The Paradox of Liberalization — Understanding Dualism and the Recovery of the German Political Economy



What do the recent trends in German economic development convey about the trajectory of change? Has liberalization prepared the German economy to deal with new challenges? What effects will liberalization have on the co-ordinating capacities of economic institutions? This article argues that co-ordination and liberalization are two sides of the same coin in the process of corporate restructuring in the face of economic shocks. Firms seek labour co-operation in the face of tighter competitive pressures and exploit institutional advantages of co-ordination. However, tighter co-operation with core workers sharpened insider–outsider divisions and were built upon service sector cost cutting through liberalization. The combination of plant-level restructuring and social policy change forms a trajectory of institutional adjustment of forming complementary economic segments which work under different rules. The process is driven by producer coalitions of export-oriented firms and core workers’ representatives, rather than by firms per se.

1. Introduction

Since the rapid changes in non-liberal market economies of the mid-1990s, doubts have emerged about the distinctiveness of the Varieties of Capitalism (VoC) literature as a useful conceptual paradigm (Streeck 2010). There were those who assumed that globalization — in the sense of market expansion, technological diffusion and closer integration — would sooner or later lead to a convergence of political economies. Institutional distinctions were merely seen as relicts from previous stages of economic development which were to be discarded in due course.

Among the more fine-tuned observers, globalization was seen as having an ambivalent effect on co-ordinated market economies. On the one hand, further opening national economies to international trade reinforces economic specialization, thereby making countries more dependent on their comparative economic advantages, and leads to further protection of institutional advantages by firms and economic interest organizations (Franzese and Mosher 2002; Hassel 2007aa; Thelen and van Wijnbergen 2003; Wood 2001). On the other hand, increasing financial internationalization might lead to the opposite effect. The rise of global investors and impatient capital even in countries with protected corporate finance would eventually lead to higher performance expectations and an increase in shareholder value expectations in firms of co-ordinated market economies (Hall and Soskice 2001; Höpner 2001).

In addition, challenges to non-liberal institutions came from other sources. First, policy changes related to the liberalization of product and labour markets during privatization of public utilities and activation of the long-term unemployed might undermine existing patterns of co-ordination. Second, structural changes to post-industrial economies also have repercussions on institutional reproduction. Third, service sector employees and the increase of female labour market participation rates in non-liberal economies might give rise to a different demand for redistribution and social security.

At the same time, there is disagreement in the literature on the interpretation of the empirical evidence. What accounts for a major institutional break? What share of a national political economy must be covered by an institutional pattern in order to classify it as distinct from other economies? Is the decline of membership in trade unions and employers’ associations a sign of decline in co-ordination? Are service sector skills more likely to be general skills? Is redistribution an indicator for co-ordination? Is micro-level co-ordination a functional equivalent to macro-level co-ordination?

Cross-country comparisons show that co-ordinated market economies today are less egalitarian than before, increasingly divergent and their institutions less encompassing (Thelen 2012). They remain different from liberal market economies in the following ways: capital markets remain underdeveloped, labour markets are more regulated, pay setting is still co-ordinated, trade unions remain much stronger and social spending generally higher. Recent discussions about the ‘commonalities’ of capitalism, rather than differences focusing on dynamic trends across all market economies, indicate a departure from the all pervasive theoretical assumptions of the comparative capitalism literature (Streeck 2010).

This article investigates the current avenues of VoC theorizing using the case of institutional change and economic performance in Germany. It questions the kind of contribution recent patterns of liberalization have had on economic institutions and performance and what processes of liberalization alter the configuration of economic institutions.

An analysis of the German political economy's transformation since reunification shows economic shocks have driven plant-level actors to pursue radical cost cutting and productivity increases by exploiting existing patterns of plant-level co-operation. Intensified plant-level co-operation led to employment guarantees for core workers which insulated them from previous demands for strong social security provisions. In turn, persistent outsourcing to low-cost countries and low-cost service sectors has added to liberalization in other parts of the economy, particularly through the use of fringe workers. Manufacturing firms, with the tacit support of their works councils, supported firms in service industries that lobbied for more liberal employment rules for non-core segments of the workforce.1 When the German government pursued activation strategies on the labour market, core firms and core workers did not veto the proposed measure of liberalization (Hassel and Schiller 2010: 123–5; Pastor 2012: 166).

Thus, sustained economic co-ordination has facilitated, and to some extent required, liberalization in some areas for cost containment, more flexible corporate finance and numerical flexibility of the workforce. As a consequence of the benefits of co-ordination, firms actively pursued a strategy of separation of the workforce, which divided employees into core and fringe workers. Liberalization did not occur despite strong resistance by key beneficiaries of social policy, but rather was accepted and supported as a precondition for sustained co-ordination.

Moreover, this article points out that the recent comeback of the German economy owes more to the institutional foundations of the ‘old’ German model than to the liberalizing policies of the early 2000s. Policy tools and firms’ strategies to overcome the crisis were built upon patterns of plant-level co-operation that German firms have pursued for the last two decades. It therefore turns out that co-ordination and liberalization are not opposites or mutually exclusive processes but complementary.2

2. Convergence and divergence in VoC, sources of liberalization and policy

There was a period of convergence in theorizing and researching the various strands of the VoC literature that took place during a phase of relative stability and continuity in advanced industrialized countries between the mid to late 1980s and the late 2000s (Amable 2003; Crouch and Streeck 1997; Hall and Soskice 2001). Many scholars’ analyses of institutional configurations in national political economies strongly emphasized the interdependence between the mode of corporate finance and the innovation and usage of human resources within firms competing in international markets. They concluded non-liberal forms of market economies displayed a number of starkly contrasting features to liberal Anglo-American countries, such as concentrated ownership of firms through block-holding, bank-finance, plant-level co-operation between workers and managers, higher levels and more specific skills in core industries and pathways of specialization in different technologies and industries.

