When metaphors become cognitive locks: occupational pension reform in Norway

ABSTRACT

The reform of the Norwegian pension system in the early 2000s sparked off a need to reform occupational pensions. In the private sector, this was done in negotiations in 2008. In the public sector, a similar attempt in 2009 failed, and the process was only finalized in 2018. We aim to show how discourse, and the way the issues were framed through narratives and metaphors, affected the reform processes in 2008/2009. We outline the significance of different constellation of actors in the two sectors, and show how this affected the ability to achieve common understandings (coordinative discourse) and popular support (communicative discourse). We then focus in particular on one popular metaphor, namely the toiler, and show how toilers emerged in different guises in both the private and public sector negotiations. A clear notion of what the toiler needed was constructed during the private-sector negotiations. When negotiations moved to the public sector, and a different type of toiler entered the picture, key reform architects were unable to adjust. The metaphor of the toiler ended up as a ‘cognitive lock’ that hampered rather than promoted reform in the public sector.

KEYWORDS:

• Occupational pension
• policy
• narrative
• discourse
• Norway

Introduction

In the past decades, welfare state research has moved from asking ‘why were welfare states developed, and why are they so different?’ towards asking about resilience and change. Noting the remarkable resilience of welfare states, even in the face of severe economic and political challenges, important works were published with titles using metaphors such as ‘frozen landscapes’ (Esping-Andersen, 1996) and ‘immovable objects’ (Pierson, 1998). Later, the narrative of stability has also been challenged, in contributions with titles like ‘defrosting the […] welfare state’ (Palier, 2000) or ‘elephants on the move’ (Hinrichs, 2009). The rapidly expanding literature on incremental change (Mahoney & Thelen, 2010; Starke, 2006; Streeck & Thelen, 2005), as well as on welfare state adaptations in light of the financial crisis (e.g. Hemerijck, 2013) and the migration crisis (e.g. Hernes, 2018) has made clear that welfare state change, to use Hemerijck’s (2013, p. 14) phrase, is ‘work in progress’.

But what happens during reform processes? How is reform successfully secured – or not? And why do seemingly similar reform processes in practice turn out quite differently? Drawing on a unique set of in-depth interviews we here address a particular case, namely adaptations in occupational pension schemes following changes in the Norwegian state old age pension schemes. These were complex processes in several respects. They involved multiple actors: government, employers’ and employees’ organizations. Also the subject matter is complex: interests are criss-crossing and technicalities plentiful.

In one of the interviews, an informant – a long-standing central Labour party politician – identified the key to success in these terms:

If problems are not understood and shared, you will never find support for a solution. Because a solution implies change. And people do not like change. So if the problem is not understood, you will never mobilise people for a solution. (…) Any change depends on you killing off the zero-alternative: [that we can] continue as before.

This policy actor is probably quite unaware of the scholarly debates over the role of discourse, narrative and ideas in policy processes (Beland & Cox 2011; Schmidt, 2002, 2014; Stone, 2012). Nevertheless, what he did was to highlight the importance of how policy problems are presented and framed in order to reach compromises and new solutions. This is the question we pursue further in this article: What is the significance of discourse and narrative in complex reform processes? More specifically, we look at how the pension issue was framed during a lengthy process of occupational pension reform and the implications this had for negotiations and the resulting agreements.

In the lingo of old age pension scholars, we look at changes in 2^nd pillar (work-place based) schemes following changes in the 1^st pillar (state) scheme. The architects of the 2^nd pillar experienced an ‘external shock’ when the 1^st pillar scheme was reformed, and were forced to adapt somehow. This took place in two distinct phases, first in the private sector, second in the public sector. We show how discourse and the way policy problems are presented through narratives and rhetorical figures, affected the reform processes in the two sectors at different stages of the process. Furthermore, we discuss why policy narratives and rhetorical figures which enabled solutions at one stage of the process, at later stages became more problematic.

In the remainder of this article, we first present our analytical framework and link it to the overall topic of this special issue. We then outline in more detail the contents of the Norwegian pension reform, and the subsequent adaptations of occupational schemes. After a short section on methods, we present our analysis, discussing, first, the evolvement of communicative and coordinative discourse (Schmidt, 2002) in the two sectors and, second, the ambiguous use of the narrative of the toiler that emerged in different guises in both the private and public sector negotiations.

What drives policy reform?

