Navigating the Representative-Politics–Liberal-Rights Dilemma: Social Policy Designs for Nonremoved Migrants

Abstract

Although nonremoved rejected asylum seekers (NRASes) are declared unwanted, the liberal state is obliged to provide them with basic social protections. We argue that various social policy designs can mediate the representative-politics–liberal-rights dilemma and allow for (limited) access to differentiated, conditioned benefits. Drawing on migration control and welfare-state literature, the findings stem from expert interviews with stakeholders and document analysis in Austria, Sweden, and the Netherlands. Welfare-enabling approaches are context specific, varying from path dependencies in Sweden to change-resistant forms of policymaking in Austria. In the Netherlands, exclusionary measures are explained by early general welfare retrenchments.

In the 1990s, the political scientist Christian Joppke (1998) raised the provocative question of why it is that liberal states accept unwanted migrants. ¹ He rejected the view, broadly shared, that states let unwanted immigrants in as a result of human rights obligations and instead presented domestic reasons. Nevertheless, in the years that followed scholars put forward the argument that the liberal state had lost the capacity for migration control due to the international human rights regime and economic globalization. In the course of international migration, failures in border control and gaps in policy formulation and implementation have been identified (Czaika & De Haas, 2013), yet domestic politics has continued to attract little attention.

This article takes up the topic of the acceptance of unwanted migrants and goes one step further to the issue of social welfare arrangements. It focuses on formal entitlements to benefits and services in the subfields of housing, financial allowances, basic education, and medical care for nonremoved rejected asylum seekers (NRASes)—an analytical category with complex and specific features. Firstly, the individuals concerned have not yet made a financial contribution to the welfare system but take from it, and secondly, they are assumed not to reside officially in the host country but cannot be removed due to legal and practical constraints. According to international and European legal frameworks (e.g., Return Directive 2008/115/EU), liberal states are obliged to supply a minimum level of social benefits and services, even for asylum seekers who are officially declared unwanted (Spencer, 2016). To put it another way: NRASES receive social benefits based on European and international provisions that must, however, be realized by the signatory states.

Politically, the theme of welfare is far from trivial. As Cvajner and Sciortiono (Cvajner & Sciortino, 2010) highlight, a sizeable, irregular foreign population challenges notions of political statehood and societal membership. Moreover, nonremoved migrants present a challenge to the legitimacy of democratic governance and representative politics (Mair, 2009; Hampshire, 2013). Alongside these areas of conflict, the theme of social benefits for noncitizens is considered to be one of the key components in the contentious politics of immigration (Bommes & Geddes, 2000). This is particularly the case with NRASES. Against the background of this challenge, in this article we deal with policy designs applied in different national contexts and policy areas to support welfare for NRASES. The concept of policy design is attentive to the means and mechanisms through which welfare is provided to target groups (Howlett & Ramesh, 2003). By their very nature, only limited welfare schemes are available for NRASES—namely, differentiated forms of universal approaches and not insurance-based benefits. The latter would require labor market participation, which is in the investigated countries ruled out by law. This restriction to certain types of welfare arrangements is important to keep in mind, as scholarly research on welfare attitudes emphasizes that universal benefits are more prone to welfare chauvinism on the part of anti-migration forces than insurance-based entitlements (Ennser-Jedenastik, 2018; Sefton, 2008).

Focusing on the analysis of policy designs, the main research questions are the following: Which policy designs are organized to more or less balance competing demands? Which types of benefits allow for better access to social welfare than others, and which types are more prone to retrenchment? What influence do specific national contexts, both institutional frameworks and politics, have on the implementation of divergent policy designs?

The article offers an innovative analytical framework in the form of a typology of policy designs. Furthermore, empirical knowledge on social policies for NRASES will be presented to contribute to the literature of welfare designs for marginalized groups. Here, besides the widely discussed group-selective approach (see Morris, 2010), we argue that two other policy designs are relevant in regulating access to social benefits: the nondebated transfer of universal measures to the specific group and the politics of exceptions from excluding propositions. The empirical findings stem from three national contexts: Austria, the Netherlands, and Sweden.

The article starts by elaborating on theoretical perspectives. In the following section provides contextual information on the three study countries, followed by the main section, which develops a typology of social policy designs and applies this to the regulations in place. In the section on understanding different policy Solutions, the outcomes are discussed. Finally, the conclusions summarize the article’s findings and points to further research challenges.

