Rolebase

The 1970s self-management wave: what it proved, and why it was shut down

Lip, General Foods, Volvo: what the 1970s self-management experiments proved, why management shut them down, and what survived.

January 30, 2026

Sandhya Domah

Key takeaways
  • Daniel Bell wrote in 1956 that the American worker had been "tamed" by the consumption society rather than by the discipline of the machine.
  • The 1972 Work in America report put numbers on the gap: 93% of urban university professors would pick the same job again, against 16% of unskilled auto workers.
  • The Gaines plant in Topeka ran on about 70 people where the conventional design needed 110, and head office still took the system back.
  • The 1970s experiments depended on a protective sponsor. The ones that last today depend on written roles and a decision rule.

In 1956 the sociologist Daniel Bell published Work and Its Discontents. He argued that the American worker had been "tamed" by the consumption society rather than by the discipline of the machine, by the promise of a better life that his wage, his wife's second income and easy credit put within reach. In Detroit, Bell watched shop-floor discontent turn into escape fantasies instead of militancy: a mechanic's shop, a turkey farm, a gas station, "owning a small business of one's own". He closed the passage with three words. "An idle dream."

Fifteen years later the shop floors stopped.

The line revolts

On 5 March 1972, 7,500 workers walked out of the General Motors plant at Lordstown, Ohio. The line there was turning out Vegas at 101 cars an hour, the fastest in the world at the time, after a speed-up decided in the autumn of 1971 that came with 300 layoffs. The strike ran twenty-two days. The settlement brought the laid-off workers back and dropped discipline against 1,400 strikers, and left the line speed exactly where it was.

Executives blamed the young first. The new generation had lost the meaning of work, they explained, in language you can read again today aimed at Gen Z.

What the Work in America report actually measured

In December 1972 a task force convened by the US Department of Health, Education and Welfare, chaired by James O'Toole, delivered a report called Work in America, published the following year by MIT Press. It reused a question that had run through two decades of surveys, compiled by the psychologist Robert Kahn: what kind of work would you try to get into if you could start all over again?

Occupational groupWould choose similar work again
Urban university professors93%
Mathematicians91%
Journalists (Washington correspondents)82%
Solo lawyers75%
Skilled printers52%
White-collar workers, cross-section43%
Skilled auto workers41%
Textile workers31%
Blue-collar workers, cross-section24%
Unskilled auto workers16%

The report went past the observation. It cited Harold Sheppard and Neal Herrick's Where Have All the Robots Gone? (1972), a survey run in 1970 and 1971 among American blue-collar workers: fewer than half said they were satisfied most of the time, and the share of positive answers tracked how much variety, autonomy and meaningful responsibility the job carried. The central chapter recommended redesigning jobs so that workers took part in the decisions that shaped them.

Two experiments that answered different questions

In January 1971 General Foods opened a Gaines pet food plant in Topeka, Kansas, designed around self-managing teams. Team leaders picked 63 operators out of 600 applicants. Teams split the work between themselves, ran their own schedules and took part in hiring. Pay followed the skills a person had mastered rather than the slot they filled, across four rates. Richard Walton told the story in the Harvard Business Review of November 1972: the plant ran on roughly 70 people where the conventional design would have called for 110, absenteeism stayed under 2%, turnover under 1%, and Walton reported $600,000 of annual savings after eighteen months.

In June 1973, in Besançon, the workers of the watchmaker Lip occupied their factory against a redundancy plan. They restarted production, sold the watches directly and paid themselves out of the takings, behind the slogan "on fabrique, on vend, on se paie" and behind Charles Piaget, a CFDT organiser. Police cleared the site on 14 August 1973 and the fight carried on in sports halls and cinemas around the town. The Dole agreements of 29 January 1974 provided for 850 rehires. Claude Neuschwander took over the watchmaking business, resigned in February 1976, and the new company filed for bankruptcy two months later. The workers occupied again on 5 May 1976, liquidation came on 12 September 1977, and in November they founded six cooperatives including Les Industries de Palente, which wound down through the 1980s.

The two stories travel together and they answer different questions. Topeka was designed by a management team, and it measured what autonomy produces when the company organises it. Lip was an industrial dispute where self-management served as a weapon, and it showed how far workers can go when the company walks away.

The day efficiency stopped counting

In 1968, in an engine plant near Detroit, management scheduled a six-week inventory and held more than fifty men at work who would otherwise have been laid off on 90% of their pay. The men organised themselves and aimed to finish in three or four days so they could take the rest of the time off. Management stopped them, on the grounds that the legitimate channels of authority, training and communication had been violated. The fifty spent six weeks working under foremen. Bill Watson, one of the workers, published the account in Radical America in May 1971.

The episode came out of no official programme. It came out of workers organising on their own, and that is what makes it useful. Management had a choice between an inventory finished in four days and its org chart. It kept the org chart.

