Immigration and political distrust in Europe: a comparative longitudinal study

The smooth and effective functioning of democratic societies depends on citizens’ trust in their government. Trust, both in institutions and among people, reduces social complexity and diminishes uncertainty (Luhmann 1979). Political trust – the belief that the political system serves the public interest - provides the foundation for democratic representation (Sztompka 2000), political participation (Marien and Hooghe 2011; Nyćkowiak 2009; Putnam 1993), coordination between citizens (Braithwaite and Levi 1998), and citizen compliance with norms and laws (Levi 1997).

Recently, democratic societies have experienced a rise of distrust in institutions as part of an emerging public malaise with politics (Hetherington and Rudolph 2015; Torcal and Montero 2006). Permeating the public discourse, this malaise is often expressed through social media (Yamamoto et al. 2016) and via polarized mass media outlets (Prior 2013). In particular, political distrust, which is a sense that trust in political institutions has been violated, has become more pervasive (see van de Meer and Zmerli 2017).¹European countries have witnessed a slow but steady decline of political trust over the past decades (Foster and Friedan 2017; Marien 2011; Norris 2011). The trend is rather disquieting as distrust is an important dimension of polixtical alienation (Fendrich and Axelson 1971; House and Mason 1975; Hunt 1982; Kahn and Mason 1987; Macke 1979; Mason et al.1985; Neal and Rettig 1967; Seeman 1975; Zeller et al.1980).

A general long term decline in political trust has partly been attributed to a rise in scepticism and post-material values changes (Crozier et al.1975; Inglehart 1997; Pharr and Putnam 2018), but this has evolved asymmetrically across time and space and the reasons for this have been widely debated (Bovens and Wille 2008; Hendriks 2009; Lipset and Schneider 1987; Marien 2011; Mishler and Rose 2001; Norris 2011).

The rise in political distrust in Europe coincides with the rise of immigration (Pennings 2017). Putnam (2007)'s claim that diverse societies ‘hunker down’ and are less trusting has raised questions about whether this is caused by decades of sustained immigration. Yet, so far, social inquiry into whether or not immigration lowers political trust in host societies is lacking. While the relationship between an individual's immigration sentiments and his or her extent of political distrust has been empirically supported (MacLaren 2012a, 2012b), the relationship between actually-occurring immigration as a demographic phenomenon and the rise in political distrust remains not well understood. The need to better understand this relationship is also crucial for policy-makers who may - either rightfully or erroneously - believe that curbing immigration inflows may be a way to restore political trust.

The aim of this article is to theoretically develop and empirically test the relationship between actual immigration trends and political distrust. In this sense, this study is not interested in individual predisposition towards immigration per se as has been established in the existing literature, but rather explore the possibility of immigration trends as a contextual-level condition in Europe which have led to distinct patterns in political distrust across time and space. The asymmetry of changes in foreign populations across European countries have generated contextual variations which are useful for a comparative study² and help avoid the interference of specific events which might occur in a single- country study such as political scandals or terrorist attacks.

Political distrust is the complement of political trust (Rose and Mishler 2011). As an attitude, political trust can be defined as the perception that the government is responsive to the citizenry, producing outcomes consistent with an individual's expectations (Hetherington 2004, 9). The presence of trust means that citizens feel that ‘their own interests would be attended to even if the authorities were exposed to little supervision or scrutiny’ (Easton 1957, 447). Easton (1957) conceptualizes political trust as a diffuse support of the political system and its modus operandi (Offe 2006, 30). Diffuse support provides a rationale for why citizens can maintain support for the political system as a whole although they may not agree with a particular policy of government. This support enables citizens to acknowledge the authority of political regimes that they may not agree with and still have faith that the government will follow fair procedures.

Political distrust is not simply having low trust (Van De Walle and Six 2014) but is a violation of trust. Rather, it is driven by a political discontent that leads to criticism of the electoral system or the policy-making process (Gamson 1968). Political distrust is ‘an unfavourable evaluation of politics’ relative to normative standards ‘by citizens who see a discrepancy between an ideal and a reality’ (Hart 1978: 28).

This article is concerned with the contextual determinants of distrust, rather than individual attributes or personality traits of the individual whereby the distrusting persons are seen as ‘misanthropic individuals with a pessimistic opinion of human nature’ (Newton 2014: 25). Political distrust is usually an indication of alienation from the political system (Miller 1974), which, in Durkheim's view, was a condition of social structure (Durkheim 1892). A prevailing sense of political distrust in a society can signal that ‘a social system is under severe strain and possibly on the verge of fundamental structural change’ (Lewis and Weigert 1985: 974).

