David Hay — Haymaker Daily: Not The Way It's Supposed To Work — the 10-year has gone from ~3.70% to 4.75% since the Fed began easing with an emphatic 50bp cut, the opposite of every cutting cycle back to 1970; blamed on recession-like deficits during an expansion plus the AI build-out's capital hunger, with 1999 the nearest analogue except that America then ran surpluses — leaving the Treasury market "extremely vulnerable" (Sep 1 2026)

A one-fact Daily, and the fact is a broken pattern. "The word 'unprecedented' is often overused during these tumultuous times. However, when it comes to what has occurred since the Fed began its latest interest-rate easing cycle, it truly qualifies." The measurement: "from the time the Fed began this easing phase, the yield on the 10-year U.S T-note has vaulted from around 3.70% to the present 4.75%," and the aggravating detail is the size of the opening move — "what makes it all the more remarkable is that this cutting cycle began with an emphatic 0.50% (50 basis points) reduction." Against the record: "going all the way back to 1970 the consistent reaction to the inception of an actual Fed rate-cutting cycle has been for longer-term Treasury yields to recede" — four Bloomberg charts are supplied because the point needs the full sequence to land. The obvious objection is raised and dismissed on its own terms: yes, there have been brief cuts into a strong economy, "such as in 1998, when the Fed cut on an emergency basis while the economy was still robust… known as a 'mid-cycle' adjustment in reaction to… the Asian Crisis and the related implosion of Long-Term Capital Management," but that one resolved the other way — "less than a year later, the Greenspan-led Fed was hiking again, and would do so three times." The diagnosis is three-part: "deep, recession-like deficits during an ongoing economic expansion," "the voracious capital needs of the AI build-out," and the timing of the easing itself — "the Fed began this easing period when the economy was not in the early stages of a downturn," which is where 1999 comes in, "when the first tech bubble (it's now necessary to distinguish between them) also continued inflating." And then the disanalogy that carries the whole note: "a massive difference versus 26 years ago is that the U.S. government was running such large surpluses back then that there were concerns all federal debt would be extinguished over the next 10 to 15 years" — against "the present pathetic state of America's fiscal condition." Conclusion: "this deep fiscal hole leaves the Treasury bond market, and the economy, extremely vulnerable to longer-term rates continuing to hit the highest levels in a generation," and the policy tell is already on the tape — "based on his recent yield-manipulation gambit, it's clear Treasury Secretary Scott Bessent is fully cognizant of those risks."

One-line take: a macro-only Daily — no security is named — and the shortest, cleanest statement yet of the bond argument that has run through this week's notes. The structure is a base-rate test: establish what normally happens (since 1970, long yields recede once a real cutting cycle starts), measure what happened this time (3.70% → 4.75%, with a 50bp opening cut that should have made the effect larger, not inverted it), and treat the deviation itself as the finding. That is a more disciplined way to use history than the usual analogy-shopping: the pattern is not evidence for a view, it is the null hypothesis the current tape has rejected. Hay then does the part most commentators skip — he steelmans the one counterexample. 1998 was also a cut into strength, but he supplies the reason it doesn't rescue the comparison: it was an explicitly labelled "mid-cycle" response to the Asian Crisis and LTCM, and it was reversed — "less than a year later, the Greenspan-led Fed was hiking again, and would do so three times." Only after clearing that does he offer causes, and he keeps them plural and ranked rather than picking one: recession-sized deficits during an expansion, the AI build-out's capital demand, and easing into an economy that was not rolling over. The 1999 parallel is offered and then partly withdrawn in the same breath, which is the note's best move: the setup rhymes (easing while a tech bubble inflates) but the initial condition is the opposite — in 1999 "the U.S. government was running such large surpluses… there were concerns all federal debt would be extinguished over the next 10 to 15 years." An analogue with an inverted balance sheet is not a template, it is a contrast, and the contrast is the argument: the same monetary posture that produced a bond rally then produces a bond rout now because the fiscal starting point is reversed. Hence the vulnerability claim — long rates at "the highest levels in a generation" threatening both the Treasury market and the economy — and hence the closing tell, which links straight back to Aug-25's defended long end and Aug-26's issuance surge: Bessent's "yield-manipulation gambit" is read as confirmation that the Treasury itself shares the diagnosis. Nothing rowed: the 10-year, the Fed, deficits and the AI build-out are discussed as macro objects; no vehicle is named and no ticker is inferred (Scott Bessent and Alan Greenspan are people, not securities).

