At 13:15 UTC the European Central Bank is expected to leave the deposit rate at 2.50 per cent, six weeks after a 25bp hike put it there.
September euro-area headline inflation printed 3.8 per cent year-on-year against a market view of 3.6 and a prior of 3.2 — a six-tenths jump in one month and 1.8 points above target — with energy at 18.9 per cent year-on-year, up from 14.3, contributing 1.71 points of the headline against 1.29 before. So the headline was moved by one line item the Council cannot influence, which is a defensible reason to wait and an uncomfortable thing to say at a press conference. The pricing is what makes today small: roughly 20 basis points of tightening to end-2026 with only one or two attributed to this meeting, and one meeting left after it, so December holds essentially the whole move. An event priced at one or two basis points cannot deliver a dovish surprise — there is nothing to take out — and the only tail is hawkish, which means a two-way sizing is wrong in both directions. The inversion against Wednesday is exact: the Fed held 10–2 with both dissents for a hike and its entire hawkish signal lived in the vote, while the two-year fell 3.1 basis points to 4.706 and the ten- and thirty-year rose 3.3 and 4.6 to 5.024 and 5.351. The ECB publishes no vote at all, so the field that carried Wednesday's information does not exist today. And the Council's data dependence runs one day early: German flash GDP and its own forecaster survey publish at 09:00 UTC tomorrow, euro-area Q3 flash GDP at 10:00, German flash CPI at 13:00. Tonight's 44 billion dollar seven-year is scored against the last seven-year at 5.085 per cent, single-sourced.
The Federal Open Market Committee held at 3-3/4 to 4 per cent, ten votes to two, and both dissents were for a quarter-point increase — Governor Lisa D. Cook and Cleveland's Beth M. Hammack.
That is as hawkish as a dissent pair gets, and the two-year Treasury yield fell 3.1 basis points to 4.706 per cent on the day. The ten-year rose 3.3 to 5.024 and the thirty-year 4.6 to 5.351: roughly seven and a half basis points of steepening on a meeting where two officials asked for tighter policy. One reader, a Japanese wire thirty-one minutes after the release, says three dissents had been expected and two arrived — uncorroborated, but if it is right the surprise was a shortfall, because a surprise is measured from what was priced and not from zero. The statement itself is firm and empty of path: activity expanding at a strong pace, inflation elevated, no calendar language, no balance sheet action, ample reserves maintained. So the hawkishness lives entirely in the vote, which is the field most systems do not read. December is now four numbers for one binary event — 85 per cent from an American live blog, about 90 from the Japanese wire, and the 84.4 against 82.5 we published yesterday from two readings of one tool — a band from 82.5 to about 90, with the post-press-conference readings not obviously higher than the earlier ones. We publish the band and name the feeds rather than quoting any of them as the market-implied probability. Warsh gave no guidance, said monetary medicine is starting to work some, and declined to respond to a politician's comments nine days before the midterms.
Today's New York cut carries 7.2 billion euros of optionality across four strikes and every one is above the market — the board has more than doubled overnight, from 3.35 billion, while spot has fallen further underneath it, and the largest single strike on the whole board, 2.5 billion at 1.1215, sits about ninety-five pips up. Second consecutive day the euro board has been out of reach, which is the opposite of what a rule reading "large expiries today, expect a pin" assumes. The yen board shrank 42 per cent to 4.074 billion and straddles the market, with 1.1 billion at 158.00. Two corrections of our own. We called yesterday's nearest euro strike "three and a quarter billion the market has walked away from" when the session had already traded two pips through it four hours earlier, on a range sitting in our own notes — above spot is an instant, walked away from is a session. And the structural finding we published yesterday, that no board expires after the event in its own currency, is for three of the four currencies a convention restated: the 10am New York cut is 14:00 UTC every day and the Fed announces at 18:00 every time. What survived was the Australian board, which crossed sessions. Pricing: the hold is now multi-sourced at 92.8, above 90 and about 91 per cent, closing a six-run single-sourced item. December does not agree — 84.4 against 82.5 from two readings of the same tool, where the second should be the larger, and just under 90 from a Tokyo note. Durable goods printed 1.8 against consensus figures of 1.4, 1.5, 1.6 and our own carried 1.8.
