The argument about hierarchy and network almost always happens inside one company. Should this firm have managers or not. That framing hides the more consequential contest, the one between a single large pyramid and a federation of small, self-governing units bound together by cooperation rather than command. The platform and ecosystem economy of the 2020s has made this the live question, and the answers are older than the technology.
Networks between units, not just within them
Inside a firm, self-management means teams that govern themselves. Between firms, the same logic scales up into federation. Paul Singer, the principal theorist of Brazil's solidarity economy, describes the shape directly. Singer (2001) argues that solidarity makes cooperation between enterprises natural, in contrast to the destructive competition that a pure market imposes. A network of cooperating units can pool what none could afford alone, share demand, and route work between members, achieving coordination without a headquarters issuing orders.
The recovered factories of Argentina put this into practice under pressure. Pizzi and Brunet Icart (2014) document not just individual occupied plants but networks of mutual support between recovered firms, an inter-firm fabric that helped these cooperatives survive where a single isolated unit might not. The unit of resilience was the network, not the firm. That is the federation logic in its rawest form: small self-governing nodes that lean on each other instead of on a central authority.
Semco built its own internal federation
Semco is usually read as a story about workplace democracy, but its structural innovation was closer to federation. The teaching case records that from 1990 the firm spun off satellites, autonomous, agile, innovative divisions such as its technology-innovation nucleus, that operated as independent units rather than departments under a central command (Siehl, Killian and Perez, 1999). Semler (1989) describes the same instinct in the founding logic: break the firm into small, self-directed pieces that see their own numbers and govern their own work. Semco effectively turned part of itself from a pyramid into a federation of small businesses, which is why it could stay adaptive at a size where most firms calcify.
This internal-federation model reappears when flat firms grow up. Groth (2020) documents Zappos, after retreating from holacracy, settling on an internal market in which teams run as small businesses with their own profit and loss and transact with one another. That is a company reorganizing itself as a federation of accountable units rather than a single chain of command, and reaching for prices and cooperation to coordinate them.
It is worth being precise about why federation coordinates without a headquarters. In a pyramid, the center resolves conflicts between units by fiat, which works until the center becomes a bottleneck and stops understanding the units it is arbitrating. A federation replaces the fiat with two lighter mechanisms: cooperation, where units help each other because their fates are linked, and prices or negotiated exchange, where units trade on terms they set themselves. Singer (2001) leans on the first, arguing that solidarity makes inter-firm cooperation natural; Zappos leaned on the second, letting internal prices carry the coordination that managers used to. Both dispense with the central command line, and both require the units to be genuinely accountable for their own results, or the discipline evaporates.
The scale problem is the catch
Federations are not a free lunch, and the honest version of this argument has to name the catch. Singer (2001) is explicit that small solidarity ventures need scale and consolidation to escape poverty and compete, pointing to Mondragón as the example of a cooperative network that reached real size. A loose federation of tiny units can be nimble and fragile at once. Without some mechanism to aggregate, whether a shared bank, a common brand, or a federating body, the network stays small and vulnerable, out-muscled by the very pyramids it was meant to outmaneuver. Singer's argument is conceptual rather than empirical, so the claim is a direction, not a proof.
History has already run part of this experiment inside Semco. The satellites were not a loss of control but a redesign of it: Semler (1989) describes deliberately breaking the company into small, self-directed pieces that saw their own numbers, and the teaching case notes those satellites became independent, agile, innovative divisions rather than departments waiting for instructions (Siehl, Killian and Perez, 1999). The firm traded the tidy legibility of a single pyramid for the adaptiveness of many small units, and accepted the coordination overhead that came with the trade. That overhead is the price of the model, and pretending it away is how enthusiasts oversell it.
The practical read for a leader thinking about structure is to widen the lens past the org chart. The choice is not only how to arrange people inside the firm. It is whether to grow by extending one command structure or by spinning off self-governing units linked by cooperation and, where useful, by internal prices. Semco's satellites and Zappos's internal market suggest the federated path can keep a large organization adaptive. Singer's caveat and the Argentine experience warn that federation without a way to reach scale stays precarious. The pyramid's advantage was always that it scaled cleanly. Beat it by building a network that can too.
Sources
- Singer, P. (2001). Economia solidária versus economia capitalista. Argues solidarity makes inter-firm cooperation natural and warns small ventures need scale and consolidation, citing Mondragón, the core theory and its caveat. https://www.scielo.br/j/se/a/Xy7BmyrV8tHfwKNVhmSXFyw/?lang=pt
- Pizzi, A. & Brunet Icart, I. (2014). Autogestión obrera y movilización social: el caso de las empresas recuperadas argentinas. Documents mutual-support networks between recovered firms, showing the network as the unit of resilience. https://muse.jhu.edu/article/540857
- Siehl, C., Killian, D. & Perez, F. (1999). Ricardo Semler and Semco S.A. Records Semco spinning off autonomous satellites from 1990, an internal federation of self-governing units. http://www.saludygestion.com/archives/Ricardo%20Semler%20&%20Semco,%20Thunderbird.pdf
- Semler, R. (1989). Managing Without Managers. Describes breaking the firm into small self-directed pieces with open numbers, the founding logic of Semco's federation. https://hbr.org/1989/09/managing-without-managers
- Groth, A. (2020). Zappos has quietly backed away from holacracy. Documents Zappos adopting an internal market of teams with their own profit and loss, a firm reorganizing as a federation of accountable units. https://qz.com/work/1776841/zappos-has-quietly-backed-away-from-holacracy