It takes change to remain the same: The transformation of Swedish government policy making in economic crises and the involvement of social partners

Abstract

The government responses to the corona crisis across the world has actualized an old debate regarding the role of the state and government policy making in economic crises. This debate should, the authors of this article argue, be understood in the context of how government policy making has been transformed in advanced capitalist economies in the past five decades and recent theoretical developments regarding the role of actors and institutions involved in the production and dissemination of ideas in such transformations. Drawing on an extensive review of policy making and changes in policy making doctrines, this article examines the role of government public investigations in the transformation of the Swedish government policy making in relation to economic crises where the state supports policy making through social partners.

Introduction

The recent government led crisis responses to the economic consequences of the Covid-19 pandemic (OECD, 2020) have actualized an old debate regarding the role of the state and government policy making in economic crises. In the policy literature, crisis plays a common role in explaining policy change (Ansell et al., 2016). Previous studies of the relationship between economic crises and government policy making, however, show mixed results as to what extent economic crises contribute to radical transformation of welfare societies and government policy making. In the aftermath of the most recent economic crisis in 2008 and 2009, for example, far-reaching welfare and policy reforms, including restructuring of collective bargaining (Visser, 2016), were implemented in several advanced economies, primarily in Southern Europe (Steinebach and Knill, 2017). In the Nordic and Continental welfare states, on the other hand, despite great variation in the types of policies implemented (Bergström, 2019; Kiess et al., 2017), crisis reactions were largely characterized by ‘crisis routine’ and incrementalism rather than by strong deviations from the status quo ante (Steinebach and Knill, 2017; Van Hooren et al., 2014), suggesting that in some countries there are certain characteristics that enable welfare states to remain the same even in times of crisis. The most common explanation to such resilience is that government crisis responses are shaped by path dependent pressures (Chung and Thewissen, 2011; Kiess et al., 2017) or psychological threat-rigidities, implying that in uncertain situations actors tend to opt for rigidity, falling back on old habits, rather than exploring new action patterns (Chung and Thewissen, 2011; Van Hooren et al., 2014).

Sweden is one of several examples of welfare states which have remained surprisingly stable, even in turbulent times (Schnyder, 2012), but while the resurgence of the Swedish model may be seen as a consequence of incremental path dependent processes (Howell and Givan, 2011), there may be other features that explain why the Swedish welfare state could go through the great financial crisis without fundamental change. The relative stability of the Swedish economy and welfare state, we argue, should be understood in the context of how the Swedish government policy making related to economic crises has been transformed in the past five decades, offering alternative explanations to the persistence and continuation of welfare policies in times of crises.

Sweden is a typical example of a small open economy, highly dependent on international trade. The dependence on export on global markets was crucial for the development of the Swedish model, emphasizing the need for continuous adaptation to shifting global market conditions. A central feature of the model was collectively negotiated wages, which put pressures on companies to rationalize and increase efficiency or to go out of business (Schnyder, 2012). State intervention in the private economy was never a hallmark of the Swedish model, but state intervention was extensive in the labour market, where active labour market policies were used in order to compensate workers displaced by market-driven restructuring and get them back to work quickly (Schnyder, 2012). Sweden benefitted greatly during the post-Second World War expansion of the global economy, but suffered the consequences of economic stagnation in the 1970s and the soaring inflation that followed from the two oil crises of 1973 and 1978, leading to a full-scale crisis in the steel and shipbuilding industries, and with double-figure inflation rates over the period. This prompted political action and new policies were developed that originally rendered the state the owner-of-last-resort as, for example, shipbuilding companies, being under intense cost pressure from competitors in the Far East, were taken over by a state-governed consortium. This attempt to rescue the Swedish shipping industry proved to be a failed campaign that cost the tax-payers a significant proportion of GDP, but the lesson learned by policy makers was that the state should preferably not intervene directly into business practices, but should instead act as a market marker that incentivizes actors, provides insurances, and at times subsidies for R&D projects, for example, and acts to build collaborative relationships within as well as between industries. This fundamental policy shift also implied a shift in the policies related to how government should manage economic crises, which gradually stabilized and today constitute the foundation for how Swedish industry and the economy are governed more widely.

The transformation of Swedish economy, we argue, was characterized not only by a decentralization of collective bargaining (Howell and Givan, 2011), but also by a decentralization of policy making in relation to economic crises from the state to social partners, reducing the burden on government expenses in times of economic crisis and therefore also reducing the need for major structural reforms. The Swedish economy has, thus, been going through a process of fundamental change that allows it to become more stable and resilient to economic crises. To explain such trajectories, Morgan and Hauptmeier (2021) suggest, attention should be paid to the social organization of ideas, referring to the stability of actors and institutions (including think tanks and public research institutes) involved in the production and dissemination of ideas in policy making. Unfortunately, the underlying political propositions and government reports that serve as the basis for government policy making are oftentimes, like much other archival material, ignored or simply forgotten as expert reports and governance agency documents are passed into legal reforms, policy declarations, and eventually, in some cases, become conventional wisdom (see e.g. Dobbin, 1994; Dobbin and Dowd, 2000). Thus, to fully understand the transformation of Swedish government policy making in times of crises, it is important that such documents are examined also in hindsight and with the benefit of the access to market data and other relevant information ex post facto. The purpose of this article is therefore to trace the roots of Swedish government policy in relation to structural transformation and economic crises, beginning in the late 1970s and continuing well into the second decade of the new millennium.

The remainder of this article is structured accordingly. First, recent theoretical developments regarding government and social partner policy making are introduced on the basis of the scholarly literature. After that, some of the conditions and qualities of the Swedish policy making milieu are presented to better explain the role of collective bargaining in the Swedish economy. The third section accounts for government propositions and government reports over the 1970–2020 period to describe in more detail how government policies relating to economic crises have been shaped by both immediate threats to economic welfare and growth, but also by forward-oriented and prescient thinking about the possibilities and limitations of the Swedish economy. In the final sections, the contributions of the article and the wider implications for policy making are addressed.