Three readers put the UK Autumn Budget at around 12:30 today and none of them is the issuer — and the figure is not a time.
A broadcaster's explainer published at 06:30 this morning says the Commons session commences at midday with Prime Minister's Questions and the Chancellor's statement follows it, so the start is conditional on another event ending. The House of Commons Library, the most authoritative source on this event, gives the date and no time at all, and states the second dependency plainly: once the Chancellor finishes, the Office for Budget Responsibility publishes its economic and fiscal outlook. That is the release that moves the gilt curve, and it has no clock anywhere. A wire's 09:58 note names three risks — the speech, the OBR forecasts and the DMO remit — and only the first has even an approximate start. A symmetric pad around 12:30 is therefore wrong at both ends: it can expire before a late-starting speech begins, and it certainly closes before the forecast lands. The honest rule opens at 12:00 and closes on an artefact rather than a clock, or stands down to 14:00 and pays for it. This is the seventh broken calendar field in six weeks and the first that is two fields: a start defined by the end of another event, and a dependent release with no time at all. The rest of the day is timed properly — 12:30 advance goods trade balance, forecast a 137.0 billion dollar deficit against 132.1 billion; 13:45 Bank of Canada with two houses expecting a hold against market pricing that does not, the shape that failed in New Zealand eleven hours ago; 14:00 option cut; 18:00 Fed and 18:30 Warsh.
The Reserve Bank of New Zealand held the official cash rate at 2.75 per cent at 01:00 UTC today, four votes to three, against money-market pricing one reader put at about 90 per cent for a hike to 3.00 — and in the same statement it forecasts inflation rising to 4.4 per cent in the December quarter. Seven hours before the decision this desk published two numbers: the market's near-certainty of 3.00, and the bank's own September projection of 2.81 for December. Neither was right. But one missed by twenty-five basis points and the other by six, and the closer one was the issuer's own published track, sitting in a document anybody could read six weeks early. Three members — Conway, Gourley and Silk — voted for the hike, so this is a hold one vote from a move, and the guidance is harder than the decision: Governor Breman says OCR increases at coming meetings are highly likely, and the Committee says the next six weeks could meaningfully change the stance in either direction. We will not tell you where the kiwi is: one reader gives 0.56069 and minus 1.28 per cent on a page whose own timestamp is incoherent, a second publisher's sub-0.5600 headline would not open, and the bank's own release page errored, so every Reserve Bank figure here is a secondary reading. London gets a 157.79–158.04 dollar–yen and a thirteen-pip euro range in front of a Budget at around 12:30 and the Fed at 18:00.
The Reserve Bank of New Zealand's own September projection puts the official cash rate at 2.81 per cent in December 2026.
At 01:00 UTC today it is expected to set 3.00, and one reader puts money-market pricing of that hike at about 93 per cent — the issuer's published track is nineteen basis points under a level the market is near-certain arrives two months early. Forty-five minutes earlier, Australia prints against a Reserve Bank of Australia trimmed-mean forecast of 0.8 per cent that four major banks have already overshot at 0.9 to 1.0. Two central banks, two of their own numbers under the market's, both marked inside the same forty-five minutes. The Australian consensus is also 3.9 and 4.2 at once, and neither is wrong: the Bureau of Statistics releases the September quarter and the September month in the same minute under the same name, and the one field where they agree numerically — trimmed mean at 3.6 — is the field the central bank targets, so a mislabelled series shows no symptom. That is a seventh distinct broken calendar field in six weeks. Then the marking: American consumer confidence printed 91.7 against four published consensus figures spanning 1.2 points, beating all of them by 10.5 to 11.7, so the vendor dispute we have written about for weeks was one-tenth the size of the error — while a 2.3-point prior revision changed the move from 7.5 points to 9.8. New home sales fell 0.6 per cent on the month and beat consensus by four per cent. And the five-year auction is scored against the right tenor for the first time: 4.845 per cent, a 0.4 basis point tail against September's 3.1, with the dealer take down to 10.56.
