How Sprawl & Suburbs Have Upended the Organizing Terrain

Unfortunately, We Can’t Copy 1930s Union Targeting Strategies

Writing in 1936, Communist labor leader William Z. Foster made the following recommendation in his pamphlet on how to organize the steel industry: “Meetings should be held especially in popular neighborhood halls, where the workers’ fraternal lodges meet, where the workers dance, where their weddings take place, and where they are generally accustomed to going.”

The almost complete disappearance of these types of halls provides a small window into how America has been reshaped over the past century — a transformation that has made scaling up union power significantly more difficult and that precludes us from simply importing old organizing methods to today.

US labor’s big breakthrough in the 1930s was rooted in a centralized political economy far different from our own. Industries tended to be clustered in a relative handful of regions and cities. Workers tended to be concentrated in dense working-class neighborhoods adjacent to their jobs. In the largest corporations, though not necessarily elsewhere, jobs tended to be in massive factories. And due to the tightly bound nature of their assembly line and production systems, strikes by even a minority of well-positioned workers could relatively easily shut down their factory and dependent production chains.

Intense employer and state repression made organizing difficult and risky. But in such a context, it was a no-brainer for radicals and the Congress of Industrial Unions (CIO) organizers to target their efforts on big, strategic workplaces in large corporations. Thus US Communists in 1933 explained that their strategy was to concentrate on those “key factories of the decisive industry which will influence not only the workers in the small factories of the same industry but all the workers in the city and often influence the development of struggles in other parts of the country.”

As the watershed Flint Sit-Down strike of 1936-37 showed, knocking out a couple “mother plants” — like General Motors’ Fisher Plant Number One — could shut down production across the nation and set off a bottom-up, worker-led chain reaction through the entire region and industry. Along these lines, the CIO single-mindedly focused on unionizing hard industry, rejecting hundreds of requests for union charters in other sectors of the economy.

But in the decades since World War Two, the United States’ socio-economic topography has sprawled out, making it harder to organize workers and harder for organizing victories to diffuse. Such changes don’t necessarily make the labor movement less potentially powerful than it was in the past. But they do require that we more rigorously adapt our organizing methods to today’s conditions. 

An extreme illustration of how things have changed was provided to me by a staff organizer with the Utility Workers of America. He explained their difficulties unionizing wind techs in Texas:

Oh, my goodness, they were spread over — you know, Texas. Good luck finding these people, right, they're all over creation. And they have much less affinity toward one another, because they probably don't even know each other. They might work for the same employer, but work hundreds of miles apart. Contrast that to a distinct power plant, where everybody’s all right there, and you can see why it’s such a challenge.

And even when these Texas wind techs do manage to win a union, the resulting ripple effect will tend to be lower than in the past because these workers are less likely to live in dense working-class neighborhoods, where the good news of the union can easily spread by word of mouth. If dominoes are stood up too far from each other, knocking one down won’t set off a cascade. 

US labor revitalization requires a clearheaded map of our decentralized terrain and viable mechanisms for deliberately scaling up worker power. Unfortunately, many activists and academics have tended to underestimate (or misidentify) these transformations and the unique challenges they pose for organizers today. So let’s examine the data to see whether a deep, and deeply consequential, socio-economic dispersal has in fact taken place.

Economic Decentralization

In the first half of the 20th century, the US economy was dominated by manufacturing, which was concentrated in a relative handful of cities and regions. In the 1930s, for example, two-thirds of factory workers were employed in three percent of the nation’s territory. Taking advantage of this industrial concentration and a favorable political context under FDR and the Second World War, a grassroots workers upsurge unionized mass production industries in the late 1930s and early 1940s. 

But employers did not wait to launch a counter-offensive. Politically, unions from 1938 onwards were race-baited in the South, red baited everywhere, and hemmed in by a slew of new anti-labor laws culminating in 1947’s Taft Hartley Act. Economically, to escape the manufacturing belt’s strong labor movements, employers dispersed across the country, especially to low-wage, anti-union areas like the South and the Sunbelt. 

Manufacturing jobs have not only decentralized since WWII, they have declined significantly in number. Figure 1, below, charts the well-known decline of hard industry and the rise of services. 

What’s often missed about this shift is its spatial consequences: services are even more geographically dispersed than today’s factories, to say nothing of pre-WWII manufacturing. In other words, deindustrialization has dramatically accelerated economic dispersal. The reason for this is intuitive: most services need to be relatively close to the people to whom the service is being provided. It would make no sense, for example, to cluster Walmart superstores or hospitals in one or two regions of the country.

