Standardised integration requirements for naturalisation: less rights and less discretion? A qualitative meta-analysis of ethnographic studies of naturalisation procedures in Europe

Abstract

Since the late 1990s and early 2000s, in what can be called an ‘integrationist wave’, standardised integration requirements for naturalisation have become increasingly common in Europe. To examine the impact of such measures, I combine original ethnographic data on institutions involved in the implementation of citizenship policies in Belgium and the UK with a qualitative meta-analysis of existing ethnographic studies of the implementation of citizenship policies. I show how, in addition to introducing new obstacles to naturalisation, standardised requirements have also reduced the discretion inscribed in earlier procedures, albeit not uniformly across different cases. The integrationist wave could thus be understood not simply as the introduction of restrictive notions of integration, but also as the systemisation of earlier evaluation practices. I further show signs of a more recent tendency for states to retreat from examining the integration of candidates to citizenship, outsourcing the evaluation directly or indirectly to private actors.

Introduction

In the past few decades, new policies have been introduced across Europe based on standardised citizenship requirements, with the literature largely regarding such policy efforts as an installation of new barriers to naturalisation. In this article I explore whether the same new policies could have also limited the discretion involved in naturalisation processes, looking at the implementation of citizenship policies.

This integrationist (Kundnani 2007; Sredanovic, 2022a) wave, also referred to as the ‘civic turn’ (Goodman, 2010, 2012; Borevi et al., 2017; Mouritsen et al., 2019; Fargues, 2020), has involved the introduction of formal integration requirements for naturalisation, often in the form of language and culture tests (but see infra). The rhetorical presentation of such policies has not been limited to claims that new citizens should demonstrate that they are ‘deserving’ or ‘compatible’. It has been further argued that integration requirements are in the interest of applicants themselves, who, according to these arguments, will acquire tools to participate in the country’s political, economic and social life (cf. the analyses in van Oers et al., 2010a; Kiwan, 2013; Mason, 2014). Nevertheless, several analyses of this wave have shown how such integration requirements introduce further obstacles to citizenship obtention. This particularly for more vulnerable groups, who are likelier to fail the integration tests or be discouraged from applying (e.g. van Houdt et al., 2011; van Oers, 2013; Peters et al., 2016; Bassel et al., 2021; Jensen et al., 2021).

As I show in this article, the integrationist approach to citizenship policies has become nearly ubiquitous across Europe, with very few countries keeping naturalisation policies without language and cultural or civic knowledge requirements. Consequently, this policy change is extensively discussed in the literature, including policy (e.g. Goodman, 2010, 2012; van Oers et al. 2010; Borevi et al., 2017; Rea et al., 2018) and normative (e.g. Bauböck and Joppke, 2010; Mason, 2014; Schinkel, 2017, 2018) analyses. A smaller but growing body of literature argues that how the policies are implemented day by day can outweigh the letter of the law in defining access to citizenship, highlighting implementation as a crucial dimension of citizenship policies (e.g. Huddleston, 2013, 2020; Huddleston and Falcke 2020). This argument follows a central finding in the field of implementation studies: the tendency of bureaucrats to develop routines and local logics that go beyond the letter of the law and are often needed to make laws applicable day by day (the classic reference is Lipsky, 1980). The implementation approach in migration studies has been consolidating as a field of study in recent years. In addition to the present special issue, a few other volumes and special issues have covered several domains of migration policies (Eule et al., 2018, 2019; Infantino & Sredanovic, 2022).

A number of ethnographic studies have covered integrationist wave-related policies such as mandatory integration courses for newcomers and citizenship ceremonies. However, most have focused on migrants’ experiences rather than on bureaucrats’ practices. Such studies have often shown critical aspects of integrationist policies. These include how they push migrants to enter less-desirable jobs (Suvarierol, 2015; Lønsmann, 2020) and encourage the division of migrants into more- and less-deserving groups (Monforte et al., 2019) and, in turn, the depiction of ‘super-citizens’ as the only ones deserving of rights (Badenhoop, 2017, 2023). Moreover, citizenship ceremonies can dismiss migrants’ experiences in the country by portraying newly naturalised citizens as if they had just arrived (Byrne, 2014).

Some quantitative studies have also shown the limits of integrationist policies in relation to their stated aims. For example, some cross-country analyses (Goodman & Wright, 2015) show that more integrationist policies increase interest in politics but not economic integration or the feeling of belonging in society. However, Bartram’s (2019) analysis of the UK shows that beyond not increasing political participation, the experience of citizenship tests and ceremonies even lowered interest in politics. Elsewhere (Sredanovic, 2022a), I have offered my critique of integrationist policies. I have shown how the simultaneously vague and complex concept of integration creates problems in implementing citizenship criteria within the everyday functioning of citizenship bureaucracies.

