Civil servants, migration and documentation

Material practices, including paperwork and digital technologies, influence the way we communicate, collect and receive information, and shape our relationship with the state (Carswell and de Neve, Reference Carswell and de Neve2020). Scholars have, as a consequence, studied how algorithms contribute to redefining “the rules of work” (Rosenblat, Reference Rosenblat2018) of public assistance or policing (Joh, Reference Joh2014), wondering about potential biases and implications for privacy. The use of biometric data is a tool for governance and migration control (Broeders, Reference Broeders2007) and documents (digital or paper) are – as argued – not passive (Hull, Reference Hull2012), but infused with politics of power that differentiate between populations and also constitute political relations (Carswell and de Neve, Reference Carswell and de Neve2020). The role of documentation becomes especially relevant when it comes to social welfare offices in which an excessive need for evidence exists, in particular when digitalisation further marginalises already precarious groups (Pors and Schou, Reference Pors and Schou2020). The place of (digital) technologies in the daily work of street-level bureaucrats, however, has rarely been investigated (Holm Vohnsen, Reference Holm Vohnsen2017; Gilliom, Reference Gilliom2001). Instead, more focus is given to the increasing standardisation and automatisation of administrative processes limiting discretion, leading to the disappearance of direct contact between clientele and bureaucrat (Bovens and Zouridis, Reference Bovens and Zouridis2002). Routines and digitalisation are thus supposed to support bureaucrats in their processing of cases. We argue that both also contribute to an increased control of clients and traceability of cases by various institutions, but do not simply replace direct interactions (Hansen, Lundberg, and Syltevik, Reference Hansen, Lundberg and Syltevik2018).

In this contribution, we illuminate how ‘hybrid’ bureaucratic work, between digital procedures, personal contacts and institutional exchanges, contributes to the expansion of migration control into social policies. This is supported by prior research at the intersection of social policy making and migration law that underlines the limited knowledge we have on processes and actors implementing welfare policies despite the ability of welfare governance studies to capture complexities of societies and their fragmentation (Carstensen et al., Reference Carstensen, Burau, Dahl and Hald2021). Some works focus on the behaviour and actions of social partners (Knotz et al., Reference Knotz, Fossati, Scalise and Hooijer2020) and how they might assist immigration control depending on the level of institutionalised cooperation (Morgan, Reference Morgan2020). Other contributions argue that ‘unwanted’ migrants are institutionally impeded from full access to welfare states (Hooijer and Picot, Reference Hooijer and Picot2015; Bolderson, Reference Bolderson2011). There is not only a high level of variation in the inclusion of noncitizens in social policies between different countries, but also differentiated belonging (Carmel and Sojka, Reference Carmel and Sojka2021; Shutes, Reference Shutes2016).

We follow the idea of Carstensen et al. (Reference Carstensen, Burau, Dahl and Hald2021) to conceptualise welfare professions as relevant collective actors that contribute to migration and social policy governance and study their (in)formal coordination and the increase in digital practices that supports uneven surveillance of the most marginalised (Henman and Marston, Reference Henman and Marston2008).

While effects of digitalisation on workers’ well-being, organisation or efficiency in the private sector (Cijan et al., Reference Cijan, Jenič, Lamovšek and Stemberger2019) are regularly studied, only a few contributions analyse the effects of digital technologies on the daily work of civil servants. One exception is Gilliom’s study (Reference Gilliom2001) on welfare bureaucracy controlling female clients claiming benefits, in which highly technologised systems often hindered payments due to petty mistakes, thus prolonging the waiting time of welfare clients, but also limiting discretion for staff.

A study of digital practices in welfare offices thus captures the outsourcing of migration control tasks that, as we argue, support political and legal restrictions affecting noncitizens and that are not simply produced and implemented by agencies that directly deal with migration enforcement (Infantino, Reference Infantino2016). We highlight new responsibilities, institutional exchange and dependencies that affect not only institutions and actors that were seemingly unconnected, but also the handling of cases and clients that find themselves within this network of responsibilities and contestations (Eule et al., Reference Eule, Borrelli, Lindberg and Wyss2019).

Social assistance policies for noncitizens in Belgium and in Switzerland

In most EU member states, welfare entitlements are contingent upon residency titles, although specific rights and access to benefits vary (Andreetta, Reference Andreetta2019; Lafleur and Vintila, Reference Lafleur and Vintila2020). Simultaneously, social assistance can have consequences on the residency titles of those who receive it: being a financial burden for the state can constitute, both in Belgium and in Switzerland, grounds for revoking one’s right to remain (Lafleur and Mescoli, Reference Lafleur and Mescoli2018; Borrelli et al., Reference Borrelli, Kurt, Achermann and Pfirter2021).

