Outcomes of Swedish migration and economics of the welfare system

Abstract This article examines the economic consequences of migration for the Swedish welfare system. The question of whether the high costs of receiving refugees undermine the sustainability of the universal welfare state has received considerable attention in the political debate. While most studies focus on the fiscal burden of refugee reception and the short-term impacts of migration on welfare arrangements, this article advocates a comprehensive, long-term assessment of the effects of migration flows on public finances. Starting from the priorities of the architects of the welfare system, who designed and expanded social policy, indicators are derived to assess strategic goals and considerations. Reproduction of the workforce by managing its age profile, increasing employment levels and the employability of the surplus workforce, and creating conditions that increase long-term growth in the economy are considered crucial for the sustainability of the system. From a purely utilitarian perspective, migration is considered beneficial to the economy, provided that migrants gain quick entry to, and a high employment rate in, the labour market.


Introduction
For many decades Sweden has been considered one of the most migrant-friendly European Union (EU) member states and one of the most diverse European societies (Schierup & Ålund, 2011). In 2020, every fifth inhabitant of Sweden was born outside the country, and over a quarter had a migrant background. Rapid demographic change has led to widespread social change, and this is perceived as posing a serious challenge to the long-standing paradigm of certainty and order that has characterized the Swedish welfare society (Hansen et al., 2017;Ruist, 2019). A dual impact has been identified. The immediate challenges posed by migration to the socio-political order were predominantly faced by welfare institutions, which were subjected to severe pressure (Beine et al., 2018;Burgoon et al., 2012;Schmidt-Catran & Spies, 2016). Advanced welfare states are often perceived as a national project tasked with forging bonds between citizens (Mau & Burkhardt, 2009). Their rights and benefits are therefore restricted to 'insiders', with the result that immigrants, as 'outsiders', represent a challenge to the durability of welfare institutions built on inclusion, integration and homogeneity. A nationalist welfare state cannot coexist with immigration and the free movement of labour (Freeman, 1986). The comprehensive Swedish social security system was originally established to provide widespread support to those temporarily excluded from regular social processes (Valenta & Bunar, 2010). The large-scale exclusion caused by migration could be seen as jeopardizing the sustainability of the system.
An additional impact identified is linked to the benefits of migration flows for the host society as migrant groups become a significant part of a social base that is strongly motivated to contribute to the preservation and sustainability of welfare institutions (Scarpa & Schierup, 2018;Steele, 2016). As employees and entrepreneurs, migrants of working age contribute to the financing of the welfare system, and a significant number will be employed in welfare organizations (Hjerm, 2004). The contributory perspective on immigration perceives inflows of newcomers to the Swedish labour market as essential, and crucial to the sustainability of the welfare system, which is one of the most ambitious and expansive social safety nets providing comprehensive social rights to all those living in the country (Esping-Andersen, 1990). It is no coincidence that the issue of migration and its impact on welfare systems attracts great attention in the political debate (Sannerstedt, 2013).
Regardless of these two more strategic perspectives on the interplay between migration and the welfare state, the immediate issues that dominate contemporary narratives on migration are the high costs of refugee reception and the unsatisfactory outcomes of integration policies in introducing new arrivals to the labour market. These two effects are believed to put severe pressure on the welfare system and the central government's finances. Many researchers who have studied the growth of anti-immigrant sentiments in Sweden (Bevelander & Hellstr€ om, 2019;Bystr€ om & Frohnert, 2013;Widfelt, 2015) have mentioned the economic costs of refugee reception as a decisive factor in the growth of discourses based on nativist and populist claims.
A comprehensive assessment of the economic effects of migration on the welfare system must include a long-term perspective that departs from the life cycle burden of both migrants and native Swedes on welfare arrangements (Hajighasemi, 2021). Furthermore, any assessment should be based on an analysis of the needs and priorities welfare system advocates have always identified as essential. The aim of this article, therefore, is to make a more comprehensive and long-term assessment of the effects of migration flows on public finances. More importantly, by referring back to the strategic priorities and considerations of the advocates of welfare states during the design and expansion of social policies, the welfare state issue can be disengaged from current populist concerns. Based on the long-term priorities outlined by the architects of the welfare state, the creation of sustainable welfare arrangements has enabled the reproduction of the workforce, increased employment levels and the employability of the surplus workforce and created the conditions for higher rates of long-term economic growth (Esping-Andersen, 1990;Hajighasemi, 2019;Hort, 2014). These variables should be considered the most crucial for the sustainability of the system. This article argues that these concerns should form the basis for assessing the effects of migration.
The next section reviews the existing literature on immigration and the welfare state. This includes a discussion on the development of the welfare state in Sweden in order to identify the priorities and considerations that played a dominant role in the design of the system and the strategies envisaged. The aim is to derive indicators that can help assess the significance of migration for the sustainability of welfare systems, their functionality and their future prospects. The results are analysed and presented, and some conclusions drawn.

