Stress testing the Nordic models: Manufacturing labour adjustments during crisis

Introduction

The Nordic countries are small, open economies accustomed to economic change. With the global financial crisis in 2008, their labour markets faced challenges as international market demand slumped. Katzenstein (1985: 24) emphasized that Nordic labour markets are equipped to handle economic shocks, as the interplay between the strong collective organizations and the collective insurance provided by the welfare state serves to cushion insecurity and muster legitimacy for swift policy adjustments through concertation with the state. But how did such cooperation influence company-level adjustments?

The Nordic labour regimes have traditionally been associated with a strong tier of workplace negotiations, within a framework of centralized coordination and state support (Kjellberg, 1998; Sippola, 2012). Regulation of collective dismissals is in comparative perspective relatively weak, especially in Denmark (Muffels and Luijkx, 2008; Rogers and Streeck, 1995); but law and central collective agreements ensure company trade unions – the local union ‘clubs’ – the right to information, consultation and negotiation in a range of areas, including labour adjustments (Bruun et al., 1992; Sippola, 2012). Thus local negotiations over workforce adjustments concerning wages, working time, use of temporary lay-offs and dismissals are a crucial arena for bargaining over crisis adjustments (Glassner et al., 2011).

In this article we study how the actors in the most internationally exposed sectors of the Nordic economies responded to the crisis at plant level. They faced different possibilities for adjustment because of national variations in the institutional regulation of dismissals and available means of workforce adjustment, but how did such differences influence the extent of cooperative adjustment, the way adjustments were negotiated and the distribution of costs? By comparing companies within a single sector with similar product market conditions we can detect the impact of differences in national institutions and regulations on the pattern of adjustment and union involvement (Marginson et al., 2004: 18–20). The decline in output in the wake of the 2008 crisis was largest in manufacturing, which was most directly hit by the collapse in export markets. Manufacturing is highly unionized, and though density has declined in recent years, the figures are still high in an international perspective, ranging between 79 percent in Sweden and 58 percent in Norway (Kjellberg, 2010; Nergaard, 2010).

Labour adjustment in the Nordic countries, inspired by the Swedish Rehn-Meidner approach (Meidner and Rehn, 1953), has traditionally emphasized external flexibility and reallocation of labour from less to more productive companies. Egalitarian, centralized wage bargaining forced companies to restructure or go out of business, since they could not resort to downward wage competition. The focus on external flexibility and mobility, combined with relatively generous unemployment benefits and reallocation of labour, has in recent years been replicated in the Danish notion of flexicurity (Ibsen, 2011; Madsen, 2004; Muffels and Luijkx, 2008). Nordic trade unions have usually rejected concession bargaining on the American model or as involved in German company employment pacts (Lehndorff, 2011; Zagelmeyer, 2011).

Labour adjustments negotiated locally between employers and trade unions are framed by institutions and regulations that both constrain and enable different cooperative dynamics and solutions for workplace actors. These institutions rest on compromises that may be subject to change, especially during economic shifts, either by tripartite or bipartite central negotiations, or through local negotiations and bottom-up pressure (Howell and Givan, 2011; Katzenstein, 1985; Mahoney and Thelen, 2010). Local negotiations are subject to a peace obligation and are embedded in relationships where bargaining usually is voluntary, informal and based on continuous exchange (Kjellberg, 1998).

In spite of many similarities, the Nordic labour markets show significant differences concerning temporary lay-off institutions and work-sharing schemes (Hijzen and Venn, 2011), dismissal regulation (Berglund et al., 2010; Sigeman, 2002), early retirement, access to company-specific training schemes and in traditions for local codetermination and tripartite cooperation. Together with variations in rules and customs regarding leadership styles (Byrkjeflot, 2001; Trygstad and Hagen, 2007) and trade union approaches, such differences can be expected to affect the pattern of cooperation, power relations and adjustment at company level. For instance, strict dismissal regulation may increase trade union bargaining power (Elster, 1992). We therefore raise three interrelated questions concerning such national variations. First, how were labour voice, cooperative relations and the extent of negotiations over labour adjustments at plant level influenced by differences in rules regarding participation? Second, how did variations regarding regulation of dismissals, early exit options, seniority, income security and temporary lay-off schemes influence such negotiations? Third, do variations in regulations, available means of labour adjustment and cooperative efforts have distributive consequences at plant level?