Some level of disagreement on the foundations and origins of diverse economic institutions has always prevailed. Different perspectives emphasized micro- versus macro-level approaches, the use of rational choice assumptions and large n-comparisons with few historical case studies. In particular, approaches focusing on the firm's role as a micro-level actor espousing rationally based preferences and assuming institutional equilibria were in contrast to macro-level studies of institutions emphasizing power resource (PR) approaches in historically unique settings over long periods of time. PR approaches perceived non-liberal economic institutions as a result of the rise of left of centre political parties in co-operation with strong trade unions. These political forces pushed for economic institutions at distinct historical turning points when business was bound (Korpi 2006). Other assessments also emphasized the political struggle, although in a different way. Manow argued for instance that non-market co-ordination was enabled by (conservative) social reforms in the case of Germany and Japan, which responded to the threat of political unrest (Manow 2001). The establishment of status protection mechanisms through social reforms allowed for trust and thereby explained the institutional fit between production regimes and social policy (Streeck 2001: 13). PR approaches identify political struggles and state intervention as the defining levers for institutional configurations.

In contrast, VoC approaches in a rational choice (VoC-RC) tradition see the evolution3 of economic institutions as a self-reinforcing process of firms’ quest for conquering market niches, innovation and productivity. To the extent that non-market institutions provide comparative institutional advantages, governments will be less prone to deregulate them in the face of globalization (Hall and Soskice 2001: 58).

However, since the mid-1990s, advanced political economies have started to display rather strong evidence of institutional change, particularly in continental European non-liberal market economies. Most countries’ governments have implemented reforms of labour market policy (Bonoli 2010), unemployment insurance (Clegg 2007) and pensions (Häusermann 2010), altering the patterns, if not scale, of social spending and the social security position of workers. Labour market regulation was weakened for labour market outsiders by facilitating temporary work, while many regulations were kept for labour market insiders. VoC literature has underlined how the protection of specific skills, inherent to the generous and far-reaching status-securing unemployment benefit systems and strong employment protection, helped workers to invest in specific skills (Estevez-Abe et al. 2001). Government tended to dismantle these provisions when they started to address low labour market participation rates and rising long-term unemployment.

Capital markets and corporate governance regulations have been the subject of intense reform pressure. Beginning in the mid-1990s, many governments liberalized capital markets towards liberal market economies (Culpepper 2010). In some cases, reform was radical and far-reaching, while in others, reform steps were less radical and incremental. Corporate finance shifted slightly towards equity finance, and some large national champions defined themselves as shareholder value firms similar to their Anglo-American counterparts.

Structural changes to the labour market towards deindustrialization and labour market deregulation also weakened the position of trade unions. Union membership figures declined substantially across almost all industrialized countries. Employers’ associations lost members, collective bargaining coverage declined and collective bargaining practices changed. While centralized collective bargaining survived in most places, the contents of collective agreements were less regulated than before and delegated more decision-making rights to a lower level.

Plant-based vocational training, another prominent feature of non-liberal capitalism, declined and a steady trend towards higher and tertiary education lured school leavers away from mid-level specific skills. In other words, the fundamental institutions’ non-liberal market economies were meant to rest upon changed profoundly in the direction of increasing liberalization and deregulation (Streeck 2009).

The literature provides different approaches to the causes, mechanisms and effects of these changes. Earlier contributions pointed out the effects of globalization as a liberalizing force because it increases the likelihood of concession bargaining through firms’ better exit alternatives. However, as Thelen and van Wijnbergen (2003) have demonstrated, though globalization increases the vulnerability of export oriented firms, their dependence on labour has grown, rather than decreased. The effects of capital market liberalization and the rise of shareholder value on the behaviour of large firms has indicated a trend towards more liberal practices, in particular off-shoring and outsourcing (Beyer and Hassel 2002).

On the whole, the discussion moved towards previously unresolved issues, such as the role of the state in modern market economies (Molina and Rhodes 2007), the role of political power relations, and the economic and political preferences of firms towards constraining regulations. These factors became increasingly important for explaining institutional change (Hancke et al. 2007).

As the transformation of coordinated market economies (CMEs) accelerated, the underlying conflict in the literature between PR approaches and VoC-RC approaches reappeared. PR perspectives would see a shift towards liberalization as a strategy pursued by business as a matter of principle to diminish the effects of constraining regulation and trade union demands for redistribution and restricted practices (Pastor 2012). Liberalization would be made possible by shifts in partisanship of governments and coalitions between business and governing parties at the expense of labour.

In contrast, authors using the VoC-RC approach would expect trends towards liberalization arising from conflicting preferences within the business community, such as financial market actors versus manufacturing firms (Hall and Soskice, 2001: 58). Financial market actors in non-liberal systems seeking new sources of corporate finance would pressure management for more short-term profits and therefore faster turnover of staff and production cycles. They would expect much less drive towards liberalization within non-liberal market economies, as long as business interests were well served by existing institutions (Wood 2001).

Both types of approaches, VoC-RC and PR, would acknowledge changes to market constraining institutions in non-liberal market economies can occur in spheres linked to production regimes, but dominated by different political preferences. They would acknowledge policy changes can be driven by actors other than those dominating a production regime. Moreover, there are a number of authors writing about institutional change in VoC who are agnostic between the two. They share the criticism that rational choice-based VoC literature tends to be functionalist, non-historical and lacking a notion of power without subscribing to the full theoretical repertoire of the PR perspective (Hall and Thelen 2009). Taking a micro-level perspective and aggregating micro-preferences to the level of collectively organized interest representation, these authors recognize the importance of political conflict and power relations in a historical context.4

Their explanatory approach model can be depicted as a third alternative focusing on producer coalitions (VoC-PC) composed of firms and workers ready to pursue their interests at the expense of other groups in the market (Carlin and Soskice 2009; Iversen and Soskice 2009; Palier and Thelen 2010; Thelen 2012).