In this article, I complexify further the implementation dimension of the integrationist wave. I explore the data available for the situation before citizenship tests became a fast-spreading policy between the late 1990s and the 2000s. Often, citizenship laws already had less systematised requirements regarding knowledge of the national language(s), history, and culture or of different forms of cultural conformity, such as the ‘assimilation’ evaluation in France (Hajjat, 2012). Accordingly, I examine the degree to which earlier naturalisation practices entailed ideas about applicants’ suitability identifiable in the integrationist wave, as well as the potential continuities between the two moments. Extensive discretion is the norm in naturalisation procedures. This is because states usually wish to maintain the power to filter out applicants even when they meet the formal requirements. From this point of view, I explore to what degree the formal integration requirements could have removed some of the discretion available to state officers in assessing citizenship applications¹. Finally, I explore how, in some cases, this process has involved some of this discretion being outsourced to other public and private actors.

The role of standardised integration requirements in reducing discretion has been occasionally described when introducing a specific policy but has not, to my knowledge, been the object of significant literature (but see van Oers, 2013). Reducing discretion can theoretically improve, to some degree, the ability of migrants to naturalise. However, in the remainder of this article, I show that this has not always been the case and that the additional barriers erected by formalised integration requirements likely outweighed the impact of reduced discretion.

In the following pages, I first discuss the integration requirements at the centre of the integrationist wave in more detail and how common these requirements are in Europe. Subsequently, I present the methodology of my original ethnographic research in Belgium and the UK and of the meta-analysis of the other existing ethnographic studies, and I discuss the different kind of data available. Discussing each of these will help to understand the degree to which (dis) continuities in terms of cultural conformity requirements and discretion might have characterised the integrationist wave. I then discuss the results of my meta-analysis and my original ethnographic research in relation to the questions and topics highlighted in the introduction.

The role of integration requirements in wider naturalisation legislation tendencies

Among naturalisation requirements, those based on language, culture and values are at the centre of integrationism. The focus on these requirements is apparent through the introduction of citizenship tests (the type of policy that has attracted the most attention – cf. Goodman, 2010; van Oers et al., 2010b), mandatory integration courses, and standardised language or education certificates (Sredanovic et al., 2018). Employment is also closely associated with integrationism, if only because mandatory integration courses commonly include ‘employability’ components (Sredanovic et al., 2018). Residence, criminal records and fees are less linked to integrationism. However, naturalisation fees have been recognised as a form of indirect selection based on income or self-support (Stadlmair, 2018). Previous studies have suggested the presence of opposite tendencies in the broader transformation of citizenship policies. On the one hand, citizenship has been de-ethnicised by introducing elements of ius soli and removing some obstacles to naturalisation. On the other, it has been re-ethnicised through measures of ius sanguinis for descendants of citizens abroad (Joppke, 2003). Inclusive measures such as the toleration of dual citizenship, gender equality, and avoidance of statelessness have been widely introduced alongside excluding measures such as integration requirements (Vink & de Groot, 2010). Consequently, in this article, I explore whether the changes in naturalisation practices are linked specifically to the integrationist turn or, more generally, to changes in attitudes towards naturalisation.

The present analysis focuses on Europe, where integrationist requirements are very common. In 2022, the GLOBALCIT (Global Citizenship Observatory) database categorised 20 European countries (out of 43) as requiring formal tests or certificates to prove language or civic knowledge. A further 20 have language, cultural or civic knowledge requirements without a formal test or certificate. Only three European countries – Ireland, Serbia and Sweden – currently have no requirements of this kind. Of these, the government elected in Sweden in 2022 has been working on introducing citizenship tests. The reasons for such widespread tendency can be identified in the political context, policy imitation, and path dependency. Most integrationist policies were initiated following the attacks on the World Trade Center and Pentagon in 2001 (Sredanovic et al., 2018), and anxieties around Muslim residents have often dwelt in the background of such reforms. Another key motivation has been a backlash against multiculturalism, even in countries where no real national multiculturalist policy ever existed (Vertovec & Wessendorf, 2010). Governments of all colours have tended to introduce or reinforce such requirements, rarely abolishing them (Sredanovic, 2016). While an EU policy for formal integration tests is not in place, the 2003 Directive on the status of third-country nationals who are long-term residents has created a framework to link permanent residence to integration requirements (Groenendijk 2004; Pascouau, 2018).