1. Key points

The claim — "unprecedented," used deliberately

The word is flagged as overused before it is used: "the word 'unprecedented' is often overused during these tumultuous times. However, when it comes to what has occurred since the Fed began its latest interest-rate easing cycle, it truly qualifies."
The whole note is one observation and its explanation — there is no portfolio action, no name, and no forecast beyond a vulnerability assessment.
Four Bloomberg charts are supplied with an explicit justification: a "four-chart sequence necessary to fully convey the point" — i.e. the pattern only reads as a pattern across multiple cycles.

The measurement — the move, and the size of the opening cut

"From the time the Fed began this easing phase, the yield on the 10-year U.S T-note has vaulted from around 3.70% to the present 4.75%" — roughly 105 basis points the wrong way while the policy rate was being cut.
The aggravating detail: "what makes it all the more remarkable is that this cutting cycle began with an emphatic 0.50% (50 basis points) reduction." A jumbo opening cut is the version of the policy that should move long yields down hardest.
Note the 4.75% level itself: it sits above the 4.6/4.7/4.8 band the archive's rates theme has been tracking as the zone where the Treasury is forced to act.

The base rate — what has always happened since 1970

"As the preceding charts make clear, going all the way back to 1970 the consistent reaction to the inception of an actual Fed rate-cutting cycle has been for longer-term Treasury yields to recede."
The word "actual" is load-bearing — it excludes one-off or emergency cuts that never became a cycle, so the comparison set is clean.
The question is then posed plainly rather than answered rhetorically: "the reasonable question is: Why is this time different?"

The counterexample, raised and then closed — 1998

The objection is granted first: "it is fair to note there have been some brief rate-cutting cycles, such as in 1998, when the Fed cut on an emergency basis while the economy was still robust."
But it was classified at the time, not in hindsight: "this was known as a 'mid-cycle' adjustment in reaction to, at that time, the Asian Crisis and the related implosion of the Long-Term Capital Management hedge fund."
And it reversed: "less than a year later, the Greenspan-led Fed was hiking again, and would do so three times" — so 1998 is not a precedent for cutting into strength and getting higher long yields; it is a precedent for the cuts themselves being withdrawn.

The diagnosis — three causes, kept plural

Fiscal: "deep, recession-like deficits during an ongoing economic expansion" — the same 6%-of-GDP figure scored in Aug-26, here converted into a term-premium explanation.
Capital demand: "the voracious capital needs of the AI build-out" — the "reverse crowding out" channel argued in Aug-27, where hyperscaler debt issuance competes with the Treasury for the same savings pool.
Cycle timing: "it could also be attributed to the reality that the Fed began this easing period when the economy was not in the early stages of a downturn" — no recession bid ever showed up to buy the long end, because there was no recession.

1999 — the analogue, and the disanalogy that matters more

The rhyme: "in that way, there are some similarities with 1999 when the first tech bubble (it's now necessary to distinguish between them) also continued inflating."
The parenthesis is doing real work — the need to number the tech bubbles is itself the observation.
The break: "a massive difference versus 26 years ago is that the U.S. government was running such large surpluses back then that there were concerns all federal debt would be extinguished over the next 10 to 15 years."
Same monetary posture, opposite fiscal starting point — which is why the analogy is offered as a contrast rather than a template. "Undoubtedly, younger people will find that nearly impossible to believe based on the present pathetic state of America's fiscal condition."

