Abstract
Some European welfare states protect outsiders from economic uncertainty better than others. Conventional theories attribute this either to the material interests of producer groups or the vote-seeking actions of political parties. By contrast, this article focuses on the interaction between trade unions and governments, drawing on empirical experience in Austria and Sweden. High levels of inclusiveness provide trade unions with a powerful interest in protecting outsiders. Governments, however, have moved towards dualistic policy choices, but when weakened by intra-coalitional divisions or a hung parliament, they retain a political incentive to negotiate political deals with trade unions, and this forces them to compensate outsiders for economic uncertainty.
Introduction
The problem of dualization occupies a central place in the comparative political economy of contemporary capitalism. Scholars focus in particular on the precarious employment and welfare standards faced by an increasingly numerous group of ‘outsiders’ (Emmenegger et al., 2012; Palier and Thelen, 2010; Thelen, 2014). Unlike ‘insiders’ with secure contracts, these are either unemployed or have precarious jobs and, therefore, often lack adequate coverage in such fundamental areas as wage bargaining, job security and welfare benefits. This dualization reflects not only profit-maximization managerial strategies but also public policy changes deregulating protection arrangements on the grounds of job creation.
The politics of labour market reform is, therefore, fundamental to the social protection of outsiders and the problem of dualization at large. But what do we know about how political actors change the social protection of outsiders over time? The principal empirical studies of comparative political economy offer two different theoretical approaches to this problem: producer group theory and partisanship theory. The first highlights the role of employers and unions in shaping producer group coalitions to achieve their preferred policy outputs (Martin and Swank, 2012; Thelen, 2014). In this perspective, the inclusiveness of producer group associations is the dominant cause of variation in the protection of outsiders. The second explanation emphasizes the vote-seeking actions of political parties. In this perspective, electoral alliances behind party coalitions are the main cause of variation in the protection of outsiders (Beramendi et al., 2015; Rueda, 2007).
These contrasting approaches share two features. First, they argue that the viability of social solidarity rests on encompassing pro-welfare coalitions. The causal mechanism is that only a united intra-labour coalition has the capacity to push for inclusive policy choices that satisfy divergent social demands. Partisanship scholars highlight the task of reconciling the policy demands of divided electoral constituencies, while producer group scholars emphasize the associational task of incorporating service-sector workers into the traditionally manufacturing-dominated industrial relations arena. Although their focus is different, they have in common a similar underlying logic. Second, they focus on either parties or producer groups to understand the national dynamics of labour market change; neither approach allows for their interaction. This is somewhat surprising, given that the number and character of actors involved in the reform process has important effects on the type of power relations underlying the final policy output. Negotiated reform with trade unions imposes constraints on the initial ambitions of national governments, while a unilateral reform strategy extends the partisan room for manoeuver.
Challenging conventional theories, I argue that variation in the protection of outsiders results not from differences in producer group coalitions or partisanship alone but from differences in the power-distributional interaction between unions and governments. High inclusiveness provides unions with a strong interest in the protection of outsiders. But the impact of union preferences has declined over the past three decades, with a shift in the balance of class power from labour to capital across the advanced capitalist economies. National governments of all partisan colours have accordingly prioritized flexibility for employers over social protection for outsiders. When they have an ideologically united majority in parliament, they may be strong enough to resist outsider-oriented policy demands from inclusive union movements and pursue a unilateral reform strategy instead.
Yet despite a decline in union power, governments are not always the driver behind the reform process: they may be unable to shape policy outputs when they are internally divided or lack a parliamentary majority. Either condition can prevent a unilateral reform strategy and create reform deadlock. Therefore, I emphasize the political opportunities for trade unions when governments are weak. Governments are then more responsive to union demands because they need support from extra-parliamentary actors to mobilize consensus. This argument draws on the social pact literature and extends to governments of the partisan right as well as the left (Baccaro and Lim, 2007; Baccaro and Simoni, 2008; Hamann and Kelly, 2007).
