Newsletter · 2026-09-29 · Ashutosh Agarwal

Warsh Hiked and the Long End Sold Off Anyway - The Long End & Fiscal Supply - Week of September 29, 2026

The Long End & Fiscal Supply for the week of September 29, 2026. Podcast synthesis on the long end selling off despite Kevin Warsh's hike, with the 10-year at 5.22% and the 30-year at 5.53% (highest since 2002), a 3bp tail on the 5-year auction, an undersubscribed Treasury buyback, Bank of America's Megan Swiber on the missing buyers, the France-Germany spread passing 100bp, and Vanguard and PIMCO starting to add duration.

The Long End & Fiscal Supply

Week of September 29, 2026: Warsh Hiked and the Long End Sold Off Anyway

Last week's issue ended on one question: could Kevin Warsh's hawkish rate hike hold down long-term Treasury yields, or was the dip just a breather? It took about three trading days to get an answer. A strong economic survey, a poorly received 5-year Treasury auction and a bond buyback that couldn't find enough sellers set off the biggest one-day drop in Treasury prices since last year's "Liberation Day" tariff shock. By Monday morning the 10-year Treasury yield was at 5.22% and the 30-year at 5.53%, both the highest since 2002. The podcasts have split into two camps. One says this is a structural repricing that no Fed hike or Treasury trick can stop. The other says yields at these levels are finally too good to pass up.

TL;DR

The Warsh experiment lost its first round. The hike flattened the curve for a few days, then the long end sold off anyway. A 5-year auction "tailed" by 3 basis points, the second-worst result in that auction's history, and 10-year yields jumped 14–17bp in one session. By Sep 28 the 10-year was at 5.22%, the 30-year at 5.53% and the 2-year at 4.91%. Chris Whalen's verdict: Warsh has "lost the 10-year and 30-year."
The problem is buyers, not just supply. Bank of America's Megan Swiber gave the clearest explanation of the week. Treasuries have stopped working as a portfolio hedge (they sell off together with stocks). Meanwhile top-rated AI borrowers offer about 100bp more than the 30-year Treasury. So investors want more pay to hold long government debt. Bessent's buyback made the point: he offered to buy about $6 billion and got only $4 billion.
Europe crossed the line flagged here last week. Saxo's John Hardy warned that the France–Germany 10-year spread "needs to stop around 100" basis points. This week the FT's Unhedged team said French yields are now "more than a full percentage point above German yields," and France now pays more to borrow than Italy.

(Quick definitions. The "long end" means long-maturity government bonds, mainly the 10-year and 30-year Treasury. A basis point (bp) is one-hundredth of a percentage point. "Term premium" is the extra yield investors demand for locking money up for years instead of rolling short-term bills. An auction "tail" means the Treasury had to sell its bonds at a higher yield than the market expected just before the sale, which is a sign of weak demand.)

What's new

1. The sell-off came back, and it was violent. The FT's Rob Armstrong described a 14bp move in the 10-year in one day, "at one point... 17 basis points," and compared it to a roughly 7% day in stocks. Katie Martin called it "the biggest drop in US government bonds that we've had since... Liberation Day." Ian Smith listed the "toxic cocktail." A business survey (the PMI, or purchasing managers' index) showed output growth "at its fastest in five years." A Fed official suggested it might take "quite a few interest rate increases." Oil rose again as hopes for peace talks at the UN General Assembly faded. And "a five-year U.S. government debt sale... didn't go particularly well." - Podcast: Unhedged (Financial Times), "Bond bombshell": listen (Katie Martin, Rob Armstrong, Ian Smith, FT journalists) - Podcast: Morning Call, "Markets, AI and rising yields test the outlook 9/28/26": listen (Morgan Brennan) - Why it matters: Last week's bulls had one argument that looked like it was working: a credible Fed would anchor the long end. Within days, yields broke through last week's highs. Brennan opened Monday's show with the numbers: the 10-year at 5.22% "the highest we've seen since 2002," the 30-year at 5.53%, the 2-year at 4.91%. WTI oil was up about 3.5% at around $95.50 and Brent near $108.