The representative-politics–liberal-rights challenge and social policy solutions

As the literature states, governments in liberal states have to cope with multiple responsibilities and competing imperatives for attaining legitimacy—on the one hand, liberal rights for all and, on the other, representative politics encompassing parties and voters who often take up an exclusionary stance on immigration and ethnic diversity (Hampshire, 2013). In more detail, Peter Mair (2009) emphasizes liberal governments’ responsibilities toward the rule of law and international treaties and to the interests of the citizenry. Consequently, human rights obligations constrain policymakers who seek to exclude migrants from access to social provisions. The migration/welfare nexus (Wright 2016) touches on overlapping frameworks, interests, and measures in the fields of migration control and social protection. As such it is embedded in this larger representative-politics–liberal-rights challenge.

The same line of thinking is used by James Hampshire in his book The Politics of Immigration (2013). He bases his analysis on the premise that immigration flows affect the statecraft of governments and ideas of legitimacy and sovereignty. Accordingly, liberal governments have to deal with representative-based and rights-based challenges. The former refers to the fact that political authorities are bound by features of democratic politics (elections, political parties, interests, and attitudes) and have to address the preferences of the majority population. Hampshire calls this phenomenon the democratic aspect of politics responsible for generating (electoral) support, trust, and confidence. In contrast, rights-based imperatives contain ideas of universalism and liberal norms and come hand in hand with human rights obligations under international and European law. The main characteristic of liberal norms is the protection of rights for all individuals and not just for national citizens, yet even so national citizens call for the exclusion of some individuals. This is precisely what characterizes liberal states (Joppke, 1998, p. 268). Therefore, basic welfare benefits do not have to be coupled to legal citizenship or residency status, rather all individuals residing in a territory should enjoy a minimum of social welfare provision. Yet as mentioned above, certain principles of representative politics necessitate national policies of restrictiveness and closure in the domain of immigration, meaning that a certain degree of restrictiveness is inscribed in the mode of majoritarian legitimacy (Hampshire, 2013, p. 157 ff).

The issue of welfare for NRASES not only touches on conflicting national interests and international obligations but also concerns two policy fields involving different actors and institutions, interests, and narratives—on the one side, the topic of migration control and, on the other, the field of welfare with its primary goal of protecting a society from destitution and the rupture of public order (Bommes & Geddes, 2000). Again, governmental actors have to strive to serve both fields—managing migration control goals and social benefits to the population residing in the country (Wright, 2016). The concrete arrangements between these policy fields determine the given quality of welfare benefits for immigrants (Sainsbury 2006). However, the knowledge available on the interplay between these two fields and how this impacts the adoption and practice of welfare provisions is far from sufficient.

Literature shedding light on the welfare-migration nexus also focuses on the so called welfare chauvinism of political and cultural forces and citizens alike. Welfare chauvinism is used to denote the attitude of certain segments of the society who are likely to view immigrants as a threat both to the values of the society and the national welfare system. Certain political actors call for different levels of entitlement between migrated and native populations, putting nationals before nonnationals. They call for welfare reductions, implying a fragmentation of social welfare along the lines of citizenship and residency status (Ennser-Jedenastik, 2018).

Regarding welfare programs for migrants, a very insightful perspective is the civic stratification approach. Lydia Morris (2010) assesses fragmentation as a policy strategy for governing irregular migration and welfare in European countries. This tool allows governments to grant distinct administrative categories of migrants lower social rights than those enjoyed by natives. A study by Sarah Spencer (2016) following this line of thinking mapped the entitlements to health care and education for undocumented minors across the European Union. Spencer acknowledged the interplay between the principle of human and refugee rights and the principle of fragmented and conditioned entitlements. The technique of conditionality couples access to benefits to certain individual characteristics and meritocratic criteria (see also Goldring & Landolt, 2013). In other words, the literature identifies both differentiation along the axis of residency status and the formulation of requirements as policy tools available to governments for dealing with its liberal obligations on the one hand and voters’ demands for closure on the other. To a certain extent, these tools may satisfy forces pushing for closure and restrictions toward migrants (Hampshire, 2013).

These insightful theoretical perspectives on fragmentation and conditionality will allow us to develop a conceptual framework focusing on various policy designs to navigate the dilemma. However, the typology presented below will transgress the existing scholarship on fragmentation and conditionality by including conditions for the transfer of universal rights to NRASES. All together, we are interested in identifying social policy designs that mediate the competing challenges and secure a certain degree of social protection for the group concerned. Before we turn to this, we provide some background information on the three study countries.

