Government Employee Retention Strategies: What Actually Works at GS-Scale Pay

Federal voluntary attrition — the OPM measure that combines quits and transfers to other federal jobs — ran 11.4% in FY 2023 (FedScope). That's elevated from the ~9% pre-pandemic baseline but well below the private-sector quit rate (~25% annualized in BLS JOLTS). The harder story is the retirement cliff: roughly 30% of federal employees are eligible to retire within five years, with concentrations in Treasury, Interior, SSA, and across technical and leadership grades. Several CFO Act agencies face 40%+ retirement exposure in their GS-14/15 ranks — the cohort that runs programs, supervises, and carries institutional memory.

State and local has a different shape. Mission Square Research Institute's 2023 data, anchored on BLS JOLTS, shows local government quits hit 6.9% in 2023 — a record since the sector began being tracked separately. The driver is the public-private pay gap. BLS Employer Costs for Employee Compensation (ECEC) data shows state and local total compensation now trails private-sector by ~15% in many metros — a gap that widened sharply post-2021.

The Partnership for Public Service's annual Best Places to Work rankings, derived from FEVS, show 20+ point engagement gaps between top- and bottom-ranked agencies. Top performers (NASA, FDIC, GAO) consistently hold talent. Bottom performers lose it. The variable is not pay — it's leadership practice at the work-unit level.

Five drivers show up consistently across FEVS open-text analysis, MSPB exit-survey work, and Partnership for Public Service interviews:

• Supervisor quality at the work-unit level. FEVS data consistently shows the largest variance in engagement is at the work-unit level, not the agency level. A great supervisor in a struggling agency retains; a poor supervisor in a strong agency loses staff.
• Recognition lag and Title 5 friction. Monetary awards are heavily constrained (5 U.S.C. § 4503) and slow. When non-monetary recognition is absent too, federal employees go years between meaningful acknowledgement — even on high-performing teams.
• Mission erosion. Federal employees consistently rank mission alignment above pay in FEVS data. When the work feels disconnected from outcomes — through endless process, OMB reviews, or churning priorities — engaged employees leave for organizations where the connection is shorter.
• Telework and flexibility friction. Post-2022, mandatory-return-to-office orders correlate with elevated quit intent on FEVS items. The cohorts most likely to leave (mid-career technical staff) are also the most flexibility-sensitive.
• Career-path opacity. Especially for new federal hires (Pathways, recent-grad). The federal HR system's emphasis on tenure and grade rather than visible career mobility loses talent to private-sector ladders that promise clearer growth.

Three patterns show up repeatedly in agencies whose retention numbers don't improve:

• The annual FEVS rollout slide deck. FEVS results presented to SES at a quarterly review, but never debriefed at the work-unit level where the variance actually lives. Staff watch the cycle repeat and disengage from the instrument itself — response rates trend down, results become less actionable, and the cycle continues.
• Generic appreciation events. Federal employee appreciation week with cake in the breakroom does not address recognition lag, supervisor quality, or workload. It signals concern without producing change, which top FEVS open-text scores as worse than doing nothing.
• Special-rate adjustments without parallel work on culture. Special rates and other authorities can patch acute pay gaps in specific job series (IT, cybersecurity, medical), but they don't move the engagement gap. Agencies that lean only on pay authorities and not on leadership development see attrition recover within 12–18 months.

Across the Best Places to Work top-quartile data and several agency-internal studies, four interventions show up disproportionately:

1. Structured stay interviews with retirement-eligible technical staff A 30-minute one-on-one between a supervisor and a retirement-eligible employee, twice a year, focused on three questions: what would extend your tenure, what knowledge transfer matters to you, and what work would you say yes to that you currently say no to. Agencies running this systematically extend median tenure of retirement-eligible technical staff by 12–24 months — and turn the conversation into a succession plan, not just a retention conversation.

2. Work-unit FEVS action within 60 days The single largest predictor of next-year FEVS scores is whether the current year's results triggered visible action at the work-unit level inside 60 days. Top-quartile agencies (NASA, FDIC, GAO) publish 'you said / we did' summaries from the work-unit supervisor to the work-unit team. Bottom-quartile agencies brief results only at the SES level. The behavior change is cheap; the institutional discipline to do it consistently is not.

3. Peer non-monetary recognition at high volume Non-monetary peer recognition sits outside Title 5 award constraints and can be deployed agency-wide without budget impact. FEVS recognition-item scores move when peer recognition becomes part of the daily rhythm — not when annual award ceremonies expand. Agencies that pair this with mobile delivery to reach field staff see the largest movement (typically +5 to +8 points on FEVS recognition items in one cycle).

4. Telework and flexibility within statutory limits FEVS post-2022 shows the work-life-balance item is now one of the strongest predictors of intent-to-stay among federal employees. Agencies that codify telework arrangements clearly (rather than leaving them to supervisor discretion) see less mid-career attrition. State and local: results-only work arrangements where mission allows, and predictable schedules where it does not (sanitation, 911 dispatch, inspectors).

Roughly 30% of the federal workforce is retirement-eligible within five years. Treating that as a retention problem alone misses the bigger lever. The employees who leave take 20–35 years of program-specific institutional knowledge with them — knowledge that the federal hiring process cannot quickly backfill given typical 90–180 day time-to-hire.

The agencies handling this best layer two things on top of retention work:

• Formal mentorship windows. GS-14/15 technical staff get assigned mentees in their final 2–3 years, with mentorship hours protected in the work plan. This pairs naturally with stay interviews — many retirement-eligible staff will stay an extra year if formal mentorship is part of the role.
• Knowledge-capture rituals. Recorded 'how this program actually works' sessions during the final 12 months. Stored in the agency's knowledge base, not a personal hard drive. Several agencies (NASA being the most-cited example) have made this a leadership expectation, not an HR initiative.

The failure mode is to wait. By the time an experienced GS-15 has filed their retirement notice, you have ~60 days. The window to capture institutional knowledge opens five years earlier, when they first become eligible.

Two metrics matter, two don't.

Track: - Voluntary attrition by tenure cohort and grade. Tenure cohorts (0–2yr, 2–10yr, 10+yr) and grade bands (GS-7-12, GS-13-15, SES) tell you whether the problem is onboarding, mid-career, or leadership. Each requires a different intervention. - Work-unit FEVS movement on the close-the-loop items ('I feel my supervisor listens to what I have to say,' 'Senior leaders maintain high standards of honesty and integrity'). These items move when action happens at the unit level — and predict attrition trends 12 months ahead.

Don't over-index on: - Agency-aggregate FEVS scores in isolation. The within-agency variance between work units is usually larger than the between-agency variance. The aggregate hides the action. - Special-rate retention. It's a short-term lever that doesn't translate into culture change. Useful for acute IT/cyber/medical gaps; misleading as a general retention indicator.