The (International) Political Economy of Falling Wage Shares: Situating Working-Class Agency

† This article was written under the auspices of ‘löneandelsprojektet’ (The ‘Wage Share Project’) that the authors conducted together with Ingemar Lindberg, Roland Spånt and Bengt-Åke Berg, and which presented its conclusions at the Headquarters of the Swedish Confederation of Trade Unions (LO) in Stockholm, 19 March 2013.

Abstract

This paper relates the financial and monetary dimensions of the contemporary economic crisis to working-class agency via a central concern of classical political economy: the distribution of surplus between the chief factors of production. The fall in the wage share of value added is now accepted as a stylised fact in the empirical economic literature. This paper argues that the punctuated pattern of the development validates the regulation theoretical narrative of an epochal shift from Fordism to finance-led accumulation. Furthermore, synthesising econometric studies supports a class-centred explanation. In the last instance, the falling wage share is due to successful transnational class rule in the form of a neoliberal hegemonic paradigm. Crucially, such class rule restructured the environment of trade unions, rendering increasingly ineffective its relational power resources. The paper concludes by considering the contradictory implications for organised labour of the current financial crisis. On the one hand, the financial crisis offers an opportunity to link its particular interests to the general interest of macroeconomic management since low wage share inhibits growth rates. But how might trade unions assert a higher wage share in the face of the structural power of (financial) capital?

Keywords:

• wage share
• inequality
• class
• income distribution
• finance-led capitalism
• neo-liberalism

Acknowledgements

We gratefully acknowledge the generous financial and logistical support of the Lindley Foundation of the Transport Workers’ Union of Sweden, as well as the logistical support of Arenagruppen and the Northern Office of the Friedrich Ebert Foundation. We alone are responsible for the contents of what is published here.

Notes on contributors

Erik Bengtsson is an economic historian at the University of Gothenburg, researching mainly wages, wage formation and functional income distribution. His recent publications include articles in Economic History Review and the Industrial Relations Journal.

Magnus Ryner is Reader in International Political Economy at King's College London. His publications include Capitalist Restructuring, Globalisation and the Third Way (Routledge, 2007) and Europe at Bay: In the Shadow of US Hegemony (Lynne Rienner, 2007 with Alan Cafruny).

Notes

† This article was written under the auspices of ‘löneandelsprojektet’ (The ‘Wage Share Project’) that the authors conducted together with Ingemar Lindberg, Roland Spånt and Bengt-Åke Berg, and which presented its conclusions at the Headquarters of the Swedish Confederation of Trade Unions (LO) in Stockholm, 19 March 2013.

1. Further measurement problems are generated by what one might broadly call ‘finance’ though very few of these, if any, point to an underestimation of the wage share. Pensions-savings is perhaps the most important of these. It is clear that pensions-savings, especially after pension funds have been integrated as institutional investors into financial markets, imply that some part of investment income benefit workers. At the same time, pension-savings in this form has further commodified labour power and increased risks in life-cycle management. But even in the limiting case of the Netherlands, which stands out as the most labour-inclusive case of integrating pensions-savings into financial markets, pensions do not ‘compensate’ for a declining wage share (Salverda 2011).

2. Duménil and Lévy use as indicator the net output of all firms minus the cost of labour to the net stock of fixed capital after depreciation. The ‘big-three’ western European countries (Germany, France and Britain) are used as proxies for ‘Europe’.

3. Brenner (2006) interprets the cause of the decline of the profit rate in the 1970s somewhat differently. For him, the main cause did not reside on the ‘supply side’ in the form of technological change, but on the ‘demand side’ and in changing terms of oligopolistic competition. The entry of East Asian competitors reduced the scope for price markups. Nevertheless, he agrees that the labour carried the burden of adjustment, and his interpretation of post-1970s developments agrees with that of Duménil and Lévy.

4. Other indicators that Duménil and Lévy invoke for the increased centrality of the financial fraction of capital are the increased ratio of net worth of financial corporations to non-financial corporations (14 per cent in the USA in 1960 and 22 per cent in 2000) and the ratio of funds held by mutual and pension funds to the net worth of non-financial corporations (35 per cent in the USA in 1960 and 160 per cent in 2000).