The analytical difference between the approaches is key to understanding institutional change, because it gives an indication of actors’ intentions as well as the intended extent and possible effects of liberalization. From a VoC-RC perspective, liberalization will remain patchy and — largely due to ‘liberal’ influences — stem from financial market industries. From a PR view, business will push for liberalization, even if it comes at the expense of benefits deriving from constraining institutions. A VoC-PC approach would assume that insiders would use their power positions to exploit cost advantages, but would also accommodate liberalizing policy change if it serves their interests.

The claim that producer coalitions are particularly well placed to shape policies in co-ordinated market economies is backed by other research. This theoretical conceptualization has a long tradition in comparative and international political economy studies.5 It has also been implicit in many of the VoC writings (Hall and Thelen 2009).

Empirically, producer coalitions (firms and their core workers) have had privileged access to policy-making arenas in co-ordinated market economies through the self-administration of social insurance schemes, in which unions and employers organizations are represented. Members of parliamentary committees for work, welfare and employment were traditionally affiliated to either unions or employers’ organizations (Trampusch 2004).

Moreover, as Chang et al. (2010) have argued, in countries with proportional representation, electoral systems policies are less likely to favour consumers and more likely to favour producers, since the competition for the median voter is diminished. The more majoritarian the system is, the more pro-consumer the policies are (p. 40).

There is some evidence for the role of producer coalitions in institutional and policy change in the German case. Carlin and Soskice (2009: 93) state that works councils representing skilled workers colluded with management on liberalizing reforms and supported flexible low-level service labour markets for two main reasons: (a) it implied cheaper services and therefore increased the real income of their members and (b) it implied that their members would bear less of the cost of prolonged unemployment. Similarly, Palier and Thelen (2010: 51) point out the dualizing nature of reforms which have protected the status and privileges of labour market insiders relatively well and at the same time provided enough flexibility to stabilize the core.

As a general argument, one might assume that dominant producer groups in co-ordinated political economies are not in favour of upsetting institutions that have served as stabilizing investments in specific skills. Rather, they have utilized proposals for liberal policy changes in accordance with their own preferences and to the disadvantage of other producer groups. Producer coalitions are therefore the best theoretical frame for allowing continued co-ordination and increasing liberalization taking place simultaneously.

In the following, I will use an analysis of the transformation of the German political economy over the last two decades as an illustration and evidence for the importance of the notion of producer coalitions as driving and shaping policy and institutional change.

3. Plant-level competitiveness and the road towards dualism

When unification hit the German political economy in the early 1990s, firms were already under competitive pressure from Japan and East Asia, as well as from an overvalued exchange rate in the European monetary system (EMS). An extraordinary pay hike added to their problems in the aftermath of the unification boom. The subsequent recession in 1992/1993 was the worst since Second World War and saw a loss of half a million jobs in the manufacturing sector. Between 1994 and 2009, the German economy devalued its real unit labour costs in relation to its European competitors by 20 per cent (Marin 2010a).

Throughout the 2000s, real unit labour costs rose slower than Germany's major competitors, including the Eurozone as a whole (Figure 1). Cost cutting was achieved through a combination of plant-level restructuring and policy change, which helped to reduce costs and increase productivity without hurting the skill base and flexibility of the manufacturing workforce.

Figure 1.

Unit Labour Costs, Annual Change. Source: OECD. Data Accessed in 2011.

Off-shoring, particularly to Eastern Europe, greatly increased in the second half of the 1990s (Jürgens and Krzywdzinski 2009). Trade with the new member states of the EU increased from 2 per cent to more than 7 per cent of GDP between 1994 and 2006. During that period, intra-firm trade represented about 21.6 per cent of imports from Eastern Europe. Goods from German subsidiaries in Slovakia and Hungary account, respectively, for 65 per cent and 40 per cent of German imports from these countries.6 ‘In sum, the pattern of intra-firm trade that has emerged between some of the older EU member states and Eastern Europe clearly suggests that off-shoring has become a significant phenomenon for European firms’ (Marin 2008: 4). Some observers have suggested that organizing production by slicing up the value chain ‘has been more important for Germany's lower unit labour costs than German workers’ wage restraint’ (Marin 2010b). According to estimates, German off-shoring to Eastern Europe boosted both the productivity of its subsidiaries in Eastern Europe almost threefold compared to local firms and increased the productivity of German based parent companies by more than 20 per cent.7

In any case relocating production to Eastern Europe made globally competing German firms leaner and more efficient helping them to win market shares in a growingly competitive world market. The efficiency gains from reorganising production were particularly pronounced after 2004 leading to a sharp fall in Germany's relative unit labour costs from 2004 to 2008 (Marin 2010b).

Marin suggests that off-shoring to Eastern Europe has also led to lower wages for skilled workers in Germany:

German firms off-shored the skill intensive part of the value chain to exploit the low cost skilled labour available in Eastern Europe. As a result, the demand for this type of labour in Germany was lower, putting downward pressure on skilled wages in Germany. Hence, off-shoring improved Germany's competitiveness by increasing German firms’ productivity and by lowering its skilled wages (Marin 2010b).

In order to restructure manufacturing plants without facing trade union opposition, management and works councils used the plant-level concession bargaining tool, often coined ‘employment pacts’, introduced by Daimler-Benz in the late 1980s. Almost half of the largest firms in Germany (55 out of 120) negotiated a company-level pact during the 1990s. Within these 55 companies, at least 156 agreements can be found (Hassel and Rehder 2001). Large firms settled agreements aimed at improving the competitiveness of plant, which led to more secure jobs. Both sides compromised: workers accepted pay cuts, longer working time and more flexible working patterns, while management guaranteed investments and promised not to resort to mass redundancies (Hassel and Rehder 2001; Massa-Wirth and Seifert 2004; Rehder 2003; Seifert 2002).

