New polling data shows complicated politics of union members

by C.M. Lewis and Kevin Reuning

Union members vote blue—at least, that is the common wisdom.

Organized labor has been a key part of the Democratic Party coalition since the New Deal. Unions’ unparalleled ability to reach and mobilize a bloc of voters has been crucial to Democratic victories, and research has shown a clear correlation between union strength and Democratic vote share. In return, Democrats pledge—with mixed sincerity—to support organized labor legislatively.

With a recent strike wave and the resurgence of organized labor’s popularity and political power, labor is back at the forefront of the agenda. Politicians are taking note. Democrats, including Democratic nominee Joe Biden, are pledging policy changes that, if enacted, would represent the most substantial pro-labor reforms since the New Deal. Given the key role of organized labor within their donor network, union expansion would stand to help future Democratic victories.

Republicans realize that, too—which is why they have waged a war against labor unions. In the past ten years Republican majorities in states like Wisconsin, Michigan, and Ohio moved to kneecap organized labor, with success in both Wisconsin and Michigan. Similar moves in Missouri have been fought off through ballot initiatives and legal action. The attacks follow a similar pattern, and model legislation promoted by groups like the American Legislative Exchange Council provides templates for legislators to introduce anti-bargaining and “Right-to-Work” laws.

But union endorsements and union money—both of which go overwhelmingly to Democrats—do not necessarily translate into union votes, and do not necessarily indicate how union members think.

The surprising defection of many union voters to Donald Trump in key states was one of the major storylines following the 2016 election. Heading into 2020, concerns linger over Joe Biden’s ability to win back union voters in states like Pennsylvania. Some union leaders report substantial support for Trump, especially in more conservative segments of organized labor like the Building Trades. Recently, the Philadelphia local for the International Association of Fire Fighters recently bucked the IAFF—an early supporter of Joe Biden—and endorsed Trump for reelection, triggering its own internal controversy.

Why would union voters defect to Trump? On the Left, many have argued that unionists rejected Clinton as an ideological choice, pointing to surprisingly strong support for Bernie Sanders in the 2016 primary. There is some reason to believe that differences on trade policy and Trump’s unfulfilled campaign promises to rebuild American manufacturing played a role in appealing to Rust Belt voters—enough to make some union members gamble on him. 

But analyzing voter behavior does not address underlying viewpoints of union members. Information about what union members think is often lacking. Most analysis of the political viewpoints of union members has not differentiated between whether the respondent is a union member or lives in a union household. That data has shown a worrying trend toward Republican support in union households. But new, publicly available raw polling data commissioned by Data for Progress gives an unprecedented snapshot of the viewpoints of union members compared to the general public.

What it shows is at sharp odds with common perceptions of union members as an overwhelmingly Democratic and progressive constituency.

Analysis of the polling data shows that active union members are actually more Republican than non-union members, even though a plurality are Democrats. This reflects, in part, a greater polarization and strength of partisan identification in union membership generally—fewer union members identify as independents than non-members, and more are likely to identify as “Strong” Democrats or Republicans.

Interestingly, the data also shows a gap between union retirees—who are more Democratic—and present members, suggesting that the shifting and more diverse demographics of the modern labor movement do not automatically translate into more Democratic support or more progressive politics overall. Although it’s difficult to identify why this is the case, some factors—like a decline in union political education programs and the changing sectoral composition of labor—can offer some possible explanations.

But partisan affiliation, while useful, does not fully explain political viewpoints. Further data allows modeling some of the more nuanced viewpoints of union members, versus the broader public. To do this we estimated regression models that allow us to measure how union status correlates with political opinions while holding other demographic characteristics constants. In the following graphs we translate this into the views we would expect an individual to hold across a variety of characteristics. In these models we hold constant the gender, age, employment status, and education of the individual.

Polling data shows that union members are, controlling for all other variables, less supportive of socialism than non-members except among Republicans. However, less support does not necessarily mean opposition. Union members also tend to report less negative views of socialism than non-union members. The above plot shows a 6 to 9 percent decrease in negative views towards socialism comparing non-union to union members.

In contrast, union members are significantly more likely than the average non-member to hold positive views of capitalism. This is contrary to public perception, especially from the Left, of unions and union members as more friendly toward socialism and more hostile to capitalism. Their level of opposition to capitalism is consistent with non-members. Interestingly, union members are more likely to express a positive or negative viewpoint, rather than remaining neutral.

Ideological frameworks like capitalism and socialism give a sense of political outlooks, but pinpointing specific issues—such as Medicare for All—gives further insight into the viewpoints of union members compared to the general public.

Union members are less supportive of Medicare for All than non-members across except among Republicans, with stronger (though still low) opposition. However, this should not be overstated—among Democrats the difference between union and non-union members is only a few percentage points, while among Republicans the difference is close to double. Union membership may not result in more progressive viewpoints among self-identified Democrats, but it does result in substantially more progressive viewpoints among self-identified Republicans.

Between strong partisan identification and polarization, and counterintuitive responses on ideological questions, there are markers that the labor movement is more politically complicated than observers think. This is important, and gives a rare look shedding light on the politics of union members on a national scale. It complicates views of the labor rank-and-file as a solid progressive bloc, or one more readily won to more Left-wing politics. 

How to apply that analysis to an institutional movement with dozens of unions is a different question. Many unions—especially education unions—may have higher concentrations of partisan support for Democrats, and some unions like the National Union of Healthcare Workers have reputations for more militant politics. Conservative members are likely more concentrated in law enforcement and corrections locals, alongside the historically more conservative Building Trades unions. In other words, there is not an equal distribution of progressive and conservative union members—they are skewed in different unions based on history, sector, and other factors. Where people work, what kind of job they perform, and what union they belong to, may matter as much or more than their membership in a union.  

However, it would be a mistake to isolate conservatism in the movement to a few entities. Even education unions like the American Federation of Teachers and the National Education Association (especially the latter) have large concentrations of more conservative members. Rebecca Friedrichs, an evangelical Christian and anti-union activist that led off the Republican National Convention (and the plaintiff in Friedrichs v. California Teachers Association), was an elected union officer before she turned against the movement. They exist everywhere—not just in the Building Trades.

There are ways to change this. Some unions—like the Communications Workers of America—have actively introduced political education programs that show promise to engage members on ideological questions with positive outcomes. Their effort mirrors past union emphasis on educating their members on the connection between their jobs and electoral politics. But for many unions, the response may be infrequent and shallow engagement with members only during election cycles—something with little potential to change underlying opinions—or to avoid contentious politics altogether. 

There is a danger to shallow engagement. Republican inroads into the movement could increase with Trump’s pro-labor rhetoric and economic nationalism from politicians like Senator Josh Hawley. Recently, a wide range of Republican officials—including Trump’s former Attorney General, Jeff Sessions—released a statement calling for more Republican engagement with the labor movement. Their outreach comes alongside a continuing trend of Democrats courting business interests; the Chamber of Commerce recently endorsed an unprecedented number of Congressional Democrats for election.

The data—and the politics it shows—show that union members are more politically complicated than many assume, that many strongly hold conservative views, and that support for Democrats cannot be taken as a given. Democrats—and union leaders and activists that want a more progressive movement—had best pay attention.

C.M. Lewis is an editor of Strikewave and a union activist in Pennsylvania.
Kevin Reuning (@KevinReuning) is an assistant professor of political science at Miami University.

1 We set these characteristics as a woman with (a) some college education that (b) works full time and (c) is 43 years old (the average age). Changing these baseline characteristics will not change the overall pattern across party and union status, but will compress it or expand it as the baseline of holding certain views shifts.