Regime
The Committee Held. The Long End Did Not.
FO BRIEF · THE FOMC, SCORED
The Committee Held. The Long End Did Not.
The Federal Reserve left rates unchanged on a 10 to 2 vote, with two members voting for a rise and a chair who would not guide December. The bond market then sent two different messages in two days: long yields rose on the decision and fell the next day as technology shares sold off. Here is what changed, what did not and how the desk’s published calls scored.
30 October 2026
|
| The five things that matter |
| 1. The Fed has not ruled out further tightening. It held its policy rate, the funds range, at 3.75 to 4.00 per cent on a 10 to 2 vote, but two members explicitly wanted a rise and the chair refused to say what happens in December. That means the data between now and December decide the next move, not a promise from the committee. 2. Bonds sent conflicting signals. Long-term yields rose after the Fed held, which is consistent with the desk’s view that the government’s supply of bonds is the central influence on long-term rates. The next day they fell as technology shares sold off: investors are returning to government bonds for protection, for the second time in eight days. That makes the direction of long-term yields less certain than the desk’s framework alone would suggest. 3. One warning sign for weaker growth has appeared, but the desk’s test is not met. The two-year yield closed through 4.65, the first half of the desk’s front-end test. But the futures market still prices more than one further rise, the second half, and the separate credit test is further away than it was. The outlook has not changed; the watch has sharpened. 4. Oil is expensive, but longer-term inflation expectations remain contained. Brent has risen sharply, about twelve dollars from 9 October to its 23 October peak and about ten on the scored settle, yet the bond market’s inflation gauge remains inside the desk’s forecast band. Refined fuels have not confirmed the broader energy shock the desk’s framework tests for, so the desk still treats the move as a supply story rather than an inflation signal. 5. The dollar’s pullback came and went. The desk said on 15 October that the dollar’s run was crowded and due a pullback. It arrived, fell short of the zone the desk named and was given back within a week. That call is scored partial, and the desk does not treat the episode as a confirmed reversal of the dollar’s broader direction. The desk’s conclusion. The economy is strong, inflation is still a problem and the committee is divided and will not guide. The desk’s scenario weights stay at 45 on the supply path, 35 on the growth path and 20 on the energy shock path because the tests that would move them did not fire, not because the evidence is comfortable. Two of its own calls are scored missed below, at full weight. None of this is trade or investment advice; the desk cannot give it, and every trading decision is the reader’s own. |
What happened
Wednesday: the committee held, and two members wanted more. The FOMC, the Fed’s rate-setting committee, kept the funds range at 3.75 to 4.00 per cent on a 10 to 2 vote, with Lisa Cook and Beth Hammack voting to raise it by a quarter point. The statement upgraded its view of activity to “expanding at a strong pace”, kept “inflation remains elevated” and said “The Committee will deliver price stability”. It said nothing about long-term yields (Federal Reserve, 28 October 2026; the upgrade as the FT characterised it, 28 October 2026). In the press conference the chair said inflation is too high and the committee has shown its readiness to act; he does not believe in forward guidance, telling the market in advance where rates go, outside a crisis and gave no signal on December (newswires, 28 October 2026). The Treasury Secretary said a further rise would “not necessarily” be a mistake (CNBC, 28 October 2026).
Two days, two messages from the bond market. On Wednesday short-term yields eased and long-term yields rose; on Thursday bonds of every maturity rallied as Alphabet closed 12.6 per cent lower on its results, ServiceNow and Fortinet fell 27 and 20 per cent and the Nasdaq lost 1.4 per cent (Treasury par curve, 27 to 29 October 2026; desk feed, daily closes, 29 October 2026; newswires, 29 October 2026).
