Multiculturalism and welfare state integration: Swedish model path dependency

Abstract

The present article offers an account of Swedish integration policies in the post-war period. The theoretical purpose is to assess Christian Joppke’s hypothesis that recent trends of integration policy convergence have rendered the national model approach analytically useless. The analysis shows that Sweden deviates, in some important respects, from the European trend by not formulating demands that link integration achievements to immigrants’ access to fundamental rights. The conclusion is that the Swedish case does not support Joppke’s hypothesis, but rather indicates that path dependency of national models is a valid explanation to ongoing developments. It is argued that the Swedish exception should be understood as an expression of the persistent impact of a policy logic according to which integration requires that all citizens have equal and universal access to certain fundamental rights. The article builds on general comparisons with European policy developments and uses Denmark as a more specific reference point.

Keywords:

• integration policies
• Sweden
• multiculturalism
• welfare state
• citizenship
• Denmark

Introduction

Since the late 1990s, an integration policy trend has been observed across Europe, interpreted by commentators as a retreat from multiculturalism and a turn towards civic integration. This policy development has provoked an intense scholarly debate concerning the usefulness of national integration models as a tool for cross-country comparison (e.g., Bertossi and Duyvendak 2012, Finotelli and Michalowski 2012, see also the introduction of this special issue). Christian Joppke has argued that Western European states’ policies on immigrant integration are converging to the extent that it no longer makes sense – if it ever did – to think in terms of national models (Joppke 2004, 2007). Others maintain that there is still enough cross-country variation for the national model approach to be valid, in terms of both describing and explaining current integration policy developments (Jacobs and Rea 2007, Koopmans et al. 2012, Banting and Kymlicka 2013, for Eastern Europe see Kušniráková in this issue).

The present article aims at contributing to this debate by discussing the Swedish policy approach, which, together with British and Dutch policies (see Meer and Modood as well as Entzinger in this issue), has commonly been characterised in the international literature as representing a ‘multicultural’ integration model (e.g., Castles and Miller 1993, Freeman 2004, Koopmans et al. 2005). The political processes following the terror attacks in 2001 profoundly changed this situation, and most commentators now agree that multiculturalism in both the UK and the Netherlands is in retreat. However, we find far less information in the literature about how the Swedish position should currently be understood. The ambition here is to help fill this gap.

Formulated more incisively, the purpose of the article is to assess whether the Swedish case supports Joppke’s hypothesis of the end of national models. In the following section, I provide a brief description of immigration to Sweden, just to set the scene. Thereafter follow four sections intended to address the main aim. In the first of these sections, I present an account of how Swedish post-war policies on immigrant integration should be characterised. Is it at all reasonable to talk about a specific ‘Swedish model’ and, if so, what does it contain? Thereafter follow two sections where I analyse how Sweden should be characterised in relation to current policy trends across Europe. Is Sweden, too, converging to the development witnessed elsewhere, or does the country constitute an outlier in relation to this trend? In the last section, I examine what conclusions can be drawn from the Swedish case as concerns the usefulness of the national model approach. To the extent that Sweden represents an exception to developments elsewhere, does this support the hypothesis of national model path dependency? In the conclusion, I discuss how evidence from the Swedish case should be understood in relation to the causes Joppke mentions when explaining the alleged policy convergence across Europe.

The article takes the Western European situation at large as a general comparative point of departure. In addition, Sweden’s Scandinavian neighbour Denmark will be used as a more specific reference point. Denmark and Sweden share a similar commitment to welfare state universalism, but represent two radically different policy approaches to immigrant integration. Arguably, these similarities and differences make the comparative outlook to Denmark useful, both in chiselling out the ‘Swedish integration model’ and in assessing whether this model should be considered helpful in describing and explaining Sweden’s current policy position.¹

