Center for American Progress Action
Unions Are Critical to the Democratic Party’s Electoral Success
Unions Are Critical to the Democratic Party’s Electoral Success
New CAP Action analysis finds that union members voted for the Democratic Party presidential candidate in much greater proportions than did nonunion members in 2020, cementing President Biden’s electoral victory and offering a path forward for the party to maintain and grow support.
In this article
Introduction and summary
Decades of research have illustrated that unions are critical to increasing voter support for Democratic candidates across a range of demographic groups. Lately, however, some observers have questioned whether this is still true, given Donald Trump’s strong performance among working-class voters in the 2016 and 2020 presidential elections.
New analysis from the Center for American Progress Action Fund finds that union members remain strong Democratic supporters. In the 2020 presidential election, union members voted for the Democratic Party candidate in much greater proportions than did nonunion members, cementing Joe Biden’s win over former President Donald Trump. Union members have voted this way for decades, and the union advantage has long been strong for Democrats across demographic groups; indeed, union voters favored Biden more than nonunion voters across most breakdowns of gender, race and ethnicity, education, and age. Notably, CAP Action’s analysis found that white working-class union voters and union voters older than age 50 provided stronger support for the Democratic candidate than their nonunion counterparts; even though on the whole, these groups favored the Republican candidate in large numbers in 2020, the union voters within them helped maintain some Democratic support. Democratic-leaning groups such as women, Hispanic workers, and the college-educated also provided strong support for Biden. The union members in these Democratic-leaning demographic groups compounded the Democratic advantage.
Specifically, the analysis finds that:
- Union women were 21 percentage points more likely than nonunion women to vote for Biden, while union men were 13 percentage points more likely than nonunion men to vote for Biden.
- The union advantage for Biden was particularly pronounced among white voters—among whom union voters were 18 percentage points more favorable than white nonunion voters—but it still held strong with Hispanic voters (13 percentage points more favorable) and other nonwhite or multiple race voters (9 percentage points more favorable). There was no statistically significant difference among Black voters, among whom both union and nonunion workers overwhelmingly favored Biden.
- Union voters were more favorable than nonunion voters to Biden among both working-class and college-educated voters, by 6 percentage points and 22 percentage points, respectively.
- Compared with their nonunion counterparts, working-class Hispanic union voters were 16 percentage points more favorable to Biden, while working-class white union voters were 6 percentage points more favorable to Biden.
- Union voters, compared with nonunion voters, provided an advantage for Biden across all age groups that only increased with age: 18–29 offered a 5 percentage-point advantage; 30–49 offered a 15 percentage-point advantage; 50–64 offered a 17 percentage-point advantage; and 65 and older offered a 27 percentage-point advantage.
Given Biden’s pro-labor stance on the campaign trail, his strong performance among voters with union ties should not come as a surprise, and his efforts to empower the working class through executive action and his legislative Build Back Better agenda could extend this advantage into the future.
Union voters are integral to the Democratic coalition
The findings of the CAP Action analysis are particularly important as the Democratic Party considers its political strategy for coming elections. Party supporters are vigorously debating which demographic groups hold the key to the party’s electoral success and how best to target them. They are weighing, for example, the trade-offs between focusing on the white working class, Hispanic and Black voters, or the college-educated, and questioning how best to appeal to these voters. CAP Action’s findings that union members across demographic groups favor Democrats indicate that, no matter where readers’ sympathies lie in these debates, they should appreciate that increasing union membership would likely improve the Democratic Party’s electoral success.
