The Willing Audience

DOGE promised two trillion dollars in savings, produced the largest peacetime workforce reduction in American history, and left the deficit larger than it found it. The audience applauded anyway.

The Headline They’re Selling You

There is a particular kind of political theater that requires no actors, no stage, and no script, only a number, stripped of its context, handed to a compliant press, and repeated until it becomes true by sheer exhaustion of the audience. On March 6, 2026, that number was 4.4 percent. Unemployment, the administration's allies announced with the serenity of men who have never once read past the first paragraph, was holding steady. What was also holding steady, buried in the same document they were celebrating, was the fact that the American economy had just lost 92,000 jobs in a single month (U.S. Bureau of Labor Statistics, 2026).

That number deserves a moment of undivided attention before the spin cycle begins. Ninety-two thousand jobs. Gone. In a single month. And yet the unemployment rate did not move, which, far from being reassuring, is itself a story worth telling. An unemployment rate can hold steady while jobs vanish for a simple and rather grim reason: people who stop looking for work are no longer counted as unemployed. The number being broadcast was technically accurate and functionally misleading, which is, one might note, a particular kind of dishonesty, the kind that comes dressed in a suit and carries a press release.

It gets worse when you read past the first paragraph, as most people are not encouraged to do. The December jobs report, which had previously been presented to the public as a modest gain of 48,000 jobs, was quietly revised to a loss of 17,000 (U.S. Bureau of Labor Statistics, 2026). A month that the public was told was productive was, in fact, a month in which tens of thousands of Americans lost their livelihoods. The January figure was revised downward as well. Combined, the revisions erased 69,000 more jobs than anyone had been told existed. There was no correction issued on cable news. There was no retraction from the accounts that had cited the original numbers as proof of a thriving economy. The revision appeared in a footnote, as these things tend to do.

Beneath even the revised numbers, a slower and more corrosive trend has been accumulating. The number of long-term unemployed, those out of work for 27 weeks or more, has risen from 1.5 million a year ago to 1.9 million today (U.S. Bureau of Labor Statistics, 2026). These are not people between jobs. These are people for whom the ordinary machinery of the labor market has simply stopped working. They are, by the accounting conventions of official statistics, still counted. But they are the kind of number that does not appear in a triumphant social media post, and so they are permitted to accumulate in silence while the headline rate holds obligingly steady.

The gap between what the document says and what was reported about the document is not a matter of interpretation. It is a matter of reading.

The Federal Workforce and the Chainsaw

Elon Musk arrived in Washington in early 2025 holding a literal chainsaw at a conservative conference, and the image was, one must admit, at least honest in its implications. Chainsaws are not precision instruments. They do not discriminate. They do not pause to ask whether the thing being cut is load-bearing. And when the Department of Government Efficiency finished its first year of operation, the federal workforce was down 330,000 positions from where it stood in October 2024, the largest peacetime workforce reduction in the recorded history of the United States (Cato Institute, 2026).

The promise that justified this was, at first, breathtaking in its ambition. Musk announced that DOGE would deliver at least two trillion dollars in annual savings. Then, as the fiscal reality of governing an actual country began to assert itself, that figure became one trillion. Then, with the quiet embarrassment of a man revising his bet after the horses have already finished, it became one hundred and fifty billion (CNN, 2025). The goalpost did not merely move. It was picked up, carried to another zip code, and planted in softer ground.

Even at that drastically reduced figure, the arithmetic was never going to work, and the reason why is not complicated. Federal personnel costs represent approximately four percent of total government spending (NPR, 2025). The other ninety-six percent, Social Security, Medicare, Medicaid, defense, and the interest payments on a national debt that crossed thirty-seven trillion dollars, DOGE never touched, never seriously proposed touching, and in several cases actively grew. The Treasury Department’s own data showed that from October 2024 through August 2025, the federal government spent six point six six trillion dollars, an increase of three hundred and seventy-six billion over the same period the prior year (NPR, 2025). The deficit did not shrink. It grew by nearly two trillion dollars over that same stretch. A person who did not know when DOGE had started its work, looking only at the spending trajectory, could not have identified the moment of its intervention. This is not a partisan observation. It is the conclusion of the Cato Institute, which stated plainly that DOGE had no noticeable effect on the trajectory of spending, but did succeed in engineering the largest peacetime workforce reduction on record (Cato Institute, 2026).