Problems must be understood and framed in a shared manner to reach compromise and establish policy change, argued the shrewd politician quoted above. Unknowingly, he echoed key insights set out in Deborah Stone’s seminal book on the role of ideas in policy change (Stone, 2012; Blum & Kuhlmann, 2019). Stone (2012, p. 159) highlights the importance of ‘stories’ in policy making, understanding stories as ‘narratives with heroes and villains, problems and solutions, tensions and resolutions’. Common themes in these narratives are change (rise and decline) and power (control and helplessness; heroes, villains and victims). In the pension reform process, which our informant refers to, the story of change as ‘stymied progress’ was paramount. The pension system was a good one, and it had served the Norwegian population well, but the combination of an ageing population and falling oil revenues meant that the system would collapse in the future unless something was done. This message was hammered home in the early stages of the process (Ervik & Linden, 2015), and was also accepted by some, but not all, trade unions. This was a narrative of taking-to-control (Blum & Kuhlmann, 2019), of the necessity of reform to maintain the pension system for the future.

The narratives of stymied progress and responsible, forward-looking politicians as central characters within these narratives helped pave the way for the pension reform. This can be seen as one in many examples of how ideas shape policies, which is an increasingly important concern in the literature on political change (e.g. Beland & Cox 2011; Beland 2007; Schmidt, 2002). In this literature, it is emphasised that in order to understand the dynamics of policy processes it is not enough to understand the relative influence of the different parties, or how institutions are designed to allow for the influence of various stakeholders. It is necessary to understand how particular policy problems are presented in ways that makes them convincing and able to sway public opinion or forge political alliances. Policy problems, in this perspective, are shaped and contested by policy actors, not given to them (Hay, 2011). How actors speak, the ideas and frames they employ, matter in explanations of how policies change (or not). In the words of Colin Hay (2011, p. 79): ‘Interests do not exist, but constructions of interests do’.

The interest in the role of ideas in policy making stems from different epistemological roots and is expressed in different terms. In a relatively recent article, Vivien Schmidt (2014) outlines that her own concept of discursive institutionalism ‘has a lot in common with approaches such as “the ideational turn” and the “référentiel” school in comparative politics; with “constructivism” in international relations; and with a range of framing and agenda-setting approaches in policy analysis’ (Schmidt, 2014, p. 190; references in the original are omitted). Stone’s work (2012) is quoted as an example of ‘framing and agenda-setting approaches’. Discussing the precise relationship between the concepts of stories, ideas, narratives and discourses is beyond our ambitions here. Given our empirical interests, we find Schmidt’s discussion of the various forms and functions of discourse particularly useful. Discourse is defined quite simply as how policy actors speak and write to each other and to the general public about policy programmes. One of Schmidt’s key points is that this is both a matter of the ideational content of discourse and an interactive process where discourse enables policy actors to ‘coordinate the construction of their policy programme and political actors to communicate it to the general public’ (2002, p. 6). However, this can usefully be supplemented by an analytical perspective on narratives as this coordination and communication to a large extent take place precisely through the means of narrative.

In the ideational dimension Schmidt (2002) distinguishes between the normative and cognitive functions of arguments. The cognitive function elaborates the logic and necessity of a policy, while the normative function demonstrates its appropriateness by appealing to shared values. A successful discourse, in Schmidt’s analyses, tends to be one that succeeds in demonstrating both the functional necessity and normative desirability of change.

The interactive dimension of discourse has a coordinative and a communicative function. The former is to provide a common language and a framework for the construction of a policy programme. The latter relates to the public presentation and deliberation of the program. In the reform of Norwegian occupational pension, as we shall see, both were crucial at different stages of the process (Ervik & Lindén, 2015). Wage negotiations (wherein pension reforms were bargained) are multi-actor processes that take place within a limited amount of time. Proceedings are undoubtedly easier when the parties have a strong coordinative discourse between them, that is, they have a shared understanding of key problems and share the same assumptions about what wages and pensions are and ought to be. After negotiations are finalised, the negotiated result is subject to balloting among all affected trade union members. This requires a strong communicative discourse, which allows union leaders to communicate to their members that the negotiated result is favourable and just.

This perspective emphasises creative and constructive dimensions of discourse. But discourse is also constraining. Ideas do not only make particular avenues of reform possible, they make others unthinkable. The idea of framing implies that an issue can be viewed from many different perspectives and may be construed as having implications for a variety of values and concerns (Chong & Druckman, 2007). Cognitive linguists have shown how rhetorical tropes and metaphors structure our perceptions, thinking and subsequently our actions (Lakoff & Johnson, 1980). In the context of pension reform Alan Jacobs, for example, has examined how ideas and images constitute cognitive frameworks that limit which causal possibilities policy makers even consider when approaching the issue of pension reform (Jacobs, 2009). Policy actors enter complex decision-making situations with pre-existing ‘mental models’ formed by ideas which channel their attention to some pieces of available information and away from others. Highly complex policy reforms are, when they are to be negotiated, broken down into a limited set of particular problem framings, which then become the issues over which negotiations actually revolve. Stone (2012, p. 168) uses the term synecdoches to refer to ‘a figure of speech in which a whole is represented by one of its parts’, making complex problems more manageable. In complex reform processes, getting all parties to agree on what constitutes a ‘typical case’, increases odds of reaching a consensus on what needs to be done.

