Dispatch Sunday Set-up

The Market Without a Balance Sheet.

Friday the rout went global and every market it touched found a bid somewhere. Semis fell 7%, Korea 5.5%, Taiwan 4.6%, the Nikkei 3%, the S&P 1.8%; gold fell 3.6% and silver 5.8% alongside the semis, which is deleveraging, not a flight. High yield widened twelve, CCC eighteen. Twos closed 4.598, nine lower on the week. Every one of those prices got made because somebody was paid to make it: a dealer running inventory, an ETF arbitrageur, a swap desk, a macro fund with a mandate to buy what the junior tranche is selling. TLT traded 41 million shares Friday. The system with balance sheets behind it absorbed a bad day and printed closes.

There is a $4.4 trillion market in this country that no longer has that. Dealers hold 0.4% of it. The banks and insurers who used to take the other side of a fund's redemption have cut their holdings by a third to a half. What replaced them owns what its shareholders own and has to sell what it is handed. It has been asked to find a bid exactly once since 2008, in March 2020, and the answer was thirteen points in nine sessions with the Fed building the bid. Nothing about that structure has been fixed. The flow side has tripled. This Set-Up is about the market without a balance sheet, because the week showed what the ones with balance sheets do under stress, and the next test runs on a different ledger.

The Dispatch publishes every trading morning. Subscribe on LinkedIn or at dispatch.thebondbro.com (https://dispatch.thebondbro.com).

———————————————————————————

THE BOARD

THE BOARD · Friday Oct 30, 2026 close · Koyfin, Freddie Mac · green = yield down / spread tighter / price up. Treasury one-day changes against Thursday's close as recorded Friday morning, one-week against October 23; credit is Friday's close against Thursday's, approximate; ETFs Friday's close.

The curve on the week. Twos 4.598, fives 4.727, tens 4.934, thirties 5.310. Nine lower at the front, nine in the belly, seven in tens and one at the back from a week ago: a bull steepener after two bull flatteners, and the first week since the long end's October high that the rally did not come from the back. 2s10s 34, 5s30s 58, 10s30s 38, 2s30s 71, eight steeper on the week. One note on Friday's marks: the Treasury ETFs closed better than the 3 p.m. yield fixes show, IEF +0.56%, SHY +0.15%, TLT +0.43%, which is month-end extension buying into the 4 p.m. close. Call the curve two to five basis points better bid than the printed one-day changes on the day, and the shape the same.

The Six Tracker, week five close. Thirties 5.310, −1 on the week, the second weekly close under 5.40 and under it every session since October 22. Freddie Mac 30-year 7.24%, −8, the second print to carry the rally. 10s30s 38, +5. Dollar-yen 157.54, +0.5%. Three of four moved the right way; the one that did not is the one the refunding reads.

Credit, Friday close. IG 89 (+3), AAA 46 (+5), BBB 110 (+3). HY 326 (+12), BB 194 (+10), B 332 (+14), CCC 1325 (+18). Treat the one-day changes as approximate; the feed's timing on credit is loose. What is not loose: HY closed inside 325 Wednesday and Thursday and closed outside it Friday, by one. The 350 score line is 24 away. Euro HY 348, flat; EM HY 374 (+8). HYG −0.19%, CWB −1.92% at z −1.7, PFF −0.80%.

Elsewhere. France 10Y 4.972 (+6) back over the US 4.934 for the second time in two weeks, Bunds 3.324 (−3), OAT–Bund 165. France 30Y 5.466, sixteen over the US long bond. JGBs −6 to −7 across the curve. Korea's 10-year −9, its 30-year −11, Hong Kong −10: Asia's bonds were bought as its stocks were sold. WTI 103.08 (+1.1%), Brent 112.57 (+1.6%), both roughly flat on the week; heating oil +120% year-to-date. Gold 3,877 (−3.6%), silver 49.12 (−5.8%), copper −1.8%. Low volatility +1.6% against the S&P at z +3.0; momentum −3.1%; energy +1.9% and staples +1.4% were the only sectors bought.

———————————————————————————

1. THE LEDGER

Start with who owns the market, because that is the whole argument. The Fed's flow of funds puts municipal securities outstanding at $4.36 trillion at the end of the third quarter of 2025, up two-thirds from 2007. The holders have changed more than the size.

Dealers held $50 billion of munis in 2007. They hold $18 billion now, 0.4% of the market, down from about 1% in 2010 and 2% before the crisis. The low was $7 billion in 2020. Property and casualty insurers, the classic relative-value buyer of long tax-exempts, held $382 billion at the 2008 peak and $207 billion now, with a step down in 2018 when the corporate rate cut took the exemption's value from 35 cents to 21. Those two left after 2007, as the thesis has it.

