Changes in migration control during the neoliberal era: surveillance and border control in Swedish labour immigration policy

Abstract

In 2008, Sweden introduced the most expansive migration policy in the OECD. The reform combines neoliberal elements with old forms of labour control in a new way. In this paper, I compare the immigration policy of 2008 with that of 1954–1972. Labour immigration from 1954 to 1972 occurred in a radically different historical context than the current market-liberal one. One important change between the two periods concerns the surveillance of migrants residing in Sweden. During the earlier period, state authorities conducted thorough surveillance of migrants residing in Sweden. The 2008 immigration policy involves private actors surveilling in a way that is unprecedented since the end of the Second World War.

Keywords:

• migration
• neoliberalism
• Sweden
• Foucault
• governmentality

Introduction

Western European states introduced restrictions on labour immigration during the 1970s, which put a halt to two decades of large-scale labour immigration. Some observers assumed that these states would continue to restrict labour immigration in the foreseeable future. However, in the 2000s, labour immigration from countries outside the EU had once again become an important political issue, and occurred in a radically different historical context than from 1945 to 1975. Labour immigration to Western Europe from 1945 to 1975 was spurred by exceptional economic growth. Migrants were often employed in large factories in the expanding manufacturing sector and workers’ organisations were growing stronger, while in some states left-wing parties were in government for long periods of time. The welfare state was expanding and a fundamental goal of economic policy was full employment. The broader institutional context was the Keynesian welfare state, which was based on a compromise between labour and capital (Jessop 2002, Harvey 2007).

This historical context has been transformed, as labour immigration has once again become an important political issue. Mass-unemployment has reappeared in Europe, and there has been an increase in social inequalities. Workers’ organisations have been weakened, and the capitalist class has abandoned the Keynesian class compromise. This transformation has been accompanied by the diffusion and implementation of neoliberal ideas. A fundamental driving force behind the neoliberal project has been to alter the power balance between classes or, in the words of David Harvey, ‘to restore the power of economic elites’ (Harvey 2007, p. 19).

The prescriptions for neoliberal reform of migration control do not deviate from the guiding principles of neoliberalism, for instance that markets should be freed from political intervention. State involvement in migration should, according to neoliberalism, be kept to minimum. People should be allowed to move as freely as possible across international borders, particularly if there is demand for their labour in another country, and there should be as few restrictions as possible on firms’ ability to recruit foreign labour. It follows that trade unions might become a target of neoliberal policy if they impose restrictions on migration.

The actual impact of neoliberalism on migration control is ambiguous and uneven across countries. Some actors have demanded more liberal labour immigration policies, for example, employer associations. However, what governments in Europe have implemented so far lies quite far from neoliberal demands. Immigration policy is sensitive to other interests than those of employers. The political right tends to be divided on the immigration issue: those that represent employers advocate for expansive policies, while the cultural conservative fraction tends to be more restrictive (Zolberg 1999). Another limitation on the introduction of more expansive policies is that populations in today’s Europe tend to be restrictive. Right-wing populist parties are gaining increasing support, and parties that want to win elections cannot disregard such restrictive tendencies in the population. Furthermore, experiences of labour immigration from 1945 to 1975 have made governments hesitant about introducing highly expansive labour immigration policies. Labour immigration from 1945 to 1975 resulted in large-scale permanent settlement of migrants in Europe, which is a pattern that some governments want to avoid repeating (Ellermann 2013).

However, there is one case where neoliberalism has had a significant impact on labour immigration policy, namely Sweden. In 2008, Sweden introduced the most expansive migration policy in the Organisation for Economic Co-operation and Development (OECD 2011). The reform was introduced by a coalition between the governing right-wing parties and the Green Party. The reform has weakened the influence of trade unions, and it has given employers a significantly larger influence on immigration (Prop. 2007/2008: 147). Furthermore, the reforms decreased the involvement of state authorities in the recruitment of foreign workers.

The radical change in migration control in Sweden cannot be explained solely by established models of immigration policy, for example that liberal democratic states tend to become more open to labour immigration over time (Freeman 1995). The profound change in migration control is influenced by the neoliberal ideological dominance that emerged at the end of the 1970s. In Sweden, some liberalising reform of economic policy was already made in the 1980s. In the 1990s, neoliberalism began to exert a broad influence on public policy reform, for instance, the reform of the school and pension systems (Larsson et al. 2012).

The impact of neoliberalism on immigration policy has not been sufficiently studied. The reason for this is probably an assumption that the restrictive forces are winning out in today’s Europe, which is not completely true. Neoliberal ideas are never implemented in a pure form and in exactly the same way in different contexts. Other factors are always limiting the influence of neoliberalism. I view the Swedish migration policy of 2008 as a bricolage, that is, a recombination of already existing institutional elements in a new way (Douglas 1986, Campbell 2004). The reform combines neoliberal elements with old forms of labour control in a new way.

In the literature on current labour migration to Western Europe there is often an implicit comparison with labour immigration from 1945 to 1975. This period is still a formative one in European migration history (Messina 2007). I will make the comparison with this earlier period explicit. I will also compare the Swedish immigration policy of 2008 with that of 1954–1972. In 1954, an aliens law was passed that was considered very liberal at the time, and in 1972 labour immigration from non-Nordic countries was restricted. One radical change between the two periods concerns the execution of the internal migration control, in other words the surveillance of migrants residing in Sweden. During the earlier period, particularly from 1954 to 1967, state authorities conducted thorough surveillance of migrants residing in Sweden. The 2008 immigration policy involves private actors surveilling in a way that is unprecedented since the end of the Second World War.

In this paper, I will focus on internal migration control and its relationship to external migration control, that is, the control of the inflows of migrants. The specific means used for internal migration control has profound impacts on migrants, ultimately influencing their mobility in the labour market, their opportunities for upward social mobility and their relationship to employers. This paper will answer the following questions: how has internal migration control been carried out in the different time periods? What accounts for differences from one period to the next? How does internal migration control affect migrants and their relationship to employers?