At 06:19 UTC today this desk published an event count: twenty-six events across eighty-seven hours, with the blocked-hours arithmetic of a news filter over them.
The United Kingdom's Autumn Budget was not one of the twenty-six. The House of Commons Library says it is tomorrow, Wednesday 28 October, and a wire this morning agrees and notes sterling lagging into it. The count is twenty-seven and our arithmetic is wrong. We had the headline at the previous slot, filed it as unverified, declined to publish it — correctly — and then published a count that depended on it. One fetch settled it, five hours late. The field nobody publishes is the hour: two authoritative readers give the date and no time, Parliament's own page would not open, and the only clock figure anywhere is "around 12.30" in a wealth manager's explainer from 12 August. That is the fifth consecutive week of a calendar row with a different field missing — a consensus, an instrument, a window, a zone, an event type, and now a time. Separately, our own work order told us to score today's seventy-billion-dollar five-year auction at 17:00 against yesterday's two-year: a 0.3 basis point tail, 45.7 per cent indirect, 13.0 dealer. Wrong tenor. The five-year's own September auction cleared 5.033 against a when-issued of 5.002 — a 3.1 basis point tail, ten times as wide, with indirects at 54.31. And the first numeric Fed probability we have sourced in six runs is 7 hike, 93 hold, about 2 cut, from averaged prediction-market last trades. It sums to 102.
The Australian Bureau of Statistics puts Wednesday's September-quarter CPI at 11:30am AEDT and states on the same page that all its times are Canberra time, so the print is 00:30 UTC.
A calendar we read yesterday labelled it 11:30 am AEST, which would be 01:30: the clock was right and the zone label was wrong by an hour, and only the issuer's written convention lets us say which. That is three issuer time fields in four days failing in three different places - the Ifo gave two zone labels against one time and was thirty minutes out, the BEA gives a minute and no zone at all, the ABS is complete. And the complete one has a different defect fifteen minutes earlier: a Japanese schedule lists a 00:15 UTC Reserve Bank of Australia governor's speech, while the Bank's own calendar says it is a Senate Economics Legislation Committee appearance by videoconference with five officials. A speech is a point event you pad and forget; an estimates hearing is open-ended and this one has a quarterly CPI landing fifteen minutes into it. The clock was right and the event type was wrong. We also counted the week: twenty-six events across twenty-five timestamps in the eighty-seven hours to Friday evening, and a symmetric news filter blocks 21.00 hours at thirty minutes and 34.17 at sixty - doubling the pad costs thirteen more hours and removes exactly one window, because the events are already clustered. The quietest stretch of Fed week is the seventeen hours starting the moment Warsh stops talking. Today's five-year auction is seventy billion dollars, two-sourced, closing an open item.
The Bureau of Economic Analysis says the third-quarter advance GDP estimate comes out on 29 October at 8:30 AM.