Immigration to Sweden

The recent immigration history of Sweden started in the earliest post-war phase with the reception of large groups of refugees, but from the late 1950s into the early 1970s immigration was dominated by labour migrants (cf. also Barou for France and Loch for Germany in this issue). Following the effectuation of an official ‘immigration stop’ in 1972, the inflow of foreign workers was more or less replaced by refugees and family members. From 1972, the share of immigrants from non-European countries also increased from less than 10% (in the period 1945–1972) to on average 40% during the decades that followed, with the exception of the 1990s when Sweden experienced a massive influx of refugees from the Balkans. Immigration to Sweden in the 2000s has continuously been characterised by large-scale family immigration and asylum migration. Sweden is one of the five countries in the European Union (EU) receiving most asylum seekers in relation to its population. The composition of immigrants coming to Sweden has recently become more diversified, due to the EU expansion eastwards in 2004, but also to the Swedish decision in 2008 to open the door to labour migrants from countries outside the EU. Today, 15% of the Swedish population (or 1.4 million) are immigrants, i.e. persons born outside the country. Additionally, more than 400,000 Swedish-born people have a ‘foreign background’ in the sense that both of their parents were born abroad. After Finns, the largest immigrant groups in Sweden include Iraqis, former Yugoslavs and Iranians (Statistics Sweden 2012).

The Swedish model of immigrant integration

In the context of post-war labour migration, Sweden started to develop a specific policy approach to promoting immigrants’ integration into Swedish society already in the 1960s. The new policy was heavily influenced by the universal welfare state principles also renowned as the Social Democratic welfare state regime (Esping-Andersen 1990). This welfare state model is characterised by its application of comprehensive, generous and redistributive benefits and welfare services that are universal in the sense that they are intended for the whole population and not only for particularly vulnerable groups (e.g., Korpi and Palme 1998).

The Swedish approach was that labour migrants should gain immediate access to welfare state membership on the same basis as natives, so as not to undermine core principles of universal egalitarianism. In 1968, it was officially declared that immigrants must be covered by the universal welfare system under the same conditions as the rest of the population. One precondition for this, however, was careful regulation of immigration in relation to the capacity to provide new arrivals with employment, housing, social care and education on the same terms as the rest of the population. Four years later, a comprehensive ‘immigrant and minority policy’, which was to be guided by ‘equality, freedom of choice and partnership’, was unanimously approved by the Swedish Parliament (Government Bill 1975/76:26).

The fundamental idea of the Swedish model for immigrant integration that emerged in the 1970s was that the general ‘integration logic’ of the welfare state was valid also for immigrants and their families. Indeed, the Swedish post-war welfare state should be seen as an integration project in itself, the aim of which was to promote the overall national identity (Borevi 2002). Corresponding to T.H. Marshall’s seminal citizenship theory, the Swedish welfare state was based on the idea that allocation of certain citizenship rights – particularly important among them being social rights to a decent standard of living – played a key role in promoting integration and fostering a sense of belonging in the national community (Marshall 1950). This logic was now extended to the immigrant population. Access to equal rights was regarded as a necessary condition for their integration into Swedish society and was simultaneously required so as not to undermine the universal character of the system (Borevi 2010, 2012).

The flip side of this rights-based integration model was an insistence on certain individual obligations, particularly the expectation of individual self-sufficiency. Characteristic of the Scandinavian welfare state system in general (but most pronounced in Sweden) was combining the idea of universal access to comprehensive and generous welfare services with the goal of full employment and economic growth (cf. Sainsbury 2012). Social insurances were carefully linked to labour market participation, which is part of what is commonly referred to as the work line principle (Halvorsen and Jensen 2004, Hvinden and Johansson 2007). For immigrants, one can say that the obligation to work – the work line – was built into the system of immigration control. In order to get a residence permit, an immigrant had to find a job with adequate terms (Lundh and Ohlsson 1994). Essentially from day one, immigrants became an integrated part of the labour market, thereby embracing the ideal of the ‘citizen as worker’. The regulation mechanism therefore constituted a basic prerequisite for Swedish adoption of an egalitarian immigrant integration approach.