I compare labour market reform trajectories in Austria and Sweden. Despite many similarities between these countries (small size, corporatist legacies, periods of centre-right coalitions, relative macroeconomic success and mature welfare states), their reform trajectories not only differed markedly but also evolved in very counterintuitive directions. Austria improved the protection of outsiders by unifying the entitlement conditions for severance pay, broadening the coverage of unemployment insurance and expanding spending on active labour market policy (ALMP); Eichhorst and Marx, 2012; Fervers and Schwander, 2015; Obinger et al., 2012). Outsider-oriented adjustment patterns appear puzzling in a conservative political economy with a segmented and male-dominated institutional legacy. By contrast, the Swedish social-democratic welfare regime shifted markedly to dualization: permanent workers still enjoy relatively high levels of job security, while fixed-term and unemployed workers faced the burden of labour market adjustment through a deregulation of temporary employment as well as cuts in benefit entitlements and ALMP spending (Davidsson and Emmenegger, 2012; Lindvall and Rueda, 2012). Hence, Sweden has become more dualistic than Austria.
The comparison is important for two methodological reasons. First, challenging partisanship theory, they show that the degree of union involvement mediates the policy impact of different party governments. An electoral logic of labour market reform, however, ignores the causal significance of union influence on public policy. Second, challenging producer group theory, they demonstrate that union influence rests on the political choice of governments to work with unions. A coalitional logic, however, misses the fundamental fact that governments may turn against negotiated reform to undermine union influence in the interest of neoliberal reform. Swedish governments, for example, gradually turned their back on the trade unions and instead pursued a unilateral reform strategy. By contrast, Austrian governments revived labour-inclusive reform patterns in labour market policy. This cross-national variation in government reliance on union support explains the diverging capacities of Austrian and Swedish labour movements to protect outsiders.
This article proceeds as follows. First, I present my argument about the power-distributional interaction between parties and unions. Second, I illustrate this argument through causal process tracing in the area of labour market policy in the two countries, drawing on primary and secondary sources as well as 24 semi-structured interviews with policy-making elites (party spokespeople, interest group representatives, government officials and academic experts). The final section considers the limits and implications of my findings.
A power-distributional framework: governments and trade unions
Governments operating in capitalist democracies have to reconcile the distributional tension between capital accumulation and social stability (Polanyi, 1944). But in the neoliberal era, they are increasingly unable to perform this task. Increasing flexibility demands for capital accumulation, on one hand, and decreasing public resources for risk protection, on the other hand, constrain the capacity of political parties to sustain social solidarity (Streeck, 2009; Streeck and Mertens, 2013). As a result, wage differentiation and labour market flexibility have become the two main policy instruments for national governments to promote job creation without raising public deficits or devaluing national currencies (Culpepper and Regan, 2014). Of course, this is not to say that partisan differences have disappeared, but divisions between left- and right-wing governments with respect to dualization have declined (Emmenegger et al., 2012: 311).
Structural constraints on the governing capacities of political parties have, therefore, underpinned the gradual shift in the balance of class power from labour to capital during the neoliberal era. While this shift was a common trend in the past three decades (Baccaro and Howell, 2011), its effect on union inclusiveness was not uniform across different national models of capitalism.
Gordon (2015) proposes an index of the ‘inclusiveness’ of union movements in 18 OECD countries in the period 1985–2005, comprising three dimensions: union density, centralization and involvement in the administration of unemployment benefits. Sweden ranks first, scoring 78.7, with very high density and a central union role in the ‘Ghent’ unemployment benefit system (Clasen and Viebrock, 2008; Rothstein, 1992). Austria ranks fifth, at 59.2 – well above the remaining countries. Austrian unions cannot rely on a ‘Ghent effect’, and density has fallen sharply in recent decades, from almost 68 percent in 1960 to only 28 percent in the early 2010s. Yet the single union confederation, the Österreichischer Gewerkschaftsbund (ÖGB), exerts an unquestioned leadership and political mandate on behalf of its affiliates, providing levels of concentration and centralization unmatched in an international comparative perspective.
The implication is that two different factors underlie outsider-inclusive union preferences: density and centralization (Gordon, 2015: 89–91). First, high density rates mean that unions organize a higher share of outsiders, directly incorporating their social demands into the representational outlook of organized labour (Becher and Pontusson, 2011). Second, strong centralization empowers the voice of those member unions that are hit hardest by the low-wage competition resulting from a growing number of outsiders. In centralized union movements, confederal authority mitigates an unequal distribution of sectoral intra-union power and thereby counteracts particularistic policy priorities driven by insider-dominated manufacturing unions (Western, 1997: ch. 3). Centralization, therefore, allows the weaker service-sector unions to push the confederal elite to resist dualization.
