Abstract
The declining gender wage gap (GWG) and the wage-productivity decoupling (WPD) are two emerging phenomena for the United States since the 1980s. This article proposes an extension to the usual decomposition method of the profit rate to account for both the GWG and the WPD on aggregate income distribution. It then uses this framework to study the US manufacturing sector from 1960 to 2017. It suggests that gender wage inequality was a source of profitability, especially before 1986, but which was slowly petered out. Since 2001, increases in the profit share have mostly been due to the WPD. These results highlight the relevance of gender-based inequality in the more traditional analysis of the profit rate dynamics.
1. Introduction
Two crucial phenomena related to wage growth have been observed in the United States persistently since the 1980s, both of them having impacts on income distribution between capital and labor, and the rate of profit. On the one hand, the gender wage gap (GWG) has been declining, possibly negatively affecting capital profitability. Elveren, Marr, and Renard (2017) find evidence that higher women’s labor force participation rates, in the presence of a GWG, leads to higher profit rates. Since women get paid less than men, capital could benefit by hiring women and reducing unit labor costs. On the other hand, wages stopped increasing at the growth rate of labor productivity since the 1970s but consistently since the 1980s, reducing output’s wage share. This wage-productivity decoupling (WPD) has been a topic of discussion since the 1990s when it was first observed that real wages stopped growing at the same pace as labor productivity during the two previous decades (Ferguson 1996; Zavodny 1999). Taken together, these two phenomena are associated with contradictory impacts on income distribution: the declining of the GWG has been associated with a burden of capital profitability, whereas the WPD has been associated with a boost of profits. Given the worthwhileness of asking about the contradictory impacts of these two wage phenomena, what has been their overall effect on profits during the neoliberal era in the United States?
This article proposes an extension to the usual decomposition method of the profit rate by including the GWG into the analysis. With it, I attempt to shed some light on how the combination of changes in both the GWG and the WPD has shaped the two aggregate outcomes relevant for the proposed discussion: the wage share and the rate of profit.
This analysis focuses on the US manufacturing business sector for the period 1960–2017 because of data availability. It applies this extended decomposition method to the rate of profit and income shares by gender and between capital and labor. Using this perspective, this article approaches the impact of the narrowing of the GWG and the WPD on both the rate of profit and the distribution of income in this sector. A plausible hypothesis is that the increase of the WPD has been a counterweight to the decline in the GWG. In other words, as the catch-up of female workers to their male counterparts is harmful to capital in terms of its profitability, the WPD could be understood as a response to maintaining a profitability level.
One of the main objectives of this analysis is to emphasize the relevance of gender inequalities when studying the rate of profit. The more traditional decomposition method considers the dynamics of labor productivity, average wage, and capital/labor ratio (Basu 2017: 1375). However, wage differences among workers also play an essential role in the tendencies of the profit rate, as in the case of gender inequalities. As Seguino (2019) points out, incorporating different gender variables into macroeconomic models improves the relevance of the theory and the policy results.
The rest of the article is organized as follows. Section 2 reviews the discussions around the factors behind the GWG and its implications on profitability and income distribution. Section 3 presents a review of the debates on WPD and the different approaches to measuring it. Section 4 exposes the extension to the traditional decomposition method that includes the GWG. Section 5 presents the data sources. Section 6 applies the decomposition method to analyze the impacts of the GWG and WPD in the rate of profit and the functional income distribution within the US manufacturing sector. Finally, section 7 concludes that there are three main periods identified in which the GWG and the WPD influenced the rate of profit of the manufacturing sector: 1960–1986, where the increasing female share of employment and a relatively stagnant GWG might have contributed to the rate of profit, 1986–2001, where this source of extra profitability was depleted by a closing of the GWG and lower participation of female workers in the sector, and from 2001–2017 where the WPD in the average workers’ total compensation helped to sustain the rate of profit.