The conclusion — vulnerability, and the policy tell

"This deep fiscal hole leaves the Treasury bond market, and the economy, extremely vulnerable to longer-term rates continuing to hit the highest levels in a generation." Both objects are named — the bond market and the real economy that borrows off it.
The closing sentence is confirmation from the other side of the trade: "based on his recent yield-manipulation gambit, it's clear Treasury Secretary Scott Bessent is fully cognizant of those risks."
That connects directly to Aug-25 (the defended long end) and Aug-26 (stealth YCC, and Japan as the precedent for where the pressure goes instead).

Housekeeping

Four Bloomberg charts of 10-year yields around prior Fed cutting cycles are referenced in the original and omitted from the saved text.
No security is named anywhere in the post; the 10-year note, the Fed, the deficit and the AI build-out appear as macro objects only, and no ticker is inferred.
Signed "The Haymaker Team."

2. In plain English

When the Federal Reserve starts cutting interest rates, it is directly setting only one rate — the overnight rate banks charge each other. The rate that actually matters for mortgages, corporate borrowing and government debt costs is the 10-year Treasury yield, and that one is set by buyers and sellers in the market, not by the Fed. Historically the two have moved together at the start of an easing cycle: the Fed starts cutting, investors conclude the economy is slowing and inflation will fall, they buy long bonds, and long-term yields come down too. Hay's point is that going back to 1970, that has happened every single time a genuine cutting cycle began. He includes four charts because the claim is about a pattern, and one chart wouldn't show it.

This time the opposite happened. Since the Fed started cutting, the 10-year yield has gone up from about 3.70% to 4.75%. That is a big move in the wrong direction, and it is made stranger by the fact that the Fed opened with an unusually large half-point cut — the kind of move that should push long yields down the most, not invert the relationship entirely.

He then deals with the obvious objection before anyone raises it. Wasn't there a time the Fed cut while the economy was fine? Yes — 1998, when it cut in an emergency after the Asian financial crisis and the collapse of the hedge fund Long-Term Capital Management. But that was openly described at the time as a temporary "mid-cycle" adjustment, and within a year the Fed was raising rates again, three times. So it isn't a counterexample to the current situation; it's an example of cuts being taken back.

Why is it different now? Three reasons, and he doesn't pick just one. First, the government is borrowing at recession-sized levels — around 6% of the economy — while the economy is actually growing. Someone has to buy all those bonds, and the only way to attract enough buyers for that much supply is to offer a higher yield. Second, the AI data-center build-out is consuming enormous amounts of capital; those companies are borrowing heavily too, competing with the government for the same pool of savings. Third, the Fed started cutting when the economy wasn't heading into a downturn — so the usual flood of nervous money into safe long-term bonds never arrived.

The closest historical parallel he can find is 1999: the Fed easing while a technology bubble kept inflating (he notes wryly that we now have to number the tech bubbles). But the parallel breaks in the most important place. In 1999 the US government was running surpluses so large that serious people worried the national debt would be paid off entirely within 10 to 15 years. Today the fiscal position is the exact reverse. Same monetary setup, opposite starting balance sheet — which is precisely why the bond market is behaving in the opposite way.

The consequence is the part to hold onto. With that much debt outstanding, every additional increment of yield costs the government — and every borrower in the economy — real money. That is what he means by the Treasury market and the economy being "extremely vulnerable" to long rates staying at generational highs. And he reads the Treasury Secretary's recent attempts to hold yields down as proof that the government sees exactly the same risk: you don't intervene in a market you think is fine.

Macro viewpoint — no security named (the 10-year Treasury note, the Fed, the federal deficit and the AI build-out are discussed as macro objects; no ticker is inferred). Summary derived from the paid Haymaker Daily (text in transcript.txt) for personal study. The four Bloomberg charts referenced in the original are omitted. Not investment advice. © Haymaker / David Hay for source material.