The debate over flatter organizations often proceeds as though the underlying question were new. It is not. Decades before anyone marketed a bossless workplace, industrial and organizational psychologists were measuring what happens when you redesign a job to give the person doing it more control. The results are some of the most robust in the field, and they reframe self-management as less of a leap of faith than its critics and some of its champions assume. Self-management is, at bottom, an argument about work design at the scale of the whole company.
The number that anchors the field
The cornerstone is a meta-analysis of the work-design literature that pooled 259 studies covering 219,625 participants (Humphrey, Nahrgang and Morgeson, 2007). Across that enormous base, fourteen work-design characteristics explained, on average, 43 percent of the variance in nineteen worker attitudes and behaviors (Humphrey, Nahrgang and Morgeson, 2007). How work is designed is not a minor influence on how people perform and feel. It is close to half the story. And within that, the motivational characteristics, the cluster that includes autonomy, did the heavy lifting: they explained about 25 percent of the variance in subjective performance, 34 percent of the variance in job satisfaction, and 24 percent of the variance in organizational commitment (Humphrey, Nahrgang and Morgeson, 2007).
One caveat has to travel with those figures. Autonomy sits inside the broader motivational bundle in this analysis and is not isolated as a standalone effect (Humphrey, Nahrgang and Morgeson, 2007). So the honest claim is not "autonomy alone explains a third of job satisfaction." It is that the family of motivational design features, autonomy prominent among them, carries a large and well-established share of the variance in how people perform and how they feel about their work. That is still a striking result, and it is the empirical soil in which self-management grows.
From a design feature to an operating system
The conceptual move that self-management makes is simple to state. Work-design research shows that loading autonomy into a single job produces measurable gains. Self-management asks what happens if you make that latitude the organizing principle of the entire firm rather than a feature of a few enriched roles. The recent well-being research treats this seriously by measuring self-management at two levels at once, the organization and the individual employee, and finds both positively related to work engagement (Morikawa, Martela and Hakanen, 2025). That is what you would expect if self-management were autonomy generalized: the thing that lifts engagement in a well-designed job also lifts it when it becomes the way the whole organization runs.
Seen this way, self-management is not a rejection of decades of management science. It is an extrapolation of one of its most durable findings. The burden of proof shifts. The question is no longer whether giving people control over their work helps, because the work-design evidence settled that at scale (Humphrey, Nahrgang and Morgeson, 2007). The question is whether the gains that show up in individual job design survive being extended to the level of the entire organization, and under what conditions.
The catch the data also carries
Extrapolation is not proof, and the same literature that supplies the promise supplies the limits. A systematic review of self-managing teams found the performance evidence inconsistent, dependent on the right input factors at the individual, team, and organizational levels (Magpili and Pazos, 2018). Autonomy is powerful, but pouring it over a team that lacks the surrounding conditions does not reliably produce the meta-analytic gains. The individual level carries its own condition: self-management improves performance only if members are competent to navigate it, with self-leadership and interpersonal skills singled out (Doblinger, 2022). Latitude and capability have to arrive together. Latitude without capability is not enrichment, it is abandonment.
This is why the strongest and the most cautious claims about self-management are both true. The design principle it rests on, autonomy and enriched work, is exceptionally well evidenced (Humphrey, Nahrgang and Morgeson, 2007). Whether a given organization realizes that potential depends on conditions the same body of research spells out in detail (Magpili and Pazos, 2018; Doblinger, 2022). The lever is real. Pulling it does something only if the rest of the machine is built to transmit the force.
The practical read
Treat self-management as work design scaled up, and use the work-design evidence to steer it. Start from the premise the meta-analysis established, that latitude over one's work is a large and reliable driver of performance and satisfaction (Humphrey, Nahrgang and Morgeson, 2007), and build the organization to extend that latitude rather than ration it. Then respect the conditions the research attaches: the input factors across levels (Magpili and Pazos, 2018) and the individual competencies that let people use autonomy well (Doblinger, 2022). The case for self-management does not rest on a hope that people are better than their org charts assume. It rests on forty years of data about what happens when you give them room, and on a sober reading of what has to be true for that room to pay off.
Sources
- Humphrey, Nahrgang and Morgeson (2007). Integrating Motivational, Social, and Contextual Work Design Features. Pools 259 studies and 219,625 participants, showing work-design characteristics explain 43% of variance and motivational features (including autonomy) explain large shares of performance, satisfaction, and commitment. https://asu.elsevierpure.com/en/publications/integrating-motivational-social-and-contextual-work-design-featur/
- Morikawa, Martela and Hakanen (2025). Self-management and employee well-being. Measures self-management at organizational and employee levels and links both to work engagement, consistent with autonomy generalized to the whole firm. https://journals.sagepub.com/doi/10.1177/23409444241307734
- Magpili and Pazos (2018). Self-Managing Team Performance: A Systematic Review of Multilevel Input Factors. Shows the performance payoff is inconsistent and conditional on multilevel input factors. https://doi.org/10.1177/1046496417710500
- Doblinger (2022). Individual Competencies for Self-Managing Team Performance. Establishes that autonomy pays off only when members hold the competencies, especially self-leadership, to navigate self-management. https://doi.org/10.1177/10464964211041114