In this context, migrants across Europe are subject to barriers in accessing full membership and rights and are required to demonstrate requirements that not all natural-born citizens necessarily hold. Moreover, the political space in which a different approach to citizenship can be conceived seems limited by an almost pan-European consensus on integrationism.

Methods

This article is organised as a qualitative meta-analysis (cf. Timulak, 2013) of the existing ethnographic data on the implementation of citizenship policies. The analysis covers the UK, Belgium, France, Germany, Italy, the Netherlands and Switzerland. I use original ethnographic data I collected in the UK and Belgium. I further use published ethnographic accounts, both to integrate the knowledge about these two countries, and to obtain knowledge about the other five countries included in the analysis.

Ethnographic accounts could be considered a ‘gold standard’ for understanding procedures in detail. By observing routines and eliciting knowledge and criteria from the bureaucrats, such accounts can offer an understanding of the procedures’ everyday workings and (potentially) their informal and tacit criteria. In this sense, ethnographic data enable more in-depth analyses than practices-on-paper data, which for example are the main kind of data used by indices such as the Citizenship Implementation Indicators (CITIMP – see Huddleston, 2013). While my focus is on ethnographic data, other kinds of data can also help open the ‘black box’ of procedures, to extend research beyond the scope of the letter of the law and regulations and statistical outcomes of policies. Court decisions can give insight into what criteria were applied beyond the letter of the law, revealing additional criteria and justifications that remain tacit in everyday practice. Examining these decisions, however, raises the question of whether the arguments presented in court by institutions are ex-post justifications. Further, the decisions that reach the courts are limited by the level of judicial review offered by the law and by the kind of cases that tend to make it to court. As naturalisation is rarely codified as entitlement, the judicial review tends to be limited. Archival data can offer further retrospective insight, again with the limitation that practices not explicated on paper are less likely to leave traces in archives.

Ethnographic data capture practices not committed to paper. These, in turn, capture most of the specific contribution of ‘street level’ bureaucrats to policymaking: the practical routines that make policies possible within organisational constraints while transforming the policies themselves (Lipsky, 1980). As these practices are often only partly standardised, qualitative approaches that allow for a thicker description from the officers are particularly well-suited.

Some further challenges that should be considered are not linked to data availability but to the fact that decisions on naturalisations in themselves can be particularly opaque. In cases where individual applications are subject to a vote, the rationale of each voter can be particularly hard to reconstruct, even more so when the vote is secret. For example, the authors working on naturalisations based on a vote in Switzerland had to show some of the criteria indirectly by highlighting statistical discrimination against Muslim applicants (Helbling, 2008; Hainmueller & Hangartner, 2013). Even in the absence of a vote, the agents’ decisions can remain obscure to the applicants if there is no law mandating the motivation for each decision.

The original data I employed for my analysis in this article originate from a research project I conducted between 2016 and 2018 to explore the implementation of citizenship policies in Belgium and the UK. The research was based on qualitative interviews. In Belgium, the project included interviews with 23 civil registers, 7 parquets (local offices of the public prosecution, which also decide on applications for the acquisition of nationality), and 6 institutions helping with nationality applications. In the UK, I interviewed 7 Nationality Checking Services (services that local authorities had the option to activate to help with the citizenship application) and 17 agents of the Nationality Team of the Home Office. I further interviewed 7 officers of UK NARIC (an agency certifying non-UK university degrees, including evaluating English language proficiency for naturalisation purposes) and 1 officer of Learndirect (an agency involved in the organisation of citizenship tests)². In the UK, the centralised naturalisation bureaucracy made the selection of interviewees straightforward. In Belgium, I aimed to cover the different kinds of territorial subdivisions (linguistic communities, regions, and the judicial arrondissements covered by different parquets) that could influence the procedures. The interviews were all conducted at the offices of the different institutions. They covered the local routines, the points of law open to interpretation, the workload and local organisation of work, and the relations with other institutions (for more details, see Sredanovic, 2022a).

In addition to my original data, I present a qualitative meta-analysis of ethnographic studies of citizenship policy implementation. I identified studies through Google Scholar (using keywords such as ‘naturalisation’, ‘implementation’, ‘qualitative interviews’, and ‘ethnography’). Further studies were identified by examining the references of already identified publications and more recent publications by the authors involved and through exchanges with other researchers in the subfield. The inclusion criteria were that the study used ethnographic methods³, involved interviews with officers (although a number of studies also interviewed applicants) or the observation of procedures, that the study was conducted in Europe, and that it is available in English, French or Italian.