The impact of migration on welfare states
The literature on the effects of immigration on welfare states reveals the complexity of the interplay between welfare regimes and patterns of immigration. Surprisingly, most studies on the subject, as Fenwick (2019) stresses, suggest that in the longer term immigration leads to increased spending on welfare but reduced spending on social welfare. One explanation for this is that immigration has a negative impact on the propensity of a large part of the middle class to demand a more generous redistribution of resources (Magni-Berton, 2014). A large proportion of immigrants, mainly those who enter the country as refugees, are low-skilled workers, and therefore employed in low-paid sectors; accordingly, at least in the short term, they depend on welfare benefits. This makes the native middle class, which is to a greater extent in the higher income brackets around or above the median, reluctant to support redistributive welfare arrangements. In the light of immigration flows, the established majority in society gradually become reluctant to defend redistributive aims, which paves the way for an erosion of the generosity of the safety net for outsiders. Reducing the generosity and attractiveness of the welfare system has also been adopted to reduce the 'welfare magnet' characteristics of the system (Razin & Wahba, 2015). The erosion of the level of welfare benefits can be seen as a 'race to the bottom' that has been taking place at least in part due to the intensification of migrant flows (Fenwick, 2019). Alesina and Glaeser (2004) work on the impact of international migration on the weakening of the welfare state has received widespread attention. Their study of US society from a multicultural perspective found that the reduction in collective solidarity was due mainly to the ethno-racial fractionalization of the population. This segmentation of society into groups with different languages, religions and cultures, in turn, explains the lack of resistance to the process of downsizing that the US welfare state has experienced. According to these scholars, a considerable proportion of the differences in welfare state spending between the United States and the European countries can be explained by the segmentation experienced by US society. Increased diversity would lead European welfare states to meet same fate as the US welfare state. In the same vein, Burgoon et al. (2012) link diminishing support for the preservation of generous levels of welfare benefits and increased indifference among some established groups regarding maintaining the redistributive characteristics of the system to decreasing solidarity caused by increased racial heterogeneity. Other studies, such as Banting and Kymlicka (2006), which have also raised concerns about reduced social solidarity and support for redistributive social programmes, have found no general or inherent tendency to undermine redistribution. Some studies have suggested that the differences in the redistributive ambitions of the US and European welfare states are the consequence not of ethnic diversity, but of the political context in which the welfare states have developed historically on the two continents (Mau & Burkhardt, 2009;Taylor-Gooby, 2005). Support for redistribution does not depend on ethno-cultural diversity or the level of migration but, as Burgoon (2014) suggests, it is related to the success of other factors such as integration policies, the inclusion of migrants in the labour market, economic outcomes for immigrants and, primarily, reducing migrants' reliance on welfare benefits.
In line with the issues outlined above, much of the welfare state literature published in recent decades has focused on the demand side. In this regard, immigration has been seen as a factor that has contributed to a deterioration in welfare institutions (Fenwick, 2019). Large-scale immigration has altered the socio-economic conditions of welfare societies, and the proportion of migrants in the total number of recipients of welfare benefits has increased steadily. This has led more and more people to doubt the purpose and foundations of welfare institutions, previously established according to the social democratic folkhem (people home) ideology, which aimed to provide a secure, equal and solidaristic society for an ethnically homogenous Swedish nation (Schall, 2016). Instead of focusing principally on ensuring the security of citizens and offering them social protection, welfare institutions have begun to allocate welfare resources to an ever-growing number of migrants. The solidarity once created by these institutions has been eroded and the established social groups' demands for, and expectations of, welfare benefits have also been challenged. As Widmalm (2018) suggests, the prevailing economism in advanced welfare societies has made citizens less solidaristic about sharing their growing wealth with those who have the greatest need for some of it.
This focus on the demand-oriented approach explains why the effects of immigration on the downward trends that welfare systems have been experiencing have received the bulk of scholars' attention. By the same token, it shows why the reluctance to preserve the redistributive ambitions of welfare arrangements has been explained solely by migration flows. Some scholars have challenged this approach. Soroka et al. (2006) studied the impacts of immigration on social spending in 18 member states in the Organisation for Economic Co-operation and Development. It found that while social spending had increased in all welfare states, the increases were higher in countries with limited migration than in countries with higher rates of immigration. These findings support the notion that ethnic diversity strengthens the propensity for welfare institutions to reduce their degree of generosity.
The narrative that paints immigration as a decisive factor in undermining the sustainability of welfare systems has been challenged even in welfare societies that have had both a redistributive welfare system and a liberal migration policy. For instance, Scarpa and Schierup (2018), who have studied the impacts of Swedish migration policy on the welfare state, found the 'Swedish government's unbending adherence to austerity politics' to be a decisive factor in weakening the Swedish welfare state's socially integrative functions. The austerity policy prevented the implementation of a more ambitious growth agenda and a dynamic, expansionary economic policy.
Economic growth, which has historically been one of the main goals of the Swedish welfare state (Hajighasemi, 2004(Hajighasemi, , 2019, has been highlighted by Rothstein (2017), who does not just reject the idea of immigration as a corrosive force in the economy, but regards it as a positive growth factor. In 2016, when Sweden had been receiving larger numbers of refugees than the other Nordic states for several years, its growth rate was increasing four times faster (Rothstein, 2017). A major reason for this was a process known as 'involuntary Keynesianism', which according to Rothstein occurred due to the fiscal stimulus effect of increased public expenditure, which was in turn caused by the large-scale intake of refugees. In other words, the refugee crisis had an expansionary impact on the Swedish public sector, mainly on human services organizations at the municipal level that had to provide comprehensive welfare services to the refugees. Positive impacts of refugee migration on the public sector have also been observed by Legrain (2016), who stresses the increased consumption and demand that are associated with receiving refugees. Fuelling consumption and demand is particularly important in times of economic downturn. Alongside these short-term considerations, Legrain emphasizes two further strategic aspects of immigration, both of which are linked to strengthening the supply side: filling the gaps in the labour market and reversing demographic trends. The success of these two aims, however, depends mainly on an effective reception system and efficient integration policies.
These issues have dominated the Swedish debate among scholars and the public. While opponents (Sanandaji, 2017) of the liberal migration policy perceive a continuation of the strategy as a threat to the sustainability of the welfare system, others (Scocco & Andersson, 2015) have identified immigration as a solution to some of the challenges facing the welfare state, notably an aging population and a growing need for labourespecially in areas that require expertise. While critics focus on the immediate negative effects of receiving refugees on public finances, others highlight the more long-term effects of migration, especially labour migration, on sustainable economic development (Flood & Ruist, 2015). The latter approach sees migration not as a burden on the economy, but as complementary to economic development by increasing revenues to the state budget from the contributions of working immigrants over their entire working lives and economically productive years in the country. This is thought to contribute to increased growth, which is crucial for the expansion of welfare and continued social development (Scocco & Andersson, 2015). Thus, from this perspective, migration is necessary for promoting investment in ambitious welfare systems.