Government and social partner policy making

Both so-called social market economies (SMEs) or coordinated market economies (CMEs), on the one hand (as in the case of Germany, Japan, or the Scandinavian countries), and liberal market economies (LMEs) (e.g. the US and other Anglo-American economies), on the other (Rueda and Pontusson, 2000), are dependent on effective and transparent policy making to function properly. Policy making results in legal reforms when new legislation is passed in parliaments, in regulatory practices, and a number of activities that promote, for example, social norms and provide incentives to make investment areas such as in tertiary education, for instance. Jenkins and Brents (1989: 894) use the term ‘policy formulation’ to denote the ‘the process of defining the goals for major policy innovations and inserting them in the political agenda’. In contrast, ‘policy making’ is defined as ‘the process of specifying the statutory and administrative means to realize these goals’ (Jenkins and Brents, 1989: 894).

In political science, there is a vast literature on policy formation and policy making, and it is beyond the scope of this article to review this literature. However, some of the contours of policy making need to be sketched to frame the empirical material being presented below. In a seminal paper, Charles Lindblom (1959: 86) introduced the concept of ‘incrementalism’ to denote that ‘policy is not made once and for all’; on the contrary, Lindblom stated, policy is ‘made and re-made endlessly’. Consequently, Lindblom (1959: 86) continues, ‘a wise policy-maker expects that his policies will achieve only part of what he hopes and at the same time will produce unanticipated consequences he would have preferred to avoid’. That is, effective policy making operates in an incremental mode wherein a succession of policies and modifications of the policy are passed in communities granted decision making authority (e.g. in democratically elected entities such as parliaments).

In economic policy, which seeks to apprehend an emergent and continuously changing socio-economic system, susceptible to exogenous shocks and disruptive events, the incrementalism view is applicable. Given the contingent nature of economic policy, Campbell and Pedersen (2015: 679) suggest that specific economic ideas influence and shape policy making practices. In this view policy ideas frame welfare state reforms and policy entrepreneurs can promote alternatives at odds with the current institutional order (Béland, 2005), providing a contrast to the dominant view of historical conditions and pre-existing institutions as an explanation to the variety of governance modes in advanced economies (Hall and Sockice, 2001).

There is an emerging body of industrial relations scholarship on the role of ideas and discourse in shaping outcomes, and on the role of agency in developing, framing, transmitting and mobilizing support for and against ideas (Cox, 2004; Kinderman, 2017; McLaughlin and Wright, 2018). While this literature has indeed contributed to the understanding of the intersection between liberal market policies and industrial relations, it tends to reproduce the notion of discourse as an outcome of power relations, where ideas and discourses primarily become rhetorical devices or discursive weapons (see e.g. McLaughlin and Wright, 2018) in the hands of more or less powerful actors. More recently, Morgan and Hauptmeier (2021) developed a framework for analysing how ideas are produced at various levels and the networks which connect them and policy makers, by different organizations and individuals, e.g. think tanks, research institutes, policy networks and any other organizations that put primary emphasis on analysing data, producing reports, making recommendations, engaging in and advising on the development of policy proposals and communicating these ideas to key audiences, arguing that an analysis of ideas does not fully capture how policy ideas are absorbed, adapted or ignored and that it is also important to take the material realities and the various actors involved in producing ideas at different levels into account. In this view, the way neoliberal market reforms are translated in different national contexts may not necessarily be explained by the composition and nature of industrial relations institutions and the existing power relationships, but by the social organization of ideas in the national context (Morgan and Hauptmeier, 2021).

For the purpose of this article, it should be acknowledged that in national contexts where industrial relations institutions are highly institutionalized and social partners are recognized as important policy makers, governments may have an interest in supporting and maintaining the functioning and operation of social partner policy making. Social partners (trade unions and employers’ associations) govern the economic landscape through negotiated collective agreements and adjust general rules to local conditions. They also serve as an important control mechanism for the enforcement of agreed rules and regulations in the economy. Social partners are not only an important instrument for securing economic development, but also a way to ensure collective voice and democracy in the workplace and therefore also in society more broadly (Acemoglu and Robinson, 2013; Ahlquist, 2017; Brady et al., 2013; Budd and Lamare, 2020; Western and Rosenfeld, 2011). Social partners are important in negotiating wages levels and increases, but are also capable of negotiating agreements on other policy areas, such as training, employee insurance and pensions. Thus, governments may have an interest in ensuring that social partner policy making contributes to the common good and the welfare of society, but also to the country’s economic growth model (Baccaro and Pontusson, 2016).

Industrial relations scholars have begun to explore the changing role of the state in the context of market liberalization. Howell (2021) for example examines how the widespread shift from wage-led growth to other forms of growth across the advanced capitalist world has encouraged changes in the role of the state in the regulation of employment relations. These roles include market making, individual employment regulation in place of collective regulation, state-directed social pacts, and redrawing the boundaries between work and non-work (Howell, 2021), suggesting that state policies in the context of market liberalization primarily reduce the scope of and substitute social partner policy making rather than enhancing it. Budd and Lamare (2020), on the other hand, examine how the state shapes industrial relations through various roles – including as a public sector employer, a regulator of labour law, an overseer of macroeconomic policy and the welfare state, and an arbiter of what constitutes acceptable economic and social relationships. In their view the state can provide incentives for inclusionary governance, alter the composition of the legislative bodies, and the nature of policies that are enacted. This also means that governments may not only have an interest in involving social partners in policy making in order to facilitate efficient implementation, governments may also develop policies in relation to social partners. On the one hand, the state has a general interest in ensuring that policies are inclusive, covering all actors in the economy. While social partner policies may be efficient and well-functioning in some areas, there may be sectors or groups which are not covered by collective agreements or other social partner policy measures. Under such circumstances governments may need to implement supplementary policies to compensate for the lack of coverage of social partner policies or provide incentives to social partners to sign collective agreements, abolish regulations that restrain social partners from expanding policies to additional policy domains, provide opportunities for derogation or try to push social partners to implement policies, for example by threats of legal intervention. As Howell and Givan (2011) note, governments can create escape routes around and beneath formal institutions such as employment protection legislation, and create carrots and sticks to persuade class actors to reconstruct industrial relations institutions. On the other hand, if social partners are not willing or able to, governments need to find alternative ways of stimulating social partners to be engaged in policy formation, what Bell and Hindmoor (2009) call governance through persuasion. Thus, the object of government policy making is shifted from the direct intervention in the affairs of market actors to that of social partners. Therefore, in contexts where social partners are recognized and have the capacity to be involved in policy making, the state becomes a market maker via or through social partners, where government attention is directed to trade unions and employer associations rather than directly to market actors.