The more general meta-analytical approach of comparing published results from different studies to produce further knowledge is applied both to quantitative and qualitative studies. Qualitative meta-analyses, such as the one presented here, are distinct because the approach is almost always interpretative. It involves ‘filling the gaps’ between different studies to identify new insights, rather than estimating ranges of statistical outcomes (Timulak, 2013). Different approaches within qualitative meta-analysis can be identified according to the general epistemological assumptions, and the kind of examination and evaluation applied to published studies (Timulak, 2013). The approach I followed is mostly close to meta-ethnography (cf., e.g. Doyle, 2003), which involves reading and comparing existing ethnographic accounts to gain further insight into the practices described. However, I modified this approach compared to Doyle’s (2003) model – hence the choice to describe the study more generically as a ‘qualitative meta-analysis’. The modifications to the model are due to the comparative scarcity of available studies of the topic, and to a research question that looks specifically at variation over time in the functioning of naturalisation. Compared to Doyle’s (2003) model, I included all the available studies that met the criteria, rather than sampling purposively. In addition, I neither produced a complete meta-coding of the included studies nor used grounded theory. Rather, I followed a specific research question limited to discretion and its variation with the integrationist wave and a linked body of existing theories of reference.

This approach is described in detail as follows. I first familiarised myself with the content of all the studies (which, in several cases, included multiple publications). Then, across different studies or within the same study, I identified pairs of ethnographic data that could answer how discretion is involved in naturalisation before and after integrationist reforms, including formalised integration requirements. Some pairs were located chronologically before and after a reform. Others were conceptually linked to practices deriving from formal integration requirements and practices that follow earlier non-standardised evaluation principles. Table 1 presents a synthesis of the studies identified⁴.

Table 1 Ethnographic studies of citizenship acquisition procedures

As the table above shows, there is now a substantial volume of ethnographic data, allowing for a certain range of cross-country comparisons. At the same time, the field of study is still developing. Relatively few countries are covered, and some have only been the object of a single study. The time gap between the first available study (Centlivres et al., 1991) and the subsequent studies is important for this analysis, as it limited the data available to compare the practices chronologically before and after the integrationist turn of the late 1990s and the early 2000s. Below, I discuss some ways to address this gap.

Before the integrationist wave and standardised integration requirements: discretion and non-systematic evaluations of integration

The integrationist wave is usually identified as having started with the 1998 reform introducing mandatory integration courses in the Netherlands (Goodman, 2010). As mentioned above, limited ethnographic data are available to understand the practices in citizenship policies before the integrationist wave in Europe. The main resources available are the study of Centlivres and colleagues (1991) on Switzerland and a number of archival and court decision-based analyses. In addition, we can examine aspects of the naturalisation procedures that have been studied in more recent ethnographic research and are linked to an earlier period, both in chronology and approach.

In their study on Switzerland, Centlivres and colleagues (1991) offered an early and in-depth analysis of the country’s naturalisation procedures. The procedures they described appear as particularly complex and invasive. Some were linked to a practice specific to Switzerland, that of informing all the citizens about the applications for naturalisation in their municipality, with presentations of individual applicants sent at home. Until 2003, full citizens could vote on each naturalisation application in some municipalities (cf. Helbling, 2008; Hainmueller & Hangartner, 2013). In addition to this practice, Centlivres and colleagues showed how applicants were examined in depth during these procedures. For example, in some cases, officers asked to see applicants’ high school reports and bank statements. Their study thus supports a view of the earlier naturalisation procedures as based on a restrictive and sometimes highly discretional definition of the suitable candidate for naturalisation. Accordingly, the integrationist idea that the candidate should learn and adhere to the language, culture and values of the country of naturalisation does not seem to constitute an addition to previous practices (at least in the Swiss context). Rather, it seems to select and formalise some of the criteria followed. Moreover, the procedures described by Centlivres and colleagues were those conducted by officers before submitting the application to a public vote. The criteria used in such votes (even in more recent times – see Helbling, 2008; Hainmueller & Hangartner, 2013) were potentially even broader in nature and more restrictive in exigence⁵. Indeed, similarly restrictive criteria can be found in archival and court sources. Reading the press of the country of origin has historically been proof of non-assimilation in France (Hajjat, 2012). Similarly, a high level of remittances to the country of origin has been used as a criterion of non-integration in Belgium (de Jonghe & Doutrepont, 2012).

As mentioned above, one limitation of these insights is that the highly restrictive criteria identified in earlier research might not result from discretionary procedures. Rather, more generally restrictive approach to naturalisation might instead have driven the use of such criteria. Accordingly, examining the less formalised procedures that continue to be used in more recent times and the more formal requirements for naturalisation can offer further understanding and potentially help capture the impact of the procedure within an altered general approach to citizenship. Moreover, it can help to add some empirical examples to an otherwise limited literature.

