Corporatism,crisis and contention in Sweden and Korea during the 1990s

1. Introduction

As a term ‘corporatism’ remains elusive, despite the common reference to Schmitter’s (1979: 93–4) formal definition. The term describes public institutions and state–society relations in both authoritarian and democratic regimes in both developing and developed states (Adams, 2002; Cawson, 1986; Chalmers, 2001; Wiarda, 1997; Williamson, 1985). Corporatism’s depiction of state and society contrasts with ideas about a plurality of interest groups governed by the state as neutral referee. Corporatism insists that the state actively intervenes to recognize, license or create particular groups that exercise an organized capacity to monopolize representation of given interests, so long as they select moderate leaders who articulate the group members’ demands.

The elusiveness of corporatism comes in part from the diversity of its ideological connotations. Liberals frown on the absence of free choice, autonomy and efficient market forces when a state incorporates business associations in order to perform some of its regulatory functions. Marxists frown on the reduced scope for revolutionary mobilization when a state incorporates union leaders, union bodies and processes of collective bargaining in order to prevent strikes, restrain wage increases and promote employment. By contrast, Catholic social commentators, military leaders, social-democratic party and union leaders, and international unionists promote corporatism (or tripartite social partnership) as an alternative to both market chaos and class conflict. For them, corporatism aims to promote harmony, achieve social stability and form public policies that develop a national economy with low inflation and high employment.

Besides ideological connotations, corporatism is also elusive because scholars use it to describe established institutions or regime types in Europe, Latin America and East Asia rather than as part of either a carefully specified normative theory about the good society or as part of an explanation about why societies, states or regimes change. In short, ‘corporatism’ slips between ideology, description, norm and explanation.

We argue that formal definitions, ideologically oriented interpretations and regime typologies of corporatism all fail to explain why union or NGO influence over the formulation of economic policy can arise, change, end or transform, either in general or in particular places and certain times. This shortcoming is particularly clear in contrast against the wide-ranging ambitions of recent theorizing about the common causal mechanisms in various forms of non-routine or ‘contentious’ politics, including peasant rebellions, revolutions, strikes, democratization, coups, ethnic conflict, civil war and terrorism (McAdam et al., 2001, 2009). The first rationale for undertaking our comparison between Korea and Sweden’s dramatic convulsions during financial crises in the 1990s is, therefore, to demonstrate this shortcoming in notions of corporatism as we analyse how, when and why our two regimes shifted.

A more obvious rationale for focusing on Sweden and Korea is that they sit at opposite ends of a spectrum variously described as stretching between state and societal corporatism (Schmitter, 1979), state-licensed and contract corporatism (Wiarda, 1997; Williamson, 1985), or authoritarian and democratic corporatism (Choi, 1989; Hundt, 2005; H Kim, 2007; Rothstein, 1991, 1992, 2002). However it is described, the regimes at each end of this spectrum faced extraordinarily severe financial crises at much the same time. What caused authoritarian Korea and democratic Sweden to endorse neoliberal views about state–society relations? What can explain this convergence from either end of the spectrum of corporatist arrangements towards a liberal centre? We question the argument that economic globalization has rendered states more dependent and beholden to the neoliberal demands of transnational capitalism, and instead advance arguments about how, when and why Korea and Sweden each renewed their traditions of economic policy formation with state bureaucracies and governments that actively contribute to the processes of economic development.

This article has three parts. Section 2 reflects further on corporatism as it identifies similarities and differences between Korea and Sweden. Using statistics from the Organization for Economic Co-operation and Development (OECD), the International Monetary Fund (IMF) and the International Labour Organization (ILO), we show how four states fared between 1980 and 2000 in terms of economic growth, the public sector, unemployment, employment protection, productivity, inflation, densities of unionization and industrial conflict. These measures offer some objective context for the narratives in Sections 3 and 4 on Sweden and Korea, which analyse episodes of regime change during the financial crises. We deploy aspects of the dynamics of contention approach to comparative politics (McAdam et al., 2009) and look more closely at interactions among insider and outsider actors, including governments, sites of contention and repertoires of contention.

Our analyses focus on three causal mechanisms that Tilly argues are crucial during democratization: changes in trust networks, cross-class alliances and the brokerage of new connections between actors and sites of contention (McAdam et al., 2001: 264–304; Tilly, 2001: 24–5, 36–7). Critics of the dynamics of contention argue that causal mechanisms are numerous and varied (Koopmans, 2003) but since global neoliberalism is a potential cause of de-democratization (Tilly, 1995a, 1995b), our three causes of contention during democratization offer plausible explanations for how, when and why our two regimes changed in the 1990s.

Given the importance of the state for corporatism and the active roles that our two states played during acute financial crises, we also focus on the possibility that outsiders’ contention is closely related to insiders’ contentious politics within established institutions (McAdam et al., 2001: 7). Strikes tend to be short, frequent and contentious in Korea, and large, infrequent and routine in Sweden, but there were few strikes in either state during the financial crises. Instead, unions in both states tested the boundaries between contentious and routine politics.