I make two main arguments. The first concerns a general dynamic in migration control, and the second the influence of the neoliberal context. My first argument is that with regard to inflows of labour migrants, expansionist movements are accompanied by strengthened internal migration control, in other words increasing surveillance of migrants who are allowed to enter Sweden. The strengthened internal migration control is to a certain extent a compensation for the opening up of borders to labour migrants. I conceptualise this opposite movement in Foucauldian terms as surveillance and discipline of migrants (Foucault 1977). Labour migrants are treated as objects of surveillance because of their status as outsiders (Bigo 2008). When the presence of outsiders – for instance labour migrants from Asia – has increased, the response has been increased surveillance.

My second argument is that different means have been used to surveil migrants in the different periods. During the period from 1954 to 1972, surveillance was mainly performed by state authorities, following migrants from the point of entry and their movements in the labour market. Since 2008, the state has, in important respects, withdrawn from this surveillance of migrants, and the task has largely been transferred to private actors. Since 2008, it is conducted more by employers and less by the state.

Border control and surveillance

States use two main strategies in their attempts to manage international migration: one focused on the border and the other on migrants themselves. On the one hand, states control the flow of migrants across their borders. This control need not literally occur at the border; it can be a form of ‘remote’ control performed before migrants reach the border (Zolberg 1999). The state might, furthermore, delegate a part of border control to non-state actors.

On the other hand, states control and surveil foreigners residing in their territory. The fundamental aim of this surveillance is to make certain that the migrant has a right to reside in the territory of the state, ensuring that the right to do so has not expired nor been lost due to some act committed by the migrant. In most cases, the right to reside is limited in time, but under certain conditions the migrant might be granted permanent status. The right to reside might be granted to migrants whose skills are needed in the labour market. In this case, an aim of the surveillance of migrants is to ensure that there is still a demand for their labour power; if there is not, they lose their right to reside.

People might be granted residence permits for reasons other than work, such as family reasons or tourism. Furthermore, the right to residence might be dependent on migrants behaving in ways that are deemed acceptable by authorities. In other words, the state defines behavioural rules that migrants must follow if they wish to stay in the destination country. If migrants do not behave in accordance with these regulations, they will lose their right to residence. In migration history a number of different kinds of behaviours have been deemed unacceptable by authorities. For example, a migrant might lose their right to residence if they are unable to keep a job, protest at the workplace, engage in political activism, commit a criminal offence or fail to adapt to the culture of the receiving country. Furthermore, what is deemed to be unacceptable behaviour changes from one context to another, as behaviour considered unacceptable in the 1930s may be considered acceptable in 2010.

The techniques for examining and evaluating the behaviour of migrants are an important part of immigration policy. These techniques examine and evaluate migrants in order to decide if they are acceptable and can eventually assimilate into the culture of the destination country. In this paper, I investigate the techniques that are used in the surveillance of labour migrants. Such techniques focus more on performance in the labour market than those used for the surveillance of refugees. These techniques might be used to discipline migrants into a tractable workforce (Foucault 1977). They are disciplinary techniques when work permits can be renewed and when authorities recurrently examine migrants’ behaviour and performance, for instance checking to determine whether a migrant has performed to expectations in the workplace. If the migrant’s performance meets expectations, their work permit will be renewed, if not the work permit will not be renewed. The disciplinary effects of these examinations are particularly strong when the migrant has a chance to receive permanent status in the future, which is the reward for passing these examinations.

In the modern world, the authority to decide to conduct this surveillance lies in the hands of states. However, they might delegate part of this surveillance to non-state actors such as employers (Lahav and Guiraudon 2000). Employers might be given the authority to check that migrants have valid work permits, and might even be involved in decisions about work permits such as whether they should be renewed. A migrant’s work permit will, in this case, be renewed if the employer is satisfied with their performance and wants to keep that person in the workforce. Such arrangements have strong disciplinary effects (Foucault 1977). Migrants will avoid behaving in ways that might be perceived as disloyal and unacceptable by their employer. They might avoid protesting against poor working conditions because they could then lose their employment and eventually be deported.

The actual execution of internal migration control can be either thorough or lenient. In some cases, internal migration control is performed leniently by authorities and is more or less perceived as a formality. In other cases, it is performed thoroughly, and it penetrates deep into the lives of migrants, carefully following them from the point of entry and through all spheres of society. A hypothesis is that expansionist movements with regard to admissions tend to be accompanied by increasing levels of surveillance of migrants residing in the territory of the state (cf. Ruhs and Martin 2008). There are several reasons for this. One reason is that employers have an interest in the surveillance of migrants (cf. Tichenor 2002). Another is that state officials perceive increasing surveillance as a compensation for the relaxing of border controls. This mechanism, for instance, operates in the Schengen Agreement, where the removal of national borders within Europe has been accompanied by an increase in surveillance. Surveillance is likely to increase when migrants are perceived as outsiders in destination countries, for example when they are low-skilled workers from North Africa. The question of who is to be an object of surveillance is influenced by ethnic hierarchies and class positions of migrants (Bigo 2008, Topal 2011). Moreover, it is influenced by beliefs about certain categories of migrants as security threats.

The politics of immigration control

Comprehensive immigration policy reforms affect the relationship between employers and native workers (Zolberg 1999). Employer organisations tend to demand expansion of admissions because it increases labour supply and lowers the cost of wages. Trade unions, on the other hand, tend to demand more restrictive admission policies in order to protect the wages and working conditions of native workers.

