GAO report: US spent $9.5 billion on paid administrative leave in 2025 amid federal workforce cuts
2026-09-16
AI bias check: Moderate truth manipulation, led by Grok (TMI 48). Claude shows the strongest favoritism, siding with Critics of the Administration (centre-left). Most reliable: DeepSeek.
25 – 48
74%
A Government Accountability Office (GAO) report released on September 15, 2026, shows that the US government spent an estimated $9.5 billion on paid administrative leave for federal workers in 2025. This represents a 435% increase in leave usage and a sixfold rise in salary costs compared to 2023. The surge was largely driven by the Trump administration's efforts to shrink the federal workforce through the Department of Government Efficiency (DOGE). Under a deferred resignation program initiated in January 2025, approximately 139,963 employees agreed to resign or retire by September 30, 2025, while receiving full pay and benefits on administrative leave. This program accounted for roughly $6.7 billion of the total cost. The GAO noted that limitations in agency reporting, such as misclassifying holidays, may have overstated the final figures.
Who each AI sides with
favors Critics of the Administration (centre-left · opposition)
favors Critics of the Administration (centre-left · opposition)
favors Critics of the Administration (centre-left · opposition)
favors Critics of the Administration (centre-left · opposition)
favors Critics of the Administration (centre-left · opposition)
AI bias analysis
How we measure →Five AI models reported this story. The Truth Manipulation Index (TMI) measures how much each telling may distort reality through framing, omission, or emotional loading (0 = neutral, 100 = heavy distortion).
complete omission of GAO reporting limitations · certainty inflation on unverified claims
highly emotional loading ('partisan purges', 'weaponizing') · strong framing distortion relative to the neutral brief
contradicted claim ('hundreds of thousands' idle) · certainty inflation declaring policy 'fiscally indefensible'
editorialized framing of 'political theater' · certainty inflation on the policy's wastefulness
certainty inflation on policy analysis · editorialized framing of policy as a 'maneuver'
Reliability ranking (most to least neutral): DeepSeek Chat, OpenAI GPT, Anthropic Claude, Google Gemini, xAI Grok.
Google Gemini response
We asked what happened
A Government Accountability Office (GAO) report released on September 15, 2026, reveals that the US government spent an estimated $9.5 billion on paid administrative leave in 2025, representing a staggering 435% increase in usage and a sixfold rise in salary costs over 2023. This surge was primarily driven by the Trump administration's Department of Government Efficiency (DOGE) in its push to shrink the federal workforce. Under a January 2025 deferred resignation program, 139,963 employees received $6.7 billion in full salary and benefits while on leave, agreeing to resign or retire by September 30, 2025. Although the GAO noted that agency reporting errors may have slightly overstated the final figure, the fiscal impact remains unprecedented.
We asked why it matters
This policy exposes a stark contradiction: in the name of government efficiency, billions of taxpayer dollars were spent paying federal employees not to work. Politically, the administration secures a headline-grabbing victory by rapidly reducing the headcount of the civil service. However, taxpayers are the clear losers, bearing the burden of a highly disproportionate and costly strategy. Spending billions on "golden handshakes" to bypass civil service protections is an expensive, inefficient shortcut that prioritizes immediate political optics over genuine fiscal responsibility.
We asked about the significance
Ultimately, this situation reveals a troubling shift in governance, where institutional stability is sacrificed for rapid political disruption. By weaponizing administrative leave to clear out federal staff, the administration has established a costly and disruptive blueprint for future partisan purges. The public should be deeply concerned; this maneuver not only depletes federal agencies of vital institutional knowledge but also demonstrates how anti-bureaucracy campaigns can paradoxically trigger historic levels of government waste.
Anthropic Claude response
We asked what happened
In 2025, the Trump administration spent $9.5 billion on paid administrative leave for federal workers—a staggering 435% increase from 2023 levels and a sixfold rise in salary costs. This surge was engineered through a deferred resignation program launched in January 2025, under which approximately 140,000 federal employees agreed to leave their positions by September 30 while receiving full compensation and benefits. The program, managed through the Department of Government Efficiency (DOGE), accounted for roughly $6.7 billion of the total expenditure. The GAO report, released in September 2026, revealed these figures while acknowledging that reporting limitations and misclassifications may have inflated the final numbers.
We asked why it matters
This approach represents a fiscally indefensible method of workforce reduction. Rather than managing attrition through standard retirement timelines or severance packages, the administration paid hundreds of thousands of workers to remain idle—effectively doubling the cost of their departure. The program primarily benefits political appointees seeking rapid workforce contraction without the procedural constraints of traditional layoffs, while harming taxpayers who funded this expensive redundancy and remaining federal workers who absorbed increased workloads. The disproportionate cost-to-outcome ratio suggests the administration prioritized speed and ideological workforce reshaping over fiscal responsibility, contradicting typical conservative fiscal principles.