The results of this analysis are consistent with long-standing academic research showing that unions both change voter preferences and can increase voter turnout—a powerful combination. Empirical studies have confirmed time and time again that union members are often more likely to vote and more likely to favor Democrats, directly tying union strength to Democratic electoral performance.1 Indeed, one recent study found that “right-to-work” laws, which reduce union membership, shrunk county-level Democratic presidential vote shares in elections from 1980 to 2016 by 3.5 percentage points compared with non-“right-to-work” border counties and lowered voter turnout by 2 percentage points.2
The union Democrat advantage is more nuanced than merely voting the union ticket, however; unions provide a range of critical support for voters, including educating them about worker issues and which candidates will support workers, providing minischools of democracy through their internal operations,3 encouraging their members to vote, and providing financial resources for campaigns. Unions’ role in educating voters may be particularly important. Union voters who support worker-oriented policies are more likely to vote for their union’s endorsed candidate, which suggests that unions strengthen the link between economic pragmatism and votes for pro-labor candidates.4 Union members prefer candidates they perceive as being advocates for the working class,5 effective at elevating working-class voices into elected offices,6 and able to increase legislative responsiveness to the needs of the poor.7 And critically, white union members also score much lower than their nonunion counterparts on survey questions used to gauge racial resentment, suggesting that unions could help decrease the potential to be swayed by the narratives of white racial grievance that Donald Trump and other Republicans promote.8
In short, stronger unions are crucial for keeping the Democratic Party coalition together and victorious. Today, however, only 10.8 percent of American wage and salary workers are members of labor unions,9 far below the 1954 peak of 34.8 percent.10 Any strategy for the Democratic Party should center strong support for unions and advocacy of policies that increase their membership, such as the Protecting the Right to Organize Act11 and the Public Service Freedom to Negotiate Act.12
For this analysis, the authors relied on the 2020 Cooperative Election Study13 (CES, formerly the Cooperative Congressional Election Study) and the 2008 through 202014 American National Election Studies (ANES). These surveys offer representative data on the voting behaviors of the American population, with ANES data allowing time series comparisons of changing voting patterns over time and the CES including a larger sample size that permits more precise studies of subsamples of voters in 2020. The authors relied on CES results for their detailed analysis of the 2020 presidential election. They used the ANES for time trends and to validate findings where possible, noting discrepancies between the outcomes of the two surveys caused by sampling inconsistencies across race and educational attainment. Both surveys indicated voters’ current union status and current employment status.
The authors restricted CES 2020 findings to validated voters and ANES time series analysis to voters who reported voting for Democratic or Republican candidates in presidential elections. The analysis includes only voters who could feasibly join unions. The authors achieved this by limiting the sample to voters who are working full time or part time (in the case of the ANES, in excess of 20 hours per week) or are temporarily unemployed, as only employees are eligible to join unions, while retirees and homemakers, for instance, are not. This ensured that the observed differences in voting behaviors were as limited as possible to the effects of union membership.
The results likely understate the increase in Democratic vote share precipitated by union membership, as removing the employment restriction from the CES analysis causes a decline in preference for Joe Biden among nonunion voters. Furthermore, although the analysis focuses on union members, the authors’ research indicates that union households vote similarly to union members and are a greater share of the population.
The authors report the results of the analysis as the “Democrat vote advantage,” measuring the extent to which Democratic candidates led among voters in a given demographic group. This measurement is the absolute difference between the proportion of votes the Democratic candidate won among a demographic group and the proportion of votes that instead went to the Republican candidate. As a result, a positive Democrat vote advantage among a demographic group indicates the Democratic candidate won a majority of that group’s votes, and a higher advantage corresponds to a greater margin of victory for the Democrat within that group. The 2020 CES analysis measures these proportions among all valid votes, including those for third-party candidates, while the ANES time series analysis excludes third-party votes.
The authors use the terms “working-class” and “noncollege” interchangeably; both refer to voters with less than a four-year college degree. The authors also use “Hispanic or Latino” to refer to voters who identified as being of Spanish, Hispanic, or Latino origin or descent.
