MPIfG Working Paper 00/7, Dezember 2000


Wage Coordination and the Welfare State: Germany and Japan Compared[1]


by Philip Manow


Philip Manow is a research fellow at the Department of Politics and Management at the University of Konstanz. Until August 2000, he was a research fellow at the Max Planck Institute for the Study of Societies in Cologne.





Is there a relation between welfare state regimes and national wage setting systems? Peter Swenson in his research on the historical dynamics of the US-American and Swedish welfare state has recently claimed that such a relation does indeed exist. The essay aims to check if this also holds true for the German and Japanese case. In the post-war period both countries have established systems of wage-bargaining that are less centralized than the Swedish system, but in which wages are highly coordinated both within and across sectors, and, subsequently, in which wage compression is relatively high as well. Thus, both countries are confronted with the same problems of wage- and welfare-drift and of firms' exit from the 'solidaristic' or coordinated wage setting that are so typical for Sweden. At the same time the German and Japanese welfare state differ from each other in almost all dimensions. Thus, both cases seem to be ideally suited to provide for a plausibility-check of the Swenson hypothesis. The essay reaches the conclusion that there is indeed ample evidence that both the German and the Japanese welfare state contributed critically to the stability of wage coordination in the era of high growth after World War II. They thus have to be understood as an integral part of the German and Japanese post-war growth model.




Gibt es einen Zusammenhang zwischen Wohlfahrtsstaatsregimen und nationalen Systemen der Lohnkoordination? Peter Swenson hat diese Frage kürzlich in seiner Untersuchung der historischen Dynamik des US-amerikanischen und schwedischen Wohlfahrtsstaats bejaht. Der Aufsatz nimmt sich zum Ziel, Swensons These anhand des deutschen und japanischen Falls zu überprüfen. Sowohl in Deutschland als auch in Japan ist zwar der Grad der Zentralisierung der Lohnverhandlung geringer ausgeprägt als in Schweden, doch die Lohnkoordinierung innerhalb und zwischen den Sektoren ist hoch und führt in beiden Ländern zu relativ starker Lohnkomprimierung. Deutschland und Japan sind folglich mit ähnlichen Problemen der "wage-" und "welfare drift" und der Gefahr eines Ausscherens hoch produktiver Firmen aus der 'solidarischen' bzw. koordinierten Lohnpolitik konfrontiert, wie sie für das schwedische Modell typisch sind. Zugleich sind die Unterschiede zwischen dem deutschen und dem japanischen Wohlfahrtsstaat in so gut wie allen Dimensionen sehr ausgeprägt. Insofern scheint der Deutschland/Japan-Vergleich eine gute Überprüfungsmöglichkeit für die Swenson-Hypothese zu bieten. Der Aufsatz kommt zu dem Schluß, daß in der Tat viel empirische Evidenz dafür spricht, daß der Wohlfahrtsstaat in Deutschland und Japan in der Phase des rapiden wirtschaftlichen Wachstums ebenfalls einen wichtigen Beitrag zur Stabilisierung koordinierter Lohnverhandlungen geleistet hat und somit integraler Bestandteil des deutschen und japanischen Wachstumsmodells gewesen ist.




The wage-welfare nexus


Academic interest in the role of employers in welfare state development has increased noticeably of late (cf. Gordon 1994; Jacoby 1999; Martin 2000; Mares 1996, 1997; Swenson 1991, 1997, 1999, 2000; Pierson 1995). This new literature provides us with many different ideas about why firms should have and actually do have an interest in (certain) social protection schemes, and why employers have historically not always taken an outright hostile stance in the social policy debates of the past. Take the following sample of explanations from the recent literature (Pierson 2000: 793-797): Employers or certain factions of them

  • have an interest in levelling the playing field with respect to national social protection standards for workers (Swenson 1991, 1999),

  • want to prevent worse things from happening and are therefore in favour of social policies that secure their influence on and control over social protection schemes (employers' strategic voting on social policy; cf. Mares 1996, 2000),

  • value the important latent functions of welfare state intervention that contribute to social peace and help secure a high level of labour productivity,

  • 'support social protection that facilitates [the acquisition of] the set of skills they need to be competitive in particular international product markets' (Estevez-Abe et al. 1999: 1; Martin 2000),

  • find extremely beneficial the possibility open to them by welfare state programs to adjust flexibly their workforce to a volatile economic environment,

  • appreciate the compensatory role of the welfare state within a liberal trade regime (Garrett 1998; Ruggie 1997; Rodrik 1997; Rieger and Leibfried 1998).

One of the most ambitious and theoretically sophisticated accounts of the positive role of business in the emergence of the modern welfare state has been provided recently by Peter Swenson (see especially Swenson 1999, 2000). In his comparison of the development of the Swedish and US-American welfare states between 1930 and 1960, Swenson emphasises that whether wage setting was centralised or decentralised had a great impact on the 'regulation of competition' among employers (Swenson 1999: 8), among big and small companies as well as among export-oriented and domestically oriented ones. In this context, Swenson stresses the important stabilising function of different welfare regimes for different wage-setting regimes. His argument, in condensed form, is as follows: The export sector in an economy has to fear that firms in the sheltered sector will give in too easily to workers' aggressive wage demands since these 'sheltered' firms can pass wage costs on to consumers via high prices. Firms competing on international markets have to fear both the negative impact of high factor prices on the prices of their products and the adverse effects on the national wage level exerted by aggressive unions in the sheltered sector. If, however, the exposed industries take the lead in wage setting within a centralised wage bargaining regime, the distributive effects differ substantially. Now the most productive and bigger firms pay 'below equilibrium' wages and thus enjoy a competitive advantage. For the economy as a whole, but not for each and every firm, wages are set below what would have been the outcome in a decentralised setting (Moene and Wallerstein1995). Yet, low-productivity, labour-intensive firms have to pay above equilibrium wages.


If labour is generally underpriced, the demand for labour increases and labour becomes scarce. Under conditions of labour scarcity, larger and highly productive firms, able and willing to pay higher wages, are constantly tempted to overbid centrally set wages, thus producing wage drift. According to Swenson, the introduction of generous and industry-wide welfare programs set up by the state can discipline employers and therefore help stabilise a centralised wage-bargaining system, which would otherwise erode given the evasion strategies of single employers. This is the reason why employers' associations, which usually are dominated by the very same companies that profit most from coordinated wage bargaining, often prove to be at least tacitly supportive of welfare state expansion. This is especially the case if entitlements benefit workers, blue- or white-collar, and are introduced in boom-periods when labour scarcity becomes a particularly pressing problem.[2] The employers who benefit from centralised wage bargaining 'acquire interests in policy designed to protect them against the results of disequilibrium' that wage centralisation itself creates (Swenson 1999: 10). Peter Swenson has provided ample empirical evidence that indeed this explanatory model can well account for the often tacitly and sometimes openly supportive position of Swedish employers in the debates over post-war social policy in the 1950s and 1960s – as well as their very aggressive anti-union, anti-welfare stance between 1900 and 1930.


However, the establishment of centralised wage bargaining presupposes the ability of employers to effectively cooperate among themselves, in particular to force unions into national wage negotiations with massive lockouts and to delegate bargaining responsibility from regional units to the national confederation. Where these preconditions were not met, employers could not help but find individual, decentralised solutions for the management of their labour markets. According to Swenson, this explains the US-American trajectory of welfare capitalism and labour market segmentalism. Numerous impediments, legal and otherwise, to both within-sector and across-sector employer coordination have left business coordination on issues of wages and employment-based 'social' benefits out of reach for American companies. Within this decentralised, non-coordinated setting, firms with production technology that demands high workplace skills and thus low labour turnover pay 'above equilibrium' wages, making employment scarce. If employers voluntarily 'overprice' labour, they create excess labour supply. The labour market does not clear, unemployment persists in equilibrium (for an overview of the large literature on efficiency wages, see Weiss 1990).


Part of employers' efforts to stabilise a core workforce is to offer better pay, more job security and more generous company-based social benefits than the average firm. They thus create internal labour markets and dualistic wage dispersion, and contribute to the emergence of a segmented industrial structure with a few large 'high-commitment' firms and many small 'low-commitment' firms. Efficiency wages and welfare capitalist benefit schemes, however, come under stress during recessions, when low wage-low benefit firms without any commitments to their workforce begin to challenge bigger firms with cut-throat price competition. In this situation it is often far too costly for big industry to revoke past commitments and to risk the destruction of the trust and loyalty that has slowly been built up over time and is so critical for the production technologies they use. Instead, employers often begin to support the extension of high social protection standards and generous social entitlements to the entire economy, thus forcing upon their less-committed competitors the same costly social protection standards. This, according to Swenson, can explain why the so-called 'corporate liberals' of large US-American companies turned out to be supportive of Roosevelt's New Deal legislation in the depression of the 1930s (Swenson 1997; Gordon 1994; Jacoby 1999), while Swedish employers came to back the Social Democrats' agenda for social policy reform not during times of economic slump, but during the high growth period of the 1950s and 1960s. Swedish employer solidarism and US-American employer segmentalism reveal the different business-cycle responsiveness of centralised and decentralised wage-setting regimes, which can account for the peculiar asynchronous development of welfare policies in these two countries. Once the long post-war growth period eased the competitive pressures on US big industry, the welfare capitalist trajectory became once again attractive for the major industrial enterprises.


Is this fascinating explanation of the US-American and Swedish welfare state trajectories applicable to other cases? How would it accommodate the German or Japanese cases, both countries which combine a relatively high degree of wage coordination and wage compression (see Figure 1) with a welfare state so different from the Swedish model and, what is more, so different from each other? Japan's firm-based welfare schemes resemble in many respects the US-model of large employer welfare capitalism (Esping-Andersen 1997), whereas business coordination and wage compression are so prominent in the Japanese political economy that Japan, like Germany, is often treated as an outright counter-model to the market liberalism of the Anglo-Saxon variant (Albert 1991; Crouch and Streeck 1996; Streeck and Yamamura 2000; Soskice 1995). Hence, we may harbour doubts whether the neat fit between wage regimes and welfare regimes so convincingly described by Swenson for the US-American and Swedish cases indeed can be generalised to other cases.



Figure 1 Wage compression and centralisation of wage bargaining, 1979-1995




Note: Wage compression is defined as the ratio between the median (D5) and the lowest (D1) income decile. Centralisation of wage-bargaining according to Iversen (1998: 58): S wjp[2]ij, where wj is the weight accorded to each bargaining level j (S wj = 1) and pij is the share of workers covered by union (or federation) i at level j.
Sources: Iversen (1998, 1999); OECD Employment Outlook, various years.



In this paper I consider whether Peter Swenson's historical account of the Swedish and US-American welfare state development can tell us anything about what appear at a glance to be deviant country cases. In particular, the early establishment of collective wage bargaining in Germany between 1956 and 1963 is compared with the establishment of collective bargaining in post-war Japan. I highlight especially the parallel establishment of collective bargaining in German manufacturing with the post-war reconstruction of the Bismarckian welfare state, and the correspondence between wage coordination Shunto-style (Takanashi et al. 1996) and the Japanese variant of company-based social protection. I argue that the parallel establishment of wage regimes and welfare regimes in Germany and Japan during a high growth period was not accidental, but backs Peter Swenson's claim that employers develop a positive interest in social policy if social policy promises to stabilise wage coordination and to lower employers' 'collective-action' costs in times of economic upswing.[4] In the last section, I briefly summarise the argument and discuss some of its broader implications.



