Sweden used to be one of the most restrictive countries in the Organisation of Economic Development and Cooperation (OECD) in terms of labour immigration policy. Labour immigration was not planned on a large scale and was not used as a remedy for problems of the economy (Boräng Reference BorängForthcoming; Cerna Reference Cerna2009, Reference Cerna2016; Hammar Reference Hammar1985). This was drastically changed in 2008. In May of that year, the centre-right coalition government submitted a new policy proposal to the Riksdag that would be very different from the previous one. The Riksdag approved the proposal, and the new immigration law came into force in December 2008, which made immigration policy more employer-driven and turned Sweden into one of the most open OECD countries. These changes were considered a ‘slight revolution’ (Interview GOV1, 23 May 2008). The central puzzle is why did one of the most restrictive labour immigration countries suddenly become one of the most liberal ones?
In contrast to the existing literature, this article starts from the premise that in order to understand this radical policy shift, we cannot limit the analysis to recent policy developments. First, we have to explain why Sweden was for several decades among the most restrictive OECD countries in terms of labour migration policy. Changes over time in the factors that kept labour migration policy restrictive can then shed light on recent changes in a more liberal direction.
To understand the Swedish case, we argue that it is necessary to consider labour market institutions and their consequences in terms of power relations on the labour market. These factors will influence the preferences, strategies and chances of success for various policy actors – including political parties. Previous research has shown that unions can have considerable influence on labour immigration policy, and that union responses to labour immigration as well as their relative success have differed depending on national contextual factors, especially on the relative power position of unions in a society (Penninx and Roosblad Reference Penninx and Roosblad2000). We also know that the challenge of globalization has caused unions in some settings – notably where unions have not been able to set the terms – to modify their responses to labour migration (Watts Reference Watts2002). However, we know less about the consequences of declining union power for labour immigration where unions have traditionally been very strong and have enjoyed significant power over labour immigration. Here we would expect the most drastic changes, precisely because immigration of labour has been previously kept at extremely low levels.
Sweden is a particularly interesting case study for two reasons. Firstly, Swedish unions have long been extremely strong, both in terms of membership and their position in society. At its peak in 1985, Swedish union membership reached 85 per cent (Kjellberg Reference Kjellberg2009) and the strong corporatist arrangements between employees, employers and the state guaranteed the unions significant influence over policy decisions and implementation (Lindvall and Sebring Reference Lindvall and Sebring2005). Moreover, unions were able to form strong alliances around labour immigration – between blue-collar and white-collar unions and among unions, political parties and government agencies. Secondly, union strength has, during the last decades, been dramatically reduced (Kjellberg Reference Kjellberg2015): the decline of corporatism in Sweden has deprived unions of privileged access to decision-making and implementing bodies (Lindvall and Sebring Reference Lindvall and Sebring2005) and the alliances around labour migration policy have weakened.
In this article, we aim to show that this has had significant consequences for labour immigration policy. Unions, and in particular the blue-collar union federation Landsorganisationen (LO – Swedish Trade Union Confederation), had exceptional power over labour migration policy for several decades. In addition, LO collaborated extensively with the government on migration, which highlights a distinctive nature of the Swedish system. Since unions and their coalition partners de facto controlled labour migration policy, there was little scope for other actors to shape policies in this domain. Consequently, during this time Sweden was one of the most restrictive OECD countries in terms of labour immigration policy (Boräng Reference BorängForthcoming; Cerna Reference Cerna2009, Reference Cerna2016; Hammar Reference Hammar1985).
However, as union power weakened, opportunities for other actors to formulate policy alternatives and to reform policy opened up. Employers and centre-right parties became more active and adopted more liberal policy positions than previously. This process culminated with the new 2008 labour migration law which made immigration policy more employer-driven and turned Sweden into one of the most open OECD countries. These changes were considered to be a ‘slight revolution’ by the former Swedish immigration minister (Interview GOV1, 23 May 2008).
The findings are based on official documents, position statements of relevant actors, party manifestos, media coverage and the literatures on labour market institutions, immigration and political representation. They are complemented by 16 original interviews with elites which offer new qualitative data as the existing literature draws sparingly on interviews. Interviewees included the former minister of immigration; officials of the Migration Board, Public Employment Service and other ministries; representatives of unions (LO, Tjänstemännens Centralorganisation (TCO – Swedish Confederation of Professional Employees) and Sveriges Akademikers Centralorganisation (SACO – Swedish Confederation of Professional Associations)), employers’ associations (Svenskt Näringsliv (SN – Confederation of Swedish Enterprise)) and political parties. The interviews were conducted in Stockholm during May–July 2008 under Chatham House rules. Therefore, interviewees are referred to using a code (e.g. GOV1). The interviews serve three objectives. First, they complement the position statements of parties, ministries, unions and employers’ associations. Second, they are key for identifying alliances between actors and confirming actors’ preferences. Third, they assess the extent to which political parties and interest groups explain the preferred and then implemented policy output.