| US Treasury yield, per cent | Tue 27 Oct | Wed 28 Oct, Fed day | Thu 29 Oct | Wed / Thu change, basis points |
| Two-year | 4.70 | 4.67 | 4.61 | minus 3 / minus 6 |
| Ten-year | 4.99 | 5.05 | 4.98 | plus 6 / minus 7 |
| Thirty-year | 5.30 | 5.37 | 5.32 | plus 7 / minus 5 |
| Treasury par yield curve, official settles. A basis point is one hundredth of a percentage point. The desk’s published lines: 4.65 on the two-year, 5.30 to 5.80 on the thirty-year. | ||||
What the futures and credit markets said. The fed funds futures, the contracts that price where the committee’s rate will be, barely moved: nothing priced for November, about three fifths of a rise by December, about 1.6 rises by March on Wednesday and about 1.4 on Thursday (desk computation from the contracts’ closes via the desk feed, 28 and 29 October 2026, on full-month averages against an effective rate of 3.88; weighting December to its meeting date reads nearer nine tenths). High-yield credit spreads, the extra yield investors demand to lend to riskier companies and one of the desk’s main indicators of economic stress, printed 3.14 on Wednesday, their fourth narrower print in a row (ICE BofA via FRED, 28 October 2026). The ten-year breakeven, the bond market’s inflation gauge, closed Thursday at 2.35 (FRED, 29 October 2026).
The rest of the week, and Friday so far. Third-quarter GDP grew at a 3.4 per cent pace against 3.2 expected and September PCE inflation, the Fed’s preferred measure, printed 3.6 per cent, in line (BEA; consensus via newswires, 29 October 2026). Abroad, the week leaned the same way: the Bank of Canada raised to 2.50 per cent; the Reserve Bank of New Zealand held with rises flagged; the European Central Bank held at 2.50 per cent with a rise debated and a December move reported as close to certain; on Friday morning the Bank of Japan held at 1.25 per cent with one vote for a rise (newswires, 28 to 30 October 2026; the detail sits in What to watch). Everything this note scores is marked at the 29 October settle. Friday is live and is not scored: at 07:24 GMT the ten-year traded near 4.93 and the two-year 4.59, seven-week lows, with Brent near 117 into the December contract’s expiry and gold under 4,000 (newswires via the desk’s feed, 30 October 2026).
The desk published the tests in advance. Here is how they scored.
Every level below was in print before the meeting, and every verdict is marked at the 29 October settle, the official closing level. The hits come first, then the partial, then the conditions that were tested and not triggered, then the misses at full weight, then the desk’s own lateness. The full evidence for each row, the registered wording, the prints and the sources, sits beneath the table.
| The call, and where it was published | Verdict | Why |
| Holders queue to sell long bonds back to the Treasury at its enlarged buybacks, ten to twenty year sector · 20 Aug, retest registered 5 and 12 Oct | Held | $40.8bn offered against the $6bn the Treasury could buy on 15 October, after $46.4bn on 1 October |
| Inflation expectations stay anchored under 2.50 through the oil move · 6 Oct | Held | Ten-year breakeven 2.35, inside the 2.25 to 2.45 band, with Brent up twelve dollars at its peak |
| The dollar’s run is crowded and due a pullback: tactical stance to reduce, the near-term call, with the longer-run view unchanged · 15 Oct | Partial | The pullback came, fell short of the 1.13 to 1.14 zone and was given back; the euro is back at 1.1156 |
| The growth-scare scenario becomes the base case: two-year through 4.65 with futures pricing under one further rise, or credit through 3.50 on four settles · 6 Oct, 15 Oct | Not met | Two-year 4.61, through; futures still price about 1.4 rises; credit 3.14 and narrowing. The front-end route is half met, the credit route is not |
| The energy-shock scenario triggers: Brent through 110 with refined products leading and the refining margin above 70, or breakeven through 2.50 with the thirty-year through 5.80 · 6 Oct | Not met | Brent 114.9 but products did not lead; the margin is near 49, not 70 |
| The thirty-year trades a 5.30 to 5.80 range; a close through either edge is a miss the day it prints · 6 Oct | Missed | Settled 5.29 on 22 October, one basis point through the floor; inside the range since |
| The queue of sellers runs across the whole long end, the longest-dated bonds · 12 Oct, reads registered 5 and 6 Oct | Missed | The twenty to thirty year sector offered $10.7bn on 27 October, the single-sector reading the desk registered |