The data in Kristol and Dahinden’s (2022) study offer an ideal comparison for the data presented by Centlivres and colleagues. It was conducted between 2014 and 2019, with a focus on francophone areas of Switzerland. Accordingly, these authors’ data capture the impact of three relevant reforms. In 2003, the Supreme Court limited the possibility of subjecting naturalisation applications to a vote of the municipality’s citizens. In 2007, a federal reform required authorities to motivate the rejection of naturalisation applications. Finally, in 2014, a federal reform standardised the integration criteria. By comparing the ethnographic data, some changes in the overall approach to migration can be identified. In the more recent study, the interviewed officers recognised human mobility as constitutive of society to a higher degree. Further, some of the more invasive procedures, such as demanding high school reports and bank statements, seem to have disappeared. Still, the interviews continue to demand a highly developed knowledge of local (municipal) culture. Kristol and Dahinden’s (2022) interviewees were also clear about expecting the applicants to know the area well, be involved in local civic life and commit strongly to Switzerland. The interviewees of Centlivres and colleagues justified the difficulty of naturalisation with the benefit to applicants of becoming citizens of a ‘well-functioning’ country. This idea resonates with how Kristol and Dahinden’s interviewees justified high expectations by referencing the Swiss capability for making civic life work well. Comparing these studies thus reveals that the wider changes in attitudes towards migration do not seem to avoid the occurrence of similar procedures reproducing similar exclusionary logics.

In Belgium, the main element remaining from the older procedures is the use of police interviews to verify the integration of applicants for nationality. Such interviews were at the centre of the country’s procedure until 2000 (see Bietlot et al., 2002). Although the citizenship law was changed in 2000 to consider the application itself as sufficient proof of integration, the integration requirements were reintroduced in 2012 as formal documentary proofs of linguistic, social and economic integration. The debate in Parliament leading to the 2012 reform touched on the need to avoid returning to the discretion involved in evaluating integration through police interviews (Apers, 2014). Despite this, in my research, I found that some of the local offices of the magistrates evaluating nationality applications (the ‘parquets’) collected data via police interviews. In some cases, the data remained unused, as the local court clarified that decisions based on such interviews would not be upheld in court. In others, the judgement of the police on integration has been the basis of rejections of the application (Sredanovic, 2020, 2022a). I am not aware of systematic research into the consequences of such interviews. However, in one of the interviews I conducted with a parquet, I saw a report that discussed issues such as the ‘adequate’ amount of clothing and shoes in the house of the applicant. The magistrate I was interviewing also seemed to find the report anomalous in its contents. Details such as these suggest that the use of older procedures like police interviews to verify integration can inject further subjectivity, if not arbitrariness, into evaluations.

In France, the naturalisation procedure has only partially moved away from the earlier form of discretional evaluation. The requirement of French language knowledge, originally verified through an officer interview, became in 2011 the requirement of a specific language certificate. However, prefecture officers still have the authority to evaluate the linguistic competence of the applicant, whose ‘assimilation’ is still verified through an interview. The French legislator has also refused to define ‘assimilation’ in detail. The ethnographic work of Hajjat (2010ab, 2012) and Mazouz (2017, 2019) has shown how the resulting subjective evaluation tends to target Muslims, who are also targeted by formalised integration criteria that refer to the veil and polygamy. In certain cases, ‘good candidates’ are somewhat identified through an ordered and readily examinable application (Hajjat, 2012). While economic requirements have been changed particularly frequently according to the party in power, ethnographic research shows restrictive ideas of self-sufficiency in everyday practice (Fargues, 2020).

Kristol and Dahinden (2020) further show how, in Swiss procedures for citizenship by marriage (subject only to procedures conducted by officers and not involving votes by political bodies), officers continue to show restrictive ideas on the deservingness of applicants. In addition, the marriage dimension tends to bring in prejudices based on gender and race.

However, the permanence of non-standardised procedures is not automatically a space used for discretion. The applicants with at least ten years of residence in Belgium can have the integration requirements (except for language knowledge) waived and substituted with a letter detailing their ‘involvement in the life of the community’. While this appears to be a measure that enables highly discretional evaluation, the option was treated as residual in all the interviews I conducted. Rather than examining the applicants in detail, the magistrates I met were willing to accept any form of involvement that could fit within the letter of the law.