Countries and data

We have chosen a case study approach that includes three European Member-States—Austria, the Netherlands and Sweden—to identify and map out various social policy designs for nonremoved migrants and to trace the factors that can contribute to understanding their different outcomes.

Heegaard Bausager, Köpfli Møller, and Ardittis (2013, p. 15) define the population of NRASes as comprising any “third-country national who is recognized by the national authorities … as being in a situation where he/she cannot be returned or removed.” Over recent decades, this group has become a quite sizeable. Although exact numbers are not available, the EU Commission estimates that there are one million rejected asylum seekers residing in the EU-territory (EU Commission, 2017). EUROSTAT uses the measure of “effective returns” and defines it as the ratio between third-country nationals ordered to leave and third-country nationals following an order to leave. In 2016, effective returns reached 50% (almost half a million non-EU citizens were ordered to leave the EU but fewer than 230,000 are known to have left). In Austria, the effective return rate is close to the EU average, in the Netherlands the rate is less than 40%, and in Sweden it is above 60% (Mananashvili, 2017).

Hosting a considerable population of NRASes in their territories, these countries have certain international and European obligations in common to care for NRASes. In particular, they have to follow the relevant European directives, most importantly, Reception Directive 2003/9/EU and Return Directive 2008/115/EU. Amended in 2013, Return Directive Art. 14 (Safeguards pending return) stipulates that member-states shall ensure that “third country nationals during the period of voluntary departure … and during periods for which removal has been postponed…” have access to a minimum set of basic rights, including access to emergency health care and essential treatment of illness, and for minors access to basic education Remarkably, this legislation does not mention access to housing or financial allowances.

Of particular relevance for policymaking in the fields of welfare and migration are anti-migration players. In Austria and the Netherlands, anti-immigration parties have been electorally successful since 2000 and have exerted a direct and indirect impact on policymaking—that is, on the policy positions of the ruling parties. In Sweden, the electoral success of the anti-immigrant the Sweden Democrats is more recent, with the party doubling its votes in 2014. Correspondingly, in Austria and in the Netherlands, migration has long been a negatively contested topic (Meyer & Rosenberger, 2015), whereas in Sweden it has become politicized only in the past few years. Nevertheless, a recent study found a negative correlation between the representation of Sweden Democrats at the municipal level and the level of social aid offered to vulnerable migrants (Tyrberg & Dahlström, 2018).

The countries under investigation share what is called the continental welfare state (Häusermann, 2010) but differ with regard to the level of universal benefits. Sweden, as part of the social democratic welfare state model provides comparatively more universal benefits than the conservative welfare states Austria and the Netherlands (Sainsbury, 2006). Although major welfare reforms to reduce spending have taken place in all three countries, the Austrian and Dutch welfare regimes are still labeled as corporatist and insurance based, and include elements of familialism; therefore, they favor those with long-term membership and participation in the labor market. In contrast, Sweden is still considered as having a more universalistic welfare regime (Borevi, 2014).

The degree of openness toward the admission and integration of migrants distinguishes the three countries. Measured by their migration policies, Sweden has one of the most liberal migration regimes, Austria is strict, and the Netherlands has shifted from liberal to strict over the last 20 years (MIPEX, 2015).

Against expectations drawn from migration and welfare literature, at present Austria has a more generous legal framework for NRASes than the Netherlands or even Sweden. The similarities in some external factors, such as EU regulations, will give us the opportunity to identify, first, national policy designs varying across countries and fields, before asking about the role of specific institutional frameworks and political determinants in order to trace the origins of diverging policy outcomes (Hampshire, 2013; Häusermann, 2010; Sainsbury, 2006).

Taking these countries, we carried out case studies to get in-depth insights into policy designs regarding the welfare of NRASes. First, we collected and analyzed legal documents—laws, decrees and administrative guidelines—that directly and indirectly regulate access to social welfare. We then carried out semistructured expert interviews between June 2016 and July 2017 (in total 25 in Austria, 27 in Sweden, and 21 in the Netherlands) to gain knowledge on the perspectives of stakeholders in the field of social benefits for NRASes (policymakers and government officials, administrators, caseworkers, representatives of NGOs, and charitable organizations at the local and national levels). A list of the referenced interviews is attached to this article. Additionally, the analysis included policy documents, parliamentary papers, academic publications, and media articles on the issue. The textual data was coded and analyzed using the method of summative content analysis to identify actors, modes of policymaking, and types and content of benefits and services (Hsieh & Shannon, 2005).