Large non-industrial corporations have, like their manufacturing counterparts, been “vertically disintegrated,” with various occupations that used to be done in house, such as cleaning or sales, farmed out to subcontractors often in entirely separate locations. Similarly, high-wage employees, like dispersed tech or white-collar workers at a company like Starbucks, often never interact face-to-face with the lower-paid employees who sell the companies’ products. 

Workplace Size Decline

One important result of manufacturing decentralization and the rise of services has also been a significant decline in the size of workplaces, especially within the largest corporations. 

A quick word of caution: we don’t have a good sense of the extent to which workplace size made large factories so strategically central in the past. Other crucially important factors include their close ties with surrounding residential communities, their highly disruptable production processes on the shop floor and beyond, and the ubiquity of big establishments within big companies, i.e. those with the deepest pockets for granting concessions. Size on its own does not necessarily make a workplace easier to organize. Unions for the past fifty years have had consistently much higher win rates at small and medium-sized establishments. Nevertheless, all other things being equal, larger workplaces do lower organizing costs by concentrating workers and, thereby, increasing their interpersonal ties and making it possible for organizers to have a wider impact.

Kim Moody’s influential 2017 book On New Terrain obscures trends towards smaller workplaces, especially in the largest American companies. As part of his case that the working class has been re-concentrated into large workplaces in large urban areas today by the logistics revolution, Moody writes that “the nation’s workplaces are not getting smaller measured by the average number of workers.” He justifies this conclusion by comparing the percentage of large workplaces in 1986 to today. But what happens when we zoom out our lens even a little further back in time? The data clearly show that since the mid-1970s the percentage of 1000+ establishments in all industries have declined, by about three percentage points, and that factory plant sizes in particular have fallen significantly. This countrywide decline in large establishments during a period of deindustrialization should not be surprising, since service sector workplaces have always tended to have fewer large worksites. 

And the available data going back to the 1930s, though somewhat spotty, suggests a far more dramatic decline in the centrality of large workplaces. We know, for example, that there’s no industry today even closely comparable to pre-war auto: in 1939, 76 percent of auto workers were employed in workplaces that had 1,000 or more employees. In contrast, the largest corporations today are generally dispersed into thousands of small and medium-sized establishments (see Table 1), making it significantly harder to scale up workplace power.

This proliferation of smallish workplaces at the heart of the US economy constitutes a dramatic shift from both the 1930s and the 1970s. By way of comparison, GM’s 69 plants employed 3,478 workers on average during the Great Depression and US Steel’s 121 plants averaged 2,159. Big companies, in short, no longer generally depend on big workplaces.

Suburbanization and the New Working Class

Economic decentralization is deeply bound up with residential dispersal and has often driven it forward. There’s little empirical basis for Moody’s suggestion that the dense urban working-class neighborhoods of old have been simply replaced by new dense working-class neighborhoods. The relationship of home to work, and of workers to other workers after their shifts are over, has been dramatically transformed since the era of US labor’s big breakthrough. 

As one geographer recently explained, “working class consciousness and organizing has tended to reach highest levels where workers live in dense communities close to their workplaces.” Capitalists learned this the hard way, as urban working-class explosions rocked intensely concentrated US cities beginning in the mid-1880s. Following in the footsteps of urban planners in places like Paris after 1848’s center-city insurrections, in America big employers and city planners reacted by increasingly pushing and pulling workers out of downtown urban centers, especially from 1899 onwards. This, however, was a slow process: for the first half of the 20th century, most workers continued to live in densely packed urban neighborhoods.

We shouldn’t romanticize that era. Working people a century ago were neither homogenous nor united on a class-wide level. Deep ethnic and racial divisions sorted workers of different backgrounds into distinct blocks, neighborhoods, or states. And women were all-too-often marginalized in a masculinist culture. But the compactness of urban working-class life facilitated an associational effervescence — religious, social, cultural, and sometimes political — among workers of all backgrounds. Well before the rise of mass trade unionism, the available evidence suggests that being part of a class-based collective was more deeply felt than today because it was more deeply rooted in the routines of daily life.

Successfully organizing one node of these tight social networks into a union could set off an organic chain reaction. Unfortunately, there aren’t many places today that resemble southern Michigan circa 1936, as described by two sociologists:

Workers from the many GM establishments—as well as workers from other Big Three plants and myriad suppliers—lived in the same neighborhoods, socialized in the same bars, and played in the same parks. The tightly coupled southern Michigan production culture therefore created enduring personal and community relationships among workers who—in a dispersed production system—would have had few useful relationships unless they deliberately sought them out. … [W]orkers from different factories knew each other, creating a city- and even regionwide interpersonal network that facilitated the type of grassroots organizing utilized by UAW organizers in Flint and elsewhere in the auto production culture.

