Unions Need to Lose More If They Want to Win More

The UAW’s defeat at a Mercedes plant in Alabama was crushing. It’s also the cost of waging risky, potentially transformative fights. If labor wants to win big, it can’t be afraid to lose big.

United Auto Workers buttons in Chattanooga, Tennessee, on April 10, 2024. (Kevin Wurm for the Washington Post via Getty Images)

There’s no sugarcoating it: Mercedes workers’ loss last week was a punch in the gut. Hopefully we can soon get some sober assessments from worker leaders and staff organizers about what — if anything — they’d try to do differently next time around.

But it’s also necessary to take a step back and acknowledge that any ambitious strategy for unionizing millions will entail lots of losses along the way. There’s an obvious way this is true: labor movements that don’t try to organize the unorganized — or that don’t go on risky strikes — never experience any big losses, they just steadily decline into irrelevance. If you unionize and strike more, your total number of losses will also rise, all other things being equal.

The point I want to make in this article, however, is more specific and less intuitive: ambitious labor movements that try to win widely actually lose a higher percentage of battles than do most unions today. Winning big and winning at scale require taking many more risks and relying less on staff. And this generally entails a higher loss rate.

As I’ll show below, one of the reasons why labor’s win rate in union elections has been so exceptionally high over the past two decades is that exceptionally few unions are seriously pursuing new organizing. And those that do are often only taking on and sticking with drives that they’re very confident will win. Any chance labor has at making a big breakthrough — any chance at decisively reversing decades of decline — requires being okay with more losses along the way.

Look at the Data

It’s helpful to compare labor’s current win rates with those of a century ago. Organized labor in the early twentieth century not only waged far more battles than today; it lost these more frequently. Though there are no records of unionization win rates back then, we can compare the success rates of strikes, as compiled by government statisticians based on whether most major demands were won. As you can see in Figure 1, there wasn’t a single year from 1900 to 1937 (the last year we have data on) where a majority of strikes were “successes.”

Figure 1: US Strike Success, Yearly (John I. Griffin, Strikes: A Study in Quantitative Economics [New York: Columbia University Press, 1939], 91, based on government reports)
National Labor Relations Board (NLRB) data from the 1960s onward show that unions tend to win a higher percentage of elections when they organize significantly less. This makes sense, because if you only undertake unionization efforts with a high likelihood of success, your universe of potential drives shrinks considerably.

As recently as 1976–1980, unions ran about five times as many NLRB drives yearly as they do today, but their win rate was only about 48 percent. In contrast, as the total number of drives has plummeted in recent decades, labor’s win rate has steadily crept upward, hovering at around 70 percent in recent years.

Were labor today to double or quadruple its amount of drives, the total number of union wins would shoot up even if success rates dropped back down to 1970s or 1920s levels. Winning at scale will very likely require losing more.

A visual depiction of these dynamics may help drive the point home. After US employers went on the offensive from 1981 onward, the total amount of union drives plummeted, and it has not yet to come close to recovering (see Figure 2).

Figure 2: Number of NLRB Certification Elections Yearly (Henry Farber, “Union Success in Representation Elections: Why Does Unit Size Matter?” ILR Review 54, No. 2 (2001), 26, based on NLRB data)

One might expect that union win rates would have similarly plummeted after 1981, since bosses were fighting harder to prevent workers from winning. But, to the contrary, union win rates have steadily risen from the early 1980s onward, as demonstrated in Figure 3. Unions started avoiding many of the hardest battles, and there were fewer overall attempts to organize. But those drives that were attempted tended to win more frequently.

Figure 3: Union Win Rate, NLRB Elections (Compiled by author on the basis of NLRB data as well as Lawrence Mishel, Lynn Rhinehart , and Lane Windham, “Explaining the Erosion of Private-Sector Unions” [Economic Policy Institute, 2020])
We can also see this same general trend — fewer attempted union elections usually equals a higher win rate — in the yearly data going back to the 1980s.

Figure 4: NLRB Election Results (Robert Combs, “Deceleration Defines 2019 State of the Unions” [Bloomberg Law, 2020])

How We Got Here

It’s obviously better for workers to win than to lose. And it goes without saying that organizers should do everything possible to win every battle, which requires (among other things) taking organizing methods very seriously and leaning as much as possible on the best practices of deep organizing. That said, post-Reagan unionism generally reflects a fear of losing that bleeds into excessive risk aversion.