2. The 5-year auction was a real warning sign. On Galaxy Brains, guest Beimnet Abebe explained what made this sell-off unusual. Normally when yields jump, buyers treat an auction as a chance to "buy cheap." This time "the rate market was [bad] and the auction was bad": a 3bp tail, "the second highest tail in the five-year auction history." His conclusion: even with oil lower at points during the week and "a Fed hike that... probably regain[ed] some credibility," yields kept rising, which "points to something that's happening that is quite structural." Investors "are demanding a higher premium to own this paper," and you can see it in real (inflation-adjusted) yields "that have continued to push higher." The one thing he could name that might stop it: "a huge correction lower in the energy complex caused by a resolution of the Iran conflict." Beyond that, "I really am not sure what stops the train." His host doubted even a peace deal would be enough: "It'd have to be like the most durable, grandest [resolution]." - Podcast: Galaxy Brains, "Bitcoin Consolidates While Rates Rise with Beimnet Abebe": listen (Beimnet Abebe, market professional) - Why it matters: Last week's newsletter named auction results as the scoreboard for a possible "buyer strike." This was the first clear failure on that scoreboard. On Thoughtful Money's weekly market recap, a portfolio manager put the auction result in real terms. The 5-year cleared around 5.1% while market inflation expectations (the 5-year "breakeven") are about 2.3%, which leaves a real yield near 2.7%, which he called "very restrictive" for "five and 10-year loans, mortgages, car loans, corporate loans, pretty much all loans." Thoughtful Money

3. Bessent's buyback didn't draw enough sellers. The 20-to-30-year buyback operation flagged last week went ahead. On Forward Guidance's weekly roundup, a panelist reported that Bessent "was ready to buy 6 billion and he only bought 4 billion," and said that "just put fuel into the fire today of the bond market." A veteran futures trader on the panel (who once did business with Salomon Brothers, "the bond vigilantes") put it bluntly: "You're going to buy 10 billion, but you're going to do a half a trillion in issuance." In his view the Treasury simply doesn't "have the gunpowder," and nothing short of "full on intervention" would change things. - Podcast: Forward Guidance, "The Bond Market Pain Isn't Over | Weekly Roundup": listen (weekly roundup panel of traders, market professionals) - Why it matters: A buyback that ends up smaller than offered means holders didn't want to sell long bonds back even at the Treasury's price. That is an awkward result for a program meant to show official support. Bank of America's Swiber said it plainly on WSJ's podcast: "even as Besson has unveiled this higher amount on buybacks... yields have continued to go up at the back end," which "suggest[s] that Treasury is going to try other things." Dennis DeBusschere of 22V Research added that the Treasury market "dwarfs any program that he is willing to use." He also noted that the full package (doubling the buyback, stepping in on the yen so Japan doesn't have to sell Treasuries, floating the use of the Treasury's cash account) makes it look as if "we were in a financial crisis" to anyone watching from outside. On The Tape

4. Bank of America explains the buyer problem. The best explanation of the week came from Megan Swiber, a Bank of America interest-rate strategist and former Fed staffer, on WSJ's Take On the Week. She ranks the drivers. First is Fed expectations ("the most important component"), which have moved "drastically" from a market that thought "there's no way... they're going to be able to hike." Second is term premium. BofA's own fair-value models and the Fed's term-premium models "have all suggested that long-end is trading relatively cheap," meaning yields are higher than fundamentals justify. Why? Treasuries have lost their "diversification benefit." They used to rally when stocks fell. Now "we're not seeing treasuries perform that important diversification benefit; they're selling off as well." And then there is competition: "30-year yields above 5.2 percent... I can go out and buy an investment-grade bond from a very highly rated [borrower] that is 100 basis points over the 30-year treasury rate." - Podcast: WSJ's Take On the Week, "Making Sense of Sky-High Treasury Yields": listen (Megan Swiber, Bank of America rates strategist, desk professional) - Why it matters: This is the difference between "too much supply" and "not enough demand," and it points to what could fix things. Swiber's view: "The change... really is demand." If the Iran war ended and oil flowed freely, "we certainly can see rates fall pretty meaningfully." But US growth and inflation outside oil still say "rates are not restrictive" and "the Fed's got more work to do." She also traced Warsh's back-and-forth: hawkish in June, "less hawkish" in July, then "pivoting back to looking to fight inflation at Jackson Hole" and September.

