Executives are slimming hierarchies and pushing transformation while employee engagement sinks. Global engagement fell to 20% in 2025, the lowest since 2020, with Gallup estimating a $10 trillion productivity cost, roughly 9% of world GDP (Gallup, 2025). Manager engagement has dropped sharply too, from 27% to 22% in a single year, even as best-practice organizations report 79% of their managers engaged (Gallup, 2025). You can restructure, automate and set new OKRs, but if people do not feel safe to raise problems, ask for help or test ideas, the engine stalls. Psychological safety is not a vibe. It is a system condition that shows up in how teams learn and how work is designed.
What the construct actually means
Amy C. Edmondson introduced and defined team psychological safety as a shared belief that a team is safe for interpersonal risk taking (Edmondson, 1999). Her field study of 51 work teams in a manufacturing company found that psychological safety was associated with learning behaviors such as asking for help, discussing errors and experimenting, while team efficacy by itself was not when psychological safety was controlled (Edmondson, 1999). Learning behavior mediated the link between psychological safety and team performance. That is the key mechanism: safety enables the behaviors through which teams improve. The study was cross-sectional and based on one firm, so it should be read as careful field evidence on associations and process, not definitive proof of causality (Edmondson, 1999).
The definitional point matters in practice. Psychological safety is not about being nice or lowering standards. It is about whether people expect that speaking up will be treated as a contribution rather than a career risk. That expectation shapes whether problems surface early, whether knowledge travels and whether experiments happen in time to matter.
What the broader evidence says
A decade and a half of research has since accumulated. The first comprehensive meta-analysis, aggregating 136 independent samples and more than 22,000 individuals, found that psychological safety correlates strongly with learning behavior (estimated corrected correlation 0.62), information sharing (0.52), engagement (0.45) and task performance (0.43). The relationship with creativity was positive but modest at 0.13 (Frazier et al., 2017). The pattern is consistent with Edmondson’s mechanism: safety shows up first in learning and knowledge behaviors, and that is how it connects to performance.
Crucially, the same meta-analysis highlights design levers leaders control. Job characteristics such as autonomy (0.47), role clarity (0.63), and a supportive context (0.49) were significant antecedents of psychological safety (Frazier et al., 2017). Interdependence showed a very large association at the individual level (0.86), but that estimate came from only five samples and was more moderate at the group level (0.40), so it should be used with care (Frazier et al., 2017). The implication is straightforward. If you want more voice and learning, design work with clarity, real discretion and visible support, then make collaboration consequential rather than cosmetic.
Flatter structures raise the stakes
Many firms are moving toward fewer managerial layers and more empowered teams. The academic synthesis on self-managing organizations describes the archetype: radical, formal decentralization of authority that eliminates the traditional manager–subordinate relationship across the organization (Lee and Edmondson, 2017). The literature they review concludes that conventional hierarchy works for reliable execution of known tasks in stable settings, but tends toward rigidity on cross-functional, nonroutine problems and reinforces status differences that can stifle development and voice (Lee and Edmondson, 2017). This is a theoretical integration rather than an experiment, but the design logic is clear. Flattening removes the safety net of supervisory oversight while increasing the need for peer coordination and candid problem solving.
That makes psychological safety a non-negotiable infrastructure in flatter systems. When authority is distributed, teams must surface uncertainty and interdependencies quickly. Without safety, decentralization becomes a rumor mill and error correction slows. With it, distributed work can do what it promises: faster learning loops at the edges where reality changes first.
Managers are the canary
The sharp drop in manager engagement is a warning light, not a side note. Gallup reports a nine-point slide since 2022, with the steepest fall between 2024 and 2025, juxtaposed with 79% engagement for managers in best-practice firms (Gallup, 2025). The gap signals design, not destiny. The same antecedents that predict psychological safety map closely to what engaged managers experience: clarity on role and goals, autonomy to act, and a supportive context that makes collaboration real rather than rhetorical (Frazier et al., 2017). Treat manager engagement as a readout of whether the environment invites or punishes learning behaviors. If managers feel exposed for admitting uncertainty, they will transmit that fear to their teams. If they feel supported to escalate risks and ask for help, they will normalize the practices that create safety.
How to build it without slogans
The research points to moves leaders can make now. Keep them practical and observable.
- Make team learning visible. Track and discuss near misses, experiments run and help requested. Edmondson’s study found learning behavior is the route by which safety relates to performance, so measure the route, not just the result (Edmondson, 1999).
- Design for clarity and discretion. Tighten role clarity and decision rights while granting autonomy where the work benefits from judgment. These job characteristics are robust antecedents of psychological safety (Frazier et al., 2017).
- Make interdependence real and supported. Create tasks that require collaboration and back them with resources and coordination routines. Remember the very large individual-level correlation with interdependence rests on a small number of samples and is smaller at the group level, so test and adjust locally (Frazier et al., 2017).
- Remove status barriers that mute voice. The synthesis on less-hierarchical organizing warns that status differences suppress problem solving across functions. Audit rituals, meeting structures and language that signal whose input counts (Lee and Edmondson, 2017).
- Hold leaders to their reactions. The fastest way to erode safety is to punish bad news. Coach and evaluate leaders on how they respond when people surface errors or uncertainty. That is when the construct is either built or broken (Edmondson, 1999).
The practical read is simple. In a year when engagement is low and hierarchies are thinning, psychological safety is a hard-edged operating choice. It pays out through learning behavior, knowledge flow and steady performance, and it is shaped by design features within managerial control. If you are removing layers or shifting to team-based models, invest first in the conditions that let people speak openly and learn quickly. The alternative is a flatter chart with the same old silence.
Sources
- Edmondson (1999). Psychological Safety and Learning Behavior in Work Teams. Introduces team psychological safety and shows, in a cross-sectional field study of 51 teams in one company, that safety is associated with learning behaviors that mediate its link to performance. https://journals.sagepub.com/doi/10.2307/2666999
- Frazier, Fainshmidt, Klinger, Pezeshkan and Vracheva (2017). Psychological Safety: A Meta-Analytic Review and Extension. Aggregates 136 samples to show strong correlations between psychological safety and learning, information sharing, engagement and task performance, and identifies job-design antecedents, with noted caveats on interdependence. https://doi.org/10.1111/peps.12183
- Lee and Edmondson (2017). Self-managing organizations: Exploring the limits of less-hierarchical organizing. Defines self-managing organizations and synthesizes evidence that traditional hierarchy suits routine execution but impedes cross-boundary problem solving and voice, underscoring the need for safety in flatter designs. https://www.sciencedirect.com/science/article/abs/pii/S0191308517300059
- Gallup (2025). State of the Global Workplace 2025. Reports global engagement at 20%, an estimated $10 trillion productivity cost, and a sharp decline in manager engagement with a large gap to best-practice firms. https://www.gallup.com/workplace/659610/state-global-workplace.aspx