The next section presents the main argument about the influence of labour market institutions and union power on labour migration. The empirical part first discusses Swedish labour migration policy from the 1960s in light of the theoretical argument. It goes on to elaborate on the recent policy changes, arguing that they should be understood against the background of the more profound shifts that have taken place in terms of union power and reorientation of policy positions. The last section offers some concluding remarks.
LITERATURE AND ARGUMENT
The main argument is that for a long time, the institutional setting of the labour market enabled union influence and constrained party politics in the area of labour migration. This introduced a bias in the policymaking process, making restrictive policy likely to prevail. The more recent liberalization of Swedish immigration policy towards third-country nationals should therefore be seen against the background of the profound changes within labour market institutions in Sweden, including a severely weakened union position.
Before developing the argument further, we consider some alternative explanations to the recent changes in labour immigration policy. The following explanations are potentially relevant for the Swedish case: (1) economic and demographic factors leading to labour shortages; (2) policy diffusion following developments in other European countries; (3) the impact of globalization on unions; and (4) the role of political parties, including their relative power and party ideology.
The most common explanation for the liberalization of labour immigration is that policymakers aim to solve the problem of labour shortages that result from either temporary booms or an increased structural demand following, for example, a changed demographic structure (Boswell and Geddes Reference Boswell and Geddes2011; OECD 2011). However, the changes in Swedish labour immigration policy cannot be sufficiently explained by these economic factors. The previous restrictive policy prevailed through periods of very low unemployment and severe labour shortages, which makes Sweden fit well into the description by Bartram (Reference Bartram2005) of those OECD countries which ‘have followed a different path when confronted with labour shortages, refraining from large-scale use of foreign labour – belying a widespread scholarly perception that importing labour is an inevitable part of capitalist development’ (Reference Bartram2005: 2). Additionally, the new liberal policy was introduced and defended in the midst of a global economic crisis that led many countries facing similar or more severe demographic challenges to limit labour immigration (Chaloff et al. Reference Chaloff, Dumont and Liebig2012).
Similarly, the liberalization of Swedish labour immigration policy can hardly be explained by policy diffusion in terms of developments in other European countries, since, as Berg and Spehar (Reference Berg and Spehar2013: 142) have noted, Sweden was ‘swimming against the tide’, liberalizing labour immigration policy at a time when many other European countries became more selective and restrictive.
Another explanation is close to the one in this article since it concerns the role of unions in labour immigration policymaking. While this article argues that a change in union power is an important predictor of liberalization, this alternative argument suggests that changes in union preferences explain liberalization: as a consequence of globalization and need for labour immigrants, union leaders have begun to support more open immigration policies (Haus Reference Haus2002; Watts Reference Watts2002). We engage with this account of the role of unions in immigration policymaking in both the theoretical and empirical sections. As we shall see, the empirical evidence suggests that changed power, rather than changed preferences, explain the policy shifts better in this case.
Finally, another strand of research points to political party and party ideology explanations (Berg and Spehar Reference Berg and Spehar2013; Green-Pedersen and Krogstrup Reference Green-Pedersen and Krogstrup2008; Green-Pedersen and Odmalm Reference Green-Pedersen and Odmalm2008; Odmalm 2011). Since parties and their policy positions have an obvious impact on immigration policymaking, we engage with this explanation throughout the article. We aim to show that while political parties certainly play an important role, policy positions and actions do not take form independently of the institutional framework in which the parties act.
In a study of Sweden’s current labour migration regime, Berg and Spehar (Reference Berg and Spehar2013) highlight the role of political party preferences based on ideological party differences. In contrast to their study, which analyses policy developments during the 2000s, our analysis covers several decades. As we shall see below, this longitudinal analysis shows that the preferences of political parties – as well as the extent to which political parties give attention to the issue – vary considerably over time, highlighting how party preferences and priorities are in turn shaped by other factors. Taking a longer time perspective allows us to demonstrate how party ideology and struggle have mostly played a surprisingly limited role in Swedish labour immigration policymaking. It is only more recently that labour immigration has been politicized by parties, thus increasing the importance of party ideology as an explanatory factor.
In this sense, labour immigration policy fits well into the description of Swedish party politics – including immigration policymaking – becoming more, rather than less, conflictual over time (Loxbo and Sjölin Reference Loxbo and Sjölin2016). Also, the immigration issue and the generated cleavages are not sufficient to transform the existing party system (adapted from Hiscox Reference Hiscox2001). For example, a newly elected centre-right government will not deal first and foremost with changing labour immigration policy. It was elected on a different platform but, due to its link with employers’ associations, it will also impact labour immigration policy (Cerna Reference Cerna2014). The sections below discuss how this can be related to the argument in this article.