2. The Gender Wage Gap
The GWG is the measure of the average differences in income between female and male workers. It has been associated with labor market discrimination, which “exists when two equally qualified individuals are treated differently in the labor market solely on the basis of their gender (race, ethnicity, age, disability, sexual orientation, etc.)” (Blau and Winkler 2018: 264). However, the reasons behind the GWG are complex.
Some economists argue that the GWG reflects differences in productivity and not labor market discrimination. Several empirical studies have controlled for different variables as the level of education, experience, and other differences in qualifications that could explain the disparities in wages by gender but that are not necessarily related to discrimination in the labor market. However, even when controlling for these factors seems to account for a portion of the wage differential, a significant share of the gap appears to respond to nonqualification variables such as unionization, race, and industry, while most of the gap remains unexplained. Different authors associate this unexplained portion with labor market discrimination (Blau and Kahn 2017: 831). In other words, some of the nonqualification variables plus the unexplained part of the gap suggest that females and their male counterparts are developing a similar job but having the former earning less than the latter. The following paragraphs take a closer look at each of these factors.
Differences in the unionization rates across genders used to have some explanatory power in the GWG. Blau and Kahn (2017) find that unionization explained around 6.2 percent of the GWG in 1980, but this factor has lost importance through time. By 2010 it completely lost its explanatory power. This finding makes sense considering the continuous decline in unionization rates in the United States, at least since the 1960s (Zavodny 1999), and the decrease in collective-bargaining coverage. As it was more common for men to be in a union (Schur and Kruse 1992), they had a higher bargaining power than women. However, these differences in unionization rates have declined to the point that by 2015 the union membership rate in the United States was higher for women than men (Card, Lemieux, and Riddell 2020), and therefore losing its explanatory power over the gap. In sum, this dynamic of unionization rates across gender is one of the reasons to explain the closing of the GWG.
Blau and Kahn (2017) also present some other nonqualification factors that help to explain the GWG and that are relevant for the analysis in the following sections. First, race increased its explanatory power in the gap, from 1.6 to 6.6 percent when comparing 1980 to 2010. Second, the wage gap also varies across industries (Fields and Wolf 1995; Horrace and Oaxaca 2001). Given these differences, industry explains around 17.6 percent of the aggregate GWG by 2010. Third, it is also crucial to notice which groups of workers have led the GWG to close. For the period 1980–2010, the GWG has closed mainly among the low- and middle-income workers.
Considering the factors associated with the GWG, the question is, What are its effects on the income distribution between capital and labor, and also on the rate of profit? Different studies have associated the presence of gender income inequality with a source of either extra benefits for capital or as a means to cut unit labor costs and be more competitive in the market. 1 For example, Seguino (2000) finds evidence that women’s lower relative wages were a stimulus to investment, exports, and rapid economic growth experienced by several Asian economies during the period from 1975 to 1990. One of the key ideas is that investment is positively linked to gender wage inequality due to higher capital return rates.
There are contradictory approaches to analyze the effect of the GWG on the profitability of capital theoretically. Seguino (2019) points out that investment is positively related to both profits and demand. This is similar to the post-Keynesian approach dealing with two different accumulation regimes: profit-led or wage-led regimes (Lavoie and Stockhammer 2013). In the first regime, a wider GWG would boost profits because of the lower unit labor costs. In the second kind of regime, a wider gender wage inequality could imply lower profits because of the shortage of aggregate demand. Therefore, the net effect is theoretically ambiguous. Nevertheless, Seguino (2019) notes that there is little evidence in the literature that the demand effect dominates the profitability effect.
Theoretical developments have also provided insights into these questions. For example, Hein (2020) proposes a neo-Kalekian distribution and growth model for advanced capitalist economies where a GWG is introduced. The model suggests that for a closed economy, narrowing the GWG will not affect income distribution between profits and wages but only improve the female wage share at the expense of the male wage share, in the case when factors that determine the firms’ mark-ups remain constant. In an open economy, the model shows that narrowing the GWG reduces the profit share. The reason is the higher costs and loss in competitiveness associated with higher female wages.