We should expect the relatively high levels of inclusiveness in Austria and Sweden to produce a vested interest in the protection of outsiders. The formation of inclusive union preferences, however, is not sufficient for the successful protection of outsiders because the gradual decline in union power has weakened the impact of organized labour in the policy-making process, opening up opportunities for governments to impose neoliberal reform ambitions against the ‘solidaristic’ preferences of inclusive union movements.
Yet despite a decline in union power, governments cannot pursue a unilateral reform strategy when they lack a cohesive parliamentary majority. Such weak governments may fall back on union involvement to mobilize an extra-parliamentary channel of consensus. This argument builds on the literature on social pacts, which also emphasizes the interaction between weak governments and trade union involvement, but extends it both by demonstrating its applicability beyond the industrial relations sphere and, most importantly, by clarifying the operational meaning of weak government: not just a minority government, as argued in the previous literature, but any type of government whose components (coalition partners and supportive opposition parties) have trouble coming to issue-specific agreement and, therefore, find it more expedient to resort to labour-inclusive forms of consensus mobilization.
Therefore, weakness, on this definition, refers to governments that lack either cohesive policy preferences or parliamentary support for legislation reflecting their policy preferences. Trade unions are influential under these conditions because their support provides weak governments with an extra-parliamentary channel of consensus mobilization. This creates a powerful incentive for governments to delegate policy-making authority to trilateral negotiations with capital and labour. Unions can credibly demand substantial concessions from a weak government because their assent to new policies can make the difference between success and failure. Strong governments, by contrast, can depend on an ideologically cohesive majority in parliament and do not depend on extra-parliamentary support for consensus mobilization.
To assess the strength of governments in Austria and Sweden, I use the partisan composition and vote shares of coalition governments as a proxy measure (see Table 1). Since the late 1980s, Austria has been governed by a grand coalition between the Sozialdemokratische Partei Österreichs (SPÖ) and the conservative Österreichische Volkspartei (ÖVP), except for a 6-year period of coalition from 2000 to 2006 between the ÖVP and the populist right-wing Freiheitliche Partei Österreichs (FPÖ), which has attracted considerable working-class support in the past three decades. 1
Governments in Austria, 1987–2017.
SPÖ: Sozialdemokratische Partei Österreichs; ÖVP: Österreichische Volkspartei; FPÖ: Freiheitliche Partei Österreichs.
Austrian grand coalitions comprise two ideologically divided and similarly strong parties. Both attributes weaken their autonomous reform capacity and thereby increase the likelihood of reform deadlocks. In addition, close ties to the social partners in a historically evolved corporatist environment pose constraints on unilateral reform ambitions. Vertical coordination usually takes place between the SPÖ and organized labour, on one hand, and the ÖVP and organized business, on the other hand. 2 We would, therefore, expect the Austrian grand coalition to facilitate union influence on the policy-making process. By contrast, the formation of a centre-right majority government implies a higher level of intra-coalitional cohesiveness. This strength perhaps best explains why the centre-right government was able to break with the long-standing tradition of social partner involvement, especially in the area of old-age pensions (Obinger and Tálos, 2006; Pernicka and Hefler, 2015). In the area of labour market policy, however, the neoliberal agenda driven by the ÖVP encountered resistance from the populist employee wing of the FPÖ. Faced with ensuing reform deadlocks, the government fell back on the well-established practices of corporatist political exchange (Afonso, 2013; Rathgeb, 2017). Therefore, intra-coalitional divisions prevented unilateral reform processes that would have excluded the one single actor that incorporated the interests of outsiders in its political priorities: the ÖGB.
In Sweden, the social democrats (Sveriges socialdemokratiska arbetareparti (SAP)) headed the government continuously from 1932 to 1976. In the period on which I focus, a centre-right coalition was in office from 1991 to 1994, headed by the conservative Moderaterna (M) together with the Centerparti (C), the Liberalerna (L) and the Kristdemokraterna (KD). The SAP returned to office with a single-party minority government in 1994–2006, giving way to another centre-right coalition in 2006. In 2014, the SAP formed a minority government in coalition with the Greens (Miljöparti, MP; see Table 2). Over this entire period, there was never a ruling coalition as ideologically divided as the Austrian grand coalitions, and Swedish minority governments were in a strong position to build a parliamentary majority for their preferred legislation. This facilitated a unilateral strategy that would exclude unions from the reform process.
Governments in Sweden, 1991–2017.