Criteria for assessing the impacts of migration on welfare systems
The purpose of this article is to examine the effects of migration on the Swedish welfare system based on criteria derived from the priorities of the architects of the welfare state that were identified as crucial for the durability and sustainability of the system. These criteria can be identified from closer study of the central goals and priorities that formed the basis for the development of the Swedish welfare system. The question of whether migration has assisted or prevented the achievement of these goals can be answered through a closer examination of the strategies outlined by the Swedish welfare system since its inception, and which have characterized the development process of various parts of the social policy system. This provides a better understanding of the perspectives of the architects of the welfare system on the challenges that have historically been regarded as the main threats to a functioning economy and welfare system.
2.2.1. 'Productive social policy': labour force supply Studies of the development of the Swedish welfare system show that a productive view of social policy has guided its design since the start of welfare state construction (Hajighasemi, 2019;Morel et al., 2012;Olsson, 1993). According to this paradigm, social policy must be seen as an instrument for promoting the needs of fast-growing industrial societies, which consist predominantly of ensuring the reproduction of the labour force and developing its skills, and measures that help to increase the level of activation in the production process. This paradigm has been likened to social investment strategies, which genuinely differ from the passive version of the welfare state that relies largely on compensatory measures of social protection (Hudson & K€ uhner, 2009;Kuitto, 2011). The latter approach has a healing agenda that seeks to mitigate the negative effects of the market economy on fragile social groups through a redistribution of resources by providing benefits and cash transfers (Giddens, 2000). For advocates of a social investment strategy, redistributive policy objectives are best achieved through efforts to facilitate economic growth and a functioning labour market. By steadily increasing the employment rate among the potential labour force in society, especially in the most highly productive economic sectors, more favourable conditions can be created to increase the income levels of households and strengthen the central government's finances. The productive approach to social policy serves this strategy by enhancing the accessibility of labour market opportunities for the potential workforce and the availability of labour for firms. The two together promote the long-term conditions for developing a balanced market for waged labour. This strategy therefore promotes social sustainability through a balanced development of households and the business sector (Ahlberg, 2009).
From this viewpoint, the welfare state is not only regarded as an apparatus for consumption tasked with satisfying the basic needs of the most disadvantaged social groups; it must also create the conditions for the emergence of a functioning and dynamic economy. A well-functioning and productive welfare state must identify the market economy's short-and long-term challenges and needs, and ensure that the necessary measures are put in place to maintain the availability of productive labour. Ensuring the orderly reproduction of the labour force has undoubtedly been one of the central goals of the Swedish welfare state (Esping-Andersen, 1990;Hort, 2014;Korpi, 1983). The state has also seen it as a central task to promote a balanced workforce with an optimal age distribution so that the needs of the labour market are met through a functional family policy that encourages childbirth. Furthermore, in the event of a serious labour shortage created by disturbances in population growth or changes in economic conditions, the productive [Swedish?] welfare state regarded itself responsible for opening up the economy to labour migration. In this regard, it should be emphasized that a crucial element of early labour market policy from the 1950s, the Rehn-Meidner model, comprised mobility-promoting policies that aimed to secure a supply of labour for the labour market (Erixon, 2007).

Women in the workforce and a family friendly labour market
Sweden's rapid industrialization in the early 20 th century was followed by equally rapid urbanization, which resulted in a decline in childbearing among urbanized groups, which no longer saw the need to have a large number of children to assist the agrarian economy (Sheiban, 2002). This trend was seen as a long-term threat to the reproduction of the labour force, one that would jeopardize the supply of labour for industrial production. To increase the birth rate, the most important task of social policy has been to promote social security for all citizens in the form of secure housing, childcare and elderly care, education and healthcare and other welfare initiatives that together enable the workforce to participate in the development of society. According to Hort (2014), the architect of this strategy was the then Minister of Social Affairs, Gustav M€ oller (1932-51), who identified the decline in fertility rates as a major threat to economic growth and social cohesion, and called on a number of experts, such as Alva and Gunnar Myrdal, to initiate a large-scale investigation of the population crisis. Their book, Crisis in the Population Question (Myrdal & Myrdal, 1934), suggested that the fall in fertility should be tackled through a 'productive social policy' strategy that perceived the problem ultimately as an issue not for families, but for the economy and the reproduction of the labour force (Hudson & K€ uhner, 2009). The productive social policy approach regarded passive support for families as out of the question. Instead, declining fertility, scholars suggested, should be addressed in combination with careful business and economic policy considerations (Hajighasemi, 2019;Morel et al., 2012). If industrial production is made more efficient by maximizing the use of technical capacity, the proper conditions can be created for real incomes to increase. This leads to greater momentum in the economy, which in turn contributes to lower unemployment and creates the conditions for couples of childbearing age to consider having children. In other words, remedying the low birth rate and bringing order to population growth were seen as economic problems that could be tackled with a social policy based on active economic measures.

Increasing employability of the potential workforce
Through this productive view of social policy, these pioneers of the welfare system showed that an orderly reproduction of the labour force was a prerequisite for growth in the economy, which in turn would ensure the long-term welfare of an industrial society. This approach was also decisive when Swedish labour market policy was designed by welfare architects in the 1950s (Hort, 2014). Competitiveness was a key element for export-oriented Swedish industry, and could be achieved through constant innovation and efficiency measures in low-productivity sectors. Unprofitable production could be shut down and labour laid off without much resistance from the industrial unions by providing mobility-enhancing measures and retraining in new skills to make people employable in high-productivity sectors, thereby relieving labour shortages in those areas (Calmfors et al., 2001). Just as with the population issue, the welfare system defined its role as protecting the standard of living of the labour force through generous welfare benefits, thereby preventing resistance to innovation. The strategically important issue of access to labour became so crucial that the Swedish labour market was opened up to women to a greater extent at a relatively early stage and with considerable enthusiasm. As early as 1939, legislation removed the formal barriers to women's access to the paid labour market. At the same time, employers were prohibited from dismissing women who got married or became pregnant. The welfare state's contribution to meeting the labour market's need to increase employment rates among women laid the foundations for a rapid expansion of family policy, with the expansion of childcare, parental insurance and special tax treatment of women (Nyberg, 1994).