Several empirical studies have also shed light on the relationship between gender wage inequality and the aggregate income distribution between capital and labor. Finnoff and Jayadev (2006) find a negative correlation between the increase in the females’ share of employment and the wage share of national income in the Organisation for Economic Co-operation and Development (OECD) countries for 1975–2000. In contrast, Zacharias and Mahoney (2009) find that women’s participation rates contributed to offset the wage share loss in the US case during the period 1982–1997. The idea is that the decline in the GWG during this period prevented the decline in the aggregate wage share. They use a decomposition methodology to separate the effects of the trends in women’s employment and wages on the labor share and the rate of profit. Tescari and Vaona (2014) perform a similar decomposition analysis to Zacharias and Mahoney (2009) of the wage share and rate of profit for the Italian economy for 1994–2005 and find similar patterns. Finally, Elveren, Marr, and Renard (2017), using panel data for twenty-one OECD countries from 1970–2008, find evidence supporting the idea that higher women’s labor force participation rates and GWG lead to higher profit rates. Their intuition is that lower unit labor costs, given the unequal pays by gender, helps to boost profits.
However, these empirical studies mentioned in the previous paragraph have a limitation, given the aggregate nature of the data: they do not control for the qualification variables as level of education or experience that affect the GWG. These studies, therefore, implicitly assume that the GWG comes mainly from discrimination in the labor market, and thus capital could save unit labor costs by hiring female workers. 2 However, as covered above, nonqualification factors such as race, unionization, the level of income, plus the unexplained portion of the gap represent most of the GWG, supporting the idea that companies could replace males with female workers and saving in unit labor costs as a result. Thus, these empirical studies should be understood as an approximation to the problem but not as precise measures, as in the case of this study.
3. The Wage-Productivity Decoupling
Since the early 1970s, wages stopped growing at the same rate as labor productivity in some industries in the United States. The WPD is a tendency that can also be seen in some of the major economies worldwide (ILO 2018; OECD 2018).
For the mainstream theory, this is a result of technical factors since it expects that each factor of production gets paid according to their contribution to the output. Cobb and Douglas (1928) tested the neoclassical theory of income distribution for the first time, and reported evidence in favor of the idea that wages are paid according to labor’s marginal productivity. However, following works have shown that neoclassical theory of income distribution has no strong empirical support since the neoclassical production functions are a transformation of an accounting identity, which make their good apparent empirical results to be due a tautology (Phelps Brown 1957; Shaikh 1974; Felipe and McCombie 2014). Other schools of thought (including the Classical, Marxist, and post-Keynesian) understand income distribution as a result of a class dispute and power relations, and not as technical market result (Cakmakli, Cantillon, and Ucal 2020). For these schools of thought, the WPD has to do with the weakening of trade unions, loss of collective bargaining power, and poor labor conditions that came along with neoliberalism (Arestis 2020; Cauvel and Pacitti 2022).
This section presents a literature review about the WPD in the US economy and the discussions about its measurement. Works, such as Cantillon and Ucal (2019), have shown that its size depends heavily on the variables used to measure it, as the use of average or medium compensation, as well as the deflator used for wages. The idea is to go in this direction and cover all the possible variables that affects its size and their implications. With the literature review, it is clear why there are misconceptions about the WPD and why some authors, such as Škare and Škare (2017) and Pessoa and Reenen (2013), find no decoupling between wages and productivity in most of the major economies in the world, and why they underestimate it for the US economy. The aim is to set the ground for the analysis in the following sections.
Ferguson (1996) presents one of the first papers that point out the increasing gap between wages and labor productivity in the United States. Considering wages for production and nonsupervisory workers for 320 nonagricultural and nongovernment industries, he finds that the wage-productivity gap between 1981 and 1988 increased by 8.24 percent. The author associates this gap with the fall in the bargaining power of workers. He uses linear regression to test this idea using variables such as unionization, unemployment, and import penetration.