M: Moderaterna; C: Centerparti; L: Liberalerna; KD: Kristdemokraterna; SAP: Sveriges socialdemokratiska arbetareparti; MP: Miljöparti.
Hence, the centre-right government between 1991 and 1994 could unite around spending cuts in response to the dramatic economic crisis that Sweden faced. The subsequent SAP minority government could also find support for its preferred policies: with a vote share of 45 percent in 1994, the party was in a hegemonic position and needed support from only one among a diverse set of opposition parties. A consequence of this government strength is that even SAP party leaders have expanded capacity to push back the policy influence of affiliated unions. Between 1998 and 2006, the SAP was in a less powerful position, albeit still dominating the legislative process. The subsequent four-party centre-right government had a majority of seats from 2006 to 2010. The capacity of the four traditionally divided bourgeois parties to formulate a common agenda was a result of their unification into the ‘Alliance for Sweden’ (Allians för Sverige; Aylott and Bolin, 2007). Notably, when this government narrowly lost its absolute parliamentary majority, it did not rely on support from the far-right Sverigedemokraterna, which first entered the Riksdag in 2010; this contrasts with the case of Austria, with the coalition between the ÖVP and FPÖ/BZÖ between 2000 and 2006 (Luther, 2003). We would expect the marginalization of trade unions under the conditions of centre-right unity to the detriment of outsiders.
While the policy preferences of Austrian and Swedish unions were similar, their political capacities to influence the reform process differed markedly. In the next section, I will demonstrate how differences in government strength affected on the capacity of national union confederations to shape the policy output in the interest of outsiders.
Contrasting union capacities to influence policy
Austria: weak governments and union inclusion
In the 1990s, Austria faced the pressing problem of reducing public debt to meet the Maastricht criteria for accession to the euro. With declining economic growth and increasing unemployment, the debt ratio surpassed the Maastricht threshold of 60 percent in 1993. In response, Austrian political actors implemented two austerity packages in 1995 and 1996 (Strukturanpassungsgesetze I & II), cutting the budget deficit from 5.7 percent of gross domestic product (GDP) in 1995 to 1.8 percent in 1997.
Immediately after the 1994 elections, the grand coalition presented a comprehensive package of spending cuts in the interest of fiscal consolidation (Rathgeb, 2017: 49). To facilitate the legislative process, the coalition did not involve the social partners in the policy preparation, for the first time in Austrian post-war history. But mounting disagreements over the design of austerity measures eventually created a reform deadlock in the budget negotiations of 1995. This led to new elections and the subsequent formation of another grand coalition without any significant changes in parliamentary seats.
The most significant reason why unions regained their influence was that the grand coalition failed to formulate a common response to fiscal consolidation. Social partner negotiations played an essential role in brokering intense political conflicts between the governing parties (Heinisch, 2001: 40; Unger, 2003: 107). Union involvement was thus a means of overcoming intra-coalitional divisions and broadening the political support base for controversial policies. As a result, the labour interest organizations remained sufficiently influential to prevent large-scale retrenchment initiatives and win concessions on behalf of outsiders. Proof of the ongoing union influence came with their successful attempt to extend the coverage of the social insurance for quasi-freelancers and the dependent self-employed (Tálos, 1999: 274–275).
Central to the reform agenda of the centre-right coalition under Wolfgang Schüssel (2000–2006) was a comprehensive pension reform, introducing a three-pillar system with increased reliance on occupational and private pension payments. However, the 2000 elections yielded the ÖVP and FPÖ equal numbers of seats, a recipe for deadlock. In 2001, the government announced that it would transform the severance pay scheme into an occupational pension pillar alongside the statutory public pension insurance. The previous scheme provided generous payments for insiders with long job tenure at one private sector workplace, with only 15 percent of annual contract terminations meeting the eligibility requirements (Obinger and Tálos, 2006: 91).
One area of contention was the minimum duration of employment for entitlement to enhanced severance pay (Pernicka, 2003; Traxler, 2001). The FPÖ advocated eligibility from the first day of employment, while the ÖVP demanded strict adherence to the government programme, according to which entitlement to severance pay would commence only after 1 year of employment. The FPÖ also demanded that employees who voluntarily terminated their employment should be covered, which the ÖVP vigorously opposed (Wiener Zeitung, 23 May 2001).