2. Initial comparisons

All forms of corporatism pursue economic development with various forms of state–society relations that contradict, rather than merely fall short of, the ideals of a perfect market. Good economic performance is the crucial test for any form of economic policy formation and implementation, but it is particularly important for corporatist institutions because neoliberalism has instilled considerable faith in market forces among leaders of transnational corporations and among bureaucrats and party leaders in the rich world. So, how well did Sweden and Korea perform before and after their financial crises of the 1990s? We first compare these two corporatist states with two market-oriented states. The point in comparing Sweden and Korea against both the UK and the US, is that Sweden has tended to converge on a less anti-union liberal centre in the UK, while Korea has tended to converge on a more anti-union liberal centre in the US. We then compare democratic-corporatist Sweden with authoritarian-corporatist Korea.

Figure 1 illustrates that the Korean economy grew at 6–12 percent per annum between 1980 and 2000, except for short, sharp recessions in 1980 (after a coup and major unrest) and 1998 (after the financial crisis). Britain and the US experienced negative growth during 1980–2 and again in 1991. Sweden mostly followed the British and American pattern, with growth at 2–3 percent during upturns and below 2 percent during downturns. However, the Swedish recession in 1991–3 was worse and lasted longer than it did in Britain and the US.

OPEN IN VIEWER

Figure 1.

Economic growth

Figure 2 on the size of total outlays of government as a proportion of GDP illustrates how Sweden and Korea have long occupied the opposite ends of a spectrum on corporatism, while the US and the UK occupy a middle ground. It is important to note that the size of the Swedish welfare state is a relatively recent development. Since 1870, total outlays of government in the industrially advanced states have expanded from around 10 percent of GDP to around 40 percent. As a proportion of GDP, the Swedish state was smaller than the US or British state before the Second World War. All three were about 30 percent of GDP in 1960, but thereafter the Swedish state has become much larger. During the financial crisis of 1990–2, it reached 71 percent and has since then declined to levels closer to Britain and the US. Statistical data from the OECD indicates that Korea’s small public sector has increased to levels closer to Britain, the US and Sweden since 1984. The difference between Sweden and Korea was 50.3 percentage points in 1992; that difference had halved to 24.2 percentage points by 2010.

OPEN IN VIEWER

Figure 2.

Total outlays by government

Figures 3 and 4 illustrate the neoliberal convergence thesis in terms of rates of unemployment and employment protection. Sweden maintained full employment for 50 years from 1941 to 1991. This was partly a matter of social-democratic ideology, pursuing genuine freedom for ordinary workers (Meidner, 1980), but more importantly it was a matter of long-term institution building (Apple, 1980; Dow et al., 1984; Higgins, 1985b) and a political culture of independence from the international edicts of neoliberalism (Therborn, 1986). Korea also maintained full employment by means of long-established institutions (life-time employment in large corporations) that sustained a political culture of independence from neoliberalism.

OPEN IN VIEWER

Figure 3.

Unemployment

OPEN IN VIEWER

Figure 4.

Employment protection

During the 1980s, Sweden and Korea each maintained unemployment levels below 4 percent. First Sweden (1992) and then Korea (1998) abandoned their maintenance of low and stable unemployment and converged on the neoliberal mainstream, as represented by the cyclical rise and fall of unemployment in the US. Indicators of the strictness of employment protection show a less dramatic convergence between our corporatist and liberal regimes.

These figures come from the IMF and the OECD, but liberals challenge the reliability of unemployment statistics in both Sweden and Korea. It is often suggested that when retrenched workers take the Swedish Labour Market Board’s courses in sought-after skills they count as students and thus disguise the real extent of unemployment. Similarly, in Korea strong family values and deep shame over unemployment disguise the extent of actual unemployment. A liberal observer might therefore argue that the levels of unemployment in Sweden and Korea during the 1980s are not entirely comparable with levels of unemployment in states where individuals are more often left to their own devices, but after the financial crises the unemployment levels are more comparable. The differences in unemployment and employment protection in our corporatist and liberal states before the 1990s may actually be less than they appear, but we argue that the timing of changes in employment levels and employment protection should be accepted as indicators of ‘regime change’.

Figures 5 and 6 also illustrate the convergence thesis in terms of inflation. The productivity of labour and rates of inflation became less important differentiators between our corporatist and liberal states and between Sweden and Korea.

OPEN IN VIEWER

Figure 5.

Labour productivity per unit of labour input

OPEN IN VIEWER

Figure 6.

Inflation

Figure 7 contradicts the corporatism–liberalism convergence thesis and illustrates a large difference between Korea and Sweden, which has long had the world’s highest density of unionization. Among OECD states, the US and Korea have always had a low union density, while Britain’s was medium-high until 1980 and fell strongly between 1980 and 2000. Korean densities increased in the years immediately after the first democratic elections in 1987 only to decline again thereafter. On strikes and lockouts, our four states mostly converged on industrial peace in the 1990s.

OPEN IN VIEWER

Figure 7.