Given this rich material, our data allows us to identify different policy designs and to investigate the impact of national particularities on the different outcomes. In the following section, we develop a typology of social policy design for NRASes before using this to present our data on policy measures in the three countries.

A landscape of diverse social policy designs

A typology of social policy designs for NRASes

The public policy approach sheds light on the policies—measures and instruments—taken by governments, parliaments, and administrative authorities to achieve goals and ends, more precisely, to address and solve societal problems. Within the analytical framework of public policy analysis, we turn our attention to a specification of the policy approach, to policy designs. Policy designs contain certain characteristics of means and measures and ways and modes of policymaking through which welfare benefits are provided or not (Trauner & Wolff, 2014; Howlett & Ramesh, 2003, p. 162).

In continental welfare regimes, welfare is supplied either through universal benefits or through insurance-based benefits (Häusermann, 2010). In the case of NRASes, the whole range of policy solutions is not available, and only universal or selective universal benefits are provided, either through cash or in-kind services.

Based on welfare and migration literature, and on the empirical insights into legal entitlements gained from our research in the three countries, we were able to develop a typology of policy designs, consisting of three dimensions. These dimensions cover the scope of provisions ranging from comprehensive benefits and services to very restricted ones for poor relief (Sainsbury, 2006; Leerkes, 2016). The following classification is based on features that allow for a differentiation of universal measures, taking into account citizenship and residency status and requirements based on individual characteristics and merits.

1. Universal policies: Benefits and services are designed for all, natives and non-natives, independent of legal status. The peculiar characteristic of universal or comprehensive policy solutions lies in the fact of no specific regulations for specific administrative categories.

2. Selective policies: Differentiation and fragmentation are a common instrument of welfare politics, in particular when based on participation in the labor market leading to insurance-based entitlements. Within migration, a selective approach to benefits and services distinguishes between natives and nonnatives. Differentiation takes place along the criteria of residency status, which includes the idea that full welfare benefits are for the native population only (Morris, 2010). Differentiation of benefits based on citizenship status are identified by Morris (2006) as a strategy for managing the conflict between taxpayers’ and voters’ interests on the one hand, and the rights of migrants on the other

3. Exception policies: Due to political reasons, policymakers turn to the denial of benefits but formulate at the very same time exceptions to these exclusionary positions (Rein, 2008). The signal to the electorate should be that the government acts harshly towards „unwanted“ migrants. The practice, however, should allow for exceptions, e.g. for those who are perceived as „deserving“ or having experienced support from society, or for those who enjoy better protection through international agreements and propositions. Exception policies try to mediate the mentioned dilemma in the highest degree. Whatever the case, exclusionary policies cushioned by exceptions include a certain degree of discretionary power by caseworkers and street-level bureaucrats (Goldring & Landolt, 2013).

Within all three manifestations of this typology, the tools of setting conditions and requirements play an important role. Although to different degrees, the mode of conditionality impacts who gains benefits and of what quality. Moreover, for all three manifestations, the type of benefit, in particular cash or in-kind forms, is also relevant for the concrete shape taken by a policy design. As Sefton (2008, p. 610) argues, providing in-kind services instead of financial allowances, for example, serves taxpayers’ interests. The taxpayer “may be prepared to pay for some kind of redistribution to the poor but only if it takes the form of providing them with specific services, such as health care, food stamps, or subsidized housing.”

Social welfare benefits for NRASes

In the following we map out the kinds of policy solutions by applying the typology developed above. Table 1 gives an overview of the policies within each policy area (housing, social allowances, basic education, and health care) for each country.

Table 1.

Overview on Social Policy Designs for NRASes (by Country, 2017).

	Universal benefits	Selective benefits	Exception benefits
Housing		NL: time (4 weeks) A: AS and NRAS S: AS and NRAS	NL: Locations for Families and Minors
Allowances		A: pocket money like AS; no cash for housing S: since 2016 no daily allowances for NRASes	NL: No cash
Basic education	Sweden NL Austria (age 15)
Basic health care	Sweden	A: basic health care NL: necessary health care

Source. Author’s illustration; AS (asylum seeker), NRAS (non-removed rejected asylum seeker).

Universal benefit transfer

A universal welfare scheme guarantees residents’ access to benefits on an equal basis, independently of individual income, ethnic characteristics, or residency status. Unified schemes care for all residents, citizens and noncitizens alike, and do not establish different categories of natives and nonnatives and among migrants. Nevertheless, constraints in the scope of benefits can be adopted, as we will see below (Häusermann, 2010; Borevi, 2014). The main characteristic of these policies is that they do not require a specific policy decision for the group in question. Policy solutions migrate from one setting to another, from all residents to specific groups (Trauner & Wolff, 2014).