Aristide Zolberg (1999) argues that immigration also has an identity dimension in that it affects the culture and way of life in the destination country. These impacts might be grossly overstated in immigration debates, but the perceptions and feelings about them are real. The perception of immigration as a threat to national culture tends to give rise to restrictive reactions, and the political form of these reactions might be culturally conservative or nationalist. On the other hand, there are also groups, such as left liberals, that tend to support higher immigration levels on cultural grounds. These groups might argue that immigration brings about positive cultural change in society by creating a multicultural society. These two conflict dimensions – one economic and one cultural – give rise to the ‘strange bedfellow’ coalitions that are characteristic of immigration politics, for instance where both left liberals and employers support the expansion of admissions.

We also need to account for changes to migrants’ rights. The position of unions and employers has historically been reversed when it come to rights (Tichenor 2002). Unions tend to demand an expansion of rights because it strengthens migrants in relation to employers. It makes migrants easier to organise and thereby protects wages and working conditions. Employers, on the other hand, tend to support restrictions on migrants’ rights. They might demand weak regulation of wages and require migrants to be tied to a specific employer.

Migration to Western Europe and unfree labour

The form and intensity of surveillance affect migrants’ rights in the destination country. There are two particular sets of restrictions on migrants’ rights that influence their position in the destination country: restrictions on their right to permanent settlement, and restrictions on their ability to change employer and occupation. In this section, I will investigate these two sets of rights. Furthermore, I will place the Swedish immigration policy of 2008 in a broader context of migration to Western Europe.

Migration to Western Europe has gone through several distinct phases since the end of the Second World War. Messina (2007) identifies three waves of migration to Western Europe. The first wave consisted of labour migration from the Mediterranean and former European colonies. This wave started after the Second World War and lasted until the 1970s. Some of the migrants were recruited through guest worker systems, while others were colonial migrants, who were often citizens of the colonial state to which they migrated.

The second wave of migration to Western Europe started at the beginning of the 1970s. It consisted of family members and dependents of earlier labour migrants. During this period, labour migrants became permanent settlers in Western Europe. This had not been the intention when Western European states opened their borders to labour migration after the Second World War.

The third wave of post-Second World War migration to Western Europe started in 1989, and it consists of irregular migration and forced migration. Messina (2007) argues that the third wave has its roots in two historical events: the restrictive immigration policies of many European governments, which produced increased rates of irregular migration, and the collapse of communism in 1989–1991, which resulted in large numbers of asylum seekers.

Messina’s (2007) main argument is that post-Second World War migration to Western Europe is a coherent phenomenon, arguing that each of the later waves was connected to earlier waves. This pattern was created by the first wave of labour migration from 1945 to 1975. Labour migration tied sending countries and destination countries to each other, and it made destination countries dependent on foreign labour.

However, it can be argued that a fourth wave of migration to Western Europe has started, as it seems relevant to describe the current labour migration to Sweden as a fourth wave. In earlier periods of labour migration, the main share of migrants originated in Europe. Since 2008, this pattern has changed, with the largest numbers of migrants instead now coming from Asia. The increased migration from Asia has no relationship with the first wave of post-Second World War migration to Western Europe, which Messina views as essential for later waves.

Hammar (2009) conducts a comparative study of the first and second periods of migration, focusing on the development of migrants’ rights. Hammar argues that the most important question to ask when evaluating a system of immigration regulation is whether and under what circumstances immigrants are granted permanent status. If this criterion is applied to Western European states, the result is two opposite types of immigration regulations: the guest worker systems of West Germany and Switzerland, and policies for permanent settlement, which are represented by Sweden and Great Britain. The other Western European states fall in between these opposite poles. It is important to note that Switzerland and West Germany also granted labour migrants permanent status during the 1970s and 80s. In fact, the expansion of migrants’ rights was a general trend in Western Europe during the second period of migration.

Labour migrants are often granted only temporary residence permits, and they do not have the right to change employer as they wish. There are neo-Marxist studies of these restrictions on migrants’ rights. Miles (1987) and Cohen (1997) argue that such restrictions turn migrants into a form of unfree or semi-free labourers, and in fact they regard these migrants as descendants of unfree labourers from earlier periods of capitalism, for example, indentured labourers and slave labourers. Cohen and Miles examine labour migration to Western Europe in the period from 1945 to 1975, during which states imposed severe restrictions on migrants’ rights, including their right to permanent settlement and their rights in the labour market.

In Miles’s (1987) study, the concept of unfree labour refers to physical and/or political-legal restrictions in individuals’ ability to sell their labour power as they wish. The ideal-typical example of unfree labour is slave labour. However, Miles argues that certain forms of wage labour can also be regarded as unfree labour. He terms these instances ‘unfree wage labour’ in order to distinguish them from slavery and other severe forms of unfree labour. An example of unfree wage labour is when legal provisions restrict the mobility of certain categories of worker to a specific sector of the labour market. Another example of unfree wage labour is contract migrant systems, wherein migrants have signed a contract that requires them to work for a certain employer during a specified period of time. When the contract has expired, or when there is no longer a demand for the migrants labour, they are forced to leave the destination country.

The concept of unfree wage labour refers mainly to legal and political restrictions on individuals’ ability to sell their labour power as they wish (Miles 1987). States often enforce such restrictions on migrants’ mobility in the labour market. However, Miles’s concept of unfree wage labour raises some problems when it is applied to Swedish immigration policy during the period from 1954 to 1972. There certainly existed such restrictions on migrants’ labour market mobility until 1967, but after that there was an expansion of migrants’ rights (AMS cirkulärmeddelande U:5/1967). Consequently, after 1967 it becomes difficult to consider labour migrants in Sweden to be unfree wage labourers.

Swedish immigration policy has been described as generous when it comes to migrants’ rights. Hammar (2009), for instance, calls it a policy for permanent settlement. This certainly holds true for some periods of time, for instance the 1970s. However, in other periods the state has reinforced restrictions on migrants’ rights. The enforcement of restrictions on rights tends to coincide with the expansion of admissions.