The ‘classical’ view of unions and immigration states that unions are expected to resist immigration since a surplus labour pool, particularly of non-unionized labour, undermines union power and could depress wages (Castles and Kosack Reference Castles and Kosack1973). Yet, Penninx and Roosblad (Reference Penninx and Roosblad2000) show that unions can pursue different strategies, depending on the degree of union power.
When unions refrain from trying to stop labour immigration, and instead bargain about the terms of that immigration, they can choose between two principal strategies: (1) demanding equal rights – including wages and working conditions – for all workers; or (2) steering labour immigration towards particular sectors – usually those with working conditions that native workers shun (Boräng Reference BorängForthcoming). The first strategy evens out differences between foreign and native workers, and will lead to a situation where employers do not have an obvious reason to prefer foreign labour. The second strategy has often included granting only temporary residence permits for employment in a specific sector or with a specific employer, thus making foreign labour equal to flexible labour. This is likely to render foreign and native workers complementary, and thus contribute to a certain structural demand for immigrant workers. Of these two strategies, the first will likely lead to the smallest number of labour immigrants in the economy, but it is probably the most difficult one to pursue (Boräng Reference BorängForthcoming).
The success of attempts to resist labour immigration and to influence the conditions under which it takes place is probably dependent on the union’s institutional strength. This strength stems both from organizational factors – such as a large membership (Rose and Chaison Reference Rose and Chaison1996) – and from the institutions affecting unions’ position in society, such as a privileged position in the institutional framework governing the labour market. This is the case when unions are included in the process of implementing labour immigration policy, are part of centralized bargaining or have a large impact on overall labour market conditions (Boräng Reference BorängForthcoming). Corporatist arrangements are crucial as they incorporate unions into the policymaking network (Katzenstein Reference Katzenstein1985).
Moreover, power could stem from alliances between different unions or strong links between labour market actors and political parties. Concerning alliances between unions, it can be assumed that actors have different policy preferences regarding high- and low-skilled immigration. Native high-skilled labour has restrictive high-skilled immigration preferences as high-skilled immigrants constitute labour market competition for them. Native low-skilled labour is likely to support more open high-skilled immigration, as these workers can benefit from increased job opportunities (Chiswick Reference Chiswick2005). However, unions representing high- and low-skilled labour can form an alliance against employers for more restrictive policies if they can agree on a bargain. For example, restrictiveness towards high-skilled immigration is offered in return for the protection of low-skilled labour against low-skilled immigration or for support of their efforts for higher wages (see Cerna Reference Cerna2009, Reference Cerna2016).
Turning to the links between labour market actors and political parties, it has been shown that in countries with strong union and left party linkages, labour can gain representation for their preferred policy (Kitschelt Reference Kitschelt1994). In sum, it is expected that a strong union position – granted by organizational factors, labour market organization or by unions’ ability to form powerful alliances – will contribute to restrictive labour immigration policies. If unions are strong enough, they can even control the issue.
What happens as union power decreases? Haus (Reference Haus2002) and Watts (Reference Watts2002) suggest that unions have changed their stance considerably regarding labour migration. Against decreased union power and globalization, unions have built alliances with employers in accepting labour migration, but have tried to set salary levels and working conditions because flows of migrants could not be stopped with restrictive policies. In contrast to previous periods, union leaders have begun to support more open immigration policies and recently unions rather than employers have been the driving force behind efforts to liberalize immigration policy (Watts Reference Watts2002). This is due to the unions’ need to develop more effective organizational strategies and diminish migrants’ precarious economic and legal conditions stemming from restrictive policies.
These studies – focusing on unions in Spain, Italy, France and the US – provide important insights into the changed role of unions in labour migration. Recalling that changed union positions were put forward as a potential alternative explanation to changes in labour immigration policy, could such changes explain the Swedish case? We suggest that we should not expect these processes to unfold in the same way in places such as Sweden, where unions have traditionally been very strong and have enjoyed an effective veto over labour migration (see Bucken-Knapp Reference Bucken-Knapp2009).
Under conditions where unions influenced labour migration from an extremely privileged position, we would expect that, when challenged, they try to defend their veto position rather than move to a strategy where immigration is accepted but the terms are negotiated. The reason is that compared to contexts where unions have always negotiated labour migration from a weaker position, losing control over labour migration would be perceived as a greater impairment compared with the previous position. We therefore expect that shifts in power can better explain the changes in Swedish labour immigration policies than changes in union preferences or strategies. This will also be further analysed in the empirical section, where we look at indicators of relative importance of union policy positions and strategies versus union power.
Additionally, when unions are so dominant that they have an effective veto over migration policy, this is likely to function as a strong disincentive for other actors – not least political ones that could potentially demand more liberal policies – to engage in this policy area, since costs are high and the chances of success are very limited. If, however, there is a shift in union power, it is likely to impact other actors’ positions and chances of success.