The savings claims themselves did not survive contact with scrutiny. DOGE maintained what it called a wall of receipts, a public accounting of its efficiency victories, which upon examination contained errors so basic as to suggest either incompetence or indifference to the difference between the two. One entry claimed four point three million dollars in savings from canceling a contract with the Federal Aviation Administration. The contract in question was not canceled. It was worth approximately one hundred and fifty thousand dollars. And nearly all of the money had already been spent before DOGE claimed credit for saving it (NPR, 2025). A professor at Northwestern’s Kellogg School of Management estimated that DOGE’s claimed savings figures were overstated by as much as eighty percent (Fortune, 2025). The White House’s response to these findings was to accuse the people reporting them of peddling falsehoods, which is, historically speaking, not the same thing as a rebuttal.

What DOGE was doing, though it was never described in these terms, was applying the logic of a private sector cost-cutting exercise to an institution that does not function like a private sector company, for the simple reason that it is not one. When Musk gutted Twitter’s workforce in 2022, the consequences were absorbed by shareholders and users. When the federal government loses a third of the IRS, the consequences are absorbed by the public treasury. When it fires the team that administers heating assistance for low-income families in the middle of winter, the consequences are absorbed by the families. The chainsaw, it turns out, does not know the difference.

They Fired the People Who Collect the Money

If one were tasked with designing a policy guaranteed to increase the deficit while claiming to reduce it, a reasonable starting point would be to fire the people whose job it is to collect the government’s revenue. This is not a hypothetical. In 2025, the Trump administration, operating under the banner of efficiency, oversaw the departure of a significant portion of the Internal Revenue Service’s hundred-thousand-person workforce, with estimates suggesting that up to a third of its employees would ultimately leave (The Global Treasurer, 2025). The Yale Budget Lab ran the numbers on what this would mean in practice. Their conclusion was not ambiguous: the IRS staff reductions were projected to produce an eight point five billion dollar reduction in collected tax revenue in 2026 alone, with that figure compounding to nearly one hundred and ninety-eight billion dollars over the following decade (The Global Treasurer, 2025). As Richard Prisinzano of the Yale Budget Lab put it, for every dollar spent on IRS personnel, a considerably larger sum comes back in (The Global Treasurer, 2025). Firing the revenue collectors does not save money. It destroys it. The administration fired the people who collect the money and then announced it had found savings.

The IRS was not an isolated case. It was a pattern. Across the federal government, agencies discovered, with the particular embarrassment of institutions that have just publicly humiliated themselves, that the workers they had fired were in fact necessary to the basic operation of the government they were supposedly streamlining. The Department of Health and Human Services eliminated the entire team responsible for administering the Low Income Home Energy Assistance Program, a program that distributes over four hundred million dollars to states to help low-income families pay heating and cooling bills. The agency was then forced to rehire a single longtime employee for two and a half weeks, specifically to run a critical formula required to distribute that funding to states that had already exhausted their prior appropriations (CNN, 2025). The team had been fired. The money still needed to move. One person was brought back, temporarily, to do what the entire team had been doing, and then presumably let go again.

The Food and Drug Administration fired staff and then rehired more than a dozen scientists at a food safety laboratory in Illinois when it became apparent that food safety work requires, at minimum, people employed to perform it (CNN, 2025). The Department of Agriculture halted plans to cut a quarter of its staff at fifty-eight facilities responsible for monitoring and responding to bird flu outbreaks, the same outbreak that had been driving egg prices to historic highs, when the connection between firing disease monitors and unchecked disease spread became difficult to ignore (CNN, 2025). Nuclear engineers at the National Nuclear Security Administration were fired and then rehired, because it emerged, with the slowness of a realization that should have arrived considerably earlier, that nuclear security is not a function that tolerates staffing gaps (CNN, 2025). The Federal Aviation Administration was specifically exempted from the deepest cuts after it became clear that understaffed air traffic control is not an abstraction but a condition with immediate and measurable consequences for anyone aboard a commercial aircraft.