Results of the analysis
Overall, union voters favored Democrats more strongly than nonunion voters in the past four presidential elections. The time series results in Figure 1 show that the union Democrat advantage remained roughly as large in 2016 and 2020 as it was in 2008 among employed voters, who overall showed a stronger preference for Democrats than the total voting population. Furthermore, even as the entire electorate shifted further to the Democrats in 2016 and 2020, the Democrat vote advantage among union voters kept pace, meaning that Biden performed the best out of the last three Democratic candidates among union voters. Rather than abandon the Democrats for Trump, union voters turned out for Joe Biden, the candidate who promised to “be the most pro-union president you’ve ever seen.”15
The union effect on the Democrat vote advantage was strong enough to flip male union voters to Joe Biden, and it compounds with Democratic preference among women. As shown in Figure 2, while Trump won among nonunion male voters, Biden won union men, demonstrating unions’ effectiveness at winning votes for the Democrats among a group that leans Republican. Furthermore, Biden’s victory margin doubled for union working women relative to nonunion working women. Working women tend to be more reliable Democratic voters, possibly due in part to the Democrats’ support for policies that uplift women, such as paid family leave; this result suggests a powerful intersection between union membership and identity that shifts voter preference further toward Democrats.
The Democrat vote advantage among union members was observed across nearly all racial and ethnic groups and was particularly prominent among white union voters, among whom Biden performed better than Trump. Figure 3 shows the results by racial or ethnic category. The sole exception to the Democrat vote advantage for union voters occurred among Black or African American voters, as nonunion Black or African American voters already favored Biden 9 to 1. The 95 percent confidence intervals of the point estimates for Biden’s and Trump’s vote shares among both union and nonunion Black or African American voters overlap, meaning a statistically significant difference in the Democrat vote advantage between union and nonunion Black voters could not be observed. Overall, white voters preferred Trump, but white union members overall preferred Biden. And white voters were not the only racial group seen as pivotal in narrowly handing Trump a defeat in 2020. Though Hispanic or Latino voters swung notably toward Trump in 2020, compared with 2016, Hispanic or Latino union members favored Biden in 2020 by nearly 40 percentage points. Union members of other or multiple racial groups also strongly favored Biden.
Focusing on voters who are Hispanic or Latino highlights unions’ importance for maintaining Democratic Party support across this demographic. Working-class and male Hispanic or Latino voters—subgroups that showed the strongest support for Trump in 2020— were much more likely to vote for Biden if they were members of a labor union.16 (see Figure 4) Although many commentators expected Hispanic or Latino voting patterns to largely emulate those of other people of color in 2020, given the centrality of white racial grievances to Trump’s campaign, others argued that many Hispanic or Latino voters would be more receptive to pragmatic messaging about jobs and the economy, especially given the pace of the COVID-19 recovery in late 2020.17 Under these circumstances, pro-worker rhetoric from Biden that reached Hispanic union members more readily than nonunion members may have had a special appeal. Indeed, union Hispanic or Latino voters proved more amenable than nonunion Hispanic voters to Biden’s campaign promises to empower and uplift workers. Notably, however, Biden won a higher proportion of nonunion Hispanic or Latino voters in the ANES sample relative to the CES sample, meaning the CES results may relatively overstate the positive Democrat vote advantage among union Hispanic voters.
Unions also helped increase support for Biden across educational attainment levels. Although noncollege-educated voters overall favored Trump,18 Figure 5 shows that unions significantly increased Biden’s vote share among the working class, helping minimize his overall loss among this group. Because the working class makes up the largest part of the working American electorate, it is crucial to any party’s election strategy. Figure 5 also shows that Biden’s margin of victory nearly doubled among college-educated union workers compared with college-educated nonunion workers.
Looking more closely at the white working class, Figure 6 shows that Biden enjoyed much higher support among white voters without college degrees who were in labor unions and stayed even more competitive among white working-class women than he did among men. While Biden did not need to win these groups outright to take the White House in 2020, his strong showing was enough to stave off Republican gains among other groups. Like the Hispanic subsample, CES results differed somewhat from ANES outcomes in estimating lower support for Biden among nonunion white working-class voters, so the CES again may relatively overstate the union Democrat advantage among the white working class.