The integrationist wave: formal barriers and a partial reduction in discretion

The changes in citizenship policies across Europe that started at the turn of the millennium have been widely analysed. As mentioned, there is ample evidence that the main motivation of the integrationist wave has been a demand for migrants to conform to ideas of national culture and values. However, compared with the highly discretional procedures that existed in many contexts even before the integrationist wave (discussed in the previous paragraph), a standardised integration requirement can reduce discretion. van Oers (2013), for example, observes how the introduction of national tests in the Netherlands and Germany assured a degree of uniformity. She also observes how in Germany it avoided (at least somewhat) previous plans in Baden-Württemberg and Hesse. The latter were openly targeting Muslim candidates and included topics regarding the candidate’s beliefs and opinions. More generally, formal requirements can remove some of the most arbitrary decisions. This particularly occurs in contexts with some subjective evaluation of the applicants, whether mandated by policy or introduced through the agents’ discretion. However, policy uniformity might still trend in a restrictive direction, simultaneously barring applicants from naturalisation whom the agents would have considered deserving, and introducing obstacles for all. Thus, although the Swiss law currently explicitly indicates that integration evaluations should be tailored to the capacities of the applicant, Kristol and Dahinden (2022) observe how this could mean further discretion rather than more equitable procedures. By contrast, van Oers (2013) reports how some of the Dutch officers she interviewed lamented that formalised requirements barred from naturalisation profiles of applicants they thought deserved exceptions. I found similar discourses among some of my interviewees in Belgium. Moreover, in contexts with limited or no subjective evaluation, there was potentially little arbitrariness to rein in. For example, van Oers (2013) reports how, before the introduction of formal citizenship tests, language knowledge was so rarely evaluated in the UK that it hardly influenced citizenship policy.

My fieldwork in Belgium and the UK showed one interesting consequence of the introduction of formal integration requirements for nationality obtention. In Belgium, this took the form of a ‘documentary approach’ (Wautelet, 2014) in which social, economic and linguistic integration must be demonstrated through different possible kinds of documentation. In the UK, the requirements are linked to a more classical combination of a language and citizenship test. The overall effect of introducing formal integration requirements was to install further obstacles in contexts where nationality was relatively accessible before the integrationist reform. This was especially the case for Belgium: the country had allowed individuals to apply for discretional naturalisation after three years and for as-of-right nationality after seven years of residence without much further requirements. However, neither the magistrates in Belgium nor the agents of the Home Office in the UK showed much will to intervene on questions of integration despite the integrationist orientation of the laws. In my interviews with magistrates in Belgium, there were occasional mentions of notions of integration and, in very specific cases, the will to go beyond the letter of the law. Some magistrates were, for example, willing to waive the requirement to prove the knowledge of one of the national languages for French, Dutch or German citizens without a high school diploma (Sredanovic, 2018, 2023). However, their overall approach was to follow the letter of the requirements. This also due to the combination of the letter of the law leaving little margin for discretion on one hand and access to judiciary review on the other (Sredanovic, 2022a). In the UK, the Home Office has much more discretion in decisions on individual applications. However, in my interviews with the agents, most of the focus was on identifying potential fraud, with little interest shown towards concepts of integration (Sredanovic, 2022a, 2023).

My data for Belgium and the UK further show the degree to which formal requirements can limit subjective evaluation. A number of interviewees I met in Belgium referenced candidates who had limited command of the national languages but fulfilled the language requirement by having been employed without interruption in the last five years. While some interviewees showed discomfort with such situations, they still prioritised the letter of the law over their evaluation (Sredanovic, 2022b, 2023). In the UK, the English language test results were observed more closely due to concerns of potential fraud, also motivated by broader anxieties around fraud in that area (cf. Harding et al., 2020). However, the hypothetical case of an applicant without competence in the English language but with a successful, fraud-free test was explicitly recognised by some of my interviewees as someone for whom there was no need to deny naturalisation.

Notably, the degree to which policy has shifted to formal tests can sometimes be ambiguous. In Italy, the transition to online applications during the COVID-19 pandemic has made space for the introduction of a questionnaire for all applicants that substitutes the interviews conducted with some of the applicants. This questionnaire focuses mostly on employment, taxpaying and respect for the law. Trucco (2022) observes how this comes close to having a formal citizenship test. In Switzerland, each municipality has some discretion in the procedures for naturalisation, including the content of citizenship interviews. Some municipalities have introduced standard lists of questions for all applicants, while others only include lists of topics that can be covered (Kristol & Dahinden, 2022). The local list of questions used in Fribourg and reproduced by Kristol and Dahinden (2022) also includes cultural questions, such as those about strictly local matters. Comparing the approaches taken in Italy and Switzerland, van Oers’ observation that national, official tests might reduce arbitrariness seems pertinent. This not only due to the variation between individual municipalities, but also due to the more restrictive and culturalist approach taken in Switzerland.