5. France has become Europe's problem. On Unhedged, Ian Smith said French 10-year yields are now "more than a full percentage point above German yields," and France "now trading above, has higher borrowing costs than Italy." Investors "don't really believe the savings plan" to get the deficit under 5% of GDP. Some left-wing presidential hopefuls have floated "debt cancellation," and "the fact that that's even circulating... is worrying people." The UK is next in line, with a budget coming up and "difficult decisions" forced by the rise in borrowing costs since the March forecasts. - Podcast: Unhedged (FT), "Bond bombshell": listen (Ian Smith, Katie Martin, FT journalists) - Why it matters: Last week Hardy said 100bp was the level where "talk of a new EU sovereign debt crisis" begins. That line has now been crossed. The one reassurance, per Smith: there is no contagion yet. "You don't see contagion from French government... widening spreads to the rest of the eurozone. If that started to happen, I think people would think, what's the role of the ECB here?"

The debate

Both sides had real coverage this week. The bears had the price action. The bulls had some notable institutional voices starting to buy.

The bears: "structural, not cyclical, and nobody can stop it"

The Fed has lost the long end. Asked whether Warsh has "already lost the 10-year and the 30-year," Chris Whalen said "I think so. Unless the Fed is willing to come in and start directly purchasing securities in size, there's not much they can do. The Treasury repurchases are way too small to be significant." In his view, the rise in long rates is "about the credibility of the United States and also the noise that's coming from the Trump administration... it has nothing to do with the Fed." (Chris Whalen, Whalen Global Advisors: analyst/commentator, opinionated but market-literate.) The Julia La Roche Show
There's too much debt to place at a good price. Sam Valverde is a former Treasury debt-management official who worked through the 2011 and 2013 debt-limit standoffs and later served as acting president of Ginnie Mae. He framed it cleanly: "not that we will have failed auctions, but... we will clear those auctions at a different price... rates will continue to increase because of the relative oversupply." He called AI borrowers "rate insensitive": "5%, 6%, 7%, 8% is not going to dissuade them," and "institutional capital... only has so much allocation to give." And with Social Security "severely impaired over the next three or four years" and little political will to act, "bond investors are looking for some indications that the fiscal outlook is going to change." (Sam Valverde, Falcon Capital Advisors, ex-Treasury: operator/insider.) Chrisman Commentary
The debt numbers have outgrown the old playbook. Reuters' Dan made the point that Warsh himself missed. At his press conference Warsh blamed long yields on economic strength, AI-driven "competition for capital," and geopolitics, but "left at least one big one off the table... debt." There is $40 trillion of total US debt, $32 trillion of it in the Treasury market, up from under $20 trillion before COVID, with deficits running "between 5 and 6 percent" of GDP (5.7% last year). Reuters Econ World
The economy is running too hot for 5% to be enough. DeBusschere estimates the US can only grow about 2% a year without inflation (a "speed limit"). With nominal GDP at "6.5% and doesn't appear to be slowing," strong data now means "tighter financial conditions. And higher 10 year yields. You will not see higher equity prices." His less obvious point, which bulls can use too: "higher rates actually lowers 10-year term premium," so real Fed tightening is the tool that works, not buybacks. (Dennis DeBusschere, 22V Research: professional strategist.) On The Tape
The futures market has priced out rate cuts for years. The Forward Guidance trader noted that the SOFR futures curve (bets on future short-term rates) shows "not a cut priced until December of '29." His favourite gauge, the December 2028 contract, "was trading 96.90" at the start of the war and is now around 95.17, which implies an expected rate about 1.7 percentage points higher. His expectation: "PCE is going to come in hot the next two months." The panel noted the SOFR curve is "pricing on a lot more hikes than the recent dots." Forward Guidance
Short-term funding is building up refinancing risk. Unhedged's Ian Smith said Treasury bills (debt maturing within a year) are expected to reach "almost a quarter" of the US debt stock in the next couple of years, a level only passed "during COVID and the financial crisis," with around $1 trillion of net bill issuance expected over the coming year. Rob Armstrong: "If you shorten your maturity stack into a rising rate environment, you're just storing up trouble for the future." Unhedged
The commentator bears. Peter Schiff (a longtime bond bear) cited a 30-year at 5.41% (a 22-year high) and a 10-year at 5.12% midweek, and argued the 30-year could reach "well north of 6%, maybe even 7%." His arithmetic: 5% on a $40 trillion debt is "$2 trillion a year... 35% of current tax revenue," more than Social Security. The Peter Schiff Show Michael Pento argues rising long rates will be "the pin that pops everything." His list: $1.6 trillion of private credit, $1.4 trillion of CLOs, $1.5 trillion of junk bonds, and "AI is raising $1 trillion a year in debt." He also says the zero-rate "global anchors" in Japanese and German bonds are gone, which weakens the yen carry trade. (Pento Portfolio Strategies: commentator; he forecasts a very difficult 2027.) The Julia La Roche Show