Drawing on two opposing theoretical traditions of macro-contextual explanations for political trust in the literature (see Mishler and Rose 2001), there are two main mechanisms through which immigration might influence political distrust in a society. While the first mechanism is based on a bottom-up, cultural theoretical interpretation of political trust as a product of social relations, the second is based on an interpretation of political distrust as a top-down driven phenomenon which depends on the performance of political institutions.

The first potential mechanism, which is based on a cultural, bottom-up theoretical interpretation of political trust, would predict that immigration weakens a society's solidarity, a crucial ingredient for political trust. This line of reasoning has been most prominently advanced by Putnam (2007) who argues that individuals who live in more diverse neighborhoods ‘hunker down’ and disengage from the collective (Putnam 2007: 149). According to Putnam (2007)'s ‘constrict theory,’ diversity not only reduces out-group solidarity but also in-group solidarity. Putnam finds that living in more diverse areas causes people to be less trusting of other people and of government. Given that political trust is thought to be based on the assumption of shared values (Uslaner 2002), a weakening of the collective identity could reduce trust in political institutions. Putnam (2007)’s work suggests that contextual diversity awakens feelings of disunity which might lead to political alienation and political distrust. Yet, when it comes to the impact of contextual immigration at the national level and its impact on political distrust, there is an important distinction. While local neighborhoods change due to immigration-driven diversity, local governments are not responsible for immigration policy as national governments are.

While Putnam proposes a mechanism through which diversity impacts trust, he focuses on the social dynamics in local neighborhoods. ||McLaren (2012a, 2012b) extends this argument by investigating how this might also lead to distrust towards national political institutions. Her central argument is that a person's sentiments about immigration are tied up with political (dis)trust, whereby having a negative view about immigration is driven by a person's diminished sense of communal or national solidarity. According to this line of reasoning, feelings of disunity experienced by individuals concerned about immigration actually provoke disunity not only between citizens themselves but also between citizen and the state.

One would then expect, according to this view, that actual demographic increases in immigration would accentuate the visibility of the immigration phenomenon and the ethnic diversity that it propagates. This argument rests upon the assumption that actual immigration trends are perceptible to regular citizens. Certainly, the public's misperceptions regarding exact statistics of immigration, a so-called ‘immigration innumeracy’, is well established (e.g. Herda 2010), this does not mean that the public is imperceptible to relative increases in foreigners. In fact, recent evidence shows that while the public may not be able to estimate the exact size of the foreign population in absolute terms, their estimates tend to correlate with relative differences in immigration trends (Lundmark and Kokkonen 2014). Based on these theoretical expectations, I derive the following hypothesis for empirical testing:

As such, increases in immigration may be perceived as poor institutional performance since many citizens believe that border protection is an important responsibility for a sovereign nation-state and that admitting large numbers of foreigners is a system failure (Czaika and De Haas 2013). For instance, in a recent study in the United States, Anglo-Americans living in areas where immigration is less strictly enforced are less trusting in political institutions (Rocha et al.2015). As the presence of ‘others’ increases in Europe, this may trigger further scepticism about the political system that admitted them. It may be hard for citizens to reconcile the presence of these newcomers with the functioning of their political system, particularly if they feel that immigration is detrimental to their country's economy or culture.

The above explanation suggests that immigration would not affect the political distrust of all citizens in the same way and, instead, immigration trends would have a differential effect depending on a person's attitudes towards immigration. If individuals perceive immigration as a threat, then as MacLaren points out, ‘the institutions that govern them are likely to be called into question’, making it more likely that citizens blame them ‘for allowing large-scale migration to take place in the first place’ (McLaren 2017: 319). In other words, if the mechanism is performance-based, we would then expect a rise in political distrust due to a perceived ineffectiveness of government to control immigration, which would mainly occur amongst individuals who perceive immigration as problematic.

Therefore, based on this institutional interpretation of political trust, I put forward the following additional hypothesis for testing:

I analyse micro data from eight biannual rounds (2002–2016) of the European Social Survey (ESS). The survey population intends to cover people 15 years and older who are resident in the country. The sample is selected by strict random probability methods and respondents are interviewed face-to-face. In each round, the minimum effective sample sizes are at least 1,500 (or 800 where the population is less than two million) in each participating country. The minimum target response rate is 70%.