The Belgian constitution states that “everyone is entitled to live in conditions that conform to human dignity” (art. 23). In 1976, public centres for social assistance (PCSAs) were created in order to distribute social assistance, meant to guarantee dignified living conditions. Citizens and (irregularised) noncitizens who wish to benefit from social assistance – medical or otherwise – have to file their request with the PCSA of their municipality. Based on a ‘social inquiry’ conducted by a caseworker, the administration examines whether the applicant meets the legal criteria and notifies the applicant of its decision by post. People’s residency status – which translates into various kinds of papers – is one of the first and most crucial steps of this inquiry, since it determines what kind of assistance could be awarded. Irregularised migrants only receive medical assistance, others might receive financial support. A letter on their ID card can signal whether they are Belgian citizens, permanent residents, EU citizens who registered, the family member of an EU resident or of a third-country national or foreign residents – most often entitling them to the minimum income scheme. An “orange card” meant that their immigration claim was still pending: users could therefore benefit from financial assistance, usually granted in the same amounts as the guaranteed income. An order to leave the country meant that their claim has been decided on and denied, excluding them from any financial assistance. In a lot of cases, however, the information on these documents was either outdated or incomplete; people lost them or failed to bring them to the administration. Immigration statuses also regularly changed while people benefitted from social assistance: residency titles were updated or withdrawn; different ones were awarded.

In some cases, receiving financial assistance could cause noncitizens to lose their resident status, and therefore, their right to public assistance beyond emergency medical careFootnote ² . EU citizens who came looking for a job or third-country nationals whose residency was accepted based on their integration into Belgian society were indeed granted their immigration claim under the assumption that they would not represent a “burden” to the Belgian welfare system. Depending on social assistance therefore meant that the aforementioned requirement was no longer fulfilled. In these cases, welfare offices had a duty to warn applicants before processing their claims, giving those applicants the opportunity to withdraw requests for assistance. If people maintained their request and eventually received financial assistance, the information would appear on the national social security database, which the migration office had access to. Although residence permits were not systematically withdrawn, an investigation was likely to occur, potentially causing people to lose their permit. In other cases, social workers tried to help their beneficiaries without a legal immigration status to successfully apply for long-term residency.

The history of Swiss social assistance began in 1905 when the ‘conference of caretakers of the poor’ (Armenpflegerkonferenz), since 1996 the Swiss Conference for Social Assistance (SKOS), came into being and tried to establish welfare based on the place of residence rather than place of origin, or ‘home’ (Schnegg and Matter, Reference Schnegg and Matter2010). In contrast to matters of asylum, social policies are a cantonal responsibility, yet cantonal courts and the Federal court have supported the financial rates set by the SKOS so far. The SKOS thus sets frameworks for its members (including most cantons, many municipalities, different federal institutions and private actors working within social services provision) and guides social services that are in charge of handing out social assistance and taking care of recipients through integration programmes and regular meetings.

Similarly to Belgium, all residents in Switzerland have the right to social assistance. It is understood as a final support net and only takes effect once no other (re)sources are available (including private funds, unemployment benefits, or other insurances). Yet, social assistance is understood as debt, which might need to be paid back in case of positive income levels, inheritance and, for foreign nationals, if they aim for naturalisation. Most importantly, foreign nationals with a (permanent) residence permit (EU/EFTA and third-country nationals) claiming social assistance may face the revocation, withdrawal or downgrading of their permits (Borrelli et al., Reference Borrelli, Kurt, Achermann and Pfirter2021; Federal Act on Foreign Nationals and Integration, FNIA, 2019 Art. 62 lit. e, Art. 63). Based on the national obligation to report (Ordinance on admission, residence and gainful employment, 2019 Art. 82b-f), several institutions, including social services, employment offices, police and prosecution (to name a few) are obliged to inform migration offices as soon as foreign nationals apply for support or come up in criminal procedures, though leaving open how this information should be given – much as, in Belgium, social services taking care of social assistance inform their applicants about this reporting obligation. The initial decision and investigation of the case is done by the cantonal migration offices, which receive information on the volume of benefits (by social services) and the level of ‘integration’ of noncitizens by several other agencies and actors that are involved within migration management.