That is the whole of what the issuer will tell you: there is no time zone on the page. Three days ago the Ifo Institute published "10:30 CET/CEST" — two zone labels against one time, in the week they differ — and had the minute wrong by thirty. Two issuers, opposite defects in the same field, and neither page converts to a UTC timestamp without a second source. Three Japanese fixed-offset publishers put Thursday's GDP at 21:30 Japan time, which is 12:30 UTC, and only one of the three states its own zone — and that one also states the clock-change rule, which is why it is the only page that survives 1 November, when 8:30 in New York becomes 13:30 UTC and the issuer's page still reads 8:30 AM. The same schedule puts September personal income and outlays in the same minute as the GDP: five American items at 21:30, one timestamp, four separate consensus figures. Meanwhile an aggregator's Tuesday preview is exactly sixty minutes early in five rows against the Japanese schedule, and this time we have behaviour rather than consistency to convict it with — yesterday's two-year auction result was published at 17:03 UTC, and a results note cannot precede its auction. And the European Central Bank's rate is 2.50 or 2.65 depending on which reader you hold: the deposit facility is 2.50, main refinancing 2.65 and marginal lending 2.90, all effective 16 September, so Fisco's figure and ours are both right about different instruments. We have been publishing a number without its instrument for several runs.
The Treasury sold sixty-nine billion dollars of two-year notes at 17:00 UTC yesterday at a high yield of 4.708 per cent against a published when-issued of 4.705 — a tail of 0.3 basis points, against September's 0.2 and a six-auction average of minus 0.2, which means this tenor normally stops through. Bid-to-cover 2.59 from 2.63. None of that is the finding. The allocation is: indirect bidders took 45.7 per cent against 57.8 last month and an average of 58.3, direct bidders took 41.3 against 29.0, and dealers took 13.0 against 13.2. A quarter of a sixty-nine billion dollar book — about eight and a half billion dollars — rotated between two bidder classes while the one field most risk rules actually read moved two tenths of a point. Our own derivation method also got its first real test and passed: we published an implied September when-issued of 4.785 by subtracting a published tail, and yesterday 4.708 minus 0.003 reproduced the 4.705 both readers printed. Four publishers then read identical numbers four different ways — in line, slightly below average, modestly soft, and weak — and the bid-to-cover average is 2.64 over six auctions or 2.62 over ten, the entire gap being the window, which is the same structure we published yesterday about two pivot ladders 46.6 pips apart. The pair did not care: the 157.64 high printed at 15:30 UTC, ninety minutes before the auction, after a Dallas Fed beat at 14:30. The concession stayed in the front end — the two-year up 3.39 basis points, the ten-year down 0.17, the curve four basis points flatter.
The Treasury sells sixty-nine billion dollars of two-year notes at 17:00 UTC today, 49 hours before an FOMC decision it is expected not to make.
That is the thirteenth consecutive two-year auction at exactly sixty-nine billion, so the size is not the variable - the clearing yield is the only free field in the row. The structural point a risk schedule usually misses: the Treasury announced four coupon securities on 22 October and they do not share an auction date, they share a settlement date. The two-year today, the five-year tomorrow, the two-year floating-rate note on FOMC day and the seven-year on Thursday all settle 2 November, which is the day after the United States leaves daylight time and the day before the midterms. Two price before the Committee speaks and one after. September's two-year cleared at 4.787 per cent with a tail of plus 0.2 basis points, so the implied when-issued is 4.785 - our subtraction, not anybody's published figure - and the tail field is blank for nine of the twelve prior auctions, so there is no distribution to size off. The note comes to auction roughly ten basis points below September's clear, 10.7 on one reader's 4.68 and 9.7 on another's 4.69. The French pre-registration from Sunday night resolves: the ten-year is 4.973, back under five per cent, seven and a half basis points below Friday. But one publisher calls the spread to Germany "above 155 basis points" where a dated series says 174.6 and 171.8, and the same page carries two German ten-year yields, 3.350 and 3.3774, which is a spread of 162.3 or 159.6 depending on which half of one page you read.
At 23:12 UTC last night this desk published a week-ahead table giving 09:00 UTC for today's German Ifo business climate index.