This trepidation is based on real accumulated experience since the 1980s. Lost campaigns have subjected workers to employer repression, prevented further drives at demoralized workplaces (or companies) for years to come, discouraged parent unions from maintaining funding for new organizing, and hurt the reputations (and job prospects) of staffers as well as elected leaders. Lost strikes, similarly, can have devastating effects. And because new organizing since the 1980s has been so resource intensive — contemporary best practice has been to hire one staffer for every one hundred workers to be organized — unions are incentivized to only take on a select number drives that they’re pretty sure will win. That’s why unions routinely say no to workers who reach out for organizing support. “Our message to the working class is ‘don’t call us, we’ll call you,’” a researcher from one such union only half-jokingly told me.

This is one of the many reasons why a worker-to-worker organizing model, as attempted at Mercedes in Alabama, is pivotal for winning widely. Part of the explanation for why the labor movement a century ago was less scared of losing was that it was far less staff intensive — averaging one staffer for every two thousand workers, a ratio replicated at Mercedes. Full-time leaders and staffers will always face stronger pressures not to jeopardize, via risky actions, the health of the union that employs them. And the more expensive it is to unionize, then (all other things being equal) the less likely it is for unions to engage in new organizing. A more bottom-up model is necessary to get labor to start embarking down the risky road of waging exponentially more battles.

At its rank-and-file oriented best, staff-intensive unionism can win significant battles. But it can’t win the war. The reason for this is simple: heavily staffed organizing is too costly to scale up — i.e. to win at the widest scope necessary. Staff-intensive unionism can’t hire enough organizers, or train enough of them quickly enough in high-momentum moments, to unionize tens of millions of workers.

In other words, scaling up requires betting more on worker leaders, even if this means a relatively higher percentage of drives may end up losing. It’s less risky to depend on staff-heavy campaigns than it is to wager on bottom-up, lightly staffed drives like Mercedes.

Faced with high stakes and powerful opponents, caution can be useful — but only up to a certain point. An excessive fear of losing is a key reason why unions have been so hesitant to seize the postpandemic organizing opening, because unions remain deeply risk averse and because a sizable number of today’s union leaders were burned by defeats in the 2000s and don’t want to repeat the experience. It’s a much safer bet to take on only the relative handful of drives that are very likely to win, rather than shooting for the moon — as the crisis of our labor movement, our society, and our planet demands.

Different Ways to Assess Victory

Too much fear of losing can also reflect narrow strategic viewpoints. Things look differently when you’re assessing the health of the workers’ movement as a whole, not just your particular union local, and when you assess outcomes not just by what you win from employers, but also by how much solidarity, capacity, and consciousness you forge in struggle.

Defeated efforts can sometimes strengthen the overall national movement if the fight inspires others. For instance, even though the spring 2021 hot shop drive at an Amazon warehouse in Bessemer, Alabama, lost its election, it played an important role in boosting the labor zeitgeist and in inspiring other Amazon workers to start organizing.

And while first contracts are an essential goal, they’re not the only ways to judge short-term success. United Auto Workers (UAW) president Shawn Fain was right to highlight the major advances Mercedes workers have already wrested through their unionization effort:

Workers won serious gains in this campaign. They raised their wages, with the “UAW bump.” They killed wage tiers. They got rid of a CEO who had no interest in improving conditions in the workplace. Mercedes is a better place to work thanks to this campaign, and thanks to these courageous workers. . . . There are more than 2,000 workers at Mercedes in Alabama who want to join our union. They aren’t going away. The sun will rise, and the sun will set, and our fight for justice for the working class will continue.


Becoming less fearful of defeat — and adopting a more expansive view of what constitutes a win — is an important step toward taking more risks, adopting new worker-to-worker techniques, and massively funding new organizing projects like the UAW’s $40 million campaign to unionize the South.

Put simply: the available evidence strongly confirms the case made by UAW’s communication director Jonah Furman after last week’s loss:

What I know: only 6 percent of private sector workers are in a union in this country.

We’re not gonna change that by playing it safe. We’re gonna lose a lot of battles along the way. What’s gonna sustain us isn’t a glide path to victory. It’s something deeper than any one vote.

And all your best laid plans and theories and handbooks are nothing compared to real life experience. Trying stuff. Taking risks. Trying again.