Research design and methods

The study was based on case studies carried out during 2010 in 15 manufacturing companies – three each in Denmark, Finland, and Sweden, and six in Norway. All plants were part of larger corporate structures; all three Swedish plants and two of the Danish, but none of the Finnish and only two of the six Norwegian, were foreign-owned. Even though the case studies were limited to a single plant, the adjustments sometimes involved other plants in the company or group, and when relevant we also gathered information about these. To ensure that the cases were typical for the phenomena under study and to enable comparison, we selected plants (in most cases with sizeable export markets) which had experienced a substantial drop in demand, leading to a substantial labour adjustment process. Products included ball bearings, engines, windmills and pumps. Two plants (NO4 and DK2) were not oriented to export markets, but both produced for the crisis-ridden construction sector and thus faced similar pressures for labour adjustment. All plants were based on fairly traditional mass production, with medium advanced technology, a sizeable blue-collar workforce and high union density. All had more than 50 employees, and all but three over 200, ensuring that the management– trade union relationship had some weight and that the number of workers affected by the adjustments was large enough to gain insight into their distributional effects. We contacted central unions and employer federations, used the Eurofound restructuring database (EMCC, 2009) and our own knowledge about the crisis at company level to find cases fitting the criteria.

Data were collected through interviews with representatives of site management (HR director and/or production/division manager) and of the main blue- and white-collar unions. Interview data were supplemented by annual reports, business data, management–union agreements and so on. A common interview guide was used in the fieldwork, and to ensure data quality all the interviews in Denmark, Norway and Sweden were summarized and emailed to the interviewees for comments. In this article we anonymize the plants, using country initials and numbers.

Background and institutional frameworks

Initial crisis responses: Central and sectoral adjustments of regulations

Besides forceful macroeconomic measures, the Nordic countries responded to the market collapse by expanding active labour market policies and education (Jochem, 2011), and partly by adjusting the duration and generosity of the unemployment benefit system. In Denmark, Finland and Norway, temporary lay-off or work-sharing schemes were also made more accessible.

Finland and Sweden eased access to unemployment benefits in 2009 (temporarily in Sweden, after being tightened severely in 2006), as was the minimum age for receipt of open-ended unemployment pensions in Finland (Jokivuori, 2011; Lovén, 2009). The Finnish criteria were further loosened in 2010. In contrast the Danish government reduced the maximum duration from four to two years in spring 2010, and tightened criteria for re-qualifying (Jørgensen and Schulze, 2011). In response, DA and CO-industri (the ‘cartel’ of LO affiliates in manufacturing) agreed on a severance pay supplement for workers with more than three years tenure raising the unemployment benefit to 85 percent of previous pay for a limited period (Ibsen et al., 2011: 333; Jørgensen, 2010).

Companies, employers’ organizations and trade unions in Denmark, Finland and Norway put pressure on the state to relax the arrangements for temporary lay-off schemes. In Denmark, temporary changes enacted in March 2009 allowed two weeks of work to be followed by one or two weeks of unemployment (Jørgensen, 2011b). Large Danish companies called for prolongation of the work-sharing period from 26 weeks to 18 months. This was supported by most trade unions on condition that the period off work did not count as a supplementary benefit period (Jørgensen, 2011a), but the centre-right government refused further changes. In Norway, temporary changes in spring 2009 increased the maximum duration of lay-offs from 30 to 52 weeks, and the period without benefits was reduced. Even after these changes, the costs to employers for hours not worked the first month were 25 percent, which was in the high end of employer costs in the OECD area (Hijzen and Venn, 2011: 12), and much higher than the 15 percent in Denmark and zero in Finland. The minimum threshold was also decreased, allowing employers to lay-off full-time workers two days each week, instead of two and a half. In Finland the minimum threshold was temporarily lowered to one day off work per week, and in January 2009 the ‘change security’ activation programme, which included retraining, was extended to temporary lay-offs (Arnkil, 2011). Company-level training was also made more accessible by reducing employer contributions for small and medium-sized companies (Miettinen, 2009).