What is the relationship between the different concepts that have been presented? How is the Foucauldian conceptualisation of discipline and surveillance related to the concepts of unfree wage labour and contract migrant? Surveillance and discipline are very general processes that operate in a wide variety of contexts, including in prisons, schools, workplaces and migration control. Surveillance and discipline are core elements of contract migrant systems, particularly when these migrants are perceived as outsiders in the destination country.

Sovereignty and governmentality

International migration is a question of both sovereignty and governmentality (Foucault 2007). Sovereign power and governmentality are in fact so tightly intertwined in the field of migration that they can be difficult to separate (Dean 2002). It is the sovereign power that decides which persons shall be allowed to enter the state, that is, exercising border control. It is the sovereign power that decides who shall be allowed to stay and who shall be deported. The starting point of migration control is this power to exclude people from the right to reside and other rights of citizenship.

Migration is an object also of governmentality, that is, ‘the conduct of conduct’ (Foucault 2007, pp. 192–193). Governmentality occurs mainly after migrants have entered the state, and there exist different forms of governmentality. Governmentalities can be distinguished by their rationalities and the techniques they use. Discipline is a crucial technique in the governance of migrants. However, it is not only migrants that are objects of governmentality: also employers and native workers are objects of governmentality. The introduction of restrictions on migration, for example, is not only about decreasing numbers: it is also a way of signalling to employers that they should recruit native workers instead of migrants, or rationalise production.

This paper investigates two different governmentalities: a neoliberal governmentality and a governmentality of social security (Foucault 2008, Triantafillou 2012). The governmentality of social security evaluates and governs migration in relation to the social security of the population. It favours state involvement in migration processes, and it can result in the introduction of restrictions on migration. However, a restrictive position does not necessarily follow from a social security governmentality: Immigration can be regarded as compatible with social security, which is illustrated by labour immigration from 1945 to 1975. Neoliberalism and social security are not successive stages in European history. In many countries in today’s Europe, the social security rationality is influencing migration policy, in combination with conservative and populist rationalities. Furthermore, this rationality might be reinvigorated also in Sweden, at the expense of the neoliberal rationality.

Neither neoliberalism nor social security per se excludes a guestworker policy or a policy of permanent settlement; both strategies can be used by the governmentalities. The question is rather with what aims a guestworker policy or a policy of permanent settlement is introduced? How is discipline and border control exercised, and how are they combined in the governance of migration?

Data and method

The method used in this paper is a qualitatively-oriented comparative method. The comparative method is often associated with comparisons of nation-states, but it is possible to compare other units such as regions, cities or historical periods (cf. Boltanski and Chiapello 2005). The qualitative comparative method assumes that contexts have a decisive influence on outcomes like immigration policy reform. Therefore, the main reason for comparing the periods from 1954 to 1972 and from 2008 to 2013 is that that they represent two different historical contexts.

The comparative method used in this paper is similar to the one used by Boltanski and Chiapello (2005), who examine the ideological changes that accompanied the reorganisation of capitalism since the 1980s. They compare 1960s management discourse with that of the 90s. Their analysis shows that in the 90s there had evolved a new spirit of capitalism, which was fundamentally different from the one that prevailed in the 60s. The ideological changes that Boltanski and Chiapello examine are a parallel to the political changes examined in this paper, although the influence of capitalism on politics is more complex and contradictory.

The data consist of documents that have been produced by the state and social actors with an interest in migration. A crucial document is the right-wing government’s proposition on revisions of the Aliens Law in 2008 (Prop. 2007/08: 147). The main social actors interested in migration, such as trade union confederations and national employer organisations, were invited to comment on the proposition. I have analysed documents produced by the migration authority and the labour market authority. The labour market authority played a crucial role in the implementation process during the 1954–1972 period, and it initially made many important policy decisions.

Labour immigration to Sweden in an historical perspective

When studying the migration history of Sweden it is important to make a distinction between immigration from Nordic countries and immigration from non-Nordic countries. Nordic countries have constituted an area of free mobility since 1954, and therefore intra-Nordic migration is not included in the following discussion on immigration policy.

Swedish labour immigration policy from 1954 to 1972 is often described in too general terms. It is described as a highly expansive immigration policy aimed at permanent settlement (Hammar 2009). There were, however, important changes in immigration policy during this period. The most expansive policy when it comes to admissions occurred from 1954 to 1967. During this period, non-Nordic nationalities could enter Sweden as tourists, and once they had entered Sweden they could search for a job and apply for a work permit. In 1967–1968, immigration policy was reformed (AMS cirkulärmeddelande U:5/1967). Until 2008, these reforms were the most important change in Swedish immigration policy since the end of the Second World War. The reforms took two directions: expanding the rights of migrants and at the same time introducing restrictions on the inflow of migrants.

Non-Nordic labour immigration continued after 1967. The aim of reform was not to stop labour migration – these reforms occurred a few years later – but to make it more acceptable to unions and the public. Another motive among state officials was to take control of migration flows, which were perceived as too large and out of control. Most of the non-Nordic migrants originated in Southern European countries (Schierup et al. 2006).

The labour immigration period that started after the Second World War ended in 1972. This occurred in a specific way in Sweden. The corporatist institutions that were in place gave unions a strong influence over work permit applications, inviting them, together with employer organisations, to comment on the applications. In 1972, unions regularly started to recommend that applications be rejected. Since authorities usually followed the opinion of unions this put a halt to the non-Nordic labour immigration.

The socio-political context from 1954 to 1972 was permeated by a social security governmentality (Foucault 2008). There was a strong belief in the interventionist state, and that the state could increase social security. The universalist welfare state was expanding, and social inequalities decreasing. Because of high demand for labour in the manufacturing industry, immigration was perceived as necessary by all the main social actors. However, it needed to be managed in a way that would not threaten the ‘Swedish welfare model’ (Bucken-Knapp 2009). When immigration was perceived as a threat to the ‘Swedish welfare model’, it promoted restrictive reactions.