Consequently, we expect that as union power diminishes, it will facilitate more liberal policy shifts – not so much by changes in union strategies as by shifts of other actors’ positions as well as their relative chance of success. These changes would be more dramatic since unions in these contexts were able to keep labour immigration artificially low. The next section examines the strong position of Swedish unions and its consequences for labour immigration policy.
UNIONS AND LABOUR IMMIGRATION POLICY IN SWEDEN
Why was Swedish labour migration policy so selective for decades, even during times of labour shortages? Unions were traditionally powerful in immigration policymaking due to their generally strong position in Swedish society. For a long time, Sweden had among the highest – or the highest – levels of union membership in the world (Visser Reference Visser2013). Moreover, Sweden has often been used as a key example for corporatist policymaking due to the power of employees and employers (Kenworthy Reference Kenworthy2003: 11). Institutional constraints have affected actors as immigration policies were negotiated in a corporatist setting between the state and the main social partners (Dahlström Reference Dahlström2004; Freeman Reference Freeman1995; Hammar Reference Hammar1999). They thus followed how other policies were proposed and decided upon (Lindvall and Sebring Reference Lindvall and Sebring2005).
Notably, Swedish unions have often had the upper hand – not least in the area of labour migration policy – due to the strong foundations provided by the corporatist setting. This can also be attributed to the strong coalition between the Social Democrats (Sveriges Socialdemokratiska Arbetareparti – SAP) and LO (Lundqvist Reference Lundqvist2004). High- and low-skilled labour in Sweden managed to keep an alliance over a long period of time, which added to unions’ strength. Two movements exist: a social democratic blue-collar union movement and a politically neutral (unitary) white-collar union movement, made up of the white-collar union confederation TCO and the professional/academic union confederation SACO (Kunkel and Pontusson Reference Kunkel and Pontusson1998). The main (low-skilled) union confederation LO was supported in its efforts by TCO and SACO.
The national agency Arbetsmarknadsstyrelsen (AMS – Swedish National Labour Market Board) was also part of the alliance around labour migration. The link between LO, the Social Democrats and AMS at the time has been described as an ‘iron triangle’ (Lindvall and Sebring Reference Lindvall and Sebring2005: 1063). AMS administered the labour migration policy of 1968, and the Social Democrats and AMS repeatedly supported LO’s position on labour migration (Waara Reference Waara2012: 145). In fact, for a long time, Swedish unions enjoyed de facto veto power over labour immigration (Brochmann and Hammar Reference Brochmann and Hammar1999; Hammar Reference Hammar1985). In sum, unions had a very privileged position in migration policymaking, stemming both from the corporatist setting and its alliances with other influential actors. As shown below, this has been of great importance in the making and persistence of restrictive labour migration policies.
Union Power and the Development and Maintenance of Restrictive Labour Immigration Policies in Sweden
The end of the Second World War left Sweden in great need of labour since economic growth was high and unemployment low (Wadensjö Reference Wadensjö2009). The Labour Force Commission started to recruit workers in Italy, Hungary and Austria in 1947. In the 1950s, it continued in West Germany, the Netherlands, and later in Italy, Belgium and Greece (Wadensjö Reference Wadensjö2009). In the 1960s, agreements were concluded with Yugoslavia and Turkey (Ministry of Justice 2001: 15–16). In 1965, a government commission for the labour market argued in favour of extensive labour immigration since the demand for labour in Swedish industry was still considerable (Ministry of Justice 2001).
However, Sweden’s relatively liberal policies ended in the late 1960s when measures for regulated immigration were introduced. Previously, people were allowed to spend three months in Sweden seeking employment, but a supplement to the Aliens Ordinance prescribed that a work permit should be granted before entry into Sweden (Lundqvist Reference Lundqvist2004: 3; Ministry of Justice 2001: 34). The following year, the Swedish parliament passed the 1968 Immigration Act, abolishing the liberal immigration system and requiring work permits for all labour immigrants prior to entry (Lundqvist Reference Lundqvist2004: 3). The employment situation and the availability of housing, education, health care and other services would determine the extent of labour immigration (Lundqvist Reference Lundqvist2004). Labour immigration was only to be allowed in exceptional cases (Ministry of Justice 2001: 37–9). Notably, these changes in Swedish labour immigration policy preceded the restrictions on labour migration which were introduced in several other European countries (including France and Germany) in the early 1970s.