Under such circumstances, policy making becomes more complex. While efficient government policies are generally seen as implemented by means of well-defined regulations or through government agencies, market policy making through social partners requires interaction with a broader set of independent actors who are not directly controlled by government agencies (Morgan and Hauptmeier, 2021). The formation of such a policy making is not momentarily enacted in policy and resolutely implemented once and for all, but economic ideas, political agendas and bargaining, institutional conditions, and a variety of other local or regional conditions shape the policy making process and the policy implementation that ensues. The state needs to collaborate with, or, alternatively, delegate responsibilities to social partners regarding how to formulate policies under specific conditions, and that outcomes from the bargaining process are, occasionally, sub-optimal in the eyes of either of the participants, consequently calling for new policies. The policy making process is thus incremental rather than linear, multi-vocal rather than autocratic, and receptive to outcomes and emerging economic and political conditions so that policies that fail to fulfil expectations are gradually renegotiated and replaced in an iterative manner. This also implies a shift in the means available for government policy makers to persuade or convince social partners to engage in policy making. As will be examined in more detail below, in the case of Swedish industry policy making over the last decades, government public investigations played a central role in the process leading to a transformation of policy making where the object of state policy becomes the social partners and less directly market actors.

The Swedish economy and policy making milieu

By the middle of the 19th century, at the beginning of the industrialization of the economy and the urbanization that followed, Sweden was one of the most impoverished countries in Europe. Eleven decades later, by the late 1960s, Sweden was together with the United States and Switzerland one of the three wealthiest countries in the world, Braunerhjelm and Henrekson (2013: 110) argue. Since this point, the growth of the Swedish economy has lagged behind many OECD countries, but it is still competitive with a for the most part positive trade balance and hosting many innovative industries. Furthermore, Sweden remains a welfare state that ensures a low degree of poverty at the same time as enterprise and innovation are actively promoted.

As has been observed in most advanced welfare states, the Swedish economy has transformed in the direction of liberalization in the last half century, primarily characterized by an emphasis on free trade and deregulation of former state monopoly product markets, exposing the economy to global competition and the economic turbulence arising from global markets, often seen as creating a need to be able to adjust to global transformative pressures. In this respect, the Swedish policy making milieu and governance structure has some specific features that need to be recognized. First, unionization is comparably high and ‘almost nine out of ten Swedish workers are covered by collective agreements’ (Baccaro and Howell, 2017: 168). Employers are also extensively organized. About 80–90% of employers are organized in employers’ associations (Medlingsinstitutet, 2018). While industrial relations structures and wage negotiations have been decentralized (Thörnqvist, 1999), collective agreements remain a central governance device, regulating wages and working conditions, as well as other policy-related elements, such as health and safety, training and social security issues in the labour market. An important premise of the decentralized wage regime is that wage increases are set at such a level that it should be possible to maintain wage levels even during times of economic turbulence. Adjustment to structural and economic changes at the firm level normally takes place through the dismissal of employees, rather than through wage cuts or other forms of precarization of employment conditions, which leads to long-term erosion of welfare and competitiveness.

The absence of state-imposed minimal wages also explains the importance of collective bargaining. With minimal wage regulations, employers could avoid bargaining by simply offering the minimal wages as defined by the government. In order for employers to avoid industrial action, they therefore need to have a collective agreement. Thus, in contrast to most other European countries, Swedish industrial relations is not based on ‘automatic’ use of administrative extension of agreements to non-organized employers. Instead, the basis for the high coverage level is sector bargaining (Visser, 2013), where unions are strong enough to force non-organized employers to apply the sector agreements and sign an ‘adhesion agreement’. While other European countries are dependent on continuing state support, Swedish industrial relations is based on a form of self-regulation where both sides gain from maintaining peace in their relationships without much state influence.

Another central feature of Swedish governance structure as regards structural change and economic crises is the employment protection legislation, enforced in the mid-1970s, which gives trade unions a relatively strong bargaining power in cases of collective redundancy, providing employers incentives to negotiate derogations from legal prerequisites in local, sectoral or central collective agreements. Thus, collective agreements serve as a governance device, buffering market actors from economic shocks, stabilize employment conditions in times of structural transformation and provide mechanisms to deal with the consequences for workers when change is inevitable.

Third, Sweden invests a high proportion of its GDP in R&D work; in fact, Lerner and Tåg (2013: 174) write, ‘substantially more’ than the United States. In addition, Sweden hosts a sizeable venture capital market (Lerner and Tåg, 2013), and the Swedish state invests ‘substantial amounts of money’ in supporting academic spinoffs (Wennberg et al., 2011: 1129). Furthermore, Swedish firms collaborate more intensively with universities than in any other European country, and Swedish academic researchers, in turn, are ‘more active at patenting their research than their US counterparts’ (Åstebro et al., 2013: 289). At the same time, a comparably low return on investment in entrepreneurial activities (Wennberg et al., 2011), and a fairly strict bankruptcy law (Eberhart et al., 2017) in comparison to the so-called Chapter 11 bankruptcy option in the US, serve to erect barriers for presumptive entrepreneurs, especially those recruited from academic research communities. To compensate for these legal and institutional barriers, the Swedish state, acting through its national and regional agencies (e.g. innovation agencies, university-based venture capital funds and holding companies, incubators, etc.), provides a dense network of business and start-up support activities. By and large, such active involvement in supporting nascent industries and start-ups has translated into a dynamic economic model. Regardless of these accomplishments, the cost that befalls tax-payers should be subject to critical reviews to avoid the state subsidizing unsuccessful entrepreneurs or inflating the size and market value of moderately promising business ventures (see e.g. Åstebro et al., 2013; Wennberg et al., 2011).