In 1995, Sweden became a member of the EU and consequently became a part of the European area of free movement. However, when it comes to labour immigration from states outside the EU, the policy changes that were made in 1967–1968 largely remained the same until 2008 (SOU 2006: 87). There has been some changes in the implementation of the 1967–1968 regulations, and it has largely been a restrictive labour immigration policy: immigration was carefully regulated by the state, and to a large extent it involved unions in the decision-making process.

Figure 1 below shows how labour migration to Sweden has changed since 2000. One important change concerns sending countries. The 2008 immigration policy promoted the globalisation of migration to Sweden, in view of the fact that the number of sending countries has increased (Castles and Miller 2009). Greater numbers of migrants now originate in countries at larger geographical distances from Sweden. Since the early 2000s, an increasing number of migrants have come from Thailand, India and China (Migrationsverket). The figure below shows that Asia has become the most important sending region. Due to the enlargement of the EU in 2004, there was a sharp decrease in the number of work permits granted to citizens of European states that year.

Figure 1. Number of residence permits granted due to labour markets reasons according to region 2000–2012.

Source: Migrationsverket.

There has also been a change in the occupational distribution of migrants. In 2012, the largest occupational groups were labourers in the agricultural, fishery and related sectors. The second-largest group was computing professionals, and the third largest housekeeping and restaurant service workers (Migrationsverket). This is different from the occupational distribution of migrants from 1954 to 1972. During this period, migrants were employed mainly in the growing manufacturing industry, and in hotel and catering jobs.

A ‘strange bedfellow’ coalition

The reform of immigration policy in 2008 was overseen by a right-wing government, which made it friendly to employer interests. However, the legal proposition received support from one party of the opposition, namely the Green Party. This illustrates a central argument in Zolberg’s theory, namely that immigration policy tends to give rise to strange bedfellow coalitions (Zolberg 1999).

The political actors participated in this coalition for different reasons. An important motive for the right-wing government liberalising admission policy was to make it easier for employers to recruit foreign workers. The economic motive behind the reform was not a problem for the Green Party, which had no objections to making it easier for employers to recruit migrants. However, a central strand in the Green Party’s support for immigration reform has included a set of cultural and ideological arguments, for instance ideas about the benefits of having a more open world and about how society is enriched by the cultural changes that migration brings about.

However, over time there emerged a consensus that Sweden needed to increase labour immigration. This view was also accepted by the Social Democratic Party and the Left Party. What created conflicts were mainly how this expansive policy should be designed (Bucken-Knapp 2009). What influence should trade unions have on labour immigration? What role should state authorities play in the recruitment of foreign workers? There have been strong corporatist arrangements in Swedish labour immigration policy, and the reforms of 2008 have rolled back these corporatist arrangements, weakening the influence of unions on labour immigration. The starting point for how the migration authority deals with work permit applications is employers’ own assessments of the need to recruit migrants (Prop. 2007/08: 147).

Migrants’ rights in the labour market

One important difference between the immigration policy of 1954–1972 and that of 2008 manifests itself concerning restrictions on migrants’ mobility in the labour market. This reflects a change in how surveillance and discipline of migrants is exercised. The immigration policy of 2008 ties migrants to a specific employer. During the period from 1954 to 1972, migrants were not tied to a specific employer, rather their work permits were tied to a specific type of work or occupation, but within that occupation they had the right to change employer as they wished (AMS cirkulärmeddelande U:5/1967).

The restriction on migrants’ labour market mobility to a certain occupation was introduced in 1927, when the first Swedish Aliens Law was passed (Prop. 1927: 198). In 1927, it was decided that work permits should be granted to foreigners for specific job categories, for example, clerical workers, electricians, maids and so on. Until 2008, this was how the labour market mobility of migrants was restricted by the Aliens Law. There have been some changes in how this restriction has been formulated in the law, and there have also been changes in how the implementing authorities interpret this section of the law. The Aliens Law has been interpreted expansively in some periods and restrictively in others. In the 1970s, for instance, there was a general expansion of migrants’ rights. During this period, migrants’ labour market mobility was restricted to a specific occupation only during their first year of employment in Sweden (Kungl. Arbetsmarknadsstyrelsens cirkulär U:1). After one year of employment, they received work permits without any restrictions at all. On the other hand, a more restrictive administrative practice emerged in the years before the 2008 immigration reform, when the implementing authorities distinguished between temporary labour migrants and permanent labour migrants (SOU 2006:87, pp. 90–91). Temporary labour migrants were tied to a specific employer.¹ In 2008, this restriction was included in the Aliens Law and was generalised to all labour migrants (Prop. 2007/08: 147, pp. 30–31). When migrants have been employed in Sweden for two years, the restriction to a certain employer expires. After two years of employment, migrants have the right to change employer as they wish within a specific occupation.

Currently, the migration authority is implementing this regulation expansively. Migrants are certainly tied to a specific employer, but they have the right to apply for a change of employer if they are able to find another employer. The largest impediment to migrants’ labour market mobility seems to be the high rate of unemployment in Sweden, which makes it difficult to find another employer. However, the current expansive practice may well become more restrictive in the future. Furthermore, since this restriction was included in the Aliens Law in 2008, it will be more difficult to change than it would be if it were only an administrative practice.

The fact that migrants are tied to a specific employer turns them into a form of contract migrants and unfree wage labourers. This description has been more relevant since 2008 than it was during the 1954–1972 period. This restriction will make migrants more dependent on employers than they would be if they were free to change employers at will. Ultimately, it will tend to discipline them into a more tractable workforce. They might avoid protesting against abuses or bad working conditions. If they do not accept their working conditions and they protest, they risk losing their jobs and eventually being deported. Furthermore, the working conditions of migrants have been weakly regulated since 2008, which has resulted in several cases of abuse (Woolfson et al. 2012). There have, for example, been cases where migrants were promised one wage level when they signed a contract with an employer, but after entering Sweden they were paid significantly less than promised. After several cases of abuse were reported, the migration authority tightened its control on employers in 2012, which has resulted in higher rates of rejection of work permit applications (Migrationsverket).