What this produced was not efficiency. It was a cycle of destruction and emergency repair, each iteration of which cost money, burned institutional knowledge that cannot be easily replaced, and left the remaining workforce in a state of sustained uncertainty that is, by any serious measure, the opposite of productive. The Partnership for Public Service estimated that the chaos of firings, rehirings, paid administrative leaves, and productivity losses could cost the government approximately one hundred and thirty-five billion dollars in this fiscal year alone, a figure that, with considerable irony, approaches the total savings DOGE claimed to have achieved in its entire existence (Fortune, 2025). The administration had set out to cut waste. It had, with impressive thoroughness, manufactured it.

“They’ll Find Better Jobs”

The Office of Personnel Management, the federal agency responsible for managing the very workforce being dismantled around it, offered the following reassurance to the tens of thousands of workers it was in the process of firing: the private sector was waiting, the jobs would be better, and the whole arrangement was really an opportunity in disguise. The move, OPM explained, was about shifting people from lower productivity jobs in the public sector to higher productivity jobs in the private sector (CNN, 2025). It was the kind of statement that could only be written by someone who had not recently attempted to find a white-collar job in the private sector, or spoken to anyone who had, or consulted the labor market data that their own government publishes.

The private sector was not waiting. Indeed, the job search platform, reported a fifty percent surge in applications from federal workers in a single month following the onset of DOGE’s cuts, a figure its own economist described as something she had simply never seen before (PBS NewsHour, 2025). That surge did not reflect opportunity. It reflected panic. Approximately sixty-eight percent of the federal workers flooding the job market in that period held at least a bachelor’s degree, entering a market where demand for knowledge work and white-collar professional roles had already been contracting before a single federal employee had been shown the door (Newsweek, 2025). Indeed’s economist was direct about what this meant: absorbing a highly educated, highly specialized workforce into an economy where those jobs were already scarce was going to be, to use her precise formulation, brutally competitive (Newsweek, 2025).

The skill mismatch problem was not a minor inconvenience to be sorted out with some résumé workshopping. Many federal jobs exist in a category that has no meaningful private sector equivalent. Regulatory specialists, nuclear technicians, public land managers, foreign service officers, disease surveillance scientists, these are not interchangeable with the roles available at companies that exist to generate profit rather than to perform the functions of a sovereign government. A biological science technician fired from Indiana Dunes National Park, who had accumulated a master’s degree, urban forestry training, and wildfire deployment experience, found himself applying to an Amazon delivery route and a line cook position at a fried chicken restaurant (PBS NewsHour, 2025). This was not a transition to higher productivity. It was a credential bonfire.

The age dimension compounded everything. The federal workforce skews older than the general workforce, with roughly twenty-eight percent of federal employees aged fifty-five or above, compared to just under twenty-four percent of the broader labor market (CNN, 2025). Older workers face well-documented disadvantages in job searches under ordinary conditions. These were not ordinary conditions. They were conditions in which hundreds of thousands of similarly credentialed, similarly aged, similarly specialized workers were competing for the same shrinking pool of openings simultaneously, in a market that a senior labor economist at ZipRecruiter described as one that had seen remarkably little hiring even before the federal flood arrived (CNN, 2025).

The Congressional Budget Office had, in fact, already done the analytical work that made the OPM’s promise look not merely optimistic but precisely inverted. Federal workers with professional degrees or doctorates were already earning twenty-nine percent less than their private sector counterparts, meaning the government’s most credentialed employees had accepted a wage penalty in exchange for stability, mission-driven work, and benefits (PBS NewsHour, 2025). The stability was now gone. The mission-driven work was now gone. What remained was a job market that did not particularly want them, offering salaries many of them had already chosen once to decline. A former budget analyst at the Department of Veterans Affairs, while working shifts at a mall clothing store to cover her expenses, was still waiting for a position comparable to the one she had lost (PBS NewsHour, 2025). The Office of Personnel Management did not follow up.

The Willing Audience

It would be convenient, and considerably more comfortable, to explain the persistence of the DOGE efficiency narrative as a simple matter of dishonesty, politicians lying, audiences nodding, the familiar choreography of modern political theater. But the research suggests something more unsettling than a lie that is knowingly swallowed. It suggests a lie that is genuinely believed, and believed not despite the contradicting evidence but in many cases because of it. The psychological literature on motivated reasoning does not describe a passive audience being deceived. It describes an active audience doing the work of deception on its own behalf.