The same trade-off played out at Topeka, more slowly. On 28 March 1977 Business Week ran "Stonewalling Plant Democracy" and described a plant that had become too threatening to too many people inside the company. Corporate functions took back pay, hiring and team organisation. In 1983 David Whitsett and Lyle Yorks reopened the file in the California Management Review and showed that the publicity built around Topeka had distorted the case and ended up damaging it inside General Foods.

The usual objection to autonomy is about performance. At Topeka the performance was there. The sticking point was scope. Once a team picks its own schedule, its own pace and its own new hires, someone has to say where its authority ends.

What is left when a fashion recedes

By the late 1970s the vocabulary had softened. First quality of work life, then employee involvement, then quality circles. The content shrank as the name became easier to sell. Teams kept control over how the work got done and handed back control over the decisions that framed it. In France, self-management stayed a political slogan carried by the CFDT and the PSU around 1970, which made the word itself untouchable for company boards.

Volvo pushed the idea further than anyone. The Kalmar plant, opened in 1974, replaced the line with teams and carriers. Uddevalla, opened in 1989, had small teams assemble complete cars. Volvo announced the closure of both in 1992, Uddevalla shut in 1993 and Kalmar in June 1994. At the time of the announcement, management rated Uddevalla's productivity comparable to its conventional Torslanda plant. The model that won the following decade came from Toyota, with the Toyota Production System and its autonomy bounded by the workstation.

A handful of companies held the line without reversing, and they share one trait. Each replaced hierarchy with something written down. W. L. Gore has run on a network of sponsors rather than a chain of command since 1958. Ricardo Semler took over Semco in 1980, cut the layers, let teams of six to ten set their own budgets and targets, and told the story in the Harvard Business Review of September 1989. Morning Star, a Californian tomato processor founded in 1970, replaced job descriptions with written agreements between colleagues, the Colleague Letters of Understanding. Buurtzorg, started in the Netherlands in 2006 by Jos de Blok with four nurses, now employs more than 10,000 people in teams of ten to twelve with no middle management.

What the 1970s experiments were missing

Topeka rested on a plant manager and a consultant. Once both had moved on, the only trace of the system was the habit of the teams, and a habit comes apart in one reorganisation. The same mechanism explains why so many liberated company programmes ran out of air forty years later. They describe precisely what to remove and leave the culture to handle the rest.

An organisation that distributes authority for the long run writes three things down.

Roles come first. Each one carries a purpose, a domain its holder decides on alone, and accountabilities the rest of the organisation can ask for. A role is smaller and more mobile than a job, one person holds several, and it changes when the need changes. That is the whole difference between a role and a job description.

The decision rule comes next. It says who decides, over what scope, and through which process when a subject runs wider than a single role. Consent decision-making answers that without demanding unanimity and without pushing the subject back up to the founder.

The place where the structure changes comes last. A regular governance meeting creates a role, amends another, works through a tension about how the organisation runs. Without it the org chart freezes while the real work drifts away from it, month after month. Sociocracy supplies these mechanisms as a package, and borrowing the useful pieces works too.

The men who finished that inventory in four days had nothing in writing to put in front of the foremen. That was the only thing they lacked.

Rolebase exists to produce exactly that document: a living org chart where every role carries its purpose, domain, accountabilities, checklist and indicators, structured meetings, and consent decisions that stay on the record. See what Rolebase does, or set up your organisation free for up to five active members.

Frequently asked questions

What is self-management in a company?

Self-management describes an organisation where teams decide how the work is split, how it gets done, and a share of the decisions that surround it, without going through a chain of command. The term covers two different things: systems designed by management, like the Topeka plant in 1971, and takeovers by the workers themselves, like Lip in 1973.

Why did the 1970s self-management experiments stop?

The economics worked. Topeka documented productivity gains, absenteeism under 2% and turnover under 1%. The blockage was about scope. A team that picks its schedule, its pace and its new hires eventually raises the question of who decides everything else. Business Week described in March 1977 a plant that had become too threatening to too many managers inside the company.

What happened to the Lip workers?

The Dole agreements of 29 January 1974 provided for 850 rehires and watchmaking restarted under Claude Neuschwander. He resigned in February 1976, the company filed for bankruptcy, the workers occupied again on 5 May 1976 and liquidation followed on 12 September 1977. The workforce then founded six cooperatives, including Les Industries de Palente, which wound down through the 1980s.

Which companies run on self-management today?

W. L. Gore has operated without a chain of command since 1958, Semco was reorganised by Ricardo Semler from 1980, Morning Star has replaced job descriptions with written agreements between colleagues since the 1990s, and Buurtzorg runs more than 10,000 Dutch care workers in teams of ten to twelve since 2006. Each put a written structure in place of the hierarchy it removed.

Photo by Birmingham Museums Trust on Unsplash.

Continue reading