In both countries, receiving assistance may cause noncitizens to lose their residence permit – in Belgium, only those without a permanent status; in Switzerland, all foreign residents are at risk. Both are decentralised federal states, often described as conservative (at times liberal) welfare regimes (Esping-Andersen Reference Esping-Andersen1990). Our results, however, become relevant for other European member states, too, given that contesting residency through welfare receipt is allowed under EU law and given that we witness a general retrenchment of welfare states and an increase of highly differentiated access (Schmitt and Teney Reference Schmitt and Teney2019). The following sections will explore how both administrations receive, process and share information in practice – and what kind of consequences this has for the administrative lives of migrants.

Digital case-making and the changing work of welfare bureaucrats

Today’s daily work of caseworkers in welfare offices includes an exchange on the legal status of noncitizens and their behaviour discussed with different agencies of public administration, as well as an increasing digitalisation of practices. Information can be exchanged personally, via phone, mails, including data and files, via paper or also via digital databases that allow several state agencies to gather relevant data on a respective client. In the specific case of social assistance, Belgian welfare bureaucracies regularly make use of, and receive notifications from, the national social security database. Introduced 20 years ago, this database contains information from various public services: it states people’s addresses and residency statutes, immigration procedures, if there are any, whether or not the persons’ health insurance is up to date, if they receive social security benefits, if they work, etc.

By introducing a few, almost automated tools for control, the database uncovered situations – such as people working while receiving financial benefits – that PSCAs had very few ways of verifying before. The database indeed compiles information provided by various public services – such as the national institute for healthcare (INAMI), (un)employment agencies or the Agency for Social Integration (SPP IS), responsible for funding local welfare administrations. Public institutions – such as the immigration office – can access the available data, but cannot directly modify its content. Because of the broad information that it contains, the database has become an indispensable part of the social inquiry that goes together with every request to the welfare office: social workers print out the database spreadsheets and put them inside the cardboard file before meeting with one of their beneficiaries. This allows them to verify that applicants meet the legal criteria for social assistance – having the right residency/immigration status – but also to gather information on the case, such as health coverage, whether people work or receive other benefits. Once a request is granted, the fact that the person receives social assistance will appear on the database a few months later – once the information is processed by the SPP IS. After this delay, migrants – whether irregularised or not – receiving assistance will be listed, just like any other kind of beneficiary. Welfare offices will also receive “warnings” if new information is available about an ongoing case: if the immigration status of the beneficiary changes, or if they start working. Digital communication between and across government agencies and policy areas also illustrates the embeddedness and overlap between different (social) policy areas (Clasen et al., Reference Clasen, Clegg and Goerne2016) – in this case, welfare, employment and migration.

In Switzerland, information flows are less automatised and rather work the other way around. Due to an obligation to report that pushes social services to inform migration offices about foreign-national clients receiving welfare (Borrelli et al., Reference Borrelli, Kurt, Achermann and Pfirter2021), municipal social services share the level of benefits received by their clients and, if asked by the migration offices, also need to write a report on their clients’ behaviour, reason for welfare dependency and prospects to re-enter the labour market. These exchanges are not regulated via a national database, but depend on the canton, due to Switzerland’s federal structure. Social services make use of a national database that stores information on foreign nationals, including asylum seekers, EU/EFTA citizens and any other noncitizens, also used by migration offices. Then there are cantonal databases in which social services can search for information, as well as internal systems that store digital case files. Only rarely do we find a shared database, as in Belgium, where migration offices can automatically see whether or not a client is depending on social assistance. This is due to social assistance being regulated by each canton, in contrast to nationally regulated asylum laws. This causes parallel systems to evolve: although partly still relying on paper files, though hoping for digitalisation to fully take over, social services need to share information with migration offices mostly via e-mail or in form of letters. Where cantons have established shared databases between different public institutions, migration officials explain that it has made their work easier, partly because it reduces the need for direct communication between them and the social services. Yet, this automated sharing of information also circumscribes the power and thus discretion of social services, unable to frame dependency in a specific way. As such, social services become an automated part of migration control, also broadening the control that clients experience. Where there is no harmonised database, information can still be selectively transferred, or its communication delayed. Further, while the existence of automated technologies may often save time and produce ready-made data, thus facilitate work, formalised ways, such as databases and standardised forms, reduce information and can cause significant and time-consuming obstacles, also explained in the Belgian context:

