Introduction
The European Union (EU) has centralised much of its regulatory policymaking in Brussels. Nevertheless, it relies on national administrative actors to put EU rules into practice (Knill and Lenschow Reference Knill and Lenschow2005, 583). Consequently, national administrators are increasingly confronted with transposition laws and EU jurisprudence besides national laws and administrative guidelines. This leads to the question: which legal authority do implementers eventually follow when there is ambiguity between national and EU legislation?
EU implementation studies only rarely address this question. Instead, the EU implementation literature has focussed mainly on EU transposition (Mastenbroek Reference Mastenbroek2005; Angelova et al. Reference Angelova, Dannwolf and König2012; Treib Reference Treib2014). This literature featured national administrations as a source of ineffective transposition (Knill and Lenschow Reference Knill and Lenschow1998; Kaeding Reference Kaeding2006; Falkner et al. Reference Falkner, Hartlapp and Treib2007; Treib Reference Treib2014). Similarly, the limited research on application of EU law has considered administrators as the cause for ineffective EU implementation (Hille and Knill Reference Hille and Knill2006; Versluis Reference Versluis2007; Dimitrova and Steunenberg Reference Dimitrova and Steunenberg2013).
Studies on national administrations in the EU context (Egeberg Reference Egeberg2008; Trondal Reference Trondal2011; Wockelberg Reference Wockelberg2014) provide an alternative view. They argue that under certain conditions, national administrations might turn into double-hatted agents who follow both national and EU masters (Egeberg and Trondal Reference Egeberg and Trondal2009). A precondition often assumed for double-hatted executive behaviour is that national administrators have a direct socialising contact with EU institutions. This assumption has led to a focus on higher civil servants who are in direct contact with EU institutions (but see Wockelberg Reference Wockelberg2014).
Consequently, EU implementation research has, so far, largely neglected the role of frontline implementers (Conant Reference Conant2012, 30). Nevertheless, the frontline of EU implementation is a particularly important level of implementation because it determines whether and to what extent EU laws reach their intended recipients. Through frontline implementation, EU law may become meaningful, even when it is not fully transposed. In turn, EU law remains meaningless if implementers do not apply it on the ground.
Tackling the gap of frontline EU implementation studies, Dörrenbächer (Reference Dörrenbächer2017) has shown considerable individual-level variation among frontline implementers’ use of EU law. However, her study investigated a single country, and did not discuss the role of bureaucratic structure for frontline uses of EU law. Consequently, this study investigates how bureaucratic structures condition the use of EU law among national frontline implementers.
In order to tackle this question, the study relies on organisational theories. Organisational theories expect that human rationality is institutionalised, embedded and contextualised (Simon Reference Simon1965; Egeberg Reference Egeberg1999, 159). Organisational variables constitute systematic biases that regulate, constitute and construct political and administrative decisionmaking (Trondal Reference Trondal2011; Henökl and Trondal Reference Henökl and Trondal2015). Following this perspective, organisational structures also condition executive behaviour towards EU law (Trondal and Veggeland Reference Trondal and Veggeland2003; Egeberg Reference Egeberg2008; Trondal Reference Trondal2011; Wockelberg Reference Wockelberg2014). Applying organisational theories to the frontline, this study expects that implementers in discretion-constrained structures use EU law only when they receive filtered instructions from national political and administrative higher-ups. By contrast, in discretionary structures frontline implementers will act as independent experts relying on EU and national law as they see fit.
By focussing on the frontline of EU implementation, this study strives to enhance our understanding of EU implementation beyond the predominant focus on EU transposition and higher-level administrators. Moreover, by applying an organisational perspective, this study responds to the appeal by Hupe and Buffat (Reference Hupe and Buffat2014, 549) to move frontline studies beyond individual-level variation. Frontline research often ignores macroinstitutional factors and comparative research is rare. EU implementation provides a particularly suitable context to start filling this gap because the same EU legal stimuli confront implementers across bureaucratic structures.
To investigate the theoretical expectation, the study draws on interview data of 42 Dutch and German migration law implementers. The field of migration fulfils the EU-level scope conditions of a highly discretionary policy field in which national frontline implementers are regularly confronted with tension between legal levels. The Netherlands and Germany share a range of background factors such as a developed national migration law, similar influx of migrants and a legalistic administrative culture. However, in the migration sector, bureaucratic structures differ considerably. The Dutch implementation structure limits frontline discretion, whereas the German structure leaves considerable discretion.
The analysis reveals that both Dutch and German frontline implementers are aware of the origins of the laws they apply. Nevertheless, they use original EU law and EU jurisprudence in highly diverse ways. Dutch frontline implementers use EU law only when it is filtered through the national ministry or their agency’s policy department. By contrast, the majority of German respondents use EU law as a parallel legal order alongside national and Länder law. Thus, German respondents sometimes consult and apply original EU law and jurisprudence along with their professional expertise. However, factors at the individual and organisational-level influence how comfortable the German respondents feel in acting as independent experts.