Why then? Full employment, domestic demand, and wage moderation in Germany


In Germany, coordinated wage bargaining and the leadership role of the metalworkers' union (IG Metall) in the annual wage rounds is said to have been first established in 1956 with the so-called 'Bremen agreement'. Coordinated wage bargaining then ultimately was established in the wake of the 1963 conflict in the Baden-Württemberg metal industry (Noé 1970). We may ask with respect to the establishment of pattern wage bargaining in Germany: why then?


In 1955, the German Wirtschaftswunder reached its first post-war peak. GDP growth achieved a stunning 11.9 per cent. For the first time since 1945, the labour market was almost swept clean, with unemployment fluctuating around 3 per cent during the summer and autumn. In absolute numbers, Germany counted less than 500,000 unemployed in September 1955 as compared to more than 17 million in gainful employment. Due to the inflationary dangers of high growth under conditions of full employment, the German central bank, the Bank deutscher Länder (since 1957 Deutsche Bundesbank), for the first time embarked upon a restrictive monetary policy, even though in 1954 the inflation rate had still been low, only about 2 per cent. Since Germany's current account remained relatively stable (but positive), the boom in the second half of the 1950s was now driven at least as much by domestic demand as by external demand (Giersch et al. 1992: 63-5). In the second half of the 1950s, Germany witnessed the onset of consumerism. Germans began to spend the incomes they earned during the Korea-boom on consumer goods. They started to buy cars, television sets, washing machines, etc. While the Korea-boom in first half of the 1950s allowed heavy manufacturing to prosper, the second half of the 1950s saw the rise of German consumer good industry.


Yet, in spite of de facto full employment and despite the union's official announcement to embark upon an expansionary wage policy (Victor Agartz's expansive Lohnpolitik), which was designed to redistribute wealth via wage demands above the level set by productivity growth plus inflation, real unit labour costs constantly fell throughout the 1950s, if less so in the second half of the decade (see Table 1).


Table 1 The determinants of West German labour costs, 1950-60 (annual average rates of change in per cent)

Earnings[a] Labour productivity[b] Inflation[c] Real unit labour costs[d]

1950-51 15.8 7.0 11.8 -3.2
1951-52 8.3 6.9 4.7 -3.3
1952-53 5.9 6.0 -1.3 1.2
1953-54 5.0 4.7 -0.3 0.6
1954-55 8.1 7.8 2.1 -1.8
1955-56 8.0 4.6 2.7 0.6
1956-57 6.6 3.6 2.8 0.0
1957-58 6.9 3.5 2.9 0.4
1958-59 5.6 6.6 1.6 -2.5
1959-60 9.0 6.8 3.2 -1.0
1950-55 8.6 6.5 3.3 -1.3
1955-60 7.2 5.0 2.6 -0.5
1950-60 7.9 5.7 3.0 -0.9

a Gross annual earnings (including social insurance contributions of employers).
b Gross domestic product (at constant prices) divided by active labour force.
c Deflator of value added (i.e. gross domestic product).
d Earnings / (Labour productivity x Inflation).
Source: Giersch et al. (1992: 72).


Giersch et al. consider four different explanations for the decline of real unit labour costs throughout the 1950s: 'organisational weakness', 'political distraction', 'economic surprise', and 'social responsibility' (Giersch et al. 1992: 73-9; see also Flanagan et al. 1983: 216-75). They dismiss the weakness argument for very good reasons (see also Paqué 1994), but reject the political distraction argument for less convincing ones. This latter argument holds that the political struggle especially for co-determination in the early 1950s meant that unions stood 'still temporarily at the wage front so as not to endanger the far more daring prospect of a thoroughly syndicalist "industrial democracy"' (Giersch et al. 1992: 75). While it is true that the conflict over co-determination (Mitbestimmung) and over work councils (Betriebsverfassungsgesetz 1952) had ended by 1953 and thus cannot explain the unions' continuous wage moderation in the following years, Giersch et al. fail to see that the conflicts over weekly working time (1956), over sick pay (1956-7), and more generally over social reform and the reconstruction of the German welfare state (pensions in 1957 and — less so — health reform in 1959-61) had become the main issues in union campaigns in the second half of the 1950s and had replaced the emphasis on redistribution via generous wage hikes (Pirker 1979; this omission is especially clear in Paqué 1995). These social policy issues were not only of general political relevance in the unions' manifestos, but especially the pension reform and the sick-pay law in 1957 had — in contrast to the conflict over co-determination — a direct impact on workers' (social) wages.


In this context it is also regularly overlooked that the Bremen agreement on reducing weekly working time from 48 hours to 45 hours was 'only part of a more encompassing social policy concept' of the IG Metall (Kalbitz 1978: 183) that carried over into the fierce fight for sick-pay reform in Schleswig Holstein's metal industry in November 1956. I would argue that the mobilization of the unions for the improvement of social protection in the second half of the 1950s made them trade real wage increases for an expansion of welfare state entitlements (cf. Eichengreen 1994), or to put it differently: especially the unions and the employers in vulnerable, export-oriented sectors were happy that the rise of the social wage due to the significant welfare state expansion in the second half of the 1950s reduced the wage push in times of de facto full employment. At the same time, the emphasis on social issues in the IG Metall campaigns almost 'naturally' established its leadership role within the unions' national confederation, the Deutsche Gewerkschaftsbund (DGB). A closer look at Table 1 reveals that the years 'in between' the struggle for co-determination and the fight for pension reform and sick pay, the years between 1953 and 1957, were the ones in which real unit labour costs showed an upward tendency.[5]


What was the pre-history of the 1956-7 conflict over working time reduction and sick pay? The IG Metall had entered the 1955 annual wage round by cancelling all collective contracts in the North Rhine-Westphalia district, by far the largest of all districts with 900,000 IG Metall members. Interestingly, the union sought to reach a different settlement for the iron and steel industry than it did for metalworking. The union wanted a moderate settlement in the more export-oriented, small-firm dominated metalworking industry and was quick to accept a moderate arbitration proposal. At the same time, the metalworkers' union hoped to achieve a higher wage hike in the iron and steel industry, where demand was more domestic in nature and where firms were much bigger (Pirker 1979: 174-75). This strategy backfired badly (Pirker 1979: 175-76). It failed in particular because the leading union representatives in the big firms of the iron and steel sector considered the IG Metall's wage demands to be out of proportion and did not back them. Especially the Arbeitsdirektoren, union members who belonged to the board of directors in the coal, iron, and steel industry thanks to the co-determination law (Montanmitbestimmung), opposed IG Metall's aggressive wage policy. The union thus was faced with a dilemma: it could not champion the cause for higher wages in sectors where firms were under strong competitive pressures from world markets. Yet the union also met with considerable resistance in those sectors where demand was more domestic in nature but where co-determination had infected union representatives with the virus of firm-level syndicalism or social partnership. All this coincided with workers' raised expectations concerning wage increases given the metalworkers' union's radical rhetoric (in times of very tight labour markets).


While union representatives in the firms were often reluctant to embark upon IG Metall's aggressive wage policy, union members often had become more radical, as is shown by the wildcat strikes in 1955 in which from time to time even work councils and (communist) workers violently clashed (e.g. during the strike at the Howaldt shipyard in Hamburg, August 1955; see Pirker 1979: 177). The IG Metall's first response to this situation was to delegate bargaining responsibility to the firm level. But this was an even riskier strategy given the workforce's radical leanings in many firms (Pirker 1979: 176, 179). Furthermore, IG Metall's claim for leadership within the DGB was hardly made more credible by the union stepping back from its former wage-bargaining responsibility and delegating bargaining to the firm level, in which case either radical (the much feared communists) or accommodative forces (believers in firm-level Sozialpartnerschaft [social partnership]) could undermine the union's position.


The IG Metall, under the leadership of Otto Brenner, who was part of the 'progressive wing' in the DGB, finally 'was condemned to act' (Kalbitz 1978: 184) if it did not want to lose out against the more moderate faction within the DGB. Aggressive mobilisation in support of significant welfare state expansion in the following years then became the solution to IG Metall's dilemma, as the Bremen agreement and in particular the Schleswig-Holstein metal worker strike in 1956-7 over sick-pay reform showed. It was to become soon clear that employers in the 'exposed sectors' of the economy favoured welfare state expansion as well because they hoped that this would weaken the upward pressure on wages in times of full employment. But the interest in short-term wage restraint was only the employers' initial motive to give up the fight against the substantial welfare state expansion in the second half of the 1950s. Employers still had to learn of the beneficial side effects of the Bismarckian welfare state on wage coordination. But insofar as the welfare state was going to become essential for the stability of medium-centralised collective bargaining and the ensuing intra-industrial wage compression, welfare state programs were more and more perceived as being in the general interest of all employers, be they in exposed or sheltered sectors of the economy (see below).


What was the prime economic rationale behind pattern wage bargaining? The large firms in the Ruhr region knew that the unions had picked them in the 1955 wage round as their prime targets in the struggle for higher wages, even though the 'dual track' strategy of IG Metall had been a complete failure. While the big firms in the west clearly could afford to pay higher wages, they were not particularly eager to do so. Hence, the 'autarchic' firms in the Ruhr region (Herrigel 1996) would most benefit from collective bargaining for the entire industrial sector with the export industries setting the pace, since this would lead to wages significantly below what they would have to pay if unions negotiated firm contracts (Swenson 1999, 2000). If, in turn, wage moderation in the export-oriented firms would be achieved by an increase in workers' social wages (via welfare state expansion), those exposed firms and workers most vulnerable to external shocks would profit disproportionately if the welfare state would provide employers with instruments to flexibly and painlessly adjust employment to sudden shifts in demand, meaning with social programs such as short-time allowance (Kurzarbeitergeld) and early retirement programs. That was exactly what Adenauer's pension reform in 1957 delivered.[6] 'Wage redistribution' in favour of the large firms in the coal, iron, and steel industries would be counterbalanced by 'welfare-redistribution' in favour of those firms that used (contribution-financed) welfare state programs to adjust their workforce to sudden shifts in demand or changes in market conditions.