Later, Zavodny (1999) found that the gap between real hourly earnings and productivity in the nonfarm business sector has widened since the early 1970s. However, “increases in real compensation per hour have more closely matched productivity gains than have wages and salaries alone, but compensation increases have lagged behind productivity gains since the mid-1980s” (Zavodny 1999: 46). Zavodny also takes a look at this gap across industries. The author finds that the cumulative difference between productivity and compensation growth between 1977–1996 differs across sectors. While the gap increases in mining, manufacturing, transportation, communication and public utilities, and wholesale and retail trade, she does not find the same tendency for finance and services industries. Her findings suggest that the WPD is smaller in industries with higher initial unionization rates, an idea aligned with Ferguson (1996).
In this same vein, Levy and Temin (2007) construct a bargaining power index using the wage-productivity gap. As in Ferguson (1996) and Zavodny (1999), the idea is that wages not growing at the rate of labor productivity reflects workers’ bargaining conditions. According to the estimations of this gap for 1945–2005, the decoupling began in the 1970s and accelerated in the 1980s. They use median annual earnings of full-time workers and the value of estimated fringe benefits, and the labor productivity in hours for nonfarm businesses.
Other works have also presented estimations of the WPD for the aggregate economy or specific sectors. Here are some relevant examples. From the Economic Policy Institute, Bivens et al. (2014) use productivity in terms of output per hour and the real hourly compensation for production and nonsupervisory workers. Using data for the period 1948–2013, they show that the gap began to increase since the mid-1970s and increased consistently since 1979. Another report by the Economic Policy Institute, Gould (2019), finds that while productivity has increased 69.6 percent for 1979–2018, the hourly pay has risen just 11.6 percent during the same period. Finally, Shaikh (2016: 60) presents that the gap between labor productivity and real compensation per production worker has increased since 1979 in the manufacturing sector for the US economy. The most intriguing fact of his work is that he presents data for this sector from 1889–2010, showing that workers’ compensation and productivity increased closely during 1889–1979 and began to diverge only since then.
Nevertheless, some authors sub-estimate the size of this gap pointing out to discussions around its measurement. There are two main critics in the literature about the measurement. The first one is about taking wages instead of total compensation of workers. As already pointed out by Zavodny (1999), the gap is much higher when it is measured using wages instead of total compensation. This difference in the patterns between wages and total labor compensation is at the center of the discussions about the WPD, as presented by the Global Wage Report 2014/15 (ILO 2015). This is a point that some authors, for example, Feldstein (2008), use to disesteem the WPD. The second one, argued in Pessoa and Reenen (2013), posits how different the decoupling conclusions are if the worker’s compensation is not deflated by using the Consumer Price Index (CPI) but the GDP deflator. In the case of the United States, they show how using the CPI as a deflator for wages, the gap for the period 1972–2010 is around 80 percent, but when using the GDP deflator, the gap is only about 13 percent. They call the former “gross decoupling” and the latter “net decoupling.” The authors argue that the latter is the relevant measure since this is the one that accounts for changes in the wage share of output. The Global Wage Report 2014/15 follows this idea but finds that the differences are less dramatic. The wage-productivity gap for 1999–2013 is almost half when the GDP deflator is used instead of the CPI (ILO 2015). Additionally, the Economic Policy Institute (2021) measures the gap for the US economy using GDP deflator and still finds it has been persistent when other variables in the measurement, different from the ones used by Škare and Škare (2017) and Pessoa and Reenen (2013), are used.