Theory section: an organisational perspective on frontline uses of EU law
Frontline implementers are public workers who carry out and enforce actions required by laws and public policies (Meyers and Vorsanger Reference Meyers and Vorsanger2003, 154). They handle individual transactions between the law and its target groups by deciding on the distribution of rights and benefits such as unemployment benefits, tax refunds and residence permits. Although final decisions are formally handed down by the executive agency, in practice it is the individual frontline implementer who decides how laws are interpreted in relation to real cases (Bovens and Zouridis Reference Bovens and Zouridis2002). Frontline implementers are formally not included in policymaking. As such, they are also not involved in EU policymaking or transposition. During implementation, they seldom have direct contact with EU institutions and the national political level, as opposed to higher civil servants. Consequently, when the responsible national ministries and higher administrators fully incorporate EU law and jurisprudence into national law, frontline implementers play no pronounced role in EU decisionmaking and implementation.
Nevertheless, taking the findings of the EU transposition literature seriously, one can assume that national legislators and administrators do not always transpose EU directives in time or fully into national law (Mastenbroek Reference Mastenbroek2005; Steunenberg and Toshkov Reference Steunenberg and Toshkov2009; Treib Reference Treib2014). Moreover, administrative guidelines are not always adjusted to EU law, or Court of Justice of the European Union (CJEU) rulings challenge national instructions, making EU jurisprudence directly applicable. These situations can create ambiguities between European and national legal guidelines that trickle down to the frontline of implementation (see Dörrenbächer Reference Dörrenbächer2017).
To investigate how frontline implementers use EU law across administrative systems, this study relies on organisational theory. Organisational theory builds on the assumption that formal structures create systematic biases in public policy because they provide cognitive and normative shortcuts and categories that simplify and guide behaviour (Egeberg Reference Egeberg1999). In other words, organisations provide cognitive maps that reduce transaction costs by providing cues for appropriate decisions (March and Olsen Reference March and Olsen1998; Trondal Reference Trondal2011).
Previous studies have shown that organisational theory proves useful in explaining executive behaviour in multilevel administrative systems (Egeberg Reference Egeberg2008; Trondal Reference Trondal2011). For example, Trondal and Veggeland (Reference Trondal and Veggeland2003) and Beyers and Trondal (Reference Beyers and Trondal2004) have shown how national administrative structures influence the extent to which national civil servants act upon national political interest or professional expertise when negotiating EU law.
Applied to domestic implementation of EU policies, Egeberg (Reference Egeberg2008) has argued that the more fragmented national ministerial agencies are, the more the Commission can steer administrative behaviour. By contrast, bureaucrats in tightly coupled agencies with limited autonomy continue to follow national ministerial signals when implementing EU policies. Similarly, Trondal (Reference Trondal2011) finds that domestic administrative structures such as vertical and horizontal specialisation condition executive behaviour towards EU law.
Wockelberg (Reference Wockelberg2014) has explored these mechanisms further by investigating the role conceptions of civil servants in the later stage of EU implementations. She has argued that a certain level of autonomy may be necessary for national civil servants to see themselves as EU servants and to make implementation choices accordingly. Although she found limited support for this hypothesis, her results show that national civil servants with limited autonomy consider themselves primarily as national servants, and civil servants in fragmented agencies consider themselves as independent experts when implementing EU law.
Applying these insights to the frontline of implementation, one can make a very similar argument, namely that the level of frontline discretion determines the use of EU law. As Bach and Jann (Reference Bach and Jann2010) and Christensen and Laegreid (Reference Christensen and Laegreid2009) have pointed out, frontline implementers operate under diverse structural discretion. As a result, the degree of discretion granted to frontline bureaucracies influences how implementers experience their legal environment and how they apply the law (Hupe and Buffat Reference Hupe and Buffat2014).
One can identify three structural dimensions that capture frontline discretion: (1) the number of national principals, (2) the monitoring relationship between national principals and frontline organisation and (3) internal frontline organisational controls. Starting with the number of national principals, multiprincipal settings encourage competition between principals who may steer in different directions (Dehousse 2008; Whitford Reference Whitford2008). This forces frontline implementers to decide which principal they follow in their everyday decisionmaking (May and Winter Reference May and Winter2009). The number of national principals differs particularly between central and federal structures. Centralised structures confront frontline implementers with a single rule-making principal. By contrast, in federal arrangements there are central and local principals. Consequently, Whitford (Reference Whitford2002) holds that decentralised implementation results in a net loss of control over implementation and frontline discretion.
Besides the number of principals, the monitoring relationship between the political principal(s) and the frontline organisation designates the extent of discretion implementers have. In tightly monitored systems, political principals rely on institutionalised oversight mechanisms such as budgetary controls and quality checks (McCubbins et al. Reference McCubbins, Noll and Weingast1987; Huber et al. Reference Huber, Shipan and Pfahler2001, 334–335). Distribution of such mechanisms across principals or a lack of institutionalisation results in discretion.
Finally, internal organisational controls limit frontline discretion (Lipsky Reference Lipsky1980). For example, the existence of explicit production targets, strict divisions of frontline labour and centralised flows of information limit the discretionary space of the frontline implementers. Similarly, the replacement of client contact with fixed computer manuals can limit frontline discretion (Bovens and Zouridis Reference Bovens and Zouridis2002). Table 1 summarises the characteristics of discretionary and nondiscretionary frontline structures.
