DOGE is not cutting the deficit

Real federal spending is higher in 2025 so far than in past years, and DOGE is making it harder to raise revenue

Elon Musk has sold his “Department of Government Efficiency” on a two-plank platform: The first is that there is waste in government. The second is that Donald Trump was elected to root it out, and that Musk is the man to do it.

The first claim is certainly true. Look anywhere in the federal government with a powerful enough microscope and you will find waste. The White House said the government was spending $8 million on subscriptions to Politico, for example, which does seem a hefty amount for taxpayers to fork over (could some departments have shared?). Infrastructure projects are plagued with cost and time overruns, too. And don’t even get me started on added health care costs.

The second of Musk’s propositions may not be quite so apt, however. That’s because in the two months DOGE has had to find the fat in government and cut it out — with nearly unchecked power to do so — it has done remarkably little of impact. In fact, Musk may be increasing bloat, rather than reducing it.

According to data collected by The Hamilton Project, the U.S. government is spending more money in 2025 than in 2024. I had seen a version of the chart below floating around online, but it did not adjust for inflation (which would naturally make it look like the government was spending “more” over time), so I did that myself.

According to the Hamilton Project data, the U.S. government has spent $2.17 trillion as of April 10, 2025, at 7:00 p.m. ET (the data, which comes from the Treasury, is updated in near-real-time). That is just shy of a 6% increase over spending on the same date in 2024 when the feds had spent “just” $2.05 trillion at this point in the year. This puts the US federal government on track to spend $8 trillion this year, barring other budget changes..

To put it another way, real government spending has gone up by 120 billion dollars since this point last year — or 12 with ten zeroes, 120,000,000,000. That is not an insignificant number; it is about ten times as much as the U.S. spent on the Agency for International Development last year, for example.

Of course, an increase in the overall budget cannot be directly attributed to DOGE, just like a total decrease can not be credited to them without proof of cuts of this magnitude. It is possible, to be generous to Musk, that he really has saved $120 billion but spending in other parts of the government has increased.

But at the very least, Musk isn’t really delivering on what he’s selling. And he may in fact be hampering his own efforts: Take for example the tens to hundreds of thousands of firings and re-hirings DOGE has been responsible for. All the paperwork the government does to affect personnel changes must come with labor and other costs. The indiscriminate cuts also cause waste by virtue of leaving loose ends for other people to tie up. This is just an anecdote of course, but several of my friends in the government are working in half-empty offices with no official directives because their programs have shut down, but they have not been fired.

Then, there are the indirect costs we all incur when the Internal Revenue Service cannot effectively enforce the country’s tax laws. The Washington Post reported this week that DOGE-enforced cuts at the IRS will cost us all $500 billion in lost tax income, or about 10% of federal tax revenues. That’s primarily because of staff and other tech cuts, as well as the reorientation of some organizational progress to data-sharing instead of enforcement for tax cheats. For those keeping score, that is nearly 4 times the total savings DOGE claims on its website (itself a likely overestimate).

So I guess you could stylize the DOGE return-on-investment calculations like so:

When you put it that way, maybe the “move fast and break things” approach that works in Silicon Valley is not a good fit for government service. When it comes to my tax dollars, I would rather the government move at the appropriate speed, and break the appropriate number of things.