Insofar as industry-sector wage bargaining with a compressed wage structure was thus externally stabilised by welfare programs, the welfare state became an indispensable part of Germany's post-war economic order, both in the sense that the political mobilisation for the reconstruction of the Bismarckian welfare state temporarily helped moderate unions' wage demands, and in the sense that the welfare state, once in place, helped compensate workers from those firms that either were no longer productive enough to afford the higher social wage or had to cope with adverse external demand shocks. The whole arrangement roughly resembled the famous Rehn-Meidner model: compressed wages exerted constant pressure on firms to increase productivity. Uniform wages within the sector resulted in redistribution from the less productive to the more productive firms. But this redistribution was balanced by a flow of welfare-income in the reverse direction, which was spent to support those workers who lost their jobs because firms went bankrupt or because of external demand shocks. Unlike social spending programs in Sweden, however, those in Germany were passive, not active, in nature. A prominent element of the Bismarckian welfare state was to reduce the labour supply via early retirement, disability pensions, or to allow for long waiting periods in case of unemployment and to provide many opportunities for retraining.[7] The passive nature of welfare state programs was due to the fact that Germany possessed no central location where policy coordination and side payments across industrial sectors, and the national coordination of wage, monetary, and social policies could have taken place. Wage bargaining remained moderately centralised and took place at the industrial sector level. No national employer association or union umbrella organisation enjoyed, as in Sweden, sole bargaining responsibility. Moreover, the prime problem for the relevant economic agents was to coordinate interests within an industrial sector, namely between iron and steel and the metalworking industries, not between different industrial sectors.[8]



Why there?


German shipbuilding, world markets and the intra-industry compromise

Both the Bremen agreement and the Schleswig-Holstein metalworkers' strike of 1956-7 were meant to lend credibility to IG Metall's claim to be the leading, pace-setting union among the sixteen DGB unions and to be also fully autonomous from directives of the unions' umbrella organisation. In Bremen and Schleswig-Holstein, the metal sector was dominated by shipbuilding. In the Schleswig Holstein metal sector, 38.9 per cent of all workers were shipyard workers, by far the largest occupational group, but many more metalworkers were employed in the supply industry. A closer look at shipbuilding's economic situation can explain why so much union activity occurred here in the mid-1950s.


In the early 1950s, the demand for ships was largely domestic. The rebuilding of the German trade fleet, made possible by generous tax subsidies and by Marshall Plan funds (Wend 2000), kept employment and production levels up during the first half of the decade. Recovery had been slow for many reasons: the destruction caused by war (which was substantial in shipbuilding compared with other industries), the 'deconstruction' after the war, an insufficient supply of steel for the shipyards, and the military restrictions placed on production that were only slowly lifted in the wake of the Korea-boom. If the year 1936 is used as a yardstick, shipbuilding had only reached 72 per cent of its pre-war production capacity by 1951, while at that point coal mining was at 103 per cent, the car industry at 166 per cent and the textile industry at 132 per cent of their pre-war levels. Industry as a whole had reached 136 per cent of its 1936 capacity level (Albert 1998). With time, external demand became increasingly important for German shipyards. Already 48 per cent of all ships built between 1952 and 1955 had been exported, and this figure increased to 65 per cent between 1956 and 1961 (Albert 1998: 85). As early as 1955, German shipyards were the world's second most important producer of ships (Kuckuk 1998: 20), and shipbuilding had the highest export share among all German industries (Kuckuk 1998: 22), having reached 77.7 per cent by 1960 (Kuckuk 1998: 22). This sector thus was especially sensible to changes in the international terms of trade and in domestic wage costs, particularly since wage costs were comparatively high in shipbuilding (between 20 and 35 per cent of total production costs; Albert 1998: 107) and possibilities to substitute labour with capital were limited.


Shipbuilding was the seventh most important export sector if measured in terms of its contribution to the German trade surplus. Yet, terms of trade had eroded steadily since the mid-1950s and suddenly deteriorated rapidly in 1957. While the closing of the Suez canal had led to a sharp increase in the demand for ships and in freight rates, in early 1957, freight rates dropped significantly and literally overnight shipyards received no new ship orders. International competition turned extremely fierce, a development long foreseeable due to the advent of a new and competitive challenger: Japan. Japanese shipbuilders had increased their world market share significantly within only a short time period, from 11 per cent (1951-5) to 32 per cent (1956-60; see Albert 1998: 85). In 1958, German shipyards started to lay off workers.


Another important factor to take into consideration is that by the 1950s all of the major German shipyards had become vertically integrated into the 'autarchic' Ruhr region's iron and steel industrial complex (Leckebusch 1963). Since 1916, Thyssen held the majority of the stocks of the Bremen Vulcan shipyard. Krupp had holdings in the Deutsche Schiff- und Maschinenbau AG, a merger of Vulcan Werke Hamburg and AG Weser, the latter was acquired entirely in 1941 (Chandler 1990: 511). The Germania shipyard had been part of the Krupp concern since 1896. Thyssen bought Blohm & Voss in 1955. Gutehoffnungs-Hütte had acquired Deutsche Werft back in 1918 (Bankverein 1954). Stinnes owned Nordseewerke Emden and the Stumm conglomerate owned the Frerichs shipyards (Leckebusch 1963). Vertical integration was supposed to help balance the business cycles via diversification (see Albert 1998: 124; Strath 1994: 75). What had been the case following World War I became relevant again after World War II: steel producers foresaw the necessity to find new markets given the significant expansion of their production capacities during wartime. Yet the vertical integration between the iron and steel industry and the ship construction industry, which was barely affected by the Allied deconcentration policy after World War II (Diegmann 1993; Wend 2000), meant that managers and unions had to take into account the different economic situations of both of these industries if wages were to be set uniformly for the entire 'metal' sector. Furthermore, it is important to note that the extensive version of German co-determination was practised in the iron and steel industry (Montanmitbestimmung), but not in the shipyards.


Shipyards were highly unionised even by the standards of the time. In all of Schleswig Holstein, union density in the metal sector was 71.9 per cent in 1956 (IG Metall 1978: 51), but at the shipyards, the best organised industry second only to the electrical industry, union density was even greater.[9] Workers were also highly politicised, as was underlined by the fact that many shipyards had been occupied by workers immediately after the war, resembling the Japanese post-war experience of unions' 'production control' (Gordon 1985: 329-66). The communist movement enjoyed a relatively strong foothold. In the 1956 work council elections at the Kiel Howaldswerft, the communist candidate won 50 per cent of the votes (IG Metall 1978: 208). Evidence of the radicalism of many shipyard workers is found in the extraordinary duration of the Schleswig Holstein strike, where even after fourteen weeks of strike union members still rejected a negotiation result with 76.2 per cent of the vote although it was endorsed by their own union leadership and where it took four strike ballots before union members finally accepted the settlement worked out between their union and the employers (Kalbitz 1978).[10] Work councils, on the contrary, had mostly pursued a policy of Burgfrieden, so called after the domestic 'armistice' between the state, the military, and the workers' movement during World War I. Given that external factors, namely the capital shortage and the Allied restrictions on production, were seen as the main obstacles towards the post-war recovery of the German shipbuilding industry, management, work councils, and local politicians formed broad coalitions and considered it to be in their common interest to put internal conflicts aside for the time being (Wend 2000). This was the shipbuilding industry's economic situation in the second half of the 1950s, when major wage increases triggered by a growing domestic demand were expected to endanger the industry's precarious international competitiveness, while a highly mobilised union membership did not want to wait any longer for its fair share of Germany's Wirtschaftswunder. As is argued in this paper the mobilization of the metalworkers' union for welfare improvements provided a solution for its dilemma: it had to avoid large wage hikes but at the same time respond to the high expectations and radical leanings among the shipyard workers in times of high growth and low unemployment. Rallying for social issues at the same time helped establish the pace-setting role of the IG Metall and was a first step towards wage coordination. Employers began to learn about the beneficial effects of both wage moderation and wage coordination.


Cartels, coordinated capitalism, and collective wage-bargaining: the metal-producing vs. metalworking industry

Whereas the Bremen agreement and the metalworkers' strike in 1956-7 represented the first steps towards introducing pattern wage bargaining, coordinated wage setting finally became fully established in the south, in the industrial region of Baden-Württemberg where mechanical engineering, automobile firms, and the electro-technical industry dominated metalworking. Again we have to step a little back in time in order to explain the rationale of wage coordination for the metal industries.


A principal problem of Germany's dual industrial structure, namely the conflict of interests between large, domestic-oriented iron and steel producers and medium-sized and export-oriented steel consumers (Herrigel 1996) that in shipbuilding was solved in the 1920s via vertical integration, had historically found a different solution in the other metalworking industries, especially the mechanical engineering, automobile, and electrical industries. While quickly gaining in economic importance, the metalworking industry had been unable to acquire the corresponding political 'interest-group influence' until the end of 1924 because of seemingly insurmountable 'collective action' problems (Feldman and Nocken 1975).[11] The lack in lobbying power was due to the profound heterogeneity of interests within the metalworking sector. For one, metalworking firms were smaller and more numerous, so that economic concentration was far less developed than in the German heavy industries of coal, steel, and iron (Chandler 1990: 550-61; Weisbrod 1978: 93-119). For another, the wide span of products ranging from automobiles and machine tools to cutlery and the corresponding heterogeneity of product markets hindered effective interest aggregation and organisational unity.


When the metalworking firms finally managed to organise themselves into the Arbeitsgemeinschaft Metallverarbeitende Industrie (AVI) in 1924-5, the way was clear for settling the serious conflict with heavy industry via a complex contractual arrangement between the AVI and the Association of German Iron and Steel Producers (Reichsverband Eisen und Stahl), the so-called AVI-agreement (Nocken 1977; Weisbrod 1978; Feldman 1977; Hoffmann 1928). The core of the compromise between both sectors was that, on the one hand, the metalworking firms would support heavy industry's campaign for protectionist tariffs on iron and steel once the Versailles treaty provision expired in 1925 that granted the victorious nations access to the German market on a most-favoured-nation basis; while on the other hand, heavy industry agreed to compensate the metalworking firms for the difference between the high German cartel prices and the world market iron and steel prices for exported products (and they supported efforts to obtain privileged access for these products especially to the French and Belgian markets). It would clearly go beyond the purpose of this paper to describe in any detail the working of this sophisticated inter-sectoral arrangement (Hoffmann 1928; Schmid 1930; Tübben 1930; Preiss 1933; Blödner 1934). Suffice it to say that this form of conflict settlement via cartel price control and inter-sectoral price compensation clearly could not be revived after 1945, when the Allied Entflechtungsverordnung (deconcentration decree) was issued, nor later, in 1957, when the passage of the Cartel Law (Gesetz gegen Wettbewerbsbeschränkungen) forestalled the further use of these extreme practices of Germany's 'organised', 'guaranteed', 'riskless' capitalism so prominent in the interwar years (see Winkler 1974; Michels 1928; Liefmann 1938). Yet, the AVI-agreement continued to be important after World War II as business leaders reminded themselves of this successful example of inter-industry coordination. In addition, the complex but highly effective cartel agreement between the two metal sectors preserved and even reinforced the dual structure of the German metal industry, since the agreement not only halted the trend towards further vertical integration, especially the integration of the metalworking industry into the huge industrial complexes of Stinnes, Thyssen, and Krupp, but even partially reversed the trend (see Feldman 1977; Weisbrod 1978; Chandler 1990).[12] The Nazis, who had extended the cartel agreements to the entire metal industry shortly after their advent to power, thereby handed down to the young Federal Republic the metal sector's dual industrial structure in an almost unaltered form (Fear 1997).