In comparison to concession bargaining in the United States, these agreements were broader and less one-sided. They included measures to improve the infrastructure, training, costs and productivity as well as technology. The workforces of particular plants were rated in benchmarking comparisons and collaborated with local management to make the most profitable bid for investments. Promises by management were not legally binding, but had a reputation for day-to-day relations with works councils.

One important component of concession bargaining was the increasing gap between core and peripheral workers through the out-sourcing process. Collective agreements were adjusted accordingly, in particular by transferring service components into other collective agreements and lower pay.8 Canteens, security and other service components were removed from manufacturing collective agreements and passed on to service sector trade unions and their collective agreements. Terms and conditions for workers in the service components of manufacturing firms drastically worsened, because their pay scales shifted from metal or chemical sector pay to service sector pay.

In the late 1990s, plant-level agreements were reached in one-third of private sector companies. These agreements provide terms and conditions which deviate from the industry-wide collective agreement. Another 15 per cent of companies simply violate the agreements, according to a survey by the union-based Institute of Economic and Social Research (WSI) (Bispinck and Schulten 2003). After 2004, plant-level bargaining was officially recognized and regulated by an innovative collective agreement in the metal sector (Pforzheim Agreement). In 2006, 1 in 10 firms in the metal sector negotiated an official derogation from the relevant agreement (Lehndorff 2010).

However, the price companies paid for plant-level agreements were tightened rules on dismissal protection for the existing workforce, rather than a more flexible regime of hiring and firing. In plant-level agreements, firms pledged to refrain from any collective dismissal for core workers for a period of several years. The flexibility firms gained from concession bargaining was internal co-operation rather than external adjustments. Unions and employers adjusted collective agreements to allow for plant-level deals. They introduced ‘opening clauses’ that allow for local bargaining, provided the business situation is bad. Pay grades became more differentiated and lower pay grades were introduced. Even the trademark 35-hour work week of German trade unionism has been effectively shattered.

Together with their works councils, many companies designed new work arrangements at the plant level. It is virtually impossible for unions to monitor and police violations of collective agreements at the plant level. Very few employees were prepared to sue a company for breaking an agreement, and unions do not have the staffing capacity to enforce or negotiate agreements in small- and medium-sized companies. Rather, firms hoped that competitive pressure, stubborn high unemployment and weaker trade unions would allow them to change the agreements, which would provide them internal flexibility to reduce labour costs and regain competitiveness.

This strategy worked with union co-operation. Unions rarely blocked workplace deals aimed at providing job security and competitiveness, and did not often talk about the deals to avoid other firms from following suit. Protection for the workforce core and the instability for fringe workers (the insider–outsider problem) were complementary to each other. Firms argued the only way to protect core workers was to look for other ways to lower labour costs — at the expense of other parts of the workforce. Flexibility was therefore achieved in an uneven pattern.

Union weakness was expressed in rapidly falling union membership rates (Hassel 2007b) and the failure to rally enough support for industrial action. While the manufacturing unions, particularly IG Metall, were capable of forcing firms to accept union demands until the mid to late 1990s (Bavaria strike in 1995), the strike weapon was seriously impaired by the 2003 Saxony strike when the union badly lost.9

The strengthening of employment protection for permanent employees has been further reinforced by collective agreements in the manufacturing sector which over time included clauses protecting long-term employees from dismissal. Repeated rounds of plant-level concession bargaining, as outlined above, led to higher levels of employment security, at least for some groups of core workers (See also Zagelmeyer 2010). For example, in September 2010, the electronics firm Siemens agreed to a deal with its works council, giving unlimited employment guarantees for almost its entire workforce of more than 120,000 employees (Spiegelonline, 22 September 2010).

4. What dualization? The transformation of the German labour market

Increasing segmentation between core and peripheral employment is partly initiated, partly reinforced by policy changes in employment protection and labour market policies. Employment protection for permanent employees has remained strong, while employment protection for ‘irregular’ contracts (fixed-term, agency and marginal work) has diminished. Over the years, three main types of irregular employment spread in the labour market: fixed-term contracts, temping agencies and low-level part-time employment. Between 1992 and 2007, these three groups increased from previously 6 per cent to 11 per cent of the working-age population (Eichhorst and Marx 2009: 13).10 Firms tended to push for these alternatives as flexibility buffers to protect permanent employment. The move towards opening an irregular employment segment already started during the 1980s, but was greatly intensified during the 1990s and 2000s.

Until 2003, marginal employment had been confined to workers putting in fewer than 15 hours per week and earning less than a low threshold of income as being exempted from social security contributions. Marginal employment status was introduced in the 1960s, when labour markets were tight and employers tried to entice pensioners, housewives and students to take up a few hours of employment without paying contributions. These groups of workers were covered by social insurance through their primary status (as pensioners, spouses or students). Over time, as contribution rates soared, employers increasingly used marginal employment to avoid paying contributions for low-paid jobs. Regulatory changes aimed to increase employers’ tax on marginal employment in order to avoid abuse while retaining the concept of subsidizing marginal employment.11

Both marginal employment and fixed-term contracts are overwhelmingly used by employers in service industries. Only about 10 per cent of marginal employment is in manufacturing, while more than 80 per cent are service sector jobs (Minijobzentrale 2010). The prevalence of fix-term contracts varies across sectors with more than 20 per cent in the service sector and less than 7 per cent in manufacturing (Statistisches Bundesamt 2010). This reflects the demand for non-standard employment in different industries. While manufacturing industries also benefited from the change in policy, service industries depended on them. In particular, part-time and marginal employment was a key policy instrument used to cut service costs.