Shifting role of the state in economic crises, 1974–2020

The 1970s: The state-as-owner-of-last-resort

The postwar period was characterized by stable growth in the Swedish economy. By the end of the 1960s and the beginning of the 1970s there were signs of slowing economic growth. This was particularly evident in the Swedish textile industry, which found it difficult to compete with new producers from low-wage countries. The centrally agreed wages contributed to cost-pressures, which had a negative effect on the competitiveness of Swedish companies. To limit imports from low-wage countries, certain quantitative trade restrictions were introduced. Despite these measures, domestic production continued to decline (Carlsson et al., 1981). Similar conditions prevailed in the steel- and the shipbuilding industries (see Schön, 2007: 437).

In 1974 the oil crisis made the situation worse. Demand stagnated, prices fell and many industries had surplus capacity. To overcome what was perceived as a temporary economic downturn, the state took over ownership of a number of crisis companies. During the 1975–1982 period, extensive support measures were implemented and several new state-owned companies were formed. The total net cost of the government emergency response amounted to SEK 60.6 billion (Regeringens proposition, 1986). The nationalization was based on an elementary idea: underpinned by state-of-the-art management methods, the companies would turn around so that they could flourish once the demand returned. This scenario never materialized: the majority of state-owned companies filed for bankruptcy, and thousands of workers were made redundant.

A public investigation of the Industrial Support Policy (Carlsson et al., 1981) concluded that the problems of the Swedish manufacturing sector were of a long-term nature and largely linked to international developments. ‘The state had misjudged the situation of the crisis-hit companies, and underestimated the longer-term difficulties caused by the international competition’, the report contended (Carlsson et al., 1981). This became the dominant explanation to the failed bridging policy. However, it took several years before that policy was abandoned.

The 1980s: Abandoning state intervention

In a bill presented to the Riksdag in 1987–1988, the Social Democratic government questioned the bridging policy of the 1970s and presented an industrial policy reform. According to the Minister of Industrial Affairs, the bridging policy kept unemployment down, but could not prevent redundancies in the longer term. As a consequence, unprofitable businesses remained in operation, and shareholders did not carry the full costs of substantial losses. Above all, the measures were described as creating an unhealthy relationship between the state and corporations.

State aid of this kind implies a risk that companies adapt to state aid rather than the commercial and industrial part of the business. The government financial costs became large and, in the long run, not socially economically justifiable. (Regeringens proposition, 1986: 79, authors’ translation)

In addition to the measures being ineffective and costly, it was also difficult to withdraw from the subsidies to which the state had committed itself to. The government thus had several reasons to abandon the bridging policy. By taking over the crisis-hit companies, the state had taken on a role to analyse and evaluate the future market conditions of each and every company to better allocate government resources to business ventures with a market potential. As a consequence, government ministries hired business analysts who produced numerous market analyses, investment calculations, and strategy reports to support decision making. Despite these attempts, most of the state-owned businesses filed for bankruptcy. According to the Minister of Industrial Affairs, Thage G Pettersson, the failure of the rescue activities was evident, and consequently proposed that other stakeholders are better suited to salvage distressed businesses: if any investor regarded a crisis-hit company as having good chances for survival, that investor would regard this as an investment opportunity. Based on these premises, the state should not take control of non-profitable businesses. Instead, these businesses should be let go, the Minister of Industrial Affairs proposed:

Obviously, it is unfortunate if a long-term competitive business is closed down due to temporary problems. But I mean that it cannot be the central responsibility of the state to make a test of whether the business is viable. If survival opportunities exist, the company should be able to obtain support from stakeholders other than the state, as efforts to save viable businesses should be profitable investments in the long term. (Regeringens proposition, 1986: 81–82, authors’ translation)

The government formulated the principles for a new policy. First, government support to crisis-hit companies would be dismantled and substituted by employment creation measures. Second, the ability to handle structural change was of central importance for the competitiveness of the economy. Unprofitable business operations must be closed down, or renewed in order for resources to be reallocated to growing sectors. State policies should support such transformation processes, but rather than to save the business operations per se, the industrial policy should support people immediately affected by the consequences of the structural transformation through a well-developed social protection system. Such a system would facilitate industry transitions and be supportive of new sectors. Furthermore, the industrial policy was premised on the participation of local community representatives and social partners. Third, limited supply of finance capital in the market was seen as one of the main causes of the crisis. By deregulating credit and equity markets, the finance capital supply would increase, which in turn would reduce the need for the state to assist in rescuing companies in distress.

Importantly, the government proposed a division between labour market policy and innovation and growth policy. Short-term and direct measures in times of financial crises would be the field of labour market policy, while more long-term and growth-oriented measures would be the field of innovation and growth. The new goal of industrial policy was to ‘achieve good long-term growth in business by promoting innovation and transformation in the manufacturing and service sectors and thereby maintaining good international competitiveness’ (Regeringens proposition, 1986: 1). The new policy doctrine implied a new role for the state in relation to private sector businesses. Instead of defensive measures to protect companies from international competition (or saving them in times of crisis), attention was turned to forward-oriented and market-based efforts, to promote renewal rather than to preserve established production facilities. Unfortunately, only a few years after the new industry policy was implemented, a new crisis emerged, which demanded new policy ideas.

The 1990s: The limited efficacy of labour market policy and the reforms that followed

The deregulation of the credit, equity and foreign exchange markets, which aimed at stimulating capital for investment in the private sector, contributed to a widespread crisis in the real estate and financial markets in the early 1990s (Englund, 1999). The crisis was founded on the currency speculation that arose in combination with a fixed exchange rate of the Swedish currency, tax cuts, and a growing public sector, which gave rise to failing public finances.

Several of the main Swedish banks suffered extensive credit losses, and the state contributed SEK 64 billion in support through the Bank Aid Committee. The new policy, not to support companies in crisis, thus came to a sorry end. However, the extensive support to the banks was justified by the crisis otherwise threatening the entire banking system, which would have had devastating and largely unpredictable consequences for the Swedish society for a considerable period of time.

Sweden suffered the worst macroeconomic crisis in the postwar period as GDP fell by 6%. Unemployment increased to 9% and almost 4% of the labour force participated in labour market measures, which meant that the total unemployment rate was almost 14%. The rising unemployment rate during the period was handled through extensive labour market policy measures and labour market training in particular. An evaluation of the training initiatives (Arbetsmarknadsdepartementet, 2000) identified several problems with training as a labour market policy instrument, a result that served as the basis for a general reformulation of Swedish labour market policy in the coming decades.