After 1945, cartelisation was no longer available as an instrument of economic coordination. 'Business coordination' of the kind embodied by the AVI-agreement clearly was ruled out by the cartel law, even if the law represented only a considerably watered-down version of what economics minister Ludwig Erhard and his ordo-liberal combatants initially had sought to achieve (Hentschel 1996; Nörr 1993). Moreover, the profound deconcentration measures imposed by the occupation forces on heavy industry (Herrigel 2000: 363-9) rendered the agreement on and enforcement of cartels a highly unlikely enterprise. Yet, the fact that the 1950s marked a sharp break with both the massive vertical integration of the early 1920s and with the cartel practices of the late 1920s is rarely emphasised enough in the literature. In particular, scholars infrequently discuss how the 'autarchic' firms in the west and the export-industry in the south once again found a modus vivendi after the war. It is seldom asked what replaced the cartel agreements plus compensatory side-payments in post-war Germany (an exception, of course, is Herrigel 1996).[13] Table 2 may give an initial, if insufficient impression of how much the surge in cartel agreements in the second half of the 1920s had been brought to an abrupt end in post-war Germany.



Table 2 Estimated numbers of cartels in Germany, 1880s-1960s

Years Number of cartels

1880-1890 50-60
1890-1910 400-500
1925-1929 2,500[a]
1952-1957 0
1957-1962 111[b]

a in 1925.
b in 1962.
Sources: Chandler (1990: 423); Klump, quoted from Overy (1996: 27); Djelic (1998: 54-5, 63); Liefmann (1938); Michels (1928).


Even if this data has to be interpreted with some caution, we know that the coverage and actual effectiveness of the AVI agreement was quite high, and we can be confident that the turn against cartels in the 1950s did mark a very substantial change in the practices governing the 'regulation of competition among business' (Swenson 1999: 8).[14] A central claim of this paper is that coordinated wage bargaining, as it came to be first established in the metal industry, proved to be a functional equivalent — at least in part — for inter-sectoral coordination via cartel agreements. And as Peter Swenson has argued convincingly, wage coordination, as a non-self-enforcing arrangement, needed reinforcement from an outside source. It was the welfare state that came to its support when the costs of coordination threatened to undermine the solidarity among employers that was so essential for the functioning of the entire system.


Again we may ask with respect to shift in the location of industrial conflict from shipbuilding in the north (1955-7) to the automobile industry and mechanical engineering in the south (1960-63): Why there? If we look at the structure of the metal industry in Baden-Württemberg (Table 3), we immediately see the need for compromise between medium and large firms if employers wanted to act collectively. As is shown in Table 3, firm size and wage levels differed considerably between mechanical engineering and automobile industries. This remains the case even if we add the electro-mechanical firms to the picture (Noé 1970: 79-80). Furthermore, the district's industrial composition also called for a compromise between primarily export-oriented (mechanical engineering) and mainly domestic-oriented firms (electrical household appliances and automobiles), even if differences here were not as marked.



Table 3 The metal industry in Baden-Württemberg, 1963/64

Mechanical engineering Automobile manufacturing

Average number of employees per firm 200 794
Share of employees in metal industry 28.9% 18.5%
Share of firms in metal industry 23.9% 3.8%
Average gross hourly wage (in DM) 3.79 4.06
Employment share of big firms (500 employees and over) 59.2% 90.2%

Source: Noé (1970: 79-80).


It was the employers who took the initiative and pressed the unions for the centralisation of wage bargaining in the early 1960s. The goal of the employers' association in the metal sector, Gesamtmetall, was to fully centralise wage bargaining on the national level. In preparing for the conflict with the metalworkers' union, Gesamtmetall had already been granted full responsibility from its regional member associations to bargain on their behalf. In 1961, employers on a nation-wide scale simultaneously cancelled all collective agreements with IG Metall in order to synchronise wage bargaining in the various districts (Tarifdistrikte).[15] After this was achieved, employers entered the 1963 wage round with the 'offer' of a general wage stop in the metalworking industries, and later proposed to increase wages by 3.5 per cent, the expected rate of average productivity growth for all industrial sectors. Only in the district of Baden-Württemberg had the union prepared for a strike, and it was here that both parties finally decided to measure their strength in 1963. Taken by surprise by massive lockouts and unable to play large and small employers against each other, the union soon had to cave in to most of Gesamtmetall's demands. The agreement reached between the two parties brought about a wage hike of 5 per cent, which was roughly equal to the productivity growth in the metal industries or to the rate of the German economy's productivity growth plus the 'natural' rate of inflation of around 1.5 to 2 per cent (Noé 1970: 316). This settlement was then quickly extended to all other metal districts. While this settlement already represented a compromise between the medium and large employers within the Baden-Württemberg metal industry that particularly benefited the larger enterprises, the extension of the settlement to other districts again benefited those (especially large) firms that otherwise would have been forced to 'share their rents' with workers much more extensively.


Although strictly speaking, this bargaining practice fell just short of representing top-level negotiations between the national headquarters of Gesamtmetall and IG Metall, centralised wage bargaining did in fact become established since employers were from now on represented in each and every regional negotiation by the same national Gesamtmetall delegation. Only later did the IG Metall centralise decision making in a similar manner by establishing obligatory consultations between union representatives from the districts and the central IG Metall wage committee before the districts could present specific wage demands.[16] But wage coordination was not important for the metalworking sector alone. The employers of the iron, steel, and coal industry (which is collectively referred to in German as the Montanindustrie) were not represented by Gesamtmetall because the special form of co-determination in this sector granted unions a seat on the executive board of these firms, so that unions would have been represented on both sides of the bargaining table. But there is ample evidence that employers of the metalworking and of the metal producing industry made a concerted effort in the Baden-Württemberg conflict and agreed upon a common strategy in the wage bargaining committee (Tarifpolitischer Ausschuss) as well as in the contact committee (Kontaktausschuss), where the Federation of German Industry (Bundesverband der Deutschen Industrie, BDI) and the Federation of Employers' Associations (Bundesvereinigung der Deutschen Arbeitgeberverbände, BDA) convened and where heavy industry was represented via the Arbeitgeberverband Eisen- und Stahlindustrie (Noé 1970: 120-1, 258-260). And although pattern bargaining, once established, did not make wages in the iron and steel industry simply identical to those set by IG Metall and Gesamtmetall in the metalworking sector, we can safely assume that the signalling function of wage agreements in metalworking effectively set the corridor for wage demands in the iron and steel sector below what otherwise, without 'the pattern', would have been the case. Especially IG Metall's dual track strategy of the mid-1950s would have been economically much more unfavourable for the large Ruhr-firms.


As became obvious with the Baden-Württemberg conflict of 1961-3, the leadership role of the metalworkers' union tended to be exploited by employers. Once IG Metall had successfully established its pace-setting role within the DGB, especially by championing social policy issues in the late 1950s, the union could be tamed by employers with the help of coordinated, sector-wide lockouts. Yet, the resulting wage restraint would not only be in the interest of big business, but would ultimately benefit the union itself, since wage and price stability under the Bretton Woods regime of fixed exchange rates would lead in the medium run to an undervaluation of the Deutsch-Mark. This would, in turn, fuel exports and bring about employment growth, which ultimately could (and did) result in union membership growth (Kreile 1978; Flanagan et al. 1983; Boltho 1994; for a dissenting view see Lindlar 1997 and Lindlar and Holtfrerich 1996).[17] Just how widespread the use of sector-wide lockouts was as an instrument to discipline the metalworkers' union can be seen in the next two tables showing the types of lockouts, number of lockouts, number of workers affected, and number of workers affected according to industrial sector.



Table 4 Lockouts in West Germany, 1949-68

Number of lockouts Affected
Firm lockouts
(in %)
Sector lock-outs
(in %)

1949-1958 7 5,543 57.7 42.3
1959-1968 15 302,371 4.2 95.8

Source: Kalbitz (1979: 26, 42).


Table 5 Sector unions affected by lockouts, 1949-76

Union Number of lockouts Affected workers in % Days lost in %

Metal industry 39 640,973 87.9 4,953,825 85.5
Textiles 12 8,317 1.1 161,129 2.8
Public services 7 17,166 2.4 469,887 8.2
Construction 5 1,650 0.2 39,276 0.6
Chemical industry 5 1,413 0.2 18,914 0.3

Source: Kalbitz (1979: 86).


Once confronted with union success in the fierce conflict over working time reduction and sick pay in the late 1950s, German metal employers now discovered the advantages of national wage coordination. They were obvious. Big industry came to benefit from wages that represented a middle road between the different 'abilities to pay' of small- and medium-sized firms, on the one hand, and larger firms, on the other. Moreover, "if wages were determined in economy-wide rather than at enterprise-level negotiations, an individual firm's investment decision no longer affected the wages it had to pay. In these circumstances, centralized wage negotiations led to a higher level of investments and, in so far as labour productivity was raised, to higher wages in equilibrium" (Eichengreen 1994: 44). Both employers and workers benefited from the new settlement.


With the establishment of pattern wage bargaining, company-based welfare schemes lost their role as a fundamental alternative to the collective, corporatist social policy as represented in the Bismarckian social insurance schemes. In particular, the Ruhr industrialists lost their interest in the old-style welfare capitalism so dominant in the Krupp, Stinnes, and Thyssen conglomerates before the 1930s. Once heavy industry had lost its control over the market via cartel agreements, it developed an interest in better controlling wages. To this end, it needed coordination with the metalworking sector. Yet, effective wage coordination meant that the Ruhr iron and steel companies became less opposed to state social policies that covered the entire industry and threatened to critically restrict the impact of company-based benefit schemes. Large firms still offered higher employment security, more generous fringe benefits, additional company pensions, and other social benefits. But 'wage drift' and 'benefit drift' now were mainly correctives for the rigidities that came with coordinated wage bargaining. Over time, this 'drift' became less important as a means to regulate the labour market and it was increasingly contained by the dynamic growth of public social programs. The fundamental conflict of the 1920s over the relative importance of the public versus the private provision of social protection was over by the 1960s. Welfare corporatism and wage coordination became 'strategic complements' for each other (Milgrom and Roberts 1990, 1994; Cooper 1999): wage coordination reinforced employers' interests in collective/corporatistic social policy, while welfare state growth increased industry's, especially heavy industry's interest in the coordination of wages.[18] Having been the fiercest opponents of the Bismarckian welfare state in the interwar years (Weisbrod 1978), employers in heavy industry slowly made their peace with the post-war compromise and even learnt to appreciate its particular advantages.


In the following section, I will briefly discuss the degree to which the parallel reconstruction of the Bismarckian welfare state had an impact on the political and economic equilibrium in industrial relations that became established in the late 1950s, early 1960s.



The reconstruction of the Bismarckian welfare state in the late 1950s


The Adenauer pension reform in 1957 brought a substantial expansionary step in welfare state spending. Expenses for pensions increased in real terms by some 27 per cent in 1957 alone. Mainly because of the pension reform, social spending as a share of total public spending jumped from 55 per cent in 1955 to 59 per cent in 1958 (Alber 1989: 84 and 78). Current pensions increased on average by 65 per cent (blue collar worker pensions) or 72 per cent (white collar worker pensions; Hockerts 1980: 422)! The contribution rate to the old-age insurance increased from 11 per cent of gross wages in 1956 to 14 per cent in 1957. Total pension expenditures doubled in only 3 years, from 6 billion DM in 1955 to 12 billion DM in 1958 (in current prices, but during a time of low inflation). Due to this significant expansionary step, Germany had the highest percentage of social spending as a share of GDP among all West European countries throughout the 1950s and until 1965 (see Table 6).



