The noticeable uptick in labor organizing in the wake of the Covid pandemic has led many journalists and academics to make big claims about where the union movement might be heading. In such accounts, a “restive” labor movement coming off of “landmark” wins is poised to ride its current “organizing and bargaining momentum” to a true resurgence. These bold claims—now admittedly dampened by the reality of the new Trump administration—give the appearance that labor’s at least been on the right path, and that what it needs is simply a whole lot more of what we’ve seen in recent years, but with more resources behind it.

The numbers, however, give us good reason to question inflated talk about labor’s recent “banner years” and “historic victories.” Despite renewed interest in unionization and high public approval of unions, union density continues its seemingly inexorable decline, reaching a record low of 9.9 percent in 2024. Even with increases in National Labor Relations Board elections and the rate of union victories in those elections, the number of workers being organized every year is not keeping pace with labor market growth or compensating for union member losses elsewhere, either through plant closings or decertifications. As labor journalist Hamilton Nolan pointed out when these numbers were released in January, “Over the course of the most pro-union presidency in my lifetime, not only did union density not rise—it declined into single digits.”

This numbers problem is no secret. Though often pushed into the subconscious for much of the rest of the year, it lurks uncomfortably behind all the cheerleading labor coverage.

If the percentage of people in unions shrinks, something clearly is not working. But that’s not the only reason we should question the idea that labor just needs to keep doing what it’s been doing. There’s also been a notable shift in who is joining unions, and where they’re joining them—and that shift contains worrying trends of its own.

From one angle, it makes perfect sense that union membership composition would be changing, given the continued decline of heavy industry and the growth in service work. But union composition is not simply changing to match the overall composition of the workforce; rather, union membership is growing in a few key sectors—mostly among urban workers in academia and healthcare—and shrinking or not keeping up with labor market growth just about everywhere else.

If we look, for instance, at a list of large-unit labor elections, i.e., NLRB elections involving 250 or more eligible voters over the past decade, this trend becomes clear. Unions have been running and winning such elections at a much higher clip since the pandemic, but this has only exaggerated the dominance of academic and healthcare elections. In 2023, 51,814 academic and healthcare workers became union members through large-unit elections, while only 6,537 workers in all other sectors did so. We are unlikely to see such an exaggerated disparity in future years, but it’s still indicative of a marked compositional evolution. Through the lens of large-unit elections, the much-vaunted recent “resurgence” in labor activity is a decidedly segmented phenomenon.

Current Issue

March 2026 Issue

Those looking for evidence that the uptick in academia and healthcare is spilling into other sectors might take some solace from 2024, when there were more workers outside of those sectors organized into unions through large-unit elections than in any other year in the past decade. But that 2024 count involved only 251 more workers than in 2014—hardly anyone’s “banner year” for labor organizing—and it was still dwarfed by academia and healthcare organizing. Projecting out from the first three months of 2025, it’s unlikely that that outcome will be repeated this year.

The overall Bureau of Labor Statistics union numbers tell a less dramatic but similar story. From 2014 to 2024, the private-sector industry with by far the largest growth in union membership was Education and Health Services, with 299,000 new members, the next being Transportation and Utilities at 68,000. Just since 2022, union membership in Education and Health Services has grown by 186,000; if that trend continues, fully one-third of all private-sector union members in the United States will be in education or healthcare by 2028.

In addition to being mostly confined to two sectors, which, according to the Bureau of Labor Statistics, comprise 15 percent of all employment in the country, recent union growth has also been a predominantly urban phenomenon. To determine this, I took Jed Kolko’s zip code classification data and applied it to all NLRB elections in 2014, 2019, and 2024. This exercise bears any number of caveats, including the key one that this data only tells us where workplaces are and that workers often travel from very far away to get to work. Still, there’s a pretty noticeable difference in the election composition in 2024, which tilts heavily urban. More or less all of the growth in labor activity after the pandemic is happening in urban settings.

Some might find no issue with labor growing disproportionately in academia and healthcare, and in urban settings, but we should nonetheless be clear when we talk about the supposed “labor resurgence” that this is a particular phenomenon, and not one that applies to the whole of organized labor.

But there’s good reason to see this trend as a problem to be overcome rather than as a neutral fact or something to be celebrated. To understand why, let’s look at a project that, as Jonathan Rosenblum and I have argued before, is existential for the labor movement: organizing Amazon workers.

In 2023, Amazon leapfrogged UPS in terms of total package volume; if Amazon continues to eat UPS’s lunch, the Teamsters are going to have a very difficult time maintaining their current UPS contract when it comes up for negotiation in 2028. The effects of a round of concessionary bargaining for this contract, the largest in the country, will resound throughout the union world. The only way for the labor movement to take fate into its own hands is to go all-in on organizing UPS’s primary competitive threat.

I ran the same zip code analysis on all Amazon Inbound Cross-Docks, Fulfillment Centers, and Sortation Centers, the big employment clusters in Amazon’s distribution network, and 92.4 percent of them are in rural or suburban areas. This is to say that the direction of the supposed “momentum” generated in the past few years is pointing away from the key facilities of this massive existential threat to labor. There is simply no vector from where we’ve been to where we need to be; organizing a company like Amazon requires a divergence from the activity of recent years rather than a continuation of it.

To be clear, it’s terrific that academic and healthcare workers have been organizing at such a rapid clip. The problem comes in seeing that organizing as representative of something broader than it is. In addition, there are other concerns to be raised about labor’s recent compositional evolution: The values and political orientations of academic and healthcare workers in predominantly urban areas are likely very different than those of rural and suburban construction or manufacturing workers. If the former becomes the face of the labor movement, will it act as a vanguard or a limitation? Will it revitalize unions or stir internal conflict? These are still open questions that will be decided in the coming years.

But to my mind, the primary problem here for the moment is simply to think that with labor heading in the right direction, all that is needed is to pour union resources into letting “a thousand organizing flowers bloom,” in labor scholar Eric Blanc’s words. That is a recipe for lopsided growth that targets the low-hanging organizing fruit and does not substantively affect 85 percent of the workforce, i.e., those workers outside of education and healthcare.

Fortunately, it is becoming clear to many people in the union movement not only that we need to organize Amazon and exert some control over rapidly changing logistical networks in the United States, but also that we need to devote resources to the problem that are an order of magnitude larger than have thus far been invested. Such a campaign is an absolute necessity at the present moment, but it is one that requires strategic prioritization that disrupts recent trends rather than furthers them along.