NextFin News - Treasury Secretary Scott Bessent said the recent rise in U.S. Treasury yields reflects global forces rather than anything specific to America's credit, brushing aside concerns that benchmark borrowing costs have climbed to levels not seen in more than two decades.
In an interview published Saturday, Bessent argued that the selloff in Treasuries is part of a synchronized global move in government bonds, not a vote of no confidence in U.S. fiscal management.
"I would be concerned if we were having some kind of idiosyncratic rise," Bessent said. "We're not seeing people selling treasuries to buy German bonds or Japanese bonds."
The reassurance arrives as the 30-year Treasury yield touched 5.69% in late September, its highest level since 2002, while the 10-year note trades around 5.25%, the highest since 2007. For a White House that has made lower borrowing costs a central economic promise, the message from the long bond has become the administration's most awkward headline - and Bessent's test for whether to worry is now live.
The Calm From the Treasury, and the Market That Disagreed
Bessent's argument is straightforward and, on its face, testable. If investors were fleeing U.S. credit specifically, capital would rotate into the other deep sovereign-bond markets. German bunds and Japanese government bonds have also risen, which is consistent with a global repricing rather than a U.S.-specific exodus. On that narrow test, the data support him: the German 10-year yield was around 3.46% on Friday, up more than three-quarters of a percentage point over the past year, and Japanese yields have climbed to multiyear highs as well.
But the global-cover argument only goes so far, and recent weeks have shown why. On Sept. 8, speaking at Southern Methodist University, Bessent signaled that the Treasury was prepared to defend the long end of the market, telling the audience he had "asymmetric information" and that traders could "bet against me if you want." The market took the dare. Over the following three weeks the 30-year yield climbed from about 5.25% to as high as 5.69%. A finance secretary warning the bond market that he is "the house" is not the posture of an official who controls the price - it is the posture of one asking it to behave.
The episode also revealed the limits of the administration's toolkit. The Treasury's defense of the long end has run through buybacks and jawboning rather than through any change in the fiscal trajectory that drives supply. A buyback can smooth the maturity profile of outstanding debt; it cannot reduce the amount of debt the government must roll over and refinance. When the market asks for a higher clearing rate on 30-year paper, a secretary can talk, but only Congress can borrow less.
The timing exposes the central tension of the administration's economic program. Treasury policy is simultaneously trying to talk down yields, finance a federal debt that has passed $40 trillion, defend an industrial strategy built on cheap capital, and keep inflation expectations anchored after the Federal Reserve raised rates by 25 basis points on Sept. 16 to a target range of 3.75%-4.00%. Those objectives can coexist in a calm market. In a market that is repricing term risk, they compete.
Why Yields Are Rising: Global Cover, Domestic Load
The global component is real and cannot be waved away. Inflation in the United States remains above the Fed's 2% target - the central bank's own September projections put core personal-consumption-expenditures inflation at 3.4% for 2026 - and energy prices have been jolted by the war in the Middle East, with Brent crude near $100 a barrel and WTI around $90. When inflation prints stay hot and oil spikes, every major central bank faces the same uncomfortable trade-off, and every sovereign curve adjusts upward. That is the environment Bessent is describing, and he is right that it is not made in Washington alone.
Yet the U.S. curve carries a premium that bunds and JGBs do not, and the spread between them is the tell. The gap between the 10-year Treasury and the German 10-year bund stood at roughly 180 basis points this week - within the range that has been normal for the past decade, but at the upper end of it. America is running a fiscal deficit that requires heavy new issuance at the same moment that the private sector is undertaking the largest capital-spending program in decades. Technology companies have committed roughly $750 billion to artificial-intelligence infrastructure this year, a buildout Bessent has praised as a productivity miracle, and every dollar of it must be financed somewhere in the same pool of savings that buys Treasuries.
The mechanism runs through three channels, and separating them matters because they have different cures. First, supply: net Treasury issuance into a market where the federal debt has crossed $40 trillion demands a higher clearing rate. Second, inflation: tariff pass-through and energy shocks keep the real rate bid higher than it would be in a disinflationary regime. Third, the term premium: the extra compensation investors demand for holding 30-year risk has been unusually low for years and is only now normalizing. Add those together and you get higher yields at the long end without any single actor "attacking" the dollar.
This is where the cyclical-versus-structural call has to be made cleanly, because the two forces point to different conclusions. The inflation and energy legs are cyclical: oil shocks fade, tariff pass-through ends, and the Fed's own forecast expects core inflation to recede toward 2.5% by 2027. Mean reversion is the base case there, and history supports it - the 10-year yield last traded near these levels in 2007, before the global financial crisis reset the entire rate regime, and it spent most of the years since well below 3%. The deficit, however, is structural. It is driven by entitlement spending, interest costs on the existing stock of debt, and a political system that has not produced a credible consolidation plan. That portion of the yield rise will not revert on its own. Bessent is right that the cyclical wave will pass; he is less clearly right about the structural one.
Market pricing makes the Fed's bind explicit. Traders have anchored the odds of zero rate cuts in 2026 near 96%, and futures-implied probabilities put the chance of a cut at the October meeting at roughly 16%. The Fed's own dot plot points to one more quarter-point increase before year-end, with the median official expecting the funds rate to sit higher at the end of 2026 than it does today. In other words, the market is not pricing a central bank that is about to rescue the bond market; it is pricing one that must wait for inflation to move. Bessent can say the rise is not idiosyncratic. What he cannot say is that it is cheap.