The complementary nature of a migrant labour supply
When domestic measures to address labour shortages appeared insufficient, the pragmatic welfare state did not hesitate to reshape migration policy to meet the needs of the expanding economy (Johansson, 2014). The rationale was that an increase in the supply of labour through labour migration could help to expand production and increase economic growth, and therefore the country's total welfare (Johansson, 2014;Penninx & Roosblad, 2000:4;Svanberg, 2011). Making economic growth the determining factor in productive social policy proved crucial for the expanding Swedish welfare system, opening it up to labour immigration (Frank, 2017). In the first three decades after the end of World War II, the huge demand for labour, especially in the burgeoning manufacturing industries, compelled the ruling Social Democratic Party government and supportive trade unions to reform the restrictive Aliens Act of 1927. This reform resulted in a new Aliens Act of 1954, which removed the requirement for special work permits for migrant workers, especially those from the Nordic and Baltic States (Spång, 2009). The decade between implementation of the new Act and the mid-1960s was a period of liberal migration policy.
The period in which labour migration was liberalized was characterized by two parallel processes: exceptional economic growth and the rapid expansion of welfare institutions, both of which led to reduced social inequalities. Trade unions, which would normally resist any liberalization of labour migration, prioritized continued high growth in the economy, which was a prerequisite for the expansion of welfare programmes, and therefore complied with the 'open-door' policy on immigration. Between the end of World War II and the beginning of the 1970s, just over 30,000 people migrated to Sweden to work primarily in manufacturing industries every year (Wadensj€ o, 2010). The welfare state, through the Swedish Labour Market Board (AMS), which dealt with labour market issues, was responsible for recruiting and, if necessary, training labour migrants. This policy was pursued until the mid-1970s, when the need for labour decreased and trade unions began to urge a more restrictive policy. By that time, labour immigration had practically ceased (Frank, 2017: 72).
The policy choices adopted by the architects of the Swedish welfare state in the four areas mentioned abovethe population issue, employability of the labour force, entry of women into the labour market and labour migrationdemonstrate how achieving a functioning economic system is strongly linked to the management of these issues. According to this approach, welfare goals can only be achieved when different sectors of the economy are provided with the conditions required to promote innovation and competitiveness. In other words, the main task of welfare states is to create the correct conditions for enterprises to have access to resources, most notably the labour they need to be highly productive and globally competitive (Hajighasemi, 2004(Hajighasemi, , 2019. Thus, starting from purely welfare state premises and priorities, by far the most important criterion for assessing the effects of migration on the Swedish welfare state is the significance of its labour market policy. Based on this approach, the impact of the extensive migration to Sweden in recent decades must be evaluated on the basis of this criterion. Migration to Sweden has been a societal phenomenon which, in addition to globalization, has had the greatest impact on socio-economic conditions and the prospects for the development and consolidation of the welfare state. The effects of migration on the welfare state can be broken down into: (a) a short-term effect, which refers to the net contribution of migration to public sector finances; and (b) a more long-term effect, which is determined mainly by the effects of migration on the supply of labour. The latter is crucial to demographics, which are threatened by an aging population. These two effects are the key criteria on which this article is based and must therefore be specified in more detail, not least because they can be interpreted differentlyespecially when it comes to net contributions.

Methodology
Based on the underlying processes outlined in the theoretical framework and the two criteria derived from the discussion on the historical development of the Swedish welfare state, this paper has sought to identify and highlight the major areas that have been affected by migration. A survey of these areas was a prerequisite for an in-depth investigation and consequential analysis of the effects of migration. These key areas were captured by reviewing the migration debate and the critical arguments that have addressed the effects of migration. Highlighting these areas has helped to derive concrete indicators from which the results section has been organized. The controversial nature of immigration and asylum issues has meant that discussion of the 'effects' of migration has been heavily influenced by anti-or pro-immigration bias. Thus, in the sections that follow, and in relation to the deeper presentation of the indicators derived from the debates, the areas or indicators are also problematized to reduce any political bias that could be associated with the issue.
The next section presents these key areas on which the migration debate has focused in recent decades. The results cover: (a) the net contribution of migrants to the Swedish welfare system; (b) the effects of immigration on population growth, and thus on the welfare dependency ratio; (c) the impacts of different categories of immigrants on the welfare system; and (d) the long-term contribution, mainly of the supply of labour linked to the demographic challenge.
The results are based partly on previous studies that have examined the impacts of migration on Swedish public finances, but also partly on data from the Swedish Statistics Agency (SCB) and Ekonomifakta, a Swedish web-based tool developed by the Confederation of Swedish Enterprises. The latter source is intended to supply information on the Swedish economy in a user-friendly manner, so its data has been treated with caution. The decisive factor for using these two sources is that they provide data on the four areas on which this paper focuses: (a) employment rates for refugees by level of education and length of stay in Sweden; (b) expenditure areas affected by the high level of refugee migration; (c) the employment gap between native Swedes and four migrant groups by length of stay in Sweden; and (d) the employment rate of refugee migrants in relation to people born in Sweden.
3.1. Some of the most debated 'effects' of migration 3.1.2. The net contribution of migrants to the Swedish welfare system In the Nordic migration debate, the net contribution to welfare is calculated by comparing the total income of migrants and the total cost of migrants to the welfare system (Andersson & Wadensj€ o, 2008: 41f;Hansen et al., 2017). If the contributions of migrants, through taxes paid and various forms of fees, are greater than the payments and other costs to the system through the consumption of welfare services, the net contribution is positive. If not, it is negative (Hansen et al., 2017;Ruist, 2014). This simple calculation has also been applied by Dustmann and Frattini (2014) in their studies of immigrants' performance in the UK. The shortcoming of this kind of comparison, which is interested only in the financial burden of certain groups in the labour force on the welfare system during their years of employment, is that it can only provide an indication of the current economic performance of the labour force. It does not give a complete picture of the cost-effectiveness of labour force change based on the lifelong costs to the welfare system. Dustmann and Frattini (2014), however, also identify the human capital savings made by destination countries. They calculate that the UK government saved £14 billion between 1995 and 2011 from migrant workers being educated abroad.
In the Nordic model (Hort, 2014), which has a strong redistributive character, another factor must therefore be considered: the lifetime cost of an individual to society. This economic calculation assesses the economic effects of an individual on a welfare system as the total net effect of the lifetime cost. In welfare societies, citizens are beneficiaries of the welfare system during their childhood and adolescence, and also in their retirement years. Only once a workforce participant's total financial contribution to the economy through production exceeds the total amount of life-long benefits received from society can there be talk about net contributions (Hennessey & Hagen-Zanker, 2020;Oxford Economics, 2018). For a welfare system such as the Swedish one, this lifetime income perspective provides a more complete and fairer picture of the economic consequences of migration for the welfare state and public finances. Compared to native Swedes, the life cycle patterns of migrants differ in three areas that can affect the economic outcomes of welfare systems.
The first, and by far the most important, factor is the question of the age of the migrant at the time of migration. Because a significant proportion of immigrants move to their destination country when they are of working age and therefore at their most employable, the average level of the cost burden on the welfare system is significantly lower. The second factor is how quickly migrants are able to enter the labour market and how long they depend on various forms of benefits and entitlements. The third factor is the level of human capital that newly arrived migrants possess. In developed economies such as Sweden's, the level of education of migrants and the recognition of their qualifications is a significant factor that can determine how quickly they are able to enter productive employment. The fourth and final factor is the extent of the welfare benefits and services that the system redistributes to newly arrived migrants. This cost is not limited to the introductory costs associated with newly arrived migrants, but includes the costs of training or retraining.
Nonetheless, studying the effects of migration on a welfare system based on these factors and from a net contribution perspective can only provide a general and incomplete picture of how migration affects the welfare state. Migrants are not a homogeneous group and there are large variations within different ethnic groups (Ruist, 2014;Scocco & Andersson, 2015). In the same way, the challenges facing welfare systems are also manifold. In addition to short-term labour needs, most welfare societies face challenges linked to low birth rates and a rapidly aging population. Thus, it is the reproduction aspect that represents the long-term perspective in the migration debate.