The flexibility enabled by such arrangements was most extensively utilized in Finland, where over 90,000 employees, just under 4 percent of the entire labour force and 31 percent of redundant workers, were temporarily laid off in 2009. In Denmark the number of employees engaged in work-sharing peaked at around 18,000 in May 2009 (Stuvøy and Jørgensen, 2009): 0.6 percent of the labour force and 9 percent of redundant workers. In Norway the peak was just over 18,000 in April 2010 (Olberg, 2010): 0.8 percent of the labour force and 20 percent of redundant workers.

Exceptionally for Western Europe, Sweden had no equivalent temporary lay-off scheme when the crisis hit. However, in response to dramatic job losses in 2009 the main manufacturing union, IF Metall and its employer counterpart, Teknikföretagen, signed a path-breaking ‘crisis agreement’, establishing procedures for negotiated work- and burden-sharing at company level. Working time could be reduced to 80 percent without compensation, subject to a company agreement (thus requiring trade union consent). Employees could be laid off for a limited number of days with a corresponding or lower cut in pay. It was the first time that a national Swedish union agreed a reduction in benefits for its members, and subsequently other unions came under pressure to follow. As a result, many workplaces settled for agreements that allowed unpaid, temporary lay-offs in order to save jobs (Henriksson and Kullander, 2011). According to a report by IF Metall (2009), almost a third of its clubs signed a local crisis agreement in 2009, covering about 60,000 workers (38% of the membership) (IF Metall, 2009). According to the white-collar manufacturing union, Unionen (2009), local branches signed 138 crisis agreements, covering nearly 14,000 members.

In the 2010 bargaining round, the crisis agreement was prolonged until October 2010 on condition that re-employment of formerly redundant workers gained priority over the use of temporary agency workers (Kullander and Henriksson, 2010). In 2012 the social partners in manufacturing proposed a revised version of the agreement as a permanent measure to handle economic crisis throughout the private sector. After revising the proposal, the government in spring 2013 decided to introduce a public scheme for short-time work, with the state, the employers and the employees sharing the costs. The scheme applies only if severe recession is expected, and if necessary structural changes are not obstructed.

In no country were pay cuts or wage freezes centrally agreed during the downturn. Pay increases were low, but various procedural innovations were developed. In Norway, the social partners in 2009 agreed an opt-out from the national pay increase for companies facing economic difficulties, if the local union accepted (Nergaard, 2011). In Finland and Sweden, similar innovations were included in some of the manufacturing wage agreements, as in the agreement for blue-collar workers in electronics, ICT and metal-working in Finland in August 2009 (Glassner et al., 2011: 312).

Employment adjustment: What scope for cooperation in hard times?

Sequence and sources of internal and external adjustments

The depth of the fall in product demand in 2009 – the median for the case companies was around 40 percent – required swift and substantial cost savings and workforce adjustments, and the available repertoire of relevant measures was crucial in the consultations. The choice of measures was influenced by the perceived depth of the crisis: many companies saw a need to adjust labour and production capacity immediately.

The measures used by the companies fall within six broad categories: cutting external employees, temporary changes in working time, internal redeployment, pay cost savings, voluntary redundancies and early retirement, and dismissals. The use of measures is summarized in Table 1. Drawing on Greenhalgh et al. (1988: 243), it depicts the short-term impact of these measures on company costs, and ranks them by their assumed contribution to integrative solutions which are perceived as the combined preferable option for both management and trade unions (Walton et al., 1994).

OPEN IN VIEWER

Table 1.

Adjustment measures.