The Swedish immigrant policy of the 1970s also contained rather radical multicultural goals. Aiming to avoid previous forms of ethnic ‘Swedifying’ (directed in particular at the Sámi minority), the policy was now oriented towards affirming and supporting immigrants’ ethnic identities. Besides the goal of fostering social equality, referred to earlier, the new immigrant policy should also promote immigrants’ ethnic identity formation. Immigrants were seen as belonging to minority cultures and languages in Swedish society, a situation that was thought to justify certain compensatory measures. Immigrants and their children were to be given a ‘real possibility to retain their own language, practice their cultural activities and maintain contact with the country of origin’ (Government Bill 1975/76:26). This new ‘multicultural’ aim guided the introduction of new reforms, including support for journals produced in immigrant languages, mother tongue instruction in the public school system and financial support for ethnic organisations. Immigrants were promised active state support ‘to retain their own language, develop their own cultural activities and maintain contact within their original country, in the same way as the majority population is able to preserve and develop its language and its cultural traditions’ (SOU 1974:69 italics added).² In another formulation, it was even stated that the so-called freedom of choice objective implied that immigrants should ‘have the opportunity to choose between their own ethnic affiliation and the Swedish majority culture’ (SOU 1974:69; Government bill 1975/76, italics added).

Comparing the Swedish approach with that of Denmark, we can first note some basic similarities. The Danish policy response to post-war immigration (which occurred much later, and in smaller quantities than in Sweden) was informed by a similar idea of welfare state egalitarianism and need for careful immigration control. Denmark also introduced a number of innovative Swedish reforms, such as the right of resident foreign citizens to vote and run as candidates in local elections, mother tongue training in the public school system and financial support for immigrant organisations (Togeby 2003, Brochmann and Hagelund 2012, Sainsbury 2012).

Still, the immigrant policy approaches of the two countries were ‘worlds apart’ from the outset (Hedetoft 2013). While Sweden explicitly embraced multiculturalism in terms of active state support and recognition of immigrants’ distinct cultures, the Danish approach is better described as assimilationist, or even anti-multiculturalist (Jensen 2010, Holtug 2013, Laegaard 2013). While the Swedish approach taken in the 1970s was justified as an effort to break with earlier assimilationist ethnic minority policies, in Denmark the active promotion of Danish cultural, linguistic and political homogeneity remained the main strategy (cf. Hedetoft et al. 2006). It is therefore symptomatic that calls to make employment benefits or residence permits for immigrants contingent on language requirements were heard in the Danish Parliament already in the late 1970s. A further important difference between the countries is that while the Swedish policy approach was adopted through consensus, in Denmark immigrant issues were politically contentious from the start (Brochmann and Hagelund 2012).

It has been argued that the compelling differences between the Swedish and the Danish cases referred to earlier reflect different national identity perceptions. Ulf Hedetoft suggests that Danish and Swedish welfare systems share a similar commitment to egalitarianism but are constructed on the basis of two different strategies of promoting consensus and social success: ‘In cultural terms, the Swedish model is geared to attempts to engender consensus, whereas cultural and identity-seeking monoculturalism is the implicit precondition for the functionality of the Danish’ (Hedetoft 2013, p. 327). Along these lines, the Swedish and Danish policy approaches arguably constitute two different models of immigrant incorporation, reflecting different ideas about national identity and active citizenship. The Swedish policy model established in the 1970s was characterised by an attempt to combine welfare state universalism with active promotion of cultural diversity, whereas the policy approach in Denmark was to include immigrants in the welfare state on equal footing with natives, on the condition that immigrants adapt to a Danish canon of cultural values and traditions.

It could be added that the different approaches to ethnic diversity in the two neighbouring countries sketched out here do not necessarily mean that Sweden has always been the more diversity-promoting of the two when it comes to concrete policy outcomes (cf. Laegaard 2013, p. 172). In fact, in some areas, the reverse might even be argued to be the case. Denmark has long applied very liberal and generous rules that allow and financially support ‘free schools’ (friskoler), while the Swedish principle (up to the 1990s) has been that pupils should receive their basic education from a comprehensive school that is common to all. Hence, although Danish school policies have never been justified with reference to ‘multicultural’ arguments, they have resulted in immigrants in Denmark experiencing greater opportunities to uphold their distinct religious or cultural identities via special schools than immigrants in Sweden have had. It is telling that Denmark is the country in Europe with the largest share of publicly subsidised Muslim private schools relative to the population size (Jensen 2010, p. 194).