Finally, the removal of discretionary power from officers can result in those officers expressing their idea of the candidates’ deservingness in other domains. Badenhoop (2023) analyses the impact of the 2005 reform that made naturalisation an entitlement rather than a discretionary concession for applicants who meet all requirements in Germany. After the reform, some officers have shifted to informally encouraging or discouraging applicants when giving information about the procedures, according to the officers’ notions about the deservingness of each applicant. Other officers began – again informally – discussing the meaning of citizenship when conferring the certificate.

Generally, the data suggest that integrationist policies have introduced new barriers to citizenship even as they often elaborate on criteria that were already followed before. Integrationist policies seem to have limited some of the discretion and left significant spaces of discretion in some cases. This occurs mostly through variations in the actual implementation of the policies, either between different local branches in the absence of centralised procedures or between different individual officers.

The retreat of the state? The externalisation of evaluation and discretion

Beyond the integrationist wave, there are signs of a potential further change in the orientation of citizenship policies across Europe. Some integrationist policies, such as mandatory integration courses, can become relatively costly for the state. This especially if they are offered free of charge to migrants. Such policies do respond to an increasingly entrenched ideological orientation towards migrations, but governments can be tempted to cut the linked costs. Accordingly, some policy changes suggest not so much a change in the overall integrationist ideology but a retreat of the state from offering integration courses and, in some cases, from examining integration. The result is a partial transfer of discretion and authority to other actors, including private ones. The Netherlands have been a prototypical case. There, the policy shifted from voluntary, free integration courses to mandatory but free courses and then to mandatory courses delegated to the market, with the costs left for the migrants to pay (Bonjour, 2018). Similarly, the 2002 citizenship reform in the UK included the option of fulfilling the integration requirement by attending an ‘English for Speakers of Other Languages’ course. However, this option was scrapped in 2013, leaving only the tests. As mentioned previously, France’s 2011 introduction of a requirement to prove French language knowledge through documentation reduced the evaluation activity of prefectural agents.

These changes in citizenship policy can be understood as a strengthening of the neoliberal component of integrationism over the cultural assimilationist component (see Schinkel & van Houdt, 2010). In the Netherlands, in particular, the end of the state’s provision and funding of such courses had strong links to neoliberal political orientations (Suvarierol & Kirk, 2015; Bonjour, 2018).

The retreat of states from integration policies linked to citizenship carries another implication. Each time an integration requirement is outsourced, public authorities surrender part of their discretional power, as (potential) subjective evaluation is usually substituted by documentation. Part of such outsourcing has a history extending back well before the integrationist wave. Whenever an academic degree, for example, is considered proof of integration, the decision on whether the applicant fulfils an integration criterion is made by the institution that emits the degree. In the larger context of the integrationist wave described here, this retreat means an additional reduction in the discretion available to the citizenship authorities. The already mentioned ‘documentary system’ introduced in Belgium by the 2012 reform is a good example of this type of evolution. The system indeed has limited space for subjective evaluations of the applicants by the magistrates (Sredanovic, 2018, 2022a).

Clearly, there is also the issue of how much subjectivity goes into evaluating the documents presented rather than the applicants themselves (e.g. van Oers, 2013 on the evaluation of diplomas to prove language knowledge in Germany). In my interviews in Belgium, this issue was mentioned by the municipalities rather than by the magistrates. A few interviewees working in the municipalities discussed how they preferred to build a convincing dossier with the applicant⁶, or how they could consider some jobs to be more linked to the knowledge of one of the national languages than others. This despite there being nothing in the laws and regulations that gives space for such distinctions (Sredanovic, 2020, 2022a). By contrast, the magistrates working in the parquets were rather uninterested in distinguishing between different kinds of documentation. The doubts they discussed were about the letter of the law and what it meant rather than whether some kinds of proof were preferable to others.

In the UK, my interviews showed a closer examination of the documentation at the Home Office, but this was mostly linked to fears of fraud (cf. Fortier, 2021; Badenhoop, 2023). This had some potential consequences: a document different from those the agents were used to could be suspected of fraud. Given the wide margin of appreciation available to the Home Office, it could theoretically result in a rejection of the application. Further, the agents I met at the Home Office mentioned how candidates with, for example, degrees from a university that appeared ‘too good’ for their profile could be subject to special attention for possible fraud. However, even cases such as these are different from a more discretional evaluation of an applicant as deserving or not.

Clearly, outsourcing the requirements does not mean a reduction of discretion in itself but merely that it has been transferred from state agents to other institutions. Degree-granting educational institutions, the organisers of language and integration courses, and other figures (e.g. employers) are free to exercise an even broader discretion in determining indirectly whether someone will qualify for citizenship. The literature includes highly critical accounts of, for example, private integration course providers (Suvareriol and Kirk 2015)⁷.

