GAO counted nearly 378,000 separations across 22 major agencies in 2025
Hiring replaced only about one worker for every three departures, reducing institutional capacity across government.
Entities and roles
- Russell VoughtTrump-appointed OMB director and acting CFPB director
- Elon MuskTrump-appointed senior adviser who led DOGE work
- Office of Personnel Managementcentral personnel office supporting government-wide reductions
- Donald J. Trumppresident authorizing the administration action
Relationships in this file
- Elon Muskcoordinated a workforce drive producing mass separationsOffice of Personnel Management
- Donald J. TrumpTrump-appointed senior adviser who led DOGE workElon Musk
- Donald J. TrumpTrump-appointed OMB director and acting CFPB directorRussell Vought
Claims and evidence
- factdocumented
GAO's payroll-data review found nearly 378,000 separations and about 127,000 hires across 22 major agencies during calendar 2025; SBA and USAID data were not included.
- U.S. Government Accountability Office: Federal Agency Workforce Changes: Update for July 2025 to January 2026Passages and source records documenting: GAO counted nearly 378,000 separations across 22 major agencies in 2025
- factdocumented
Elon Musk was the Trump-appointed senior adviser who led DOGE work.
- The White House: President Trump participates in a press conference with departing DOGE adviser Elon MuskAppointment or biography record identifying Elon Musk as Trump-appointed senior adviser who led DOGE work
- factdocumented
Russell Vought was the Trump-appointed OMB director and acting CFPB director.
- United States Senate: Roll Call Vote 37: Confirmation of Russell VoughtAppointment or biography record identifying Russell Vought as Trump-appointed OMB director and acting CFPB director
Measurements
Federal workforce separationsnearly 378,000 people - Basis
- Observed
- Population
- 22 major federal agencies
- Period
- 2025-01-01 to 2025-12-31
- Method
- GAO analysis of federal payroll data.
- Uncertainty
- Rounded count; excludes SBA and USAID because comparable data were unavailable.