Swiss social workers reflect on the same issue. The need to log every step of their work into an online system severely reduces their time for clients, shaping them more into technocrats than being frontline workers (see Bovens and Zouridis, Reference Bovens and Zouridis2002). The lack of time to really assess and work on a case also challenges potential contestation, something that has also been highlighted in the Danish and U.S. contexts (Holm Vohnsen, Reference Holm Vohnsen2017; Gilliom, Reference Gilliom2001). Standardisation increases the screen-level time for bureaucrats and digitalisation pushes them towards system maintenance, rather than handling single cases (Bovens and Zouridis, Reference Bovens and Zouridis2002). Social workers consistently highlight that their job has gone from being mainly “social”, although there was of course always paperwork involved, to mainly administrative. They refuse their role as mere paper-pushing bureaucrats, who ask for documents, fill in forms and templates, and write reports and official notices. This shift in technologies, which reduces the necessity for direct physical exchange, affects the self-understanding of social workers.

In large Belgian cities, every staff member handles at least 100 cases, of which an average of 20 need to be called in for a meeting every month. This is similar to Swiss social services. These meetings mostly consist of bringing paperwork: bank statements, school transcripts, medical certificates, job applications, etc. Social workers, on their end, check the database for changes.

Romuald’s actual understanding of social work is to accompany clients to other appointments and to follow their cases more physically, something he is unable to do, due to his caseload. Similarly, Marlene, working in a Swiss social service, acknowledges:

The reduction of social workers’ ‘physical’ activity is caused by the expectation to excessively control clients, which is facilitated through digital information storage and standardisation (see also Henman and Marston, Reference Henman and Marston2008). Digitalisation, as Gilliom (Reference Gilliom2001) already underlined, may thus reduce the ability for bureaucrats to react in an individualised manner and adapt their work to the client. Discretion can be hindered or more guided – a goal that was institutionally aimed for and was investigated by various researchers focusing on bureaucracy (Evans and Harris, Reference Evans and Harris2004; Bovens and Zouridis, Reference Bovens and Zouridis2002), while coming at a cost for the individual client that faces very specific circumstances.

Despite the masses of data that are collected, welfare offices still rely on direct contacts (interviews and field notes, Belgium and Switzerland), needed to assess individual situations and decide on support or sanctioning. Due to the complexity of cases that cannot be depicted by database entries, we argue that especially the bureaucratic work of social assistance has not (yet) become a fully automated and standardised field of computer-controlled decision-making, but rather an entangled hybrid practice shaped by new rules to communicate with other agencies, internal digitalisation processes and the professional rationale of individual service.

Administrative systems that are supposed to routinise procedures in order to counteract heavy workload can create professional disagreement among social workers, sometimes even increasing workload. The level of routinisation and standardisation in migration decisions and welfare has not (yet) reached the level of entirely human-free systems, yet digital technologies become tools of increased control that support data and information sharing with other institutions and also allow agencies to exert more power upon certain individuals. These exchanges support migration policies further undermining the right to receive welfare for foreign nationals: while foreign nationals may receive social assistance, it can endanger their stay – and today’s fast information flows allow for an enhanced and quick control.

The creation of evidence in digital times and client control through documentation

Besides information exchange and digitalisation, welfare bureaucrats’ daily work consists of investigating every individual request, collecting and examining documents, writing reports and recommendations (Andreetta, Reference Andreetta2019). After a request is granted, additional reports have to be written every time new information comes to light. Those reports are, however, supposed to be discussed with beneficiaries – who can then provide additional information, explain inconsistencies, or challenge the information received by the welfare office.

Belgian and Swiss welfare offices also get notices when people work or are listed as beneficiaries of other social benefits: other state institutions thus also provide social workers with information that they need to act upon. As soon as individuals receive money, Swiss social services expect to be informed in order to re-assess the budget that individuals receive (field notes, 2020), e.g. if their clients received an income, lottery winnings, inheritances or money from friends. The system is itself already suspicious and prone to collecting as much documentation as possible, something that is facilitated by digital storage and databases. Evidence is also created through registering people’s debts (including loss certificates, or unpaid bills), something which will also affect the right to stay on the territory. In the Swiss case, even if someone paid back all debts, it will still be registered in a national database, unless one pays a fee for its deletion and the creditor agrees to it (interview with Barbara, Swiss cantonal integration centre 2020).