The bulls: "5%+ is finally worth owning, and the long end is starting to price a slowdown"

Vanguard is telling clients to add duration. On Monday's Morning Call, Vanguard's Joyce Wong said "we aren't terribly concerned about this recent rise in rates. We think that it is likely to stay rather contained." Her case is that "the economy is strong. Jobs are good. Rates probably should be at this level," and investors are badly positioned. "Our studies at Vanguard have shown that investors are very short duration, mostly concentrated... in cash, money markets." Her advice: "think about adding a little bit of duration." (Joyce Wong, Vanguard: buy-side professional.) Morning Call
PIMCO is cutting its bet against long bonds. Jeff Snider reported that PIMCO "has begun trimming its underweight position in long-term U.S. government bonds as yields above 5% create what the firm considers better value." Its CIO said investors with an intermediate time horizon "can now find good value." Snider's caveat: "That isn't a declaration that yields have peaked." Eurodollar University
The curve is sending a slowdown signal. Snider's main argument, recorded just before the midweek sell-off, was that the gap between 2-year and 10-year yields had narrowed to about 20bp, "the narrowest spread in roughly a year and a half." That happened because the 10-year fell, not because the 2-year rose. "The front of the curve sees the Fed. The back of the curve, that sees demand." Oil works "like a tax," so it can push short rates up (through the Fed) while pulling long rates down (through weaker future demand). His supporting evidence: yields on the weakest junk bonds (CCC-rated) "approaching 1,100 basis points" over Treasuries, and AI project financing "becoming more expensive and harder to distribute." Liz Thomas (RiskReversal Pod) cited the 2s10s at 18–19bp and has bought the 10-year Treasury herself. (Jeff Snider, Eurodollar University: commentator/macro analyst.)
The Fed may have to ease off sooner than priced. Liz Thomas's view: if oil stays high, "the Fed's going to keep hiking, full stop... a very quick way to have a recession." Her base case is that "the Fed is going to have to tap the brakes on this hiking cycle," especially "if and when the war deescalates." Whalen, a bear on long rates, also makes a contrarian call for cuts "sooner than anyone realizes." (Liz Thomas: market strategist.) RiskReversal Pod
Even skeptics want to lock in these yields. The veteran trader on Forward Guidance, bearish on where yields go, still said "if you can get... above six on a 30 year... or five and a half on a 10 year... I'd throw a couple mil at that all day long... it's risk-free." He also flagged TIPS (inflation-protected Treasuries) with real yields around 2.83–2.84%, "not that far away from a three," as "extremely interesting." Forward Guidance

The read: The bears won the week. The bulls who are actually putting money in (Vanguard, PIMCO) are buying on value and time horizon, not because they think yields have peaked. Nobody on the bull side is calling a top. They are arguing that 5%+ now pays you to wait. The one bull catalyst every podcast agrees on is an end to the Iran war and a drop in oil. Until that happens, the bears have momentum.

Trades & positioning in play

Add duration gradually at 5%+. Vanguard's Wong suggests "adding a little bit of duration," and PIMCO is trimming its long-end underweight (via Snider). Both are value-based entries with patience built in, not calls that yields have peaked. Morning Call · Eurodollar University
Stay short. The money managers on McNamaraOnMoney are sticking to 4–5-year maturities rather than buying the 30-year at "almost five and a half percent." They cite inflation "around three and a half percent," Q3 GDP "tracking for 5%," and a US "trust issue with the rest of the world" given "a $2 trillion per year annualized deficit." McNamaraOnMoney
TIPS near 3% real. The Forward Guidance panel pointed to real yields around 2.83–2.84% as close to an attractive round number. Forward Guidance
Curve flattener / inversion watch. Snider's trade is a curve that flattens because the long end rallies, with inversion "within the market's grasp." The midweek sell-off pushed against it, but the 2-year at 4.91% against the 10-year at 5.22% still leaves only about 31bp between them. Eurodollar University