The impact of immigration on the dependency ratio
The most crucial issue in examining the impact of immigration on the long-term sustainability of welfare systems is the effects of immigration on population growth and the age structure, and thus on the welfare dependency ratio (Ruist, 2014;Sanchez Gassen & Heleniak, 2016;Scocco & Andersson, 2015). A major characteristic of immigration is its age structure, in that the younger age groups, notably people between 20 and 40, are by far the most represented. With large-scale migration, this age group increases rapidly, and thus the proportion of people of working age. As a consequence, the proportion of people who need welfare benefits and welfare services the most, those aged 0-14 and 65 and over, decreases. With increased migration, the dependency ratio on welfare usually falls. This, in turn, can lead to faster economic growth and result in sharp reductions in public expenditure as a percentage of GDP (Malmberg, 1994).
The shortcoming of reasoning based solely on the demographic effects of migration is that it only shows population figures and ignores other important factors, such as the employment rate and level of education. The latter factors are particularly important in the Nordic welfare regime, which has stronger redistributive policy ambitions. Given that the transfers in these systems are extensive and the level of generosity quite high, they depend strongly on high employment levels and a low degree of dependence on welfare benefits. The more people are supported by various forms of public transfers and the longer the dependence on these benefits, the more negative will be the effects on public finances, and the greater the distributional policy consequences for the welfare state.
This leads the discussion to a central issue in the migration debate: the labour market's integration of migrants, or how quickly migrants are able to enter the labour market and how dependent they remain on contributions from the welfare system. Facilitating integration through participation in the labour market not only reduces the strain on the welfare system, but also accelerates integration by positioning migrants to achieve better health, housing and education outcomes. Furthermore, the sooner migrants become self-sufficient, the longer they are able to pay taxes and contribute to public finances and the social security system (Bakker et al., 2017;Hansen et al., 2017;Ruist, 2014). The pace of the economic integration of new arrivals, and thus their capacity for self-sufficiency, determines the effects of migration on public finances. This ability varies between different migrant groups. Therefore, it is important to have an accurate picture of the conditions for the integration of different groups into the labour market.

The impacts of different immigrant categories on the welfare system
In any discussion about the effects of migration on welfare societies, a distinction should be made between three different categories of migrants, each of which has its own conditions and therefore affects the economy and finances of the host country in different ways.
The first group consists of highly educated immigrants who are desirable and often welcomed by destination countries, as they are usually a highly productive group (Globerman, 2019). Highly educated people are often active in technology, IT and science, and can therefore contribute to innovation and industrial development. There is competition between high-tech economies to attract highly educated experts. This group of migrants has positive effects on the average productivity and income of destination countries and directly or indirectly help to increase central government tax revenues (Globerman, 2019:26;Oxford Economics, 2018).
The second category of migration is labour migration, which often occurs when a labour market is suffering from labour shortages, especially in sectors that are expanding. In these economies, labour migrants are welcomed for their contributions to productivity development and growth, as expanding economies have the potential to immediately employ migrant workers in the production process and, as a result, absorb part of their net contribution to the entire economy. Increased productivity in sectors that have a tailwind of growth provides income for the entire economy and the state is also allocated extra tax revenues (Nyman & Ahlskog, 2018;Oxford Economics, 2018).
The third category is refugee migrants. In most cases, this is an unplanned and sudden migration that is not coordinated with the labour needs of the host country. Refugees are often heterogeneous in terms of their level of education, qualifications and work experience, and are often distributed across the age range. The unplanned aspect of refugee migration means that migration has not taken place on the basis of labour market needs in the recipient country. Refugees can therefore be employed quickly to a much lesser extent than the other two groups, and often tend to depend on public support for longer periods of time (Hennessey & Hagen-Zanker, 2020).
The debate on the consequences of immigration for the labour market and the welfare state usually focuses on the effects of migration in three main areas: wage levels, employment and public finances (Globerman, 2019). With regard to the effects of migration on the wages or employment situation of the existing workforce, the most important question is whether immigration, either planned or sudden, puts the labour market of the host country out of balance (Kerr & Kerr, 2011;Peri, 2016). If the labour market is prepared for migration, it has greater opportunities to adapt to changes in the supply of labour and thereby increase production volumes according to the availability of labour. Newly migrated labour then has a theoretical chance of quickly entering a production cycle that is as productive as before, as companies have time to expand capacity in the production of goods and services, which means that the new workforce does not have to compete with the existing one. This prevents both wage cuts and an increase in unemployment, as the equilibrium of the labour market remains unchanged (Globerman, 2019:11-12).
It is in the event of sudden migration flows, mainly of refugee immigration, that wage cuts or an increase in unemployment may become apparent. According to Ruist (2019: 121), who has studied the labour market effects of sudden migration on several countries, such effects in most cases have been quite limited. Based on these case studies, it appears that if the labour force increases by 1 per cent, either unemployment increases by 0.2-0.3 percentage points or wages fall by 0.2-0.3 per cent. In the case of large-scale refugee migrationespecially that which has taken place in European states, including Sweden, over the past two or three decadesthe conditions for rapid entry into the labour market have been absent. It took up to 10 years for the average employment level in this group to reach 50 per cent. Therefore, the net contribution to public finances by refugees in the initial years of their stay has been clearly negative.