Stone (2012) further highlights the advantages of ambiguity: symbols such as metaphors and synecdoches, ‘can mean two (or more) things simultaneously’ (Stone, 2012, p. 178). Ambiguous symbols thus help to blur disagreements, unite broad coalitions, and allow politicians to make broad statements in public before developing details of new initiatives behind closed doors. Later in the article we develop this argument through the analysis of the synecdoche of the toiler, a rhetorical figure that was used to represent what was at stake in the pension issue, but with different effects in the private and public sector processes respectively. Our main contribution to the literature on the role of discourse in political reform processes is to highlight the limit of ambiguity. Ambiguous symbols allow actors to fill them with their own content, and in reform processes, actors can spend a lot of time developing and fine-tuning shared understandings. By the end of such processes, the ambiguous symbol is in fact far from ambiguous for insiders. When new actors arrive, however, they may adopt the symbol but reject the content. If the former insiders are not able to disengage from the implicit understandings they have helped develop and take on board the new perspectives, it will be difficult to find common ground. When this happens, the metaphor has become a cognitive lock, in Blyth’s (2001) sense. Our analysis gives an example of this.

The Norwegian institutional context

The pension reform in Norway, effective from 2011, implied four broad changes to the existing system (Pedersen, 2017):

• A strong link between life-long earnings and pensions. Pension accrual starts from the first ‘penny’ earned, and all years count equally. Pension accrual is 18,1 per cent of annual pensionable incomes up to a ceiling of 7,1 times the National Insurance’s base amount (app. 1,5 times average full-time income). Prior to the reform, full pension accrual was reached after 40 years and benefits were calculated based on the best 20 income years.

• Retirement age is fully flexible between 62–75 years on actuarially neutral terms, leaving behind the former fixed pensioning age at 67 years. Rules for combining work and pension for individuals older than 62 have become far more flexible, to encourage longer working careers.

• Pensions are reduced in proportion to observed increased in longevity, hence transferring the risk of changes in longevity from the state to the pensioners.

• Pension payments are indexed with the development of wages minus 0,75 percentage points, removing the previous wage adjustment guarantee on running pensions.

The pre-2011 system aimed at securing a given proportion of life-time earnings in old age. The new system has no such ambitions, but provides each with an individual notional ‘pension savings account’ (NDC-system). When a person retires, her/his annual pension is determined by the size of this account (pension wealth) and the number of years expected to live past retirement (the longevity adjustment). The replacement rates that can realistically be achieved in the new system are modest, even after relatively long working careers. Full time workers will in most cases have less than 50 per cent of lifetime incomes from the state system. This has opened up an increased room for occupational pension (Grødem & Hippe, 2018).

Public-sector employees in Norway are covered by an occupational pension scheme that historically predates the National Insurance. This scheme provided top-up pensions from the National Insurance, so that the individual would be guaranteed 66 per cent of her final wage upon retirement (for incomes up to a ceiling), provided she had 30 years of accrual in the public sector.

In the private sector, coverage of occupational pensions were scattered until 2006, when it was made mandatory for employers to establish occupational pension schemes. A recent change in legislation had opened up for tax deductions for a wider menu of occupational pensions, which prompted many employers to transform their established defined benefit (DB) schemes to defined contribution (DC) schemes (Pedersen, Hippe, Grødem, & Sørensen, 2018). While DB-schemes are coordinated with National Insurance pensions to guarantee a certain benefit level (typically around 66 per cent), DC-schemes are typically mandatory savings schemes with no risk pooling and no coordination with other benefits. DC-schemes are favoured by many employers because they make it easier to predict future costs. They also fitted the new pension system, because they were not coordinated with the National Insurance pension and thus did not interfere with the new work incentives. The legislation that made occupational pensions mandatory was a response to demands from the social partners, yet most observers see it as part of the pension reform, at least in the political sense: by giving trade unions this victory, politicians may have hoped to secure their support for the larger pension reform.

The negotiated early retirement scheme, AFP, covered about 60 per cent of all employees, and was expanded successively in the 1990s (NOU 2004, p. 1: 312). In 1998, it lowered the retirement age to 62 for those covered. State subsidies were provided to cover the costs for those who retired at 65 (public sector) or 64 (private sector), but retirement earlier than this had to be paid for by the employers. These provisions were made in negotiations between the social partners, which explain the different outcomes in the two sectors.

AFP in both sectors, and the DB-nature of the occupational pension in the public sector, were essentially incompatible with the incentives structure in the new pension system. The new principle of flexibility on actuarially neutral terms could not coexist with negotiated arrangements that guaranteed retirees a certain benefit level. The schemes had to be adapted, and – as the schemes were ‘owned’ by the social partners – these adaptations had to be made through tripartite negotiations.