However, the main catalyst for cost cutting was the change in labour market policy in 2003. While initially driven by the need to curb public spending, activation policies turned out to be a major programme for subsidizing low-skilled employment. Fiscal constraints were the key facilitators for policy change (Hassel and Schiller 2009, 2010). German unification saw unemployment benefits and spending on active labour market measures in the East explode. Contribution rates for unemployment insurance doubled, and subsidies by the federal budget to the labour agency rocketed. Social expenditure as part of total government spending stood at 51.2 per cent in 1991 and increased to 51.2 per cent in 2000 (OECD StatExtracts; data extracted on 27 August ).

When the dotcom boom collapsed in the early 2000s, public finance problems accumulated on several frontiers: unemployment rose again in both Eastern and Western Germany, long-term unemployment accounted for an increasingly higher share of the unemployed, reunification costs rapidly increased, the Stability and Growth Pact started to kick in and tax reforms (the Eichel tax reform) reduced tax revenue, particularly for local authorities.12 All these developments put enormous pressure on the government to restructure social spending.

The Hartz IV welfare reforms cut the maximum duration of unemployment benefits and limited earnings related transfers to the first year of unemployment (18 months for those over 55). The new long-term unemployed benefit (exceeding 12 months of unemployment) was a means-tested flat rate payment and set at what was universally seen as a low level of social assistance (it can be topped up temporarily if a claimant previously received considerably higher unemployment benefits). The reform further introduced major in-work benefits for part-time and low paid employees. Since all transfer recipients are required to take any job offered to them to prove their willingness to work, and since no statutory minimum wage has been set, wages can be set at extremely low levels and be topped up by transfer payments. About 28 per cent of long-term unemployed benefit recipients are employed in work and receive benefits at the same time. In June 2010, this group amounted for 1.4 million employees.13

At the same time, benefit system reforms have not altered the high-tax wedge that burdens low-skilled low-paid work, an obstacle towards a more employment friendly system. Germany remains the OECD country with the highest marginal tax rate for low-paid employment. Social security contributions are set at a proportional rate and kick in at a comparatively low threshold. The reason for non-progressive social security rates is primarily due to the insurance-based welfare state, which draws on employers and employee contributions equally. This is also partly the reason for the high number of marginal jobs described above. While marginal employment is exempted from contributions, full-time employment for low-paid workers is taxed at a rate of 36 per cent (Immervoll 2007). The strong pressure on unemployed to take up low-paid employment and a new system of topping up income with partial benefits create strong incentives for low-skilled workers to take up part-time employment for very low wages and simultaneously draw social security benefits.

As a result of policy change, the trend of declining male employment rates, particularly elderly men, reversed beginning in 2003. Both overall employment rates and the absolute number of people in employment have increased. A study commissioned by the Bertelsmann Foundation in 2010 summarizes: ‘Germany reached a historical high point of employment in 2009 and exceeded other countries’ employment rates. At the same time, levels of inactivity have declined’ (Eichhorst and Marx 2009: 4). Labour market participation of women and elderly workers increased. The employment rate increased by 4 percentage points between 2004 and 2008 and unemployment levels are below average (which is mainly due to short-term working and the effects of the financial crisis). Inactivity is no longer a problem of the German labour market.

However, the structure of the German labour market has dramatically changed in the process. The number of full-time jobs has decreased by 20 per cent, while the number of part-time and marginal employment has drastically increased. The rate of part-time employment doubled between 1991 and 2007 and number of marginal employment rocketed.

The economic upswing after 2005 showed a different trend. From 2006 on, unemployment decreased faster than any other time in post-war German history, from 4.8 million unemployed on average in 2005 to 3.2 million in 2008, the lowest level since 1992. This is more remarkable, because the German definition of the unemployed and ‘able to work’ includes all benefit seekers capable of working more than three hours per day. This is more than 90 per cent of all those who claimed social assistance in 2005.

Eight hundred ninety thousand new jobs, 210,000 full-time and 590,000 part-time, were created during the most recent economic upswing in 2006/2007. In contrast to earlier periods, the share of full-time jobs has increased again. The share of ‘proper jobs’ compared to marginal employment are significantly higher than in previous economic upswings (Koch et al. 2009: 236). Long-term unemployment also decreased faster than in previous periods of economic recovery (Gartner and Klinger 2008: 442). In 1999, 1.7 per cent of economic growth was needed to create additional employment; this threshold now stands at 1.3 per cent (Gartner and Klinger 2008: 445).

However, many full-time positions are now offered as temporary jobs or agency work. Between 2006 and 2007, agency work increased by 64 per cent. The initial pay rate for agency work is 7 Euro and therefore below the rate unions want to see as a minimum wage (Vanselow 2009: 3).

Survey data show that the unemployed are increasingly willing to take jobs below their skill levels, for lower pay and with worse terms and conditions. Those employed are also more willing to accept concessions in exchange for job security (Kettner and Rebien 2009: 6–7).

The downside of the labour market activation is the rapid increase of low pay. Since the mid-1990s, low pay has been constantly rising. Between 1995 and 2010, the share of low paid among all workers shot up from 17.7 per cent to 23.1 per cent (Kalina and Weinkopf 2012: 5).14 In European comparison, Germany and the Netherlands were the only countries in which the share of low-paid jobs increased between 1995 und 2000. In 2000, only the United Kingdom (19.4 per cent), Ireland (18.7 per cent) and the Netherlands (16.6 per cent) had higher shares of low-pay employment than Germany (Bosch and Kalina 2007: 27). The share of low paid has since increased and was only topped in 2005 by the United States (Carlin and Soskice 2009: 77). Low pay is not confined to the unskilled; the share of low-paid skilled workers rose from 58.5 per cent in 1995 to 70.8 per cent in 2007 (Kalina and Weinkopf 2009: 6). Low pay is gradually diffusing into the core of the labour market; whereas it used to be concentrated in atypical work, it is now found in full-and part-time employment and standard jobs. The majority of those low paid are women, though the share of men is rising. In terms of quantity, marginal employment has been the most important form of irregular employment. In 2007, marginal employment accounted for 7 million employees, with the highest concentration in retail (Minijobzentrale 2010).

