5–29. DOI:https://doi.org/10.1177/0020715207088585.

37.

Solodoch

(2021) Regaining control? the political impact of policy responses to refugee crises. International Organization 735–768. DOI:https://doi.org/10.1017/S0020818321000060.

38.

Stockemer

Niemann

Unger

et al. (2020) The “refugee crisis,” immigration attitudes, and euroscepticism. International Migration Review 54: 883–912. DOI:https://doi.org/10.1177/0197918319879926.

39.

Thielemann

Armstrong

(2013) Understanding european asylum cooperation under the schengen/dublin system: A public goods framework. European Security 22: 148–164. DOI:https://doi.org/10.1080/09662839.2012.699892.

40.

Toshkov

Kortenska

(2015) Does immigration undermine public support for integration in the european union? Journal of Common Market Studies 53: 910–925. DOI:https://doi.org/10.1111/jcms.12230.

41.

Trauner

(2016) Asylum policy: The eu’s ‘crises’ and the looming policy regime failure. Journal of European Integration 38: 311–325. DOI:https://doi.org/10.1080/07036337.2016.1140756.

42.

United Nations High Commissioner for Refugees (2023) Refugee population statistics database. Available at: https://www.unhcr.org/refugee-statistics/download/?url=sH5pnE.

43.

Van Der Brug

Harteveld

(2021) The conditional effects of the refugee crisis on immigration attitudes and nationalism. European Union Politics 22: 227–247. DOI:https://doi.org/10.1177/1465116520988905.

44.

Voss

(2005) Multicollinearity. New York: Elsevier. pp.759–770. DOI:10.5860/choice.42-5629.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.11 MB

91.58 MB

0.00 MB

3.62 MB

0.03 MB