As labor rights advocates anxiously await Supreme Court judgment in Janus v. AFSCME Council 31—which would bar unions from automatically deducting dues for “fair share” fees and drastically reduce the ability of unions to support their efforts—just last week, on May 21, the Supreme Court unveiled a decision that bans class-action lawsuits for violations. Justice Ginsburg, writing for the dissenters in last week’s case, argued that the result will be a continued underenforcement of the very laws that are meant to help workers. Unions have been a critical aspect of upward mobility and building an American middle class, making blue-collar jobs well-paying, with good benefits and sufficient time off. But lax enforcement of labor laws and a hostile business environment have made it difficult for unions to maintain their membership and keep up with the changing structure of the labor market.
None of this has stopped a swelling of actions by workers in recent months. Highly visible labor actions, particularly by public school teachers, have been taking place across the United States in the early part of 2018. Remarkably, many of these teachers’ strikes have been in so-called right-to-work states, where the legal framework makes it difficult for unions to collect membership dues in order to support organizing efforts. Among the teachers’ demands is, of course, pay increases to address long-term wage stagnation and improved funding for schools. But, more broadly, these actions have demonstrated the social-justice roots of the labor movement: In addressing economic inequality, the direct impact of wage stagnation for teachers, and how the underfunding of public education affects future generations, it’s clear how union organizing attempts to intervene in inequality.
This connection between unions and inequality is not just aspirational, as new research published as a working paper for the National Bureau of Economic Research shows. Economists Henry Farber, Daniel Herbst, Ilyana Kuziemko, and Suresh Naidu have compiled survey data from a variety of historical sources to expand on what we knew previously about rising and declining union density over the 20th century.
Using Gallup polling data that asked about union representation in households starting in 1937, the researchers show union density (the percentage of the workforce that belonged to a union) was more than 10 percent before the passage of the Fair Labor Standards Act the following year. That number went up to almost one-third of households in 1950 and has been slowly declining since then. (The data are confirmed by the Census, which began collecting data on union membership in its Current Population Survey in 1973.) That unions have been declining in their size and power is nothing new. Here’s where it gets interesting.
The inverted U of union density (from its slow rise from 1937 to its peak to its drop) is matched by its inverse in economic inequality as measured by the top 10 percent and top 1 percent shares of family income over time from 1920–2016. The correlation between unions and economic inequality is undeniable.
But is there causation?
Over this long time period, the economy has also gone through significant structural shifts, like increased adoption of technology, outsourcing, and a better-educated workforce. In 1976, only 14.7 percent of adults over the age of 25 had a college degree, and this rose to 23.6 percent by 1996, then further to 33.4 percent by 2016. Many economists, led by the work of David Autor, have argued that growing income inequality can be explained by a concept called skill-biased technological change, or SBTC. Autor’s theory around SBTC holds that the rise of high-skilled workers has caused job polarization—with more demand for very high-skill jobs and some demand for low-skill jobs and a hollowing out of midskill jobs—leaving those with less education behind, resulting in the inequality we see today.
The decline of unions was thought to be another effect of SBTC: Since unions have historically represented workers with lower levels of education, and as more educated workers became paid more, they would opt out of joining a union, or so the logic went. In this sense, the decline of unions is not inherently a bad thing, because it means fewer Americans need the bargaining power afforded by unions, since increasing their skill is presumably giving them the ability to command higher wages without collective bargaining.
The problem is the evidence that education will help workers get higher wages has not borne out for many. For example, women and black Americans get lower returns to education compared to white men, so increasing their education level has not been able to fully address pay inequality.
While SBTC is not accepted as unquestionable truth in economics, its message permeates in economic policy. It has led to policies, like $300 billion for questionably effective worker-training programs, that put education and workforce development above any systematic improvements to the jobs workers currently occupy and improvements in workers’ agency to demand more vis-à-vis their employers. Promoting education as the tool to overcoming economic inequality is convenient because it puts the onus on individuals to pull themselves up by their bootstraps and doesn’t call into question structures and barriers that limit economic opportunity.
But the new data from Farber, Herbst, Kuziemko, and Naidu throws a wrench in the SBTC explanation of rising inequality. They find that the education level of union members also followed a U-shape curve from 1936–2016. Early in the American labor movement, union households tended to have higher education levels—think of the guild model of union membership, with only very highly skilled craftsmen. But as union membership increased, beginning in a period where labor activism was a social movement, unions started representing more workers with lower levels of education, a more “working class” population. The correlation between years of education and union membership declined from 1940 to the height of the labor movement in the 1950s and ’60s. But beginning in the mid-1960s, the same time as union density was beginning to decline, unions increasingly represented workers with more education. That’s probably because of the combination of exerted attacks on unions’ ability to organize with the overall restructuring of our labor market—outsourcing and subcontracting isolated low-skill workers. Unions were not able to maintain their representation of the workers who would benefit the most from collective bargaining.
What does this mean for the theory that low education and a lack of skills explain growing wage inequality? It means that while education and skills training matter to worker productivity, they’re not sufficient to ensure economic security without workers’ power to bargain for better wages too.
The recent teachers’ strikes are emblematic of the modern labor movement. Teachers are by and large an educated workforce. However, even these highly educated workers are now also facing the pressure of wage stagnation and economic insecurity. Likewise, teachers are predominately women, and women face an even greater wage gap at higher levels of education. Unions have helped women close the gender wage gap, but their declining power limits their ability to do so. (The same is true for black Americans, who face greater wage gaps at higher levels of education, and for whom unions have been crucial in closing racial wage gaps.)
Greater education and the need for more workers to receive it are not adequate explanations of inequality. The decline in union density is. Providing opportunity to American families will require a robust labor movement that balances corporate power and pushes back against the wage stagnation affecting most workers.
