Wage Stagnation and Productivity: Challenging the Conventional Analysis

Stagnating real wages may have contributed to the slowdown of US productivity

In much of the advanced world, we have witnessed at least three decades of stagnating real wages and massive reductions in the labor share in income. Many analyses have documented these trends, reflecting on their causes and effects from very different standpoints. In the US economy, where the trend toward wage stagnation seems to be particularly strong, it goes together, according to Temin (2015), Storm (2017), and Taylor and Ömer (2020), with a ‘dualistic’ tendency of the economy, with growing polarization between a limited number of high-wage and high-productivity sectors and a growing mass of workers employed in low-productivity and low-wage sectors. Wage stagnation, Taylor and Ömer (2020) note, is also the basis of the growing inequality in personal and family incomes recorded in the USA as well as in many other societies.

The mainstream explanations of these trends generally rely on either globalization (Elsby et al, 2013) or technical progress, especially in the form of jobs-displacing automation (Autor and Salomons 2018, Acemoglu and Restrepo, 2017). Acemoglu and Restrepo (2020) also indicate automation as the main cause of growing intra-wage inequality, affecting especially low-qualified workers. The prevailing idea in the literature, it results, is that a combination of market mechanisms and exogenous shocks have made labor (and especially poorly qualified labor) superabundant with respect to other factors and the economy’s requirements.[1] Implicit or explicit, the message of the analyses that see the change in distribution as essentially the outcome of market forces is that there is little that policies can do, apart from somehow palliating the worst social consequences of the technological or commercial shocks.

Generally, however, these analyses fail to give an account entirely consistent with the data. As shown by Mishel and Bivens (2017), other waves of automation in the past did not result in a decrease in the wage share but rather went together with increased compensation and enhanced work conditions. Globalization and automation, note Stansbury and Summers (2020), cannot easily account for the concomitant extraordinary growth of profits unless other factors are also taken into account. Especially difficult to explain is the fact that the decline in the wage share and the stagnation of real wages have gone together in some periods (and most notably the 2010s) with a sustained growth of employment, also (and especially) of the low-wage variety, which testifies to a strong rather than weak demand for labor.

This leaves room for other, more ‘institutional’ explanations of the distributive shift. The latter may indeed be seen mainly as the product of social forces, i.e. changes in policies, institutions, and collective behavior that have produced a long-lasting loss of bargaining power for workers. It is not only a question of the increase in the oligopolistic power of employers, which enables them to reap as profits most of the productivity gains (Covarrubias et al 2019). Especially, as analyzed by Weil (2014), the diffusion of low-wage jobs in the US economy is related to the profound restructuring of the organization of business that has taken place in a variety of industries and firms since the 1990s, that included outsourcing activities and workers towards firms in which labor is comparatively less protected, and use of employment agencies. Weil especially focuses on the pressure of capital markets on big corporations to produce ‘value for investors’ and the role of information and communication technologies in enabling the coordination among workers of different firms. But policies have played an essential role in shaping the new institutions now prevailing on the labor market. As noted by Stansbury and Summers (2020) “changes in policy, norms, and institutions” have produced a “legal and political environment … in favor of shareholders and against workers.” The authors enumerate the laws “undermining unions’ ability to fund themselves and the increasing corporate use of union avoidance tactics, both legal and illegal.” The policy-induced loss of workers’ power over many decades can explain the comparatively prosperous periods as the 2010s have been characterized by wage stagnation.

This different conception of distribution as essentially shaped by social and institutional forces – which appears as supported by empirical evidence – has very different implications in terms of policy. Before coming to that, it is worth noting that, however divergent the diagnoses, not only is the decades-long trend itself towards wage stagnation little disputed in the literature, being evident in data, but there is also a certain consensus that something should be done. Many different analysts have advocated the need for measures that could at least in part check the tendency to wage losses, if only to avoid the worst effects of growing inequality in terms of social disintegration.

Against this background, it would be legitimate to expect that the recovery in nominal wages which has materialized in recent months in the USA should be saluted with relief and approval by economists of different persuasions, as a first sign of a changed situation and the beginning, perhaps, of an inversion of the long-term trend.

