The Math Is Not Hyperbole
Marc Goldwein, senior policy director at the Committee for a Responsible Federal Budget, put it plainly this week: the country is now spending significantly more just to service past debt than it invests in its own future. The mechanics are brutal and simple. The U.S. is on track to borrow more than $2 trillion this year. Roughly half of that new borrowing goes straight to interest on debt already outstanding. You borrow to pay interest, which grows the debt, which grows the interest, which forces more borrowing. That is the technical definition of a debt spiral -- and by the Old Goat's read, arguably a death spiral, because every one of the traditional exits is blocked at the same time.
The 30-year Treasury yield hit its highest level in nearly two decades this week -- meaning the cost of new borrowing is accelerating at precisely the moment the government needs to borrow the most. Treasury Secretary Scott Bessent's own target -- cutting the deficit to 3% of GDP by 2028, down from over 6% when Trump took office -- is now mathematically difficult to hit without a war ending, rates falling, or a growth surprise that isn't currently on the board. Bessent has acknowledged as much himself.
Three Revenue Pillars, All Down at Once
- Tariff revenue, reversed. The Supreme Court ruled Feb. 20, 2026 that IEEPA does not authorize the tariffs Trump imposed, including the "Liberation Day" duties. Those tariffs had raised upward of $160 billion; roughly $81 billion has already been refunded this fiscal year, with total exposure potentially running far higher as the Court of International Trade works through the refund process. The fiscal offset the 2025 tax cuts were counting on didn't just shrink -- it reversed into an outflow.
- DOGE savings, unverified. The Department of Government Efficiency promised $1 trillion in federal savings. The Government Accountability Office's Aug. 6, 2026 audit of DOGE's own "Wall of Receipts" found the agency could not substantiate its methodology for the large majority of the savings it claimed on federal contracts, and could not verify 96% of the method behind its claimed grant savings -- including credit taken for leases that were already being phased out before DOGE existed. Whatever DOGE's real net effect on the budget is, it is not the verified pillar of fiscal consolidation the administration has presented it as.
- Growth, dragged. Elevated energy prices and defense spending tied to the Iran war are acting as a fiscal multiplier in the wrong direction -- raising costs while the growth that was supposed to lift tax revenue hasn't materialized at the pace the administration projected.
Update, August 22, 2026: Where the $40 Trillion Actually Came From
One caveat first: not everyone measures it the same way. Counted as federal debt alone -- stripping out the state and local debt the $40 trillion headline blends in -- the number is closer to $32 trillion, per Paul Krugman's own accounting. Doesn't change the direction of anything above. It does mean the figure depends on which one is being cited.
The more useful addition is causal. Bobby Kogan and Jared Bernstein -- he ran the Council of Economic Advisers under Biden -- have modeled how much of the current deficit is tax policy rather than spending. Their finding, in Kogan and Bernstein's own words: primary deficits (the gap before interest costs) are "now projected to be roughly 3 percent of GDP indefinitely." Krugman, citing the same modeling, put it plainly: the deficit "would be far more manageable" without four rounds of Republican tax cuts -- Bush 2001, Bush 2003, the 2017 Tax Cuts and Jobs Act, and 2025's One Big Beautiful Bill, which Kogan and Bernstein separately price at $5.8 trillion in lost revenue (adjusted for economic growth), or $6.6 trillion if its temporary provisions are made permanent. That is a tax-policy story before it is anything else.
It also closes a loop with the earlier point about who holds the debt. Domestic holders own roughly 70% of it -- much of that through pension, mutual, and insurance funds -- per Treasury and Committee for a Responsible Federal Budget data, and Federal Reserve Distributional Financial Accounts data puts the top 1%'s share of all household financial assets at on the order of a third. (Old Goat's synthesis, not a claim made by Kogan, Bernstein, or Krugman:) cut taxes concentrated at the top, borrow to cover the resulting gap, then pay interest on that borrowing to the same households who disproportionately hold the debt. The mechanism isn't hidden. It's arithmetic -- now with a name and a modeled number attached to the tax-cut side of it.