Fixed-term employment has increased significantly over the last two decades. In 2008, about 2.7 million of the 30.07 million employees were on a fixed-term contract, which accounted for a share of 5.7 per cent (Statistisches Bundesamt 2010). This number excludes trainees and students. Temping agency work also increased rapidly. In 2009, 1.6 per cent of all employees worked for temping agencies, an increase of 53 per cent over 2 years (Eichhorst and Marx 2010). In total, however, fixed-term employment and workers employed by temping agencies still account for less than 10 per cent of the workforce (Eichhorst and Marx 2009: 14).

The increase of irregular and marginal employment must be seen in the context of firms’ attempts to increase or maintain job security for core workers. Governments, employers and unions jointly preferred the deregulation of the peripheral labour market over the deregulation/liberalization of employment protection for the core workforce (Hassel and Schiller 2010: 122).15 As a consequence, the dualism of insiders and outsiders on the labour market has deepened, regarding both the number of workers affected and the degree of regulatory differences.

Overall, the experience is therefore mixed. More employment is combined with low pay and insecure employment. Studies have shown that low paid employment does not usually serve as a path to better paid work. In a survey of 30,000 low paid full-time workers in 1998/1999, only 13 per cent managed to find better work by 2005 (Koch et al. 2009: 249–50).

Activating the long-term unemployed has therefore not solved the structural problems of the German labour market. In 2010 Germany had among the highest unemployment rate among the unskilled in the western world.16 The labour market is increasingly segmented into core and periphery. Underemployment has emerged and women with children work very few hours compared to mothers in other countries. The Hartz-IV reforms have thereby introduced a form of negative income tax (or in-work benefits) in which low paid employment is topped up by benefits. In absence of a minimum wage, these forms of combined income further drive down wages already at the bottom end of the labour market.

5. The effects of dualization and the challenge of the service economy

So far, I have argued that firms reacted to economic shocks by fostering co-operation and off-shoring non-core parts of the production either abroad or by subcontracting to cheaper service suppliers. Government policy has liberalized employment legislation and social policy for non-core workers. Liberalization and co-ordination can therefore go hand in hand, leading to a segmented and dualist political economy. In this section, I will argue that dualism between different segments of the economy is also complementary and mutually dependent. In other words, Germany's competitiveness in manufacturing sectors does not only depend on collaboration with works councils at the plant level but also on liberalization of the service economy. Moreover, the same institutional set up which protects exporting industries helps to liberalize the service sector.

The development of wages in the service economy is one example of how dualization feeds directly into the cost cutting of manufacturing firms. In contrast to other European countries, manufacturing cost cutting in Germany was helped rather than counteracted by service sector pay setting. In many other countries of the Eurozone, pay restraint was achieved in the exposed sectors, but not in the sheltered sectors. Therefore, pay rises in services outstripped the manufacturing sectors.

In Germany and Austria, cost cutting in manufacturing was accompanied by an even fiercer cost cutting in services (Johnston 2009). Over time, service wages fell relatively to manufacturing wages, even though these sectors were sheltered sectors not under international competition. Figure 2 shows the development of service sector wages in relation to manufacturing wages. In all parts of the service sector, wages fell below manufacturing wages. The figure illustrates how hourly pay in finance, insurance and real estate was higher than manufacturing pay, right up until the late 1980s. Since the late 1980s, hourly wages in services have grown even less than in manufacturing despite the persisting pay restraint in manufacturing. In the hotel and restaurant sectors particularly, wages today are less than half than what they are in manufacturing industries. During the 1970s they were at 80 per cent of manufacturing pay.

Figure 2.

Hourly Wage in Services as Share of Manufacturing Wages. Source: Klems Database. Data Accessed in 2011.

The origins of this particularly severe wage restraint in the service sectors are not obvious. Johnston (2009: 26) attributes it to the pattern bargaining of Germany's wage bargaining system:

A similar constraint on sheltered sector wage setters might also exist in countries where inter-industry coordination of wage bargaining remains strong. Austria and Germany provide notable examples: both have pattern bargaining systems where wage-setters in all sectors shadow the metalworking sector. The metalworking sector (IG Metall in Germany and GMT in Austria), leads negotiations, setting wage increases equal to the increase in the national aggregate labour productivity rate. All other sectoral unions then shadow these increases, using them as a target, but rarely reaching them unless their sectoral productivity levels permit it.

In addition to pattern bargaining, which prevents service sectors from catching up with wage developments in manufacturing, liberalization policies themselves have contributed to a wage decline in services. Atypical employment, such as marginal or part-time work are concentrated in the service economy and heavily oversubscribed by women workers (Seifert and Keller 2011). High labour costs have driven retail and restaurant firms to staff predominantly female part-time workers, who work in small workplaces with no union representation. Wages have stagnated or even fallen behind as a consequence of fierce competition in services, low unionization and outsider staffing.

Three more reasons can account for weak service sector pay. First, in the wake of reunification, collective agreements were hastily transferred to the Eastern states. In manufacturing, trade unions ensured that wage levels were at an appropriate level, compared with the West. In many services sectors where unionization was weak, unions settled for very low wages in order to reach an agreement, since many employers were rather reluctant to enter collective bargaining. They thereby established a low wage floor for service sector pay. Second, the absence of a national minimum wage, which prevents wages from falling to incredibly low levels, has contributed to a downward wage drift. Third, a range of subsidies for low wages encouraged service sector workers to accept low wages in exchange for additional transfers. Activation policies, intended to increase labour market participation, came at the price of low wages concentrated in services.