The 2008 settlement in the private sector negotiations transformed the AFP-scheme from an early retirement scheme to a top-up pension, available from age 62, paid on actuarially neutral terms on top of the pension awarded from the National Insurance. In this way, it complemented the main principles of the reform instead of undermining them. This was a displacement (Mahoney & Thelen, 2010) of the old AFP, introducing a fundamentally different scheme, albeit with the same name. The full effects will only be felt after a long phasing-in period, as considerable state subsidising over a long period was put in place following the settlement between the state and the social partners (Grødem & Hippe, 2018).

The architects behind the pension reform pushed for a similar solution in the public sector, but this was met with resistance from the trade unions. This was partly anchored in a phrase that was written into a parliamentary compromise early in the reform process (2005), which stated that public sector occupational pensions were not to be touched. Also, some argued that because public sector employees gave up part of their wages to fund the occupational pensions, the scheme was protected by the constitutional ban on introducing laws retroactively. Nevertheless, the trade unions accepted that occupational pensions should be reduced in proportion to increased life expectancy, and they also accepted the indexing rules – in other words, they accepted the two cost-reducing mechanisms. This, in their view, was compatible with the promise from 2005. They however refused to give up the guaranteed level at 66 per cent of the final wage at the point of retirement, and they also managed to keep the public sector AFP-scheme intact as a subsidised early retirement option between the ages of 62 and 67.

Between 2008 and 2018, then, employees in the public and private sectors related to two separate pensions systems. This was unsustainable, for various reasons. First, the new system granted private-sector employees choices and a flexibility that was not extended to public-sector employees. Public sector employees could not combine earnings and pensions, and they could not increase their pensions by working longer. Second, the longevity adjustment placed a ‘time bomb’ under the public sector scheme (Pedersen, 2018). This principle ‘ate away’ the guaranteed level, and implied that the public sector scheme was a DB-scheme only by name. Over time, this would lead to intolerably low pension for future cohorts. This insight brought the parties back to the negotiating table in 2018, and a solution was found when the social partners accepted to transform both AFP and the occupational pension schemes to NDC-schemes to be drawn between ages 62 and 75 on actuarially neutral terms.

This return to the negotiation table can, in Streeck and Thelen’s terms (2005), be interpreted as an instance of policy drift: the public sector pensions were not adequately adapted to changing circumstances. The question is, then, why did private sector unions agree to adapt their pension straight away, while public sector unions resisted for a decade?

Below, we examine the role of discourse and narrative in the two reform processes, with an emphasis on the constellation of actors and the use of ambiguous metaphors. We suggest that the lack of a coordinating discourse and a unified pension narrative in the public sector is an important part of the explanation. Furthermore, the ambiguities inherent in the narrative of ‘the toiler’ which proved so useful in the private sector negotiations, served to unmask diverging interest when transferred to the public sector process. By choosing this approach, we are not denying that other factors mattered. Arguably, more was at stake in the public sector: life-long occupational pensions were on the table, while the private sector ‘only’ needed to settle the issue of early retirement. This was however not a trivial issue for private sector unions: indeed, other contributions have presented the 2008 negotiations as the biggest miracle of the pension reform (Grødem & Hippe, 2018). A proposal to reform the private sector AFP in 1998 led to loud protests from the trade unions, and a deadlock where trade unions refused to even discuss pension reform. None of the key actors in the reform process were confident that the 2008 negotiations would produce a solution. Stakes were high in both sectors, although there were more money on the table in the public sector negotiations.

Data and methods

Our analysis is based on qualitative interviews with informants who were central in the pension reform process. The interviews were carried out by one of the authors and a colleague. Among our informants are key bureaucrats in the Ministry of Labour and Social Affairs, LO (The Norwegian Confederation of Trade Unions) officials and staff (including former leaders Valla and Flåthen), key politicians (including former Prime Ministers Stoltenberg and Bondevik, former Minister of Labour Hansen, former Ministers of Finance Johnsen and Schjødt-Pedersen), as well as central players in unions outside LO, and in the Confederation of Norwegian Enterprise (NHO). We also talked to a few independent experts. All in all, we carried out 15 interviews with 18 informants between November 2015 and September 2016. One interview in the Ministry included two civil servants. In the interview at NHO, three persons participated. Only one of the respondents asked to participate did not respond to the invitation. The interviews were recorded and fully transcribed.

We made it a key aim of the interviews to pinpoint the process in as much detail as possible: who talked to whom when, under what circumstances and institutional conditions, promoting which interests etc. By the end of the data collection, therefore, the interviewers were confident to have obtained a good overview of both meeting places, actors, and the interests the actors sought to promote in different settings. The interview data are complemented by a strategic selections of newspaper articles and documents from the involved parties that in particular were used to inform us about their communicative discourse. As interview data were less rich on the communicative aspects of the public sector process, we searched a media data base (Atekst) for relevant articles in four main newspapers (Aftenposten, Dagbladet, Dagens Næringsliv, VG) in 2009 using search terms such as pension reform, occupational pension and early retirement schemes (AFP). We also browsed the unions’ websites for information material addressed to members, where particularly information found at UNIOs website were used to supplement interview data.