Banks did not. The ledger says they came in. Bank holdings doubled and more from $204 billion in 2007 to a $630 billion peak in 2021, on yield and on the liquidity rules that made munis a reasonable place to park deposits; they have fallen 42% since, to $367 billion, on unrealized losses, deposit outflows and the same 2018 rate cut. So the "since 2008" line is right for dealers and insurers and wrong for banks by thirteen years. The Dispatch prints the ledger as it reads. The conclusion survives the correction: dealers, banks and P&C together held about 21% of all muni holdings in 2021 and about 14% at the last print. That is a four-year, one-third reduction in the share of the market held by balance sheets that buy because something is cheap.

What replaced them is flow. Mutual funds went from $372 billion to $828 billion. ETFs went from $600 million to $163 billion, 3.4 times what they were going into 2020. Households, the largest line, hold about $2 trillion directly, much of it through separately managed accounts that behave like funds on the way out. A fund owns what its shareholders own. When they redeem, it sells, at whatever price the remaining bid will pay, and the remaining bid is now mostly other funds. That is the market without a balance sheet: long-only, flow-priced, no one paid to take the other side when the flow reverses.

———————————————————————————

2. THE LAST TEST

The structure has been tested once. From March 6 to March 19, 2020, the national muni ETF fell 13.1%. The 7–10 year Treasury ETF, which carries about the same duration, fell 3.5%. The 20-year-plus Treasury ETF, with three times the duration, fell 11.1%. A six-duration asset fell as much as an eighteen-duration one, and more than the Treasury it is supposed to track, because the selling was redemptions and the buying was nobody. Muni funds saw record outflows in those weeks. MUB traded at a discount to its net asset value of several percent, which is the ETF balance sheet telling you it has run out. Dealers held $16 billion going in and were not going to take a $4 trillion market's redemption wave onto their books at any price.

The bid that ended it was built, not found. March 20 the Fed opened the money-market facility to muni paper; March 23 it added variable-rate demand notes and muni commercial paper; April 9 it announced a $500 billion Municipal Liquidity Facility. From the March 19 low to April 30 MUB returned 9.8%, IEF 4.8%. The round trip looks like a V on a chart and a liquidity event on a desk: the market needed a balance sheet, had none, and borrowed the Fed's.

Here is the ledger going into that test and going into the next one. March 2020: dealers $16 billion, banks $508 billion, P&C $285 billion, mutual funds $831 billion, ETFs $48 billion. Now: dealers $18 billion, banks $367 billion, P&C $207 billion, mutual funds $828 billion, ETFs $163 billion. The dealer line is the same. The two balance sheets that could absorb a redemption wave are each about a third smaller. The ETF sleeve, which transmits a redemption to the cash market in seconds, is 3.4 times larger. The muni market has not had a major credit event since Detroit in 2013, outside Puerto Rico, and it has not had a liquidity event since the Fed stood one down. Both are untested at these weights.

———————————————————————————

3. THE LAG, MEASURED

The practitioner's complaint about this market is that it lags Treasuries mechanically: the buyers who would arbitrage the ratio are gone, so munis follow the rate move late and partially, then overshoot when the flow turns. Ten years of daily data say that is right, with one amendment.

From October 2016 through Friday, 2,513 sessions, MUB moved 0.37 for every 1.00 in IEF on the same day, with a daily correlation of 0.49. Over a week the beta is 0.46; over a month, 0.59, with the correlation at 0.76. Regress the muni move on the Treasury move and its five prior days and the lags add 0.20 to a 0.38 same-day coefficient: about a third of the muni response to a Treasury move arrives in the following five sessions. In 2020–21 the same-day beta fell to 0.22 and the lagged terms carried more than the same-day one; the market was pricing flow, not rates. Up and down capture are symmetric at about 0.4 on the biggest Treasury days. On total return over the ten years MUB is +20%, IEF +6%, TLT −17%, which is the tax-exempt carry and the shorter duration, and is why nobody who owns it for income has complained.

The amendment: since 2024 the same-day beta has risen to about 0.5 and the lagged terms have gone to zero. The rolling 60-day beta is 0.61 now and peaked at 0.72 in the April 2025 tariff week. Munis now move with Treasuries on the day, at half the size. The reason is in Exhibit 1: the ETF sleeve is large enough that its arbitrage sets the cash price intraday. The ETF replaced the dealer as the price-setter. It is a balance sheet that works in one direction. It can make the market efficient on a normal day, and in March 2020 it was the thing trading at a discount. Friday's print, MUB +0.31% against IEF +0.56%, a capture of 0.55 on 85% of the duration, is the beta, not news. The news is what the beta does when the flow is one-way.