How did the government justify tying migrants to a specific employer? The government argued that such restrictions are in line with the fact that work permits are granted in order to remedy recruitment difficulties of specific employers (Prop. 2007/08: 147 p. 30). It is a strictly economic argument that favours employer interests.

There are three different trade union confederations in Sweden, one of which is Landsorganisationen (LO). LO has experienced the hardest consequences of the immigration reform. It organises the majority of the traditional working class, and it is within these occupations that there have been the most cases of abusive and even slave-like working conditions (Woolfson et al. 2012). LO’s attitude towards the 2008 immigration policy has been ambivalent. The union has accepted the argument that Sweden needs to increase labour immigration; however, it wants a completely different immigration model (LO 2013). LO wants increased involvement of state authorities and tighter border controls, and it wants to ensure that migrants are channelled into occupations where there are actual labour shortages. On the other hand, LO has demanded an expansion of migrants’ rights, for instance that the contract between the migrant and the employer becomes legally binding, so that on arriving in Sweden migrants are guaranteed to receive the wages they were promised when they signed the contract in their home country.

Swedish trade unions, which were highly influential in the past, are bearers of a social security rationality. Demands for the tightening of border controls are motivated by social security considerations. This does not imply an ignorance of the social security of migrant workers. On the contrary, the aim of trade unions is to gradually include migrant workers in the same social protection schemes as native workers. However, this can only be achieved if migration is controlled and adapted to the needs of the labour market (cf. Tichenor 2002).

The national employer organisation, Svenskt Näringsliv (SN), opposed the practice of tying migrants to a specific employer, arguing that it did not promote mobility in the labour market (Prop. 2007/08: 147). This seems to run contrary to standard views on employers’ attitudes towards migrants’ rights. However, we need to distinguish between the general organisation for Swedish capital, on the one hand, and specific employers and sectors in the economy, on the other hand. SN represents the general interest of its members, or more precisely how it perceives this interest in a specific context. Ideologically, SN has developed in an ultra-liberal direction since the 1980s, and this influences its standpoints on immigration. Although SN opposed the decision to tie migrants to a specific employer, such a restriction is in the interest of the sectors and firms that recruit migrants, as it provides them with a more stable workforce.

Guest worker policy or a policy of permanent settlement?

An important question when analysing immigration policy concerns the regulations related to the duration of migrants’ stay: do they receive residence permits for a limited period of time, or are they granted a permanent residence permit? Under what conditions might they be granted permanent status? Hammar (2009) makes a distinction between two different types of immigration regulation: a policy of permanent settlement and a guest worker system. In a guest worker system migrants are recruited to work for only a limited period of time. Hammar argues that Sweden’s immigration policy was a policy of permanent settlement. However, this description is too general for the whole period from 1954 to 1972. It ignores the important changes in immigration policy during the 1960s. After 1967, Sweden certainly entered a path that can be described as a policy of permanent settlement (AMS cirkulärmeddelande U:5/1967). Migrants who arrived in Sweden after 1967 were in practice granted permanent status. However, before 1967 Swedish immigration policy also contained elements of a guest worker policy: it was a blend of guest worker policy and a policy of permanent settlement (Frank 2005). Some elements support Hammar’s argument such as a relatively generous policy of family immigration, whereas other regulations are characteristics of a guest worker policy. For example, migrants were granted work permits valid for only a short duration (Frank 2005).

The time limits on work permits (and residence permits) can create a more tractable and vulnerable workforce than would be the case if the migrants were granted permanent status. This is particularly the case when renewals of permits are conditioned on migrants having demonstrated good behaviour. In this case, the authorities investigate the behaviour of migrants when they apply for renewal of permits. If the authorities conclude that migrants have not shown good behaviour, the work permits are not renewed and the migrants must return to their home country.

This type of surveillance was used on migrants from the Mediterranean from 1954 to 1967. When migrants entered Sweden they were granted work permits with a short time limit: their first work permits were valid for three to six months (Frank 2005). After the migrants had resided in Sweden for approximately a year, the validity of their permits increased to one year. These restrictions in work permits enabled authorities to adjust the size of the labour force depending on the current demand for migrant labour. If there was no such demand applications to renew work permits could be rejected. The authorities could, furthermore, determine whether the migrant had shown good behaviour. A migrant’s residence permit could be denied if there was evidence that they had been unwilling to work, had not adapted to the Swedish way of life or was regarded as a security risk (SOU 1967: 18, p. 199). Renewals of residence permits were particularly rejected on the basis that foreigners had generally misbehaved (SOU 1967: 18, p. 199). The residence permits of most of the labour migrants that arrived in Sweden from 1954 to 1967 were renewed, mainly because there was a high demand for labour. However, migrants’ behaviour was recurrently examined in this way by the authorities, which constituted a system of surveillance and discipline of migrants. It was a generalised discipline that was exercised according to standardised principles (Frank 2005). The purpose behind this standardisation was to ensure that migrants were treated in the same way in different locations. If migrants were allowed to be treated differently in some locations, this was potentially divisive and a threat to native workers. In this respect, it was a disciplinary technique that fitted neatly with the social security governmentality.

