This is a Monday note, so it covers two and a half days: Friday’s US employment report, Friday’s New York close, the weekend, and what Tokyo has done with it since the open. The short version is that the calendar handed the dollar the best payrolls number relative to trend in a year — 165,000 against a reported 65,000 consensus, 3.4 times the twelve-month average — December hike pricing went up by seven to ten points depending on whose feed you read, and dollar-yen still stopped roughly eleven pips below its 200-day moving average and closed under 158.50. The range it has been stuck in since late September is intact. Our companion piece this morning takes apart the print itself; this one is about what the print failed to buy.
The line it stopped under is five numbers spanning nine pips
Friday’s high is already two numbers, from the same publisher, half an hour apart. One Gaitame.com note timed 06:49 JST today — 21:49 UTC Sunday — puts it near 158.58. A second note from the same house at 07:45 JST, 22:45 UTC, puts it around 158.56. Two pips, two desks, one tape. The close is worse documented: the first note says the pair was bought back to near 158.50 into the close and the second gives only a 0.5 per cent gain on a closing basis with no level at all.
Now the 200-day. The second note puts it around 158.69 today. The same house’s week-ahead piece, written to the 5 November close, puts it at 158.66. Our own file carries 158.60, 158.63 and 158.65 from three days of Tokyo publishers last week. That is five readings of one moving average spanning nine pips.
So the sentence “Friday’s high failed eleven pips below the 200-day” is true on today’s reading and becomes “failed two pips below it” on the lowest one in our file. Both are defensible and they describe different events. We are not going to pick, and we will say what we think the honest version is: the pair did not trade above 158.70 and did not close above 158.50, and anything finer than that is a claim about whose moving average you subscribe to rather than about the market.
If your handling treats a 200-day break as a state change, this is the week to find out which vendor’s line your system is actually using, because the answer changes the sign of what happened on Friday.
And the week’s high may already be above it, on one reading
Gaitame’s week-ahead note says dollar-yen reached 158.764 at some point last week. Our own Thursday article published the week’s high as 158.67. Those two figures are 9.4 pips apart and, critically, they sit on opposite sides of the 200-day readings above.
On the 158.764 reading the line was cleared during the week and did not hold, which makes Friday a second failure at a level already proven. On our own 158.67 the line was grazed and never cleared, which makes Friday the first real test. These are different stories about the same five sessions and we cannot tell you which one is right. The 158.764 figure is single-sourced and it contradicts a number we published ourselves, which is the combination we have learned to flag rather than resolve.
What we can say is that three separate quantities — the week’s high, Friday’s high and the 200-day itself — are all uncertain to about ten pips, and the thing people want to know is whether one is above another. When your error bars are the same size as the distance you are measuring, the measurement is not available at the precision the question assumes.
Dollar up, front end up, long end pointing both ways
On Friday morning this desk published a six basis point bull steepening as the December repricing: the two-year down 10 to 11bp to 4.46–4.47 per cent, the ten-year down only 5 to 4.83–4.84, with a published ten-year minus two-year spread of +0.38 points. We argued that a flight to quality flattens, and that this one had instead taken money out of the next two meetings and left the term premium alone.
The print gave back about half of it. One US reader has Friday’s two-year rising to 4.51 per cent from 4.46, five basis points, and the ten-year to 4.86 from 4.84, two basis points — so roughly 46 per cent of the two-year move and 40 per cent of the ten-year move came back, and the curve re-flattened by three basis points. The proportionality is the interesting part: the print did not change the shape of Thursday’s move, it partially rewound it.
And the long end is reported in both directions. The Tokyo note says the ten-year “fell back” from near 4.90 per cent. The US reader says it rose two basis points on the day. Both can be true of one session — up on the day having come off a 4.90 per cent intraday high — but if anything you run keys on the sign of the ten-year’s daily change, it mattered this weekend which page you read. Neither publisher is wrong and neither tells you the other half.
December is four numbers, and for once our pre-print mark was the one that held
On Friday morning, two hours before the print, we published that everyone had the December hike probability at around 70 per cent and that the three readers carrying it disagreed about what it had fallen from. Here is the post-print board, with the feed named where the publisher names it, because none of these is “the” probability:
- 76.8 per cent for a 25bp hike, attributed to CME FedWatch as of Friday 6 November — up from 66.9 per cent a week earlier on the same feed.
- “Roughly 77 per cent”, attributed to CME FedWatch, in the body of one article — whose own summary field says “close to 80 per cent”. One article, two figures, three points apart.
- 70 per cent earlier Friday rising to 78 per cent after the data, with no feed named at all.
- About 80 per cent, in a Tokyo desk note written 06:49 JST today.