The long-term fiscal impact of immigration
The reasoning so far suggests that of the three major categories of migrants, it is mainly refugee migration that accounts for the negative net contribution to public finances. To examine this further, it is useful to outline where the revenue for the welfare state comes from and how it is used. In the Nordic model, welfare systems are funded mainly by general taxation on incomes, consumption taxes and employers' social security contributions. In Sweden, these three sources of income together make up about half of the total value of everything produced in the country (Sj€ oberg, 2005). These funds are used to finance social insurance systems, or transfers that include pensions and various types of benefits, as well as public consumption that consists mainly of the costs of healthcare, providing care for the elderly and children, and providing schooling and other activities that are part of the public sector (Esping-Andersen & Myles, 2009). The nature of the welfare system is such that everyone who works, owns assets or undertakes economic activities contributes an annual net sum to the treasury so that everyone can then consume social insurance and welfare services. For some of the population, the net contribution is positive, meaning they give more to the system than the benefits and services they receive. Others have a negative net contribution, as they are in greater need of welfare. The redistributive and egalitarian nature of the Nordic welfare systems aims to achieve a situation in which some groups' net contributions are positive and others negative (Esping-Andersen, 2016). It is obvious that the average resident of working ageindividuals aged between 20 and 65will make a positive net contribution to public finances, while children and pensioners are a net cost to society (D'Albis et al., 2018;Storesletten, 2000).
Given that people in the 20-35 age bracket constitute by far the largest group in almost all migration flows, be they economic migrants or refugees, there is a vast potential among this group to contribute positively to the recipient country's public finances. As Storesletten (2000) stated, most migrants at the time of migration are over the age of 20, which means their costly years of healthcare, childcare and education have already been financed by others. Recipient countries often receive labour that is at its most productive age, with the potential to be active for a lifetime with little need to finance their reproduction costs. This means most migrants can enter the labour market directly, earn an income, pay taxes and thus contribute to the state treasury. Thus, the vast majority of migrants have the potential to make a positive net contribution to the public finances of the recipient country if they can quickly find employment.
While studies on the fiscal impact of migration in central and southern European states, for instance in Spain (Izquierdo et al., 2010), France (Chojnicki et al., 2018) and even the EU as a whole (Nyman & Ahlskog, 2018), have found it to be largely beneficial for public finances, the Nordic countries report predominantly negative impacts. Studies on Sweden (Ruist, 2014), Denmark (Hansen et al., 2017) and Norway (Furlanetto & Robstad, 2016) have all found that non-Western immigration had negative impacts on the fiscal balance. The common denominator of all these studies is that they examine the short-term effects of the net fiscal impact, often in the first few years following migration. Studies have assessed the long-term effects of migration to a much lesser extent. According to Ferrie and Hatton (2015), this is mainly because of the difficulties in quantifying the long-term impacts of migration on the fiscal system of the host country. A comparison between the three groups of migrants in Swedenhighly productive migrants, labour migrants and refugeesshows that from a purely short-term net contribution perspective, the first group has highly positive effects, while refugees have a negative effect and labour migrants have a weak positive effect. Ruist (2014) found that labour migrants from the rest of the EU paid in about two per cent more than they received from Swedish public finances in 2007. In other words, their net contribution of two per cent was slightly positive.

Results
The effects of migration on the welfare state are both short term and long term. The short-term effects are the net effect of migration on public sector finances, mainly the costs of refugee reception and the costs associated with assisting refugees' entry into the labour market. The long-term effects refer mainly to the increased supply of labour and remediation of the crucial demographic challenge caused by an aging population. Based on the indicators introduced in the previous section, the results are organized around the profit and loss effects of migration on the Swedish welfare system and their impact on the system's long-term sustainability.

The fiscal impact of refugees in Sweden
Refugee migration to Sweden has had negative effects on the government's finances, mainly during the initial years of refugees' residence in the country. Based on data from 2015, Ruist (2018) shows that refugees accounted for 5.1 per cent of revenues to public finances and 7.4 per cent of the costs in that year. Refugees accounted for an estimated 7 per cent of the population, and therefore had a negative net contribution of 2.3 per cent of public spending. The figures also show that refugees' over-representation on the cost side was not particularly large, at only 0.4 percentage points. A striking factor was the under-representation on the revenue side, at 1.9 percentage points. This can be explained by the lower employment rate of refugees as well as the lower incomes of those in employment, which together resulted in a net contribution to the state treasury that did not reach the level of the share of the population. In other words, the burden of refugees as a group on public finances does not depend on any particularly high-cost items, but consists mainly of the financial assistance provided to new arrivals who are unable to support themselves. This cost item has been high mainly during the years when refugee migration was increasing (Ruist, 2018).
The low employment rate among refugees in their first years in the host country means that average incomes were low in these years. Although incomes grew steadily as the years passed, it took almost 20 years for the average refugee in Sweden to make a positive annual net contribution to public finances. The simple explanation for this is that a larger proportion of migrants have been active in sectors that are socially necessary but low paid, in service occupations such as healthcare, transport, restaurants, and so on (Fr€ odin & Kjellberg, 2018). The positive annual net contribution is then estimated to last for 15 years. This is an average, but given the large variations in the refugee population in terms of employment rates and income levels, the positive annual net contribution is significantly higher among the group with high potential than the group with low potential. Figure 1 shows a notable difference in employment frequency in groups with upper secondary or post-secondary education, compared with those with pre-upper secondary levels of education. While the employment rate after 15 years among the first two groups is close to the national level, employment rates among the least educated never reaches 50 per cent of that level.