The power of unions, and particularly the LO, was very important for the way this policy area developed. First, the 1967 measures were partly ‘a result of energetic demands by the trade union movement’ (Ministry of Justice 2001: 16). During the debate surrounding the 1968 Immigration Act, which abolished the previously liberal immigration policy, the employers’ association for the private sector, the Svenska Arbetsgivareföreningen (SAF – Swedish Employers’ Confederation), protested against the suggested changes and argued in favour of a continued liberal policy – but they were unsuccessful when facing the ‘iron triangle’. AMS and the Social Democrats supported the union position (Waara Reference Waara2012: 145), and Swedish labour immigration policy changed from liberal to restrictive. The fact that the SAF lacked equally powerful allies has been highlighted as an explanation for employers’ failure in promoting liberal labour migration rules (Waara Reference Waara2012).
As a consequence of the new 1968 regulation, Sweden’s labour immigration became very selective and its labour market was almost closed to non-Nordic labour migrants from 1972. Only 300–400 permanent labour immigrants from outside the Nordic countries (later EU/EEA) were admitted into the country each year. Labour immigration constituted the smallest percentage of all migration types since the 1970s; instead, family reunions and asylum-seekers dominated immigration inflows (Ministry of Justice 2001: 17). Temporary work permits were given for all skill levels but were granted mostly to experts and key personnel in industry, research, culture and sports (Ministry of Justice 2001: 30). Although the numbers of temporary work permits were higher than the permanent residence permits for workers, and increased over time (at least from 1994 onwards; see Swedish Migration Board 2011), the total yearly numbers of labour migrants were comparatively very low until the 2000s (Boräng Reference BorängForthcoming).
Already before the 1970s, unions had managed to impose heavy conditions on the recruitment of foreign workers (Brochmann and Hammar Reference Brochmann and Hammar1999: 176; Hammar Reference Hammar1985; Lundqvist Reference Lundqvist2004). The LO was opposed to a guest-worker system because it tended to upset conditions for those in regular employment and risked endangering union influence at the workplace (Lundqvist Reference Lundqvist2004: 11). Swedish labour migrants who wished to settle were given permanent residence permits and labour migrants were also allowed to bring family members (Borevi Reference Borevi2012). This is all the more remarkable considering that for European countries that did not have colonial immigration, guest-worker schemes were, at the time, the norm (Hansen Reference Hansen2016).
Consequently, unions adopted the view that immigrants should enjoy the same standards as Swedes in terms of housing, education and social benefits. If that was not available, immigrant labour would have to wait. Companies would not be allowed to pass the costs on to the taxpayer (Lundqvist Reference Lundqvist2004: 8). As a result, the LO advocated controls that would enable the unions to deny work permits. Employers who failed to meet union standards on pay or working conditions could expect to be denied permits. According to Lundqvist (Reference Lundqvist2004: 7), unions were ‘particularly tough on low-wage companies’.
Moreover, immigrant workers were to have the same working and wage conditions as native workers, and in the case of unemployment, foreign workers were to enjoy the same unemployment benefits as all other union members (Hammar Reference Hammar1999: 175). LO also demanded – and forced the employers to accept – mandatory union membership for immigrant workers (Waara Reference Waara2012).
LO thus went for the first strategy discussed in the theoretical section: namely to demand wages, conditions and social rights for foreign workers equal to those of native workers – and, with the support of AMS and SAP, it was successful. This functioned as a guarantee that migrants would not be paid less than others at the same skill level, which could undercut wages (see Hansen Reference Hansen2016). The introduced policy was for a long time not seriously challenged. Overall, no attempts at liberalization took place in the 1990s (Cerna Reference Cerna2014; Ministry of Justice 2001). This restrictive approach to labour migration stands in stark contrast to the Swedish approach to humanitarian migration, which used to be among the most open in the OECD.Footnote 1
The theoretical section argued that when unions have an effective veto over migration policy, this will probably have an impact on the preferences and strategies of other actors. Therefore, we now turn our attention to the political parties and their preferences until the beginning of the 1990s, when the labour market institutions in question still remained intact.
If unions in practice controlled a policy area, it would serve as a disincentive for other actors to engage in this field. It is thus important to note that political parties used to be very passive on the issue of labour migration. When parties did discuss migration, they usually focused on refugees and family migrants (Demker and Malmström Reference Demker and Malmström1999). The labour immigration issue was virtually non-existent in election manifestos and election debates until the 1990s (Brandorf et al. Reference Brandorf, Esaiasson and Håkansson1996: 23f). At least until the 1990s, there was a considerable degree of cross-party consensus on immigration (Dahlström Reference Dahlström2004; Green-Pedersen and Krogstrup Reference Green-Pedersen and Krogstrup2008; Green-Pedersen and Odmalm Reference Green-Pedersen and Odmalm2008: 371; Hammar Reference Hammar1999; Odmalm Reference Odmalm2011). This description is sustained by a recent longitudinal analysis of parliamentary opposition in Sweden which shows that before the 2000s it was uncommon for opposition parties to reject government bills on immigration/integration policy (Loxbo and Sjölin Reference Loxbo and Sjölin2016).