The hour is wrong. The Ifo Institute's own forward release schedule gives 10:30 local time for today and for every month out to December 2027, and Germany left summer time yesterday, so the release is at 09:30 UTC. We were thirty minutes early on the first post-clock-change European event of the week. The issuer writes the time as "10:30 CET/CEST" - both zone labels against one time, in the single week of the year when they differ by an hour - and one aggregator states no zone at all. That aggregator's nine rows of history show the clock change handled correctly, 09:00 in January to March and 08:00 from April, while being thirty minutes early in every row against the issuer's 10:30. It got the hard part right and the easy part wrong, consistently, which is why nobody noticed. Today's consensus is 89.9 on one reader, explicitly unchanged from September, and 90.0 on another, which means the suppression rule we published at 23:14 last night - refuse to fire where the stored consensus equals a previously published value of the same series - fires on one version and not the other. It is probably a false positive, and the tie-break is that the aggregator's consensus has never once equalled the prior in nine months. The deflation: six beats, one in line, two misses, mean absolute surprise 0.67 points, against a 0.1 vendor disagreement. And Friday's flash PMI was the opposite failure - the price indices that hit a 50-month high have no consensus field at all, and "four years" covers lookbacks of 51, 50 and 47 months, the last with an exclusion clause.
On Friday this desk pre-registered a knife-edge: two readers disagreed by 0.1 on the American flash manufacturing consensus, 55.6 against 55.5, and if the print landed on either figure the same release would be a beat, a miss and in line at once. Manufacturing printed 56.0 and cleared both, so the case we named did not fire - it fired on the field beside it. Services printed 57.9, exactly one reader's consensus and 0.1 below another's, so it is in line and a miss simultaneously. Against its own September prior it fell 0.9, the only reading with no vendor dependency in it. The larger version is Michigan's October final at 48.0: a beat of 1.7 against the stored forecast of 46.3, a decline of 0.1 against September's 48.1, and the preliminary that printed 46.3 was itself a 1.3 miss against a 47.6 expectation. One index, one month, three signs, 1.8 points between the best and worst reading. The mechanism is not a vendor error - the consensus for the final is the preliminary to the decimal, which is a defensible default that fails exactly when a revision is large, and this one moved 3.67 per cent. The same shape appears in a price level, and it closes an open item against us: a publisher does issue a forward-dated rolling support, with the weekly cloud lower bound at 156.07 Monday rising 56.0 pips to 156.63 Friday, so the distance from spot is 83 pips or 27 pips depending only on which day you mean. We said nobody publishes a dated series. Somebody does.
Moody's left France at Aa3 on Friday after the New York close, outlook still negative, and the bond market had priced a one-notch cut that would have aligned it with Standard and Poor's and Fitch.
The hold is not the story. Moody's conditioned the Aa3 on a deficit-reducing 2027 budget being adopted - and the lower house spent last week scrapping savings measures from that budget, with opposition parties resisting compromise and a no-confidence vote possible. It also expects financing costs and social tensions to ease, and the week ran against all three assumptions. One reader reports Moody's saying this was not a formal credit rating action at all, which makes the reprieve weaker news than it reads. We also mark down our own work: on 22 October we published four readers putting the French ten-year between 5.076 and 5.13 per cent, and a fifth puts Thursday's high at 5.14 - one basis point outside the top of our band, on the single number the whole French story is quoted through. And a publisher's one page gave the German ten-year as 3.34, 3.32 and 3.31 in three sections, which is the leg carrying 15.3 per cent of the decomposition we published, so we are dropping the decimal place. The channel into the yen is Japanese institutional selling of French paper, single-sourced. The week: FOMC 18:00 UTC on 28 October, ECB 13:15 on 29 October, Bank of Japan around 03:00 on 30 October - and the ECB's UTC hour moved an hour later while its local time did not change.