Table 1 Nondiscretionary and discretionary structures
How do differences in structural discretion relate to the use of EU law? On the basis of the discussion above, it can be expected that frontline implementers who experience control from only one national political principal are not used to legal stimuli outside the central national legislation. In line with the principal-agent theory, national parent ministries as direct principals control the implementing agent (Epstein and O’Halloran Reference Epstein and O’Halloran1994; Huber and Shipan Reference Huber and Shipan2002; McCubbins and Schwartz Reference McCubbins and Schwartz2009). This formal relationship suggests that implementers draw on national law as the single centre of executive authority without relying independently on EU law (Trondal Reference Trondal2011). In combination with tight national oversight and organisational control, a single national principal will discourage implementers from relying on original EU law. Consequently, implementers may not be aware of the origins of the laws they apply. Furthermore, under legal tension such frontline implementers will not consult original EU directives or CJEU rulings. Instead, they will prioritise national political and administrative instructions. Among higher civil servants, Trondal (Reference Trondal2011) labels such administrative behaviour ministry-driven. Similarly, Trondal and Veggeland (Reference Trondal and Veggeland2003) refer to administrators who typically give priority to national interests as government representatives. Applied to the frontline, this leads to the first hypothesis:
Hypothesis 1: Frontline implementers who operate in a nondiscretionary structure do not use EU law independently of national law and instructions.
Next, implementers who operate in discretionary structures may be used to varying legal stimuli. Limited national and organisational monitoring and multiple principals leave frontline implementers with discretion to use the different legal authorities alongside each other. Thus, in discretionary structures implementers may use EU law parallel to national law by developing strategies to cope with legal tension and ambiguity in line with their professional norms (Christensen Reference Christensen1991; Wockelberg Reference Wockelberg2014) and capacity limitations (Lipsky Reference Lipsky1980; Tummers et al. Reference Tummers, Bekkers, Vink and Musheno2015). Trondal (Reference Trondal2011) calls such administrative behaviour compound behaviour. It implies that implementers are aware of the different legal sources and use their professional expertise when consulting original EU law alongside national law and administrative instructions. Thus, the following can be expected:
Hypothesis 2: Frontline implementers who operate in a discretionary structure use EU law independently and alongside national law and instructions.
Finally, Trondal (Reference Trondal2011) identifies a third executive behaviour towards EU law, namely: Commission-driven behaviour. This behaviour captures national administrators who act as supranational actors (Trondal and Veggeland Reference Trondal and Veggeland2003). These actors may bypass domestic government by fully shifting their accountabilities to the EU level. However, studies have shown that Commission-driven behaviour is rare (Trondal Reference Trondal2011; Wockelberg Reference Wockelberg2014). Frontline implementers are even less likely to engage in such behaviour than higher-level administrators are because they lack the direct contact with EU institutions. Thus, direct steering by EU institutions and socialisation effects are unlikely. Table 2 summarises the expected uses of EU law at the frontline of implementation.
Table 2 Conceptualising frontline uses of European Union (EU) law
As Egeberg (Reference Egeberg1999, 161) has stressed, administrative structures are not the only factor that affect eventual decisions. Organisational theories do not preclude that individual-level variation, sectoral and cultural aspects influence frontline implementation. Bureaucratic structures constitute merely a systemic bias that conditions the decision-making process. Without neglecting that there may be complementing explanations, this study investigates how this systematic bias affects EU implementation at the frontline.
Moreover, frontline discretion may not only result from organisational structures, instead, national principals may also deliberately delegate discretion to implementers (Epstein and O’Halloran Reference Epstein and O’Halloran1994; Huber and Shipan Reference Huber and Shipan2002). For the purpose of this study, the policy field of migration is selected, in which national principals typically aim to stay in control of EU law interpretation. Although Dörrenbächer and Mastenbroek (Reference Dörrenbächer and Mastenbroekin press) have shown that national transposition actors also sometimes delegate EU discretion down to practical implementers, they highlight that transposition actors typically do so to maintain administrative and legal structures. In the field of migration, discretion is typically not delegated to explicitly encourage frontline implementers to circumvent national instructions or to leave interpretations to the frontline.
EU-level scope conditions: the field of EU migration law
To investigate the hypotheses, EU-level factors need to remain constant and fulfil some scope conditions. The first condition is that a multilevel legal context should exist with legal ambiguities between the EU and national law. Second, EU-level institutions need to have limited power in implementing EU law and in directly steering national implementers. Finally, frontline implementers across bureaucratic structures need to determine similar issues with comparable levels of tension between national and EU law.
This study investigates frontline use of EU law in the field of migration. Migration is not only a highly topical field of EU law but also provides for the EU-level scope conditions. First, migration is a relatively newly harmonised policy field (Dörrenbächer Reference Dörrenbächer2017). Most EU legislation only provides minimum criteria and leaves considerable discretion to national legislators and implementers (Hartmann Reference Hartmann2016). Contrary to regulations that are directly applicable, the most important EU legal instruments in the field of migration are directives that need to be incorporated into national legislation. National transposition actors are often reluctant to adjust national laws to EU obligations, which can lead to legal tension between legal levels (Dörrenbächer et al. Reference Dörrenbächer, Mastenbroek and Toshkov2015). In turn, insufficient transposition and vague EU rules have provoked a large body of CJEU rulings that annul or challenge national laws. In sum, these characteristics create ambiguities between national rules and vague EU rules that may trickle down to the frontline.
With respect to the second scope condition, the EU has no direct control tools to steer frontline migration implementation. Although the EU established supranational agencies such as FRONTEX and the European Asylum Support Office to assist member states with external borders in their day-to-day border controls, the distribution of residence permits and visas remains the task of national and local administrations. Domestic migration offices operate at a distance to EU institutions and require considerable national frontline activities (Christensen and Laegreid Reference Christensen and Laegreid2009; Eule Reference Eule2014).