With regard to NRASes, social policymaking takes place in other settings and is then transferred to nonremoved rejected asylum seekers. However, the transfer of universal policies concerns only a few domains. The most illustrative example of a universal benefit is basic education. In Austria, it is the practice that all minors, independent of their status, enjoy identical treatment as set out in constitutional law on schooling. The same kind of regulation is present in the Netherlands. The Swedish law on education even states that persons who cannot be removed have to be considered as residents and therefore enjoy the same right to education. A recent amendment to the law on schooling explicitly gave the children of undocumented migrants the right to education, as extends to all other children living in Sweden (Ändring i kap. 29 av Skollag 2010:800/SFS 2010, Kap. 29, §2). Differences in age limits exist between the three countries. Here, Austria is the most rigid, and only minors up to the age of 15 are covered by this principle; in the Netherlands and in Sweden the age limit is 18.

To a lesser extent, basic health care is provided through the transfer of universal benefits. In Sweden, access to health care is universal; in Austria and the Netherlands health care is restricted to certain services. In Sweden, all children under 18 have free access to health and dental care. In 2013 a statutory law set new conditions and gave all undocumented migrants the same access to health services as asylum seekers. This change to an inclusive approach was the result of public outcry, primarily from physicians and NGOs (interview 1). In the Netherlands health care is limited to what are termed necessary health needs. In Austria, if health needs exceed basic services, additional services may be granted after examination of the specific case (interviews 2, 3, 4, and 5). The use of discretionary power plays here a decisive role.

Taken together, certain benefits are provided on a comprehensive basis, mainly without specific decisions. Policy transfer means that access is simply granted. As the transfer mechanism does not require decisions it has, ultimately, not led to contestation. In this sense, transfer of universal policies has turned out to be a successful strategy for reconciling both sides: to date, governments have avoided negative contestation in these fields and have recognized international obligations.

Selective benefits

The policy design of differentiation gives rise to selective benefits and services for specific categories of people. Producing specific groups and, in turn, treating them hierarchically is the main rationale behind the politics of differentiation. When applied to migrants, citizens and noncitizens are not treated equally but differently. Following this logic, benefits for citizens are more inclusive than for noncitizens, benefits for regular migrants are more comprehensive than those for nonregular migrants, and so on (Morris, 2006). The tool may take many forms. One is to provide a different quality of benefits for different groups, another is to grant different forms of services to different groups—for instance cash or in-kind benefits (Häusermann, 2010).

How is the politics of differentiation regarding NRASes reflected in policies in the countries investigated? The answer is twofold and reveals two steps: (a) the differentiation of benefits between asylum seekers (AS) and rejected asylum seekers pending return (NRASes) and (b) differentiation within the NRAS population by setting a range of conditions and requirements.

In Austria, transposing Reception Directive 2003/9/EU into national law has taken the form of the adoption of the Agreement on Basic Welfare Support (Grundversorgungsvereinbarung, BGBl. Nr. I 80/2004). This is the specific legal framework for both asylum seekers in need and nonremoved rejected asylum seekers and sets out that the category of aliens includes both asylum seekers whose status is being determined and rejected asylum seekers who cannot be removed for legal or practical reasons. Welfare support is provided predominantly through in-kind services; cash transfers are given only to a limited extent. To attain benefits, the individual has to demonstrate that he or she is financially in need and requires social protection by the state. Additionally, two conditions have to be met: active cooperation with return procedures and remaining in the same province wherein the asylum application was filed (Rosenberger & Koppes, 2018). Recent amendments to the Aliens Law (Fremdenrechtsänderungsgesetz 2017) have introduced special facilities for rejected asylum seekers, pending return. This legal change is a shift from the practice of accommodating asylum seekers whose cases are pending and rejected asylum seekers in the same institutions. The new policy follows a political interest in singling out the group of rejected asylum seekers and treating them worse than asylum seekers pending a decision on their application.