As a consequence, manufacturing and low skilled service sector firms now work under different institutional regimes. Manufacturing is organized around a body of skilled high productivity core workers which is protected against economic insecurity. Low skilled services operate under conditions that are similar to labour markets in liberal market economies.17

The capacity of service sector unions, primarily Verdi, to protect and raise wages by campaigning for a national minimum wage, for instance, is thus severely limited by the opposition of manufacturing unions.18 The same is true for limits on wage subsidies and the creation of social security exemptions for low skilled jobs. Both are accepted and even encouraged by manufacturing unions, whose members benefit from low cost services.

Finally, the wage restraint in manufacturing and wage decline in services are directly linked to weak domestic demand and explain the export dependency of the German economy. The incremental but increasing institutional specialization into a two sector dual economy has led to a development trap of an export dependent equilibrium.

This in turn creates several problems for the domestic economy. One problem is the training young school graduates. Low pay in services has reduced the incentives for school leavers to enter three year apprenticeships in the service sector, especially if the pay for trained workers barely exceeds pay for the unskilled. It also undermines co-ordination in services, as workers do not trust employment security or skill based employment in services. As the economy deindustrializes, this schism will either be reinforced into a dual economy or turn into a major source of conflict due to the service industries’ different skill and pay structure needs. Combined with different types of demand for social security and redistribution by service sector employees (particularly part-time women) with higher turnover, this might turn into the biggest challenge.

6. Germany's recovery in the financial crisis

When regarding the trends of the last two decades, it becomes clear how the German economy owes its recovery to the process of fostered co-ordination in a context of service sector liberalization outlined in the preceding sections.

The German economy was hit comparatively late by the financial crisis. The outlook for the German economy was relatively optimistic until the fall of 2008. The Council of Economic Advisors forecasted 1.8 per cent growth for 2008 (SVR 2008). This supported the government's view that the crisis would be confined to the US and other financial centres. However, the German economy began to shrink in the last quarter of 2008. In early 2009, exports and manufacturing collapsed. By the second quarter of 2009, the German economy had shrunk by more than 6 per cent in comparison to the second quarter of 2008, which registered a more disastrous performance than the countries which had ‘caused’ the crisis (Bogedan et al. 2009).

Since then, the German economy has seen an extraordinary development. It had the steepest decline, followed by the fastest recovery, among all OECD countries (Figure 3). The effects have been most dramatic in the manufacturing sector, given the extreme dependency on the export of manufactured goods.

Figure 3.

Quarterly Growth of Real GDP. Source: OECD. Data Accessed in 2011.

The recovery has been helped by the contribution of the German welfare system's automatic stabilizers and the two stimulus packages, November 5 2008 in the amount of 11.8 billion Euro and on January 27 2009 of approximately 50 billion Euro. In total, the German contribution to global demand stood slightly above the OECD average (Hassel and Lütz 2011). Of these, the German equivalent of the ‘Cash for Clunkers’ programme (amounting to 5 billion Euros) subsidized car manufacturers worldwide. In particular, the cash for clunkers programme protected core skilled workers in export oriented industries.

At the same time, the labour market was relatively protected from the slump's fall-out. The elasticity of employment relative to the gross domestic product (GDP) was the second lowest among the EU (European Commission 2010), meaning the GDP loss did not translate into job losses. As a result, Germany was the only major country which emerged from the crisis with lower unemployment levels than before the crisis (Figure 4).

Figure 4.

Unemployment Rates 2007–2010 (Q3). Source: OECD; Note: 2010 Refers to Q3. Data Accessed in 2011.

As the OECD points out, the single most important explanation for the gap between the business slump and employment outcomes is the reduction of working hours (Lehndorff 2010). Manufacturing firms hoarded their permanent staff by employing various measures: they cut back on overtime, used deposits on working-time accounts, reduced working-time and used the public short-time provisions, which were extended as part of the stimulus package. In total, these measures were used by about a fifth of all firms. According to a plant-level survey by WSI, 30 per cent of all firms used working time accounts in order to avoid dismissals. This was by far the most important adjustment mechanism. Other mechanisms were job rotation (14 per cent), extra holidays (13 per cent) and pay cuts (11 per cent) (Bogedan et al. 2009).

Labour hoarding enabled German firms to rebuild capacity quickly, as demand on world markets picked up. Unlike in liberal market economies, where downswings are immediately translated into redundancy, labour hoarding stabilized demand and protected the skills of the workers concerned. This was also reflected in the change of unit labour costs. While hoarding labour first pushed up unit labour costs in 2009, it decreased in 2010. Labour hoarding was only possible due to the plant-level agreements German firms had negotiated with their core workforce beginning in the late 1980s. They had introduced annual working time accounts, which could now be used to balance working hours throughout the year.

In the midst of the financial crisis, the German economy reported a remarkable recovery of the competitive position of German firms, higher than average growth and the highest employment levels ever (Möller 2010). We can therefore recognize the two components which contributed to this remarkable development: first, German firms used flexible adjustment tools which they had developed over the two decades since the post-unification crisis. Second, public policy, particularly the specific measure of short term working, contributed to employment stabilization during the crisis.

However, analysis of the recovery only briefly mentions the extent to which wage subsidies for the low skilled, the lack of a minimum wage and wage decline in the service sector have served as a cost containing environment, allowing export-oriented firms to contain their wages and unit labour costs. The increasingly dualist nature of the German economy has created an export-oriented high skill industry which depends on a domestic environment of low cost services to control labour costs. This model, which is questionable in its social and economic long-run effects, is specific to the interactions of wage bargaining institutions, social and employment policies and training institutions.