Table 6 Germany's rank in social spending among OECD countries, 1950-80 (in per cent of GDP)

Germany Rank OECD average OECD maximum OECD minimum
(in %) (in %) (in %) Country (in %) Country

1950 14.8 1 9.3 14.8 D 5.7 N
1955 14.2 1 10.2 14.2 D 6.8 CH
1960 15.4 1 11.4 15.4 D 7.5 CH
1965 16.6 2 13.4 17.8 A 8.5 CH
1970 17.0 5 15.8 20.0 NL 10.1 CH
1975 23.5 5 21.2 26.8 NL 15.1 CH
1980 23.8 6 22.8 32.0 S 13.8 CH

Source: Alber (1989: 40, Table 4).


It was only later, in the second half of the 1960s, that first Austria, then the Netherlands and later Sweden surpassed Germany and became the leaders in social spending (Alber 1989: 41). Hence, the high growth period of the German economic miracle, its Wirtschaftswunder, was also a time of big (social) spending, and apparently the success of German industry on world markets remained largely unaffected by the above-average welfare costs. It is particularly interesting to note that in the reform debates of 1956 and 1957, few complaints were made about the sharp increase in social spending caused by Adenauer's pension reform (Hockerts 1980). The main concern was that the new 'dynamic' pension would endanger price stability, that it would set in motion a scala mobile between prices, wages, and the 'dynamic' pensions. Inflation, not the immediate steep increase in social spending, was the most troubling question raised (Hockerts 1980; BDA 1956a, 1956b). In response to the pension reform plans of the Adenauer government, the BDA recommended that firms should begin to 'exercise restraint' when deciding upon new voluntary company-based social programs, given the significant expansion of the public commitments as introduced with the reform (BDA 1956c: 4). Furthermore, employers called for a binding upper limit on the replacement level of private and public pensions combined and for the possibility to adjust past private commitments (downwards) to the new situation (BDA 1956c: 4). This clearly indicates how much the growth of public social programs in the late 1950s began to crowd out the private schemes. Revealingly, this process did not provoke much employer protest.


It is important to note that in Germany, as in Sweden, the most significant period in the post-war growth of the welfare state took place in times of high economic growth and de facto full employment, when strong wage pressures seemed to endanger international competitiveness. This sets Sweden and Germany clearly apart from the US, where the New Deal Keynesianism was primarily a response to the severe economic crisis of the 1930s (Swenson 2000). The political and economic logic of welfare state expansion and contraction thus seemed to have been diametrically opposed in these countries.


With the onset of the social reform debate in the second half of the 1950s, German unions shifted their emphasis from wage demands to demands for general 'social policy' issues like working time reduction, sick pay, or early retirement (Pirker 1979). The 1956 Hamburg DGB conference was entirely devoted to the discussion of social policy issues. The new DGB chairman, Willi Richter, elected at the Hamburg conference, was a Social Democratic MP and an expert on social policy. With the 1956-7 Schleswig-Holstein strike, unions had visibly added 'social concerns to the predominantly economic issues which were considered "strike-worthy" by the labour movement' (Markovits 1986: 193). Apparently, the idea of substituting political demands involving the extension of welfare entitlements for wage demands pointed to a solution for the dilemma in which IG Metall found itself after 1955. Both the intense conflict over sick pay (during which, in 1956-7, the metalworkers' union of Schleswig-Holstein staged a strike that lasted 16 weeks, which has since proven to be the longest strike ever in the history of the Federal Republic; cf. Markovits 1986: 190-5) and the significant extension of workers' pension benefits through the pension reform of 1957 have to be seen in this light.


The German post-war political economy came to be based on a quasi-corporatistic compromise among employers, unions, and the government even though central corporatistic coordination, active macro-economic management, a 'loose' monetary regime, Social-Democratic hegemony, and a commitment to full employment all remained alien to 'Modell Deutschland'.[19] In particular, workers could expect to benefit from the 'industrial achievement-performance model' as established with the Bismarckian model. Initially, the welfare state helped balance unions' wage demands and the necessity to remain competitive in world markets. In giving social policy issues top priority, the IG Metall had almost 'naturally' established itself as a pacesetter, since no other union could afford not to fight for the extension of 'social achievements' like the 45-hour week, sick pay, or increased employment protection. Employers, in turn, had a hard time justifying why workers should be treated differently across industries.What was at stake were questions of justice, fairness, workers' dignity and status, rather than the more mundane issues of productivity, business cycles, and income distribution. It was only later, in 1963, when the metal employers had already learnt of the strength of the metalworkers' union through the Schleswig-Holstein strike, that a concerted and organised lockout in the Baden-Württemberg region established pattern wage bargaining (Kalbitz 1979; Noé 1970).


From the employers' standpoint, welfare expansion was viewed less critically as long as it helped moderate unions' wage demands. Employers valued the flexibility which the various welfare state programs offered to them when coping with market volatility. This was especially true for the shipbuilding firms who were confronted with 'severe market fluctuations' (Strath 1994: 74) within a 'rapidly changing industry' (Strath 1994: 73), Furthermore, the generous and dynamically growing welfare state disciplined employers who were tempted to abandon employer 'solidarity' in wage setting and to pay wages above the level set by collective agreements (Swenson 2000). Generous welfare benefits, directly linked to wages, constrained both 'wage drift' and 'welfare drift'. Also, by coupling the interests of the large firms in the Ruhr region with the economic interests of the small and medium-sized enterprises and the exposed firms, companies like Thyssen, Krupp, and other representatives of German heavy industry were integrated into the postwar order. The Ruhr magnates gave up their old paternalist Herr-im-Hause policy and slowly reduced their elaborate 'welfare capitalist' systems. Part of these firm-specific voluntary benefit packages had been comprised of continued wage payments in case of sickness exactly like those for which the IG Metall had called the Schleswig Holstein-strike. Again, it was obvious that big firms were only slightly affected negatively by the strike's outcome.


The next section will briefly discuss just how well the Japanese post-war settlement of industrial conflict may fit into the picture of the parallel and interdependent formation of wage and welfare regimes.



Japan's post-war compromise: the spring wage offensive, keiretsu, and the company-based welfare state

To 'regulate competition among employers' meant something quite different in Japan. Unlike the German and Swedish cases, Japanese employers neither had to worry much about the 'exposed' versus 'sheltered' divide, nor about a profound conflict of interest between large and small firms. The major industrial conglomerates, the keiretsu, organised both export-oriented and domestically oriented firms (horizontal keiretsu; Westney 1998; Nakatani 1998), and the large employers and small firms were often linked to each other via complex contractor-supplier relations (vertical keiretsu). This prevented sharp intra-industrial conflict from arising. In particular, large companies were able "to operate their own type of wage restraint on small companies by setting as monopsonists the prices at which they are prepared to buy from them" (Soskice 1990: 42). Moreover and probably even more important, the Japanese state with a panoply of protectionist and regulative measures insulated the home market, encouraged the formation of cartels, and eased 'the great pressure on the smaller (firms) to modernise or to raise wages due to the 'higher wages in the country's rapidly expanding larger firms' (Pempel 1998: 61). Thus, despite the fact that Japan's industrial structure was markedly segmentalised, competition between low- and high-commitment companies was less of a problem (Pempel 1998). Cartels and encompassing protectionism remained legitimate instruments of economic policy in post-war Japan. Compared with the Federal Republic, the deconcentration measures and the anti-cartel legislation initiated under US occupation had far less success in installing a liberal market economy (Pempel 1998: 93; Herrigel 2000).


More of a problem for Japanese industrialists was the regulation of the fierce oligopolistic competition among the major industrial groupings and the negative effect union militancy in the public sector was having on wage negotiations in the private sector. Whereas the latter problem was solved over time by the harsh anti-(public)union stance of the conservative LDP government that introduced compulsory arbitration for public enterprises and national corporations and finally revoked public workers' right to strike in 1973, the former problem, dubbed the problem of 'excessive competition', faded when finally the coordination of wage setting became established. Shunto, the annual "spring offensive" in which management and trade unions conduct wage negotiations throughout the country, came into being in the late 1950s (Hamada and Kasuya 1993 177-78). As was to become apparent during this period, the Shunto wage offensive was not only an instrument by which unions tried to achieve labour movement unity, it could also serve as a regulative device that promised to ease the competitive pressures among employers (Koshiro 1983: 212). Shunto improved the 'ability of the large companies to concert their wage policies and bargaining tactics both within and across industry boundaries' (Brown et al. 1997: 161), something of paramount interest to employers in times of high growth and labour scarcity. Moreover, because export industries would have the leadership role in the spring offensive, the impact of public sector union radicalism could be effectively contained after 1965 (Koshiro 1983: 220-222).


From the unions' point of view, the spring wage offensive was an opportunity to make 'workers focus more strongly on their relative (interfirm) wages and thus to encounter the efforts of large scale firms to insulate their own labour markets' (Brown et al. 1997: 160). Seen from the employers' perspective, this attack on firm-specific 'insulated' labour markets finally provided the large industrial groups with a means to avoid the costly mutual overbidding under conditions of full employment. Yet, what was true in the German case was true in the Japanese case as well: it took employers some time to discover these beneficial side effects of coordinated wage bargaining. In 1956, in the initial Shunto wage round, unions managed for the first time to push through their coordinated wage demands against strong resistance from Nikkeiren, the national employers association (Takanashi et al. 1996: 2-3). As employers were to learn from this and coming wage rounds, unions also helped ease the competition among employers by levelling wage differentials between the major firms, and once the link between wage demands and productivity growth had been accepted, wage coordination Japanese style displayed most of the beneficial effects already pointed out above with regard to coordinated wage bargaining in the Swedish and German cases.


The most important of these advantages was that wages were set below equilibrium for the most productive firms, especially if 'major firms in the private and open sectors of the economy' took the lead in the annual wage rounds (Brown et al. 1997: 160; my emphasis). For large firms, it was essential that 'as long as wage negotiations and strikes were coordinated, management did not have to worry about losing its competitive position', while unions 'being ultimately organised along the company's line, also had a strong interest in not damaging the relative position of their own company' (Takanashi et al. 1996: 24). Again, Japanese employers, like their Swedish or German counterparts, also changed their attitude towards state social policy once wage coordination became entrenched. Japan's dual industrial structure had been replicated in the dual structure of the Japanese welfare state. Most large firms had chosen to opt out of the Employee Pension System (Kosei Nenkin Hoken, KNH/EPS), which had been founded in 1941 and rebuild in the early 1950s, while workers in the many small- and medium-sized firms were usually members in the EPS, if not in the National Pension System (self-employed). Within this dual system, lower premiums, less co-payment, and higher benefits prevailed in the state-regulated company schemes and the contracted-out company schemes in health and pension insurance were even more generous (Employee Pension System, Employee Health Insurance and Employee Pension Funds), while residual provisions, lower benefits and longer qualifying periods were to be found in the public programs (National Pension System [NPS, 1959] and National Health Insurance [NHI, 1938/58]; see Estevez-Abe 1996; Estienne 1999). Yet, smaller firms in which the workers were in the 'lower' tax-financed, flat-rate social protection tier enjoyed a higher degree of public subsides or enjoyed particular tax privileges for the pension plans that were tailored to their needs (Tax Qualified Pensions, introduced in 1962).