Read-throughs

Mortgages and housing. Whalen's rule of thumb: "take basically the 10-year treasury and add two points," which puts a standard 80% loan "well over 7%." He expects "a lot of consolidation" and job losses in the mortgage industry ahead of the MBA annual meeting in Chicago. The Julia La Roche Show BofA's Swiber explains why the Fed has to push harder than usual: with so many fixed-rate mortgages, "there's not this total pass-through to many consumers," so tightening works mainly "through financial conditions" and especially stock prices. WSJ's Take On the Week
Politics and the midterms. Raymond James' Ed Mills (Washington policy analyst) said that with the midterms "less than a month and a half away," mortgages "north of 7%" and a Fed hiking against the president's wishes "create political problems for Republicans." He added that "if you didn't have Treasury Secretary Besant trying to control the long end of the curve, and it's seemingly as if it's going against him, this wouldn't be as much of a political problem." Morning Call
Stocks. So far stocks are holding up. Carson Group's Ryan Detrick noted that tech "took back the baton" in what is historically one of the year's worst weeks, and that October and November are the best two months of a midterm year. The Unhedged team said the most common listener question was "when does this nightmare in the bond market spill over into stock markets?" Liz Thomas warned that a hiking cycle looks fine 12 months out, but "What happens 16 months out? What happens 18 months out?"
Credit and AI borrowing. This week the AI-debt story and the long-end story became the same story. Swiber puts top-rated corporate bonds at about 100bp over the 30-year. Valverde calls AI issuers "rate insensitive." Warsh himself listed "competition for capital" from hyperscaler debt as a driver (per Reuters). Snider sees the flip side: AI financing "becoming more expensive and harder to distribute," and warns the AI build-out "can turn from a reflationary force into a credit tightening mechanism."
Gold. Gold broke below $4,250 an ounce "for the first time in six weeks," at around $4,183 on Monday, and miners fell with it. Morning Call Views differ on why. Liz Thomas says higher yields are a headwind for a metal that pays nothing, and that "central banks had to start selling gold in order to raise cash because of these higher oil prices." She still added to gold as a fear hedge. RiskReversal Pod The guest on Money Metals' Weekly Market Wrap argues the opposite: central banks had "their largest quarter on record in the second quarter," and the People's Bank of China is sitting "at all-time lows of treasury bonds" while buying gold. Money Metals' Weekly Market Wrap Whalen's commentator view: "When people sell treasury securities, they're buying gold."
Oil and diesel. Oil is still the biggest single driver of bonds in both directions. Diesel at "nine bucks a gallon" in some places and about $7 in Chicago (Forward Guidance) feeds into food, freight and fertilizer. Galaxy Brains flagged a planned trucker strike on October 1 over diesel costs.
Europe and the UK. France–Germany is above 100bp and France now pays more than Italy. The UK budget will show the damage from higher borrowing costs since the March forecasts. The Unhedged hosts also noted that the US now spends more on interest than on defence, and that Congress's budget forecasts assume yields "much closer to 4% than 5%."

What changed vs last week

The Fed-credibility bull case took its first real hit. Last week: a hawkish hike, a twist flattener, and the 10-year back under 5%. This week: the 10-year at 5.22% and the 30-year at 5.53%, the highest since 2002, with the Fed hike already in the price. Last week's newsletter noted that the 30-year had "touched 5.34%." It is now about 20bp above that.
The auction scoreboard turned bad. A 3bp tail on the 5-year, the second-worst in that auction's history. Swiber and Valverde both describe it as a demand problem. Auctions will still clear, but "at a different price."
The buyback went from "ineffective" to "undersubscribed." Last week's verdict was that it was too small. This week, sellers filled only about $4 billion of a roughly $6 billion offer.
OAT–Bund went past 100bp. Hardy's crisis-talk threshold has been crossed, and France now trades wider than Italy.
Institutional bulls began buying. PIMCO trimming its underweight and Vanguard recommending duration mark the first real move by large asset managers toward long bonds in this sell-off.

Next week's watch-list: whether the 10-year holds above 5.2% and the 30-year above 5.5%; the next auction results; hot PCE inflation data (the Forward Guidance trader expects it); any Iran ceasefire talks and the oil price; October Fed-hike odds; France's budget credibility and the UK budget; the October 1 trucker strike.