Unable to find a compromise, the government eventually delegated the issue to the social partners. 3 The ÖGB used this opportunity to push for an encompassing coverage of the severance payment scheme (Pernicka, 2001). After long negotiations, the social partners jointly proposed a new scheme, entitling all private sector employees to accumulate individual savings from the first day of employment onwards and across different employment relationships over time. In return, the WKÖ managed to reduce the costs of severance pay by fixing an employers’ contribution rate of 1.53 percent of the gross monthly wage. 4
The new severance scheme (Abfertigung neu) involved stricter regulation of the use of temporary forms of employment because it obliged employers to accumulate severance pay provisions for short-term contracts that had not been covered by the previous scheme. At the same time, it translated into slightly reduced employment protection for permanent workers, who had previously enjoyed higher dismissal payments after long job tenure.
Another point of contention inside the coalition concerned eligibility for unemployment benefits. According to the 2000 and 2003 government programmes, the coalition would introduce an obligation for the long-term unemployed to work in municipal community services in return for continued benefit receipt and state-financed pay top-ups (Ein-Euro Jobs). This closely resembled the German Hartz-IV legislation, but it was eventually blocked by the FPÖ, as the ÖVP’s party whip and chief ideologue, Andreas Khol, told me in an interview. 5 His account was corroborated by another interview with the Head of Department for Labour Market Policy in the Austrian Ministry of Social Affairs. 6 Instead of pursuing a unilateral approach, the government developed activation demands in cooperation with the social partners (Afonso, 2013: 163–164).
However, the 2003 pension reform powerfully demonstrated that union involvement is not an institutionalized constant in Austrian politics. Pension policy is not my main focus in this article, but it provides an illustrative scenario showing that without a divided government, the outcome may not include unions, even in the ‘country of corporatism’ (Traxler, 1998). Government unity in the case of the pension reform rested on two factors. First, Schüssel called new elections in 2002 in response to prolonged internal turmoil within the FPÖ (Luther, 2003) and won an unprecedented victory for the ÖVP, with a swing of 15 percentage points in its favour, while the FPÖ lost two-thirds of its seats. This shifted the balance of power radically, and made Schüssel more determined than ever to pursue his unilateral reform ambitions against organized labour. Second, the introduction of a three-pillar system was an indispensable part of the government’s claim to put an end to the ‘debt policy’ (Schuldenpolitik) of the previous grand coalition and reach a sustainable fiscal situation in the public pension system (Müller and Fallend, 2004: 815). The presentation of the draft bill in March 2003 led the ÖGB to organize mass demonstrations and industrial action – very unusual in the consensual culture of Austrian post-war politics. What was remarkable was that the law took effect not only despite fierce union protests but also against the opposition of the WKÖ and influential figures within the governing parties. Notwithstanding, the government emphatically rejected a joint request by the social partners to present an alternative proposal.
The distributive outcome of the unilateral 2003 pension reform differed from the labour-inclusive reforms observed in the area of labour market policy: it substantially reinforced social segmentation in public pension entitlements by extending the reference period for benefit calculation from the highest paid 15 to 40 years. This came at the expense of public pension entitlements for outsiders with periods of unemployment or parental leave (Pernicka and Hefler, 2015: 44). The labour market reforms, by contrast, lacked any notable departure from the incrementalism of the previous grand coalition and included outsider-oriented concessions to the ÖGB instead (Lindvall, 2010: Ch. 4). Notably, dualistic policy proposals were also on the table of the government’s labour market programme, but the coalition partners eventually lacked the internal cohesiveness to pass this in parliament, which created opportunities for organized labour to influence the reform process.
The SPÖ-led grand coalition since 2007 under Alfred Gusenbauer revived labour-inclusive reform strategies in virtually all economic policy areas. Mutual reform blockages, however, created mounting tensions and culminated in the call for new elections after only one and a half years (Rathgeb, 2017: 52). Nevertheless, the grand coalition legislated a high number of policy changes at this time, which suggests that the social partners somewhat offset the government’s difficulties in carving out policies by assuming a leading role in the formulation of reforms (Tálos, 2008: Ch. 5). In 2007, for example, the unions demanded extended coverage of unemployment insurance and educational leave in exchange for the employers’ demand for tightened availability requirements on the part of registered job seekers (Obinger et al., 2012: 184).