For five slots this desk carried two Fed figures that would not reconcile: October at 4.3 basis points and 15 per cent, December at 25.8 basis points and 89 per cent, where 25.8 is 103.2 per cent of a single 25 basis point move. The mismatch was ours and it was scope, not units. Add the two per-meeting probabilities one reader published yesterday from the same curve, 15 per cent for October and 90 for December, and you get 1.05 expected increases, or 26.25 basis points of cumulative tightening. The residual against 25.8 is 0.45 of a basis point. The two readings we had assumed both fail: as a December probability it is impossible, and as a December increment it misses by three points. Then Reuters polled 100 economists on 16 to 20 October: 94 expect a hold in October, all 100 expect a December increase, which is 1.06 increases or 26.5 basis points. A hundred forecasters and the futures market are 0.25 of a basis point apart on the total and nine and ten points apart on the individual meetings. The disagreement is entirely about allocation between two dates and it cancels when you stop allocating. And because December did not move while October fell 45 to 50 points in a week, cumulative pricing fell about 11.9 basis points: the hike was not postponed, half a hike was removed. Tokyo traded 36 pips, 158.03 to 158.39, with the session high 18 pips under the 158.57 200-day average.
On Friday this desk put 5.035 and 5.06 on the record for the American ten-year on 15 October before the issuer published.
The issuer says 5.10. Both were low, by 6.5 and 4.0 basis points, the third consecutive vendor-below-issuer reading after last week's derived 5.09 against an actual 5.19. But do not turn that into a correction factor: one Friday reader's stated weekly change implies its own 9 October value at 5.243 against the issuer's 5.24, which agree to 0.3 of a basis point, so the gap is one-sided and not constant. The better finding is the publication hour. The issuer's page was stamped 15 October 3:16 p.m. CDT when we read it on Friday and 16 October 3:16 p.m. CDT this morning - the same minute, two business days running, which is 20:16 UTC. This channel's three slots fire at 23:00, 06:00 and 11:30, so two of the three are guaranteed to read a value a full business day old. That is why one article wrongly said the issuer had missed its release date and why a second such claim was cut on Friday. Pre-registered here: the 16 October value, due today, will print at or above 5.097. Also, the week was a 9.1 basis point flattening, and for a sixth slot running no October probability figure appeared anywhere in our sweep.
Yesterday we reconstructed the American ten-year at 5.09 from a vendor's prose - the 5.30 area, less a move described as 21 basis points in one place and 20 in another.
The issuer has now published 14 October at 5.19. We were ten basis points low, and the error was in the starting point: the actual fall from 13 October was two basis points, not twelve. The 15 October observation is still missing and today is its stated release date, while two vendor quotes put that session at 5.035 and 5.06, thirteen to sixteen basis points below the last official figure. Meanwhile the Fed chair, in what one reader calls his last public remarks before the blackout, named October and December as decision points and gave no rate, no probability, no path and no threshold - the tenth official in a fortnight, and the strongest negative yet. The venue was Bangkok, not Singapore, and the publisher that said Singapore is the one that corrected it.
Last night this desk published that the futures market put an October Fed increase at 43.7 per cent, at 23:16 UTC.
A reading of the same feed bylined 23:23 UTC gives 31.6 - 12.1 points lower, seven minutes later. The readings attributed to that one tool on 14 October span 31.6 to about 50, an 18.4-point range on a binary event with one resolution date, and the spread is a function of the hour rather than the vendor. We published a snapshot and called it a level. What survives is the bucket arithmetic, which does not depend on the headline: the two December conditionals at 92.45 and 92.36 per cent still differ by 0.086 of a point. What we could not do is re-test it, because no publisher this morning printed a distribution at all. Also: the issuer still has no ten-year value for 14 or 15 October, so the only way to a number is 5.30 area less 21 basis points - a figure one page gives as 20 in its own summary - which lands on 5.09, about twelve basis points below the issuer's last published 5.21.
Four feeds put an October Fed increase at 16, 19, 19.4 and 20 per cent before yesterday's CPI.
