“We will make mistakes,” Elon Musk said at a press conference a month ago. “Some of the things that I say will be incorrect and should be corrected.”
The billionaire entrepreneur, who is unofficially in charge of the federal cost-cutting initiative known as the Department of Government Efficiency (DOGE), was certainly right about that. News outlets have repeatedly identified embarrassing and consequential mistakes in DOGE’s data on purported spending cuts, including contracts that had not been awarded yet, contracts that were not actually canceled, contracts that were terminated before Trump took office, contracts that were counted multiple times, conflation of contract caps with actual spending, the inclusion of past spending in estimates of future savings, and overvaluation of contracts, such as the notorious data entry error that transformed an $8 million Immigration and Customs Enforcement contract into an $8 billion cut.
While conceding his fallibility, Musk promised to be “as transparent as possible.” Toward that end, he said, “we post our actions” on the DOGE website and X account, striving to be “maximally transparent.” That characterization of DOGE’s activities has proven to be less accurate.
Contrary to Musk’s promise, Reason‘s C.J. Ciaramella notes, the Trump administration tried to shield his project from Freedom of Information Act (FOIA) requests by moving the U.S. DOGE Service (USDS), formerly the U.S. Digital Service, from the Office of Management and Budget, which is bound by FOIA, to a separate slot within the Executive Office of the President. As Ciaramella reported, a federal judge rejected that dodge on Monday, saying “USDS is likely covered by FOIA” and therefore must comply with public requests for information about its work.
So far, DOGE has withheld most of that information. Although the website that Musk touted currently claims $115 billion in “estimated savings,” the details of that calculation remain mysterious.
The website says the total includes “asset sales, contract/lease cancellations and renegotiations, fraud and improper payment deletion, grant cancellations, interest savings, programmatic changes, regulatory savings, and workforce reductions.” But the information posted on the site relates to just a couple of those categories, listing “contract terminations,” “lease terminations,” and “grant terminations.” Together, DOGE’s numbers indicate, these account for one-third of its total “estimated savings,” and there is ample reason to be skeptical even of that part.
In some cases, DOGE has revised or deleted erroneous line items after journalists pointed out its mistakes. But DOGE recently made that corrective process more difficult by omitting federal identification numbers from its latest batch of canceled grants, which was posted on March 2. Those numbers, The New York Times reports, could initially be found in the corresponding source code, but DOGE “deleted this identifying information from the code later in the week.”
A White House official who “asked not to be named” said DOGE is withholding the information “for security purposes.” But without the identifying grant numbers, news organizations cannot verify the savings DOGE is claiming. If DOGE “is now going to fill its site with uncheckable claims,” the Times notes, the “Wall of Receipts” on the site “loses its value.”
DOGE’s list of savings from 5,356 “contract terminations,” which it says total about $20 billion, still includes links to information about the underlying contracts. That information has allowed journalists and analysts to expose errors and exaggerations. The list includes, for example, a $1.9 billion IRS tech support contract with Centennial Technologies that was terminated during the Biden administration. After the Times noted that mistake, the contract disappeared from DOGE’s list, but now it is back.
On February 19, when DOGE was claiming $16.5 billion in contract savings, NPR found that the actual number, based on confirmed cancellations, was about $2 billion—88 percent less. DOGE’s hyperbole was so pervasive that Manhattan Institute budget expert Jessica Riedl, in a February 28 interview with New York Times columnist David French, described its work as “government spending-cut theater,” saying “most of what is claimed to be spending cuts are just accounting errors.”
DOGE also lists savings from 793 “lease terminations,” which it says total about $500 million. Each item includes the leasing agency, the location and size of the space, the annual lease cost, and the “total savings.”
Unlike DOGE’s list of purported contract savings, its list of 7,488 “grant terminations,” which it says total about $17 billion, does not include information beyond the source agency (e.g., “Department of Education”), the “total contract” amount, the claimed savings, and the date the line item was “uploaded.” But the Times, which copied some of the relevant source code before the identifying numbers were deleted, found mistakes that suggest the grant list suffers from the same problems as the contract list.
“At least five of the 20 largest ‘savings’ appeared to be exaggerated, according to federal data and interviews with the nonprofits whose grants were on the list,” the Times says. In one case, DOGE claimed it had saved $1.75 billion by canceling a grant from the U.S. Agency for International Development (USAID) to “a public-health nonprofit called Gavi, the Vaccine Alliance.” But “the grant had not been terminated,” and USAID, in any case, “had already paid out all the money it owed.” In other words, “even if the grant had been terminated, the savings would have been $0.”
Needless to say, this is not what maximal transparency looks like. And DOGE’s error-riddled lists, which add up to $37.5 billion in claimed savings, leave out two-thirds of the $115 billion in total “estimated savings.” That means most of that calculation is impossible to check.
Part of the remaining $77.5 billion, DOGE says, comes from “workforce reductions.” That makes sense in light of the Trump administration’s efforts to reduce the number of federal employees through layoffs and severance offers. But it’s not clear how much DOGE claims to have saved in this category.
DOGE also says it has saved money by attacking fraud and improper payments. That is a potentially worthwhile effort, since the Government Accountability Office has estimated that “the federal government could lose between $233 billion and $521 billion annually to fraud.” But it is clearly not true that DOGE already has “found hundreds of billions of dollars of fraud,” as President Donald Trump claimed in his speech to Congress last week.
How much fraud has DOGE actually identified? All we know is that it must be considerably less than $77.5 billion, the unitemized amount of claimed savings.
DOGE also counts “asset sales” as part of its “estimated savings.” Again, the amount is unclear, and this characterization is misleading in two ways: Asset sales are a source of revenue rather than savings, and any given sale is a one-time event, meaning it cannot help bridge the gap between spending and revenue beyond a single fiscal year.
There is a more serious conceptual problem with including “regulatory savings.” While lifting unnecessary burdens on businesses would certainly be welcome, the resulting private savings would not amount to a federal spending cut. We don’t know how consequential that conflation is because we don’t know how much of the claimed savings falls into this category.
“Programmatic changes,” another component of “estimated savings,” is pretty vague. But it sounds plausible, depending on the details—which, again, we don’t know.
Any actual reduction in spending would help curtail borrowing, so it makes sense to include “interest savings.” But that can’t amount to much in the context of a $2 trillion annual budget deficit and $29 trillion in publicly held debt.
By insisting on “competence and caring,” Musk thinks, he can “cut the budget deficit in half” by the time DOGE sunsets on July 4, 2026. That always seemed improbable, given that such a reduction would amount to about 63 percent of discretionary spending.
If you take DOGE’s numbers at face value, it looks like Musk is well on the way to his target. But it’s clear that we can’t take DOGE’s numbers at face value.
The post Elon Musk, Who Promised To Be ‘Maximally Transparent,’ Makes DOGE’s Numbers Even Harder To Check appeared first on Reason.com.