Importantly, information from the database or by other actors is assumed to be true. Therefore, if warnings are incorrect, or the received social assistance budget wrongly calculated, beneficiaries end up having to provide evidence – in the form of documents – that the information is false.

Proving that something did not happen turns out to be difficult for beneficiaries, not to mention that it forces them to produce yet another document. However, offices ask for the production of evidence through paperwork or online documents. They need to scan, save and sort such documents that create a case file and thus become the story of each client. Due to the ability to store and organise huge quantities of data, but also the need for evidence and documentation, we gain insights into digitalisation processes from their absurd side. Foreign nationals become fearful of registering and making use of support, because it remains unclear what kind of information will be used how, and which kind of support impacts upon their residence permit:

In Switzerland, the receipt of social assistance directly affects foreign nationals’ right to stay, no matter if they have a permanent residence permit or have lived in Switzerland for decades (FNIA, 2015). Legally, the (significant and long-term) dependency on welfare leads to the revocation, non-prolongation or even downgrading of (permanent) resident permits. Following their duty to inform beneficiaries of their rights and duties, most welfare offices (in both countries), when faced with a request from EU and third-country nationals with a residence permit, warn them of the risks – sometimes causing beneficiaries to withdraw their request. Social services and legal counselling offices thus problematise this clash of rights and restrictions to their claimants, at times even proposing that clients do not apply for assistance, despite their right to receive financial support.

The digitalisation and data exchange drastically increases the documentation of each case and thus allows bureaucrats (and in particular migration offices) to trace individuals and control them, something that creates unease within welfare offices. In Belgium, updates or alerts also indicate when a new decision has been made at the migration office, whereas in Switzerland it is often noncitizens who take the migration office letters to the welfare offices. These decisions are partly based on prior reports done by welfare offices making them into tools for ‘postentry migration-control aims’ (Ataç and Rosenberger, Reference Ataç and Rosenberger2019, 2). Here, information flows and migration law undermine certain social and welfare rights, due to the consequences that the former have on foreign nationals and their right to remain. However, there is also the possibility to contest such policies, using digital practices and information flows to the clients’ advantage.

Contesting documentation and practices of control

While controlling and gathering evidence are significant aspects in the daily work of those granting welfare, social work is also embedded in a culture of care. The social workers that we shadowed regularly pointed to the tensions between their professional ethos, the values embedded in their training, and the system of control that “working for the state” entailed. This section will show how both digital and analogue paperwork can sometimes be used in favour of the client, by using delays in database information transfers, or by writing paperwork that will eventually help advance their immigration case.

Sarah’s example demonstrates that while the introduction of technologies has changed the daily work and practices of civil servants, in welfare administrations, social workers still have discretion in the way they deal with requests, yet sometimes are also hindered through routinised work steps. They know time frames in which migration offices might react and can write reports explaining or contradicting information from the database: they can encourage clients to amend or withdraw their requests. Such withdrawal is, however, criticised by Swiss social workers, since they are worried that people will not be able to make ends meet. The fear of registering and non-take up of support, as Swiss staff assume, will create a new group of marginalised and poor individuals. At the same time, welfare offices also negotiate what kind of information they are willing to hand out:

Social workers are well aware that their information matters, even though they cannot be sure to what extent their reports are taken into consideration, causing noncitizens to lose their permit. As such, they negotiate what kind of paperwork and contents they are willing to share, to the extent that legal requirements allow them to.

Communication across different state institutions finally includes the transfer of case files to welfare courts, in case applicants decide to challenge administrative decisions. In Belgium especially, social assistance claims have become increasingly judicialised, and applicants without a permanent resident status have regularly gone to the European Court of Justice – see, for example, Grzelczyk v Centre public d’aide sociale d’Ottignies-Louvain-la-Neuve (2001) or Abdida v Centre public d’aide sociale d’Ottignies-Louvain-la-Neuve (2014) – in order to claim social assistance ‘against the state’. In such cases, social workers can choose to write reports ‘in favour of the user’ in order to help further their case in court (Andreetta, Reference Andreetta2019). As explained in the case of Grzelczyk, local offices indeed rely on state funding for the benefits that they grant: if such funding is denied on the grounds that the applicant is not eligible due to their immigration status, the only way for welfare administrations to qualify for state reimbursement is through a judicial decision. In such contexts, case workers use their discretion – both in face-to-face interactions with users, and in the documents that they write – in order to aid despite, or outside of, migration control.