This is not so, as is well known. On the contrary, tight labor markets and rising wages, supposedly mainly brought about by excessively generous fiscal policy (especially in the USA), are now blamed, by more than one commentator, as the main source of unsustainable inflationary pressures. Among the most vocal advocates of the need to rein in wage growth, through a sustained increase in interest rates aimed at increasing unemployment, is the same Larry Summers who in Stansbury and Summers (2020) had shown how dramatic has been the loss of workers’ power in the recent decades. It is very difficult to imagine that such persistent causes of weakness as analyzed there have suddenly evaporated, albeit under the impetus of the exceptional post-covid conditions, so as to result now in an excess of workers’ power, with the inflationary consequences that this supposedly entails.

Though evoked as a threat to price stability, the occasional tightness of the labor market, which, as shown by Storm (2022), is in fact mainly the result of the enduring restrictions covid has imposed on labor supply – the continuing infection risks, the problems associated with childcare, the people kept away from actively seeking work by long-covid symptoms – is actually only a secondary cause of inflation and, especially, does not seem able to invert in any significant way the long-term trend towards real wage losses.

Actually, real wages have not risen in 2021 and are falling right now, both in the USA and in Europe, despite the historically low rates of unemployment. Moreover, the interest rates hikes currently taking place in the USA – and likely soon to be followed by similar rises in Europe – will in all probability be effective in raising unemployment, although this might occur through channels different from those traditionally indicated by textbooks (i.e., not through the disincentive to investment, but possibly through the impact on credit-financed consumption and the housing market). Combined with the economic consequences of the war in Ukraine and the current international instability, this is not unlikely to generate a recession. However ‘soft’ may be the landing, it will not increase labor’s strength. The trend towards compression of real wages and impoverishment of labor does not seem to be close to an end.

If wages are a problem, it follows, it is not the rise in nominal wages that should capture public attention, but rather the enduring weakness of real wage growth.

It is not only for reasons of fairness and social cohesion – which of course are of paramount importance because a social arrangement based on the increasing impoverishment of the largest part of the population obviously cannot be desirable – that wage stagnation should be regarded as a serious problem. It also produces economic effects that may have negative long-lasting consequences. In our new INET working paper, we focus on one of such possible effects and explore how stagnating real wages may have contributed in recent decades to the slowdown of US productivity.

In much literature, the relationship between wages and labor productivity is seen as unidirectional. If productivity growth may fail in some circumstances to be transmitted to wages, the possibility of reverse causation, whereby the dynamic of wages may be an important determinant of productivity growth, is generally little explored in the mainstream approach. This is due to the conception of distribution as endogenous to market forces (the relative scarcity of factors and their efficiency). Wages may be affected by productivity since slow productivity growth leaves less room for wage increases, but cannot generally exert an independent positive influence on productivity. Rather, if imperfections in the market mechanism or the interference of labor market institutions determining excessive workers’ power should cause wages to grow exogenously, this would imply increasing unit labor costs for the firms, loss of competitiveness, and loss of employment, with negative effects on growth. Productivity, in this theoretical framework, is typically seen as an indicator of technological efficiency, with its growth determined by technical progress or other supply-side factors as the accumulation of human capital.

In our paper, we start from a demand-led growth perspective, in which aggregate demand has a prominent role in determining both output growth and the growth of resources. Such an approach, we show, implies that productivity should be reconsidered, both in its definition and in its determinants. Far from being merely a measure of technological efficiency, observed output per unit of labor is also a reflection of the intensity of the use of resources, an intensity which is determined endogenously by the conditions in which the economy operates. The system is not endowed with mechanisms that continually ensure the tendency to produce close to the efficiency frontier, but the conditions of demand may well produce inefficiencies and waste of resources. It is within this flexible theoretical framework, in which productivity becomes a largely endogenous variable, that we explore the possible endogeneity of productivity growth to the dynamics of wages.