The transformation implied that labour market training would be offered to job seekers only in specific circumstances and if there was an explicit need from employers. Instead, policies aiming at stimulating employers to employ various vulnerable groups were initiated. As a consequence, the expenditure on labour market training decreased since 1998, while the expenditure on labour-stimulating measures, such as reduced employer contributions or support for particularly disadvantaged groups, stabilized. The state, thus, became more restrictive with the use of training as a labour market policy instrument in favour of employment subsidies aiming at stimulating employers to employ particularly vulnerable groups.

The late 1990s: The rehabilitation of social partner policy making

The banking crisis in the early 1990s marked the starting point of a movement towards an increased emphasis of social partner policy making. At the end of the 1980s, the collective bargaining system broke down (see e.g. Lundh, 2008). The central employers’ association, SAF (Svenska Arbetsgivarföreningen), wanted to decentralize wage negotiations (Thörnqvist, 1999). After the state had threatened to impose a wage freeze (see Ahlberg and Bruun, 2005), the parties agreed upon a new negotiating order and signed the Manufacturing Sector Agreement in 1997, which stipulated that the social partners in the export-oriented manufacturing sector would set the mark for wage increases for other sectors to follow (Styhre and Bergström, 2019). In this way, the risk of inflation-driven wages in the economy would decrease, while at the same time the central government would assume a role to maintain economic stability and the Riksbank was given the task to maintain inflation at 2%. The signing of the Manufacturing Sector Agreement implied a new vertical division of labour between the state and the social partners, where more space was given to social partner policy making.

In this new policy regime, the state assumed a new role, using its role as a public employer to influence and shape social partner policy making. As a consequence of the banking crisis, major structural changes were introduced within the government sector. Inspired by free market advocacy, state monopoly markets for telecommunication, railway, postal services and energy were deregulated and several former state authorities were transformed into government-owned companies. In order to deal with the expected changes, the state as an employer negotiated and signed a job-transition agreement with a consortium of public sector trade unions. This agreement formed the basis for the establishment of a transition service organization for state employees – Trygghetsstiftelsen – in 1990. The purpose was to ‘prevent as far as possible that civil servants who are laid off become unemployed’ (Walter, 2015: 58). In practice, this meant that state employees who were made redundant were offered counselling services to help them find new jobs. It was seen as important to take good care of employees, since the perceptions of the state among citizens, both as an employer and as a public authority, would be affected (Regeringens proposition, 2009). The formation of the job-to-job transition agreement was therefore a mechanism that enabled the state to set a good example, using its capacity as an employer in state authorities and public agencies.

The transformation of state authorities into state companies during the 1990s served similar purposes. The most apparent example was the transformation of the telecom monopoly market, where the state-owned company Telia settled a collective agreement, which granted redundant workers a generous transition package, supporting them to find new employment inside or outside the company. Similar agreements were signed in other state-owned companies, the Swedish Postal Services and Vattenfall, and later on also in larger privately-owned companies like Ericsson and Swedbank. The state thus became a fugleman, showing the way for other companies on how they could take responsibility when dismissing workers. However, not all companies and not all occupational groups were covered by job-to-job transition agreements.

The 2000s: Expanding social partner policy making by persuasion

By the end of the 1990s, there were five functional job-to-job transition agreements in place, covering primarily white-collar workers in the private sector, voluntary or non-profit organizations, cooperatives and the state sector. Blue-collar workers and municipality and county council employees were not covered by these agreements. In 2000, the government therefore decided to appoint a government public investigation into the effects of the already established job-to-job transition agreements, to examine whether such agreements would potentially cover the entire labour market. An important proposition was that the social partners, both in the private and public sectors, should bear the responsibility and costs of job-to-job transition and that no government subsidy would be provided. The collective agreement at the state-owned company Telia was staged as a good example.

The public investigation argued that the job-to-job transition agreements would not replace but merely reduce the burden on public labour market policy (Arbetsmarknadsdepartementet, 2002), and concluded that the financial effects mainly arise from reduced costs for unemployment benefits and labour market policy measures, as well as increased revenues in the form of income tax, social security contributions and consumption taxes. The investigation also noted difficulties in the interaction between the public labour market policy programmes and the bilateral transition service organizations and that social partners claimed that they had to pay for measures that should, in their view, have been financed by the state. The investigation concluded that any inconsistencies between government and social partner bodies should be resolved through collaboration.

After a period of intense negotiations between the Confederation of Swedish Enterprise and the Swedish trade union confederation LO, a new job-to-job transition agreement for blue-collar workers was signed in 2004. Blue-collar workers were now also covered by a job-to-job transition agreement. Furthermore, in 2006, an agreement was settled between the transition service organizations and the Public Employment Services agency, regulating how they should cooperate in cases of major redundancies.

The dismantling of the role of the state in structural change was confirmed by the Long-Term Investigation in 2008 (Finansdepartementet, 2008). Structural transformation driven by globalization was now portrayed as something permanent and ongoing, legitimizing social partner policies facilitating mobility in the labour market.

As demand and production technology are constantly changing, the economy will constantly be in need of change. If this transformation pressure is counteracted, the industrial and employment structure is preserved, leading to lower economic growth and thereby reduced consumption and reduced welfare. In a changing world, it is therefore important that resources are mobile and not tied to the production of a particular product or service. (Finansdepartementet, 2008: 42, authors’ translation)

Global specialization and free trade were seen as requiring a constant ability to change. The investigators argued that the costs of structural change, as a result of increased free trade and global specialization, are small in relation to the increased income for the country as a whole. Social partner policies were described as an important prerequisite for managing the structural transformation in the economy. The financial crisis in 2008 became a test of the new Swedish transition model, in which the state had a more marginal role to play.

The 2008 financial crisis and beyond: The crisis without state intervention

The financial crisis began in September 2008 and was initiated by a financial bubble in the US, mainly related to the mortgage lending market and the securitization of sub-prime loans. Although the Swedish economy fared relatively well in relation to other countries, the crisis still hit Sweden hard. Industrial production collapsed by more than 30 index points.