The contracting-out option had been given to business in 1966 in the form of Employee Pension Funds (Kosei Nenkin Kikin, KNK/EPF). A firm could contract out of the Employee Pension System if the company union consented and certain requirements were fulfilled (e.g., 30 per cent higher benefits than those paid by the EPS). The firm had to hand the income-related part of the EPS over to a life insurance company or trust bank (Estienne 1999: 53-54; Watanabe 1998). That is, firms were not allowed to run these schemes by themselves (for instance in the form of book reserves). This allayed union concerns about the potential misuse of company pensions as a paternalistic device to control and discipline workers, and it also coincided with the Ministry of Finance's (MOF) interest in supporting the domestic life insurance industry and the banking sector.[20] What is more important in the context of our argument is that this stipulation subjected EPF pensions to tight state regulation that secured the pooling of funds, set a uniform revenue rate (5.5 per cent) for the pension funds of all firms, and later secured the indexation of entitlements as well, strictly linking the 'private' EPF pensions to the development of the public EPS pensions (Estienne 1999: 94). While the Ministry of Finance mainly pursued a 'developmentalist' strategy in tightly regulating the investment strategies of trust banks and life insurance companies and in channelling pension capital assets into targeted industries, it also helped stabilise the industry-wide standardisation of wages and company-based social benefits. Moreover, the state regulation of the 'private' contracted-out schemes ensured that company welfare became more than a simple fringe benefit, more than voluntary commitments that could be revoked during an economic downturn. While companies could stabilise their workforce by offering more favourable provisions as compared to the public schemes, the state continued to tightly regulate company welfare, to force uniform standards upon the larger part of the industry, and to lend credibility to the private promises of business. The state played the role of a 'direct structurer and overseer of occupationally-linked pensions' (Shinkawa and Pempel 1997: 170; Estienne 1999: 94). In this context it is important to distinguish between the public provision and the public regulation of social welfare, the latter often being highlighted as a distinctive feature of the East Asian model of the welfare state (Jacobs 1998; Goodman et al. 1998). The common assumption that the private provision of social benefits automatically is proof of the liberal, 'commodifying' character of a welfare state regime (for Japan see e.g. Esping-Andersen 1997: 183) ignores this important difference and substantially underestimates the degree to which 'private' welfare schemes can diverge from a pure market equilibrium. While they seem to resemble US-American welfare capitalism in many ways, Japan's 'private' company pension schemes actually differ quite profoundly, due to the tight state regulation and their character as 'defined benefit' schemes in contrast to 'defined contribution' schemes like, for instance, 401k pensions (Watanabe 1998: 9). Thus, in a detailed assessment of the Japanese pension funds, Jean-Francois Estienne concludes that 'contracting out' in Japan 'hardly resembles the Anglo-Saxon version since the mandatory state regime and the private pension funds are organisationally linked and performance related' (Estienne 1999: 94, translation by the author). A closer look at the Japanese 'welfare production regime' (Estevez-Abe et al. 1999) reveals that the Japanese welfare state came to underpin business coordination much more than purely voluntary schemes would have been able to do.



In this paper I have considered the extent to which Peter Swenson's historical account of Swedish and US-American welfare state development can shed light on the situation in other countries. Insofar as future research confirms the findings of this paper, namely that the German and Japanese post-war experiences provide considerable evidence that welfare and wage regimes in these countries have developed in close correspondence with each other, the academic debate will have to focus on a new set of causal links between capitalist production and social protection. Peter Swenson's work and recent contributions in a similar vein by other scholars (e.g. Isabela Mares, Margarita Estevez-Abe, Cathie Jo Martin) call for a new political-economic view of the welfare state, a perspective that would have to take into account

  • the importance of industrial structure for the position of employers towards public social policy,

  • the impact of (different forms and degrees of) trade openness on employers' interest in business and wage coordination and, consequently, on employers' interest in social spending programs,

  • the ability of employers to coordinate within and across industrial sectors and the role that the welfare state plays for the 'regulation of competition' among employers.

Such a new perspective would also require a fresh look at the nexus between economic growth and the growth of social spending (Castles 2000), which could challenge the old claim made in industrialisation theory that the wealth created by economic growth was a precondition for subsequent welfare state expansion. The argument presented here implies an alternative explanation, namely that high growth periods in centralised/coordinated production regimes weakened employers' anti-welfare stance in the social reform debates of the times. Peter Swenson's contribution to the literature on comparative welfare regimes represents both a veritable challenge to what is held as conventional wisdom in this literature and a fruitful scientific program for future research in this field.



    Alber, J. (1989)Der Sozialstaat in der Bundesrepublik 1950-1983, Frankfurt: Campus.

    Albert, M. (1991) Capitalisme contre Capitalisme, Paris: Seuil.

    Albert, G. (1998) Wettbewerbsfähigkeit und Krise der deutschen Schiffbauindustrie 1945-1990, Frankfurt: Peter Lang.

    Baldwin, P. (1990) The Politics of Social Solidarity. Class Bases of the European Welfare State 1875-1975, New York: Cambridge University Press.

    Bankverein (1954) Wer gehört zu wem? No place, no publisher.

    BDA [Bundesverband Deutscher Arbeitgeberverbände] (1956a) Probleme der Sozialreform. Entwurf, Cologne: mimeo.

    BDA (1956b) Stellungnahme und Vorschläge der Bundesvereinigung der Deutschen Arbeitgeberverbände zur Rentengestaltung, Cologne: mimeo


    BDA (1956c) Niederschrift über die Sitzung des Ausschusses für Sozialversicherung und betriebliche Sozialfürsorge am 30. Juli 1956 in Köln, Cologne: mimeo.

    Blödner, P. (1934) Preis- und Produktstabilisierung durch Kartelle. Eine Untersuchung ihrer Problematik am Beispiel der deutschen Großeisenindustrie, Dissertation, Martin-Luther Universität Halle/Saale.

    Brown, C., Nakata, Y., Reich, M. and Ulman, L. (1997) Work and Pay in the United States and Japan, New York: Oxford University Press.

    Castles, F. (2000) 'The dog that didn't bark: Economic development and the postwar welfare state', European Review 8 (No. 3), 313-32.

    Chandler, A.D., Jr. (1990) Scale and Scope. The Dynamics of Industrial Capitalism, Cambridge, MA: Belknap Press.

    Cooper, R.W. (1999) Coordination Games. Complementarities and Macroeconomics, New York: Cambridge University Press. Crouch, C. and Streeck, W. (eds.) (1996) Modern Capitalisms, London: Sage.

    Diegmann, A. (1993) 'American deconcentration policy in the Ruhr coal industry', in J.M. Diefendorf, A. Frohn and H.-J. Rupieper (eds.) American Policy and the Reconstruction of West Germany, 1945-1955, New York: Cambridge University Press, 197-215.

    Djelic, M.-L. (1998) Exporting the American Model. The Post-war Transformation of European Business, Oxford: Oxford University Press.

    Esping-Andersen, G. (1997) 'Hybrid or unique?: The Japanese welfare state between Europe and America', Journal of European Social Policy 7, 179-89.

    Estevez-Abe, M. (1996) 'The welfare-growth nexus in the Japanese political economy', paper for the 22nd Annual APSA Convention, San Francisco, CA.

    —, Iversen, T. and Soskice, D. (1999) 'Social protection and the formation of skills: a reinterpretation of the welfare state', paper for the 95th American Political Science Association Meeting, Atlanta, GA.

    Estienne, J.-F. (1999) Réforme et Avenir des Retraites: Les Enseignements de l´Exemple Japonais. Paris: La Bibliothèque d'Économie Financière.

    Fear, J. (1997) 'August Thyssen and German steel', in T. McCraw (ed.) Creating Modern Capitalism. How Entrepreneurs, Companies and Countries Triumphed in the Three Industrial Revolutions, Cambridge, MA: Harvard University Press, 185-226.

    Feldman, G.D. (1977) Iron and Steel in the German Inflation 1916-1923, Princeton, NJ: Princeton University Press.

    Feldman, G.D. and Nocken, U. (1975) 'Trade associations and economic power: a comparison of interest group development in the German iron and steel and machine building industries 1900-1933', Business History Review 49, 413-445.

    Flanagan, R.J., Soskice, D.W. and Ulman, L. (1983) Unionism, Economic Stabilization, and Income Policies. European Experience, Washington, DC: The Brookings Institute.

    Garrett, G. (1998) 'Global markets and national politics: collision course or virtuous circle?' unpublished manuscript, Yale University, New Haven, CT. Giersch, H., Paqué, K.-H. and Schmieding, H. (1992) The Fading Miracle. Four Decades of Market Economy in Germany, New York: Cambridge University Press.

    Gillingham, J. (1991) Coal, Steel and the Rebirth of Europe, 1945-1955. The Germans and French from Ruhr Conflict to Economic Community, New York: Cambridge University Press.

    Goodman, R., White, G. and Kwon, H.-J. (1998) In Search of an East Asian Welfare State, London: Routledge.

    Gordon, A. (1985) The Evolution of Labor Relations in Japan. Heavy Industry, 1853-1955, Cambridge, MA: Harvard University Press.

    Gordon, C. (1994) New Deals: Business, Labor and Politics in America, 1920-1935, New York: Cambridge University Press.

    Hall, P.A. and Franzese, R.J. (1998) 'Mixed signals: central bank independence, coordinated wage bargaining, and European Monetary Union', International Organization 52, 505-535.

    Hamada, K. and M. Kasuya (1993) 'The reconstruction and stabilization of the postwar Japanese economy: Possible lessons for Eastern Europe?' in R. Dornbusch, W. Nölling, R. Layard (eds.) Postwar Economic Reconstruction and Lessons for the East Today. Cambridge, MA: MIT Press, 155-187.

    Hentschel, V. (1996) Ludwig Erhard: Ein Politikerleben, Munich: Olzog. Herrigel, G. (1996) Industrial Constructions: The Sources of German Industrial Power, New York: Cambridge University Press.

    — (2000) 'American occupation, market order, and democracy: reconfiguring the steel industry in Japan and Germany after the Second World War', in J. Zeitlin and G. Herrigel (eds.) Americanization and Its Limits. Reworking US Technology and Management in Post-war Europe and Japan, Oxford: Oxford University Press, 340-99.

    Hockerts, H.G. (1980) Sozialpolitische Entscheidungen im Nachkriegsdeutschland. Alliierte und deutsche Sozialversicherungspolitik 1945 bis 1957, Stuttgart: Klett-Cotta.

    Hoffmann, E. (1928) 'Die Organisation der Verbindung der Eisen schaffenden und der Eisen verarbeitenden Industrie', Dissertation, University of Munich.

    Holtfrerich, C.-L. (1998) 'Geldpolitik bei festen Wechselkursen (1948-1970)', in Deutsche Bundesbank (ed.) Fünfzig Jahre Deutsche Mark. Notenbank und Währung seit 1948, Munich: Beck, 347-437.