Finally, union members of all ages preferred Biden to Trump. Among voters ages 18 to 29 and 30 to 49, Biden won handily, with union members providing additional support compared with nonunion members. Working voters ages 50 to 64, as well as those older than age 65, largely went to Trump, but union voters at these ages flipped to Biden. (see Figure 7) Although Biden enjoyed steadier support among younger voters overall, younger people made up a much smaller pool of working voters than those older than age 50, making the inroads to these voters Biden won through unions crucial to his 2020 victory.
These results cannot be attributed solely to the possibility that people who join unions are simply more likely to be left-leaning or to favor pro-worker voices. Figure 8 shows that, in 2020, former union members voted much the same as nonunion workers.
The effects of active union membership, including greater access to information on candidates’ labor policies, encouragement from union leadership, and participation in a local democratic institution, worked together to award Joe Biden an advantage among union members in 2020. These effects do not seem to be limited to active union members within a household: As seen in Figure 9, Biden’s performance was much the same among voters with union members in their household regardless of whether they were union members themselves. This outcome indicates that unions have a much further reach than their membership numbers alone imply.
Across most demographic groups, unions increased voter preference for Biden in 2020. This support could prove crucial to cementing Democratic support in upcoming elections. Union voters helped boost Democratic Party support among voters young and old, female and male, white and Hispanic. They helped Biden increase his advantage among a number of already Democratic-leaning groups, such as college-educated voters, and helped Biden remain more competitive among groups that Trump won, particularly the working class.
CAP Action’s new analysis shows that no voters can be taken for granted; they need to be recruited and retained. Unions help retain voters across demographics and educational attainment, which is why Democratic Party strategists and politicians should recognize unions’ importance and support policies to strengthen them.
About the authors
Aurelia Glass is a research assistant for Economic Policy at the Center for American Progress Action Fund.
David Madland is a senior fellow and the senior adviser to the American Worker Project at the Action Fund.
Ruy Teixeira is a senior fellow at the Action Fund.
- Tom Juravich and Peter R. Shergold, “The Impact of Unions on the Voting Behavior of Their Members,” ILR Review 41 (3) (1988): 374–385, available at https://www.jstor.org/stable/2523904; Richard B. Freeman, “What Do Unions Do… to Voting?” (Cambridge, MA: National Bureau of Economic Research, 2003), available at https://www.nber.org/papers/w9992; Benjamin Radcliff, “Organized Labor and Electoral Participation in American National Elections,” Journal of Labor Research 22 (2001): 405–414, available at https://link.springer.com/content/pdf/10.1007/s12122-001-1042-7.pdf; Roland Zullo, “Union Membership and Political Inclusion,” IRL Review 62 (1) (2008): 22–38, available at https://www.jstor.org/stable/25249183; James Feigenbaum, Alexander Hertel-Fernandez, and Vanesa Williamson, “From the Bargaining Table to the Ballot Box: Political Effects of Right to Work Laws” (Cambridge, MA: National Bureau of Economic Research, 2018), available at https://www.nber.org/system/files/working_papers/w24259/w24259.pdf.
- Zullo, “Union Membership and Political Inclusion”; Feigenbaum, Hertel-Fernandez, and Williamson, “From the Bargaining Table to the Ballot Box.”
- David Madland and others, “Unions Are Democratically Organized, Corporations Are Not” (Washington: Center for American Progress, 2020), available at https://www.americanprogress.org/issues/economy/reports/2020/10/16/491731/unions-democratically-organized-corporations-not/.
- Paul Clark and Marick F. Masters, “Competing Interest Groups and Union Members’ Voting,” Social Science Quarterly 82 (1) (2001): 105–116, available at https://www.jstor.org/stable/42955705.
- Monica Bielski Boris and Robert Bruno, “‘Not Ready to Make Nice’: The Politics of Identity and Why Union Voters Wanted a Class Champion in 2008,” Labor Studies Journal 35 (1) 2009: 94–115, available at https://journals.sagepub.com/doi/10.1177/0160449X09355581.