In the Netherlands, the Linkage Act (Koppelingswet) of 1998 differentiates between housing benefits for asylum seekers and rejected asylum seekers. At the time it was passed, the act amended several laws with the goal of linking claims to social benefits and services to a valid residency status. After an asylum claim has been rejected, there exists a legal term of four weeks within which to effectuate departure (RvA 2005 – Regeling verstrekkingen asielzoekers en andere vreemdelingen). However, this exclusionary policy came under pressure from local and international actors. In 2009, the European Committee on Social Rights ruled that the Linkage Act was in violation of the European Social Charter. In September 2012, the High Court ruled that families and children without legal residency status may not be left on the street. Eventually, the government turned from strict regulation to the policy tool of exceptions and installed special family facilities for migrants without legal status (Kos, Maussen, & Doomernik, 2015).

In Sweden, the Reception of Asylum Seekers Act (Lagen om Mottagande av asylsökande, LMA) of 1994 regulates social support for persons seeking asylum or subsidiary protection. Under this law, access to accommodation and financial allowances did not differentiate between asylum seekers awaiting a decision and rejected asylum seekers. In 2015, when Sweden accepted the highest rate of asylum seekers (per capita) in the EU, public opinion and the government’s position on admitting asylum seekers and its return policies have undergone a fundamental shift. Within weeks, the rationale of openness changed toward one of partial closure and conditionality (interviews 6 and 7). In this vein, the government issued a draft for an amendment to the LMA, which came into effect in June 2016. The amendment concerns those cases in which the right to support is terminated. Rejected asylum seekers have a period of two to four weeks within which to leave their accommodation after the decision to expel them has entered into legal force. Additionally, NRASes lose the right to a daily allowance. But certain exceptions are stipulated to mitigate the harsh consequences, as the next subchapter illustrates.

Denials cushioned through exceptions

When policymakers turn to measures excluding residents from social protection, a frequent side strategy is the adoption of exceptions (Spencer, 2016, p. 1624). As our empirical findings suggest, exceptions are a strategy to curb the risk of destitution and homelessness or, in the case of children, to avoid breaching moral principles or international obligations. In the case of NRASes, this toolkit is used in the subfield of housing. Demographic characteristics are used for defining and legitimizing exceptions.

Strict regulations in the Netherlands are not applied in the case of minors. The Linkage Act states that for minors, education is accessible in the same way as it is for citizens. Exception rules exist also for accommodation for families; families with children under the age of 18 have the right to accommodation in a so-called Family Location. Nonremoved migrants who cooperate actively with their return procedures may also have access to accommodation in Freedom Restricted Locations (Van der Leun & Bouter, 2015).

Regulations in Sweden stress the rights of the minor in all the policy fields under investigation. Adults living together with their children or with a child for whom they can be considered guardians are exempt from the restrictive regulations of 2016, as described above. In Austria, children are less relevant at the legislative level but are relevant in practice. In particular, stakeholders in the field emphasize that homelessness of families should be avoided; hence, greater discretionary generosity in granting basic care is shown when families are involved (interviews 2, 4, and 8).

The literature points out that exception as a policy design can be a strategy for dealing with contradictory functional requirements but that it does go hand in hand with vagueness and inconsistencies in regulations. Martin Rein (2008) identifies ambiguity and vagueness as characteristics of so-called problematic policies and policymaking. Although his assessment is not clear about the consequences of ambiguity and vagueness, Rein notes that in certain situations, clarity could be “costly and the only pragmatic course to follow is by the use of ambiguity, viewed as a strong precondition to achieve some measure in building a political coalition to promote collective action” (Rein, 2008, p. 392). Differentiation through group-specific regulations, applying a range of conditions, and adhering to the rule of exceptions may therefore result in confusing situations that are far from a transparent process based on clear decisions. However, it might also be a strategy for navigating the competing demands of liberal obligations and representative politics.

Policy developments after 2015

Since the arrival of large numbers of asylum seekers in 2015 and a high degree of the politicization of asylum, the issues of enforcing return and social welfare for NRASes has become prominent in political debates (Lutz, 2018). With NRASes protected by European law, on the one hand, and negatively mobilized for domestic electoral purposes, on the other, liberal governments face an even greater challenge of achieving a tradeoff between the inclusionary human rights regime and some exclusionary features of representative politics.

The policies for housing and allowances had shown a significant degree of consistency and institutional durability; however, in the aftermath of 2015 they have seen a lot of changes. The strict policy in the Netherlands meant that selective approaches and poor relief measures already dated back two decades; whereas, in Sweden and in Austria changes are now underway (modest ones in Austria and strong ones in Sweden). In Austria, the welfare of NRASes was addressed with a package of laws that entered into force in October 2017. In particular, the amendment to the Aliens Law 2017 puts the emphasis on detention and sets up separate facilities for those who are due to be returned within a short period of time. In the interests of preventing absconding, the law stipulates that rejected asylum seekers may be held in detention centers or have to live in specific return facilities until their voluntary or forced return is concluded. In all three countries, the limitation of social protection is achieved by differentiating more clearly between ASes and NRASes and intensifying the use of the condition of cooperation with return (Rosenberger & Koppes, 2018).

