The long bond’s big news.
In fact, I’d wager you’re like me in having read more than you care to over the past week or so about 30-year US yields and America’s long-term borrowing costs more generally. I imagine you’ve also heard the word “duration” enough times in recent days to last you the duration of your enjoyable life.
But steel yourself, because you’re going to be subjected, both here and in the mainstream financial media, to a lot more in the way of such editorializing. Because the discussion’s both amenable to politicization and related to AI.
When it comes to competing in 2026’s attention economy, a subject that’s both a hot-button political issue and an AI story simultaneously is a veritable golden goose.
For the layperson, it’s enough to know that nominal (i.e., “all in”) long-term US borrowing costs are the highest in decades and that America just paid the most at auction to sell 30-year bonds in… well, in almost 30 years. But for those interested in a little nuance, the figure below’s worth a look.
That shows you the breakdown of 2026’s 40bps increase in 30-year yields. Long story short: It’s all reals.
This year’s increase has pushed 30-year reals to 3% for the first time since 2008. That repricing has a lot to do with AI, and here I’m not talking about the upward pressure on yields from a quarter-trillion in hyper-scaler IG issuance.
“The repricing in real rates can be partly attributed to anticipation of a future AI-driven productivity boom given the widely-held assumption that technological progress is a primary driver of long-run growth,” BMO’s US rates team remarked, in their weekly, reminding investors that market-based proxies for long run neutral are now disconnected from academic models of r-star.
The figure above shows you the panned-out view. 30-year breakevens have gone exactly nowhere in years, whereas reals are up sharply.
The r-star debate (which is active at the Fed) is best conceived as a background discussion. In the foreground are all the factors you’ve heard so much about in recent days.
“‘Who buys the bonds?’ is morphing into ‘crowding-out risk’ as credit and ‘yieldier‘ stuff continues to challenge a Treasury Department looking for buyers of duration,” Nomura’s Charlie McElligott said, in an August 14 note. The threat of additional bear steepening on the curve and attendant “‘loss of control’ fears” are likely contributing to the bid for gold, he said.
“In addition to concerns about a higher r-star, hyper-scaler debt issuance appears to be actively underpinning 30-year real rates,” BMO’s team went on, in the same note mentioned above. “The logic holds that AI infrastructure financing is adding substantial non-Treasury duration supply to US fixed income markets, with consequences for the outright level of yields as well as the shape of the yield curve and term premium.”
I believe there is a 30-year TIPS auction later this week, but it may be a re-issue. It could be interesting — perhaps just slightly more interesting than your usual TIPS auction anyway — to see what happens there.
I appreciate all of the bond coverage. Like many, I do not understand everything there is to know about bonds, but too often I feel like the talking heads in the news and on T.V. know even less than I do. I hear people saying they would not touch bonds in the current environment, but if you can lock-in a high rate on a government bond, or a highly rated corporate, and plan to hold until maturity, that is a good thing — assuming the issuer can remain solvent! As you mentioned, 30-year rates are the highest they have been in almost 30-years. One question to answer is how much higher rates may go, and the another is who is more likely to go belly-up over the next 30-years: Amazon, Alphabet, or Uncle Sam?
In my infantile view. Long bonds can go to 5% to cover debt growth + 3% to inflate our way out of said debt + 2% to cover duration + unknown for irrational exhuberance. 5% debt growth may be low estimate. Therefore at 11% 30 year bond and 13% mortgages, I am an aggressive buyer of credit assets.
Someone get Warsh on the phone and tell him the Fed should start buying hyperscaler bonds to help lessen the competition for Uncle Sam’s treasuries. If the Fed is buying is buying private sector bonds instead of treasuries, that means it won’t cause inflation, right?
Aside. I am old and in my youth I grew up skipping school to listen to the radio. The soapy quiz shows were big in the morning. One of these was a real tear jerker called “Queen for Day.” Ladies from the lower economic stratum would be chosen to vie for a fancy prize (usually a Speed Queen washer/dryer combination) they couldn’t otherwise ever afford. That picture of Scott you posted here just puts me in mind for all the quiz show memes I listened to on Chicago radio in the 1950s. That toothy grin. Sorry, strictly Kitch.