- Aaron J. Sojourner, “Do Unions Promote Members’ Electoral Office Holding? Evidence from Correlates of State Legislatures’ Occupational Shares,” ILR Review 66 (2) (2013): 467–486, available at https://www.researchgate.net/publication/254070533_Do_Unions_Promote_Members%27_Electoral_Office_Holding_Evidence_from_Correlates_of_State_Legislatures%27_Occupational_Shares.
- Michael Becher and Daniel Stegmueller, “Reducing Unequal Representation: The Impact of Labor Unions on Legislative Responsiveness in the US Congress,” Perspectives on Politics 19 (1) (2021): 92–109, available at https://www.daniel-stegmueller.com/SB_representation_April2020_web.pdf.
- Paul Frymer and Jacob M. Grumbach, “Labor Unions and White Racial Politics,” American Journal of Political Science 61 (1) (2021): 225–240, available at https://onlinelibrary.wiley.com/doi/epdf/10.1111/ajps.12537.
- U.S. Bureau of Labor Statistics, “Union Members Summary,” Press release, January 22, 2021, available at https://www.bls.gov/news.release/union2.nr0.htm.
- Gerald Mayer, “Union Membership Trends in the United States” (Washington: Congressional Research Service, 2004), available at https://www.everycrsreport.com/files/20040831_RL32553_e5e58e1832de83247c0883a6fd691bc84691a745.pdf.
- Protecting the Right to Organize Act, H.R.842, available at https://www.congress.gov/bill/117th-congress/house-bill/842 (last accessed December 2021).
- Public Service Freedom to Negotiate Act, H.R.3463, 117th Cong., 1st sess. (March 11, 2021), available at https://www.congress.gov/bill/116th-congress/house-bill/3463.
- Brian Schaffner, Stephen Ansolabehere, and Sam Luks, “Cooperative Election Study Common Content, 2020,” available at https://doi.org/10.7910/DVN/E9N6PH (last accessed September 2021).
- American National Election Studies, “Time Series Cumulative Data File (1948-2016),” available at https://electionstudies.org/data-center/anes-time-series-cumulative-data-file/ (last accessed October 2021); American National Election Studies, “2020 Time Series Study,” available at https://electionstudies.org/data-center/2020-time-series-study/ (last accessed October 2021).
- The Associated Press, “The Latest: Trump predicts he’ll ‘so easily’ win Michigan,” November 3, 2020, available at https://apnews.com/article/election-2020-donald-trump-virus-outbreak-district-of-columbia-elections-3f42f2ddb730a8dff916b54be74672a0.
- Figure 6 includes only respondents who replied “Yes” when asked if they are of “Spanish, Latino, or Hispanic origin or descent,” whereas the Hispanic category in Figure 5 includes only respondents who selected “Hispanic or Latino” from a mutually exclusive list of possible racial or ethnic categories. The former definition of “Hispanic” was favored for the analysis of Hispanic voters only, as it is a more precise definition that does not exclude white Hispanic people.
- Thomas B. Edsall, “Democrats Are Anxious About 2022 — and 2024,” The New York Times, March 10, 2021, available at https://www.nytimes.com/2021/03/10/opinion/democratic-voters-anxieties.html.
- Although other polling sources corroborate a Trump gain over Biden among noncollege-educated voters, CES data relatively overstate this advantage compared with Catalist, which only indicated a 3 percentage-point to 5 percentage-point advantage for the Republicans in 2020 over the Democrats. See Yair Ghitza and Jonathan Robinson, “What Happened in 2020,” Catalist, available at https://catalist.us/wh-national/ (last accessed November 2021).
The positions of American Progress, and our policy experts, are independent, and the findings and conclusions presented are those of American Progress alone. A full list of supporters is available here. American Progress would like to acknowledge the many generous supporters who make our work possible.
Senior Fellow; Senior Adviser, American Worker Project
Former Senior Fellow