One more structural piece, tying back to the Tokyo intervention and Warsh sections below: AI-infrastructure borrowing is now a second large customer for the same capital Treasury needs. Nine major tech companies spent roughly $600 billion on infrastructure over the past year, hyperscaler bond issuance is on pace to roughly double in 2026, and Goldman Sachs projects debt will fund more than a third of hyperscaler AI spending by 2027 -- while Treasury has to refinance $9.7 trillion in debt coming due this fiscal year on top of covering a CBO-projected deficit near $2.1 trillion. (Old Goat's read:) that's a cleaner explanation for why the bond buyback below hasn't calmed the 30-year yield than "the market didn't believe it" -- there's a second real borrower in the room a buyback funded by more short-term issuance doesn't touch.
It also independently corroborates the Warsh line in the next section. Jared Bernstein, on Kevin Warsh's Fed chairmanship, in a conversation posted to Krugman's newsletter this month: "Kevin Warsh has gotten off to a bit of a shaky start, and I think that's spooking markets a bit."
The Tokyo Intervention
One line from this week's reporting deserves more attention than it has gotten: in early August 2026, Bessent helped engineer the first joint U.S.-Japan currency intervention since 2011, buying yen to arrest its slide. Japan holds roughly $1.1 trillion in U.S. Treasuries -- the largest single foreign holding in existence -- and a sharply weakening yen pressures Japan's Ministry of Finance to sell Treasuries to raise dollars and defend its own currency. Bessent has separately been pushing the Fed to expand a lending facility (FIMA repo) that would let Japan support the yen without having to touch its Treasury holdings directly.
Bessent intervened in a foreign currency market specifically to keep one of the Treasury market's largest customers from having to sell. That is the tail managing the dog: the U.S. shoring up its own sovereign debt market by propping up the currency of its biggest foreign creditor.
This is, in the Old Goat's read, the clearest evidence yet for the thesis this Old Goat has been tracking all year: the bond market is not a backdrop to Trump-era policy, it is the operational constraint on it. The 30-year yield now does more to shape what the administration can and cannot do than any single executive order.
Warsh's Balance Sheet
Fed Chairman Kevin Warsh, who took over in May 2026, has made shrinking the Fed's multi-trillion-dollar portfolio of government bonds and mortgage-backed securities a top priority -- a task force he's convened is due to report by year-end. Bessent himself has said publicly it could take up to a year for the Fed to settle on a direction here. The tension is real: that Fed portfolio has functioned as a buyer of last resort for Treasuries. If Warsh pursues outright sales rather than letting the portfolio run off through maturities, that adds fresh Treasury supply into a market already straining to absorb roughly $2 trillion in new issuance a year -- more supply, lower prices, higher yields, a deeper hole. Traders describe an aggressive QT path as far more disconcerting than a passive one. Warsh is simultaneously expected to hold the inflation line and to weigh dumping trillions in bonds onto a stressed market. At $40 trillion in debt, those two goals may not be reconcilable.
What This Means for the $300 Billion Iran Number
Put the debt picture next to the $300 billion Iran-reconstruction figure from the G7 Memorandum of Understanding this Old Goat has tracked since June, and the number gets harder to take at face value. Iran MOU Tracker → A government borrowing $2 trillion a year and refunding $160 billion in struck-down tariffs is not in a fiscal position to also commit $300 billion to a reconstruction fund as a hard, near-term obligation. That doesn't mean the commitment is fake -- the MOU text is real and documented -- but it strongly suggests the number functions as negotiating architecture rather than money that is actually going to move on the timeline implied.
The Gulf Read
The sovereign-wealth relationships this Old Goat has tracked through the Kushner/Witkoff orbit -- Qatar, Saudi PIF, the UAE -- read differently against a $40 trillion debt load than they do as a story about individual self-dealing alone. Orbit Tracker → A country that has run out of domestic fiscal room to fund its own obligations has a structural reason, not just a personal one, to court foreign sovereign capital willing to roll American debt or invest in American assets. That is a materially darker read than where this Old Goat started on those relationships, and it does not require assuming any individual deal was corrupt to be true -- it only requires taking the $40 trillion number seriously as a constraint on everyone in the administration, not just as a talking point.