The financial burden of migration
The perceived financial burden resulting from large intakes of refugees has been one of the major concerns that has fuelled the immigration debate in Swedish society. Sanandaji (2017), who has been critical of Swedish immigration policy, highlights the high costs of receiving and integrating refugees into the labour market. He claims that each refugee costs the Swedish state an average of SEK2-4 million from the time of their arrival until they become self-supporting. This puts the annual cost of refugees as a group at about SEK50 billion. In stressing the burdensome characteristics of migrants, Sanandaji (2017) even estimates the annual cost implications for the welfare system of the total population with a migrant background at SEK100-150 billion. These sums, especially the latter, refer to the costs of the benefits and welfare services the state is required by law to offer the entire population, regardless of background.
The fact that these costs are high is linked to the Swedish welfare system's ambitious redistribution policy. Approximately half of all public expenditure is accounted for by the expenditure item 'social protection' (for the entire Swedish population), which consists of unemployment, sickness and parental insurance as well as pensions and care for the elderly. If the costs of education and healthcare are added, these welfare transfers and services amount to 70 per cent of total public expenditure, and in 2019 amounted to nearly SEK700 billion (Statista, 2020). If this is the basis for the calculations, the total costs of welfare transfers to people with a migrant background (SEK100-150 billion, as stated above) account for 14-21 per cent of the total cost of welfare transfers. Given that people born abroad make up 21 per cent of the population in Sweden (SCB, 2020a), the total costs this section of the population imposes on the welfare state is broadly proportionate to its share of the population.
However, the central government budget shows that two areas of expenditure -'Migration' and 'Establishment of newly arrived immigrants'increased sharply in 2016 and 2017 after Sweden experienced extensive immigration in 2015. Figure 2 shows that the annual cost of migration was almost SEK10 billion in 2013, but increased massively in 2016 to just over SEK41 billion. It then decreased sharply, however, and by 2019 was just under SEK12 billion. Another item of expenditure that gradually increased over the period was the establishment promotion of newly arrived immigrants, which increased from just under SEK5 billion in 2011 to a peak Source: Based on data from SCB Statistics Sweden (SCB, 2020a) and Ekonomifakta (2020).
of SEK24 billion in 2017. This expenditure also decreased slightly a year later, to just under SEK17 billion.
The negative net contribution of refugee immigration to the economy, which affects the overall statistics on migration, can be explained by the relatively large gap between the employment rate of those born in Sweden and those born abroad (Ruist, 2018). In recent years, this gap has stretched to around 15 percentage points. Although the level of employment among immigrants in Sweden is in line with that of other EU member states, it seems to attract more attention in Sweden than in the rest of Europe. Several factors can explain this gap. First, the employment rate in Sweden historically has been high (Hajighasemi, 2004), which requires migrants to also achieve the same high level of employment. In 2016, for instance, the employment rate in Sweden was 81.2 per centalmost 10 percentage points higher than the EU average (Dessimirova et al., 2017). Second, there has been a sharp decline in the number of low-skilled jobs in Sweden. The proportion of jobs that do not require any qualifications is relatively low in Sweden (4.8 percent). The only country with a lower proportion is Norway (3.8), while Germany and neighbouring Denmark each has a significantly higher proportion of jobs that require no qualifications, at 7.7 percent and 11.4 percent, respectively (Ackum et al., 2019: 30). The relocation of lowskilled jobs from the Swedish labour market to low-wage economies has worsened conditions for newly arrived groups with low levels of education, or with education that does not match current needs, making it more difficult to gain a foothold in the Swedish labour market (Hajighasemi, 2004).
Despite this employment gap, the proportion of migrants of working age in the entire migrant population is significantly higher than the proportion of native Swedes. In 2010, the dependency ratio among immigrants was 57, meaning that for every immigrant of working age there were 0.57 who required support. The figure among native Swedes was 0.76 (Scocco & Andersson, 2015: 20). This can be explained by the age distribution within the two groups. Historically, the vast majority of people who have migrated to Sweden have been young, and this has contributed to a large proportion of migrants being of working age. A recent study of Sweden (Gustafsson et al., 2017) shows that the age of migrants when migrating to Sweden is of great importance to their establishment in the labour market. The younger migrants of working age are (i.e., those in the age group 20-35), the faster they establish themselves in the labour market. The difficulty gaining a foothold in the labour market increases markedly for those from middle-and low-income countries aged 40 or above. Most of the migrants from these countries aged 50 or above are never offered employment in Sweden (Gustafsson et al., 2017).

Differences in labour market integration of migrant groups
Thus far, the employment statistics used to compare migrants' performance in relation to the native population have covered the entire migrant population, both newly arrived and established, in the same category. For a comparable measure of migrant groups' labour market performance, however, a comparison should be made between established migrant groups and native Swedes. This issue is important because Sweden has been identified as the EU member state that performs least well in creating conditions for newly arrived refugees' rapid entry into the labour market. Entry into the labour market is facilitated gradually over many years, especially for those who are better placed educationally. One EU study (Pasetti, 2015) on the integration outcomes of Turks and Iranianstwo established non-European ethnic groups in Swedenfound the performance levels of these two groups to be in line with ethnic Swedes or even higher. Labour force activity was 65 percent among Iranian immigrants and 62 percent among Turkish immigrants. Among ethnic Swedes, the figure was 63 percent. There were differences in job characteristics among the groups, as the Turkish and Iranian migrants worked in the service sector to a greater extent than the native Swedes. Regarding employment type and sector, while 80% of Turkish immigrants were employed in relatively low-skilled occupations, up to 45% of Iranians were employed in high-skilled occupationsa comparable proportion to that of native Swedes (46%) (Pasetti, 2015).
Migrants' socio-economic background, level of education and reason for migration all play a decisive role in how they establish themselves in their new country over time (Ackum et al., 2019). Figure 3 compares how the employment gap with the native population for migrant groups from Bosnia-Herzegovina, Iran, Turkey and Iraq decreases over-time. The first observation is that there is a clear difference between the groups, where people from Bosnia-Herzegovina and Iran seem to have been integrated to a greater extent than people from Iraq and Turkey. Another trend is that the employment gap, which is initially relatively high for all groups, narrows over time. In fact, regardless of country of origin, the employment gap falls continuously. Figure 3 demonstrates the failure of Swedish labour market institutions to introduce early efforts to accelerate the establishment of new arrivals in the labour market. As time has passed, however, Swedish efforts appear to have led to more stable matching, mainly of those groups with higher levels of education (Ackum et al., 2019). Late entry into the labour market has had long-term negative effects not only on the first generation of migrants, but also on the second generation, which has fallen behind in relation to the native population. A study of the incomes and level of education of 33-year olds in Sweden shows that, among non-Europeans, only those with parents born in Iran were on a par with those born in Sweden. Among all other non-Europeans, levels of both education and income were noticeably lower (Katz & € Osterberg, 2020).