Finally, in order to ensure comparable frontline decision-making contexts with similar levels of legal tension between national and EU law, this study focusses on family migration. Family migration is one of the most common reasons for migration in the EU (Kofman Reference Kofman2004). The most prominent EU legal instrument is the Family Reunification Directive that regulates third-country family reunification (Strik et al. Reference Strik, De Hart and Nissen2013). Other relevant EU norms for family migration are the Turkey Association Agreement and Charter of Fundamental Rights. Across bureaucratic structures, frontline implementers evaluate similar conditions for family visa and residence permits. Moreover, the CJEU has challenged national laws of both countries included in this study, suggesting that there are comparable levels of legal tension.
Country selection
For the purpose of this study, frontline implementaters in the Netherlands and the German Bundesland of North Rhine-Westphalia (NRW) are investigated. This country selection allows controlling for a range of context factors that could influence uses of EU law. First, Germany and the Netherlands have comparable percentages of foreigners living on their territory. Second, they have well-developed national migration laws and an established implementation apparatus. Third, both countries are reluctant transposers of EU migration laws, shown by the transposition of key EU legislations such as the Family Reunification Directive or the Returns Directive (Strik et al. Reference Strik, De Hart and Nissen2013; Dörrenbächer et al. Reference Dörrenbächer, Mastenbroek and Toshkov2015). Fourth, with respect to EU law, Falkner et al. (Reference Falkner, Hartlapp and Treib2007) classified both countries into the so-called “world of domestic politics”. Following this classification, application of EU law typically runs smoothly once EU law is transposed into national law. Finally, both countries share a legalistic administrative culture in which the rule of law is deeply embedded in administrative practice (Painter and Peters Reference Painter and Peters2010, 22). With respect to family migration, the CJEU has challenged the transposition laws of both countries (e.g. Chakroun,Footnote 1 K&A Footnote 2 and Dogan Footnote 3 ).
Beyond these similarities, the two countries employ highly diverse structures for the implementation of family migration law. As will be discussed in more detail below, implementation in the Netherlands is centralised and implementation in Germany is highly fragmented, suggesting good conditions to observe differences in structural discretion. For the German case, the Bundesland of NRW was chosen. The size of NRW is comparable to the Netherlands and limiting the study to one Bundesland allows holding constant political and legal conditions at the Bundesland level.
Method and data
In order to investigate the established hypotheses, interviews with 42Footnote 4 Dutch and German migration law implementers inform the analysis. In-depth qualitative data were required because explanations and examples of how respondents use EU law in their everyday tasks are largely unquantifiable. The qualitative design allowed filling the established typology of uses of EU law with meaning, beyond predefined behavioural categories. In addition, the interview data allowed tracing the relationship between macrolevel bureaucratic structure and microlevel descriptions of individuals’ uses of EU law. In this way, the study goes beyond correlation designs by tracing the causal mechanism (George and Bennett Reference George and Bennett2006) between structural discretion and uses of EU law.
The explanatory condition – discretion in bureaucratic structure – was operationalised by interviewing four employees of the policy department (beleidsafdeling) of the Dutch migration agency and three representatives of two district governments (Bezirksregierung) in NRW. These respondents constitute the link between the political level and the frontline and provided overviews of bureaucratic structures and monitoring tools. In addition, organisational control mechanisms were discussed in the frontline interviews. The information was crosschecked with administrative documents.
To investigate frontline uses of EU law, 14 caseworkers of the Dutch migration agency and 21 German caseworkers in NRW were interviewed from spring 2015 to the beginning of 2016. In the Dutch sample, respondents from all three departments of the central migration agency, which handle family migration, were interviewed. For the German sample, respondents from 10 local foreign offices that decide on family migration residence permits and visa were interviewed.Footnote 5 The 10 offices include local offices in large and small cities as well as on the countryside.
The number of German respondents was slightly higher because of the great variation between local offices that required additional crosschecking of the findings. Interviews lasted up to 190 minutes. In both countries, interviews were conducted individually or in groups of two to three. Besides accounting for the time constraints of the respondents, group interviews helped to determine whether responses differed when their colleagues were observing them (see Dörrenbächer Reference Dörrenbächer2017). Moreover, this data collection method had the advantage of interviewing respondents in their usual working environment where respondents typically share offices with colleagues. Another advantage of the group interviews is that they allowed observing discussions among colleagues, which helped to uncover variation in approaches to EU law. A disadvantage of the group interviews was that junior caseworkers participated less actively in the discussions than their more advanced colleagues. With respect to the reported use of EU law, the analysis revealed no systematic difference between individual and group interviews. Interviews were semistructured, conducted in Dutch and German, respectively, recordedFootnote 6 and transcribed into English.
In order to operationalise use of EU law, the interview guide included questions on the role of EU law in respondents’ daily work. This allowed determining the extent to which respondents are aware of the origin of EU law in their field of work. To investigate whether respondents consult original EU legislation, the interviewer asked respondents to describe the practices following the emergence of new CJEU rulings and EU directives. In addition, respondents were asked whether they recalled situations of legal conflict and how they handle such situations.