Densities of unionization

Figure 8 draws on ILO statistics about strikes and lockouts, excluding disputes shorter than one day and fewer than 10 workers in Korea, Sweden and the UK, fewer than 100 workers in the USA, and all political strikes in the UK. Figure 8 on the other hand illustrates the convergence thesis. After widespread industrial peace in the 1950s and 1960s, and an unexpected surge in unofficial strikes in the 1970s, Sweden’s most extensive industrial dispute in 1980 was something like an earthquake that had aftershocks in 1984, 1988, 1990 and 1995 (Törnqvist, 2007).

OPEN IN VIEWER

Figure 8.

Number of workers in longer, larger industrial disputes per 1000 workers in the total workforce

The British strikes of the 1970 and 1980s largely ceased in the 1990s (Lyddon, 2007). The pattern of industrial conflict in the US resembles those of Britain and Sweden. The decision to exclude smaller stoppages from the strike statistics after 1981 tempers the observation that fewer stoppages took place in the 1980s but it is fairly clear that as elsewhere many fewer stoppages have taken place since 1990 (McCartin, 2007).

The Korean pattern differs from the others because its authoritarian corporatist labour market instituted ‘yellow’ or compliant unions that prevented stoppages. It was not until the democratization movement began to make headway during the 1980s that independent unions could form and organize extensive stoppages in 1987–90. Since then stoppages have dropped away much as they have elsewhere (B Lee, 2007).

These figures describe an uneven convergence on a liberal centre. Workers in both Sweden and Korea continue to benefit from employment protection measures against unfair dismissal, for high health and safety standards, and effective rights to organize collective bargaining, which remain clearly better than employment protection rights in Britain and the US. Britain’s density of unionization has fallen close to the low levels in Korea and the US, but Sweden’s density of unionization remains high. Data on the outlays of government illustrate a convergence between a large public sector declining during the 1990s in Sweden and a small Korean public sector that continued to increase as it became more prosperous. Industrial conflict remained important in Sweden but fell to low levels in Britain, the US and Korea.

These statistics describe general differences between the states in our study but tell us nothing about why or how growth faltered, inflation rose and full employment became a lesser priority, and they cannot explain how our corporatist regimes responded to the problems posed by financial crisis.

3. Swedish contention and financial crisis

On 7 February 1990, Kjell-Olof Feldt, the treasurer of the minority social-democratic government, announced a crisis package that included command-economy proposals to freeze wages, prices, rents, dividends and taxes, and even ban strikes during 1990 and 1991 (Feldt, 1991: 456–69). Amid these dramatic proposals, an unheralded declaration of intent to apply to join the EU was almost an aside, meant to underline how serious the government was about controlling inflation. A week later, the Riksdag (parliament) rejected the crisis package bill, the government resigned, and Feldt left politics altogether. After another week, the social democrats formed a new minority government.

The only aspect of all the drama (Vandenberg, 1990; Vandenberg and Dow, 1991) that had longer-term importance was the government’s sudden declaration of intent to apply to join the EU. In retrospect, the government had effectively declared it no longer had faith in the capacity of the Swedish Employers’ Association and the union peak bodies to negotiate non-inflationary national wage agreements. Feldt’s successor, Allan Larsson, made further attempts to institute public controls over wage-setting but they also failed. In January 1991, Larsson said in his budget speech that reducing inflation had to come before all other goals. Many interpreted this as the final and formal end for the long-standing priority of full employment (Svenning, 2005: 53–5). When the bourgeois parties replaced the social democrats at the September 1991 elections, they had no qualms about making reduced inflation the first priority of public policy.

During 1990, major developments took place in the routines of industrial relations. In 1983, the large engineering corporations (such as Volvo, Electrolux and Ericsson) had managed to persuade the metalworkers to settle wage contracts separately from the centralized negotiations between the Swedish Employers’ Association and the union peak federations. In 1990, the employer federation finally ceded to the engineering corporations and declared that it would no longer negotiate wages or conditions at a national level with the Confederation of Unions. It backed this up with a decision to close its offices of statistics and negotiations, and a year later it withdrew all representatives from the boards, committees and inquiries that had built up around tripartite ‘democratic corporatism’ over almost 90 years (Rothstein, 2002).

Serious economic problems also developed during 1990. The real-estate price bubble burst in May. By October, real-estate slumps in Stockholm, London and Brussels sent Sweden’s largest finance company, Nyckeln, bankrupt. The collapse of Nyckeln and other finance companies revealed extensive exposure in the partially privatized Nordbanken, SE-Banken, Första Sparbanken and Gotabanken. During 1991–2, the social-democratic government paid 65 billion crowns (about US$7 billion) to Nordbanken in return for shares when it bought the bankrupted Gotabanken. It also demanded shares from Första Sparbanken in return for guaranteeing its loans to that bank. When the economy recovered and the value of the banks’ shares also recovered, the government sold its shares and recovered about half of the money it had given the banks (Bäckström, 1997).

The breakdown in the centralized negotiation of wage-setting, the collapse in real-estate values and the banking crisis all contributed to the currency crisis of September and November 1992. The final factor was a change in government in September. Devaluations are most commonly undertaken by new governments that seek to differentiate themselves from their predecessors and to compensate for the loss in international competitiveness usually supposed to have been caused by the inflation that had accumulated under former governments. Contrary to international speculation, however, both the social democrats and the incoming bourgeois government repeatedly disavowed any recourse to devaluation. Instead, they conformed to the stable-currency and low-inflation criteria for EU membership.