	Type of measure	Subtype	Company
			DK	FI	NO	SE
>———-Increasing integrative bargaining solutions——— →	1. Cut external employees			2	2,3,4,6	1	>——Increasing short-term savings ——→
	2. Temporary working time adjustments	a) Flexible working time arrangements	2	1,2,3	2	1,2,3
		b) Temporary lay-offs, work-sharing	2,3	1,2,3	1,2,3,4,5,6
		c) Training courses without normal pay	2	1,2		1,2,3 a
		d) Reduced working time and pay			1,2,3,5	1,2
	3. Internal redeployment		1,2,3	2	2,4,6	1,2
	4. Changes in pay systems		2,3
	5. Exits and early retirement	a) Voluntary redundancy with severance pay	1,2,3			1,2,3	←Increasing long-term savings —<
		b) Early retirement incentives	1,2			1,2,3
	6. Dismissal	a) Phasing out temporary employees		2	2,4,6	1,3
		b) Dismissals	1,2,3	2,3	1,2,3,4,5,6	1,2,3 b

a

In SE1, training measures were funded by the operating budget. In SE3 an externally funded project covered the training costs, while the wage costs were underwritten by the company.

b

In SE2 only a few white-collar employees were dismissed.

Discussion and conclusion

As expected within the Nordic two-tier systems, company-level negotiations between management and trade unions proved important in achieving swift workforce adjustments. Differences in dismissal regulations and in the existence and form of state-sponsored temporary adjustment measures influenced the extent and form of cooperation, the way these adjustments were negotiated and the distribution of costs.

We find that the company actors make their adjustments in articulated interaction with central actors and institutions. The Nordic countries showed different patterns of institutional adaptation, depending on the arrangements prior to the crisis and the subsequent level of tripartite cooperation and state support. In Sweden, with no state support, collective bargaining and autonomous sectoral actors were decisive. When facing market shocks, these contingencies worked as beneficial constraints, spurring cooperation at company and sectoral levels, enabling the creation of innovative agreements at both levels. In Norway and Finland the state-funded temporary lay-off measure shaped the cooperative efforts, but even so there was significant renewal in local agreements, and central actors showed ability to respond swiftly by adjusting such arrangements as temporary lay-off rules. These changes underpinned local choices and flexibility. Norway faced a smaller unemployment crisis, and had ‘unlimited’ funds to use for temporary lay-off arrangements. Denmark seems to exhibit an intermediate approach. The budget crisis and government reluctance to contribute caused a political blockage in the established tripartite channels, restricting cooperation at national and sectoral levels. Since dismissal rules were the most lenient amongst the Nordic countries, and the work-sharing scheme was more restrictive than in Finland and Norway, in 2010 Dansk Metal negotiated increased severance pay, increasing the cost of dismissing employees.

In spite of such differences, we found largely comparable results in terms of social outcomes. Still, while one should be cautious with generalizing from outcomes when using a case study design, our cases follow the national statistics in that temporary lay-offs were used more in Finland and Norway than in the other countries. In all cases, across the countries, the actors knew that the alternative would be dismissals; the priority was to save jobs. Further, while seniority rules meant that the youngest workers were worst affected in Norway and Sweden, greater employer influence in Denmark and Finland seemed to create a more differentiated pattern of dismissals: the ‘least productive’ tended to be dismissed, regardless of age. Indeed older workers, with narrower skills, might be most vulnerable. The early retirement systems in these two countries seemed to smooth conflicts in case of such choices, as in DK2.

While the extent of management–labour cooperation and trade union voice in handling labour adjustments during the crisis varied between companies in the four countries, the most striking variation was in the extent of strategic renewal and institutional innovation. The negotiation of local and central agreements facilitating temporary cuts in working hours and pay in Swedish companies was remarkable in several respects. This was the first instance in modern history where Swedish metalworker unions ceded acquired rights and indirectly accepted cuts in pay, a significant departure from the traditional Rehn-Meidner model with its emphasis on external flexibility as means of adjustment. The gain was that the unions achieved a more even distribution of losses among the workers retained, and as such an integrative solution. While companies in all countries emphasised the importance of retaining company-specific skills, the Swedish move towards more emphasis on internal working time flexibility resembles the German model of local employment pacts. Eventually, tripartite talks in the wake of the crisis agreement led to the government proposal described above.

