Amid Union Decline: State-level Unionization and Overwork of American Workers

Labor Union Photo
Photo by ibreakstock on Adobe Stock
View the research summary as a pdf.

 

Overview: 

Over the past several decades, overwork has increased across most demographic groups in the United States, while organized labor's power has steadily declined. Historically, labor unions played a central role in defining standard work hours and shaping social norms around the forty-hour workweek. This study examines the association between changes in state-level union density and the likelihood of overwork from 1983 to 2019.

Methodology:

The analysis uses data from the 1983–2019 Current Population Survey Outgoing Rotation Group (CPS-ORG). The sample includes 2,821,901 wage and salaried workers aged 18 to 64, not currently in school, not institutionalized, and not employed in farm occupations. Only odd years were used to avoid duplicate respondents. 

The dependent variable in this analysis is overwork, defined as working more than 45 hours per week. The main independent variable is state union density (percent of all workers who are union members).¹ Two-way fixed-effects models were applied, controlling for year- and state-fixed effects, along with individual- and state-level covariates. 

  • Individual controls: education, age, race and ethnicity, marital status, parental status, number of children, occupation, industry, sector, and metropolitan residence.
  • State-level controls: proportion of college graduates, industrial composition, racial demographics, GDP per capita, unemployment rate, and state economic policy conservatism. 

The models estimate the net association between changes in union density and overwork while accounting for state-specific and time-varying conditions. Other analyses test for nonlinear associations, gender differences, and robustness to different definitions of overwork (41, 50, and 55 hours per week).

Findings:

Three dominant themes highlight how media narratives shape public understanding of book censorship and LGBTQ representation.  

  1. Negative Association Between Union Density and Overwork 

    There is a robust negative association between state union density and the likelihood of overwork. A one percentage point increase in state union density is associated with a 0.13–0.14 percentage point decrease in men’s likelihood of overwork and a 0.07–0.09 percentage point decrease for women. From 1983 to 2019, the overall decline in union density contributed to an estimated 12–22% increase in overwork, depending on gender. 

  2. Conditional Relationship by Level of Unionization 

    The effect of union decline varies by the level of unionization. The negative association between union density and overwork is stronger in highly unionized states and becomes statistically insignificant in loosely unionized states. The inflection point occurs around 10% union density; below this threshold, further union decline has little measurable effect on overwork. 

  3. Period Differences 

    The negative association between union decline and overwork is stronger during 1983–1999 than during 2001–2019. This pattern reflects the diminishing influence of unions as overall density declined over time. 

  4. Robustness and Supplemental Analyses 

    The results hold across men and women, racial and ethnic groups, and occupational categories. Furthermore, the association remains consistent when limited to full-time workers, nonunion workers, or private-sector employees. The net negative association between union density and overwork holds across demographic groups, and it is robust to alternative definitions of overwork and model specifications.

Implications:

The findings confirm that the decline in union density is linked to an increase in overwork across U.S. states. This relationship reflects the collective influence of unions in a labor market, not just individual union membership. 

Consistent with power resources theory, higher union density strengthens workers’ bargaining position, limiting employers’ ability to demand excessive hours. From a moral economy perspective, unions cultivate egalitarian workplace norms that reinforce the standard workweek and fairness in labor practices. As unionization declines, these norms weaken, and overwork becomes more common. 

The authors note that the rise of overwork partly reflects workers’ increasing income insecurity and employment volatility, which accompanied the collapse of a union-based moral economy. Where union influence is strong, collective bargaining can restrain overwork; where it is weak, competitive pressures and cultural norms of “ideal workers” lead to longer hours.

Policy Relevance:

This research highlights unions’ role in shaping labor-market behavioral norms and in institutional constraints on long work hours. Because the association is stronger in higher-union contexts, policies or political changes that materially alter state-level union strength (for example, changes to collective-bargaining regimes, right-to-work adoption, or other labor laws) can have measurable effects on workers’ hours and the prevalence of overwork. The study implies that addressing overwork through workplace and state-level institutional levers (rather than only individual- or firm-level interventions) may be consequential for work–family balance and public health outcomes linked to long hours.

Key Takeaways:

  • There is a consistent negative relationship between state union density and overwork from 1983 to 2019.
  • Union strength matters: the effect is largest where unions were historically strong and minimal where they are weak.
  • The impact of unions on overwork extends beyond individual members, reflecting broader labor market and cultural dynamics.
  • Declining union power contributes to a shift toward longer work hours, reduced worker leverage, and erosion of the 40-hour norm.

Authors: 

Yurong Zhang, doctoral candidate in the Sociology Department and graduate research assistant, University of Kansas, yurongzhang@ku.edu. 

ChangHwan Kim, Ph. D., professor in the Sociology Department and director of graduate studies, University of Kansas, chkim@ku.edu

Reference:

¹Hirsch, B. T., & MacPherson, D. A. (2003). Union Membership and Coverage Database from the Current Population Survey: Note. ILR Review, 56(2), 349-354. https://doi.org/10.1177/001979390305600208 (Original work published 2003) 

Learn More: 

The full version of this paper was published in Social Science Research. https://doi.org/10.1016/j.ssresearch.2025.103178. 

Institute for Policy & Social Research, University of Kansas, ipsr.ku.edu