The strategy for coding the interviews was to sort responses systematically according to themes that correspond with the conceptualisation of the three indicators of uses of EU law: awareness, consultation of original EU legislation and giving priority to EU law (see Table 1 supporting material). Moreover, analysing the interviews allowed tracing how respondents embedded their use of EU law in their perceived discretion. The following section first discusses the two settings by characterising them as discretionary and nondiscretionary structures. This is followed by a presentation of uses of EU law among implementers in the two systems.
Structural discretion in Dutch and German migration offices
Starting with the number of principals in the Dutch system, the first aspect to note is that implementation of migration law is highly centralised. The central frontline organisation is the Immigration and Naturalization Service (IND), which is responsible for processing visa applications and residence permits. The IND stands directly under the Ministry of Security and Justice as single political principal.
In contrast to the Dutch structure, German frontline migration administrations are steered by a number of different political principals. The main frontline organisations are the Ausländerbehörden, which are subdepartments of the German counties and cities (Kreise and Kreisfreie Städte). There are 82 offices in the Bundesland of NRW (correspondence with Ministry of Internal Affairs in NRW). Three political actors can affect the activities of the Ausländerbehörden, namely the federal legislator, the Land and municipal leaders.
As regards the monitoring relationship, the Dutch case again fits into a nondiscretionary structure. The IND, as a contract agency (agentschap), receives a central state budget and a framework document lays down the hierarchical relationship with the ministry (van Thiel Reference van Thiel2001, 8). The IND policy department constitutes the link between the political level and frontline implementation in the decision-making units. When new EU legislation comes about, the departments of the parent ministry consults the IND policy department to draft the transposition laws and implementation instructions for the frontline (IND PolDep 1, 2).
By contrast, the monitoring relationships between the Ausländerbehörden and the different political principals are fragmented. Similarly to the Netherlands, new migration laws including the transposition of EU laws are the responsibility of the national level. However, the federal level has no direct tools to monitor frontline implementation. The Ministry of Interior of the Bundesland interprets and clarifies federal migration laws through decrees. For the Ausländerbehörden, implementation is a delegated task without autonomy (Pflichtaufgabe nach Weisung) meaning they have to obey Länder decrees.
In order to monitor the activities of the Ausländerbehörden, NRW delegates the subject supervisions to five district governments (Bezirksregierungen). These intermediate administrative bodies communicate between the Ausländerbehörden and the Ministry of the Land (NRW BZR1, NRW BZR2). The Bezirkregierung can be compared with the policy department of the IND. However, contrary to the policy department, the Bezirksregierungen are only connected to the Ministry of the Bundesland. In addition, they have limited sanctioning power and are less directly involved in policymaking of the ministry (NRW BZR1, NRW BZR2). The Bezirksregierung can conduct local investigations of files and respond to complaints. However, each year only one or two Ausländerbehörden per district are investigated. The current refugee crisis limits these controls further (NRW BZR1, NRW BZR2). The Bezirksregierungen have no direct sanctioning power and only report to the Ministry of the Land and the major of the relevant municipality. Municipalities and counties operate the Ausländerbehörden and incur personal and financial consequences. Thus, political monitoring is fragmented, suggesting larger margins of structural discretion for frontline migration organisations in NRW than for their Dutch counterparts.
With respect to internal organisational controls, the first aspect to note is the high division of labour within the Dutch IND. The IND employs around 3,000 employees in a hierarchical structure (IND PolDep1). At each decision-making location, there are several teams of circa 30–40 decisionmakers (IND PolDep2). Strategic and operational managers oversee the day-to-day work in the teams, but they are not directly involved in substantive implementation (IND A.2). For substantive matters, senior decisionmakers supervise new employees and link the frontline to the policy department (IND C.1).
The IND divides frontline workers into decisionmakers and practitioners. The latter prescreen visa and residence applications (IND A.7). Decisionmakers decide on the more complex cases and may issue rejections. The focus in this study is on the latter. Among employees, there are different specialisations based on the grounds for migration (e.g. family migration, highly skilled migration, labour migration) and decision-making stage (e.g. first instance and appeal decisions) (IND PolDep1, 2). Furthermore, there are specialised units for follow-up controls (traject controle), medical exceptions and hardship clauses (IND A.1). In line with a discretion-constrained organisational structure, caseworkers rarely have personal client contact, but take decisions mainly based on documents (Bovens and Zouridis Reference Bovens and Zouridis2002).
Finally, there is a fully developed system of quality checks through peer review. The IND computer system randomly selects a number of decisions per week, which are controlled by decisionmakers and seniors at other locations (IND A.1). The IND also established benchmarks of the number of hours frontline implementers have at their disposal per decision. Although these benchmarks are handled in a flexible way, they give the seniors and managers control over frontline activities (IND B.2). The IND policy department distributes implementation instructions including instructions for EU law and jurisprudence though the organisation’s intranet (IND PolDep1). Furthermore, the IND’s judicial department keeps track of all jurisdictions and offers advice to decisionmakers (IND B.1).
Coming to internal organisational control in the Ausländerbehörden, there is huge variation. Similarly to the IND, in big city offices, there is a relatively high division of labour. Subdepartments, working groups and team leaders oversee the casework (locations A, B, C). Moreover, large offices have judicial departments and research departments that screen new legislation and advice employees.