Policy change

Since the late 1990s, citizenship and integration policies in (Western) European countries have undergone considerable change. There has been a new insistence that newcomers acquire the language of the host country and learn about its history, norms and traditions, most conspicuously expressed by the introduction of various types of integration tests at different stages of the immigration process – initial entry, renewed residency and naturalisation (Goodman 2010, cf. also the other contributions in this issue).

Two interpretations of the ongoing policy processes have had a significant impact in the scholarly discourse: first, that these processes constitute a pervasive backlash from multiculturalism (e.g., Entzinger 2003, Joppke 2004, Koopmans et al. 2005), and second, that they represent a general trend of policy convergence, which has rendered previous notions of national integration models obsolete (Joppke 2004, 2007). These claims have not remained unchallenged. Keith Banting and Will Kymlicka maintain that the idea that policy developments constitute a backlash for multiculturalism is based on the assumption that civic integration is inherently incompatible with multiculturalism. They refute the validity of this assumption, pointing to Canada and Australia where multicultural policies have always been combined with strong integration policies, focusing on employment, learning the national language and shared liberal values. As regards developments in Europe, Banting and Kymlicka hold that ‘compatibility depends on the level of pressure brought to bear on immigrants; and the openness of the national identity of the country to diversity’ (Banting and Kymlicka 2013, p. 588). Joppke’s claim that ongoing developments render all discussion about national models obsolete is also rejected with reference to the fact that there is still considerable cross-country variation in the design and content of civic integration measures (Jacobs and Rea 2007, Banting and Kymlicka 2013). The aim of the following discussion is to see whether the Swedish case can provide some insights into Joppke’s two claims.

As described in the previous section, Sweden had set out on a policy track in the 1970s with the ‘multicultural’ ambition to support immigrants’ ethno-cultural identities. The latter ambition came to be subject to critical re-evaluations already in the mid-1980s. Two government-appointed investigations criticised the goal of actively helping immigrants maintain their ethno-cultural identities, arguing that this in effect risked jeopardising their swift and successful integration into the Swedish labour market and mainstream society: ‘Measures of a minority-supportive nature compete in terms of time and money with measures to promote a long-term, cautious and natural adaptation’ (SOU 1984:55, p. 263). The discussion resulted in a declaration, in a government bill from 1986, that the immigrant policy ‘does not aim at supporting immigrants as collective identities’. Measures should solely target individuals, aiming at promoting their integration as full members of society (Government Bill 1985/86:98, p. 17). Thus, immigrants should no longer be regarded as ‘minorities’, a term that is now reserved for groups that ‘have lived a very long time or always in the country in question’. As a consequence, the label of the policy area was changed from the earlier ‘immigrant and minority policy’ to simply ‘immigrant policy’ (Soininen 1999, Borevi 2010, 2012).

A decade later, in 1997, the principle of not directing specific measures at immigrants qua ethnic groups was again declared, in a government bill proposing a new ‘integration policy’. Policy measures targeting ethnic categories were now criticised for working in a stigmatising manner, thereby reinforcing the notion of ‘us’ and ‘them’ (Government Bill 1997/98:16). The government portrayed the 1997 integration policy as representing a paradigmatic shift in focus, where previous policies were rejected as a failure. In substance, however, it constituted little more than a confirmation of the change in course that had already been made in the 1980s (Borevi 2012).