    IG Metall (ed.) (1978) 'Streik der Metaller Schleswig-Holstein 1956/57', Dokumentation, Frankfurt.

    Immergut, E.M. (1999) Liberalism, Associations, and the Social Wage: A Case Study of the Politics of Sickness Pay. Paper prepared for the workshop 'Liberalism and Change: Political Rights and Economic Capacities in Germany and the United States' January 22-23, University of Berkeley, California.

    Iversen, T. (1998) 'Wage Bargaining, Hard Money and Economic Performance: Theory and Evidence for Organized Market Economies', British Journal of Political Science 28, 31-61.

    Jackson, Chr.R. (1993) Infirmative Action: The Law of the Severely Disabled in Germany, Central European History, 417-455.

    Jacobs, D. (1998) 'Social welfare systems in East Asia: a comparative analysis including private welfare', CASE paper CASE/10, Center for Analysis of Social Exclusion, London School of Economics.

    Jacoby, S.M. (1999) Modern Manors: Welfare Capitalism since the New Deal. Princeton, N.J.: Princeton University Press.

    Kalbitz, R. (1978) 'Biographie über den Streik der IG Metall in Schleswig-Holstein 1956/57', in IG Metall (ed.) Streik der Metaller Schleswig-Holstein 1956/57, Dokumentation, Frankfurt, 181-217.

    Kalbitz, R. (1979) Aussperrung in der Bundesrepublik, Cologne: EVA. Katz, L.F. and Summers, L.H. (1989) 'Industry Rents: Evidence and Implications', Brookings Papers on Economic Activity, 209-275.

    Kaufmann, H.M. (1969) A Debate over Germany's Revaluation 1961. A Chapter in Political Economy, Weltwirtschaftliches Archiv 103 (2), 181-212.

    King, G., Keohane, R.O. and Verba, S. (1994) Designing social inquiry: scientific inference in qualitative research, Princeton, NJ: Princeton Univ. Press, 1994.

    Koshiro, K. (1983) 'Development of Collective Bargaining in Postwar Japan', Taishiro Sh. (ed.), Contemporary Industrial Relations in Japan, Madison, WI: University of Wisconsin Press, 205-258.

    Kreile, M. (1978) 'West Germany: The Dynamics of Expansion', in P.J. Katzenstein (ed.) Between Power and Plenty. Foreign Economic Policies of Advanced Industrial States, Madison, WI: University of Wisconsin Press, 191-224.

    Kuckuk, P. (1998) 'Westdeutscher Schiffbau in der Nachkriegszeit. Ein Überblick', in P. Kuckuk (ed.) Unterweserwerften in der Nachkriegszeit. Von der Stunde Null zum Wirtschaftswunder, Bremen: Edition Temmen, 11-36.

    Leckebusch, G. (1963) Die Beziehungen der Deutschen Seeschiffswerften zur Eisenindustrie an der Ruhr in der Zeit von 1850 bis 1930. Schriften zur Rheinisch-Westfälischen Wirtschaftsgeschichte, Cologne: Rheinisch-Westfälisches Wirtschaftsarchiv zu Köln.

    Liefmann, R. (1938) Cartels, Concerns and Trusts, New York: Dutton.

    Manow, P. and Seils, E. (2000) 'The unemployment crisis of the German welfare state' West European Politics, special issue on 'Restructuring European Welfare States', M. Rhodes and M. Ferrera (eds.), April 2000, 138-160.

    Mares, I. (1996) 'Firms and the welfare state: the emergence of new forms of unemployment', Discussion Paper FS I 96-308, Berlin: Wissenschaftszentrum Berlin für Sozialforschung.

    — (1997) 'Business (non) coordination and social policy development: the case of early retirement', paper for the conference 'Varieties of Capitalism', June 6-8, 1997, Wissenschaftszentrum Berlin für Sozialforschung, Berlin.

    — (2000) 'Strategic alliances and social policy reform: unemployment insurance in comparative perspective', Politics & Society 28, 223-244.

    Markovits, A.S. (1986) The Politics of the West German Trade Unions. Strategies of Class and Interest Representation in Growth and Crisis, New York: Cambridge University Press.

    Martin, C.J. (2000) Stuck in Neutral. Business and the Politics of Human Capital Investment Policy, Princeton, NJ: Princeton University Press.

    Michels, R. (1928) Cartels, Combines and Trusts in Postwar Germany, New York: Columbia University Press.

    Milgrom, P. and Roberts, J. (1990) 'Rationalizability, learning and equilibrium in games with strategic complementarities', Econometrica 58, 1255-78.

    — (1994) 'Complementarities and systems: understanding Japanese economic organization', Estudios Economicos 9, 3-42.

    Moene, K.O. and Wallerstein, M. (1995) 'How social democracy worked: labor-market institutions', Politics & Society 23, 185-211.

    Nakatani, I. (1998) 'The economic role of financial corporate grouping', in P. Drysdale and L. Gower (eds.) The Japanese Economy, Part 1, Volume IV: The Nature of the Japanese Firm, London: Routledge, 176-209.

    Nocken, U. (1977) 'Inter-industrial conflicts and alliances as exemplified by the AVI-Agreement', in H. Mommsen, D. Petzina and B. Weisbrod (eds.), Industrielles System und politische Entwicklung in der Weimarer Republik, Dusseldorf: Droste, 693-704.

    Noé, C. (1970) Gebändigter Klassenkampf. Tarifautonomie in der Bundesrepublik Deutschland. Der Konflikt zwischen Gesamtmetall und IG Metall vom Frühjahr 1963, Berlin: Duncker & Humblot.

    Nörr, K.W. (1993) An der Wiege deutscher Identität nach 1945: Franz Böhm zwischen Ordo und Liberalismus, (Schriftenreihe der Juristischen Gesellschaft zu Berlin, Heft 129), Berlin: de Gruyter.

    Overy, R. (1996) 'The economy of the Federal Republic since 1949', in K. Larres and P. Panayi (eds.) The Federal Republic of Germany since 1949. Politics, Society and Economy before and after Unification, London: Longman.

    Paqué, K.-H. (1996) 'Unemployment and the crisis of the German model: a long-term interpretation', in H. Giersch (ed.) Fighting Europe's Unemployment in the 1990s, New York: Springer, 123-155.

    — (1998) 'Zur Zumutbarkeit von Arbeitsplätzen: Bestandsaufnahme und Reformvorschlag', in E. Knappe and N. Berthold (eds.) Ökonomische Theorie der Sozialpolitik, Heidelberg: Physica, 71-89.

    Pempel, T.J. (1998) Regime Shift. Comparative Dynamics of the Japanese Political Economy, Ithaca, NY: Cornell University Press.

    Pierson, P. (1995) The Scope and Nature of Business Power: Employers and the American Welfare State, 1900-1935, unpublished manuscript, Harvard University and Russell Sage Foundation, Cambridge, MA.

    Pirker, T. (1979) Die blinde Macht. Die Gewerkschaftsbewegung in Westdeutschland, Teil 2: 1953-1960. Weg und Rolle der Gewerkschaften im neuen Kapitalismus, Berlin: Olle & Wolter.

    Preiss, K. (1933) 'Das System der Ausfuhrrückvergütungen in der deutschen Eisen- und Metallindustrie', Dissertation, University of Kiel.

    Rieger, E. and Leibfried, S. (1998) 'Welfare State Limits to Globalization' Politics & Society 26, 363-390.

    Rodrik, D. (1997) Has Globalization Gone Too Far?, Washington, DC: Institute for International Economics.

    Ruggie, J.G. (1997) 'Globalization and the embedded liberalism compromise: the end of an era?' MPIfG Working Paper 97/1, Cologne: Max Planck Institute for the Study of Societies, Internet Version: wp97-1_e/wp97-1_e.html.

    Schmid, R. (1930) 'Das AVI-Abkommen als Mittel verbandsmäßiger Exportförderung', Dissertation, University of Cologne.

    Schmitter, P.C. (1979) 'Still the Century of Corporatism?' in P. Schmitter and G. Lehmbruch (eds.) Trends towards Corporatist Intermediation, London: Sage, 7-52.

    Schneider, M. (1980) Ausperrung. Ihr Geschichte und Funktion vom Kaiserreich bis heute, Cologne: Bund Verlag.

    Shinkawa, T. and Pempel, T.J. (1996) 'Occupational welfare and the Japanese experience', in M. Shalev (ed.) The Privatization of Social Policy? Occupational Welfare and the Welfare State in America, Scandinavia and Japan, London: Macmillan, 280-326.

    Soskice, D. (1990) 'Wage Determination: The Changing Role of Institutions in Advanced Industrialized Countries' Oxford Review of Economic Policy 6, 36-61.

    — (1995) 'Finer varieties of advanced capitalism: industry versus group-based coordination in Germany and Japan', manuscript, Wissenschaftszentrum Berlin für Sozialforschung, Berlin.

    Strath, B. (1994) 'Modes of governance in the shipbuilding industry in Germany, Sweden, and Japan', in J.R. Hollingsworth, P.C. Schmitter and W. Streeck (eds.) Governing Capitalist Economies. Performance and Control of Economic Sectors, New York: Oxford University Press, 72-96.

    Streeck W. and K. Yamamura (eds.) (2000) The Political and Social Origins of Nationally-Organized Capitalism: Japan and Germany in Comparison (forthcoming). Swenson, P. (1991) 'Labor and the limits of the welfare state: the politics of intraclass conflict and cross-class alliances in Sweden and West Germany', Comparative Politics 23, 379-399.

    — (1997) 'Arranged alliance: business interests in the New Deal', Politics & Society 25, 66-116.

    — (1999) 'Varieties of capitalist interests and illusions of labor power: employers in the making of the Swedish and American welfare states', paper for the 'Conference on Distribution and Democracy', Yale University, Department of Political Science, November 12-14, 1999.

    — (2000) Labor Markets and Welfare States: Employers in the Making of the American and Swedish Systems (forthcoming). Takanashi, A. et al. (1996) Shunto Wage Offensive, Tokyo: The Japan Institute of Labor.

    Tübben, W. (1930) 'Die nationale und internationale Verbandspolitik der Schwerindustrie vor und nach dem Kriege', Dissertation, University of Wurzburg.

    Watabane, N. (1998) 'Occupational Pension Systems in Japan', Japan Labour Bulletin 38, 5-14.

    Weisbrod, B. (1978) Schwerindustrie in der Weimarer Republik. Interessenpolitik zwischen Stabilisierung und Krise, Wuppertal: Peter Hammer.

    Weiss, A. (1990) Efficiency Wages. Models of Unemployment, Layoffs and Wage Dispersion, Princeton, N.J.: Princeton University Press.

    Wend, H. (2000) Recovery and Restoration: U.S. Foreign Policy and the Politics of Reconstruction in the West German Ship Construction Industry, 1945-1955, New York: Praeger (forthcoming).

    Westney, D.E. (1998) 'The Japanese business system: key features and prospects for change', in P. Drysdale and L. Gower (eds.) The Japanese Economy, Part 1, Volume IV: The Nature of the Japanese Firm, London: Routledge, 151-175.