However, in small cities and counties the division of labour is much lower (location J, I, H, G). Frontline workers in these offices are all-rounders and decide on all migration-related issues of their municipality, ranging from family reunification over deportation to citizenship, etc. At the same time, the county offices have more employees in the middle civil service compared with bigger offices, which employ mainly higher civil servants for the same tasks (location I). The office leadership is flexible in organising frontline tasks. Contrary to the IND, caseworkers in NRW have daily personal client contact. There is no intranet that connects the Ausländerbehörden within the Land, let alone within Germany.
In sum, because of the close supervision by the central ministry and the internal control tools of the IND, the Dutch implementation structure can be classified as a nondiscretionary structure in which frontline implementers have limited room for manoeuvre. This makes the Dutch case a least likely case for independent use of EU law. By contrast, frontline implementers in NRW operate in a discretionary structure in which frontline implementers are less closely watched by a number of different principals. Internal organisational controls vary depending on the size of the municipality or county and office leadership. Table 3 summarises the differences in structural discretion.
Table 3 Bureaucratic structures in Dutch and German migration administrations
Note: NRW=North Rhine-Westphalia.
Beyond the differences in structural discretion, there are some differences between German and Dutch frontline implementers regarding their educational background. German frontline workers are civil servants in the Middle or Upper Intermediate Service (mittlerer or gehobere Dienst). They typically have a Diploma or Bachelor’s degree from an administrative school (Fachhochschule für öffentliche Verwaltung). The training includes legal specialisations for regulatory agencies. However, according to the respondents, the curriculum typically does not include specific trainings in migration law. The Dutch civil servants typically have a social science background but increasingly there are also caseworkers with a legal-university background. Generally, these differences in education do not point at distinct patterns towards the use of EU law. Similarly to the German caseworkers, Dutch frontline implementers usually have no prior expertise in migration law. Following the respondents, caseworkers receive training in using legal texts and for deciding on individual cases on the job through supervision from their peers or complementary seminars on specific legal topics. In both systems, frontline implementers receive typically no training specific for EU law.
Uses of EU law across implementation structures
To explore the expectations, the following section investigates use of EU law by examining awareness and consultation of EU law and practices in case of contradictions between EU and national law. Table 4 summarises the results.
Table 4 Uses of European Union (EU) law by Dutch and German migration law implementers
Note: NRW=North Rhine-Westphalia.
Awareness
Starting with the Dutch respondents, the first aspect to note is that all respondents were aware of the origins of the laws they apply. All respondents specified CJEU rulings and EU directives when outlining decision-making practices. Three respondents argued that though they were aware of EU law, it played no critical role for them. According to them, EU law is not real law but only jurisdiction (IND B.1); in other words:
For me personally it (EU law) does not play a role. Of course, I know that the laws we have (…) derived from EU law but in fact, this is now national law, so it does not really matter where this law comes from. (IND A.3)
Nevertheless, the majority of Dutch respondents (11) argued that EU law plays a major role for them. These respondents identified national laws that changed because of EU law. In addition, three respondents highlighted that EU law gave them new room for manoeuvre (IND B.3; IND B.4; IND A.5). Overall, even in nondiscretionary structures, respondents experienced an influence of EU law on their work.
As regards decisionmakers in NRW, the interviews showed that respondents of all foreign registration offices but one agreed that EU law affected their working practices. Similarly as in the Dutch case, many respondents listed substantive influences. Respondents in eight of the 10 foreign registration offices referred to EU law mainly as a source of new complexity in their work (NRW I.1; NRW J.2; NRW A.1; NRW A.2; NRW F.1; NRW F.2; NRW C.1; NRW C.2; NRW E.1; NRW B.1).
Overall, respondents in the Dutch and German systems were aware of the substantive influence of EU law on migration law. However, respondents particularly from the smaller Ausländerbehörden doubted their own expertise regarding EU law more than did the Dutch respondents:
(…) honestly, I have to say, I do not always have everything in my head. A lot of these things (EU law) just slip through. Sometimes these issues are also not really forwarded to us. (NRW J.2)
This perception of low EU expertise may limit independent use of EU law, constituting a challenge for the established expectations.
Consultation of EU law
The next indicator for different uses of EU law is whether frontline implementers ever read and consult original EU legislation and judgments. Across the two systems, the interviews showed considerable variation. Dutch respondents rarely indicated that they read original EU legislation. All respondents explained that when EU rulings emerge, they only consult the national policy instructions based on the ruling: “We only use the elements which we get from the policy department” (IND A.1). Similarly, another respondent explained: “You cannot expect from a decision maker to read all these judicial texts and then just apply it. For that you really need the specialist of the IND” (IND B.4). One of the IND frontline managers confirmed this practice as the desirable organisational practice:
(…) our employees are not hired to read independently all the jurisprudence and to interpret them on their own. This would cost way too much time. There are people hired for such things in the policy department. Our people here are hired to take decisions and not to read jurisprudence. (IND A.4)
Only when CJEU rulings come about for which there are not yet any national instructions, respondents described that they sometimes look up the original judgement. However, this only happens when migrants or their lawyers explicitly rely on new EU rulings to argue in favour of their case (IND C1; IND B.4; IND A.3; IND B.2; IND B.1). In such situations, they hold the case until they receive instructions from higher administrative levels, as explained by this respondent:
I would tell them that the ruling has so far no direct influence on our national policies. I cannot tell them what that ruling means for them. So, a decision maker would not look into the original EU law to adjust his practices. (IND C.2)
With respect to German respondents, practices were diverse. The leader of one of the larger offices argued that he would not expect his employees to look into EU rulings themselves, and therefore relied on a prescreening procedure (NRW E.1). Nevertheless, when new EU legislation or jurisdiction emerged, most respondents argued that they would first of all read the original documents or they would consult interpretations of legal journals to learn more about EU rules (e.g. NRW D.1; NRW H.1; NRW H.2; NRW A.1; NRW A.2; NRW J.2; NRW F.1; NRW F.2; NRW I.1; NRW B.1). The respondents in NRW were used to looking into different legal sources and their interpretations. However, there was variation across offices. The differences are illustrated well by a respondent who recently transferred from a bigger city office into a small county office:
In my old office, I was part of the team: legal certainty. We filtered new legislation and prepared the other caseworkers when new EU rules came about. We gave them briefings so that everyone acted upon the new rules in the same way. Here, we also get the information, but we are free if we actually read the new directives and decrees, and even if I apply them in practice. Nobody really checks that. (NRW J.2)
The quote indicates that frontline implementers in smaller offices have considerable discretion at their disposal regarding consultation of EU law.