| The desk’s own delivery: the fortnight scorecard owed on the 23 October settle; the refresh of the Record, the desk’s published scorecard, owed 19 October | Late | Four sessions on the scorecard; the Record refresh, owed 19 October, was published on 29 October. Logged outside the call count |
View the full scoring evidence and sources for every row
| Ten Dollars Offered. One Accepted. · 20 Aug, Brief, retired 29 Sep, reinstated 5 Oct, retest read registered 12 Oct · HELD, the buyback queue in the ten to twenty year sector The call: holders are queueing to sell long bonds back to the Treasury at its enlarged buybacks, and the Treasury buys the maximum it is allowed. The test: the next ten to twenty year operation against the $20bn line registered on 5 October and carried to this operation on 12 October. The print: on 15 October the operation was offered $40.836bn against a $6bn maximum and the Treasury bought the full $6bn in three of 42 issues (FiscalData, 15 October 2026), nearly seven times the maximum, after $46.4bn on 1 October. The verdict: held for this sector. The 12 October note carried the test forward and asked this print to speak for the whole long end, which it cannot; the other sector is scored separately below. The Treasury’s quarterly refunding announcement on 4 November decides whether the enlarged buybacks continue. |
| The Price of Duration. · 6 Oct, Premium · HELD, inflation expectations through the oil move and the meeting The call: in the Premium’s base case the ten-year breakeven stays anchored under 2.50, with a 2.25 to 2.45 band beneath it. The test: the breakeven against the line and the band through the oil move and the meeting. The print: December Brent went from 104.72 on 9 October (the exchange settlement in the 12 October Brief) to 116.7 on the 23rd and 114.9 on the 29th on desk-feed closes; the breakeven went from 2.33 to 2.32, 2.34, 2.38 and 2.35 on the 29th, inside the band throughout and fifteen basis points under the line (FRED, 9 to 29 October 2026). The real ten-year yield, the ten-year less the breakeven, fell from 2.91 to about 2.63: most of the fall in nominal yields over the fortnight was a fall in real yields, and there was no sustained breakout in longer-term inflation compensation. The verdict: held, inside the band and under the line. One reading, not scored: Wednesday’s six basis point ten-year rise split evenly between the breakeven and the real yield, neither the flat breakeven the Premium’s test for telling supply from inflation was written for nor pass-through from oil into inflation expectations, and the desk records it as mixed. The Record keeps that test registered against 2.50 and unscored. |
| The Print Was Hot. The Tape Sold It. · 15 Oct, Brief · PARTIAL, the stance to reduce: the pullback came and was given back The call: the dollar’s run was crowded and due a pullback, so the tactical stance moved to reduce on the 14 October settle, with the structural view unchanged and governed by the two-year through 4.65 with the futures market pricing less than one further rise. A dollar index close through 103 scores the stance early at full weight; the structural stance resumes if a euro retracement into 1.13 to 1.14 fails with the two-year above 4.65 and a second rise still priced by December. The test: the index against 103 and the euro against 1.13 to 1.14. The print: the index closed 101.81 on the 23rd from 102.53 on CPI day and 102.04 on the 29th; the euro reached 1.1221 on the 23rd and closed the 29th at 1.1156, the level it stood at when the stance was written (desk feed, daily closes, 14 to 29 October 2026). Speculators’ bets against the euro in the futures market, the gross short, set a record of 360,945 contracts as of 20 October (CFTC, released 23 October 2026). The verdict: partial, under the Record’s own definition of a move delivered and then materially given back. Neither the early condition nor the resumption condition has printed. Tonight’s positioning report, as of 27 October, says whether the record short began to clear into the meeting. |