Differentiating between ASes and NRASes is a policy trend that goes against the findings presented by Heegaard Bausager et al. (2013, p. 15). The authors emphasize that in most countries, NRASes are often covered by regulations already in place for asylum seekers. Our findings suggest an inclination toward specific and restrictive policy programs, not at least as a response to calls from migration and return politics. Governments aim to make life more difficult for those already in the country and send restrictive signals to newcomers (interview 9). Based on our interviews with experts, we conclude that cuts to social provision for NRASes are motivated less by reducing welfare costs than by gaining an advantage from appearing tough on the issue of asylum. Both policy aims are in place; however, in the aftermath of 2015, aims around migration control have prevailed over those around welfare protection. Putting this in the terminology of the representative-politics–liberal-rights challenge, the pendulum has swung toward prioritizing representative politics.

Understanding different policy solutions

As demonstrated, the policy solutions allowing for social benefits differ between domains: education and medical care are more accessible through the use of the transfer of universal benefits than housing and financial allowances. Social policy solutions in these areas have a specific feature: they are realized almost exclusively as in-kind benefits. Moreover, they vary between the three national contexts; although for different reasons, Sweden and Austria are more open and supportive. The Netherlands takes a very strict stance in the areas of accommodation and allowances, precisely in those areas in which European law does not explicitly state minimum rights.

What seem most relevant in understanding the different policy solutions are national particularities with regard to politics and policymaking. In the Netherlands, the restrictive policy turn occurred within the arena of the welfare state; in Austria the sustained permissive policies for NRASes are the result of a certain type of policymaking that takes place outside of public contestation. The policymaking-argument regarding Austria is supported by the fact that both countries share a corporatist, insurance-based welfare system in general but differ greatly in their social treatment of NRASes.

In the Netherlands, since the late 1980s a shift from welfare to workfare was introduced in order to shrink the Dutch welfare state (Sanandali, 2013). The Linkage Act was passed in 1998 and accompanied a general attitude toward tightening welfare spending and privatizing certain welfare schemes and toward targeting irregular migrants. In this context, concerning welfare benefits for irregular migrants, the aim was retrenchment.

In Austria, the legal regulations governing social support of NRASes date back to 2004. To recap, in 2004 the federal government, by then a coalition between the Conservative Party (ÖVP) and the far-right, anti-migration Freedom Party Austria (FPÖ), transposed the Reception Directive (2003/9/EU), which laid down minimum standards for the reception of asylum seeking persons into national law. The policymaking process involved the federal government and representatives of the nine provinces. After long and tense negotiations, these actors come up with the Agreement on Basic Welfare Support, which was given the status of a constitutional law. This status implies an institutionalized form of decision-making by consent that is characterized by a veto-player setting. To change the substance of the agreement would require the consent of all federal and regional actors in charge. In other words, the legal character of the policy leads to policy continuity rather than to alterations. Although the accommodation of both AS and NRASes became politicized in 2016 and was a prominent issue in the 2017 general election campaign, the constitutional character of the Agreement on Basic Welfare Support meant that it was not possible to effect profound changes to the provisions for those who cannot be removed.

Until very recently, Sweden was famous for its comprehensive welfare regime that provided residents with support and care independent of legal status. This included asylum seekers whose applications had been refused and irregular migrants (Borevi, 2014). This inclusive welfare scheme and the immigration regime have been changed in the aftermath of the immigration of refugees in 2015. A closer look makes it obvious that the altered legal situation pointing toward restrictiveness is also accompanied by a range of exceptions to curb the negative effects. Moreover, a few years before, Sweden expanded medical care and education for irregular migrants, including rejected asylum seekers, to grant them the same access as citizens and other migrants. In sum, the picture is manifold. We can conclude that while the debate has clearly changed, in fact, access to welfare for nonremoved rejected asylum seekers still reflects the universal norms. The welfare state is there, albeit with some ruptures. This may lead to major discursive policy gaps, which means that the policy talk differs largely from policy adoption (Czaika & De Haas, 2013).