The useful finding is the unnamed intraday pair. It puts the pre-print mark at 70 per cent, which is exactly the figure we published at 11:40 UTC on Friday, two hours before the release. This desk has spent several weeks being wrong about numbers of this kind; this is one where our mark was the one that held, and we will take it without pretending it was skill — we published the consensus of three readers and the consensus happened to be right.
The 66.9 per cent week-earlier figure is the genuinely new information, and it breaks the tidy story. We reconstructed last week as a decay from 80 to 70.9 to 70 per cent. Anchor that at 66.9 per cent a week ago and it is not a decay at all — it is a round trip up and back, and now up again past where it started. Do not quote any single one of these as the market-implied probability.
Monday is thin and every bit of the week’s information is on Tuesday
Today, in UTC: 01:30 China October CPI, forecast 0.8 per cent against 0.9 prior, in the same minute as China October PPI, forecast 5.3 per cent against 5.0 — the only genuine same-minute pair today, and a producer price index accelerating through 5 per cent is the more interesting half. 05:00 Japan’s preliminary leading index, forecast 116.5 from 118.0, and the coincident index, forecast 119.4 from 118.8. 07:00 German September trade balance, forecast 18.0 billion euros against 19.4 billion revised. 13:40 Bowman, a voter. 15:00 US final September wholesale inventories, forecast 0.6 per cent against a 0.6 per cent prior — a release whose forecast equals its own prior, which is a news window you can safely not have. 17:15 Barkin and 17:25 Bailey, ten minutes apart. 18:00 the three-year note auction, the first $58bn of the week’s $125bn. Goolsbee speaks with no published time, which is its own kind of event.
Then at 23:50 UTC tonight the Bank of Japan publishes the Summary of Opinions from the 29–30 October meeting — twenty minutes before this slot’s window reopens tomorrow, with persistent BOJ hike expectations already named by two Tokyo publishers as a yen-supportive factor. On Thursday at 01:30 UTC, Asada speaks, with a press conference at a time nobody has published; he dissented on the September hike, which makes him the board member whose words move the terminal-rate argument rather than the next-meeting one.
And the rest of the week is one event. US October CPI at 13:30 UTC on Tuesday, for which no page we opened publishes a numeric consensus. The ten-year auction follows at 18:00 Tuesday, the US bond market is shut on Wednesday for Veterans Day while equities trade, the 30-year goes at 18:00 Thursday, Japan’s CGPI lands 23:50 Wednesday, and PPI closes the week at 13:30 Friday.
Which brings the sizing question, and it is the only question this note is really asking. The forecast range for today is 157.800 to 159.000 — 120 pips. The week’s is 156.40 to 159.20, 280 pips, with a break above 158.65–158.76 said to open 159.00–159.20 and a break below 157.25–157.40 said to open 156.65–156.92. The 25-day sits at 157.97 and the 100-day at 159.13 on the 5 November close, so spot is between them and the 200-day is the nearer wall.
Friday was as strong a catalyst as this calendar produces and it closed inside both ranges. A payrolls print 100,000 above consensus, at 3.4 times trend, with both wage fields accelerating and the front end selling off five basis points, did not clear 158.70. So the question for Tuesday is not whether CPI runs hot. It is whether anything on this calendar can break a six-week range, and the base rate you now have for that is one very good attempt that failed. We would size the CPI window as a range event with a break option, not as a breakout waiting for a trigger — and we would rather be wrong about direction than wrong about which regime we are in.
What this does not tell you
Every dollar-yen level above comes from one publisher’s two notes. That is two desks, not two houses, and it is thinner corroboration than it looks. Three other publishers had Monday-morning quotes and none of the pages would open — two client errors and one robots refusal — so we have not reproduced their levels, and OANDA Japan’s dollar-yen note for today failed as well. There is no independent third read on Friday’s close in this article.
“The range since late September” is one house’s characterisation and we did not derive the range’s own high and low ourselves. Treat it as a description, not a measurement, and do not place orders on its edges on our word.
The 158.764 weekly high is single-sourced and conflicts with a figure we published. The S&P 500 close of 7,376.86, up 0.31 per cent, is single-sourced, and we do not have Friday’s Nasdaq or Dow closes at all — so the equity leg of this story is missing, which matters, because an AI-driven selloff was the yen’s support all last week. The Nasdaq’s 5.52 per cent fall on 5 November is likewise from one week-ahead note.
We have no option expiry board for today and no resting-order levels: neither note we could read carried any, and the publisher that has been the only source of a strikes board is not in this run. So we cannot tell you where the week’s optionality sits, and a 200-day argument with no expiry map underneath it is weaker than it reads.
Finally, there is no consensus for Tuesday’s CPI on any page we opened, which means we cannot define a hot print for you. Anyone telling you where CPI needs to land to break 158.70 is using a number we could not find.