5. Using GDP per worker, agricultural shares, industrial shares, capital–labour ratios and information and communication technology (ICT) services as proxies for technological change; trade openness and terms of trade as proxies of globalisation; financial globalisation as proxy for financialisation; and government consumption and union density as welfare state effort, Stockhammer (2013: 54–62) found that in the period 1990/4–2000/4 financialisation had the largest negative impact (1.5 per cent) of a series of key variables for a hypothetical average country. By contrast, globalisation and welfare state retrenchment explained about 0.5 per cent each and technological change made a positive contribution to the wage share. This sample included 71 advanced as well as developing economies. Running the same test only for 28 advanced capitalist societies in the period 1980/4–2000/4, financialisation had the largest negative impact of 3.3 per cent. By contrast, the impact of welfare state retrenchment was 1.9 per cent, globalisation 1.3 per cent and technological change 0.7 per cent.

6. Hancké (2012) uses wage share data from the OECD from 1970 to 2010 and shows that the wage share has actually fallen more in what the Varieties of Capitalism school calls ‘coordinated market economies’, countries such as Germany and Sweden, than in ‘liberal market economies', such as the USA and the UK. Hancké’s explanation is that in the coordinated market economies, there have been switches to non-accommodating monetary policy in the period, and the encompassing (Olson 1982) and ‘responsible' trade unions have adjusted their wage policy accordingly, following wage restraint strategies that have led to large decreases in the wage share.

7. Glyn (1997) finds a strong correlation between unemployment and manufacturing wage shares in the 1980s, and makes a Marxian argument about the industrial reserve army. Unemployment is included also in a row of other (non-Marxist) empirical studies: for example, in Daudey and Decreuse (2006) and Hancké (2012), it has a negative effect on the wage share.

8. See, for example, Schor (1985) on the use of restrictive economic and monetary policy to increase unemployment to decrease wage pressure, and the studies referred to in Note 7 for econometric evidence on the negative relationship between unemployment and the wage share.

9. Please recall Note 5, and that Stockhammer identified welfare state retrenchment as the second most significant factor negatively affecting the wage share.

10. Indeed, early reflections on the implications of ICT and computer technology and the coming of ‘post-industrial society’ predicted a massive increase of the public sector (for example, Servan-Shreiber 1969).

11. Rodriguez and Jayadev (2010) use United Nations data for developed and developing countries – the latter is uncommon in the literature – mainly since 1960. They find a statistically significant decreasing trend after 1980 and constancy or a weak positive trend before 1980. They also look for trends per sector and find that within-sector declines are important: they explain 80 per cent of the overall decline in the wage share from 1980 to 2000; the other 20 per cent is explained by sectoral shifts (12). Azmat et al. (2011) use sectoral wage share data for 18 OECD countries between 1970 and 2001 with a focus on ‘network industries': electricity, gas, telecommunications, postal service and transport. They show negative effects on the wage share of privatisation in these sectors.

12. Young (2010), following up on a previous study by Solow for the 1929–54 period, looks at 35 industries from 1958 to 1996. He finds that within sectors the wage share has decreased but at the same time the sectors with higher wage shares (service sectors) have grown as a share of the economy, so the total economy's wage share looks more constant than the within-sector shares do. From 1980 to 1996, Young (2010: 97) finds about a ten percentage point drop in the manufacturing sector wage share but more or less constancy in the service sector.

13. Guscina (2006) and Ellis and Smith (2007) are good examples of how the results from the econometric literature are reconcilable with the story that we tell: their independent variable in focus is the alleged switch from labour-augmenting to capital-augmenting technological change, but they do not have a measure of this. Instead, Ellis and Smith (2007) include a linear time trend in their regressions and Gustina (2006) uses a temporal division in the regressions between the pre-1985 and post-1985 periods, assuming that the revolution in ICT, which supposedly is capital augmenting, started around 1985. Of course, their results of a decreasing wage share over time/after 1985 are equally supportive of our argument about an epochal shift in capitalism, as with their story of technological change.

14. Arpaia and Pichelmann (2008) work from a neoclassical perspective but do not actually present any evidence for their interpretation, and in their conclusions can only claim that it ‘appears' – from their perspective – that skill-biased technological change together with institutional changes and wage moderation strategies is the best explanation for the fall in the wage share. The inclusion of institutional factors and wage moderation too, without any empirical testing of the relative importance of the different factors, means that this paper should not be taken as strong support for the technological change explanation. Which it has in common with the papers referred to in Footnote 13 that also fail to provide actual empirical evidence for this favourite explanation of neoclassical economists. The OECD economists Bassanini and Manfredi (2012) provide slightly stronger evidence but their proxy for technological change is productivity growth and what they prove is that in years with higher productivity growth the wage share has fallen more. This is hardly strong evidence for the technological change explanation, and again the problems of measurement highlights the need, in relation to the extant econometric literature, for an historical–structural and epochal interpretation like ours of the falling wage share: the process is difficult to reduce to separable, small, measurable units.