For many years, the ruling Social Democrats (and the left) followed the unions’ view and were opposed to changes in labour migration (Interview POL1, 29 May 2008; Berg and Spehar Reference Berg and Spehar2013). While party manifestos discussed refugee migration extensively, references to labour migration were kept minimal. One explanation is that labour immigration – unlike refugee immigration – involved ‘bartering with the trade unions’ (Odmalm Reference Odmalm2011: 1082).
Though it is unsurprising that left parties supported the unions’ view, it is notable that centre-right parties – in contrast to their present liberal positions on labour migration – during this period often also held views that were not radically different from those of the unions. For example, the present Centre Party position, which is very pro-labour migration and strongly challenges the position of LO (Berg and Spehar Reference Berg and Spehar2013), stands in stark contrast to its party manifestos from 1970 and 1981, where the union position was not challenged at all. Similar to the view of LO, the Centre Party in 1970 stated that ‘[f]oreign labour shall get help to adjust to Swedish society and be given equal rights as native labour’ (Centre Party 1970). The 1981 party manifesto argued that, with the exception of refugee migration and Nordic migration, immigration must be regulated and ‘based on the possibilities to create the same economic and social security for immigrants as for Swedish citizens’ (Centre Party 1981). The party emphasized that whereas labour migration could only be accepted if it was compatible with the labour market situation and availability of housing, refugee policy should be ‘generous’ without such constraints (Demker and Malmström Reference Demker and Malmström1999).
As the 1968 selective policy was about to be introduced, one could expect that the party with the closest ties to the employers’ federation – the Moderate Party (Moderaterna, until 1969 Högerpartiet) – would have supported the employers’ attempt to keep the liberal policy. However, the party generally agreed with the view in the 1968 government bill that immigration had to be regulated, and it specifically mentioned the need to give immigrants an equal standard and level of security to that of natives. The party also supported the labour market partners – both employers and unions – having a strong influence over immigration (Moderates 1968). In the party manifestos, little explicit attention was given to labour migration. Additionally, the question was not mentioned in the election manifestos during the 1970s and 1980s, which was remarkable given the very low unemployment during parts of this period (Moderates Various years).
Similarly, the Liberal Party did not object when the selective policy was about to be introduced in 1968. In line with the unions’ view, immigration had to be adjusted to the need for labour in order to avoid increased unemployment or a downward pressure on wages (Liberal People’s Party 1968).
Overall, very little activity and open conflict existed around the issue. This was visible in that no changes in labour immigration policy occurred when centre-right parties were in government: 1976–82 and 1991–4, and there were no serious attempts to liberalize labour immigration policy until the 2000s.
In sum, unions – especially the LO – had a strong influence over labour migration policy. Political parties were largely inactive on this issue, and there were no serious challenges to the union position. This follows the argument that if an actor or group of actors have a de facto veto power over an area, this could serve as a disincentive for other actors to actively promote alternative policy positions.
The Weakening of Unions and Recent Migration Policy Developments
As described above, union power over migration policy in Sweden was based on a number of factors: high levels of union membership, corporatist institutions, and unions’ ability to form strong alliances around the issue of labour migration. Over the last decades, however, union strength has been weakened in all three respects.
While the unionization rate in Sweden is still comparatively high, there has been a significant decline over the last 20 years. Union membership dropped from 85 per cent in 1994 to 71 per cent in 2008 (Kjellberg Reference Kjellberg2015: Table 31). The decrease in union membership was intensified after 2006, due to the centre-right government’s reform of the finances of unemployment insurance (Kjellberg Reference Kjellberg2011), but was significant even before this reform. Importantly, changes in membership differ substantially between the different union movements, changing the relative unionization rate between the blue-collar and the white-collar movement. The decline in membership has hit LO particularly. In 1987, LO had 2,021,998 members, a figure that dropped to 1,564,409 by 2006. TCO saw a much less dramatic decline during the same period, and SACO significantly increased its membership (Kjellberg Reference Kjellberg2015: Table 34). Among young blue-collar workers, union membership fell from 75 per cent in 1995 to 36 per cent in 2010 (Larsson Reference Larsson2010).
Over the last two decades, significant institutional changes in Sweden also took place (Lindvall and Rothstein Reference Lindvall and Rothstein2006; Lindvall and Sebring Reference Lindvall and Sebring2005). First, there is the decorporatization of public administration boards. The system of interest organizations being represented as board members in public agencies was abolished in 1991, after that the employers’ confederation (SAF) unilaterally decided to withdraw from the system. For LO, this meant a substantial loss in terms of representation on boards and committees. It also reduced the legitimacy of the national administrative agencies, such as AMS (Lindvall and Rothstein Reference Lindvall and Rothstein2006: 51), thus weakening the ‘iron triangle’ of LO, SAP and AMS.