Does immigration help alleviate fears of a labour supply crisis?
Despite the difficulties the labour market has faced in activating the relatively large untapped reserve of labour that refugee immigration has contributed to Swedish society, the mitigating effects of refugee migration on demographic trends and population structure are becoming increasingly apparent (Anderstig, 2012;Sanchez Gassen & Heleniak, 2016;Str€ omb€ ack, 2016). In addition, increased migration also contributes to a reduction in the supply ratio in Sweden, which can alleviate fears of a growing supply crisis for the Swedish welfare state.
The elderly support ratio, or the number of older people in relation to the number of people of working age, has risen from 29.8 per cent in 2007 to 35.5 per cent in 2020 and is forecast to rise still further to 43 per cent by 2050 (SCB, 2020b). For the combined group of pensioners and people aged 0-19 years, the dependency ratio is expected to increase to 85 per cent by 2050 (SCB, 2020b). This increasing dependency ratio will have a dampening effect on economic growth if working life is not extended or the level of activity among reserve labour resources does not increase. For this to happen, it is not only immigration to Sweden of either economic migrants or refugees that will be needed. Conditions must also be created for the employment level among migrants to increase faster than it is currently. Above all, the waiting time to enter the labour market must be radically shortened. Figure 4 shows that it takes almost eight years for refugees as a group to reach an employment rate of 50 per cent. This late entry not only burdens the welfare system with providing income support, but also reduces GDP growth and tax revenues (Hansen et al., 2017;Ruist, 2018).
The rapid activation of newly arrived refugees will be particularly important as, according to forecasts, half of the population increase in Sweden in the coming decades is expected to take place over the current retirement age of 65. The increase in life expectancy and the growing number of elderly people will lead to a higher dependency ratio and present a major challenge for the various societal functions and institutions that provide care and services to the group. In other words, the growing number of elderly people will put a strain on the pensions system, and increase the costs of healthcare and supporting the elderly (Anderstig, 2012).

Conclusions
This article assesses the effects of migration on the welfare state on the basis of criteria developed through consideration of the priorities and assumptions of its architects during the system's long process of construction. These criteria are reproduction of the labour force, creating the conditions for economic growth, and securing financing for, and the delivery of, welfare services. At an early stage in the expansion of the welfare state, labour migration played a central role in Sweden's strategy and its tradition of productive social policy. Access to a workforce that can meet the needs of the labour market in order to keep it highly productive and internationally competitive has long been strategically important.
With this as a central point of departure, an aging population is considered by far the most important motivation for increasing immigration. From this perspective, migration can be justified both from a short-term perspective, as its net contribution to the public finances can be crucial for the financing of welfare, and from a longterm perspective, as it can have clearly positive effects on the supply of labour. This is mainly for demographic reasons as the vast majority of migrants are of young working age. Among migrant groups, two categories are clearly favourable to government finances: highly educated migrants and labour migrants. Objections are often raised to the third categoryrefugee immigrantswho are argued to have high introduction costs, mainly in the initial years of residence.
A one-sided focus on the average cost burden of refugee migrants that only compares their costs during the years of stay in Sweden with the costs of the native population during the same period is highly misleading. Such a comparison ignores the extensive costs to which comprehensive welfare systems are exposed. For the Swedish welfare system, with its generous benefits and welfare services, life cycle welfare expenditure includes a social safety net during childhood and adolescence. This provides a more comparable picture of migrants' actual burden on welfare programmes in relation to citizens covered by social protection from 'the cradle to the grave'. The significant number of refugees who migrate as adults imposes no costs at all on the public finances of the host country during these years. Thus, if their costs to the welfare system are related to their age, the average total cost burden on the welfare system will be significantly lower than that of the native population.
The costs of refugee migration to the welfare system can be reduced drastically if the labour market is successful at employing this labour reserve efficiently. Although refugee migration is often about sudden movements, which can make rapid activation measures difficult, the country's long experience of labour and refugee immigration means that preparedness for the reception of new groups could have been significantly greater in 2015. A shortening of waiting time for migrants' economic integration through their more rapid entry into the labour market and thus their faster independence from contributions from the welfare system would have had a crucial impact on public finances.
In sum, and as Scocco and Andersson (2015) and Ruist (2019) note, the effects of immigration on the economy are exaggerated in the political debate. The growing opposition to immigration can be explained by the failure of the political establishment to implement the rapid inclusion of newly arrived migrants into the labour market. The literature on the impacts of migration does not find any trends that could seriously threaten the sustainability of welfare states. Modern welfare states do not experience any dramatic economic problems due to immigration. In economic terms, immigration can affect central government finances by a few percentage points, plus or minus, depending on the success of the employment policy and whether the labour market succeeds in quickly absorbing new migrants, but can by no means be considered a threat to financial stability.
In the socio-political debate, however, economic arguments, especially when focused only on the short-term and current cost aspects of refugee migration, can be very effective at convincing groups that are dissatisfied with the ongoing restructuring of the economy. The growing and obvious dissatisfaction with the scope of immigration and the failure of the establishment to include migrant groups in society are established facts. However, linking this dissatisfaction to the threats immigration might pose to welfare institutions is not based on genuine evidence. Furthermore, such an approach cannot be justified using the strategies and approaches of those who designed and developed Swedish welfare. There may be increasing fear of growing multiculturalism, and its sociocultural impacts on the preservation of the cultural and ethnic aspects of Swedish society may be of concern to some, but growing multiculturalism poses no threat to the sustainability of welfare institutions.