The AI Question Bessent Will Not Answer
The second half of the story is the one Bessent has spent months declining to entertain: whether the AI buildout itself is a financial-stability risk. In July, a draft report inside the Treasury Department warned that the AI market was repeating the 2000 dot-com bubble and that a burst would ripple through the broader economy. The department disavowed the document as the work of a career staffer that did not reflect its views, but the fact that it reached the secretary's desk is itself data. The dot-com analogy is imperfect - today's AI leaders are cash-generative incumbents rather than loss-making startups - but the question underneath it is not.
The administration's public line has been consistent since. At a Sept. 15 House Financial Services Committee hearing, Bessent rejected calls from AI executives for a slowdown in frontier-model development.
"They could stop any time they want to," Bessent said. "They could stop their progress any time they want to."
The exchange came after executives at leading U.S. AI companies endorsed slowing development over existential-safety concerns. Bessent's point was that the industry is voluntary and self-correcting; the bond market's point is that capital is not. Once a data center is financed and a GPU order is placed, the spending has a momentum of its own, and the lenders who underwrote it are exposed to the returns it produces.
The connection between the two topics - yields and AI - is the cost of long-duration money. A $750 billion capex program, much of it financed through project finance, asset-backed structures, and private credit with limited transparency, is acutely sensitive to the discount rate. Every 50 basis points added to the 30-year bond raises the hurdle rate on data-center economics and forces a re-underwriting of the AI investment thesis. The higher yields go, the more AI spending looks like a bet on future productivity that must be proven rather than a certainty that can be financed at any price.
This is where the administration's two priorities can collide. It wants AI investment to accelerate - the growth and productivity story depends on it - while also wanting yields to fall. If the bond market is signaling that the savings pool cannot fund both the deficit and the AI buildout at 4% money, then talking down yields without addressing either the deficit or the pace of capex is an attempt to wish away a price signal. The draft Treasury report's dot-com analogy may have been disowned, but its underlying question - what happens to the lenders when the assets do not produce the expected cash flows - is now being answered by the market, not by a memo. That is the second-order transmission from yields to the real economy: higher rates do not just make borrowing expensive; they make the AI growth story itself more expensive to believe.
What the Bond Market Is Actually Telling the White House
There is a political-economy reading that should not be ignored. Concerns about the affordability of fuel and other essentials have made Republicans in Congress nervous about the November elections, and a Treasury secretary telling the public not to worry is as much a political signal as an economic one. Reassurance is cheap; yields are not. The administration's economic standing rests on the voter's experience of prices and borrowing costs, and both remain elevated. A 30-year mortgage rate that tracks a 5.25% 10-year yield is a household-budget issue long before it is a macro issue.
The strongest counter-thesis to this reading is also the simplest, and it deserves to be stated plainly: if yields are rising across Germany, Japan, and the United States, then singling out American policy is misplaced. A global risk premium driven by war, sticky inflation, and a repricing of term premium is not something any single finance ministry can fix, and blaming the White House for a worldwide move is politics, not analysis. Bessent's bund-and-JGB test is a legitimate way to make that point, and on the direction of travel he is correct. The world is repricing sovereign debt together, and the U.S. is not uniquely broken.
The answer is that both things can be true. The direction is global; the amplitude is partly domestic. The spread between U.S. and German 10-year yields is the tell, and it is the relative move, not the level, that would prove Bessent wrong. His claim is that the rise is not idiosyncratic - that is, not specific to the United States. The falsifying signal is concrete: if the 10-year Treasury rises more than 75 basis points faster than the German bund over a rolling month, or if the U.S.-Germany 10-year spread widens beyond roughly 250 basis points, the "global trend" defense no longer holds and the rise becomes exactly the idiosyncratic scenario Bessent said would worry him. At roughly 180 basis points today, the market is not there yet - but it is watching, and the threshold is observable week by week.
What to Watch
Short term, the path of oil is the swing factor. A durable de-escalation in the Middle East would pull diesel and crude lower and take some inflation premium out of the curve; a wider war, or a disruption to the Strait of Hormuz, does the opposite. The next personal-consumption-expenditures print will tell whether inflation is still sticky above 3% or whether the tariff pass-through is fading as the Fed expects. A string of prints at or below 0.2% month over month would be the first real evidence that the cyclical leg is turning.
Medium term, the scrutiny shifts to the hyperscalers. Third- and fourth-quarter earnings will show whether AI capital spending is producing returns or merely accumulating assets on the balance sheet. If revenue per dollar of capex disappoints, the equity market will reprice the buildout, and the bond market will ask how much of it was debt-financed. That question travels: a repricing in AI equities would tighten financial conditions, pull down the collateral values behind private-credit exposure, and feed back into the very yields that started the move.
Long term, the structural question is the deficit and the term premium. If the government continues to issue heavily into a market where investors demand a higher premium for duration risk, yields can stay elevated even without new inflation shocks. That is a slower, quieter problem than a crisis, but it is the one Bessent's reassurance does not reach. A fiscal framework that credibly slows the growth of debt - and the Treasury buyback program can help manage maturity risk but cannot substitute for it - would do more for yields than any number of interviews.
Base case: the 10-year yield plateaus in the 5%-5.3% range as oil stabilizes and the Fed holds at 3.75%-4.00%. Upside case: disinflation accelerates and the Middle East de-escalates, allowing a move toward 4.5%. Downside case: inflation re-accelerates or the deficit trajectory worsens, pushing the 10-year toward 5.75% and the 30-year back above 5.7%.
Bessent is right that the world is repricing sovereign debt together, and his bund-and-JGB test is the correct way to watch it. But the bond market is not asking whether America is uniquely broken; it is asking whether the savings pool can fund the deficit and the AI revolution at once. The administration's answer so far has been to talk - and yields have kept rising anyway.
Explore more exclusive insights at nextfin.ai.