In addition, the alliance around the labour migration issue within the union movement had weakened. SACO, and to some extent TCO, deviated from LO’s view and adopted a more liberal position towards labour migration rules (Interview UNI2, 23 May 2008). For instance, SACO (and also TCO) supported more open immigration policies for low-skilled workers, which could be seen as complementary to their (high-)skilled members. In contrast, LO still resisted a more liberal immigration policy as the reform was expected to affect mostly its low-skilled members (Interview UNI1, 9 June 2008).
In sum, membership losses, decorporatization and weakened alliances resulted in a very different situation from the previous one in which policy had been formulated, adopted and maintained. Within this institutional framework, actors were more likely to suggest and implement changes to the labour migration regulation than before.
Having withdrawn from the corporatist bodies, SAF sought a new role as a public policy advocate, focusing its efforts on public opinion and lobbying on various political and ideological issues (Pontusson and Swenson Reference Pontusson and Swenson1996). Employers became more active on the labour immigration issue than previously (Cerna Reference Cerna2014; Johansson Reference Johansson2012). In late 2001, the employers’ association, the Confederation of Swedish Enterprise (Svenskt Näringsliv – SN) tried to ease labour migration restrictions. A 2003 report by the SN presented the private sector employers’ view of a need to allow more immigrant workers into Sweden, under the condition that they had a job.
Nonetheless, the regional labour market boards, the Social Democratic government and unions still criticized these attempts by employers (Interview EMP1, 2 July 2008). The Swedish union movement continued to be concerned about the risk of ‘social dumping’, i.e. that immigration might lead to the deterioration of wages and working conditions for Swedish residents (Storrie Reference Storrie2005; Interview POL1, 29 May 2008). Even though the main union confederation, LO, has mostly worried about low-skilled immigrants, it has tried to resist any change in labour immigration policy (Interview UNI1, 9 June 2008). However, the interviews revealed that SACO was more concerned about open high-skilled immigration than low-skilled immigration (Interview UNI3, 22 May 2008). This follows our expectations about actors’ preferences: SACO represents high-skilled workers and thus preferred to restrict high-skilled immigration in order to protect its own members.
The centre-right parties, who had previously given little attention to labour migration, became more active around 2002 (for overviews of the positions of different parties, and alliances between them, see Berg and Spehar Reference Berg and Spehar2013; Bucken-Knapp Reference Bucken-Knapp2009; Cerna Reference Cerna2009). In 2003, SAP could not block an alliance between four centre-right opposition parties and the Greens; demanding a parliamentary committee on labour migration to review the immigration legislation. The committee (called KAKI in Swedish) consisted of representatives of all seven parliamentary parties, as well as some experts from relevant ministries, the Migration Board and AMS (Storrie Reference Storrie2005). The 2006 Committee Report called for a certain liberalization of labour immigration (EMN 2007; Ministry of Justice 2006).
During this period, the question of intra-EU migration also became important. Sweden was one of the three countries that allowed workers from new EU member states to access its labour market from 2004. The decision was preceded by heated discussions between unions, employers and political parties (see Berg and Spehar Reference Berg and Spehar2013). SN’s more liberal views clashed with Social Democrats in power and LO, who supported transitional measures (OECD 2011). Labour market access to new EU workers was granted in the end because the Social Democrats did not receive a parliamentary majority.
After the 2006 election – when the centre-right Alliance (including four parties: Moderates, Centre Party, Liberal People’s Party and Christian Democrats) had formed the new government – clear steps towards a more liberal legislation of third-country labour migration took place. The government, together with the Greens, put forward a proposal (Prop. 2007/08:147) that was largely based on the committee’s recommendations, but reflected the pro-capital side of the (centre-right) government in which employers could decide on labour market needs and apply for work permits without union approval (Interviews GOV1 and GOV2, 23 May 2008). The parliament voted on it in November 2008, and the law came into force on 15 December 2008.
The new law appeared very pro-capital as it was up to individual companies/employers to determine labour market needs. The demand-driven system allowed labour immigrants of all skill levels with a job offer to come to Sweden. The Swedish Migration Board examined whether conditions (such as salary and terms of employment) corresponded to the Swedish collective agreements. While the board used to have a more administrative function, it is now in charge of examining applications and granting work permits (Interview GOV3, 27 May 2008). In contrast, the Public Employment Service (previously AMS) no longer played a role (Regeringen 2008; Interview GOV4, 29 May 2008).
The 2008 law changed the unions’ role drastically. For almost four decades, unions had exercised a de facto veto – their opinion was required and consolidated practice treated a negative union opinion as a veto (OECD 2011; Schierup and Ålund Reference Schierup and Ålund2011; Interview GOV4, 29 May 2008). Since the new law entered into force, some claim that ‘the unions’ role is reduced to that of toothless consultation’ (Schierup and Alund Reference Schierup and Ålund2011: 59). More specifically, unions could be asked for their advice and approval in some cases of influence, but not in every work permit case (Interview UNI1, 9 June 2008). The OECD (2011) conducted an evaluation of the new policy, which stated that ‘Sweden has become one of the most open countries for labour migrants’ (OECD 2011: 69).