| The Price of Duration. · 6 Oct, Premium, with The Print Was Hot. The Tape Sold It. · 15 Oct, Brief · NOT MET, the growth-scare scenario as base case: the front-end route half met, the credit route not The rule: the growth-scare scenario becomes the base case if credit spreads print through 3.50 on four consecutive settles, or if the two-year closes through 4.65 while the March contract prices less than one further rise. The print: the two-year settled 4.61 on Thursday, through the line for the first time since it was drawn, but March priced about 1.4 rises; credit narrowed to 3.14 by Wednesday (Treasury par curve, 29 October 2026; desk computation, 28 and 29 October 2026; ICE BofA via FRED, 28 October 2026). The verdict: the rule needs both conditions, so it is not met, and the desk keeps its original threshold unchanged rather than reading one condition as the whole. The weights stay at 45, 35 and 20. The return condition in the other direction, two credit settles inside 3.10 with claims steady and the two-year holding 4.80, is not met either: credit is four basis points from the first and the two-year nineteen short of the third. The line is also the floor of the Premium’s 4.65 to 4.95 band for the two-year; it broke on the 29th as a description of the front end, the shortest maturities, and the Record scores that row when it next marks. |
| The Price of Duration. · 6 Oct, Premium · NOT MET, the energy-shock scenario The rule: the shock scenario triggers if Brent is through 110 on the December contract with refined products leading and the desk’s refining margin above 70, or if the breakeven goes through 2.50 with the thirty-year through 5.80 in the same week. The print: Brent went through 110, to 116.7 on the 23rd and 114.9 on the 29th, but products did not lead: on a provisional series the margin fell from 65 on 14 October (the November contracts at exchange settlement) to about 49 on the 29th (the December contracts at daily closes) as crude rose and gasoline and diesel did not (desk computation, 23 to 29 October 2026; a contract roll and a change of source sit inside that comparison, and the exchange-settlement series is appended when the chart feed is restored). The breakeven closed the 29th at 2.35 and the thirty-year at 5.32, neither near its condition. The verdict: crude led, margins compressed, not triggered. Of the four conditions the 14 September Premium, The Tide Line, registered for standing the shock path down, the pipeline restart has printed; the margin through 55 reads as printed on the provisional series and is scored when the settlement series is appended; Brent under 95 and a shipping-lane accord in the Gulf have not printed. |
| The Price of Duration. · 6 Oct, Premium · MISSED, the thirty-year range, on 22 October The call: the thirty-year trades a 5.30 to 5.80 operating range, and a close through either edge is a range miss scored the day it prints. The test: a settle through 5.30 or 5.80. The print: on 22 October the thirty-year settled 5.29, one basis point through the floor, on a session the desk’s regime did not describe: an equity-led flight into bonds while oil spiked (Treasury par curve, 22 October 2026). The verdict: missed at full weight and resolved on that date. One basis point and one session are enough under the note’s own wording. The same note also wrote the floor conditionally: a close below 5.30 with credit quiet is a range miss, and with credit widening it reads as the growth path inside the framework. Credit was widening at 3.40, and the desk scores against the wording that reads against it. The regime’s retirement test, a close below 5.30 on a downside labour or inflation surprise with the term premium, the extra yield investors demand for holding long bonds, under 0.90, is not met: there was no such surprise, weekly jobless claims printed 194,000 that morning (Department of Labor, 22 October 2026) and the term premium’s last estimate is 0.99 (New York Fed via FRED, 16 October 2026). Since the miss the thirty-year has settled 5.33, 5.31, 5.30, 5.37 and 5.32, inside the range throughout. The range stands as published until a Premium redraws it. |
| Ten Dollars Offered. One Accepted. · 20 Aug, Brief; the reads registered 5 Oct and completed 6 Oct, carried in Duration Cleared. The Crowd Did Not. · 12 Oct, Brief · MISSED, the queue across the long end The call: the queue of sellers runs across the whole long end, not just one sector. The reads registered for the twenty to thirty year sector: offers above $20bn confirm the queue across the long end; offers near $10bn say the 1 October print was a single-sector event; between the two, the claim stands at the sector level only. The prints: the 8 October operation drew $14.9bn, between the reads, and was scored partial on 12 October; on 27 October the sector’s next operation was offered $10.746bn against the $6bn maximum, the Treasury buying the full $6bn in six of 35 issues (FiscalData, 27 October 2026). The verdict: missed on the desk’s own wording: the queue is a ten to twenty year phenomenon, and the longest sector offers the Treasury less than twice what it can buy. Offers in a buyback measure what holders will sell at the Treasury’s price, not why. No published note named the 27 October operation as the test; the desk applies the reads written for this sector to it on its own motion and takes the harder reading against itself. |
