Every agency eventually works through a period it did not plan for. The cause varies — reorganization, budget uncertainty, hiring constraints, mission realignment, an attrition wave — but the pattern holds. The mission does not pause. The people carrying it absorb the difference.
For anyone deciding where a constrained development dollar goes, the practical question is narrow: which investment protects the agency’s ability to keep delivering while conditions shift? Recent public-sector workforce data point to an answer that is easy to miss, because it sits one layer below where most workforce strategy gets aimed. The answer is the first-line supervisor.
The Finding: The Sector Gap Is Largely a Workplace-Experience Gap
Gallup’s analysis of the 2025 federal workforce changes produced something rare in public-sector workforce research: a clean comparison group. Drawing on Gallup Panel data from 2022 through Q1 2026, Bureau of Labor Statistics employment series, and Office of Personnel Management administrative records, researchers used a difference-in-differences design to compare how federal employees’ experiences moved against state and local government employees over the same window (Gallup, May 2026).
The gaps were real. At the widest point, in Q2 2025, the share of engaged federal workers fell six percentage points more than among state and local peers. Job satisfaction ran roughly 10 to 15 points below state and local peers through the rest of the year. Federal employees were about eight to nine points more likely to report high burnout in Q2 2025, and about eight points more likely to be actively job searching in Q1.
Then Gallup added four workplace-perception measures to the model — whether employees trust leadership, whether they believe their organization cares about their wellbeing, whether they feel connected to their organization’s culture, and whether they feel their workplace treats people with respect. Once those perceptions are accounted for, Gallup reports, the relationship between employee outcomes and federal status “shrinks considerably.”
That is the finding worth sitting with. Being a federal employee during a period of workforce change was not, on its own, what predicted who struggled. How the workplace felt was.
Connected Employees Did Not Crash. Unsupported Ones Did.
Gallup states the mechanism plainly: employees who felt more connected and supported were less likely to experience sharp drops in engagement. The disruption did not land evenly. It sorted people.
The recovery data reinforce the point rather than soften it. Most employee-experience measures rebounded by Q4 2025 and held into Q1 2026, and federal job-search behavior became statistically indistinguishable from state and local peers by Q4. A period of significant change did not produce a uniform or permanent decline. It produced a split — and workplace experience determined which side of that split an employee landed on. The variable in play is not whether turbulence occurs. It is how much delivery capacity survives it.
The Supervisor Decides Which Group an Employee Joins
Trust, connection, respect, and the belief that the organization cares are not conclusions employees reach by reading a memo from headquarters. They form through the person who assigns the work, explains the change, and fields the question no one at the top ever hears.
Gallup’s broader research quantifies that relationship. The manager accounts for 70% of the variance in team engagement. Employees who strongly agree they feel supported by their manager are about 70% less likely to experience burnout. Employees who strongly agree their employer cares about their overall wellbeing are three times as engaged and 71% less likely to report burnout.
Gallup’s own guidance to agencies navigating continued change puts manager training focused on supporting teams through uncertainty alongside clear communication and reinforcing mission and purpose. The supervisor is the point at which an agency’s stated intent either becomes an employee’s daily experience or does not.
The Cost Side: The Buffer Is Under the Most Strain
The layer doing the absorbing is also the layer under the heaviest load. Across the U.S. federal, state, and municipal government workforce, 35% of managers reported burnout, compared with 23% of individual contributors (Gallup, collected quarterly in 2023; 5,410 government respondents; margin of error ±4.0 percentage points).
That twelve-point spread is an operational risk, not a wellbeing footnote. It sits on the one layer the agency relies on to hold everything together. When managers burn out, the buffer thins at exactly the moment it is needed, and entire teams feel it.
The countervailing evidence is equally specific. Gallup’s manager-development research reports that upskilling managers produced 10% to 22% higher engagement among the managers themselves, 8% to 18% higher team engagement, 21% to 28% lower turnover, and 20% to 28% performance improvement relative to peers. For scale on the downside, Gallup estimates that in an agency of 10,000 employees with an average salary of $50,000, low engagement costs roughly $66 million annually.
The Capability Gap Predates Any Particular Disruption
If supervisors are the buffer, the next question is how well prepared they are before turbulence arrives. The most rigorous public-sector measurement available comes from the Public Health Workforce Interests and Needs Survey (PH WINS), and its scope should be stated precisely rather than stretched. PH WINS 2024 surveyed the state and local governmental public health workforce: 56,595 respondents across 48 state health agencies and approximately 1,200 local health departments. It is not a governmentwide federal sample and should not be read as one. Within that population, 28% identified as supervisors, managers, or executives.
The peer-reviewed analysis published in Frontiers in Public Health found that 46% of those supervisors reported their agency did not provide leadership training when they entered their current supervisory role. Another 41% received no organizational training at entry, 27% receive no ongoing leadership training, and 31% receive no ongoing support as a supervisor. Nearly half — 46% — had five or fewer years of supervisory experience before stepping into the role (Popalis et al., 2025).