Our analysis takes inspiration from a conception of distribution – typical of the classical political economy of Smith, Ricardo, and Marx – as affected by social and institutional forces that are at least in part exogenous both to output growth and to productivity growth. This theoretical conception is perfectly compatible, it is hardly necessary to note, with the above-mentioned analyses that have documented the weight of political and institutional factors in shaping the trend of distribution in recent times.

In this theoretical context, wages may affect productivity not only indirectly, by expanding or depressing aggregate demand, but also directly. Following the analysis of Paolo Sylos Labini (1984, 1993), we analyze and distinguish in the paper two different direct effects of wages on productivity: one acting through induced technical progress – labor-saving innovations fostered by high wages – and a second one acting through the incentive that high wages represent towards a more efficient use of the labor input through reorganization of the production processes.

The first effect is triggered by an increase in the price of labor relative to the price of machines. Sylos especially underlines the role of competition, both in the domestic and in the international market, that forces firms, in a high-wage environment, to innovate in order to defend or increase their market shares. For its characteristics and its dynamic character, such an effect is very different from the neoclassical substitution mechanism based on the production function. The second effect is instead induced by an increase in wages relative to the price of output. This incentivizes firms towards a more efficient use of labor, through reorganizing production processes or work practices. Sylos labels this effect as the ‘organization’ effect. This effect is also conceived as acting dynamically since it entails a particular kind of innovation, i.e., ‘organizational’ innovations. Differently from the innovations entailing increased mechanization, these organizational innovations do not necessarily involve high investments in fixed capital and may thus take place more immediately (empirically, indeed, Sylos distinguishes the organization from the mechanization effect also for its faster action).

Our hypothesis is that these two effects of wages have each contributed to the productivity slowdown observed in the US economy. The dramatic distributional shift against wages has in our view reduced the incentive for technical innovation, at least in some sectors or sub-sectors of the economy. Moreover, it has constituted a strong incentive for firms in different sectors to build their business strategies by taking advantage of the mass availability of low-wage labor. Such strategies may involve using massively the labor input, often with the intermediation of other firms, which also implies the redistribution of employment, within the same sector, between firms with different work arrangements and different levels of productivity, thus entailing transformations along the various phases of the value-chains. To the workers, this may produce long and little-paid working hours, sometimes with the intermediation of types of contracts where the worker appears as self-employed or rather has multiple employers.

By analogy with farming techniques, we propose to label these strategies as based on ‘extensive’ vs ‘intensive’ use of labor in the production processes and the business organization. Extensive use of the labor input implies using cheap labor abundantly (in terms of heads or working time). Although such phenomenon has affected the various sectors with different intensity, we believe that the drift towards a more extensive use of the labor input is not limited to traditionally low-wage and low-productivity sectors, but may well have affected all those sectors which are susceptible to different forms of organizing production, including some important sectors in manufacturing, traditionally a high-productivity branch that has lately shown a slowing down of its productivity dynamics.

In the paper, we conduct two different exercises to test empirically our theoretical hypothesis on the US economy. In the first place, we propose a shift-share analysis of the dynamics of both productivity and wages in the USA in recent decades, to identify the main sectoral sources of change in the two variables. We also investigate the sectoral contributions to the cumulative growth of the wage-productivity gap. Through the shift-share analysis, we find that structural change may explain only a minor part of the weak dynamics of productivity, which is mainly due to factors acting within each sector. We also find that after a sharp change in distribution against wages, some historically high-productivity sectors switched towards slower productivity growth. This supports our hypothesis that the anemic growth of productivity may be partly due to the trend toward massive use of cheap labor. On the basis of this analysis, we advance the hypothesis that the disappointing growth in productivity is not only a structural phenomenon due to the growing weight of low-wage and low-productivity sectors but is also due to the recent trend towards the extensive use of labor even in traditionally high-productivity sectors. Indeed, after the 2008-09 crisis, a shift to a regime of low productivity growth seems to take place also in manufacturing.