Discussion: variable discretion and concepts of integration

As I anticipated in the introduction, my analysis is compatible with the existing assessment of the integrationist wave in that it was motivated by the demand for cultural conformity and that it introduced new obstacles to access to citizenship. However, there is another layer to consider: the degree of discretion involved in the different approaches. In addition, while states are nowhere close to having surrendered their discretional authority on citizenship procedures, some of the discretion, decisions and services linked to naturalisation have been transferred to other actors. I have shown examples of this from the Netherlands, UK and Belgium. I have also highlighted, however, that there is ethnographic evidence from the UK, Belgium and Germany that officers can move from discretionally evaluating the applicants to discretionally evaluating some of the external documentation submitted.

While I have discussed elsewhere (Sredanovic, 2022a) in detail my ethnographic data from Belgium and the UK, some details can help illustrate the points discussed across this article. First, in both the countries the new formal integration requirements were built on previous, less codified ones. However, as discussed such less codified requirements were minimal in the UK (a rarely used examination of the knowledge of English) and more developed in Belgium (in particular through police interviews until 2000). Second, the integrationist wave has brought codified integration requirements in both countries, expressed through standard tests in the UK and through a highly codified law in Belgium. As discussed, the examination of integration itself is characterised by low discretion in both countries (Sredanovic, 2022a). However, extensive discretion is available to and used by the Home Office in the UK, especially for what concerns fraud suspicions. In Belgium, while the discretion available to each magistrate is limited, the way in which the law is interpreted and applied changes a lot across the territory (Sredanovic, 2020, 2022a). Finally, in both cases part of the citizenship attribution process has been externalised. In Belgium this involves for example the integration courses, which are often offered by local NGOs, with variations in the content of the course (Sredanovic, 2022a), even if all recognised courses are proof of social and linguistic integration. In the UK, on the other hand, only the material administration of standardised tests has been externalised. Further, when test centres have been suspected of colluding to pass candidates without a sufficient knowledge of English, the reaction of the Home Office has been to increase drastically the controls on the process (Sredanovic, 2022a).

A further question concerns the implications of how the degree of discretion has changed. In implementing policies, some discretion is generally required and often positive (e.g. Pratt, 1999). However, any degree of discretion tends to be filled by criteria of deservingness beyond those defined in policy (e.g. Lipsky, 1980). As I have discussed elsewhere (Sredanovic, 2020, 2022a), citizenship is a relatively simple policy domain that does not require as much discretion, compared for example to many social policies. Issues linked to democratic equality are also particularly important in ensuring equal opportunities in access to citizenship. There are clear alternatives to standardised tests and documentary requirements to this end. These include moving away from integration as a vague policy concept and as a pre-condition to full rights. This can be accomplished by anchoring citizenship access to other criteria that are less difficult to determine and require less evaluation, such as the length of residence.

Conclusions and possible further lines of study

In this work, I have discussed how the integrationist wave and its emphasis on standardised integration requirements could have resulted in, among other consequences, the reduction of discretion in naturalisation procedures. With the corpus of data currently available, there is evidence that restrictive ideas about acceptable candidates for citizenship existed before formal integration requirements were introduced. Such formal requirements did also reduce the space for discretion, albeit not uniformly.

To demonstrate the existence of highly restrictive ideas of assimilation in pre-integrationist naturalisation procedures, I referred to ethnographic data from Switzerland, archival data from France, and court data from the UK and Belgium. More recent ethnographic data from Switzerland show a relative continuity in the ideas about ‘worthy’ applicants. Moreover, ethnographic data from the Netherlands and Germany and my ethnographic data from Belgium show that integrationist standardised requirements tend to introduce new barriers and limit officer discretion. Data from France reveal how a more formal requirement can coexist with the continuation of discretionary evaluation by officers. Finally, my data for the UK show how even a combination of discretionary and standardised integrationist requirements does not necessarily result in the use of the discretion available nor an interest in integration on the officers’ part.

Increasingly, the growth of implementation-focused migration policy studies underscores the importance of bureaucratic routines beyond the letter of the law as well as their importance to the extensively debated integrationist measures in citizenship policies. From this point of view, two lines of research appear promising. First, ethnographic research on a larger number of countries should allow us to go beyond the focus on a small number of countries where we have ethnographic data about citizenship procedures. Second, archival research into bureaucratic practices before the late 1990s should allow us to put recent transformations in focus. In particular it would allow us to better trace the continuities and discontinuities between ‘pre-integrationist’ and integrationist naturalisation criteria.