Practices in case of legal tension
Turning to situations in which EU and national law compete with each other, five Dutch respondents explained that they have never encountered such situations. The fact that they rarely read original EU laws and do not search for interpretations of EU law outside the national administrative instructions may explain this perception. Nevertheless, the majority still recalled situations in which EU law demands different practices than national laws. In the context of family migration, respondents referred mostly to the more liberal approaches in the European Charter of Human Rights or the Turkey Association Agreement.
When asked which laws they use under legal ambiguity, Dutch responses were very homogenous. Although some respondents pointed out that EU law stands in principle above national law (IND A.1; IND A.5), none of them would independently apply EU law if it conflicted with national law. Instead, in such situations, respondents notify the policy department of their agency. A typical response was: “Nobody would start to make their own interpretations of the EU judgments. We always wait for official instructions” (IND A.4). This was justified with respondents’ own understanding of their task as implementers: “We are implementers and I have to stay within my level” (IND A.1). Other colleagues added the following:
We have a more reluctant approach. We follow the instructions that we get once a ruling or law is relevant for our department. So for example with ChakrounFootnote 7 , we got a clear guideline of what the ruling means for our daily practice. Once there are guidelines we take them of course into account in our decisionmaking. (IND A.2)
Coming to practices in NRW, responses were much more diverse. Only few respondents with little experience did not recall situations of conflict between EU and national law (NRW J.1). Most respondents recalled situations where they encountered that EU directives or judgements were not incorporated in time or only vaguely into national law.
There was considerable variation in how frontline implementers described their use of the different legal levels in case of conflict. One respondent argued that he interprets EU law independently (NRW A.1). He would do so even if it meant bypassing national law. The respondent explained that he does not consider it fair that the individual migrant has to pay for the failure at the national level in transposing directives and rulings on time (see Dörrenbächer Reference Dörrenbächer2017). Respondents from the Bezirksregierung confirmed that because of the supremacy of EU law it was expected from the city and district offices to rely on EU law when transposition actors did not incorporate EU law within the deadline.
Despite this expectation, most German respondents adopted more reluctant approaches. Typically, the respondents indicated that they would not directly rely on directives that were not yet part of German law. They argued that this goes beyond their competences. Nevertheless, most German respondents explained that they would directly apply CJEU rulings.
Generally, respondents described that they often put national law, Länder decrees, EU directives and CJEU judgements side-by-side to take decisions. In this way, they try to fill in gaps in the different legal levels. By “putting the laws next to each other” (NRW J.2) they explained how they distill the principles of EU law and interpret them in light of national law. Many respondents also explained how they sometimes created a “line of paragraphs” (NRW C.1; NRW C.2) out of the different legal levels to take decisions in individual cases, if the individual legal levels left them with ambiguity. In this respect, most respondents argued that EU law:
(…) constitutes a parallel legal text for us (…). You cannot just open the law book and check how to deal with a case. It is a patchwork of all the different directives and EU case laws. (NRW H.2)
Using EU law alongside national law is experienced as a new challenge by many implementers, and as one respondent explained: “applying EU law is the examination of the masters” (NRW H.1). In addition, some respondents considered EU law as frustrating because they felt discouraged by the different national principals to independently rely on new EU law (NRW F.1; NRW F.2). As a result, one office leader described the emergence of EU rulings as a trial and error process: “If you lose the second case in front of the administrative court, due to an EU ruling, you will notice that maybe in the future you have to consider this aspect” (NRW B.1). At the same time, particularly in small offices, implementers feel insecure and overwhelmed by the many legal levels:
I have to have so many more things in the back of my mind and an understanding of where to find all the different laws and I need an understanding of how it is meant for each context. Often it is a matter of knowing where the discretion actually is. This creates some insecurity. It would be easier if I had the law book where I can read literally what conditions there are and what exceptions there are. EU law makes it again more blurry. (NRW J.2)
In small offices, implementers explained that they often ask for advice from their district government in case of legal ambiguity. In large offices, the office leadership and caseworkers themselves often solve legal ambiguities (NRW BZR.2).