Constructing an efficient reform narrative

Policy discourse, as noted, performs two different interactive functions (Schmidt, 2002). It serves a coordinative function in providing a language through which agreement on a policy programme can be reached, and a communicative function in providing a means of persuasion whereby public legitimacy can be created. Typically, each will be important at different stages of the policy process. First policy actors construct a shared policy programme through coordinating discourse, and then political actors communicate it to the public (Ervik & Lindén, 2015).

Coordination

Wage negotiations are formal processes between a limited number of actors, to be undertaken within a limited period of time. The actors represented discrepant interests, and went into the negotiations with different purposes. To reach agreements they depended on reaching a common vision of the purpose of reform, by means of developing cognitive arguments about the necessity of changes and normative arguments about their appropriateness. In short, there was a need for a coordinating discourse that could construct a space for maintaining key principles along with fundamental reorganization of the schemes.

We find that there were significant differences between the private and public sector in terms of providing an institutional context that was supportive of generating such a shared coordinating discourse. Two aspects were crucial: the numbers of actors involved, and the extent to which the respective actors had been involved in developing a coordinative discourse over time.

In the private sector, LO had a dominant position. In Norway, parties in the export-oriented industry always negotiate first, and their settlement provides a template for negotiations in other sectors (Andersen, Dølvik, & Ibsen, 2014). Thus there were relatively few actors involved: LO, NHO (the main confederation of employers) and the Labour-led government. According to our interviews, NHO took little interest in pensions in the negotiations. Its main interest was to avoid increasing employers’ pension expenses. This enabled an informal coordinative discourse that took place outside of institutionalised negotiations.

This was not the first time LO and the Ministry of Labour and Social Affairs met to discuss pensions. High-ranking LO officials had been actively involved in outlining the details of the NI pension reform (Grødem & Hippe, 2018). LO historically has close ties with the Labour party, and in the pension reform process, this connection was activated. Some informants were surprisingly frank with respect to the nature of this process. Bilateral talks between selected Labour Party ministers, central civil service actors and the LO leadership went on through informal channels, and what was presented as a result of formal negotiations had in reality been negotiated informally beforehand at top level between a few government members and the LO leadership.

So, in the course of that process we developed what became the end result. That was one of those behind the scenes-processes. And it ended with LO and NHO presenting a deal which in reality was a slightly rewritten version of what we had developed on beforehand. […] That process with LO … that was actually quite constructive, I’ll say. (Civil servant, Ministry of Labour and Social Affairs)

Fewer actors and the involvement of top-level actors also meant that the private sector occupational pension negotiations in many ways continued an already existing coordinating discourse, established during the NI pension reform. The actors involved were not new to pension, but had been strongly engaged in the issue over time. Long time engagement breeds both understanding of the key principles of the reform and commitment to seeing them fully implemented. A coordinating discourse and shared problem definition was already established. The need to achieve a model based on actuarial neutrality was agreed upon.

The public sector setting for coordinative discourse was a different one. Not one, but four main trade union associations sat at the negotiating table (LO, YS, Akademikerne and UNIO). Positions diverged between and within the unions. Some (including large LO organizations, such as Kommuneforbundet, which organizes municipal workers) represented low-paid public sector workers who in reality gained little from the occupational pension schemes when these were coordinated with the National Insurance pension. Others, such as UNIO, represented large professional groups who often had shorter working careers (due to more years in education, and also because many teachers and nurses e.g. were women) and for whom the occupational pension scheme constituted a significant part of their total pension income. Changing the pension system for members of one union but not the others was never an option: the public sector occupational pensions have always been unified, and the unions adopted a ‘one for all and all for one’-strategy from the outset. The involvement of several organisations with diverging agendas however allowed for new coalitions and power plays. The informants outlined how Kommuneforbundet supported a reformed system, while UNIO was staunchly against. NTL (another LO branch, organizing state workers) hesitated, but eventually sided with UNIO. In that situation, the representatives from Kommuneforbundet bailed, not wanting to go alone for a controversial result that the other unions rejected.