The mechanism has been documented with considerable precision. Research on partisan motivated reasoning identifies three interlocking processes through which politically engaged people evaluate information (Vail et al., 2023). The first is congruence bias, the tendency to seek out, accept, and internalize information that confirms what one already believes. The second is disconfirmation bias, the active rejection, criticism, and counterarguing of information that challenges existing beliefs. The third, and most counterintuitive, is the polarization effect: the combination of the first two means that when a partisan encounters information that contradicts their position, they do not update their belief. They become more confident in it (Vail et al., 2023). Present a Trump supporter with the revised December jobs number, the one that turned a reported gain of 48,000 into an actual loss of 17,000, and the research predicts not a moment of pause but a hardening. The correction becomes evidence of media bias. The footnote becomes proof of a conspiracy. The mechanism does not require bad faith. It only requires identity.

And identity, in this context, is not a casual thing. Research tracking Trump supporters across three years found that those whose personal identities had become deeply fused with Trump were significantly more likely to internalize his claims as their own beliefs, independent of whether those claims were factually supportable (Moniz & Swann, 2024). More troubling still, accepting one piece of misinformation did not merely satisfy the belief; it primed the acceptance of the next. Convincing supporters of one false narrative made them measurably more receptive to subsequent ones, with each accepted claim deepening the fusion and narrowing the aperture through which new information could enter (Moniz & Swann, 2024). The 4.4 percent unemployment headline is not, for this audience, a partial truth being strategically deployed. It is a complete truth, because the architecture of their information processing was constructed to receive it as one.

This is what makes the gap between the headline and the document something more than a communications problem. When an administration announces that unemployment held steady at 4.4 percent while 92,000 jobs vanished, while December’s reported gains evaporated into actual losses, while long-term unemployment climbed quietly toward 1.9 million, it is not simply hoping the audience won’t read the footnotes. It is operating with a sophisticated understanding, whether conscious or instinctive, that a meaningful portion of its audience has been psychologically prepared to treat the headline as the whole story and to experience any challenge to that headline as an attack on themselves. The partial truth is not an oversight. It is a delivery mechanism, and the willing audience is not a passive recipient. It is a co-producer of the narrative, doing the cognitive labor of completion that the administration does not need to perform itself.

What This Actually Is

Let us, for the sake of generosity, assume that everyone involved believed what they were saying. That the chainsaw was wielded in good faith. That the wall of receipts was maintained with honest intentions. That the Office of Personnel Management genuinely believed the budget analyst at the VA would land somewhere better. Assume all of it, and the conclusion does not change by a single decimal point. The deficit grew. Spending went up. The workers did not find better jobs. The agencies rehired the people they fired. And the Cato Institute, an organization that has spent decades arguing for smaller government with considerably more rigor than a man carrying a chainsaw at a conservative conference, concluded that DOGE had no noticeable effect on the trajectory of federal spending, but did succeed in producing the largest peacetime workforce reduction in the recorded history of the United States (Cato Institute, 2026).

That last part is worth sitting with. The largest peacetime workforce reduction in American history. Not a trimming. Not a restructuring. A historic gutting of the civil service, accomplished without a corresponding reduction in what the government spends, without a measurable improvement in what the government delivers, and without the private sector absorption story that was used to justify the human cost of doing it. What was produced, in the assessment of the Partnership for Public Service, was approximately one hundred and thirty-five billion dollars in potential losses from disrupted productivity and lost revenue collection, a figure that approaches, with considerable irony, the total savings DOGE ultimately claimed after revising its ambitions downward from two trillion dollars to one hundred and fifty billion (Fortune, 2025). The administration set out to save money and may well have spent more than it saved in the act of trying.

The long-term unemployment figure that appeared without fanfare in the March 6 report, 1.9 million Americans out of work for six months or more, up from 1.5 million a year prior, does not exist in a vacuum (U.S. Bureau of Labor Statistics, 2026). It exists alongside a federal job market that shed 330,000 positions, a private sector that was not equipped to absorb the overflow, and a white-collar hiring environment that economists described as frozen by uncertainty even before the federal flood arrived. The number being broadcast was 4.4 percent. The number accumulating beneath it tells a different story, and it will continue accumulating long after the press releases have been recycled.