    Winkler, H.A. (ed.) (1974) Organisierter Kapitalismus. Voraussetzungen und Anfänge, Goettingen: Vandenhoeck & Ruprecht.

    Wrede, J.M. (1933) 'Die internationale Verbandspolitik der nordwesteuropäischen Eisenindustrie in ihrer Bedeutung für die deutsche Eisenindustrie', Dissertation, University of Würzburg.






    1   I am grateful to Jim Mosher for fruitful discussions and to Peter Swenson, Michael Wallerstein, Kathy Thelen, Katharina Bluhm, Sue Giaimo, Gary Herrigel, Eric Seils, Martin Höpner and Helmut Voelzkow for their valuable comments. Many thanks go also to Dona Geyer and Cynthia Lehmann for the language editing.

    2   Given that big firms within a decentralised setting would be forced by their workforce to engage more in 'rent sharing' than small firms (cf. Katz and Summers 1989), in particular if union membership correlates positively with firm size. For more on the 'local pushfulness' of decentralized wage bargaining, see Soskice (1990) and Iversen (1999).

    3   How could primarily tax-financed social programs have been able to exert "disciplinary pressures" on employers in the Swedish case? In answering this question, Swenson focuses in particular on the important introduction of the contribution-financed and income-related ATP pensions in the 1950s "that resembled Bismarckian social insurance more than they did the supposedly socialist model of the Scandinavian welfare state" (Baldwin 1990: 146). We can of course also substitute Swenson's argument with the traditional argument that wage moderation in centralised wage bargaining is exchanged for welfare state expansion (Goldthorpe 1984). In this setting wage moderation was important for employers because it made investments profitable and because it enabled them to finance further investments out of these profits (Eichengreen 1994: 41), and wage coordination was important since it made sure that wage moderation in one sector would not lead to higher profits (without higher investments) in other sectors of the economy (see for an extremely elegant account of this labour/capital-game Eichengreen 1994; on the last point see p. 44).

    4   The reader will notice that this paper does not contain a complete and fully equal comparison of German and Japanese 'industrial relations cum social protection' in the post-war period. The following narrative is much more detailed in the description of the German post-war settlement. The parallel Japanese case rather serves as an additional plausibility check in a case that is as critical for Peter Swenson's argument as the German case (King, Keohane and Verba 1997).

    5   I will not discuss here at any length the two other explanations with which Giersch et al. (1992) try to explain the persistent gap between wages and productivity in the 1950s. 'Economic surprise', which they believe played a role, is largely incompatible with rational expectations, especially if we try to explain persistent differences over a time period of 10 years. 'Social responsibility' presumably played a role, yet this explanation clearly overlaps with the one endorsed here, which holds that the IG Metall leadership tried to keep wages internationally competitive, but at the same time sought to increase workers' welfare and to avoid an increase in unemployment.

    6   Besides the substantial increase in benefits, the Adenauer reform contained three elements that were quite significant for unions and workers: First, by strictly linking benefits to contributions, the reform put strong premiums on stable career paths, on high incomes and on the early acquisition of skills. It thus came to benefit in particular the core union clientele, the highly qualified core workforce (Facharbeiter). Second, by also entitling blue-collar workers to a pension at the age of 60 after a span of unemployment of more than 12 months, the Adenauer reform gave workers the assurance that they would not - under any circumstances - have to go through a longer period of unemployment toward the end of their career. Third, the indexation of pensions "surmounted problems of intergenerational equity" (Eichengreen 1994: 49). "Wage moderation now which translated into higher incomes later might benefit future generations at the expense of present ones, a fact which might cause those currently working to hesitate to defer their gratification. Governments therefore indexed pensions not just to changes in the cost of living, but to increases in the living standards of the currently employed" (ibid.).

    7   Cushioning the effects of economic downturns and/or compensating for job losses if a firm at the low productivity end of the market was no longer able to pay the 'above equilibrium' wages was an effective strategy as long as a particular sector (here, for example, the manufacturing sector) was still growing or as long as employment was continuing to shift from agriculture to manufacturing (Paqué 1996, 1998). But once the employment shares of manufacturing shrank in comparison to services, where they remained moderate due to high non-wage labour costs (especially social insurance contributions), the whole arrangement mainly led to an increase in labour market hysteresis and ever-increasing levels of structural unemployment (Manow and Seils 2000).

    8   This is not to say that co-ordination between industries does not take place (Soskice 1990: 44-46). Since wage settlements in other industries follow the key settlement in engineering, the wage hikes in the engineering sector "have to be in line with existing inflation plus economy-wide labour productivity growth (to avoid the risk of a deterioration of international competitiveness as a result of the Bundesbank responding to higher inflation by tightening monetary policy and thus bringing about an appreciation)" (Soskice 1990: 45). This adjustment mechanism, however, only worked in the post-Bretton-Woods era.

    9   For Germany as a whole, IG Metall's membership in the shipbuilding industry was second-highest among all branches of the metal industry in 1961 (53.7 per cent) and highest in 1963 (56.9 per cent) (Noé 1970: 340).

    10   In this context it hardly surprising that the sectorwide lockouts in the Bremen shipbuilding industry in 1953, the first of their kind in any industrial sector in postwar Germany, ended up being just short of an utter fiasco for the employers (Kalbitz 1979: 151).This is one of the reasons why employers in the metal-working industry finally chose the district in northern Baden-Württemberg as the battleground in their fight with IG Metall: this was the only district where employers seemed to be capable of carrying out district-wide lockouts (Noé 1970: 254).

    11   1.8 million were employed in the metal-working sector and 300,000 in the metal-producing industries in 1930 (Wrede 1933: 137).

    12   Part of the AVI agreement was that the Vereinigte Stahlwerke had to sell its machine-tool firm, Thyssen Maschinenbau, in 1926, and that heavy industry promised not to go further into metal-finishing, since the metal-working industry feared (and justifiably so) that an effective price cartel for steel and iron would increase incentives for the steel magnates to further integrate vertically. Metal-finishing subsidiaries of the large steel and iron producers would then have been able to gain an unfair advantage over their independent competitors by being provided with raw material below the cartel price. In 1926, the Thyssen Machine Company was sold to an independent competitor, the DEMAG, 'which immediately shut it down and sold its electro dynamo production to Siemens-Schuckert' (Fear 1997: 213).

    13   Perhaps the foundation of the European Coal and Steel Community (ECSC) in 1945 could also be considered a reason for the changes in heavy industry's political and economic situation in the Ruhr area after 1945. I do not have the space here to fully take the impact of the Schuman plan on the German metal producing industry into account. Yet, the AVI agreement of the interwar years had also been only part of a broader international steel cartel between Germany, France, Belgium and Luxembourg, which lasted until 1937. As has been convincingly shown by John Gillingham (1991), the interwar cartel has to be seen as the forerunner of the ECSC. Thus, in economic terms, not all that much had changed for the Ruhr when the European Coal and Steel Community was founded.

    14   Despite the lack of detailed information on the extent of the cartels' market control, we have quite good evidence that the iron and steel cartels of the interwar years were much more effective than they had been before or during the First World War. They now also controlled most of the markets for semi-processed metal-products. For instance, the pre-war cartel agreements on rolled steel (Walzeisen) had covered an average of 37 to 43 per cent of total production. The coverage increased rapidly after the first AVI agreement came into force (in April 1925 7 per cent of the market for rolled steel was covered by the cartel; in June 1925: 36 per cent; in January 1926: 90 per cent; Preiss 1933: 22-3). A similar picture emerges for other iron and steel products (see, for example, Weisbrod 1978: 101-2; Hoffmann 1928). The establishment and enforcement of cartel agreements became much easier in the second half of the 1920s because of 'the increased concentration within industry and the resulting economic power of the merged enterprises' (Chandler 1990: 512), itself a result of the huge industrial mergers that had taken place in the wake of hyper-inflation.

    15   One important contextual factor for the employers' more aggressive stance in wage bargaining was of course the revaluation of the Deutschmark in 1961, which made German exports more expensive and thus wage control more important. Regarding the political conflict over the Deutschmark exchange rate and the 1961 revaluation, see Kaufmann (1969) and, for a less convincing account, Kreile (1978).

    16   The history of the establishment of German pattern bargaining highlights the importance of the employer's capability to act collectively for the functioning of corporatist labour-capital co-ordination (Soskice 1990) as compared to the factors usually emphasised in the corporatism literature, such as union centralisation and the monopoly of representation enjoyed by employer associations and union umbrella organisations (Schmitter 1979).

    17   The Germans called this 'Mengenkonjunktur' in the 1960s. I think it is interesting to note that the Japanese have coined exactly the same term ('boom in volume') for the pattern of post-war growth with real wage stability, a stable currency and strong export growth. To the extent that wage co-ordination caused employers to get interested in wage compression (see Swenson 1999, 2000), wage co-ordination also met the traditional union call for 'wage solidarity'.

    18   It was not only simple welfare state growth that mattered in this respect, as the conflict over sick pay also made clear. With the law on sick pay (1957) the state for the second time in post-war Germany transferred "social welfare obligations to employers" (Jackson 1993: 453) and thus interfered with their sovereignty to freely design the relation with their employees. The first instance had been the Law of the Severely Disabled (1954), which forced employers to hire quotas of the disabled (mainly war veterans). While one of the hottest points of contention over Weimar's social policy was the extent to which the state could legitimately interfere with employer autonomy, both the Law of the Severely Disabled and the conflict over sick pay made clear that the state was willing to redefine the "workplace as having obligations beyond simply making a profit for its owners" (Jackson 1993: 453). A liberal conception of a strict separation between the private market sphere and the public sphere had lost out against a conception that saw work and the employment relation as embedded within the broader moral fabric of the society (Immergut 1999). From this it was clear that employers would only have a very limited range of options to privately and autonomously design company welfare schemes.

    19   In other words my argument is in line with Barry Eichengreen's (1994) parsimonious explanation for Europe's post-war growth. I claim that in the German case there is indeed ample historical evidence for "implicit contracts between government, employers and unions on wage moderation to maximise social gains in the long run" (Paqué 1995: 19, who denies that such evidence exists). Yet Paqué might be partially right in his critique since 'contract' does not seem to be the right word given that an important part of the entire arrangement (both in Germany and Japan) was the ability of employers to force unions via massive lockouts into national wage coordination and into the acceptance of 'productivity growth + natural inflation' as the benchmark for wage demands.

    20   The Japanese government had taken measures to support the life insurance industry prior to 1964, however. In 1947 it had licensed life insurance companies to sell Group Insurance Plans, and in 1962 the government introduced Tax Qualified Corporate Pension Plans which again only life insurance companies and trust banks were allowed to handle.

Copyright © 2000 Philip Manow

No part of this publication may be reproduced or transmitted without permission in writing from the author. 
Jegliche Vervielfältigung und Verbreitung, auch auszugsweise, bedarf der Zustimmung des Autors.

MPI für Gesellschaftsforschung, Paulstr. 3, D-50676 Köln, Germany




MPIfG:  MPIfG Working Paper 00/7

[Zuletzt geändert am 29.03.2007 11:00]