The magnitude of the gap is affected by the group of workers included into the measurement of the gap, a difference not considered by the mainstream literature. The wages and compensations of workers with managerial positions tend to be higher and evolve differently compared to nonsupervisory workers (Duménil and Lévy 2011; Elsby, Hobijn, and Sahin 2013; Giovannoni 2014). This point could be addressed using two different approaches. The first one is by directly including into the data only nonsupervisory workers. The Economic Policy Institute (2021) shows that the wage-productivity gap is dramatically higher when accounting only for nonsupervisory workers. This group of workers is the one that has suffered more from losing the fruits of economic growth. By taking into account the evolution of the compensation of workers as a whole, you miss the increasing inequality within companies in the United States. A second approach is to use median labor compensations, instead of the average. These two measurements depend on the concentrations of income among workers. The use of average compensation underestimates the gap with productivity compared to the median compensation. With this distinction, the Schwellnus, Kappeler, and Pionnier (2017) find that for the period 1995–2013, for 24 OECD countries, the gap of real median compensation is significatively broader than when using the average.
Finally, Baker (2019) points out the relevance of the measurement of output considered for productivity changes. Because GDP does not account for capital depreciation, it is more precise to measure productivity using NDP. Also, he discusses the point made by Pessoa and Reenen (2013) about the different deflators for wages. He recognizes that the CPI generally shows a higher inflation rate than the output deflator (GDP or NDP). However, he points out that part of the reason is how the deflators are constructed. While the CPI is a fixed-weight index, the NDP deflator is a chained index. This difference tends to make the former larger than the latter, even if they are including the same basket of goods and services. A second factor that accounts for that difference is they include a different mix of goods and services. The NDP deflator includes investment goods, which price has grown less rapidly compared to consumer goods.
Table 1 summarizes the different measurements that could be used for each variable to estimate the WPD. As revealed by the literature review, the conclusions about the magnitude of the decoupling change according to the chosen measurements for the variables. First of all, considering all workers, instead of only nonsupervisory workers, underestimated the gap (Economic Policy Institute 2021). In the case of workers’ income, the productivity gap is usually higher when using wages instead of total compensation (Zavodny 1999; Feldstein 2008; ILO 2015). Taking the average income instead of the median underestimates the gap (Schwellnus, Kappeler, and Pionnier 2017). For the deflator for workers’ income, the use of CPI shows a much wider gap than using one of the output deflators (Pessoa and Reenen 2013; ILO 2015; Baker 2019; Economic Policy Institute 2021). Finally, no relevant differences in the gap seem to be reported when measuring productivity per hour or per worker.
Different Measurements to Estimate the Wage-Productivity Gap.
4. Decomposition Method
Basu (2017: 1375) presents a literature review about the quantitative empirical research in Marxist political economy. In the section dedicated to decomposition analysis, there is a subsection dedicated to medium-run decomposition analysis. He mentions that to study the drivers of profitability over a medium-run time scale, the usual analysis is done using a decomposition of the profit rate into the share of profit and the output/capital ratio (see, e.g., Shaikh 1992; Duménil and Lévy 2002; Bakir and Campbell 2006; Basu and Vasudevan 2013). Additionally, the author presents that the decomposition is sometimes taken further from these two components into subcomponents. From here, a more detailed analysis of the profit rate tendencies is usually done by decomposing it on the labor productivity
This approach is the more general framework to analyze the rate of profit in particular and the disputes for income distribution in general. As it is, it does not account for differences in wages among workers. The proposal here is to disaggregate wages by gender. With this, it could be possible to account for gender wage discrimination and differences in labor shares of employment by gender.
It is possible to redefine equation 1 by including the average wage for females
The GWG (
It is also possible to define the female share of employment
Replacing
Equation 5 shows that with an absence of a GWG, when
However, in the presence of a GWG, it is known that
This approach can also be used to decompose the wage share, taking into account gender wage differences. It allows analyzing the impacts of the GWG and the female share of employment on the wage share. Besides, it is also possible to account for the dynamics of wages and productivity. Both need to increase at the same rate in order to keep the wage share constant. An increasing wage-productivity gap means that a higher portion of the output goes to profits:
Finally, this approach makes it possible to analyze the decoupling of labor productivity and wages by gender. This decoupling in part reflects that the share of the total output is received by female and male workers separately as in equations 7 and 8, respectively. They provide information about the distribution of income, not only by class but also by gender:
5. Data Sources
Data are restricted to the period 1960–2017 for the US manufacturing sector. This includes corporate and noncorporate manufacturing businesses. Data on the GWG comes from two different sources. The Comparative Family Policy Database has a series of wages by gender in the US manufacturing sector from 1960 to 2008 (Gauthier 2010). An estimate for the period 2009–2017 is done by using data from the US Department of Labor on median annual earnings by sex. 3 Data on the number of workers in the manufacturing sector by gender comes from the US Bureau of Labor Statistics. The CPI comes from the Federal Reserve Economic Data (FRED).