| The delivery · owed on the 23 Oct settle; the Record refresh registered 15 Oct, owed 19 Oct · LATE, four sessions The scoring note for the fortnight to 23 October was written on the 23rd’s settles, put through the desk’s editorial gates on Monday and Wednesday and not published; its rows appear in this table, re-marked to the 29 October settles, four sessions after the Monday publication it was written for. The Record’s refresh, registered on 15 October for the 19th, was published on 29 October: the Record is now scored to 9 October and revised 29 October. The tape behind these rows ran on FO News Live as it printed and the registered levels stood in the notes, so a Premium reader held both halves of every score on the day; what was late is the desk’s scoring of the two together, and that is what this row logs. The Record logs the desk’s lateness in a process log outside the call count, and this row is the entry the log will carry. The ledger’s earlier entries stand as logged: one session on 14 September, seven on 29 September, one on 5 October, Friday to Monday on 12 October and on time on 15 October. |
| A note on the format |
| FO Research is introducing a clearer format, and this Brief is the first in it. The conclusions now come first, followed by what happened and a concise scorecard, with the complete registered evidence available to every reader beneath it. The Premium analysis continues with the reasoning, the market implications and the exact conditions that would change the desk’s view. The research and scoring standards are unchanged: every call is still dated before the data and scored after it at the level its note published, with the misses at full weight. The Analysis Reports will follow in the same shape. |
| Continue Reading · FO Brief |
| The rest of The Committee Held. The Long End Did Not. is for Premium subscribers. |
| You have read what happened and how the calls scored. The mechanisms, the market implications and the exact conditions that would change the desk’s view continue behind the desk. |
| ✓Is the Fed finished? Why two dissents, a chair who will not guide and a Treasury Secretary who would not object to a rise matter more than the hold, and what would make December a full rise. |
| ✓Why long yields rose on a hold and fell the next day: the two forces fighting over the long bond, and what the desk needs to see before it redraws a range it has just scored itself missed on. |
| ✓Is the economy weakening? What credit at 3.14 says, which condition of the growth test was met, which was not and why the weights do not move on one session. |
| ✓Is the dollar turning? The euro back where the stance was written, a record short into tonight’s report and what a renewed euro rally would be judged against. |
| ✓Is oil becoming an inflation shock? A deal timetable against a strike timetable, Brent against the inflation gauge in one table and the margin question Shell’s record quarter raises. |
| ✓What would change the desk’s view: the registered triggers that move the weights, the range and the stance, each with its level and where it stood on Thursday, and the dated watch list from tonight’s positioning report to the December meetings. |
| Recently Published |
| 15 October 2026The Print Was Hot. The Tape Sold It.September CPI scored, the dollar’s failed high, the tactical stance to reduce and the window rule for the record. |
| 12 October 2026Duration Cleared. The Crowd Did Not.The retest scored: the thirty-year cleared at 5.618, the Treasury bought its full buyback maximum, the record euro short and the desk’s own misses. |
| 6 October 2026The Price of Duration.The regime named, the range redrawn to 5.30 to 5.80, the reweight on the fired tell and the majors mapped to the paths. |
| Every published desk call is dated before the print and scored after it, misses included. Inspect the record before you pay for it → |
| Subscribe to Premium |
| Editorial-grade macro research, written for sophisticated readers. Notes land when the data warrants. New to the mechanics? The Learn section is open to everyone. |
The research is Premium. The record is public.
Every Brief and Analysis in full: the analysis and market implications, the conditions that would change the desk’s view, the scenario maps, the tactical reads and the PDFs; the scorecard stays public. Inspect the record before you pay for it.
Subscribe to PremiumNot ready? Get the free read by email. New to the mechanics? Learn is open to everyone.