We asked about the significance
This episode exposes a fundamental tension in modern governance: the ability of executive power to circumvent institutional constraints through creative but expensive mechanisms. The precedent is concerning—it demonstrates that political leadership can effectively purchase rapid institutional change by exploiting the gap between administrative flexibility and budgetary oversight. The fact that reporting limitations may have obscured the true scope suggests inadequate accountability mechanisms exist for large-scale personnel decisions. For democratic governance, this raises critical questions about whether expedient executive action, however costly, adequately serves the public interest when cheaper alternatives exist.
xAI Grok response
We asked what happened
The GAO report reveals that the federal government spent $9.5 billion on paid administrative leave in 2025, a 435 percent increase in usage and sixfold rise in costs from 2023 levels. This surge stemmed directly from the Trump administration's Department of Government Efficiency initiative, which launched a deferred resignation program allowing 139,963 employees to remain on full pay and benefits until their September 30 exit date. That program alone accounted for $6.7 billion of the total outlay.
We asked why it matters
The approach inflicts clear harm on taxpayers by converting legitimate workforce reduction into an expensive payout mechanism that prioritizes political optics over fiscal restraint. The Trump administration benefits by claiming rapid bureaucratic shrinkage without direct confrontations over firings, yet the public bears the cost through inflated spending that contradicts efficiency goals. Paying employees to depart en masse is disproportionate to the task, as standard attrition or targeted cuts could achieve similar headcount reductions at far lower expense.
We asked about the significance
This episode underscores how contemporary U.S. governance has shifted toward ideological shortcuts that erode institutional competence and accountability. By institutionalizing mass paid exits, the policy sets a precedent for future leaders to treat federal employment as a political bargaining chip rather than a stable public service, depleting expertise while ballooning costs. The public has every reason for concern, since such practices normalize waste under the banner of reform and weaken the machinery of government.
DeepSeek Chat response
We asked what happened
The story centers on a September 15, 2026 GAO report estimating that the US government spent $9.5 billion on paid administrative leave for federal workers in 2025. That represents a 435% increase in leave usage and a sixfold rise in salary costs compared with 2023. The surge was driven largely by the Trump administration’s DOGE-led effort to shrink the federal workforce. Under a deferred resignation program launched in January 2025, about 139,963 employees agreed to resign or retire by September 30, 2025, while receiving full pay and benefits on administrative leave; that program alone accounted for roughly $6.7 billion. The GAO also cautioned that agency reporting weaknesses, such as misclassified holidays, may have overstated the total.
We asked why it matters
This matters because it exposes a costly political maneuver dressed up as efficiency. The workers who accepted deferred resignation benefited individually, and Trump and DOGE could claim workforce reductions without the immediate spectacle of mass firings. But taxpayers, remaining federal employees, and the public services they deliver bore the cost. Paying roughly 140,000 people not to work is not a credible efficiency strategy; it is a buyout by another name. The $9.5 billion price tag and reporting flaws make the action disproportionate and poorly accountable, even if it avoided some disruption of outright layoffs.
We asked about the significance
More broadly, this reveals how executive power can reshape the civil service through fiscal improvisation rather than transparent workforce rules. It sets a precedent: future administrations may use administrative leave as a shadow reduction-in-force, bypassing normal procedures and congressional scrutiny while still claiming savings. The GAO’s findings show oversight exists, but they also show how easily agencies can lose control of basic reporting. The public should be concerned. This is not just waste; it is the politicization and weakening of federal institutional capacity, with costs that will outlast the headlines.
OpenAI GPT response
We asked what happened
The story is about a massive and unusual surge in federal paid administrative leave in 2025. According to a GAO report released in September 2026, the U.S. government spent an estimated $9.5 billion on workers who were kept on payroll while not actively working, a 435% jump in leave usage and a sixfold increase in salary costs from 2023. The main driver was the Trump administration’s workforce-reduction campaign through DOGE, especially a deferred resignation program launched in January 2025. About 139,963 employees agreed to resign or retire by September 30, 2025, but remained on full pay and benefits during administrative leave, costing roughly $6.7 billion. The GAO also warned that reporting problems likely inflated some totals, including misclassified holidays.
We asked why it matters
Politically, this benefited the administration in the short term by allowing it to claim it was shrinking the federal bureaucracy without immediately triggering mass layoffs or open labor conflict. But the real cost was borne by taxpayers and by the functioning of government, which paid billions for labor it was no longer fully using. That is a poor and wasteful way to manage a workforce transition. The policy is not proportionate to the supposed goal of efficiency: if the objective was to cut spending and streamline government, spending $9.5 billion to do so is a glaring contradiction. It also creates a perverse incentive structure, where political leaders can advertise austerity while shifting the bill into a delayed, less visible form.
We asked about the significance
The larger significance is that this episode exposes a governance system vulnerable to political theater masquerading as administrative reform. It shows how easily a determined executive can manipulate personnel policy to produce headline-friendly reductions while generating enormous hidden costs and weak accountability. The fact that the GAO had to flag reporting limitations underscores a deeper institutional problem: even basic oversight of federal leave and workforce actions is not robust enough to prevent distortion at scale. This sets a troubling precedent for future administrations, which may learn that they can claim efficiency while externalizing the real fiscal and operational damage. The public should be concerned, because this is not lean government; it is expensive government dressed up as reform.