Remarks to a Fixed Income Desk That Didn't Ask For My Opinion
Let's start with the number nobody wanted to say out loud on Friday: five point two.
That's how many billion dollars the Treasury actually repurchased in Wednesday's expanded buyback operation — the one authorized back in August, cap doubled to $4 billion per bucket, running September 9th through November 4th, sold to all of you as the circuit breaker for the long end. Ten point five billion was on offer. They got five point two. Below the six billion cap. Below half of what was tendered. You do not need a bond desk to tell you what a failed buyback looks like. It looks like nobody showing up to sell you what you offered to buy.
And the 10-year didn't wait for the ink to dry before answering back — 4.97% on Friday, the highest print since October 2023, before Thursday's CPI-adjacent chop pulled it back down into the high 4.7s. Same week, the 30-year sat parked around 5.24%, and I want you to notice something about that number: it's basically unchanged from where it was a month ago, at 5.3%-plus territory, meaning the entire apparatus of buybacks, of jawboning, of a new Fed chair supposedly restoring credibility — none of it has actually moved the long end anywhere. It moved the vocabulary around the long end. The yield itself just sat there like it knew something.
Here's the part I actually want on the record. This isn't an American problem wearing an American flag. Two weeks ago, Japan's 10-year JGB crossed 3% for the first time since 1996 — three decades — while the 2-year touched 1.81%, a 31-year high. UK 10-year gilts hit 5.25%, a level last seen in June 2008, in the teeth of the global financial crisis, and the 30-year gilt printed 5.89%, the worst since March 1998. Germany's benchmark, the anchor of the entire eurozone curve, moved up in sympathy. This is not four sovereign debt markets independently discovering four separate domestic problems. This is one global repricing of duration, happening simultaneously, across every major currency bloc, and the coincidence is the story.
So ask yourselves what's actually competing for capital in that market right now. It's not just sovereigns rolling debt into a higher-rate world. It's AI hyperscalers, who have collectively issued north of $1.5 trillion in new corporate debt this year — a figure that would have sounded like a typo eighteen months ago — financing data center buildouts and GPU clusters against revenue streams that are, charitably, still maturing. Every one of those bonds gets priced off the same curve the Treasury is trying to defend with a buyback operation that just failed to clear half its own offer size. You cannot out-buy a market that's absorbing sovereign refinancing needs, a trillion-and-a-half in fresh AI paper, and — starting Wednesday — a Fed that's sitting at roughly 90% odds of hiking into all of it, after Friday's core CPI print beat expectations at 0.3% month-on-month.
I'll say the quiet part plainly, because someone in this room needs to hear it stated rather than implied: a rate hike into this bond market doesn't relieve the pressure on long-end yields, it validates it. The whole hiking cycle was supposed to be the mechanism that eventually let the curve normalize downward once inflation credibility returned. Instead we're getting the hike and the credibility question at the same time, in a market that's already telling you, through a failed buyback and three separate developed-market yield records set in the same fortnight, that duration itself is the asset nobody wants to hold at these prices — regardless of which central bank is setting the front end.
What happens Wednesday afternoon almost doesn't matter as an isolated event anymore. Twenty-five basis points on a fed funds rate that's already sitting at 3.5 to 3.75 is a rounding error next to a 10-year that flirted with 5% on its own, driven by a debt-management operation that couldn't find enough sellers to fill half its own tender. The Fed can hike, hold, or cut, and the structural question underneath all of it stays exactly where it was on Friday morning: who, in a world issuing sovereign debt, war-driven energy premiums, and trillion-dollar AI credit lines simultaneously, is actually supposed to be the marginal buyer of thirty-year paper at these yields? Bessent's buyback was supposed to be part of the answer. Wednesday's auction results suggest it isn't one yet.
Go back to your desks. Watch the long end, not the fed funds rate. The rate decision is theater we've all agreed to watch together. The bond market is where the actual vote is happening, and it's telling you the ballot's still open.