The rest of the data comes from the Bureau of Economic Analysis National Income and Produce Accounts (BEA-NIPA) tables. The level of output comes from table 6.1: National Income Without Capital Consumption Adjustment by Industry (B, C, and D). The capital stock comes from table 3.1: ESI Current-Cost Net Stock of Private Fixed Assets by Industry. On wages, there are two possible approaches: either using total compensation of employees, which includes wages and salaries, and supplements to wages and salaries, or only wages and salaries. Because this article aims to approach the rate of profit in its broader sense (output minus labor compensation), the analysis is done using the total compensation of employees. The data used comes from table 6.2 (B, C, and D). The output deflator for nonfarm business comes from table 1.9.4: Price Indexes for Net Value Added by sector. To control for part-time and full-time workers, I use table 6.5: Full-Time Equivalent Employees by Industry.
6. Decomposition Analysis
The evolution of the GWG is presented in figure 1. Four main periods seem relevant in the movement of this variable. Between 1960 and 1980, the gap was relatively stagnant. From 1981 to 2000, the GWG gap declined. For the period 2001–2012, it was relatively stagnant again, and then began to continue falling since 2013. A significant part of the decline in the GWG is explained by the increasing level of schooling and years of experience in the aggregate economy, as Blau and Kahn (2017) find. However, as their article also points out, even when controlling for qualification variables, there is still a decline in the GWG associated with changes in nonqualification factors as the evolution of the unionization rates across genders.
Gender wage gap: US manufacturing sector, 1960–2017.
The second central variable for the analysis is the share of female workers in the total number of employees in the manufacturing sector. As shown in figure 2, it presents an inverted-U shape for most of the period. The share of female workers increases steadily from 1960 to the beginning of the 1990s. Since then, its share began to fall up to 2013, where it started to grow again up to the most recent data in 2021.
Female share of employment: US manufacturing sector, 1960–2021.
With the data on the GWG
Figure 3 suggests that gender wage discrimination might have provided additional benefits to the rate of profit throughout the entire period. However, those extra benefits, the difference between the actual and counterfactual profit rates, have been getting lost over time as the GWG have been closing. Some caution should go into interpreting these results. As discussed in section 2, the GWG responds to multiple factors, including, at least partially, differences in qualifications. This means that for capital it is not always possible to replace a male worker with a female counterpart and save the total amount of the GWG on average. Nevertheless, the nonqualification factors explain an even more significant portion of the GWG. First of all, this article focuses only on the manufacturing sector, which sets aside a portion of the gap explained by industry. Second, part of the gap and its decline is explained by the drop in the differences in unionization rates by gender. Third, discrimination based on race also has an important role. Finally, most of the gap remains unexplained, something associated with labor market discrimination. Therefore, this analysis should be taken as an approximation and not as an exact measure of the problem, considering the limitations of the data. The consequence of this exercise is a possible overestimation of the extra benefits obtained by capital coming from the gender wage discrimination.
Actual and counterfactual rates of profit: US manufacturing sector, 1960–2017.
Figure 4 presents a clearer picture of the patterns in the ratio of the difference between the factual and counterfactual rates of profit, which I am calling extra benefits. From 1960 to 1986, the extra benefits increased. This was the same period when the GWG was stagnant, and the female labor share of employment rose. The combination of these two effects might have increased the additional benefits by reducing unit labor costs. The additional gains began to fall afterward as the GWG narrowed, and the female labor share of employment began to decline. The combination of these two phenomena might have decreased the extra benefit for capital coming from the gender wage discrimination.
