Return-to-office edicts and AI pilots have not stemmed a quieter crisis: shrinking motivation. In 2025 only 20% of employees worldwide reported being engaged at work, the lowest level since 2020, with Gallup estimating a $10 trillion drag on output, or 9% of global GDP (Gallup, 2025). Manager engagement fell sharply too, down to 22% from 27% in a single year. The gap between common practice and what is possible is glaring. In the best organizations Gallup tracks, 79% of managers are engaged, nearly four times the global average, which suggests a design problem more than a generational mood (Gallup, 2025).
The question is where to pull. Three levers stand out in the strongest research: give people real discretion over how they shape their jobs, make it safe to speak up and learn, and, where the work warrants it, move authority out of the traditional chain.
Start with job crafting, and measure engagement
Job redesign can sound like a months-long project. It does not have to be. A meta-analysis of 14 job crafting interventions shows that even light-touch programs reliably nudge people to tailor their tasks, relationships, or perspectives in ways that fit their strengths and values (Oprea et al., 2019). The effect sizes are small but significant: overall job crafting behavior increased by Hedges' g of 0.26, work engagement by 0.31, and contextual performance by 0.39. The last estimate deserves caution because the confidence interval begins near zero, which means the true effect could be minimal in some samples (Oprea et al., 2019).
The mechanism matters more than the magnitude. The meta-analysis finds that gains in performance are fully mediated by engagement. In plain terms, getting people to craft their roles only shows up in performance if it raises how energized and absorbed they feel at work. That is a design cue. Instead of chasing output directly, leaders should sponsor structured, voluntary experiments that help employees adjust task boundaries, redesign handoffs, and align responsibilities with personal strengths, then track engagement changes as the primary signal. Oprea and colleagues also report a utility analysis indicating potential economic returns for healthcare organizations, though the exact figures vary by context and should not be generalized beyond the studied settings (Oprea et al., 2019).
This is not a silver bullet. The estimated effects are modest, and the interventions studied differ in intensity and duration. But they are implementable within a quarter, and they point in the right direction: agency first, outcomes through engagement.
Make risk-taking safe, or nothing sticks
Discretion without safety is a trap. People will not propose a new workflow, admit a miss, or test a different customer script if speaking up carries career risk. In a field study of 51 work teams inside a manufacturing company, Amy Edmondson found that team psychological safety, defined as a shared belief that the team is safe for interpersonal risk-taking, was associated with higher learning behavior. Team efficacy was not, once psychological safety was accounted for. Learning behavior in turn mediated the link to performance (Edmondson, 1999). The study is cross-sectional and drawn from one firm, so it cannot claim causality across industries, but it gives a clear mechanism. If you want the benefits of empowerment to show up on the bottom line, build the routines that make inquiry normal.
That can be as prosaic as a standing debrief where the first agenda item is a process improvement someone tried, or a weekly check where a frontline team names the riskiest assumption they plan to test next. The point is to make visible, repeated room for fallible humans to surface problems and ideas. Without that climate, job crafting decays into private coping and performance plateaus.
When empowerment becomes structure
There is a limit to how far empowerment-through-programs will take you when the underlying structure remains manager centric. Michael Lee and Amy Edmondson define self-managing organizations as firms that decentralize authority radically, formally, and systematically across the organization, eliminating the subordinate manager relationship as the default (Lee and Edmondson, 2017). Their synthesis of the literature is not experimental, but it draws a sharp line: managerial hierarchy suits reliable execution of known tasks in stable settings. It also tends toward rigidity, reinforces status differences, and inhibits the kind of cross-functional problem solving that complex, nonroutine work requires (Lee and Edmondson, 2017).
The lesson is not that every company should abolish managers. It is that empowerment should be written into the operating system where the work benefits from it, rather than added as a surface program. That means shifting certain decisions from approvals to clear decision rights owned by teams, publishing transparent roles and accountabilities instead of relying on title and tenure, and replacing performance management by attendance with performance defined by outcomes everyone understands. In domains where reliability trumps novelty, keep hierarchy. In domains where the work evolves weekly, move authority closer to the problem.
A practical sequence for 2025
Start with the crisis you already see. Manager engagement has fallen faster than that of their teams, and the global engagement level sits at one in five employees, with a sizable economic cost reported by Gallup’s survey of more than 260,000 respondents (Gallup, 2025). Treat this not as a messaging issue but as a design audit.
- Run a 90-day job crafting sprint. Offer voluntary workshops and coaching that help people identify tasks to expand, reduce, or swap, and relationships to reconfigure. Track engagement before and after, and treat engagement as the leading indicator, consistent with the mediation pattern in the meta-analysis (Oprea et al., 2019).
- Institutionalize psychological safety. Bake explicit learning routines into team cadence. Use debriefs, premortems, and visible error reporting to signal that voice is rewarded. The Edmondson study links safety to learning and learning to performance, even if the evidence comes from a single firm and time slice (Edmondson, 1999).
- Decide where to decentralize. Map the work. Where tasks are repeatable and the environment stable, keep managerial oversight. Where problems cross functions and evolve quickly, formalize team decision rights and reduce layers, in line with the conditions where hierarchies underperform that Lee and Edmondson summarize (Lee and Edmondson, 2017).
None of this removes the need for managers. It clarifies the role. In empowered systems, managers spend less time granting permission and more time setting context, building capability, and stewarding interfaces between teams. The Gallup data on best-practice organizations, where nearly eight in ten managers are engaged, shows what is possible when the job is designed to develop people rather than police them (Gallup, 2025).
The thread through all three levers is agency with guardrails. Let people shape their work, create the psychological conditions for learning, and push authority to where the knowledge lives. Measured in engagement first and performance next, that is what empowerment looks like when it earns its keep.
Sources
- Gallup (2025). State of the Global Workplace 2025. Reports global engagement at 20%, an estimated $10 trillion productivity cost, and a steep drop in manager engagement, highlighting a design gap between average and best-practice organizations. https://www.gallup.com/workplace/659610/state-global-workplace.aspx
- Oprea, Barzin, Vîrgă, Iliescu & Rusu (2019). Effectiveness of job crafting interventions: a meta-analysis and utility analysis. Finds small but significant gains in job crafting, engagement, and contextual performance, with performance effects fully mediated by engagement. https://doi.org/10.1080/1359432X.2019.1646728
- Edmondson (1999). Psychological Safety and Learning Behavior in Work Teams. In a cross-sectional field study of 51 teams, team psychological safety is linked to learning behavior, which mediates its relationship with performance. https://journals.sagepub.com/doi/10.2307/2666999
- Lee & Edmondson (2017). Self-managing organizations: Exploring the limits of less-hierarchical organizing. Defines self-managing organizations and synthesizes evidence that hierarchy suits reliable execution in stable contexts but can inhibit cross-functional problem solving and development. https://www.sciencedirect.com/science/article/abs/pii/S0191308517300059