The crisis became a perfect illustration of the Swedish economy’s dependence on global markets and, in particular, the dependence on developments in the United States. At the time the main actors in the Swedish automotive industry, Volvo Cars and SAAB Automobile, were owned by US companies, Ford and General Motors, respectively. During the autumn of 2008, the blue-collar workers’ trade union proposed to government several measures which in their view could protect jobs in the automotive industry. The proposals did not result in any government intervention (or any other immediate actions by the state). In the absence of government support, social partners in the manufacturing industry signed a temporary framework agreement on short time working (STW) in March 2009, stipulating that working hours could be temporarily reduced and wages would be reduced by 20%.

As a response to the crisis, the government implemented other measures, primarily related to job creation and stimulating demand. The directive to the Public Employment Services agency, to prioritize matching activities and the long-term unemployed, remained intact. In 2009, however, labour market policy was adapted to the deteriorating economic situation and more resources were allocated to people unemployed for shorter periods. The Public Employment Services agency was to supplement and cooperate with the transition service organizations, in order to be able to provide ‘good support’ to individual workers who were made redundant (Statskontoret, 2011:11).

Despite these attempts, the consequences were significant. More than 210,000 employees were laid off in 2008–2009, employment decreased by 2.6% and about 24% of men and women aged 15–24 were unemployed. About 20,000 employees in the Swedish automotive industry lost their jobs (Lavén and Bergström, 2014). Most of the automotive industry recovered, especially after Zhejiang Geely Holding Group, a Chinese car company, acquired Volvo Cars in March 2010. However, SAAB Automobile was forced into bankruptcy in December 2011, following a long period of unprofitable business operations. To save SAAB from bankruptcy, voices were raised for the reintroduction of a more active government industrial policy, including proposals that the state should take over the ownership. Already in 2009, however, the government announced that neither the state nor the tax-payers should own car plants, a declaration consistent with the industrial policy enacted in the 1980s. The bankruptcy resulted in 3000 employees losing their jobs. The dismissed were supported by the bilateral transition service organizations and most of them found new jobs within a year. However, the head of the transition service organization for blue-collar workers criticized the Public Employment Services agency for not helping the dismissed SAAB workers: ‘The tardiness of the PES to prepare and organize help in a reasonable way is remarkable. In addition to tardiness, there were also flagrant mistakes and misjudgements’ (TSL, 2013).

Such critique prompted Riksrevisionen, the Swedish National Audit Office (NAO), to initiate an examination of the role of the PES in large-scale redundancies (Arbetsmarknadsutskottets betänkande, 2014/15:AU11). The NAO found that the participation of the Public Employment Services agency in the transition of redundant workers was consistent with the prevailing labour market policy (Riksrevisionen, 2014), which since 2007 had focused on the long-term unemployed, rather than dismissed workers. The NAO also stated that there was room for improvement regarding the cooperation between the Public Employment Services agency and the bilateral transition service organizations (Skr, 2014/15:42), recommending the Public Employment Services agency inform its employees about the role of the transition service organizations in connection with major redundancies, and to review the agreements with the transition service organizations. The government, however, emphasized the complementary role of the bilateral transition service organizations in relation to the work of the Public Employment Services agency, while emphasizing the importance of collaboration between them, ‘whenever necessary’ (Skr, 2014/15:42). The labour market policy’s focus on long-term unemployed made it difficult, at an early stage, to set up measures to support workers notified as redundant. The role of the state was, thus, defined primarily as ‘complementary’ to the bilateral transition service organizations offering support to redundant workers.

The financial crisis in 2008–2009 consequently passed more or less without state intervention. To a large extent, the dismissed workers were handled through policy measures organized and financed by the social partners. When the social partners asked for additional support to mitigate the effects of the crisis, the government declined. The role of the state was mainly limited to cooperating with the bilateral transition service organizations, but the working methods for this collaboration were not clear. The role of the state in connection with major workforce reductions and closures was thus in principle completely dismantled.

At the end of the crisis, a new job-to-job transition agreement was signed in 2010 and a new transition service organization for the municipalities and county councils was established in 2012 (Walter, 2015). As a result, the majority of employees in the Swedish labour market were now covered by transition agreements.

The 2010s: Supplementing and prompting social partner policy making

The 2010s were characterized by relatively stable economic development. Announcements of redundancies remained at a relatively low level compared to previous decades. Other concerns dominated labour market policy making. However, a number of amendments in the government policy related to structural change were made, establishing the state as a supplement to social partner policy making and prompting social partners to expand policy making to new areas.

In 2012 the government formulated new guidelines for the Public Employment Services agency. The assignment was partly a result of the events following the SAAB bankruptcy in 2011. According to the new guidelines the Public Employment Services agency should work with a so-called assessment support to identify job seekers at risk of long-term unemployment. In addition, persons at risk of long-term unemployment should be offered individualized support at an early stage. Furthermore, in 2013 the Public Employment Services agency was requested to report on their performance regarding collaboration with other actors, such as municipalities and the bilateral transition organizations, to facilitate labour market transition for the people who move from activities within the transition organizations to the activities of the PES (Arbetsmarknadsdepartementet, 2013).

In 2011, the blue-collar workers’ trade union confederation, LO, suggested that the state should reintroduce the short-time working scheme that was abandoned in the 1990s (Larsson et al., 2011). In response, the government introduced a new law on short-time work in November 2013 (Finansdepartementet, 2013). According to the law, the state would grant support for short-term work if there was a particularly deep recession, and if the support was not considered to counteract a socio-economically desirable structural transformation or cause other significant socio-economic disadvantages. To be eligible for the subsidies, the employer must sign a central collective agreement.

In 2014, the government modified the conditions for tax exemptions for bilateral transition service organizations, removing the obstacles for social partners to renegotiate their already established policy measures and expand the scope of future collective agreements, e.g. to include additional services such as preventive measures and training to better facilitate labour market integration for dismissed workers. Under the previous legislation, the bilateral transition service organizations would only be exempt from income tax for activities directed at employees who were already made redundant. In the new legislation, the scope of the tax exemptions was expanded, enabling tax exemptions for transition service organizations that offer support or training to employees who are unemployed, or at risk of unemployment within a period of five years. The social partners could now come to agreement on preventive measures to support workers at risk of unemployment, even if they have not yet been formally notified. The state thus removed some of the obstacles for social partners to expand policy making to other policy domains.

