The bond-pain case, tested
2026-09-26 · One subject in depth · The bond-pain case for higher US yields
How it is made. Aavistus Daily is written and voiced by our own systems, every morning, from data we gather ourselves. The voices are synthetic. Every fact is traced to a source on this page, and every dated call is settled in public on the calibration page. The presenters, Nora and Callum, are named characters, not people.
Aavistus is a new framework, still under development. We will make mistakes, and every dated call is settled here in public so you can see them; the aim is to keep calibrating until the record beats the average forecaster, not to sound right.
A fixed-income trader made a specific case yesterday that US yields keep rising until something breaks — we test the numbers, agree on direction, disagree on the terminal path.
The argument
On Forward Guidance yesterday, a fixed-income trader who goes by DCP told hosts Felix and Quinn that the US ten-year Treasury closed at 5.22%, that the thirty-year bond future has fallen nearly ten full points inside ten weeks, and — the part worth arguing with — that the thirty-year yield needs to push into the sixes and front-end Secured Overnight Financing Rate futures need to rise another fifty basis points before the market finds its balance point. His break comes through an AI-capex plateau or a private-credit accident.
Where we agree, where we don't
On direction, we agree. Yesterday we said the base case is the US ten-year holding in the range we set out then, with the American equity market taking a modest de-rating rather than a full break. The disagreement is on the terminal case. The 10Y-2Y Treasury spread at 0.31 and the 10Y-3M at 0.94 — both wider on the week — are consistent with rising term premium, not with the front end pricing more hikes. It is the long end doing the tightening, not policy.
Credit spreads back the mechanism but cap the story. The high-yield option-adjusted spread stood at 2.8% on Thursday, wider by a tenth on the week. You do not get a private-credit accident at 280 basis points over Treasuries; the spread has to widen materially first, and it hasn't started. So the pain being felt is duration pain, not default pain, at least not yet.
One piece of DCP's view worth steel-manning: his ultimate trade, once the hiking case is priced out, is a generational long on short-term interest-rate futures. On that shape we agree entirely; where we differ is on how many more basis points the front end has to price before the reversal begins.
What's actually happening at the long end
Debt supply and an inflation-persistence premium — not the policy path. Under any credible entitlement-spending trajectory the primary market has to absorb heavy issuance every year for a decade, and labour-force growth is the release valve that lets the US grow into its debt. Neither is trading noise for now, but both strengthen the case that term premium — not the policy path — is where the yield story continues to live.
The tell that would flip our range view is unchanged: five consecutive daily closes on the ten-year above 5.4% with no break in the S&P.
From where we sit
Euro area core inflation came in at 2.4% on Friday's Eurostat print, down a tenth on the month — the ECB does not have a US-style core problem. But Spain at 4.6% headline and 3.6% core, and Norway at 3.5% core, up four-tenths on the month, mean the peripheral-and-Nordic story diverges from the German-French core. The ECB meeting on 29 October is not the non-event the market is pricing. Our earlier view that an October hike is no longer off the table stands; the trigger that would confirm it is a hawkish speech on the Spanish or Norwegian core reading before the meeting.
On the Nordic banks: yesterday we said the Norges Bank hike does not change the read on Nordea and DNB from where we sit, and that view stands. Nordic real-estate credit does need US yields not to break higher — European commercial-property spreads sit closer to the US long end than to the domestic policy path, and the case for adding there is not there until the tell above trips.
On the megacaps: our Wednesday view that the American megacap trade is priced for perfection stands. Meta's rip this week is what the top of a single-name flip looks like, not a new leg for the trade. The read only changes if Apple, Microsoft or Alphabet rip on similar news inside the next fortnight — none of them has moved with Meta this week, and that is the tell.
Watch list
US non-farm payrolls on Thursday 2 October — soft keeps the range case comfortable, hot gives DCP's terminal case room. Nord Pool Monday–Wednesday averages for the western Nordic zones — the base case is that this week's surge passes as wind returns in early October; what would flip it is SE3 or Finland running above €100/MWh through Wednesday. And ECB commentary — a hawkish warning from a Council member on the Spanish print or the Norwegian core reading is what would tell us the 29 October meeting is live in a way the market is not yet respecting.