The micro data has been pooled with macro-level country data from Eurostat.³ The countries included in this analysis are the 17 Western European countries: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland and the United Kingdom.⁴ The dataset is composed of 17 countries and 8 time periods, there are 111 country-year aggregate units with a total of 187,651 individual-level observations, after applying list-wise deletion.⁵

The dependent variable is a factor constructed using two measures of a person's political distrust: distrust in parliament and distrust in politicians (Cronbach's alpha: 0.83). A factor analysis (varimax rotation) reveals that the items load onto one main factor, which has a variance of 2.25 This indicates that while distrust in parliament is indicative of dissatisfaction in policy-making and distrust in politicians measures dissatisfaction with political actors, there is coherence around single-dimension construct of political distrust.⁶

The independent variable is foreign nationals⁷ as a proportion of the country's total population. The availability and quality of migration data is an important concern. For the purposes of this study, stock measures are a more convenient and reliable measure across countries than immigration inflows because what constitutes immigration is not uniformly defined and records of migration tend to represent the legal or immigration policy framework of the country and do not adhere to international standardizations. The advantage of using a foreign stock measure is that the criteria remain uniform across all countries and over time. The measure still has some drawbacks, namely that the size of the foreign national population can change owing to other factors than migration flows, such as natural increases in the foreign population and legislation on naturalization. Moreover, like all official migration data, it tends to underestimate foreigners who are undocumented or migrate seasonally.

Central to this analysis is the notion that an individual's concerns about immigration amplify the relationship between immigration in a country and their distrust in political institutions. For this I use two measures of concerns of immigration: (1) concerns that immigration threatens the country's economy and (2) concerns that immigrations threatens the country's culture. Each of these is then interacted with the stock foreigners and introduced into the model as a cross-level interaction term.

I also use a standard series of control variables that affect a person's level of political trust. I control for a person's age and time spent in education which are coded in years. Dummy variables for females, whether the person is unemployed, or lives in an urban area. A dummy variable for whether or not the person is a first or second generation immigrant is also included. Furthermore, the person's orientation on the left-right political scale is included as well.⁸ I add a dummy variable for subjective low income, which indicates if the respondent is struggling to live on his or her current income.⁹

I also introduce a series of country-level controls. I control for political efficacy since aspects of institutional performance are found to positively determine political trust both within and across countries. The measures of institutional trustworthiness come from the Worldwide Governance Indicator (WGI) database. The aggregated data is based on the perceptions of various stakeholders, which include household surveys, private sector experts and NGOs. Each country is then given a score ranging from −2.5 to 2.5 standardized units across various criteria.¹⁰ A control is also introduced for the extent to which the country's immigration policy is increasingly restrictive. To be comparable across countries, this measure does not show how restrictive the country is overall, but rather the number of policy changes in immigration policy which made it more or less restrictive during the period of 1994–2014. The level of restrictiveness depends on the number of changes that were introduced.

I also control for the country's GDP per capita, unemployment rate, and inequality (Gini coefficient). Dummy variables for each bi-annual round of the survey (2002–2012) are also included to adjust for any overall time trend. Data for all four of these control variables comes from the Eurostat database. A table with the descriptive statistics of all variables used in the analysis and a correlation matrix can be found in the appendix.

As an empirical strategy, I employ a multilevel modeling approach that is tailored to the structure of repeated cross-national cross-sectional data. This approach is referred to as comparative longitudinal research design because while the individuals (micro-units) in the survey are not repeatedly observed, the countries (macro-units) are repeatedly observed.¹¹ Following Fairbrother (2014), I employ a three-level hybrid model whereby individuals are nested within time (country-years) which are then nested within countries. Macro explanatory variables are then split into two different components: the cross-time mean for each country (the cross-sectional between-country difference component, BE) and demeaned variations for each observation at country-year (the longitudinal within-country change component, WE).¹² The advantages of this three-level multi-model is that it distinguishes between-country effects and within-country change, while controlling for compositional differences at the individual level (see Fairbrother 2014 for a detailed discussion, see Schmidt-Catran and Fairbrother 2016; Ziller 2015 for applications). Using multi-level analysis also helps avoid committing an ecological fallacy which occurs when data is analyzed at one level but conclusions are formed at another level (Hox 2010). Therefore, multi-level approaches are the most commonly used method in comparative attitudinal studies (for overviews see Gelman and Hill 2007; Steenbergen and Jones 2002).

Table 1 reports the coefficients for the null model and introduces individual-level controls for the dependent variables. Model 0 shows individual variation in political trust to be mostly at the individual level, but that it also exists at the country-level and the country-year level, justifying a multilevel approach (Snijders and Bosker 2012). Including individual-level controls in Model 1 improves the model fit as indicated by a higher pseudo-likelihood and by a reduction in the residual variance. Model 1 also confirms, as has been extensively demonstrated in the existing literature, that socio-demographic controls are strong predictors of political distrust.

