THE NUMBER EVERY STRATEGIST QUOTES — AND NO TRADER USES.
The Bond Bro Dispatch · Special Edition · September 2, 2026
Term premium hits the front page. A 20-year desk audit of the models everyone is quoting.
Yesterday it was CNBC, all day. This morning it graduated: Bloomberg's front page is running "US 30-Year Average Yield Highest Since 2004" next to a headline about Treasury losses, with a subhead that says — in Bloomberg's own words — investors are demanding more compensation to hold longer-maturity debt. That sentence is the term premium. The word is suddenly everywhere, the number behind it comes from two models, and in twenty-plus years on institutional desks I never watched anyone trade off either one. This edition is the audit.
WHAT THE NUMBER ACTUALLY IS
A 10-year yield is two things glued together: the market's guess at the average path of the funds rate over ten years, plus everything you charge for locking up for ten years instead of rolling — inflation risk, supply risk, the risk the bid isn't there when you need it. That second piece is the term premium, and the problem TV skips is that it is unobservable. No screen prints it. It is a residual: what's left after a model guesses the unguessable — where the funds rate averages out to 2036.
Two models produce the number that gets quoted. ACM — Adrian, Crump and Moench, the NY Fed's model — is purely statistical: the curve itself is the only input, published daily. Kim-Wright, from the Fed Board, runs similar machinery but anchors the expected rate path with forecaster surveys. Historically they disagree by half a point or more, which tells you the precision to attach to any single headline number. Except right now:
When two models built on different philosophies land within a few basis points of each other, the directional claim is as robust as this framework gets: the post-2024 rise in long yields has been premium, not the expected path of policy.
THE DESK NEVER NEEDED THE MODEL
Here is the part the panels won't say: trading desks never used these models, because the desk trades the observable expressions of the same forces, in real time. Auction tails — when buyers demand a concession to take the bonds. Long-end swap spreads — the rent the system charges to warehouse duration. Dealer positions against balance-sheet caps into supply weeks. The foreign bid in the TIC data. That is the term premium, priced live. The models are the smoothed, lagged, academic restatement of what the tape already told you.
Every line the model asserts, the tape confirms independently. That triangulation — models, desk observables, and the global co-movement in Exhibit 1 — is a defensible argument. A single model output isn't.
THE FED CUT. THE CURVE DIDN'T COME WITH IT.
One correction to the popular framing, straight from the tape: this is not a long-end steepening story. Over the last twelve months the belly led the selloff — 3s repriced +86bp against +33bp in the long bond, and 30s10s actually flattened. The accurate description: the Fed cut, the front end followed policy down, and every point from 2s out repriced higher anyway. The whole curve now sits 40 to 155bp above a falling funds rate, with CPI at 3.4 and crude up 57% on the year. A market that believed in the cutting cycle does not price 2s at 4.41 against a 3.75 funds rate. That is the expectations leg and the premium leg pushing the same direction — and neither one rallies back on a soft print.
THE VERDICT — WHERE THIS LANDS FOR MUNIS
Long muni yields price off a Treasury long end whose move is premium-driven, and premium-driven yield is sticky — it unwinds on supply, balance sheets and the buyer base, not on two soft inflation prints. The z-score board says munis are the most dislocated sector on the tape: MUB at −3.7 standard deviations, short munis at −4.4, cheapening harder than Treasuries, credit, or MBS. And the raw ratios say the cheapening has a shape:
Read the ladder: 57% in 2s and 62% in 5s is still rich — the tax exemption is fully priced and then some. But 70% in 10s, 80% in 20s and 87% in 30s, with the whole complex sitting at −3 to −4 standard deviation prices, is the buyer's side of the curve cheapening in real time while the Treasury long end it prices off is being held up by premium, not expectations. A 4.58% AAA 30-year tax-exempt against a 5.29% Treasury is a 7.7% taxable-equivalent at the top bracket. That is what the entry point looks like while everyone else is learning the vocabulary.
Pressure Gauge, carried from the Aug 30 letter: Term Premium/Long End — PRESSURE, trading against an official seller with a stated price objective. This morning's tape does not argue with that reading; it underlines it.
PREVIEW — THE PREPAY SETUP
Here is where the ratio ladder stops being an observation and starts being a setup — in gas prepays, the corner of the market this desk spent years in. The trade needs four things at once, and the tape has all four. Rich front-belly ratios keep the issuance arb alive, so supply keeps printing. Belly USTs near 4.6–4.8% set the absolute-yield foundation. The 10Y-and-out ratios are cheapening live. And bank funding appetite — the other side of every prepay deal — is elevated at these rates. Structured bank risk in tax-exempt form, at ratios the generic curve can't touch, in a market where evaluated pricing is weakest exactly where this paper lives. The CUSIP-level work — which guarantors, which puts, which vintages — is Vault material. This is the map; the coordinates are behind the door.
Sources: Koyfin global yields, curve spreads, credit and sector ETF pulls, 9/2/2026 6:05 AM ET; Bloomberg front page 9/2/2026; NY Fed (ACM) and Federal Reserve Board via FRED (Kim-Wright), late-Aug 2026 levels; AAA muni curve — retail national benchmark (Hennion & Walsh), 9/1/2026.
Produced with AI assistance. All data selection, analysis, conclusions, and final editorial judgment are the author's. All content is reviewed and approved by Positive Carry LLC prior to publication.
The Bond Bro Dispatch is published by Positive Carry LLC. All content is general market commentary provided for informational and educational purposes only and does not constitute investment advice, a recommendation, an offer, or a solicitation to buy or sell any security. Nothing herein is tailored to the circumstances of any recipient. Data are drawn from sources believed reliable; accuracy and completeness are not guaranteed. [email protected] · © 2026 Positive Carry LLC. All rights reserved.