In sum, we can conclude that a downscaling occurred in Sweden in the mid-1980s (and was reaffirmed in the mid-1990s) as concerns the multicultural ambitions that had been formulated a decade earlier. These developments have certain similarities with processes in other European countries where multiculturalism had been explicitly embraced, notably the Netherlands (Entzinger 2003 and in this issue). Developments in Sweden, however, differed from the Dutch counterpart in (at least) three important ways. First, the multicultural critique took place a decade earlier in Sweden than in the Netherlands. Second, the Swedish backing away from multiculturalism came nowhere near the ‘wholesale retreat’ from multiculturalism that occurred in the Netherlands. In Sweden, most measures that were formerly presented as a way to help immigrants retain their cultural identities – notably mother tongue education and support to immigrant organisations – were in the end seen as constituting valuable support to immigrants’ integration into mainstream society and therefore remained intact. Hence, the Swedish multicultural ‘downscaling’ mostly concerned how measures were justified and was intended to send a signal to immigrants that they should not expect any state support for long-term preservation of their ethno-cultural identities. Third, the Swedish case differed importantly from developments in the Netherlands by not combining the critique against multicultural policies with a new surge of efforts to identify and consolidate a sense of national identity, via introduction of various types of integration requirements and tests for new arrivals and other non-citizens. It is in relation to this aspect that Sweden seems to deviate from developments in most other European countries. We will turn to this now.

Sweden in relation to the civic integration policy trend in Europe

‘Civic integration policy’ is the label attached to the ongoing policy changes across Europe since the turn of the Millennium and is defined by the European Council in 2004 as signifying the importance of immigrants integrating more fully into the mainstream of society. The council emphasised employment as a key part of integration, but also declared that ‘respect of liberal democratic values’ and ‘a basic knowledge of the host society’s language, history and institutions’ were indispensable to integration. As a fourth principle, the council mentioned anti-discrimination laws and policies as essential to improved integration (Joppke 2007). As already stated, one of the most characteristic features of the civic integration policy trend is the introduction of integration requirements that new arrivals must meet in order to gain access to various rights.

This widespread use of mandatory courses, assessments and requirements in the context of immigrant integration is heavily informed by a larger activation and workfare paradigm, where stricter demands are being formulated for the allocation of social allowances, and beneficiaries are required to make an active effort to become self-supporting (Ferrera and Rhodes 2001, cf. Joppke 2007, p. 14). Two features, however, importantly distinguish immigrant integration requirements from activation policies in general.

First, the civic integration programmes include demands of a different character than those tied to activation policies intended for natives. New arrivals are typically not only required to become economically self-sufficient and adhere to a prescribed work ethic, but also to acquire a certain degree of acculturation: learning the receiving country’s language; gaining knowledge about its history and societal institutions; adapting to – or indeed, swearing loyalty to – certain norms and values that are perceived as fundamental to the national identity (e.g., Goodman 2010, Orgad 2010). Second, the sanctions involved in civic integration policies are typically more far-reaching than are those tied to activation programmes intended for natives. Integration requirements are often linked not only to access to social benefits, but also to the right to stay in the country or to eligibility for national citizenship (e.g., Bonjour 2010). Hence, an immigrant who fails to comply with such demands runs the risk not only of experiencing material loss or hardship, but also of being denied three types of rights: permission to enter the country, secure residence status and national citizenship status.

Evaluating the Swedish situation, the country seems to constitute an outlier in relation to important parts of this civic integration policy trend. Indeed, the general activation policy trend is clearly visible also in Sweden. Since the 1990s, Sweden has developed introduction programmes for new arrivals that contain strong activation elements (Borevi 2012). Furthermore, since December 2010, the mandatory element has become more pronounced with the introduction of a centralised administration implementing the principle that new arrivals’ non-participation is sanctioned by withdrawing or reducing benefits (Government Bill 2009/10:60), whereas previously the municipalities made these decisions locally (Qvist 2008). Hence, the introductory programmes contain a certain de facto mandatory element, as programme participation becomes compulsory for those unable to support themselves financially (cf. Hagelund and Brochmann 2010, p. 153).