This elaborate system of surveillance was in place until 1967, when Sweden introduced more stringent border controls (AMS cirkulärmeddelande U:5/1967). Concurrent with the intensification of border controls that year, immigrants obtained greater protection and an improved position in the labour market. Firstly, migrant workers from non-Nordic countries were allowed to obtain work permits with a longer period of validity than previously, which meant that they were subjected to much less discipline and control from the authorities. Secondly, migrants were guaranteed that their work permits would henceforth be prolonged, which in practice meant that they were granted permanent status. This guarantee, together with the longer period of validity of work permits, represented an important move away from those elements of a guest worker policy that had been part of the authorities’ handling of migrant workers from the Mediterranean. Thirdly, after 1967 it became easier for migrants to change occupations as they wished.

Accordingly, in 1967, there were more stringent regulations of new labour migration and an expansion of migrants’ rights. The immigration policy of 1967 implied a radical change of strategy in the management of migration movements from an emphasis on the techniques of disciplinary power to an increased use of border control. These new regulations meant that careful control of non-Nordic citizens was carried out before they were admitted to the Swedish labour market, which also meant that similar exhaustive controls following the granting of a work permit were no longer necessary. Recurrent examinations were deemed unnecessary, and the problems of trying to control a large number of immigrants in the labour market could thus be avoided. Attention and resources could instead be transferred to border controls.

The immigration policy of 2008 is different from both the policy of 1954–1967 and 1967–1972. The immigration policy of 2008 has increased the involvement of employers in the surveillance of migrants. Furthermore, the possibility of staying in Sweden is more dependent on migrants’ ability to sell their labour power. Since 2008, work permits are valid for up to two years (Prop. 2007/08: 147), which is longer than was the case from 1954 to 1967. However, permits will not be granted with longer duration than that of the planned employment in Sweden. The same principle is applied when work permits are renewed. This implies that the duration of permits can be significantly less than two years. When migrants have been employed for two years they must apply for a new work permit, and they might then be granted a work permit for two more years. After four years of employment in Sweden migrants can be granted permanent residence permits (Prop. 2007/08: 147).

Although migrants can be granted work permits valid for up to two years, the 2008 policy rests on the assumption that it is the demand for labour that determines how long migrants will be allowed to reside in Sweden. If there is a demand for six months, they will be allowed to reside in Sweden for six months. If there is a demand for one year, they will be allowed to reside in Sweden for one year. In this regard, there is a strong element of a guest worker policy in the 2008 immigration policy. I stated earlier that the relationship between state authorities and migrants from 1954 to 1967 was characterised by surveillance and discipline. In the immigration policy of 2008, this exercise of disciplinary power has largely been transferred to employers; it is exercised less by the state and more by employers. The fact that work permits are tied to a specific employer, and that the duration of migrants’ stay is strongly dependent on employers’ need for workers, gives evidence of such a change. If migrants want to continue to stay in Sweden, they have a strong interest in adapting to the needs of employers. If they do not, they risk losing their jobs and eventually being deported. It is a more flexible form of power that is adaptable to the specific circumstances at different workplaces, for instance when it comes to the ethnicity and class position of migrants. Consequently, it allows for differential treatment of migrants to a larger extent than from 1954 to 1972.

A large share of the migrants that arrived in Sweden from 1954 to 1972 became permanent settlers, partially because of the high demand for labour. It is too early to draw any conclusions about the immigration policy of 2008 and whether it will lead to large-scale permanent settlement of migrants. Many migrants will never reach the limit of four years of employment. Furthermore, since 2012, the migration authority has begun to reject a larger share of applications for renewals of work permits (Table 1). This more restrictive stance is a consequence of a tighter control over working conditions in low-wage sectors.

Table 1. Applications for renewals of work permits 2009–2013.

Year	Number of applications	Approved	Rejected	Others*
2009	2483	2344	28	111
2010	2435	2323	45	67
2011	4414	4146	108	160
2012	5744	4958	484	302
2013	7931	6430	1077	424

* Applications that have been discarded/rejected or left without measure.

Source: Migrationsverket.

Conclusion

The immigration policy of 2008 is not a neoliberal immigration policy in the sense that it introduces free mobility to Sweden; it is a bricolage, that is, a recombination of different institutional elements in a new way. Some elements are borrowed from older institutional models in European migration history, while others are newer in the sense that they are an integral part of the neoliberal political context in which the reform was made.

Neoliberal ideas have become a political doxa to borrow a concept from Bourdieu (1977). They have become taken for granted and partially embodied by actors in the political field, particularly by actors on the political right but also on the left, for instance, the Green Party supported the reform of 2008. Neoliberalism would not have achieved such a dominant position if it did not articulate some fundamental concepts that appeal to people’s instincts, desires and widely held values (Harvey 2007). The political ideals of individual freedom and liberty have, as Harvey argues, played a legitimising role in the neoliberal project. Freedom can, however, mean many different things. In the case of the immigration reform, it has meant freedom for employers to recruit migrants without restrictions imposed on them by the state and unions. It has meant freedom for foreign nationals to move to Sweden under the precondition that the migrant can find an employer that will hire him/her. The reform combines such neoliberal elements with old forms of discipline and control of migrant workers. It has borrowed elements from the classical contract migrant model such as tying migrants to a specific employer. Historically, this practice has been more common in Central Europe than it has been in Sweden. Furthermore, the radical changes of migration control do not constitute a complete break with the reforms that were introduced in the 1960s. Trade unions still play a role in the immigration policy of 2008, although it has been significantly weakened. Trade unions are invited to comment on the wages and working conditions that are offered to migrants when work permit applications are processed. Although this procedure is just a fragment of how unions were involved in migration control in the past, it provides a justification of the reform in the face of political opposition.