In the public sector negotiations, it is clear that a coordinating discourse was missing: not only between the unions and the ministry, but also between the unions themselves. There was no consensus between the unions on what the nature of the problem was. For Kommuneforbundet, the problem was that the occupational pension offered little to their low-paid members; for UNIO the reform represented a threat to the guaranteed replacement rate. The civil servants, on their hand, expressed frustration that after almost a decade of intense work with the pension reform, they had to start nearly from scratch with new actors who neither shared their problem definition nor their visions for the future pensions:

I mean, the first meeting we had about public sector occupational pension some of the organisations came in saying they did not understand why they were here! [they claimed] we had the [2005 parliamentary settlement], which said that public sector occupational pensions shall be continued – so what are we doing here – and there was this protection in the Constitution and … we can’t do anything, can we, to the public sector pensions (civil servant, Ministry of Labour and Social Affairs)

Following preparatory work groups, the main agreement was scheduled to be reached through a relatively short period of collective agreement negotiations, first between the parties, subsequently with the aid of the State Mediator. The parties thus had a short deadline for establishing a common ground for coordination. Levels of stress and sleep deprivation ran high as the process went on. Importantly, UNIOs leaders had not been included in the informal talks that had preceded the NI pension reform, nor had they been part of the pension commission. They also did not have the same type of close personal relationships that existed between the leaderships in LO and the Labour Party. There was, in short, not the same kind of continuity from previous processes to build on, as in the private sector. ‘We did not see the realities of the world looking a bit different. We thought we could repeat the 2008-AFP’, said one civil servant informant. They could not.

Communication

As noted above, the results of tripartite negotiations are balloted among all union members. Negotiators not only had to achieve agreement at the top-level, they had to be confident that they would subsequently be able to communicate the agreed-upon models to their members in a convincing manner. This required an efficient communicative discourse.

The ballot was a genuine hurdle, also in the private sector negotiations. Our informants from LO and the Labour party told us that they were by no means confident that the negotiated result – with the new AFP – would survive the ballots (Grødem & Hippe, 2018). But LO had a plan: after the negotiations were finalised, a massive information campaign was initiated. The message from LO was that the state had been exceptionally generous in the negotiations, and that private sector workers would be given the opportunity to draw pension while working as much as they pleased. The campaign worked, as the negotiated result passed the ballot by a wide margin.

In the negotiations in the public sector, two alternative models were on the table. The government promoted ‘påslagsmodellen’, the top-up model, in which occupational pensions would be paid on top of the state pension, with no coordination or guaranteed levels. The white-collar union UNIO championed ‘bruttomodellen’, the aggregate model, in which state pensions and occupational pensions combined should guarantee a given replacement level. UNIO’s claim was epitomized in the slogan ‘66–65-30ʹ, which proved to constitute a very effective communicative discourse. The figures refer to the established parameters of the DB public occupational pension scheme: 66 per cent of final salary, when retiring at 65, after 30 years of accrual. It was actively used: ‘They had that on buttons […] they had it on their lavatory walls, glued up there, they sent it out to all their troops … ’ a former Labour executive marvelled in his interview. This slogan proved extremely efficient, probably for two reasons: two-thirds of final wages is a good pension, and it is a principle that is easy to understand. ‘66–65-30ʹ made the complexity of pension comprehensible – you work 30 years, you retire at 65, you get 66 per cent of your wages. ‘It was impossible to resist. We could not respond with påslagsmodell’, says one of our civil service informants. Maintaining the status quo of what was generally regarded as a good pension scheme with defined benefits thus became the raison d’être of the UNIO campaign. Meanwhile, LO was marred by internal division and kept a low profile; it was not necessary for LO to rally its members to battle for the status quo. UNIOs leaders gained confidence from a unified rank-and-file equipped with ‘66–65-30ʹ-buttons, and against this effective communicative discursive measure, the Ministry was effectively checkmate.

Constructing a narrative: a fair pension for the toilers

So far, we have looked at how the (institutional) conditions for creating a shared coordinating discourse differed between the private and public sector negotiations. We also saw how the trade unions created a communicative discourse vis-à-vis their members which facilitated support for the compromise negotiated in the private sector and against a similar solution in the public sector. In this section we will move a step down from the institutional context of discourse and look in more detail at some of the narratives that were at work in the discourses.

Pension policies are extremely complex, and non-experts find them hard to understand. Actors need to consider incentive structures as well as distributional effects of new schemes, macroeconomic aspects vis-a-vis social justice, risk-sharing mechanisms between those still working and those currently retired, the distribution of burdens and advantages between different generations, and a long list of other normative and technical issues. In order to cut this complexity down to a size where it could be dealt with in negotiations, the issue had to be simplified. This simplification happened through the construction of relatively coherent narratives, where a particular symbol played a key role: the toiler. This is an example of processes pointed out by scholars from different intellectual traditions (Lakoff & Johnson, 1980; Stone, 2012): language, metaphors, examples and narratives that are applied influence how issues are received, thought about and acted upon. In other words, policy actors’ use of narratives, including metaphors, work to frame a policy issue in a particular way, focusing some aspects of the matters at stake at the cost of others. Some informants also reflect upon this, underlining the strategic effects of such framing:

With pension you can nearly always define the things in your way, because it is so decisive which perspective you have. Because you cannot describe … decisions or processes in so many dimensions at the same time, you know, you have to choose. […] So you could in a way define a little, right, so when we were to sell this afterwards no one thought it was unfortunate that there was a bit of vagueness with respect to whether those were accumulated [figures] or annual effects, right. (trade unions representative).

