Between 1979 and 2007, the share of national income going to the top one-percent of wealthy households grew from 9.6 percent to 20 percent. This shouldn’t surprise even the most casual consumer of economic news. Many scholars attribute this upward redistribution of wealth in part to a decline in labor’s share of national private sector income and an increase in capital’s share. Labor’s share of income, which has historically hovered slightly above 50 percent, has fallen 6 percentage points across the U.S. private sector since the late 1970s. Since capital income is concentrated among wealthier households, a relative increase in capital income benefits the richest households most of all.

In her June 2013 paper, “The Capitalist Machine: Computerization, Workers’ Power, and the Decline in Labor’s Share within U.S. Industries,” social scientist Tali Kristal focuses on the role unions play in this phenomenon. Kristal introduces the theory of “class-biased technological change,” which states that decades of technological change precipitated the decline of labor unions and weakened workers’ ability to bargain for a larger piece of the economic pie. First, new technologies lead to job losses in previously highly unionized sectors, like manufacturing, as work becomes more mechanized and production moves to lower-wage regions around the world. New technologies require new skills, a fact which can create a wedge between workers with and without those skills, polarize wages, and degrade workplace solidarity. Moreover, Kristal argues, new technologies empower employers to exert greater “technocratic control” over employees and engage in union-busting tactics. Drawing on the belief that class struggle drives the income distribution process, Kristal concludes that a shift in the class’ relative power leads to a shift in relative income.

Alternative theories attribute labor’s declining income share to factor-biased technological change: as new technologies improve a firm’s productive capabilities, the returns on equipment grow relative to the returns on labor, incentivizing producers to substitute labor for equipment. Using data on capital investment, compensation, unionization, and import penetration by low-wage countries, Kristal finds some evidence to support this.

However, the predominant effect of technology is indirect and mediated by declining unionization. As she writes, “labor’s share declined only in core unionized industries, despite the massive flow of computer technologies across all industries.” Turning to the manufacturing sector, she finds that a 20-percentage-point decrease in union density from 1978 to 2002 correlates with an 8.4-percentage-point decrease in labor’s share of income. Moreover, when controlling for changes in union density, the effect of technology on labor’s share nearly disappeared in industries with historically low union density, such as trade, services, finance, insurance, and real estate.

Certainly, labor’s relative share provides only one take on income equality. As James Heintz notes, labor’s overall income decline can hide the growing wage gap between categories of workers, particularly low- and high-skilled workers and supervisory and non-supervisory workers. Supervisory income may be more appropriately counted as capital income or overhead costs. Disaggregating labor income by supervisory and non-supervisory workers, Heintz finds that income for non-supervisory workers since 1964 has declined at a faster rate than overall labor income. In other words, as overall labor share has declined, income inequality within labor has left the rank-and-file with even less.