In February 2018, the government initiated a public investigation to analyse how the state-funded short time working (STW) scheme could be made more competitive in comparison to other countries. The investigation was presented in late 2018 and the implementation of a reformed STW scheme was included as one of the priority areas of the new Social Democratic/Green government formed in January 2019, supported by the Liberals and the Centre party. Another priority area in the formation of the new government was the initiation of a government public investigation on a ‘Modernized employment protection legislation’, stipulating that if social partners did not come to an agreement, the proposals of the new public investigation would be enforced. Thus, the government put pressure on social partners through the threat of legal reforms. In the spring of 2020, when the public investigation was to be presented, the economic consequences of the corona crisis had already been revealed and in response the government introduced an extensive package to support firms’ ability to deal with the economic effects of the crisis (STW schemes, compensation of revenue losses, loan guarantees, etc.). These measures expanded the state arsenal of policies and may be seen as supplementary to the already established social partner policies (IER, 2020). Thus, the recent policy response to the pandemic may not be seen as a return to state policies of the past (the government as a last resort), or a deviation from market policy making, but as an example of government policy making, being supplementary to social partner policies and measures.

In December 2020, a consortium of trade unions and employers’ associations signed a new general agreement on job security, restructuring and employment protection, which extends employer-financed schemes for workplace training, life-long learning and job-to-job transition in exchange for reformed rules in the employment protection legislation. In an official letter to the State Minister and the Minister of Labour Affairs, the social partners write: ‘in order for the parties to finally adopt the agreement, it is assumed that the state implements certain changes in the employment protection legislation, that the state introduces a new and parallel public student support system and that the state introduces a new public job-to-job-transition support for employees in companies that are not covered by collective agreements’ (Dahl et al., 2020: 1). In January 2021, the government initiated three further collaborative investigations, where government experts together with representatives from the social partners explore the possibilities to implement the demands from social partners. In other words, the government policy to threaten social partners was turned back to the state. The social partners take on a role to formulate the policies for the state to implement, rather than the other way around. The outcome of the last round of investigations remains to be seen.

Over a period of five decades, government policy making related to economic crises and structural transformation has changed substantially. Instead of directly intervening in structural change and labour markets, government policies are now operating through or via social partners. The transformation did not mean a change in the type of policies enacted. However, the role of the state has shifted to supporting, supplementing and cooperating with social partners.

Discussion and conclusions

In this article, we have analysed how government policy responding to economic crises has been adjusted and modified over the past five decades, from the 1970s bridging policy in which the state took over ownership of crisis companies, to the 2010s, where the state has withdrawn and intervenes only in the case of negotiation failures as social partners are given greater responsibility for formulating and implementing policies. The empirical material reveals a relatively slow and incremental transformation process, which of necessity intensifies during major crises. The oil crisis in the 1970s prompted the bridging policy, the banking crisis in the mid-1990s initiated a process of transformation of the industrial relations system, and at the time of the 2008–2009 financial crisis, responsibility for direct measures in connection with major closures and reductions had been largely left to the social partners. Since then, state policy has further developed into a supplement to the joint actions of the social partners, illustrated by the extensive policy measures introduced to maintain employment during the coronavirus pandemic. More importantly, this process reveals a successive shift in government policy making, where the object of government policy making becomes the social partners rather than directly intervening in the relationship between market actors, employers and employees. This shift implies not only a different role for the state to shape and regulate industrial relations (Budd and Lamare, 2020; Howell, 2021), it also involves the introduction of a new type of governance, where government policies are directed towards incentivizing or persuading social partners to come to agreements or, as a last resort, to threaten them with legislation when they fail to come to agreements. The government used its role as an employer to set an example for other actors to follow and the state also provided supplementary actions and policies in sectors or policy areas where social partners failed to come to agreements. While this reveals an incremental process of the shifting role of the state and an expansion of social partner policy making, taking over more and more responsibility for policy areas that previously used to be part of government policy making, it also reveals how government policy making is increasingly done through or via social partners, implying a new set of tactics and strategies.

There are arguably several plausible explanations for this development. The changed role of the state can be treated as a consequence of an institutional learning process (see e.g. Bartley, 2007; Blyth, 2002; Douglas, 1986), or ‘institutional adaptiveness’ (Casper, 2000), wherein actions are implemented, evaluated against outcomes, and translated into more specific policy proposals. Previous studies have emphasized the importance of the organizational and institutional machinery that generates data, research, policy recommendations and other ideas that influence policy making (Campbell and Pedersen, 2015), and Morgan and Hauptmeier (2021) illustrate how think tanks and research institutes were important elements of the social organization of ideas explaining the transformation of the German and the US economy. In this article we have explored the role of government public investigations in the institutional adaptation of government policy making. As shown, evaluations identified limitations in the 1970s bridging policy, including high costs, inadequate use of tax-money, insufficient effects, and unintended consequences, legitimizing a reformulation of the role of the state in relation to economic crises. New guidelines and ambitions were formulated and implemented, which were later reviewed and evaluated to once again form the basis for new measures. Later, the emergent experiences of social partner policy making were gathered and evaluated in another government public investigation, promoting social partners to expand negotiations to cover other sectors and occupational groups. More recently, social partners were threatened with reforms of employment protection legislation unless they came to an agreement on how regulations could be amended to serve not only their own but also societal interests. Government public investigations have, thus, played a crucial role in the formation of this new policy making regime where the government abandons its own role in policy making.