The British government and other participants in the partial pegging regime of the European Exchange Rate Mechanism (EERM, a forerunner to the European Monetary Union) also sought stable currencies and made the pursuit of low inflation the first priority of public policy. The collapse in real-estate values and the associated banking crisis, however, sustained the risk of devaluation. Britain led the European defence of the EERM as it invested its gold reserves heavily in maintaining the value of the pound but on 16 September 1992, the speculation proved too great. The pound was allowed to float completely.

In Sweden, the Riksbank continued to resist the same course and this attracted the attention of speculators. On 17 September, the Riksbank set the daily instrumental interest rate at 500 percent per annum, and it was prepared to set the marginal rate at 4000 percent if necessary (Bäckström, 1997; Dennis, 1998: 49–56). The Riksbank succeeded in the short term and the daily instrumental rate fell to 20 percent. By November, however, the number of bankruptcies and the strain on the banking system forced it to allow the krona to float completely. Resisting devaluation because of its potential to cause inflation resulted in higher interest rates and a loss of competitiveness for small export-oriented firms, but more employment was lost in the public sector. During 1990–4 about 185,000 jobs were lost in industry and another 100,000 in private services (such as banks), but cuts to public spending saw 253,000 jobs lost in the public sector (Ehrenberg and Ljunggren, 2006: 126).

Poor economic circumstances saw the social democrats returned to office in September 1994. During the lead-up to the referendum held in November, the newly elected and popular social-democratic prime minister, Ingvar Carlsson, led a successful Yes campaign to join the EU. Along with Finland and Austria, Sweden officially joined the EU in 1995. How did the government manage to persuade a majority of voters to agree to join the EU? Why did Sweden abandon its long-standing commitment to full employment as the first priority of public policy and then turn to the neoliberal public policies required to join the EU? What consequences did this shift in public policy have for the patterns of protest against and support for the government? To answer these questions we turn to our three dynamics of contentious politics: changes in trust networks, changes in cross-class alliances and the brokerage of new connections.

Swedish citizens’ networks of trust remain well integrated in the public sphere of debate, policy formation and policy implementation. This is a legacy of the way the labour movement and other social movements mobilized from 1890 to 1920 to pursue democratization as a ‘joint project’ with liberal elites who dissented against an anti-democratic conservative regime (Collier, 1999: 83–5). Unions and social movements developed strong links between workplaces, local organizations, cooperatives and neighbourhood organizations and their respective central committees or bodies, which had a strong national presence (Rothstein, 2002: 295). Unions in Sweden remain both strongly decentralized, with strong local negotiation of piece rates for example, and highly centralized, with well-staffed national offices and extensive engagement in industry-wide wage negotiations and all manner of public inquiries and boards of public authorities (Higgins, 1985a; Kjellberg, 1997, 2007).

During the 1990s, however, membership in many different civic associations declined somewhat from very high levels. In 1993, 92 percent of the adult population belonged to at least one civic association and by 2001 this had declined to 89 percent (Persson, 2003: 11). Older people belonged to civic associations in greater proportions than younger people (Vogel et al., 2003: 58, 64, 70–1). There is, however, an important nuance in this picture of high but declining social capital in networks of trust.

Bo Rothstein (2002: 330–1) argues that networks of trust among citizens in the wider community remain strong and vibrant, but that since the end of the 1980s, unions are less well integrated in public decision-making because national union leaders and corporate leaders no longer trust each other. He attributes the breakdown in elite trust to, first, the unions’ mobilization for economic democracy and the social democrats’ legislation of extensive union rights at work from the mid-1970s, and second, the employers’ neoliberal counter-mobilization during the 1980s. From our dynamics-of-contention perspective, ideological mobilization and counter-mobilization is an important but insufficient explanation of the differential changes in the networks of trust because it relies on the actors’ views of the world. A fuller explanation needs to include consideration of how fragmentation within both the union movement and the Employers’ Association affected their relations with each other, with the government and with sites and repertoires of contention.

Since the early 1970s, technological development and structural change in the global economy had pushed the global competition-exposed engineering employers to dissent from the approach that the Employers’ Association took in the central wage negotiations with the various union peak bodies. The engineering employers disputed the wisdom of accepting that ongoing wage-rate compression was a fair price for no-strike clauses in central agreements. Besides a wider spread of wage rates, they demanded that other employers and unions allow wages in the competition-exposed sector to be the benchmark for national wages in accordance with national productivity increases. When this failed to happen, they sought greater flexibility in local piece-rate setting in order to cope with the merging of blue- and white-collar tasks and to compete for scarce skills in a full-employment economy. In 1983, they used high wage offers to persuade the metalworkers to negotiate independently of the central wage-setting procedures, but it was not until 1990 that they could finally persuade the Employers’ Association to abandon national wage-setting altogether.