———————————————————————————

4. THE WEEK THAT HAD A BALANCE SHEET

Now hold the muni ledger next to the week. Friday was the worst session for the AI complex since the build started trading as a credit: semis −7.0%, software −3.4%, internet −3.4%, technology −4.9%, Korea −5.5%, Taiwan −4.6%, the Nikkei −3.0%, Nasdaq −2.9%. Momentum lost 3.1% to the index, IPOs 4.6%. Gold and silver sold with it, which is what a margin call looks like, not a hedge. Credit widened twelve in HY and eighteen in CCC, the first widening since the Fed and the one that took both Gauge step-downs off the table.

And the Treasury market did its job. The front end was bought all week, twos −9, SHY and IEI at z +1.7 and +2.0 Friday. The long end was left alone: thirties −1 on the week, +1 on the day by the fixes, TLT at z +0.7 against IEI at +2.0. The flight stopped at tens. That is the term-premium read the Gauge has carried since October 18: the long bond does not get bought as a hedge against the junior tranche, because the long bond is where the supply is. The refunding on Wednesday is the first day the market says whether the supply has changed. France made the same point from the other side: OATs +6 while Bunds −3, the 10-year back over the US at 4.972 against 4.934, and a 30-year at 5.466 that is sixteen basis points over the US long bond. Germany is the senior tranche; France is still being priced as the junior one, and a week of bought Bunds did not change which side of the line it sits on.

Every one of those prices was made by a market with a balance sheet. The Treasury market has primary dealers obligated to bid at auction, a futures basis with $1 trillion of leveraged money in it, swap desks, and ETFs that can create and redeem against the deepest cash market in the world. The equity market has the same. The muni market has $18 billion of dealer net positions against $4.4 trillion, a bank bid that is a third smaller than its peak and an insurance bid that is nearly half. On the week when the market with balance sheets absorbed a −7% semis day and printed a nine-basis-point rally in twos, the one without was +0.31% on Friday and was not asked anything.

———————————————————————————

5. THIS WEEK

Monday Nov 2 — ISM manufacturing. First read on Friday's credit widening: HY from 326, CCC from 1325. Apple and Palantir report after the close; the junior tranche's weekend.

Tuesday Nov 3 — Midterms. The Dispatch has called this a credit event since October 18, because the build's junior tranche is priced off permission duration, permits, interconnects, rate cases, local votes, and both parties arrive with a mandate to slow it. It is also a muni event: the ballots carry state and local bond authorizations, and the tax exemption goes back on the table in whatever tax bill the next Congress writes. RBA overnight.

Wednesday Nov 4 — The refunding, 8:30. Coupon sizes, the buyback schedule and the language on "at least the next several quarters" are the three things the long end has been waiting since September to read. ISM services. The first session the Term Premium dial can earn its step-down on supply rather than on a hedge.

Thursday Nov 5 — Bank of England. Jobless claims. Freddie Mac mortgage rate, the third print to carry the rally or not.

Friday Nov 6 — October payrolls. The Term Premium re-score at the close: under 5.40 with HY inside 325, after the refunding, is the step-down. The Credit dial needs HY back inside 325 and a close that stays there.

Then — November 27, S&P on France. The Anthropic S-1, which was due by late October under the fifteen-day rule for a mid-November listing and has not appeared on EDGAR as of Sunday.

———————————————————————————

THE PRESSURE GAUGE

Front End / Policy: GROWING, held. Twos 4.598, nine lower on the week, fifteen inside the 4.75 dial and twenty-five from the 4.85 reversion. The front end caught the flight all week. December is still priced after the hawkish hold. Payrolls Friday are the next test.

Term Premium / Long End: PRESSURE, held by one basis point. The October 18 rule is two Friday closes under 5.40 with HY inside 325. Thirties closed 5.310, the second close under 5.40. HY closed 326, after +12 on the day. One basis point outside is outside; the rule does not round, and it should not, because the tape that put it there was a credit widening into a global equity rout, which is the opposite of what the step-down is meant to recognize. On the week thirties fell one while twos fell nine. The dial steps down on a November 6 close under 5.40 with HY inside 325, after the refunding has printed.

Credit: PRESSURE, held. HY 326 (+12), back outside the trigger after two closes inside it; CCC 1325 (+18), the three-session tightening run reversed. B 332, BB 194. Two closes inside and one outside is not a reversion. The 350 score line is 24 away.