There is a word for the practice of firing the people who collect revenue, watching revenue collection decline, and declaring the exercise a success. There is a word for revising your savings estimate from two trillion dollars to one hundred and fifty billion while the deficit grows by nearly two trillion in the same period. There is a word for rehiring, at additional cost and with destroyed institutional knowledge, the workers you fired three months earlier because the essential functions of government do not pause for ideological theater. The word is not efficiency. It has never been efficiency. What it is, stripped of the language of reform and the imagery of chainsaws and the promise of a better job that did not materialize, is the public absorbing the cost of a performance staged for an audience that was never meant to read the footnotes.

The Bureau of Labor Statistics published the footnotes on March 6, 2026. They were available to anyone willing to read past the first paragraph. Most people were not encouraged to do so, and that, perhaps more than anything else in this document, is the point.

References

Cato Institute. (2026, January 6). DOGE produced the largest peacetime workforce cut on record, but spending kept rising. Cato at Liberty Blog. https://www.cato.org/blog/doge-produced-largest-peacetime-workforce-cut-record-spending-kept-rising-0

CNN. (2025, February 6). Attention federal employees: It may be hard to get a private sector job right now. https://www.cnn.com/2025/02/06/business/private-sector-job-market/index.html

CNN. (2025, June 24). Trump administration scrambles to rehire key federal workers after DOGE firings. https://www.cnn.com/2025/06/24/politics/doge-fired-workers-rehired

CNN. (2025, December 24). Elon Musk’s DOGE tally: The federal workforce is down while government spending is up. https://finance.yahoo.com/news/elon-musks-doge-tally-the-federal-workforce-is-down-while-government-spending-is-up-192850019.html

Fortune. (2025, April 28). DOGE’s mass federal workforce cuts may cost taxpayers $135 billion this fiscal year alone. https://fortune.com/article/doge-mass-federal-workforce-cuts-taxpayers-billions/

Moniz, Philip & Swann, William. (2025). The Power of Trump’s Big Lie: Identity Fusion, Internalizing Misinformation, and Support for Trump. PS: Political Science & Politics. 58. 1-6. 10.1017/S1049096524001203.

Newsweek. (2025, April 8). Fired federal workers flood “brutally competitive” job market. https://www.newsweek.com/fired-federal-workers-flood-brutally-competitive-job-market-2055185

Newsweek. (2025, June 19). Fired federal workers struggle to get new jobs in private sector. https://www.newsweek.com/fired-federal-workers-employees-jobs-market-2087217

NPR. (2025, March 6). DOGE wants to cut $1 trillion this year. But it’s not looking at big spending drivers. https://www.npr.org/2025/03/06/nx-s1-5318072/how-much-money-has-doge-saved-budget-deficit-congress

NPR. (2025, October 1). DOGE cuts to federal government staffing and spending are being undone. https://www.npr.org/2025/10/01/nx-s1-5558298/doge-fiscal-year-savings-budget-rehired-government-shutdown

PBS NewsHour. (2025, April 4). Fired federal workers struggling to land new jobs in tightening white-collar sector. https://www.pbs.org/newshour/show/fired-federal-workers-struggling-to-land-new-jobs-in-tightening-white-collar-sector

The Global Treasurer. (2025, April 28). Are DOGE’s cost-cutting measures actually costing taxpayers billions? https://www.theglobaltreasurer.com/2025/04/28/doge-federal-workforce-cuts-cost/

U.S. Bureau of Labor Statistics. (2026, March 6). The employment situation — February 2026 (USDL-26-0367). U.S. Department of Labor. https://www.bls.gov/news.release/empsit.nr0.htm

Vail, K. E. III, Harvell-Bowman, L., Lockett, M., Pyszczynski, T., & Gilmore, G. (2023). Motivated reasoning: Election integrity beliefs, outcome acceptance, and polarization before, during, and after the 2020 U.S. Presidential Election. Motivation and Emotion, 47, 177–192. https://doi.org/10.1007/s11031-022-09983-w