When Elon Musk stood at a lectern in front of a mock chainsaw at a February 2025 Conservative Political Action Conference event and promised to cut two trillion dollars from the federal budget, the crowd was ecstatic. The number was arresting. The federal government spends approximately $6.8 trillion per year. Two trillion would represent a cut of nearly thirty percent — the most significant reduction in the size of the American state since the New Deal was constructed. It would, Musk said, be accomplished in a matter of months by a lean team of engineers applying Silicon Valley efficiency to Washington bureaucracy.
Fourteen months later, the Department of Government Efficiency has claimed $150 billion in savings. Independent analysis by the Congressional Budget Office, the Government Accountability Office, and four separate nonpartisan budget institutes finds the actual verified savings to be between $12 and $22 billion — roughly one percent of what was promised. The gap between the promise and the reality is not a rounding error. It is the story.
The Arithmetic of the Promise
The two-trillion-dollar figure was never grounded in a specific plan. When economists at the Penn Wharton Budget Model asked the DOGE team in January 2025 to identify which programmes would be cut to achieve the target, the response was a list of "efficiency improvements" and "waste elimination" that totalled approximately $180 billion in aspirational savings — none of it fully costed, much of it involving programmes that cannot be eliminated without congressional action, and some of it involving contracts that the federal government is legally obligated to honour.
The federal discretionary budget — the portion of spending that can theoretically be reduced by executive action without changing law — totals approximately $1.7 trillion. Non-defence discretionary spending, the portion most accessible to DOGE's cuts, is $900 billion. To cut two trillion from total spending would require slashing the entire non-defence discretionary budget twice over, eliminating all military spending, and still coming up short. The number was never possible. It was a branding exercise.
What Was Actually Cut — And What It Cost
The cuts that did happen fall into several categories, and their true cost-benefit analysis differs significantly from the DOGE dashboard metrics.
Foreign aid was the most visible target. The administration shut down USAID with near-total effect in February 2025, cancelling approximately $60 billion in active contracts and grants. The claimed savings were real in the narrow sense: the contracts were cancelled. But the downstream costs — to the 10,000 USAID staff who were terminated, to the international programmes that provided food security and disease prevention to 50 million people, and to the diplomatic damage of abruptly abandoning partner countries — were real too, and none of them appear on the DOGE dashboard.
The Head Start programme, which provides early childhood education to 750,000 children from low-income families, was cut by 47 percent through a combination of budget impoundments and grant rescissions. The claimed savings: $4.3 billion. The estimated social cost, based on peer-reviewed studies of Head Start's long-term educational and economic effects: $18–24 billion in reduced lifetime earnings, higher special education costs, and increased criminal justice involvement over 30 years. The DOGE methodology does not account for returns on investment.
Don't miss the next investigation.
Get The Editorial's morning briefing — deeply researched stories, no ads, no paywalls, straight to your inbox.
The Human Cost: Casework Files
Behind the budget numbers are the case files of federal agencies in administrative disarray. The Department of Veterans Affairs, which DOGE targeted for a 15 percent workforce reduction, processed 340,000 fewer disability claims in Q4 2025 than in the same quarter of 2024. Average claim processing time rose from 105 days to 147 days. Veterans' service organisations estimate that 45,000 veterans are currently waiting for decisions on appeals that would qualify them for disability payments they are legally entitled to.
The Social Security Administration, operating with 7,000 fewer staff than in 2024, reported average wait times for disability determination hearings of 22 months by February 2026 — a record. An estimated 230,000 pending disability cases are backlogged beyond the 18-month legal maximum. Federal courts in four jurisdictions have issued orders requiring the SSA to process cases faster; the SSA has responded that it lacks the staff to comply.
The Internal Revenue Service — a favourite DOGE target — lost 18,000 staff, primarily in enforcement and audit functions. The direct revenue impact is counterintuitive: the IRS generates approximately $5 in recovered taxes for every $1 spent on enforcement. The CBO estimates that the DOGE-directed IRS cuts will reduce federal revenue by $70–110 billion over ten years — a net fiscal loss that exceeds the claimed savings by a factor of four.
Elon Musk's Conflicts: The Ledger
No serious accounting of DOGE can ignore the conflicts of interest of its founder. Elon Musk, as of January 2025, was simultaneously the head of DOGE, the CEO of Tesla, SpaceX, and X (formerly Twitter), and the owner of companies with approximately $15 billion in active federal contracts. His companies' contracts with the federal government increased by $2.4 billion in the twelve months following his appointment. SpaceX won three additional NASA contracts totalling $1.8 billion. The FAA, which had been a persistent regulatory obstacle to SpaceX's launch programme, approved fifteen new launch licences in the six months after DOGE's establishment — compared to four in the previous equivalent period.
The Office of Government Ethics, which would ordinarily oversee such conflicts, was effectively neutralised: its director was placed on administrative leave in February 2025, and the agency's operating budget was cut by 40 percent. No formal recusal order was ever issued. No disclosure form was ever filed. A lawsuit brought by the Citizens for Ethics and Responsibility in Washington (CREW) demanding conflict-of-interest documentation is proceeding through federal court but has not yet produced a ruling.
The Exit and the Legacy
Musk announced in March 2026 that he would be "scaling back" his DOGE involvement to focus on Tesla, whose stock had fallen 43 percent since his Washington appointment as management distraction and consumer boycotts took their toll. The announcement was characterised in some quarters as a victory for accountability. It is more accurately characterised as the withdrawal of a temporary operator, leaving behind a permanently altered institutional landscape.
The federal workers who were terminated cannot be quickly rehired — institutional knowledge, once lost, takes years to rebuild. The contracts cancelled for NGOs, research institutions, and international partners have in many cases been awarded to alternative providers or simply abandoned. The regulatory agencies hollowed out by DOGE's workforce reductions will operate at reduced capacity for years regardless of any future policy reversal.
What Efficiency Actually Looks Like
The frustrating irony of DOGE is that genuine government efficiency reform is a legitimate and important goal. The federal government does spend money on outdated IT systems, duplicative programmes, and contract management that would make private sector counterparts wince. A serious efficiency effort — one focused on technology modernisation, procurement reform, and programme consolidation — could plausibly save $50–80 billion annually in genuine long-term costs.
The 2022 Inflation Reduction Act allocated $80 billion over ten years to IRS modernisation — the kind of institutional investment that generates measurable returns on investment. DOGE cancelled that programme in January 2025. It is perhaps the most expensive "efficiency" measure in the initiative's history: by cancelling a programme designed to generate $5 for every $1 spent, DOGE sacrificed an estimated $400 billion in revenue for the symbolic satisfaction of cutting the tax collector.
The chaos DOGE leaves behind is not the chaos of disruption — the productive disorder that precedes transformation. It is the chaos of demolition: the dismantling of functional systems without any architecture for what comes next. Trillions were promised. Billions were claimed. Tens of billions in social, economic, and institutional costs were generated. The balance sheet, when it is finally properly drawn up, will not look like efficiency. It will look like the most expensive public management experiment in American history — and one of the least successful.