In Austria and Sweden, the issue of return became prominent in the aftermath of the so-called refugee crisis in 2015, when politicians across the political spectrum called for the return of rejected asylum seekers to secure the rule of law and society’s capacity to successfully integrate new migrants (interview 7). Against this background, the pendulum has swung toward more-restrictive policy proposals and the legal situation concerning welfare support has shifted toward exclusion. Thus, the ongoing challenge to redefine the role of the nation-state very much affects nonremoved migrants. Tough rules toward nonremoved migrants and signaling harshness are thought to have a positive influence on national features such as legitimacy, sovereignty, and state-craft (Bauböck & Scholten, 2016).

Conclusions

This article investigated how governments of liberal states cope with welfare support for NRASes against the backdrop of the representative-politics–liberal-rights dilemma. The paper drew on policy analysis literature to identify and discuss several policy solutions across welfare subfields in three countries. Due to a ban on participation in the labor market, social protection for NRASes has to be based on either universal entitlements (for all) or specific entitlements (for specific groups), which are granted under certain conditions and/or are viewed as an expression of means-tested benefits. Moreover, these benefits are provided mostly on an in-kind basis. The results demonstrate that universal policies to manage the aforementioned dilemma take manifold forms and are derived from specific national contexts.

In all three countries, we identified the impact of EU directives on national provisions. Nevertheless, the actual generosity or restrictiveness of social support varies between national contexts; the countries under investigation show substantial differences in providing social welfare for NRASes. Austria counts as more liberal than the Netherlands; Sweden practices a greater level comprehensive policy transfer within the fields of education and medical care but undertook action to restrict access to housing and allowances for those migrants who do not live with a child. Legislation in Austria and Sweden has addressed NRASes in need in the same way as asylum seekers or residents with a status. Regarding subfields, all three countries provide access to health care and education on a policy-transfer basis—that is, services are provided to all groups without differentiation. Restrictions pertaining to housing and financial allowances—subfields that are not specifically regulated at European level—are more greatly elaborated than others. Besides the Netherlands, in which welfare cuts already came into force in the 1990s, a more pronounced restrictive development has characterized Austria and Sweden since 2015. The tendency is to push this group even further to the margins of the welfare state, firstly, by adopting more-fragmented benefits and services, and secondly, by making them more conditional. Access is being constrained. At the same time, legal restrictions contain a range of exceptions, making the situation confused but also allowing for social support. Access is partially enabled. When policymakers start to regulate access to social benefits for non-removed migrants separately, then the scope of those benefits becomes mostly more limited. Restrictive regulations have followed from public contestation and are not only exerted by right-leaning populist actors but also by government coalitions of center and left parties.

Like differentiation between status groups, the setting of conditions signals to the constituency that migrants have to be in a certain way (poor) and have to act in a certain way (cooperate with return authorities) to be eligible for support (Goldring & Landolt, 2013). Political actors use different policy solutions to navigate between restrictive opinions and the social needs of migrants.

The article faces some challenges and limitations. First of all, across Europe the issue of NRASes and their social protection is volatile, a moving target (Lutz, 2018). In particular, in Sweden legal regulations and formal entitlements themselves have changed during the period of data collection. This implies that not only has the legal situation become more restrictive but it is still undergoing a process of restriction. A further and even more important limitation of the article is the problem of absconding among nonremoved rejected asylum seekers. In this article, we were only able to address the situation of those people who show up for social benefits and services. This study does not reflect the social situation of nonremoved persons on the ground at all; rather it is a study of formal entitlements and the politics behind them. As we know from official statistics and several studies, only a small proportion of the people who fall under the category of the deportation gap (difference between the number of deportation orders and effective returns) show up for social benefits and services (Rosenberger & Koppes, 2018).

These limitations are a mandate for further research. In particular, it would be productive to include the views and strategies of migrants to gain a more complete picture of the relevance of diverse policy designs. Moreover, it would be fruitful to study the effects of the very recent legal and practical restrictions on risks and threats such as homelessness, destitution, and disruption to public order.

Funding Statement

Research for this article would not have been possible without the funding granted for the project “Inside the Deportation Gap: Social Membership for Non-Deported Migrants “(P 27128-G11) by the Austrian Science Fund.

Acknowledgments

I am very grateful for the valuable contributions of Ilker Ataç, Sabine Koppes, Victoria Reitter, and Theresa Schütze for their feedback, cooperation, and support in data collection. The author expresses her gratitude to the members of the workshop in Vienna in December 2017 and to the anonymous referees for their constructive comments on the article.