Peter Swenson and Jonas Pontusson (2000) argue that it was a cross-class alliance, between the engineering employers, the metalworkers and a bargaining group of white-collar unions in private industry that brought the peak bodies’ centralized wage negotiations to an end. The timing of the changes to wage-setting procedures illustrates the agency of the engineering employers’ active choices because the imperatives of technological and economic structural change developed gradually over a long period. How did this cross-class alliance affect other actors and relations between actors, sites of contention and repertoire of contention?

The most direct effect of the cross-class alliance between the engineering employers, the metalworkers and the white-collar engineering workers was to ally actors in export-oriented manufacturing against three groups of actors shielded from global competition. One, they allied against both large retailers who could pass the cost of higher wages on to their customers and the large timber and paper manufacturers whose wage costs were small compared to their capital costs (Swenson and Pontusson, 2000: 92–4). Two, they allied against public sector employers who could either trim their capital maintenance costs in workplaces such as schools and hospitals, or lobby for higher taxes to cover wage increases. Three, they allied against public sector unions that mobilized around feminism (Higgins, 1996) to attack the gap between men’s private sector wages and women’s public sector wages for comparable work.

From the mid-1970s to the end of the 1980s, inflation aggravated tensions between the engineering employers and other parties to wage contracts. They insisted that manufacturing productivity had to be the benchmark for wages across the economy while other employers were more interested in avoiding strikes or could avoid or pass on their increased wage costs. The public sector unions mobilized against the gendered pay gap between the public and private sectors. Between 1976 and 1982, successive bourgeois and socialist governments used currency devaluations to adjust for inflation. The social democrats, liberals and conservatives’ turn to the EU saw socialist and bourgeois governments abandon devaluation as a last resort and join the trade-exposed employers and unions’ cross-class alliance against non-trade exposed employers and unions.

A second effect of the pro-EU cross-class alliance was that militant unions outside the trade-exposed sector began to consider collaboration with the anti-EU leftists or greens in parliament and began to support diverse groups opposed to the EU. The municipal workers drew on feminism to motivate campaigns against privatization and authoritarian managerial practices (Higgins, 1996), and paid local officials to help organize the Stockholm social forum (interview, Ann-Terese Mörch, 26 March 2004).

The Union of Commercial Employees organized a successful consumer boycott of Toys ‘R’ Us (Vandenberg, 2006) and mobilized its younger members to support the clean clothes campaign against child labour (interviews, Stefan Carlén, 26 February 2007; Sven Englesson, 1 March 2004). The blue-collar Confederation of Unions also attempted to forge connections with the campaign for a Tobin tax. Unions formed connections with anti-globalization and anti-EU groups in the Toys dispute and other campaigns, but their achievements were defensive rather than constructive. They were a matter of resistance against unwelcome incursions on established rights rather than a matter of, say, advancing industrial democracy or scaling back the prerogatives of management. Part of the reason why the pro-EU cross-class alliance was stronger than the anti-EU groups was that they presented themselves as competing for a proud and wealth-generating place in a rapidly globalizing world, while anti-EU groups reverted to a national sense of identity to resist the insecurity that flows from globalization and integration with the EU.

Korean contention and financial crisis

A high-capacity, repressive state suppressed civil society for several decades after the founding of the anti-communist Republic of Korea in 1948. The state repressed trade unions due to suspicions that they were communist sympathizers. Independent unions were destroyed, and replaced by loyalist, ‘yellow’ organizations that supported the intelligence agencies in monitoring workplaces. These belonged to the Federation of Korean Trade Unions (FKTU) and were incorporated as a highly subordinate actor in the development model that emerged in the 1960s. The model also entailed a small number of family-owned and managed conglomerates – the chaebol, which received privileged access to resources in return for supporting the regime and for acquiescing to state designation of priority industries (Hundt, 2009).

Individuals were promised ‘lifetime employment’ and a set of basic rights in return for quiescence, but this came at the expense of collective rights to organize and bargain. Korean unions failed to play the advocacy role expected of them, and working conditions subsequently remained relatively poor even in recent times when evaluated by indicators such as hours worked each week, fatal injuries in the workplace and treatment of staff (Buchanan and Nicholls, 2003: 206, 222; Kong, 2006: 377; Ogle, 1990).

Korean civil society proved to be resilient despite decades of repression, and key constituents of the minjung (people’s) coalition such as churches (Wells, 1995) and students (N Lee, 2007) played a crucial role in building class consciousness among workers in the 1970s and 1980s (Koo, 2001). The linkages between minjung forces and workers had at least two discernible effects.

First, it greatly improved the organizational capacity of workers, and fostered the formation of independent unions that were willing to undertake the advocacy role neglected for so long. From a low base and despite restrictions, which allowed only one union to organize in each workplace, the independent unions enjoyed strong growth during the 1980s. The number of unions in Korea more than doubled after 1986, reaching more than 6000 by 1988 (Burkett and Hart-Landsberg, 2000: 157). The independent unions formed their own peak body, the Korean Federation of Trade Unions (KCTU).