However, the situation in Sweden deviates from that in many other countries by formulating neither demands nor sanctions that go beyond economic goals or incentives. Hence, while allocation of social benefits is linked to participation in introduction programmes, individuals are not required to pass any integration or language tests; individuals failing to attend the programme do not run the risk of being denied a prolonged residence permit. A similar approach can be identified in relation to Swedish citizenship regulations. Sweden does not apply any official language requirements for citizenship. There is also no knowledge-of-society requirement to be eligible for Swedish citizenship; one is required neither to complete a civic education programme nor to document one’s knowledge by passing a test. The stipulated residence requirement of five years, introduced in 1975, still applies, and in 2001 Sweden introduced an important liberalisation of its naturalisation rules, in the form of a right to dual citizenship.

In comparison, Denmark stands out as Sweden’s absolute opposite. Denmark belongs to one of the countries in Europe that has gone furthest in introducing civic integration policies, as regards both formulating rigorous integration demands and linking integration achievements to immigrants’ access to rights. To gain permanent residence status in Denmark, new arrivals must have participated in the mandatory introduction programme and also obtained a passing mark on a language test plus have a total residence period in Denmark of five years. As regards naturalisation to Danish citizenship, the rules have been sharpened considerably over the past 15 years. For instance, the residence requirement has been raised from seven to nine years, and applicants must declare their allegiance and loyalty to the Danish nation; they must not have received social benefits for more than six months of the preceding five years, and they must pass a language test and a test on Danish culture and history (Ersbøll 2006, Adamo 2008). Finally, regarding integration requirements for family members to gain admission to the country, Denmark has introduced a number of restrictive reforms, including an age limit of 24 years for marriages with third-country nationals and a required bank deposit (of approximately €7400) before family reunification is possible (e.g., Vad Jønsson and Petersen 2012). Denmark also had plans to introduce a ‘points system’ for permanent residence; however, these were cancelled by the new Social Democratic Government that took office in 2011.

In sum, Sweden seems to constitute an outlier in relation to the ongoing civic integration policy trend by not having introduced requirements directed solely at immigrants that go beyond economic incentives and that use as sanctions access to rights to: (1) admission to the country, (2) secure residence status and (3) national citizenship status. Denmark, on the other hand, embodies an exceptionally outspoken variant of the civic integration policy trend, with the development of a number of integration requirements closely linked to access to various rights.

Understanding the Swedish exception as national model path dependency

How, then, should we understand the above-described situation? What conclusions can be drawn in relation to Joppke’s argument that recent policy changes have rendered the national model approach useless as an analytical tool? My conclusion is that the above-described policy developments in Sweden – as well as those in Denmark – do not lend support to Joppke’s hypothesis, but rather suggest that the idea of national model persistence is fruitful, both in describing and explaining current policy positions.

Following Marshall, the core idea of Swedish welfare state universalism is that integration presupposes that citizens enjoy equal access to a bundle of fundamental rights. Rights are particularly crucial for integration, as they provide the necessary integrative glue for the entire society. They enable individuals to regard themselves – and be regarded by others – as full and legitimate citizens (Marshall 1950, Borevi 2010). The current civic integration policy trend entails a true challenge to this integration approach, as it builds on the opposite logic. Here, fulfilment of certain obligations is instead emphasised as the main strategy for promoting integration. Instead of forming part of the integration process, access to rights is intended to work as an incentive and inspiration, a prize that individuals will be awarded after successfully fulfilling the goals formulated in the integration obligations.

In the following, three examples of empirical evidence are presented to support the argument that the persistence of a Swedish model of immigrant integration, where access to rights is seen as indispensable to integration, contributes to understanding the Swedish exception.