This type of world was stretched apart by the mass suburbanization. Even a cursory look at the data on population density in urban areas registers a dramatic decrease from 1950 to 2000 — a 25.8 percent drop, on average, according to one dataset. Chicago, for example, has become almost as sprawled out as Los Angeles, dropping from 17,409 inhabitants per square mile to 12,746. 

As car commuting, isolated tract homes, and television became the norm after WWII, working people came to live and socialize further from each other as well as from their places of employment. As Eric Hobsbawm notes, the late 20th century was marked by “the breaking of the threads which in the past had woven human beings into social textures.”^⁠ Though white and white-collar workers were the first fractions to leave the city, suburbanization has become the norm for the American working class as a whole. In many metropolitan areas, a majority of low-income households live in the suburbs, as do a majority of Americans. 

The extent of this spatial-social fragmentation is well captured by changing commute distances. In 1934 Pittsburgh, more workers walked to work than drove. As late as 1950, a majority of US working-class families did not own a car. Even in car-centric Detroit, mid-century UAW shop stewards lived on average 3.4 miles from work. 

In contrast, the average American today commutes 20.5 miles to work each way — a 27 percent commute time increase since 1980 (the first year the US Census began tracking the figure). Though there’s no older nationwide data for the United States, findings from Britain reveal that a huge shift has taken place over the past century — and that Americans have exceptionally long journeys to work by international standards (see Table 2). As Robert Putnam explains in his bestselling book Bowling Alone, “each additional ten minutes in daily commuting time cuts involvement in community affairs by 10 percent.”

Table 2: British Commute Distances, One-way

1890-99: 2.23 miles

1930-39: 4.34 miles

1999-98: 9.07 miles

2019: 11.5 miles

It’s not just unions that have paid the price. As Putnam details, the United States since the 1960s has witnessed an equally dramatic erosion of all forms of associational life, strong interpersonal networks, and church attendance among Americans of all backgrounds. Recent studies have confirmed these findings, including the fact that personal discussion networks have shrunk by one third since 1985. Though territorial sprawl is certainly not the only factor driving this atomization, it is clearly a phenomenon that organizers have to acknowledge and find strategic answers to.

What About Logistics?

Not everybody agrees with the picture of geographic dispersal I’ve painted above. Some authors and organizers in recent years have insisted that the rise of logistics — transportation, warehousing, etc.— that manage just-in-time supply chains has recreated organizing conditions similar to the 1930s by geographically concentrating large numbers of workers with a high degree of structural power. Like the “clusters of auto assembly plants of yesteryear in Detroit or the steel mills in Gary” writes Kim Moody, today’s warehouse and transportation workers, as well as their inter-linked white-collar and point-of-sale peers, are “tied together by urban-based concentration” as well as “the modern just-in-time supply (production) chain.” Concluding that “Amazon can be today what General Motors was for organized labor in the 1930s,” Kim Moody argues that modern supply chains, due to their numerous chokepoints, “have reproduced the vulnerability that capital sought to escape through lean production methods and relocation.” 

This analysis suggests a relatively unproblematic and straightforward approach to scaling up: organize drivers and warehouse workers at the nucleus of contemporary capitalism, and hope to detonate a working-class explosion like that seen in the 1930s. Take John Womack’s case for focusing on employees with the power to disrupt supply-chain choke points:

Unionizing movements happen in a kind of avalanche. That one rock falls, and that moves other rocks, and eventually the whole side of the mountain gives way. And that’s what you want in a company like Amazon or in the Big 3 automobile industries. But Amazon, in particular, depends on those warehouses … And if [unionists] can organize several of the big warehouses, that, I think, would make a big, big difference.

Much of this is accurate. Organizing drivers and warehouse workers is pivotal because just-in-time supply chains provide them with significant leverage. And logistics workers are economically central: three of the eight largest companies are explicitly in logistics (Amazon, UPS, FedEx) and the big retailers on that list (Walmart, Home Depot, Target) are also central drivers of the logistics revolution.

But a closer look under the industrial hood shows why we shouldn’t try to replicate the old hyper-targeted industrial strategy of the Communists and CIO. For starters, organizing today’s logistics hubs is made more difficult because metropolitan areas are less dense than they used to be. One of the reasons why the unionization win at a big Amazon fulfillment center like JKF8 on Staten Island was so exceptional is that workers were obliged to commute hours to work each way, leaving them with less time for off-work socializing and with fewer personal ties with their equally dispersed co-workers. Organizing something like systematic house visits to talk about the union with all employees — a crucial step in the summer of 1936 drive leading to the Flint sit-down strike — is much more cost-intensive in today’s suburbanized context. “Urban-based” means something quite different socially than it did ninety years ago.