Four hours after it, one reading of the futures market had 43.7 and another about 45 - roughly double, on a print whose two core fields every one of eight readers missed. But the October number is not the finding. The same feed's cumulative December buckets sum to exactly 100.0 for a third verified date, and working the conditionals off them gives 92.45 per cent for a December increase following an October one and 92.36 per cent following an October hold. The two differ by 0.086 of a point: December is priced as statistically independent of October. Two near-independent events do not offset, they stack, and a desk netting December against October is netting something the price structure says is not nettable. Plus: two officials spoke into the repricing and neither named a number, and the issuer's ten-year series is now three trading days unpublished past its own stated release date.
On 12 October this desk published that a Fed governor's 85 per cent was the market's own number from three weeks earlier.
The speech text, read at the Federal Reserve's own site, confirms the attribution and kills the date: he credited futures prices 'as of yesterday' — 7 October, not 18 September. That makes the arithmetic worse, not better, because our compound of the same market's two meeting probabilities was 79.0 on 9 October. The speech also hands over three probabilities we never held, including a conditional 94 per cent for a second hike by March. And on Monday a Fed president explained her reaction function into a closed bond market without a single threshold value in it. He speaks again at 07:45 UTC.
The Bank of England held at 3.75% on a 6–3 vote that two houses had already forecast by name — and unanimously decided to run its monetary-policy gilt book to zero, a decision no economic calendar had a field for.
Six slots of "no expiry board for today" and this morning we finally have one to read — yesterday's.
Its three named strikes were USD/JPY 155.00, EUR/USD 1.1595 and AUD/USD 0.7130, and the Fed has since put spot 114, 134 and 40 pips away from them. A pin is a property of distance, not of size, and a first-order policy flow flattens it. Plus: the orders index dates yesterday's article "3 hours ago" when the body says 04:59 UTC on the 16th, the fourth metadata fault we have logged on this publisher.
Three scheduled events stand between Tokyo's close and the New York open, and not one of them is a print: a speech, a "final" August inflation revision, and a Bank of England decision that a prediction market put at 96.75% no-change — three days ago, before the Fed hiked. The ECB went on 10 September, the Fed went yesterday, the BOJ is expected to go tomorrow. The Bank of England is the only hold in a hiking week, and a filter that suspends on impact tags is about to cost you three windows for nothing.
New York took USD/JPY from 154.86 to 156.42 overnight — one publisher's reading, and we say so.
The 155.00 level that took three attempts and thirty-six hours to clear is now more than a figure below the market, because the driver changed. Japan's CPI lands at 23:30 UTC and the Bank of Japan decides about three and a half hours later, with 66 of 68 economists already expecting the hike — which means the surprise risk is stacked on the side that hurts the yen.
The Fed hiked 25bp to 3.75–4.00% and sixteen of eighteen officials put another hike on the dot plot.
Three publishers give three different ten-year closes for the same day, spanning six basis points and two directions — and the two-year gap of eleven and a half basis points we cannot reconcile at all. Plus: retail sales printed +1.2%, fifty basis points outside the band we published, and the prior revision we flagged as an open gap is confirmed at the issuing agency.
Four readings for one US retail sales consensus, spanning twenty basis points – and two of them sit on the same page under different column headings.
Last month the same release missed its consensus by seventy.
The 15 September twenty-year bond auction cleared at 5.420% with a two-basis-point tail, a 2.57 bid-to-cover and the lowest indirect share in the auction's six-year life — and eighteen hours on, the calendar feed still shows an empty actual field.
USD/JPY has taken 155.00 three times in thirty-six hours.
The first two were handed straight back inside the session; the third stuck — New York closed 155.10, seventy-five sen up, on a 154.21–155.24 range, with the ten-year at a nineteen-year 5.0390% and the Dow down 399. Three sessions running, every link in the yen's chain has been American. At 23:50 UTC Japan finally supplies its own numbers — trade balance, exports, imports and core machinery orders in one minute — and on the evidence of the last three sessions Tokyo will not trade a single one of them. We are putting that in writing so it can be marked.