The 2008 election led to the formation of another grand coalition. Immediately after taking office, the SPÖ-ÖVP government under Werner Faymann invited the social partners to design a tripartite policy response to the global recession. The coalition increased the ALMP budget by €400 million (44%) in 2009 to offset the corrosive socio-economic effects of the crisis (Atzmüller et al., 2012: 27 f). As a result, in 2010 and 2011, Austria achieved the third and fourth highest ALMP expenditure rates in the OECD, in proportion to the unemployment rate and GDP (Bundesministerium für Arbeit, Soziales und Konsumentenschutz (BMASK), 2012, 2013). Further upskilling investments came into force in 2013: the qualified employees’ grant (Fachkräftestipendium) and educational part-time work (Bildungsteilzeit; Allinger, 2013).
Sweden: strong governments and union exclusion
In the early 1990s, Sweden faced the collapse of an enormous inflationary bubble, which was widely attributed to the deregulation of credit markets without exchange controls and the stimulus from the ‘tax reform of the century’ (1989–1990). Unemployment rose from 2 to 11 percent, resulting in a budget deficit of 12 percent in 1993.
From 1991 to 1994, the centre-right minority coalition under Carl Bildt mobilized political support for two austerity packages in 1992, which included various social security cuts. These ended the inflation indexation of unemployment benefits, cutting future entitlements by default. Notably, these two packages were supported by the SAP as part of a general transition to a hard currency regime and in response to escalating public deficits.
The 1994 elections raised strong hopes for the main union confederation Landsorganisationen i Sverige (LO). After the short ‘bourgeois’ interlude, the return of the SAP brought the ‘political arm of the labour movement’ back to power. Certainly, formal party-union ties remained relatively strong, given that the SAP derived a considerable share of its votes and funding directly from the LO (Allern et al., 2007). At the same time, the SAP received over 45 percent of the vote, which allowed the formation of parliamentary majorities with the support of only one of the five opposition parties. Therefore, partisanship theorists and trade union representatives alike would have expected labour market policy to shift in the preferred direction of the unions, that is, towards the expanded protection of outsiders (Beramendi et al., 2015).
Yet despite some concessions to the LO, the SAP became undeniably committed to the rapid achievement of ‘sound public finances’ by moderate cuts for the weakest segments of the labour market. After having contributed to the severe economic crisis, the SAP sought to regain economic credibility by vigorously tackling the inflationary public debt crisis through a strategy of orthodox economic adjustment. In fact, thanks to rigorous crisis packages, it turned a budget deficit of 12 percent of GDP in 1993 into a surplus of 2.7 percent by 1997, according to OECD statistics. While the fiscal indicators for this adjustment path look quite impressive, behind these aggregate figures the SAP contributed to rising levels of labour market segmentation by deregulating fixed-term contracts and cutting unemployment support (Thelen, 2014: 173–175).
The politics behind these reforms reveal a great deal about the causal role of government strength. Unlike the reform deadlocks found in Austria, the SAP was in a hegemonic position and could choose support from the opposition party that was closest to its ideal policy preferences. In 1995, therefore, it established cross-class cooperation for the first time with the bourgeois Centerparti to obtain support for unilateral reforms in the areas of job security and unemployment insurance. In both cases, the SAP moved in the preferred direction of the employers, not the unions. First, it weakened managerial constraints on the use of fixed-term contracts, from 1.77 to 1.44 according to OECD statistics. Second, it cut the replacement rates of a number of welfare benefits (from 80% to 75% in unemployment benefits). Aided by an economic recovery and faced with massive union protests, the SAP subsequently agreed to restore the top replacement rate to 80 percent, while extending the qualification period to 6 months at the same time. But it refused to re-index the unemployment benefit ceiling to inflation, thereby effectively ensuring an ongoing decline in benefit generosity.
Moreover, the SAP did not reverse the trend of falling ALMP spending. Notably, the spending cuts observed after 1999 must be placed in the context of a simultaneous rise in fiscal surpluses (Calmfors, 2012). As LO economists complained (Andersson, 2008), the share of participants in ALMPs went down, especially among those unemployed who were not members of a union-run benefit fund and thus in the weakest labour market situation. Transition to a hard currency regime during the late 1980s and early 1990s brought a massive cut in ALMP measured as the percentage of GDP per unemployed. After an increase in the mid-1990s, the spending level once again more than halved within 5 years, declining from almost 0.4 percent in 1999 to 0.14 percent in 2004. Despite another modest increase at the end of their tenure, the spending pattern under the SAP (1994–2006) was sharply downwards.