The results in Table 2 show that actual immigration trends are generally unrelated to political distrust in Europe. In Model 2, the coefficients for foreigner group size (between and within country) are introduced, reducing both the country and country-year variance components when compared with Model 1 in Table 1. Still, we can observe that both coefficients for foreigner group size are statistically insignificant in Models 2–4. This means that neither the average foreigner group size in the country, nor the annual change in foreigner group size predict whether an individual is distrusting of their country's political institutions.

In sum, these results suggest two important conclusions: (1) Immigration is not a leading cause of political distrust in Europe, yet (2) there is evidence of a relationship, conditional on a person's attitudes towards immigration. While the conditional relationship is significant and robust,¹³ it remains relatively small in substantive terms. This is because the model shows that a person's attitudes about immigration conditional the relationship between foreigner group size is significantly related to political trust only when attitudes towards immigration are extremely negative (8 or higher). The proportion of Europeans who feel so strongly threatened by immigration is relatively small (less than 15%). In other words, for 85% of Europeans, continuing immigration will not bring about more distrust towards national political institutions.

Using the benefits of a multi-level research design, this study provides new insight into how sustained immigration is related to European citizens’ relationship with politics. The findings fail to support the hypothesized relationship between a growing foreign population and higher political distrust amongst its citizens. In other words, there does not appear to be a culturally driven phenomenon where actual immigration leads to a general weakening of trust between the governed and governing. This finding then implies that mass immigration is not the cause of the continued decline of political trust in Europe.

Yet this is not to say that immigration trends are entirely unrelated to political trust. Sustained immigration as it occurred in many European countries is likely to have reflected poorly on the perceived performance of domestic political institutions in the eyes of some citizens. In fact, based on the results of this study, I contend that for individuals who oppose immigration, continued immigration, has likely been perceived as a negative aspect of performance of their governments and institutions which allow the process to continue.

Rather, based on our findings, we speculate that certain (but not all) citizens may be perceiving immigration as domestic system failures because they still cling to the normative expectations that it is the responsibility of their national political institutions to control immigration. It would appear that it is the discrepancy between an individual's ideals about how much immigration should occur in their society and the realities of its actual occurrence what gives rise to political distrust.

In sum, the findings in this paper support a theoretical interpretation of the relationship between immigration and political distrust, which is driven by the performance of political institutions rather than a cultural interpretation. Nevertheless, it would be premature to discount the cultural mechanism entirely, based on these results. The lack of empirical support for this possible mechanism in this study could be due to a series of limitations. Firstly, it may be that cultural change occurs at a much slower pace than is perceptible in the time frame of this study (15 years). It might also be, as the study begins in 2002 when most European countries already have sizeable foreign populations, that most cultural changes had already occurred. Finally, the reason that immigration might seemingly not be related to political trust might be due to public misperceptions about immigration trends. Individuals who are against immigration might have a heightened awareness regarding its occurrence in society. By the phenomenon being more perceptible to individuals who are more anti-immigration, this could be an alternative explanation for the differential role that immigration has on their political trust.

Despite these limitations, the findings of this study raise some important questions about the role immigration has had in the decline of political trust in Europe. Why should we be concerned if immigration is breeding political distrust among a sub-group of European citizens? Firstly, because being distrustful of politics influences a person's political behavior such as their electoral choices (Chanley et al.2000; Hetherington 1999; Ziller and Schübel 2015). The issue of immigration is just one factor contributing to political distrust, it also coincides with other important developments in liberal democracies such as globalization, offshoring, and a growing social distance between citizens and the political elite. If anti-immigration sentiment becomes more pervasive in the public and mass immigration persists, then a growing and sustained political discontent could become embedded into a ‘generalized climate of suspicion leading to alienation and passivism’ (Sztompka 1999: 22). Much is still to be learned about the citizens’ norms and expectations of democracy and the way in which immigration, along with other conditions of our global system, translate into social and political grievances. Future research should investigate how these grievances are formed and to which political actors blame is attributed.

My sincere thanks go to Peter Kemp for his advice and support throughout this research. I am also grateful to the anonymous reviewers, Erzbet Bukodi, Isabel Shutes, members of the Politics of Social Policy Research Group at the University of Oxford for their useful comments. I would also like to thank Christine Jeannet for her careful reading of this work.

No potential conflict of interest was reported by the author.

Anne-Marie Jeannet is a postdoctoral researcher at the Migration Policy Centre at the European University Institute. She is also a research affiliate at the Carlo F. Dondena Centre for Research on Social Dynamics at Bocconi University in Milan. She received her doctorate from St. Cross College at the University of Oxford. Her research interests focuses on the political sociology of immigration, inter-group relations, and the influence of the mass media.

Anne-Marie Jeannethttp://orcid.org/0000-0002-0773-1485