Importantly, and in contrast to the findings in Haus (Reference Haus2002) and Watts (Reference Watts2002), there was little evidence of the LO changing its strategies or positions. Certainly, there was no ‘unlikely alliance’ between LO and SN. The strategies and position of LO thus remained similar to the previous ones, but with the profound changes that had taken place – the decorporatization of public administration boards, the declining union membership, the discontinuation of a strong alliance with high-skilled unions – its influence has diminished.
This decreased influence is also mirrored in how other actors have become more active and challenged LO’s position. The earlier period’s cross-party consensus and limited interest in labour migration stand in stark contrast to more recent policy debates where parties have adopted strong and divergent positions.
Like LO, the Social Democrats did not significantly change their view. Even though they eventually accepted the need for migrant labour due to an ageing population, they were concerned about wage dumping and the threat to welfare as a result of more open labour immigration and wanted to keep the existing system, in which the Swedish AMS determined labour market shortages and consulted unions and employers (Interviews POL1, 29 May 2008; POL2, 22 May 2008; Berg and Spehar Reference Berg and Spehar2013). The Social Democratic Party emphasized that Sweden had to draw on unemployed native workers and immigrants already in the country before opening its borders to more labour immigrants (Interviews POL1, POL2).
However, the positions of the centre-right parties shifted to more liberal views on labour migration (Berg and Spehar Reference Berg and Spehar2013). Real pressure from centre-right parties did not start until 2002 (Berg and Spehar Reference Berg and Spehar2013; Cerna 2014). Whereas little open conflict over labour migration took place in previous decades, the new preferences and activity of the centre-right parties set them apart from the left and the union approach that labour immigration should continue to be restricted.
DISCUSSION AND CONCLUSIONS
The radical change in Swedish labour migration policy, from very selective to very open, has been called a ‘slight revolution’ (Cerna Reference Cerna2016). Drawing on original data and secondary sources, this article has shown how this recent policy change was preceded by more fundamental changes in the Swedish labour market.
For decades, important characteristics of the Swedish labour market – in particular a very strong union position within a comprehensive corporatist setting – worked against a liberal labour migration policy. The centre-right parties, which could have been expected to push for liberalization, showed very little interest in the issue, whereas the employers’ federation occasionally did demand more liberal rules – but were defeated by the ‘iron triangle’ of LO, the Social Democrats and AMS. LO was further supported in its efforts by the issue alliance with the other two union confederations, TCO and SACO.
However, the institutional setting changed, which, together with lower union membership and weakened alliances, resulted in unions’ loss of power over the labour migration issue. Lindvall and Sebring (Reference Lindvall and Sebring2005: 1071) suggest that in Sweden, ‘coordination and mediation between interest organizations has become more politicized – drawn into an ideological struggle between left and right’. This conclusion also seems valid for labour migration policy. From having been an issue with very little conflict, and where centre-right parties could even echo the positions of unions, it has become a topic where these parties position themselves strongly against unions and the left parties. Party struggle thus seems to matter more than it used to, so that party ideology has become more promising as an explanatory factor to immigration policymaking (see Berg and Spehar Reference Berg and Spehar2013). Rather than being characterized by weakening left--right mobilization between the main parties, as has been suggested is the case for some policy areas by Oskarson and Demker (Reference Oskarson and Demker2015), the area of labour migration fits better into the description of Swedish parliamentary politics (Loxbo and Sjölin Reference Loxbo and Sjölin2016) that points to increased politicization and partisan struggles.Footnote 2 While the authors do not aim to explain why conflict has increased, our study suggests that shifting institutional conditions on the labour market opened up for more partisan conflict in labour immigration. Future research should examine whether shifting institutional conditions could also be a valid explanation for other policy areas.
In contrast to Haus (Reference Haus2002) and Watts (Reference Watts2002), we do not find signs of any ‘unlikely alliances’ between unions and employers emerging as a consequence of a weaker union position. Unions and employers’ associations have very contrasting positions regarding labour immigration. This shows that we should not expect processes to unfold in the same way in different contexts when unions are weakened. Unions starting from a very strong position may be much more hesitant to change their strategies or to enter ‘unlikely alliances’ than unions used to negotiating labour migration from a weaker position. We invite future research to examine whether these conclusions hold in a wider set of countries.
Moreover, while Berg and Spehar (Reference Berg and Spehar2013) find that Swedish parties have recently formed ‘unholy coalitions’ in favour of liberal policy, we find that a longstanding consensus existed in favour of restrictive policies among parties with very different ideological views. While we should not underestimate the importance of political party preferences, this article illustrates how the institutional environment helps to shape those preferences.