Another explanation can be found in the emergence of the EU state economic aid rules that prescribe that the state can no longer offer support to companies in distress as the Swedish government did in the 1970s. Swedish EU membership, beginning in 1994, thereby provides an institutional explanation for why industry policy and rules supportive of structural transformation have changed. The integration of European monetary policy, in particular the Stability and Growth Pact, also means that Sweden (despite not being part of the European monetary union and issuing its own currency) can only to a lesser extent pursue a divergent monetary policy. If measures to manage economic crises are primarily financed by the state, the state’s expenditures would increase in the face of a system-wide economic crisis, which in turn triggers budget deficits and, in the long run, results in cuts in the public sector – so-called austerity policies (Jordà and Taylor, 2016; Major, 2014; Schui, 2014). Transferring the responsibility for managing job-to-job transition to the social partners can thus be regarded as a way to ensure that crises are taken care of where they arise so that they do not generate excessive costs for the state. From this perspective, new institutional conditions shape structural change in the private sector, otherwise potentially handled in a different way. It should be noted, however, that European membership has not prevented the Swedish government from developing its policy making model that relies on social partners negotiating collective agreements, and with Swedish state agencies only serving a supportive role, intervening in the process only when asked for by either negotiating party. Although not necessarily intentional, this iterative policy making process in which the state has sought to find a constructive balance in relation to social partners, has evolved for more than five decades.

This article contributes to debates in the policy making literature regarding the changing role of government policy making in the context of advanced capitalist welfare states by providing a case that explores the ways in which state policy making can take shape when social partners are recognized as and have the capacity to become an integral part of policy making. As demonstrated in the empirical section, the Swedish government, representing both social democratic and centre-right governments, has over time acted on the basis of the joint policy idea that international product market competition is an effective mechanism for optimizing the use of economic resources, but that this pressure on corporations to optimize their performance needs to be combined with policies that counteract the adverse effects of the economic system of competitive capitalism. As shown, policies to deal with the consequences of economic crises, such as job-to-job transition previously delivered through the public employment services, were successively handed over to social partners, while at the same time government policy measures related to economic crises, for example early retirement schemes, were abandoned or withdrawn. At the same time the state adopted a role to persuade, prompt or pressure social partners to engage in or expand their scope of policy making to new domains.

This shift of policy making has several possible implications. Previous studies suggest that in times of crisis governments tend to fall back on old habits or are pressured by path dependent processes (Van Hooren et al., 2014). In the Swedish case, where crisis measures are financed and organized by social partners, the government does not need to increase spending on crisis measures. As a result, the state can give priority to groups or policy areas not covered by social partner agreements and may instead use government resources to expand investments in measures to promote economic growth. Most importantly, the decentralization of crisis policy making to social partners implies that government spending in other social policy areas is protected, reducing the need for social policy reforms and austerity measures, ensuring the continuation and stability of the welfare state. The transformation of government policy making in times of crises thus offers an alternative explanation to the relative resilience of the Swedish welfare state in times of crises. This study thus contributes to debates regarding the relationship between economic crises and policy making (Steinebach and Knill, 2017; Van Hooren et al., 2014).

The contribution lies not only in the empirical detail of how such transformations may unfold, it also contributes to recent debates regarding the explanatory power of ideas and discourse in such transformations (see e.g. McLaughlin and Wright, 2018; Schmidt, 2017). Indeed, economic policy ideas were important, but the way these ideas were communicated, in this case through seemingly neutral government public investigations, played an important part in the transformation of government policy making. As shown, government public investigations served the function not only to promote and legitimize government policies, but also to deconstruct previous policies, legitimize the withdrawal of government policy making, and at the same time impose threats to engage social partners to expand their policy making. Thus, in contexts where social partners are recognized as policy makers, government public investigations may also be seen as part of what Morgan and Hauptmeier (2021) call the social organization of ideas, contributing not only to change or disrupt institutions, but also to maintain them. The analysis of policy making should, thus, not be limited to the ideas promoted through skilful rhetoric or used as discursive weapons in political battles (McLaughlin and Wright, 2018), but to understand the transformation of policy making and its various modes in advanced capitalist societies, it is also important to pay attention to how ideas are produced and organized and the dynamics that unfold in the context where they are promoted.

The decentralization of policy making to social partners, however, puts governments in a dilemma. In times of crises governments are often pressured by public opinion and media to show that they are doing something to mitigate or resolve the crisis (Ansell et al., 2016). The decentralization of crisis policy making, however, implies that the government no longer has the tools available for dealing with the crisis, which would otherwise provide politicians with the appearance, at least, of having the ability and authority to manage crises that threaten the economy and the society. This inability may in turn reduce public trust in the state and the government and reduce the chances for re-election. Thus, the decentralization of policy making to social partners may also have political consequences, as Héritier (2017) puts it, ‘politics determines policies’ and ‘policies determine politics’, but the absence of policy making also determines politics. The withdrawal from policy making in economic crises, while maintaining the fundamentals of the welfare state, may make governments more vulnerable to political change.

In this perspective, policy formulation and policy making (Jenkins and Brents, 1989) is a politico-economic process that is dependent on a variety of contingencies. Swedish industrial policy shifted over time and in accordance with new economic ideas, wherein the deregulation of the finance market in the late 1980s is arguably the most important policy change, and the outcomes have not always been possible to anticipate. In the end, the government has gradually changed from being an owner-of-last-resort of defaulting or distressed companies, to an active market maker that primarily operates through social partners. This policy making trajectory should not be mistaken for what Sutton and Dobbin (1996) and Dobbin and Sutton (1998) refer to as the ‘weak state’ theory, which fails to recognize the state’s role in ‘shaping private enterprises’ (Dobbin and Sutton, 1998: 472), perhaps not so much ‘through direct mandates’, but on basis of ‘moral suasion and the manipulation of market incentives’ (Sutton and Dobbin, 1996: 798). The state may govern product and labour markets on the basis of a leniency preference principle, but the consequences are substantial and material just the same.

It should be noted that the findings are limited to policy making in relation to economic crises. We are convinced, however, that similar processes and developments may be found in other policy areas, such as pension schemes and unemployment benefits, where social partners take on an increasing share of policy responsibilities. While Sweden may be construed as an extreme case, with its rather well-known institutional conditions, there may be important lessons to bring to the analysis of other countries going through similar institutional transformation processes. Future studies should explore the sources of stability in times of change in other countries and the different ways in which the state can incentivize or persuade social partners to engage in crisis policy making. The lessons learned from the Swedish case seem to be that it takes change to remain the same.