In terms of interests and objectives, there was no salient difference between the actors in the different countries. Both sides shared the overriding goal of rescuing the site and retaining competences and crucial company-specific skills, on which production and the remaining workplaces would be reliant. Both sides also soon realized that external flexibility would be necessary to rescue as many jobs as they deemed possible, but views naturally diverged as to where and how the line was to be drawn – and the division of costs – depending on which alternative options were available. The conflicts of interests between management and labour over such distributive issues took different forms, depending on the alternatives that were at hand. In all cases, the division of costs for employers and the losses of income or jobs for workers were at the heart of discussions, but how the burdens were to be distributed among the workers was also contentious. In practice, such disagreements first and foremost played out with regard to the number of dismissals as against alternative adjustments (reduced pay, working time, lay-offs, severance pay, retirement) and with regard to the criteria and selection of those dismissed. For the employers such questions were primarily functionally related to skills and productivity, whereas for the unions they were more a matter of justice and equality. The combination of common and conflicting interests meant that the actors both were dependent on each other, and, to different extent, had control over parameters that could be used in exchange processes and as power resources.

The study reveals interesting variations in the power relations and the extent and depth of cooperation. Swedish unions, and to considerable extent also the Norwegian, were markedly more involved and influential in decision-making over labour adjustments than were the Finnish and Danish. With broadly the same general rights of participation, the main source of such variations stems in our view from differences in regulations and criteria for collective dismissals. The access or not to publicly funded schemes for lay-offs or reduced working hours also influenced the options and constraints facing the actors. Besides reducing the level of conflicting interests between management and unions, the buffer function of such schemes apparently implied that the depth of union involvement in decision-making was less profound than in Sweden.

The requirement of union consent regarding the application of seniority criteria in Swedish (and Norwegian) companies, and the absence of external public buffers in Sweden, entailed that the mutual interdependence between the actors and the pressure to find negotiated solutions were stronger than in the other countries. Swedish unions thus had control over parameters that were of utmost importance to management, at the same time as they faced stricter economic constraints. Paradoxically, it can be argued that the Swedish company actors, constrained by lack of short term, external flexibility schemes, engaged in rather balanced, integrative internal adjustments, notably for those who remained in the companies.

In Norway, where the central collective agreements allow much less local working time flexibility, the incentives to engage in such negotiations were much stronger during the crisis, also because of the costs and losses in pay associated with the temporary lay-off scheme. Such pressures were considerably higher in the early phase before rules were relaxed, prompting a number of local agreements on internal work-sharing schemes which were preferred by both sides because they involved fewer costs. Further, the unions controlled the conditions for flexible application of the seniority principle, but the access to external labour hoarding co-funded by the state eventually made the pressure to negotiate internal flexibility less urgent than in Sweden.

This was even less so in Finland, where the issue of dismissal selection was mainly under management control and the alternative of temporary lay-offs was virtually costless for the companies. Similarly in Denmark, managements’ prerogative to decide the number and selection of dismissals, and the somewhat strictly regulated work-sharing scheme meant that the parties had less incentive to negotiate over further internal flexibility beyond the already quite flexible working-time schemes. This implied that there was little to negotiate about, and the plant-level unions therefore played a minor role, as seen also in another recent study (Ibsen, 2011).

In sum, the cooperative system of labour relations still played an important role in accommodating large-scale workforce adjustment in Nordic manufacturing, but the extent of local trade union participation in such processes varied significantly between the countries. The differences in cooperation between the local social partners, and in trade union influence, can according to our analysis mainly be attributed to national differences in regulation of dismissal selection criteria, and in the availability of short-term work or temporary lay-offs. These differences provided variations in the need for cooperative solutions and in the actors’ control over issues of interest to their negotiating partner. The lower level of trade union participation in Denmark and Finland reflects the fact that employers enjoy much greater discretion and flexibility in dismissal selection than in Norway and Sweden. In the latter the strict seniority criteria implied that employer decisions were reliant on trade union consent, which granted the unions a stronger say. In Sweden, the absence of a temporary lay-off scheme, a significant disadvantage for employers’ ability to save company-specific skills, also posed greater pressure on the actors to negotiate mutually acceptable solutions, showing how the actors in the Nordic economies, when facing crisis and inflexible institutional mechanisms, are able to adjust by way of swift institutional changes.