The first example concerns the reasoning of the Swedish official inquiry on citizenship, with representatives from all parties in the Parliament, which in 1999 rejected a proposal to introduce official language requirements for Swedish citizenship eligibility. In the opinion of the committee, such a reform was not needed, as good teaching facilities together with the required five years of residence would ensure that the absolute majority of those granted Swedish citizenship also had acquired sufficient language skills. What is more, the committee held that an official language requirement could even work to obstruct integration: ‘the committee regards citizenship as a path to societal cohesion and as an essential part of the integration process. Increasing the qualification demands would instead have the counterproductive result of decreasing cohesion in the nation as a whole’ (SOU 1999:34, p. 318). Arguably, this citation clearly expresses the idea that rights are indispensable not only for the individual integration process but also for sustaining the entire national identity. The Swedish reasoning could be compared with the prevailing view in Denmark, which instead is in line with the core ideas of the common civic integration policy trend. In 2005, when Denmark decided to sharpen its language and conduct requirements, the changes were justified as follows: ‘Danish nationality should be something to strive for; a carrot for foreigners to adapt to Danish society, be independent, learn Danish and be able to socialise with the Danes’ (Ersbøll 2006, p. 131).

The second example comes from the reasoning behind the Swedish decision not to introduce a requirement for citizenship concerning civic skills. The issue was considered by a government-appointed inquiry in 2010. The final report concluded that such a reform should not be introduced in Sweden because this would imply a break with the Swedish model of inclusion: ‘It is not for the state to lay down conditions for citizenship requiring a passed civic education. Just like a democratic state should do for all other citizens, all new citizens should be included generally, blindly and equally, without testing their level of knowledge. Besides, anything else would be a historical breach of the supporting, solidarity-based and inclusive idea underlying the Swedish people’s home and welfare state’ (SOU 2010:16, p. 25).

The third example concerns the 2010 introduction of a requirement that new arrivals be economically self-supportive prior to being granted family reunification. Different from the previous two examples, this reform arguably signifies a clear move towards the European trend of emphasising rights as a reward for integration achievements. It formulates a link between integration achievements and admission rights, and is justified as a method to ‘create an incentive and a motivating force’ for new arrivals to quickly integrate into the labour and housing market. In particular, the requirement is expected to encourage new arrivals to settle in areas with good opportunities for acquiring their own housing and a job (Government Bill 2009/10:77). Still, due to a political process that arguably indicates the persistence of the Swedish ‘rights-based’ integration logic, the practical impact of the reform has been very limited. The proposal was highly disputed and received harsh critique from both opposition parties and within the right-of-centre Alliance Government. One salient argument was that the right to reunite with one’s close family was indeed indispensable to the integration process (Parliamentary Motions 2009/10: Sf11, 2009/10:Sf9, 2009/10:Sf10). As a result, amendments were made to the original proposal, including an extensive list of exemptions from the support requirement, which in the end meant that very few people were actually covered by it.³

Conclusion

The purpose of the present article was to examine whether the Swedish case supports Joppke’s hypothesis of the end of national models. The above analysis suggests a negative answer to this question. Sweden deviates from policy developments elsewhere in a manner that could indeed be understood as national model path dependency. In this conclusion, I will reflect further on how the Swedish exception could be explained. Christian Joppke identifies three main causes for the current integration policy: (1) lack of public support for ‘multicultural’ policies; (2) shortcomings and failures that are linked to such policies, especially socio-economic marginalisation of migrants and their children; and (3) a ‘new assertiveness’ of the liberal state in imposing the liberal minimum on its dissenters (Joppke 2004, p. 243). In the following paragraphs, the Swedish case will be related to these three causes.

Starting with public opinion, it should be admitted that few opinion polls ask specific questions about support for ‘multicultural’ policies. Understood more generally, however, popular perceptions of immigration and integration policies seem to be intimately connected with current policy changes. For example, Entzinger (see in this issue) maintains that the sudden increase in anti-immigrant feelings, notably against immigrants associated with Islam, was one of the main reasons why the Netherlands, in the 2000s, came to adopt one of the most severe assimilation policies in Europe. In this regard, evidence from Swedish opinion polls shows a different development. The share of respondents agreeing that ‘there are too many foreigners in the country’ has steadily decreased, from 52% in 1993 to 36% in 2009, and the support for immigrants’ right to freedom of religion has remained stable since 1993, hovering around 40% (Demker 2010). In a Danish‒Swedish comparison, the authors report that more people in Denmark than in Sweden see immigration as a threat, despite the fact that the share of non-Western immigrants is lower in Denmark than in Sweden (Lolle and Torpe 2010). In sum, public opinion in Sweden does not seem to present the degree of public distrust of immigration that in other contexts has been linked to ongoing integration policy changes. On the other hand, it is difficult to decide whether public opinion should be regarded as an independent or a dependent variable; the increase in anti-immigration feelings among the population in many countries could also be interpreted as a result of changes in elite discourses on immigrants.