Inside a warehouse itself, as sociologist Nantina Vgontzas has uncovered in her rigorous research on Amazon, a flexible, redirectable labor process means that, unlike in an assembly line, small groups of workers refusing to work can’t shut the place down. Even more importantly, as she notes, the “dispersal of the labor process across warehouses” means that no single workplace can paralyze the broader distribution network. There are too many different nodes through which shipments can be rerouted: “In most cases, management can simply block off that warehouse by pressing a ‘little red button,’ which tells the system to not send any new orders to that warehouse.” Due to this optimized-and-dispersed labor process within and between Amazon’s 1200-plus workplaces, “a few strategically placed workers can no longer shut down an entire warehouse or distribution channel, as they could in the golden days of manufacturing.”

The unprecedented importance of logistics in our contemporary economy results precisely from imperatives to connect sprawling networks of decentralized factories, stores, and offices. Like in the rest of the economy, tapping the structural leverage of warehouse and transport workers thus requires deliberatively finding ways to scale up on a highly decentralized terrain. To quote Vgontzas:

While the form of power present in today’s warehouses can be likened to that in yesterday’s factories … the scope of its activation will need to be far more diffused. Amazon’s fulfillment network will not be shut down by workers strategically placed along the assembly line in several key facilities, as in the interwar strikes that led to the initial unionization of mass industry. It will require a much wider level of coordination within fulfillment centers, across regions, and between nodes in Amazon’s network, from fulfillment centers to the tech offices that optimize fulfillment work.

And the same approach to proactively scaling up is true for deliberately unionizing widely beyond logistics, an industry that doesn’t carry anywhere near manufacturing’s numeric weight nor its magnetic role in the broader community. Whereas manufacturing workers made up 34 percent of the US workforce in 1939, transportation and warehouse workers make up 4 percent of the workforce today.^⁠ 

Conclusion: A New Worker-to-Worker Model

It’s true that because logistics workers link together worksites in so many industries — especially retail — they have lots of potential leverage to support other unionization efforts. To quote a Starbucks worker organizer in Boston: “A big part of winning a first contract, I think, will be organizing the supply chain and getting solidarity from Teamsters, who deliver things to our stores. That’s how we could really bring business to a halt.”

Targeting strategic workplaces and companies is still an important tactic; some jobs and some industries are still more strategic than others, due to their disruptive capacity, economic centrality, social weight, etc. But given logistics workers’ relatively low numbers, and an overall context of territorial dispersal and social fragmentation, we should expect that triggering and spreading union avalanches today will require initiatives to scale up that are less hyper-targeted, more geographically spread out, more economically varied, and more digitally enabled than was necessary or possible in the 1930s. And we need to place more of a focus than previously on recreating a deeply felt class culture, via high-profile union battles, insurgent anti-establishment electoralism, proactive socializing, and viral class-struggle social media agitation. New conditions require a new organizing approach.

Though decentralization since the 1930s means that we can’t just copy and paste strategies from that era to today, we also shouldn’t throw the baby out with the bathwater. One crucial strength of Great Depression-era unionism was that it was a worker-to-worker model that relied far less on staff than is currently the norm. Economic transformations, in short, have made such a grassroots approach more relevant than ever.

People like to say that unionizing mega corporations like Amazon and Starbucks today is a David vs. Goliath battle. But at least with Goliath, David only had to concentrate on fighting one fight. With Amazon and Starbucks, you have to wage countless separate union battles in every corner of the US. In that sense, mass unionization today is more like confronting the Borg from Star Trek or the Mind Slayer in Stranger Things, hive minds that act through a legion of separate bodies. 

How can unions effectively wage that kind of many-fronted battle? In today’s decentralized conditions, the only way to build worker power at scale is by relying less on paid full-timers and more on workers themselves. Though staff-intensive organizing is often very effective, and often does a great job at training workers to lead their co-workers in struggle, it costs too much to grow widely. The good news is that recent bottom-up struggles in cafes, journalism, auto, and higher-ed have developed a scalable, digitally enabled approach to building worker power across wide spatial divides.

What does this new approach look like exactly? I’ll be digging deep into this question over the coming months in this Substack — and in my forthcoming book We Are the Union: How Worker-to-Worker Organizing Can Transform America (UC Press). For now, here I’ll just leave you all with a brief summary of how today’s worker-to-worker unionism differs from its 1930s’ predecessor:

[This is a working paper — feel free to share on social media, but don’t republish anywhere without my permission, thanks! For a version with citations included, see here]