Inflation, Growth / Labor, Cross-Asset: PRESSURE. September PCE 3.6% headline, 3.1% core. Crude +1–2% Friday, heating oil +120% on the year. Gold and silver sold with the semis. The rout went global; Treasury ETFs were the only thing bid, every bucket between z +0.7 and +2.0.

Composite: PRESSURE. Two dials sat at their step-down conditions Thursday night. Friday's rout took both away, Term Premium by one basis point on HY and Credit by twelve. The rule held because the tape went the other way. Reverts to ELEVATED on CCC inside 1125 and EMB inside plus or minus one standard deviation for two sessions. Neither is close.

———————————————————————————

THE KICKER

Friday showed what a market with a balance sheet does: it takes a −7% day in its most crowded trade, widens credit twelve, buys the front end nine on the week and prints closes. The $4.4 trillion market without one was not asked, and has not been asked since the Fed answered for it in 2020. The dealers who would have answered hold 0.4% of it. The banks and insurers hold a third less than they did. The funds and ETFs that replaced them hold what their shareholders hold and will sell it the day their shareholders say so, into a bid made of other funds. Ten years of data say munis follow Treasuries at half the size and a third of it late; one test says that when the flow turns, the beta is thirteen points in nine sessions. Watch three numbers into the refunding: thirties against 5.40, HY against 325, and MUB against IEF on the next bad day. The first two tell you whether the market with a balance sheet is done repricing. The third tells you whether the one without has been asked yet.

The Vault version, The Market Without a Balance Sheet (https://dispatch.thebondbro.com/upgrade), carries the full holder ledger back to 2006, the ten-year lag study with the regression output, the redemption-wave arithmetic at today's holder weights, and the Excel behind all three. It goes into the Vault with this edition.

Earlier this week: Sorted by Seniority (https://dispatch.thebondbro.com/p/dispatch-morning-brief-fc49) (Oct 30) · The Medicine (https://dispatch.thebondbro.com/p/dispatch-morning-brief-4e76) (Oct 29) · Wrong Culprit (https://dispatch.thebondbro.com/p/dispatch-morning-brief-8bcf) (Oct 28) · Eighty Billion, Minus Eighty-One (https://dispatch.thebondbro.com/p/dispatch-morning-brief-62fd) (Oct 27) · Parabolic and Zero (https://dispatch.thebondbro.com/p/dispatch-sunday-set-up-d3b9) (Oct 25, last Sunday's Set-Up) · The Senior Bid (https://dispatch.thebondbro.com/p/sunday-set-up) (Oct 18, where the Term Premium rule and the midterms-as-credit-event call were set)

The Dispatch runs every trading morning at dispatch.thebondbro.com (https://dispatch.thebondbro.com). The Vault (https://dispatch.thebondbro.com/upgrade) takes the same tape a step further: the analysis behind the levels, and the Toolkit (https://dispatch.thebondbro.com/p/vault-toolkit-v1), three Excel calculators that run on your own holdings, brackets and cash.

———————————————————————————

Sources: Koyfin global yields, yield curves, corporate credit OAS, commodities, FX, FI ETF, US sector, factor, country, world-indices and world-economics dashboards, Friday Oct 30, 2026 close (exported Sunday Nov 1, 02:34 ET); Treasury one-day changes are Dispatch calculations against Thursday's close as recorded in the Oct 30 edition, one-week against the Oct 23 close as recorded in the Oct 25 Set-Up; credit one-day changes likewise, and approximate. Koyfin daily price history, MUB, IEF and TLT, Oct 31, 2016 – Oct 30, 2026, dividend-adjusted closes; lag-study statistics are Dispatch calculations on log total returns. Federal Reserve Financial Accounts of the United States (Z.1), table L.212 Municipal Securities (L.211 before 2013), releases of Jun 9, 2011, Sep 16, 2016, Sep 20, 2019, Sep 9, 2022 and Jan 9, 2026; pre-2010 figures from the earlier vintage are approximately comparable for the sectors cited and not for households or totals. Federal Reserve press releases, Mar 20, Mar 23 and Apr 9, 2020 (MMLF, CPFF, MLF). Freddie Mac Primary Mortgage Market Survey, Oct 29, 2026 (30-year 7.24%). BEA, September 2026 PCE. Week-ahead calendar per Treasury, BLS, ISM, Bank of England and RBA schedules as reported. Anthropic S-1 status per EDGAR, Nov 1. Prior Dispatch editions as linked.

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.

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.