This informant clearly saw this ability to manoeuvre information as an advantage, particularly in communicative settings (‘when we were to sell this afterwards … ’). The ability to ‘define a little’ however also makes it difficult to arrive on a unified understanding, if parties come to negotiations tuned in on quite different aspects of the reform complex.

Anyone who champions social policy reforms must present a narrative that constructs the target population in a way that allows them to sell the reform to the public, e.g. by constructing groups as deserving or undeserving. This is rarely a problem for those who wish to expand old age pensions: old age pensioners are typically seen as the most worthy claimants of all (van Oorschot, 2006). But when flexible retirement became a main topic, the issue of deservingness became contested. At what age should people be exempt from the obligation to work? To what extent is a long retirement period a social right, and to what extent is it a luxury individuals should pay for themselves? Early retirees may be construed either as deserving tired workers, or as undeserving privileged elders who should instead contribute in the labour market (Blum & Kuhlmann, 2019). In order to defend the ability to retire at 62, at least for the older cohorts, the unions’ negotiators in both sectors called up this key rhetorical figure: the toilers, sliter in Norwegian. Sliter is a noun constructed from the verb slite, which means to work hard. A toiler is someone who has worked hard, but been poorly or moderately rewarded for this through his or her working life. The toiler has thus earned a moral right to fair treatment in old age – it is the epitome of the deserving claimant. The image of the ‘toiler’ can be characterized as a synecdoche, in Stone’s (2012, pp. 158–59) language: ‘figures of speech in which a part is used to represent the whole […] [Important] because we often make policies based on examples believed to be representative of a larger universe’. By letting the toiler represent the complexity of the pension issue, pension debates are contextualised in a normative universe of justice and deservingness.

This hard working but exhausted character figured prominently in both the private and public sector pension negotiations. However, there were different kinds of toilers being presented in the two contexts which allowed for different claims and outcomes. This illustrates how frames, narratives and synecdoches developed in one discourse, can acquire a different type of effect when transferred to a new institutional context. The ‘toiler’ trope was sufficiently ambiguous to travel between contexts (cf. Stone, 2012, p. 178). It allowed the parties to talk about different things in different contexts while referring to the same source of moral authority (who would deny a hard-working toiler the right to a decent pension?). In discourse theory, this sign (‘the toiler’) might be termed a floating signifier – a signifier that is overflowed with meaning because it is articulated differently within different discourses (Torfing, 1999, p. 301).

Private sector LO represents male-dominated industries, and its concept of the toiler was a male industrial worker who had worked since an early age, at least 40 years (meaning full accrual in the NI system). He had worked hard and long, and in his early 60s, he was exhausted. Securing this hard-working and exhausted man a decent pension, even when he had no choice but to retire at 62, became the crucial parameter in the private sector negotiations. In the words of two informants:

What LO keeps saying is … we have this standard industrial worker who has worked for 40 years by the time he is 62, and then … we have to pursue an AFP in his first year of pensioning that is as good as the one we have (former minister, Labour).

And eventually it’s all about that. How are we to secure that they who retire at 62 are not worse off than what they would be in the old system (civil servant).

The toilers in the private sector were eventually secured through an arguably very generous transition arrangement, which effectively postponed the effects of the longevity adjustment and allowed early retirement on advantageous terms for the older cohorts. LO thus accepted the key principle of the pension reform, that later retirement should be rewarded by higher pensions, but sheltered the toilers in the older cohorts from the full implications of this principle. In this sense, the retrenchment reform also had expansionary elements. This made the narrative of (some) recipients’ deservingness crucial.

Also in the public sector process that followed the year after, a hardworking and exhausted figure played a major legitimizing part in the unions’ discourse. But as different unions sat at the table, so did a different kind of toiler. In the public sector, female workers dominate. Many of them are professionals with moderate salaries, such as teachers and nurses. It is not only physical work, but also ‘large demands to ongoing involvement and personal engagement in other people’s lives, health, upbringing, aging and living conditions’ that leads to these workers’ ‘real and legitimate need for early retirement’ (UNIO, 2011). This standard worker has not worked 40 years. She has entered the labour market later due to education, she has worked part-time juggling work and family, she cannot meet the NI-requirement of 40 years of accrual to qualify for full pension. The public sector occupational pension has compensated for shorter time in employment as it requires only 30 years of accrual and calculates a 66% replacement level based on final salary. The 66–65-30-slogan was therefore also a shorthand for the rights of the public-sector toilers.