In one dimension, the neoliberal reform is a failure. The working conditions of migrants are weakly regulated in the 2008 immigration policy. This has led to several cases of abuse and even slave-like working conditions. Therefore, in 2012, the migration authority tightened its control of migrant-recruiting firms in low-wage sectors, which resulted in an increased number of rejections of applications for renewals of work permits. Widespread abuse and slave-like working conditions were hardly the intention of the reform. The introduction of the reform was characterised rather by a belief that there was no need for extensive control of employers. The argument of the government was that employers themselves had the best knowledge of labour shortages (Prop. 2007/08: 147 p. 26). There was no need for state authorities or trade unions to control this, or suspect that the weak regulations of working conditions could be misused. This could be interpreted as a naive belief in employers’ willingness to do well. A more elaborate interpretation is that this part of the reform was inspired by dominant neoliberal ideas about capitalist firms, namely that firms should have as much influence as possible on their own affairs, and conversely that the state should intervene as little as possible. The neoliberal argument is that this will work to the benefit of the common good. This argument did not stand the test against Swedish political realities, and the consequences proved to be unacceptable in a Swedish context where relatively high social standards still prevail.

With regard to the surveillance of migrants, the 2008 reform represents a radical change of how migration control was carried out in the past. Table 2 summarises the findings. In the past, mainly state authorities performed surveillance of migrants. During the period from 1954 to 1967, the state granted migrants work permits with only a short duration during their first years in Sweden, as there were never any guarantees that migrants would be allowed to continue to reside in Sweden. The recurrent examinations of migrants made it possible for authorities to determine whether they were able to keep a job, that they had not misbehaved, that they were prepared to adapt to the way of life in Sweden, and so on.

Table 2. Comparison of immigration policies of 1954–1972 and 2008–2013.

	1954–1967	1967–1972	2008–2013
Duration of work permits	Short time limit on work permits	Permanent work permits when entering Sweden	Time limit on work permits dependent on planned employment in Sweden
Restrictions on labour market mobility	Work permit within a specific occupation	Work permit within a specific occupation	Work permit tied to a specific employer and occupation
		After one year of employment in Sweden migrants are free to sell their labour power in all sections of the labour market (1970)
Guest worker policy or policy of permanent settlement?	Blend of guest worker policy and policy of permanent settlement	Policy of permanent settlement	Elements of guest worker policy; permanent status after four years of employment
Surveillance and discipline of migrants	Increased surveillance and discipline	Decreased surveillance and discipline	Increased surveillance and discipline
Agents performing surveillance	Mainly state agents	Mainly state agents	Increased involvement of employers

In 1967, Swedish immigration policy was reformed. Border controls were tightened, and surveillance and discipline were concurrently decreased. From the perspective of state officials, border control and surveillance are somewhat interchangeable strategies. The tightening of border controls created a space for decreasing surveillance. The thorough surveillance that was performed before 1967 was perceived to be unnecessary, and attention and resources could be transferred to border controls. In 1967, Swedish immigration policy became a policy of permanent settlement. The rights of migrants were expanded and they were almost immediately granted permanent status.

The 2008 reform involved employers in surveillance in a way that is unprecedented in the post-war era. Since 2008, there is a strong relationship between migrants’ right to residence and their employment. Migrants are, in principal, allowed to reside in Sweden as long as they have a job. This is a deeply disciplinary system that provides employers with a tractable and vulnerable workforce. Many migrants strive to satisfy employers’ demands and to adapt to the working conditions provided in order to avoid the risk of losing their jobs and eventually being deported. Furthermore, migrants are tied to a specific employer, which increases their dependence on the employer. During the 1954–1972 period, migrants had the right to change employers as they wished within a specific occupation. This restriction was relaxed during the 1970s. During this period, there were strong administrative regulations against tying migrants to a specific employer.

How do we account for differences between time periods? I have argued that a change of context has strongly influenced how the surveillance of migrants is carried out. The immigration policy reform in 2008 took place in a market-liberal historical context, with weakening corporatist institutions and the increasing privatisation of social life. The change of context influences actors’ ambitions, strategies and beliefs. A general ambition of right-wing parties has been to privatise functions that were previously performed by the state, which has been done in several areas besides migration policy. Another ambition has been to weaken the influence of unions, which held a strong position in Sweden. This change of context explains the increased involvement of employers in immigration control.

However, there are also similarities between the periods of 1954–1967 and 2008–2013. In both periods, borders were opened to labour migration and surveillance was increased. Surveillance tends to increase when migrants are perceived as outsiders. One reason for this is that state officials tend to perceive surveillance as a compensation for the relaxation of border controls. Another reason is that employers have an interest in the surveillance and discipline of migrants.

How should we understand the peculiar mixture of free and unfree elements in the 2008 immigration policy? On the one hand, it aims at freeing migration from political restrictions, and, on the other hand, it introduces unfree labour forms. The neoliberal governmentality is abundant with such apparent contradictions between liberal and illiberal practices. As Foucault and others have shown, neoliberal freedom cannot be separated from power and dominance (Dean 2002, Foucault 2008, Triantafillou 2012). Neoliberal freedom is a peculiar form of freedom that presupposes disciplined and flexible subjects, which have the capability to exercise this ‘freedom’. Those that lack the capability to exercise ‘freedom’ – whether it is the unemployed, the poor or migrants – have regularly been subject to coercive measures. It is at this point that governmentality and sovereign power criss-cross in the field of migration. The neoliberal governmentality uses sovereign power in order to shape disciplined and flexible subjects. Those migrants that do not meet employers’ demand for a flexible workforce are excluded from the right to residence.

Disclosure statement

I do not have any financial interests or other benefits arising from the direct application of this research.

Notes on contributor

Denis Frank (PhD) is a senior lecturer in sociology at the University of Gothenburg. He is currently leading a research project, funded by Riksbankens Jubileumsfond, which examines the globalisation of migration to Sweden.

Acknowledgements

This work was supported by Riksbankens Jubileumsfond under Grant P12-1028:1. I thank the participants at the Working life seminar at the University of Gothenburg for valuable comments on an earlier version of this paper. I also thank the anonymous reviewers for valuable comments.

Notes

1. During this earlier period, the Social Democrats led the government. Therefore, this change was not created by the right-wing parties alone.