Ratio of the difference between actual and counterfactual rates of profit over the actual: US manufacturing sector, 1960–2017.
Taking a look at the WPD also provides important insights. As covered in section 3, the dimension of the decoupling depends on the different measurements for each the variables to be considered. The present analysis uses the following measurements. I consider all workers, that is, supervisory and nonsupervisory. For worker’s income, I use the average of the total labor compensation. For the deflator of worker’s income, I present the results of using the CPI and the net output deflator, considering the debate proposed by Pessoa and Reenen (2013) and the response by Baker (2019). For output, I use net output instead of gross. And finally, productivity is measured as output per worker, and because I am using full-time equivalent employees, it is equal to the productivity per hour.
There is a limitation with this approach. As said before, taking the totality of workers, instead of nonsupervisory workers, and the use of average worker’s income instead of the median underestimates the WPD. Different authors have pointed out that the compensation of low- and middle-income workers has been stagnant while high managerial salaries have had a significant increase (Duménil and Lévy 2011; Elsby, Hobijn, and Sahin 2013; Giovannoni 2014; Economic Policy Institute 2021). This leads to an overestimation of the labor share for the immense majority of workers. However, given the available data, it is not possible to control for these factors in the present work.
Figure 5 compares the evolution of workers’ compensation with labor productivity using two different deflators: the CPI and the Net Output Deflator. In the first case, the decoupling began in the 1970s and has grown in a sustained path since then, a result coherent with Shaikh (2016: 60). In the second case, the decoupling started in 2001. Though the data are not presented in this article, it is worth pointing out that the decoupling is higher in both cases when only wages and salaries are used instead of workers’ total compensation, which is in line with Zavodny (1999). Using wages and salaries, the decoupling began in 1970 using the CPI and in 1986 using the output deflator.
Wage-productivity gap: US manufacturing sector, 1960–2017. Output is deflated using the Net Output Deflator, and labor’s total compensation is deflated using Consumer Price Index and Net Output Deflator.
The wage share, measured as the portion of workers’ total compensation in the net output, is presented in figure 6. It can be seen that the share began to decline drastically after 2001. The reason is that this was the point at which the total compensation of workers stopped increasing along with productivity, as seen in figure 5 in the case of the Net Output Deflator. As Pessoa and Reenen (2013) point out, a fall in the wage share can only be observed when the WPD is measured using the same deflator for both wages and output.
Total compensation of employees as a share of net output: US manufacturing sector, 1960–2017.
After analyzing the WPD and its impacts on the wage share, it is worth asking about the gender aspect of this issue. More specifically, the GWG and the female share of employment provide a new angle to understand the problem. Figure 7 shows the real total compensation of workers by gender. It can be seen how total female compensation grew along with productivity up to 2002, while male compensation was relatively stagnant (which is especially true when looking at wages instead of total compensation). This suggests that the decline in the GWG followed roughly the rhythm of increasing labor productivity up to 2002.
Wage-productivity gap: US manufacturing sector, 1960–2017. Labor’s total compensation is deflated using Consumer Price Index.
Some interesting patterns can be seen regarding the share of output received by female and male workers using equations 7 and 8. The female wage share, shown in figure 8, closely follows the inverted-U shape of the female share of employment. On the other hand, the male wage share has declined since the 1970s. The reason is mainly the increasing participation of female labor in the manufacturing sector and the stagnation of male wages relative to labor productivity.
Total compensation of female and male workers as a share of net output: US manufacturing sector 1960–2017.
It is possible to apply the same counterfactual exercise to analyze the wage share using equation 6. Figure 9 shows that assuming the absence of the GWG
Actual and counterfactual wage shares using total compensation of employees and net output: US manufacturing sector, 1960–2017.