Second, the revitalization of the union movement provided anti-state forces with a critical mass of support in their quest for political change. The minjung alliance and KCTU-aligned unions formed the vanguard of the democratization movement in the summer of 1987. The movement also gained support from the middle classes, indicating a widespread demand within Korean society for political change (Chung, 1997; H Kim, 2007: 213–14).

In the more open social and political conditions of the democratic era, a new group of civic organizations emerged, with a focus on middle-class issues such as human rights, the environment and gender equality. Two of the most long-standing and successful ‘comprehensive citizens’ movements’ to emerge at this time were the People’s Solidarity for Participatory Democracy (PSPD) and the Citizens’ Coalition for Economic Justice (CCEJ). They could be distinguished from the independent unions and minjung partners by their degree of class consciousness, strategies and priorities for social and economic reform. Conservative politicians and media encouraged this splintering of civil society by depicting students and the KCTU as dangerous radicals, and cited a propensity for violent protest and support for dialogue with North Korea as evidence that these elements could not be trusted (Jee, 1997: 151–2; JK Kim, 2008: 305–6).

In contrast, the middle-class civic groups eschewed violence as a tactic, accepted the basic framework of the new democratic political system and interpreted their role as acting as a check on executive power. For the PSPD, founded in 1994, the primary focus of its activities was the ‘monitoring of power’ (Yoon, 2001: 69). The focus of the CCEJ was economic justice and procedural democracy. The themes of participation, accountability and transparency guided its projects to monitor the attendance and activity of National Assembly members, to raise standards of corporate governance and to protect consumer rights (Burkett and Hart-Landsberg, 2000: 194–5; Multinational Monitor, 1995: 27).

The effects of democratization on the Korean development model were becoming apparent by the mid-1990s. The government applied for membership of the OECD in 1996. The application was rich in symbolism: Korea had transformed from one of the world’s poorest states in the early 1960s to a member of the ‘rich world club’ only three decades later, with income levels reaching US$10,000 per person. However Korea was required to further open its economy, including the financial system, to international markets. The government embraced a liberalization of the economy because a trade-dependent state should foster competition rather than avoid it.

The government also sought to increase the ‘flexibility’ of the labour market, including the dissolution of employment guarantees that had accrued in earlier decades. The manufacturing sector witnessed particularly strong improvements in wages and conditions due to successful campaigns by unions aligned to the independent KCTU.

In response to complaints from the chaebols that wage increases were destroying their competitiveness, the government passed a raft of amendments to the Labour Relations Law in late 1996. The aim was to make it easier for firms to lay off workers (Kong, 2006: 371). Unions strongly opposed the removal of employment protections, and staged a series of strikes to repeal the rollback of protections. For the first time, the state-aligned federation and the KCTU joined forces to defend labour rights. Under pressure to end the strikes, the government agreed to repeal the most restrictive elements of the new legislation. In March 1997, a compromise was reached whereby unions acceded to limited job security in return for greater freedom to organize. The government also granted pay increases to both public and private sector workers (Buchanan and Nicholls, 2003: 218).

Evidence that the chaebol-centred development model was faltering was not difficult to find. A mid-sized chaebol, the Hanbo Group, declared bankruptcy in early 1997. Hanbo’s loans came at low interest rates and allowed the group to record extraordinarily high levels of debt, even by the standards of the chaebols (Schopf, 2001: 709–10). The government nationalized Hanbo, but its collapse shook investor confidence. The collapse of Kia Motors followed months later. This loss of confidence became especially important when investor panic spread across Southeast Asia in the second half of 1997. Within months lenders were scrutinizing the debt levels, investment patterns and corporate governance practices of the chaebols. By year’s end, Korea was forced to apply to the IMF for a loan to repay the costs of actual and potential corporate failures.

In return for a bailout loan from the IMF, the government agreed to a range of changes: spending was reduced, reforms were enacted to increase the quality of loans and new regulations were introduced to improve standards of corporate governance. The IMF also recommended that the flexibility mechanisms that were originally mooted in late 1996 be introduced without amendment (Hundt, 2005: 245–7).

Amid the financial crisis Kim Dae-jung, who survived assassination and imprisonment at the hands of authoritarian regimes, won the presidency at his fifth attempt. Unlike his predecessors, he was somewhat amenable to bringing non-traditional elites into government. The trust networks of civic movements thus began to move into the public sphere at the same time as the state–chaebol networks of privilege and corruption were coming under scrutiny. The new government recruited civic groups to support elements of its chaebol reform agenda such as the enhancement of corporate governance. The ‘small shareholders’ movement’, for instance, sought to ensure that chaebols called annual general meetings as required in their statutes, other groups campaigned for the appointment of outside directors and some civic groups provided input to social policy by suggesting incentives for women to re-enter the workforce. Some representatives from civil society even served as Cabinet ministers (H Kim, 2000: 603–4; 2007: 216–19; JK Kim, 2008: 79–80; Lim and Jang, 2006: 449–51).