The SAP achieved the most significant labour market reform of its tenure in 2001–2002, before losing the 2006 elections (Lindvall, 2010: 168). Its autonomous reform capacity, however, was inhibited by a change in the parliamentary balance of power. In contrast to its hegemony from 1995 to 1998, its electoral losses in 1998 prevented another cooperation agreement with the Centerparti. Therefore, the political priority remained the same – ‘sound public finances’ – but the capacity to achieve this in the legislative process was weakened by its reliance on support from the Vänsterparti (Left Party), which rejected another round of cuts (Aylott and Bolin, 2007: 624). The reform raised the benefit ceiling twice and placed higher demands on the active job search of unemployed workers, while abolishing the possibility to re-qualify for unemployment benefits through participation in ALMPs. The rationale behind this reform was based on calculations from experts showing that tightened work obligations would allow for a cost-neutral rise in the benefit ceiling. Still, drawing on the data set of Van Vliet and Caminada (2012), the level of net replacement rates for average production workers fell from 82 percent in 1994 to 72 percent in 2006 because of the non-indexation of benefit levels to wage increases.
The 2006 election victory of the centre-right bloc reinforced the exclusion of organized labour from the policy-making process. It gave the Allians för Sverige a majority of seats for the first time since 1932. This paved the way for a meticulously planned ‘make work pay’ agenda, with cuts in access and generosity of welfare benefits to finance income tax cuts for low-income and mid-income workers as well as payroll tax cuts for employers. The rationale was to enhance work incentives by increasing the gap between wage levels and transfer income, while pitting the interests of workers with stable jobs (insiders) against those at risk of unemployment (outsiders).
Immediately after taking office, the government cut the replacement rate from 80 to 70 percent for 200 days of unemployment and to 65 percent for unemployed with children after 300 days; it reduced the maximum benefit ceiling from 730 to 680 SEK to the 2002 level while extending the reference period for the calculation of benefit levels from 6 to 12 months. Second, the government tightened eligibility criteria in various ways, especially for university graduates out of work. Third, and most controversially, it increased and differentiated the level of fees for membership in the union-run and voluntary unemployment benefit funds (Clasen and Viebrock, 2008: 444–445). As a result, the number of members covered by unemployment insurance fell from 3.8 million to 3.3 million, union density decreased from 75 to 68 percent, while the LO lost 15 percent of its membership within 1 year after the reform took effect (Kjellberg, 2011). Finally, the government doubled the maximum duration of temporary contracts to 24 months and abolished the obligation on the part of the employer to justify the use of temporary work.
The legislation of a reform package of this profound scope within a period of 2 months after its election victory suggests that the Allians was not in any way bothered by the opposition of all three union confederations. In fact, Mats Wadman, Senior Official at the Ministry of Employment from 1988 to 2007, told me in an interview that his department had to push these changes through parliament without accepting the incorporations of any union demands that would have led to a deviation from the initial government proposal. 7 In the 2010 election, the Allians lost its majority despite a modest increase in its vote, as the radical right-wing Sverigedemokraterna entered parliament. If anything, however, the strength of the government in its first tenure was sufficient to cause the significant reinforcement of a dualistic reform path.
Comparing national reform trajectories
In contrast to the traditional image of progressive labour market policies in Sweden, the country now has a more dualized job security framework, grants income support to a substantially lower share of unemployed and spends fewer public resources on ALMP than in Austria.
Swedish governments have radically deregulated temporary employment contracts, while leaving restrictions on permanent contracts almost untouched. In Austria, by contrast, regulations on temporary contracts were enhanced, while protections for permanent employees have been somewhat weakened. These opposing trends, as indicated by OECD data, are clear from Table 3.
Employment protection legislation strictness, 1990 and 2013.
Source: OECD.Stat, EPL database (version 3).
The drastic reduction in ALMP spending in Sweden has already been noted. Figure 1 shows the almost continuous reduction from the late 1980s, whereas spending in Austria has risen from its previous low level. Indeed, per capita expenditure in Sweden has now fallen below the level in Austria.
ALMP spending (% of GDP, divided by unemployment rate), 1985–2010.