While the properties of the toiler differed between the private and public sector, the discursive emphasis on their situation created a precise locus for struggle in both sectors. It was the needs and deservingness of specific types of toilers that were addressed during negotiations. The needs of future pensioners (and other kinds of workers) were simultaneously sidelined. ‘All the emphasis was on early retirement and the first cohorts. [We] didn’t think about what this would be like in ten or twenty years’, says a civil servant informant. In both sectors an effect of placing the emphasis on toilers about to retire, was that the situation of younger cohorts was obscured. The short-term effects on the pension of cohorts that were close to retirement age were more visible than the long-term consequences for younger cohorts.

Conclusion

In this paper, we have used Schmidt’s (2002) analytical framework on the role of discourse and Stone’s (2012) concept of narratives to illuminate the outcomes of the negotiations between the social partners and the state in the wake of the NI pension reform.

We have considered both the ideational and the interactive aspects of discourse: how problems are understood, and how these understandings are communicated among the actors and to the wider public. The institutional set-up of negotiations in Norway demands effective coordinative as well as communicative discourse: these are processes with many veto players. In principle, any union or employer’s association can say no. This calls for strong coordinative discourses. At the same time, the negotiated result is subject to balloting among the rank-and-file, which creates a need for effective communicative discourse. Negotiators must be able to explain what they have achieved (or given up) and why the negotiated result deserves their members’ support.

In the private sector, LO was the dominant non-state actor. It benefited from deep historical traditions of collaboration, as well as long-term involvement with the NI pension reform to which they had developed a shared commitment with the Labour Party government (cf. Grødem & Hippe, 2018). The framework of an efficient coordinative discourse was thus already in place when the negotiations started. In the public sector, more actors were involved, many with limited experience from and ownership to the NI pension reform. The time frame for establishing a coordinating discourse was shorter, at the same time as the ideational distance between the parties was larger in the first place.

Yet there is more to this story than whether there was a coordinative discourse or not. We suggest that the strong coordinative discourse in the private sector negotiations became an obstacle in the public sector negotiations. This is because discourse creates a problem definition and a shared sense of which aspects of a situation are relevant and which are not. This appears from our interviews to have been a real challenge for the advisors in the Ministry of Labour and Social Affairs: LO’s top executives were involved in discussions of the new National Insurance pensions, and these executives’ understandings were carried over into the negotiations on occupational pension in the private sector, which were led by male-dominated, blue-collar LO-affiliated unions. Civil servants adopted a narrative, a way of framing the issues at stake, which had evolved to fit these unions’ concerns and interests. When other unions with other interests and a different membership entered the scene, the civil servants were unable – in the short term – to adjust. The established narrative did not account for part-time work, punctuated working careers, and working careers shortened by long education.

This story also illustrates some potential side-effects of using effective, yet ambiguous synecdoches (typical examples). We have showed how the rhetorical figure of the toiler served a key function in both processes. The toiler highlighted the inherent deservingness of old age pensioners, by emphasizing the many years of hard work and the understandable exhaustion of typical pensioners. Exhausted 62-year olds were worthy claimants, and the ambiguity of the symbolic figure made the rhetoric available to practically all trade unions. These weary workers who were exhausted at 62 were present both in the public and private sector negotiations, but in very different guises. In the private sector, he was a manual worker who had started working at 16; in the public sector, she was a well-educated teacher or health worker with family obligations in addition to a demanding job. Giving the toilers the dignified retirement they deserved thus required different practical and technical solutions, a fact that was obscured by the similarities in rhetoric.

We have shown, then, both the potentials and limitations in appealing to value-laden, and ambiguous, metaphors and synecdoches in complicated reform processes. Ambiguity obscures differences, which is often desirable to reach agreement through coordinative discourses or amass popular support through communicative discourse. In complex reform processes, where multiple partners are involved over time and the issue at hand is highly technical and complex, this ambiguity can also constitute less flexibility to deal with the complexity of interests and concerns involved. What happened in our case was that the metaphor of the toiler created a cognitive lock (Blyth, 2001). An image of the toiler, and a clear notion of what the toiler wanted and needed, was constructed during the private-sector negotiations – down to the last technical detail. When negotiations moved to the public sector, and a different type of toiler entered the picture, key reform architects were unable to adjust. The story thus shows how narratives can promote as well as block reform. Synecdoches simplify issues but can also highjack attention, and ambiguity can oil the wheels as well as put a spanner in the works in complicated reform negotiations.

Additional information

Funding

This work was supported by the Norges Forskningsråd (NO) [238205/H2].

Notes on contributors

Anniken Hagelund

Anniken Hagelund is professor of sociology at the University of Oslo in Norway. She has published widely on the role of discourse in policy processes within social policy and migration policy. Currently she is working on a project on media discourses and migration.

Anne Skevik Grødem

Anne Skevik Grødem is a sociologist and research professor at the Institute for Social Research, Norway. Her main research interests include family policies, comparative welfare studies, poverty and marginalization, and migration as a challenge for national welfare policies. She is currently working on a project on the Norwegian old age pension reform.