Next, Joppke suggests that the increasing salience of immigrants’ socio-economic marginalisation, expressed through immigrants’ high rates of unemployment and growing reliance on social benefits as well as problems connected to residential segregation, can account for the trend of European policy convergence. Problems experienced in Sweden seem to be very similar (or in some respects even worse) compared to the situation in other European countries. For example, Swedish employment rates for immigrants (64% compared to 82% for natives) situate the country at the bottom of OECD countries (e.g., Segendorf and Teljosuo 2011), and there are significant problems associated with residential segregation, exemplified by the 2013 riots in several Stockholm suburbs. The present analysis supports the general idea that insufficient socio-economic integration among immigrants contributes to triggering policy change. In the mid-1980s, the ‘multicultural’ ambitions formulated in the 1970s were significantly downscaled as a direct response to reports of immigrants’ emerging problems with integrating on the Swedish labour market. Further, the growing sense of crisis as concerns immigrants’ lack of socio-economic integration, since the late 1990s, has led to a number of policy reforms that have important similarities with developments in other European countries. However, the present analysis showed that developments in Sweden differ importantly from those elsewhere by rejecting requirements that go beyond economic incentives and that use as sanctions access to: (1) rights to admission to the country, (2) secure residence status or (3) national citizenship status. Hence, Sweden seems to have adapted to a common trend, but in a path-dependent manner.

The third cause that Joppke mentions to explain the current European integration policy trend is ‘a new assertiveness of the liberal state in imposing the liberal minimum on its dissenters’ (Joppke 2004, p. 243). Arguably, this formulation is better understood as a description of one important feature of ongoing European policy developments, rather than as an explanation for why they occur. The present analysis shows that, thus far, Sweden has shown conspicuously little engagement in formulating integration demands to ensure that new arrivals comply with a set of liberal norms and values. Hence, the idea of regarding immigrants as a category that by default should be suspected to be ‘liberal dissenters’ seems to have had much less impact in the Swedish context than elsewhere. While there may be manifold reasons for this, one interpretation is that it could be taken to support the persistence of a Swedish integration model in which the notion of universalism plays one important role.

In sum, the present article has shown that, to the extent that we are concerned with analysing Swedish and Danish integration policies, the national model approach would still seem to be most useful. In particular, the conclusion is that the path dependent impact of the Swedish policy logic, according to which integration requires that all citizens have equal and universal access to certain fundamental rights, contributes in explaining why the country in some important respects deviates from the European integration policy trend.

Acknowledgements

I would like to thank Bo Bengtsson and three anonymous referees for their valuable comments on earlier versions. The article is part of the research project ‘From multiculturalism to assimilation? Swedish integration policies from a European perspective’, funded by The Bank of Sweden Tercentenary Foundation.

Additional information

Notes on contributors

Karin Borevi

KARIN BOREVI is Researcher and Lecturer at the Department of Government, Uppsala University.

Notes

1. The present analysis is based on extensive text analytical empirical research on the Swedish political discourse on immigrant integration during the period 1968–2013, including official inquiries, government bills and parliamentary records (see Borevi 2002, 2010, 2012, 2013).

2. SOU (Statens Offentlig Utredningar) refers to Swedish Government Official Reports.

3. Of all family reunification cases handled by the Migration Board in 2011 and 2012, less than 1% was actually covered by the support requirement (Svenska Dagbladet 3 August 2013).