In order to ease the implementation of the IMF’s rescue package, the government invited the rival union federations to partake in tripartite talks. Both federations had supported Kim Dae-Jung during the election campaign (Kong, 2006: 377–8), and this was an attempt to establish democratic corporatism around a pact between labour, government and business. The 6 February pact expanded basic labour rights, allowed for the formation of a teachers’ union and allowed laid-off workers to join a union. Unions won the right to establish a political party, and the KCTU won belated recognition as a peak body. The pact largely removed the expectation of job security, although the government established a multi-billion dollar fund to compensate workers for job losses resulting from restructuring (Ha and Lee, 2007: 910–11).

The unions had a range of motivations for participating in the talks. First, the union federations argued that the chaebols were largely responsible for the financial crisis. For instance, they argued that the embattled Kia Motors was a well-run firm experiencing only temporary liquidity problems, in part due to the bigger chaebols’ monopolization of credit markets. The unions recognized the value of positioning themselves on the side of reform during a time of national crisis (Burkett and Hart-Landsberg, 2000: 200). Union leaders proposed that chaebol restructuring proceed via job sharing and wage freezes. In other words, while agreeing to the new ‘flexibility’ mechanisms, unions sought to save jobs. The public was wary of any group that appeared to share the chaebols’ opposition to reform. As one union official recollected: ‘There had been and was a very strong mood among the members to fight, but members were worried about how their actions would be interpreted within the overall context of the economic crisis itself. They found it difficult to be seen as the people who rock the boat at a time of difficulty’ (quoted in Neary, 2000: 4).

Second, the talks went some way towards fulfilling the aspirations of union leaders to alter their status as outsiders to routine politics. Unions emphasized the symbolism of the talks, and Yoon Young-mo of the KCTU argued that: ‘Labour could now be included and recognised as a pillar of the society, as a value in itself as an organised progressive force. In this way you could philosophically change the way in which labour was perceived in society’ (quoted in Neary, 2000: 4).

Third, unions could change their organizational strategy. Instead of operating solely at the enterprise level, they recognized an opportunity to create industry associations (Kong, 2006: 376). Some of the legal barriers to creating such structures were removed, although problems remained in terms of collecting membership fees at the industry and national levels. Consequently the fledgling industry unions were ill equipped to undertake nationwide negotiations over wages and conditions with government and employers (Buchanan and Nicholls, 2003: 221).

Despite the unions’ willingness to compromise, the government played the most forceful role in the restructuring process. Kim Dae-jung proposed that the chaebols undertake asset swaps to reduce overcapacity and indebtedness. The state dictated the terms of restructuring, and the Minister for Finance and Economy warned: ‘the public must understand that reducing employment by 10 to 20 percent through restructuring is a way to prevent a situation where 100 percent of jobs are lost to unemployment in the future’ (quoted in Song, 1999: 91).

This left unions with less autonomy than initially envisaged. Problems arose when the union federations returned to their members with details of the pact. Delegates voted to reject the agreement, which placed the brunt of the burden for reform on workers (Chang and Chae, 2004: 432–4). With their constituents opposed to the proposed model of restructuring, some union leaders saw no option other than to return to the non-routine politics of strikes, sit-ins and protests to press their claims on government. The government reverted to authoritarian means of coercion – such as the violent suppression of strikes and the arrest of union leaders – to halt industrial disputes (Ha and Lee, 2007: 912). Two examples illustrate the government’s response to the unions’ reversion to non-routine politics.

First, Hyundai Motors announced in early 1998 that 10,000 workers would need to accept voluntary redundancies for the company to survive. More than 8000 workers opted for early retirement. Individual members acted in what they perceived as their self-interest, but the Hyundai union viewed voluntary redundancies as a direct threat to its strength. The union proposed saving jobs through wage cuts and job sharing. When the company rejected these options, 30,000 union members went on strike in July 1998 to oppose further redundancies and factory closures. After holding out for two months against the company and government, the union capitulated. In the face of the government’s threat to break the strike via heavily armed riot police, the union agreed to the sacking of more than 200 workers and the dispatch of more than 1200 others on extended unpaid leave – bringing the total reduction in the workforce close to the 10,000 mooted only months earlier. Two-thirds of members rejected it, but the deal was ratified (Neary, 2000: 3–5).

A second dispute illustrated the partial revival of links between unions and other elements of civil society. The student movement supported a strike by subway workers in April 1999 by sheltering workers on a university campus. The students, however, proved incapable of protecting the workers from the repressive powers of the state. Student activists claimed that their efforts to support workers did not attract widespread support: ‘There are a whole generation of young people in the universities for whom the student activists are very unpopular. They are too militaristic and their songs and language and idea have nothing to do with them’ (quoted in Neary, 2000: 7).

The failure of the campaign damaged efforts to create links between the unions and affiliated civic groups. Both federations had threatened to quit the tripartite talks on several occasions, but were persuaded to return. Under pressure from the government to rein in militancy, the KCTU cancelled three general strikes in 1998. The subway strike was intended to form the ‘first wave’ of a series of rolling strikes in the spring of 1999, but the harsh treatment of the subway workers dissuaded secondary action by affiliated unions. The KCTU cancelled the strike, and also quit the talks.