7. Conclusion

Maintaining and regaining competitiveness for manufacturing firms has been a driving force in restructuring the German political economy. Firms have responded to economic shocks by restructuring; while public policy has aimed to accommodate and support manufacturing competitiveness with social policy liberalization which has helped to contain costs in the service economy. Both have contributed to deepening dualism.

Thus, the adjustment trajectory of the German political economy entails continued co-ordination and liberalization. These are not opposites; rather, they are complementary. Sustained co-ordination requires increasing liberalization for the labour market fringe. The result is an increasing inner core of predominantly manufacturing firms who hire a mix and match of core and fringe employees for their plants.

The combination of the two is the most important underlying factor for Germany's recent recovery following the financial crisis. The competitiveness of German firms hinges not only on wage restraint and plant-level cost cutting exercises in the manufacturing sectors, but also on cost cutting service supplies that facilitate wage restraint.

Analysing the complementary workings of co-ordination and liberalization helps understanding the apparent contradictory accounts of the transformation of German capitalism. Moreover, it points to the drivers of change. Both features of the German political economy — continued co-ordination and liberalization of services — are best explained by highlighting the role of producer coalitions in policy and institutional change.

On the whole, the business community has not pressed for wholesale deregulation of the labour market as the PR approach would expect. At the same time, some segments of the business community, representing the core of manufacturing industries, pursued strategies to regain competitiveness at the local level. Management allied with core workers in its quest for productivity increases. Policy changes accommodated these strategies and enabled cost containment in the service sector.

The preceding interpretation of the German case suggests an emerging dualism of co-ordination and liberalization that is based on sector specific cross-class producer coalition preferences. The theoretical framework and evidence implies that this development might be a typical if not stable pattern of adjustment of a co-ordinated market economy to a series of economic shocks. It therefore allows a more nuanced perspective on institutional change in advanced political economies and provides a starting point for further analysis.

As structural changes towards service economies continue and the role of manufacturing firms in national political economies changes, the relative political weight of manufacturing coalitions vis-à-vis service sector actors will shift. It remains to be seen to what extent preferences of high productivity services such as business services will follow the trajectory of co-ordinated market economies or rather opt for further liberalization. In any case, a coalitional approach explains policy change and business preferences better than the theoretical alternatives.


  1. 1

    The empirical evidence for actors’ preferences for welfare and employment regulation is based on a research project about the origins of welfare liberalization in Germany. More than 40 in-depth interviews with key policy makers were conducted including union and employers’ representatives. See Hassel and Schiller (2010).

  2. 2

    See also Herrigel (2010) with a related argument.

  3. 3

    This refers to institutional change rather than the origins of institutions. As Hall and Thelen point out: ‘In sum, although some see the varieties-of-capitalism approach as insufficiently political because it focuses on the ways firms coordinate their endeavours construed in equilibrium terms, it deploys an understanding of institutions that anticipates a lively politics, marked by experimentation, negotiation and conflict, even in cases of institutional stability’ (Hall and Thelen 2009: 14).

  4. 4

    Streeck (2009, 2010) argues that market expansion is an inherent tendency of market societies and is largely disconnected from immediate business preferences.

  5. 5

    In international political economy studies, Ronald Rogowski's work is a key example (Rogowski 1989). In comparative political economy studies, Peter Swenson (2002), Gourevitch and Shinn (2005), Mares (2003) and others have worked on this perspective.

  6. 6

    All data are from Marin (2008: 4).

  7. 7

    See for the general argument OECD (2007).

  8. 8

    See Doellgast and Greer (2007) on the process of outsourcing and renegotiations of agreements in the German telecommunication industries.

  9. 9

    For an analysis of the strike in 2003, see Raess (2006).

  10. 10

    Particularly agency work now comprises 2 per cent of the working age population where as marginal employment increased from 1 per cent to 4 per cent during that time.

  11. 11

    With regard to further regulatory changes, the regulation of fixed term employment has been loosened drastically. Using the OECD scale of 0 to 4, regulatory tightness of fixed term employment was relaxed from 3.5 in 1990 to 0.75 in 2008. Similarly, agency work underwent massive deregulation from 4.0 to 1.75. At the same time, permanent employment was more strictly regulated from 2.58 to 3.0.

  12. 12

    Their revenue from income tax declined from €21.3 to €19.8 bn; revenue local business tax declined from €19.3bn to €15.2 bn between 2000 and 2003. In 2003, 90 per cent of all local authorities in the state of Northrhine-Westfalia were unable to cover their expenditures. See Hassel and Schiller (2010: 181).

  13. 13

    Datensammlung Erwerbstätigkeit und Leistungsbezug nach dem SGB II, 2007–2010, Sozialpolitik aktuell, Universität Duisburg-Essen.

  14. 14

    Low pay was defined as pay below the threshold of 2/3rd of the median hourly pay.

  15. 15

    Hassel and Schiller describe the reaction of works councils in large manufacturing plants to social policy reforms. As an expert emphasized: ‘I noticed at the time from feedback from works councils that there was little enthusiasm to engage for higher social transfers.’ (Hassel and Schiller 2010: 122).

  16. 16

    According to OECD data, unemployment among the unskilled in Germany stood at 17 per cent compared to an EU21 average of 13.7 per cent in 2010. OECD (2011). Table D. Employment/population ratios, activity and unemployment rates by educational attainment, 2009.

  17. 17

    This is reemphasized by the fact employment protection is annulled for minimal offences against employers. A string of court decisions in unfair dismissal cases have held up the view that a shop worker with 30 years tenure can be fired for taking food or minimal amounts of money from the till.

  18. 18

    See on this 2010: 125) and Dribbusch (2004). The position of manufacturing unions on the minimum wage has changed in recent years and manufacturing unions have moved with the mainstream political parties towards favouring a national minimum wage. This was however after pay in services has considerably declined.