Another indicator of dualization is the trend in recipient rates for unemployment benefits and assistance. Table 4 shows that during the Great Recession, the proportion remained high and stable in Austria, while stricter eligibility rules contributed to a substantial reduction in Sweden. The main reason why the unemployment insurance in Austria provides income support to a much higher ratio of unemployed lies in the relatively open access to supplementary unemployment assistance (Notstandshilfe), which is only slightly less generous than the first-tier unemployment insurance (over 90% of the previous benefit level). On average, 38 percent of the unemployed received these benefits from 2008 to 2013; without them, the mean recipient rate would decline to 51 percent. Sweden stands out in terms of unequal access to unemployment benefits. Drawing on a study of the umbrella organization of Swedish unemployment benefit funds, the share of unemployed without social rights to benefit receipt increased from 30 percent in 2005 to almost 50 percent with the onset of the Great Recession in 2008 (IAF, 2009).
Percentage of unemployed receiving benefits and assistance, 2008–2013.
Source: AMS (Arbeitsmarktservice) for Austria, own calculation; Arbetsmarknadsrapporten (2014: 20) for Sweden.
Conclusion
The power-distributional interaction between an inclusive union movement and a politically weak government has been shown to be the dominant explanation for the divergent trends in protection of outsiders from economic uncertainty in Austria and Sweden. In both countries, an inclusive union movement incorporated the interests of outsiders into the producer group arena. But it was the presence of a politically weak government that created the necessary opportunities for unions to achieve policy influence in the legislative arena. In this article, I have, therefore, explored how variations in the strength of governments influenced the success of trade unions in enhancing the protection of outsiders.
The first potential limitation of my argument emerges from its empirical focus on two small West European states with similar corporatist legacies. It might be argued that my analysis lacks explanatory merit for other countries that are non-corporatist and large in size. This objection, however, would ignore the fundamental fact that corporatist legacies were not sufficient for the policy-making influence of organized labour. On the contrary, my argument regarding variations of government strength draws on the social pact literature, which primarily addresses countries without corporatist legacies, such as Ireland, Italy and South Korea (Baccaro and Lim, 2007; Baccaro and Simoni, 2008). In all these cases, national governments resorted to political exchange with unions despite the absence of institutionally entrenched channels of corporatist policy-making. Therefore, tripartite policy-making forums appear less influential for the reform strategy pursued by national governments than in the post-war Keynesian era.
A second potential limitation is more difficult to refute. As emphasized above, my case selection covers only countries with relatively inclusive union movements. Thus, they had stronger incentives to mobilize political support for outsiders than unions in countries with lower levels of inclusiveness. By implication, where unions lack members at risk of unemployment as well as the centralized structure to give voice to precarious sections of the labour force, they might well have a strategic incentive to prioritize the interests of insiders to protect the rights of their remaining core membership. Thus, my argument that union influence is crucial for the protection of outsiders does not necessarily hold in cases with more exclusive labour movements (Gordon, 2015; Thelen, 2014), though recent research suggests that they also use any remaining influence to resist growing inequality in employment and welfare standards (Benassi and Vlandas, 2016; Naczyk and Seeleib-Kaiser, 2015).
My core argument is that political economists should treat government strength as one of the key contextual considerations in future research. To date, it seems that the literature has not sufficiently absorbed this thesis from the social pact literature. But in the age of dualization where partisan left-right differences appear to be in decline (Emmenegger et al., 2012: 311), we must consider not only partisanship per se but also differences in the ability of parties to ‘go it alone’ when we assess government reform trajectories (p. 311). Differences in government strength may, therefore, help us to understand the conditions under which either parties or producer groups drive the politics of labour market adjustment. When governments are internally united and externally encompassing, they are likely to dismiss the policy demands from organized interests to prevent a distortion of their political agenda. This is especially the case when these demands come from trade unions, whose structural power and electoral significance have faded over time. But when they are weakened by internal divisions or a hung parliament, they face a political incentive to reach out to unions because labour-inclusive negotiations provide them with an extra-parliamentary channel of consensus mobilization. This interaction between government strength and union involvement seems to be a crucial and underexplored part of the debate about the political economy of dualization in contemporary capitalism.
Footnotes
Acknowledgements
I would like to thank Hanspeter Kriesi, Pepper Culpepper, Lucio Baccaro, Wolfgang Streeck, Johannes Lindvall, Carlo Knotz and two anonymous reviewers for their generous and helpful comments. This article has also benefitted from stimulating exchanges with Anja Sahora, Fabio Wolkenstein, Max Schaub, Lukas Hakelberg and Sanna Salo.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This article was derived from my PhD research at the European University Institute in Florence, for which I received a Stipendium from the Austrian government.