The failure of the strike marked the end of the contentious episode in Korea. The cross-class coalition failed to maintain its unity and sense of purpose in the face of state repression and the state’s maintenance of links with the chaebols, the IMF and some middle-class social movements. For their part, affiliated civic groups found fault with the unions’ strategies, highlighting the philosophical differences between the two wings of Korean civil society. Park Won-soon, the leader of one middle-class group, argued that the main problem was the unions’ inability to persuade the wider public that their campaign was worthy of support: ‘In the French case, the public will support subway workers if they go on strike, but it’s not that case for us. . . . Labour has to present evidence that strikes are in the interests of society, not just the working class’ (quoted in Yoon, 2001: 72).

5. Conclusions

We have argued that it was changes in trust networks, shifting cross-class alliances and the brokerage of new connections between actors and sites of contention, rather than an inevitable or general collapse of corporatism that saw Sweden and Korea each abandon full employment regimes and converge on neoliberal public policies during the financial crises in 1990–3 and 1997–8.

Our comparison of episodes of contention during these financial crises shows that unions lost influence over public policy formation when their respective governments sought membership of supranational organizations. The turn to international bodies also had the effect of fragmenting the union movement in each state.

Prior to the financial crisis in Sweden, the blue-collar Union Confederation had managed to maintain some solidarity across the public and private sectors even as high inflation had aggravated its long-standing insider status. Inflation strained relations between the unions and both the Employers’ Association and the social-democratic government, which undertook liberalizing reforms during the 1980s. After the financial crisis, the metalworkers remained insiders when they joined a pro-EU cross-class alliance with the engineering corporations, the social democrats, liberals and conservatives while public sector unions turned to EU-sceptical outsiders among the left and green parties and diverse newly established social movements. This particular cross-class alliance explains change in a more satisfying way than generalizations about a capitalist state that incorporates unions and protest groups in order to promote the interests of transnational corporations and neoliberal globalization at the expense of workers’ and ordinary citizens’ interests in more jobs, better wages and better welfare.

Prior to the financial crisis in Korea, insider unions in the pro-government FKTU joined the outsider unions in the militant and independent KCTU to support the first president to win an election from opposition. There was also a broad alignment between all unions and the middle-class civic movements calling for greater corporate accountability. After the financial crisis, the independent unions abandoned the beginnings of social-corporatist tripartite discussions and reverted to their outsider status as they again deployed a familiar repertoire of contention against the new government, which reverted to a familiar repertoire of authoritarian repression.

This shift from a brief experiment in democratic corporatism before the financial crisis to neoliberal practices during and after the crisis is less clear than the shift in Sweden, where the metalworkers chose to remain insiders and part of a pro-EU cross-class alliance. In Korea, comparable groups of workers in competition-exposed manufacturing chose to revert to contention against the government from the outside. In both cases, governments benefited from the fragmentation of unionism as different unions pursued routine or contentious politics.

We have focused on three particular causes (changes in trust networks, cross-class alliances and the brokerage of new connections between actors and sites of contention) for a specific process (the decline of contrasting forms of corporatism), rather than attempt to use a general (or ideological) cause of corporatism to explain recent developments. Our decision to compare Sweden against Korea might at first glance seem to reflect a conventional and nation-state-centric approach to comparative politics. The statistical data in Section 2 reinforce that impression. The narratives in Sections 3 and 4 depart from the conventional approach to comparative politics in two ways. We eschew general causes of differences between nation-states and we compare specific episodes of financial crisis during which governments used their connections with international financial bodies to establish a new regime for governing their domestic political economy. We argue that the diverse actors involved – governments, unions, NGOs and various corporations – were active proponents of contrasting means of coping with globalization.

Our analyses of networks of trust among unionists and other social movements and between union and corporate leaders produced nuanced pictures of why some unions chose insider negotiations with governments while others chose outsider contention as financial crises imposed serious threats to workers’ livelihoods. The analyses of cross-class alliances produced nuanced pictures of why various unions chose different means of attempting to capitalize on the opportunities posed by the financial crises. The governments of both Korea and Sweden, however, proved adept at brokering new connections between domestic and international sites of routine politics. These connections constituted a somewhat optimistic and cosmopolitan sense of identity founded on the nurturing of globally competitive manufacturing sectors. In the circumstances of financial crisis, this cosmopolitan identity prevailed against nationalist identities that failed to broker strong connections in either national or international sites of contention.

These two governments succeeded in transforming national regimes of corporatism into more internationally oriented regimes of economic government but this does not mean that other governments might emulate these states or that their experience will be repeated elsewhere. It suggests that both insiders and outsiders in other regimes elsewhere can learn from how particular actors interpreted their circumstances successfully or not, and prevailed or did not prevail against other actors, and in future attempt to interpret their own necessarily different circumstances in fruitful ways. Our claims about how to explain shifts in the labour market regimes of Sweden and Korea are specific to the financial crises of the 1990s. We have developed ideas about contentious politics as we deployed causes of democratization to explain shifts in the national regulation of the effects of globalization and we have shown how the politics of contention offers explanations of change that are more satisfying than the descriptiveness of ‘corporatism’.