By now, the isolated effects of the GWG and the WPD in the level of profitability and income distribution have been covered. The first phenomenon might have helped sustain profits by saving in unit labor costs by hiring female workers during 1960–1986 as more women entered as workers in this sector. Nevertheless, these extra benefits began to disappear as the female wages got closer to their male counterparts. It was also shown that the WPD began to gain relevance in sustaining the profit rate after 2001. With these two processes in mind, it is possible to create a second counterfactual rate of profit that combines both effects.
Figure 10 combines the assumptions of absence in the GWG and wages growing along with productivity (constant wage share after 2001). The graph suggests that the US manufacturing sector might have gotten extra benefits from these two sources. During 1960–2000, these additional benefits came mainly from the gender wage discrimination and afterward from the WPD. This means that part of the fall in the profit rate experienced before the new millennia could be associated with closing the GWG. After this period, the extra benefits come mainly from the decoupling between wages and productivity, despite the persistence of gender income inequality. In sum, the results suggest that the US manufacturing sector might have transited from getting extra benefits by exploiting cheap female labor to a period of constraining labor compensation by providing increments below the productivity growth and therefore reducing the wage share. Table 2 presents a summary of the results.
Actual and counterfactual rates of profit. US manufacturing sector 1960–2017.
Summary of the Results.
7. Conclusions
The narrowing of the GWG and the WPD are two relevant phenomena to understand the struggle for income distribution in the US economy. The GWG has been associated in the literature with higher profitability, investment, and competitiveness due to the lower unit labor costs. The WPD has also been associated with a way to maintain or boost profitability.
This article proposes an extension to the usual decomposition method of the profit rate that takes into account labor productivity, average wage, and capital/labor ratio (Basu 2017: 1375). It introduces the GWG and the female share of employment into the equation. It then applies this method to analyze the possible effects of the evolution of this wage discrepancy across genders and the WPD in the profitability and income distribution of the US manufacturing sector.
The results suggest that three consecutive periods could be defined for the US manufacturing sector. The first one, 1960–1986, in which the combined effects of gender wage discrimination and the increasing participation of female workers in this sector might have become a source of extra benefits for capital. In the second period, 1986–2001, the combined effects of narrowing the GWG and female workers leaving the sector reduced the extra benefits that capital could have gotten by hiring cheaper labor. This might have pushed capital to change the strategy to maintain profits. In consequence, the third period, after 2001, is when the WPD gained more relevance to sustain profits, suppressing the wage share of output. One of the main takeaways from this analysis is the relevance of including gender inequalities when studying the dynamics and tendencies of the rate of profit.
There are several caveats and shortcomings to the analysis as the nature of available data sets some limitations. First, I analyze the entire manufacturing sector, which includes corporate and noncorporate businesses. However, the analysis of the profit rate is generally more relevant for only the corporate sector since it is the one whose primary motivation is profitability. Second, as discussed above, the factors behind the GWG are complex. And given the nature and availability of these data, it is not possible to control for the qualification variables that explain part of the gap in order to obtain cleaner results. In consequence, the analysis of the extra benefits coming from gender wage discrimination could be overestimated. Third, given that I considered the average labor compensation of all workers in the sector instead of focusing on nonsupervisory workers, the results could be underestimating the impacts of both the GWG and the WPD in the functional income distribution and the rate of profit.
The present article also opens possibilities for future research. I focus here on the manufacturing sector. However, it is well known in the literature that the dynamics of the GWG and the WPD vary across sectors. Therefore, further research could shed light on the interaction between these two variables in the rest of the economy. Additionally, given the availability of new data sets, it could be possible to have better estimations of the impacts of the GWG and the WPD on the income distribution between capital and labor.
Footnotes
Acknowledgements
I am grateful to Kristin Munro, Sara Cantillon, and Aaron Pacitti for their valuable comments and suggestions on previous visions of this article. Any remaining errors are my own responsibility.
Author Note
Daniel Ossa is now affiliated with Department of Economics, University of Denver, Denver, CO, USA.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