Considering the central role of manufacturing in the US productivity slowdown, in the second place we focus our analysis on this sector, estimating the effects of wages on productivity through the productivity equation originally proposed by Paolo Sylos Labini. Our estimation of Sylos Labini’s productivity equation confirms the existence of two direct effects of wages, one acting through the incentive to mechanization and the other through the incentive to the reorganization of labor use. We also include two indicators that measure labor “weakness” in terms of duration and reasons for unemployment and we find that both stagnating wages and increasing labor insecurity do indeed contribute significantly to the decline of productivity.

This shows that the persistent deterioration in real wages and the wage share does not only raise social issues but may also produce permanent macroeconomic scars, in terms of long-term negative effects on the growth prospects of the economy.

The question is what could and should be done. As remarked above, if wage losses are regarded as the spontaneous outcome of market forces, they may be countered only in part, for example. by enhancing education and training programs for workers. By contrast, the idea that distribution is especially the product of social forces implies that society has the means, through the adoption of suitable policies and the reform of institutions, to change the course of events. As noted by Stansbury and Summers (2020), for example, “progressive institutionalists have long argued for pre-distribution alongside redistribution, strengthening worker power by changing the structure of labor market institutions…. Strengthening worker power can be an important countervailing force against firms’ dominance in product and labor markets… Overall, we believe that increasing worker power must be a central and urgent priority for policymakers concerned with inequality, low pay, and poor work conditions. If we do not shift the distribution of power toward workers, any other policy changes are likely to be short-term and insufficient.”

The problem of wages and work conditions is now crucial. Higher wages and safer and better jobs would not only re-orient firms towards higher productivity and more efficient organization. They would possibly also represent a serious incentive towards enhanced labor supply, thus easing one of the constraints currently concurring to inflation. The lasting wounds that the regime of low wages inflicts on the economy, in addition to what it entails for the greatest part of society, should be a matter of serious concern.


Acemoglu, D. and Restrepo, P. (2017), “Robots and jobs: evidence from US labor markets”, NBER Working Paper 23285.

Acemoglu, D. and Restrepo, P. (2020), “Unpacking skill bias: automation and new tasks”, NBER Working Paper 26681.

Autor, D, and Salomons A. (2018), “Is automation labor-displacing? Productivity growth, employment, and the labor share”, Brookings Papers on Economic Activity, Spring, 1-63.

Covarrubias,M., Gutiérrez, G. and Philippon, T. (2019), “From Good to Bad Concentration? U.S. Industries over the past 30 years”, NBER Working Paper 25983.

Elsby M.W.L., Hobijn B. and Şahin A. (2013), “The Decline of the U.S. Labor Share”, Brookings Papers on Economic Activity, 1-52.

Lawrence, R.Z. (2015), Recent declines in labor’s share in US income: a preliminary neoclassical account, NBER Working Paper 21296

Stansbury, A.M. and Summers, L.H. (2020), “The declining worker power hypothesis: an explanation for the recent evolution of the American economy”, NBER Working Paper 27193

Storm, S. (2017). The New Normal: Demand, Secular Stagnation and the Vanishing Middle-Class. Institute for New Economic Thinking, Working Paper 55;…

Storm, S. (2022). Inflation in the Time of Corona and War, Institute for New Economic Thinking, Working Paper 185;

Sylos-Labini, P. (1984), Le forze dello sviluppo e del declino, Editori Laterza, Bari.

Sylos-Labini, P. (1993), Progresso tecnico e sviluppo ciclico, Editori Laterza, Bari.

Taylor, L. and Ömer, Ö. (2020); Macroeconomic Inequality from Reagan to Trump: Market Power, Wage Repression, Asset Price Inflation, and Industrial Decline, Cambridge University Press, Studies in New Economic Thinking.

Temin, Peter. 2015. The American Dual Economy: Race, Globalization, and the Politics of Exclusion, Institue for New Economic Thinking Working Paper No. 26;…

Weil, D. (2014), The Fissured Workplace. Why Work Became So Bad for So Many and What Can Be Done to Improve It, Harvard University Press.


[1] See also Lawrence (2015) who maintains that the analytical tools of standard neoclassical theory are more than sufficient to explain the persistent contraction of the labor share if one assumes a reduction in the elasticity of substitution between capital and labor and a labor-augmenting character of technical progress.

Share your perspective