Overall, the analysis indicates that whereas EU law has substantive effects on their work, the Dutch respondents typically do not struggle with EU law. Instead, they rely on the filtered instructions form the ministry or the IND policy department. Instead, in the more discretionary German structure, independent compound uses of EU law were sometimes required from the frontline implementers. However, practices of the German respondents were highly diverse. Although the discretionary structure forces implementers to sometimes act as independent experts, respondents differed in how comfortable they felt when doing so. Respondents in small Ausländerbehörden approached EU law more in line with their Dutch counterparts. This is interesting, as they operated under most different discretionary conditions.
Dutch respondents understood their task as implementers in direct accountability to the national legislator, by contrast respondents in the small Ausländerbehörden struggled with the diversity of legal steering signals coupled with a lack in expertise and confidence to apply EU law. Frontline implementers in larger German offices saw themselves more often as being competent to decide on their own. This active approach can be explained by the fact that larger offices provide a more institutionalised support structure for frontline implementers but are still subject to little external supervision.
Conclusion
This study investigated how bureaucratic structures condition frontline uses of EU law. Following an organisational perspective (Egeberg Reference Egeberg1999; Trondal Reference Trondal2011; Wockelberg Reference Wockelberg2014), the article added macroinstitutional factors and comparative research to frontline studies (Hupe and Buffat Reference Hupe and Buffat2014, 549). It was hypothesised that when frontline implementers are confronted with multiple political principals with few monitoring tools and little organisational controls, they use EU law more independently of national instructions than frontline implementers in discretion-constrained structures.
Using interview data from Dutch and German frontline migration implementers, the study largely supports these expectations. Dutch frontline implementers, who operate in a nondiscretionary structure, largely acted as national servants. By contrast, many German respondents used original EU law parallel to national law. The differences between the Dutch and German civil servants in this study resemble the differences between the Danish and Swedish implementers studied by Wockelberg (Reference Wockelberg2014). Generally, these findings support Trondal’s (Reference Trondal2011) and Egeberg’s (Reference Egeberg2008) claim that organisational variables condition executive behaviour towards EU law. Nevertheless, the study challenges Egeberg’s (Reference Egeberg2008) assumptions that direct socialisation contact between national administration and EU institutions is required for independent use of EU law. Instead, this study supports Wockelberg’s (Reference Wockelberg2014) observation that implementers in the later stages of EU implementation also sometimes use EU law independently of national instructions.
However, going beyond Wockelberg (Reference Wockelberg2014), this study showed that even though the German respondents operated under considerable discretion, they did not always consider themselves competent enough to act as independent experts. Instead, German respondents, particularly those in small local offices, resembled their Dutch colleagues in their reluctant approach towards EU law. Yet, the rational for reluctance towards EU law in these most different structures was different. The Dutch respondents did not consider it as their task to consult original EU legislation and CJEU rulings by contrast, the respondents in the small German offices struggled with EU law. The findings of this study have implications for organisational theories in the EU context. They suggest that whereas discretionary structure may be a necessary condition for independent use of EU law, it is not a sufficient condition. In discretionary structures, internal organisational structures and microlevel factors complement explanations for the use of EU law (see Dörrenbächer Reference Dörrenbächer2017).
Finally, this study provides new insight into EU compliance and implementation studies, that go beyond the transposition phase (e.g. Knill and Lenschow Reference Knill and Lenschow1998; Kaeding Reference Kaeding2006; Falkner et al. Reference Falkner, Hartlapp and Treib2007; Versluis Reference Versluis2007). The Dutch case suggests that discretion-constrained administrative structures promote harmonised uses of EU law because implementers in such contexts approach laws and administrative guidelines in a consistent way. Nevertheless, the downside of this behaviour is that implementers may also follow noncompliant national instructions. Moreover, incorporation of CJEU rulings on the ground is delayed. By contrast, the German respondents may be more Europeanised as their discretion forces them to sometimes consult original EU law. This provides the chance to correct vague and noncompliant transposition on the ground. However, as the interviews have shown, it is questionable whether frontline implementers have the appropriate competences to interpret original EU laws. Although the EU could tackle this issue by offering national implementers more guidance on how to use EU law, the diversity of frontline activities in the discretionary structures still poses risks for harming legal certainty and accountability. Overall, the respondents of this study were very aware of the origins of the laws they apply, and their use of law has considerable implications for the final outcomes of EU law. Nevertheless, the potential to correct for problematic transposition at the frontline should not be overestimated.
In order to hold EU-level factors constant, this study focussed on the specific policy context of migration law which is a highly politicized field. This case study approach limits the generalisability of the findings. In more technical policy fields, implementers might be less aware of EU law due to limited public debates. At the same time the politicisation may make frontline implementers more risk adverse which may limit their independent use of EU law. To explore the generalisability of the results further, an avenue for future research would be to extend the study across policy areas. Moreover, future research should pay more attention to combining macrolevel with microlevel explanations for frontline use of EU law. A fruitful approach could be a cross-country survey research among national implementers, which combines individual-level and structural explanations for attitudes and approaches towards EU law.
Acknowledgement
The author would like to thank the discussants and participants of the workshop “(non)compliance with EU norms, rules and values” at the University of Duisburg-Essen 6–7 April 2016 who commented on a previous version of this article. The author would also like to thank the three anonymous reviewers for their constructive comments that have helped the author to considerably improve the article. A special thanks also to all respondents.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0143814X17000095