The downstream effect is measurable in the same dataset. Among non-supervisors, 14% reported dissatisfaction with their supervisor and 18% said their supervisor lacked effective people-management skills. Among those intending to leave within the year, 23% cited dissatisfaction with their supervisor as a contributing factor. Agencies where a larger share of supervisors received training and ongoing support saw higher staff satisfaction, with ongoing support showing the strongest association (adjusted odds ratios 1.14 to 1.38, all p < 0.005). The design is cross-sectional, so these are associations rather than proof of causation.
Public health is one corner of government, and the figure above belongs to that corner. The mechanism it documents is not confined there: technical experts get promoted for technical proficiency, inherit responsibility for people, and receive no preparation for the second job. Every agency can check its own version of that pattern in a single afternoon.
The governmentwide federal equivalent is older than most people assume. The last primary measurement is the Merit Systems Protection Board’s May 2010 study A Call to Action: Improving First-Level Supervision of Federal Employees, which reported from the 2007 Merit Principles Survey that “[l]ess than two-thirds of supervisors said that they received training prior to or during their first year as a supervisor,” and that of those who did, almost half received one week or less. That finding is nineteen years old, and the item was not carried into MSPB’s 2021 Merit Principles Survey. There is no current governmentwide federal number. Any percentage presented as one should be asked to name a study, a year, and a sample — the figures in general circulation name none of the three.
Why This Argument Survives a Change in Administration
Every workforce investment eventually has to justify itself to a decision-maker who did not authorize it. Most people-development spending struggles at that moment, because it was argued on grounds the next decision-maker may not share. The supervisor case does not carry that weakness, for three reasons.
First, it is a continuity-of-operations argument, not a policy argument. Whatever an agency is directed to do, and whoever directs it, the instruction reaches the workforce through supervisors. A supervisor layer that cannot explain a change, hold a team together, and keep output steady degrades the execution of every directive — expansion, consolidation, realignment, or contraction alike. The capability is directionally neutral; it makes whatever comes next work better.
Second, it is a cost argument with a documented denominator. Burnout, disengagement, and turnover are incurred whether or not anyone budgets against them. An agency that declines to fund supervisor capability has not avoided the cost, only moved it off the training line and into recruiting, onboarding, and lost institutional knowledge.
Third, it is measurable. Engagement, burnout, retention intent, and satisfaction with supervision are already instrumented through existing federal workforce surveys. An agency can baseline the supervisor layer, invest, and show movement inside a single budget cycle.
What This Looks Like as a Line Item
The evidence supports a narrow, unglamorous set of investments rather than a broad culture program:
- Train at the point of promotion. The gap opens on day one of the supervisory role. New-supervisor onboarding should cover the basics no one was hired for: setting expectations, running a real one-on-one, resolving conflict, and managing performance.
- Make support ongoing rather than one-time. Ongoing support showed the strongest association with staff satisfaction in the PH WINS analysis. Coaching and peer networks outperform a single course at appointment.
- Equip supervisors specifically for communicating through uncertainty. This is the skill the 2025 data identified as decisive, and it is rarely taught.
- Protect the supervisors themselves. A 35% burnout rate in the layer you are relying on is a capacity problem. Workload review and manager-level support belong in the same investment.
- Measure at the supervisor level. Roll-up agency scores hide the variation that matters. The team is where the buffer either holds or fails.
Bottom Line
Turbulence is not an anomaly to wait out. For federal agencies it is a recurring operating condition, and the evidence is now clear that its effects are not distributed evenly. The distribution is decided close to the employee — by whether the person they report to was prepared to lead them through it.
That is the buffer. Agencies that fund it keep their delivery capacity intact through change. Agencies that do not end up paying for the same disruption twice: once when it happens, and again when the people who absorbed it walk out the door.
Sources
Makridis, C., & Lloyd, C. (2026, May 13). “What 2025 Federal Reforms Did to Worker Engagement.” Gallup. Engagement, job satisfaction, burnout, and job-search gaps between federal and state/local workers; the workplace-perception mediation finding. https://www.gallup.com/workplace/708866/2025-federal-reforms-worker-engagement.aspx
Petrucelli, L., & Ritz, M. (2024, March 24). “What Is Driving Federal Government Burnout?” Gallup. Manager burnout (35%) vs. individual contributor burnout (23%) among U.S. federal, state, and municipal government workers; data collected quarterly in 2023, n = 5,410, ±4.0 percentage points. Also the source for the manager-upskilling outcomes, the 70% engagement-variance figure, the 70% burnout reduction associated with manager support, and the $66 million low-engagement cost estimate. https://www.gallup.com/workplace/612518/driving-federal-government-burnout.aspx
Popalis, M., Leider, J. P., Mason, A., Najjar, M., Robins, M., & Castrucci, B. (2025). “Bridging the leadership gap: agency training and support shapes non-supervisors’ perceptions of their leaders.” Frontiers in Public Health, 13:1610400. doi:10.3389/fpubh.2025.1610400. Peer-reviewed analysis of PH WINS 2024 (56,595 respondents; 48 state health agencies and approximately 1,200 local health departments). SCOPE NOTE: state and local governmental public health workforce only — not a governmentwide federal sample. https://doi.org/10.3389/fpubh.2025.1610400
de Beaumont Foundation & Association of State and Territorial Health Officials. Public Health Workforce Interests and Needs Survey (PH WINS) 2024 — survey background, methodology, and findings. https://debeaumont.org/phwins/