The Bank of Japan raised its policy rate to a level Japan has not seen since 1995, and the yen fell. Not a little: USD/JPY was 155.87 in the Tokyo morning and printed 157.33 within the hour of the announcement, with EUR/JPY through 180 for the first time since 7 September and GBP/JPY up to 210.26. Every yen cross on the board went the same way at the same time. This is the post-decision measurement that both of our articles this morning said in print we did not have, so here it is — and the thing worth your attention is not the direction, it is that the hike itself was never the flow. The dissent structure was.
A hundred and ninety pips, and the hike was the least of it
Take the sequence as the intraday feeds published it, in UTC. Around 03:30 the Bank announced 1.25% by seven votes to two. By roughly 06:40 one Tokyo feed had USD/JPY at 157.33 against a 155.87 morning level, and described it as the first print above 157 in about two weeks. EUR/JPY went from 178.94 to 180.67 and GBP/JPY from 208.25 to 210.26 over the same window. That is a broad yen sell-off, not a dollar rally, and it happened on the morning the yen’s own central bank tightened.
The explanation the Tokyo desk published is the right one and it is worth stating flatly: two of nine board members voted against the hike, and the market read that as the near-term follow-up being harder than it looked. What repriced was not the 25 basis points that happened. It was the 25 basis points after that. A hike delivered into full pricing transfers no information; a vote count does, and a vote count is the one field your execution stack does not have.
If you were long yen into the decision on the theory that a 31-year high is yen-supportive, you were not wrong about the policy. You were wrong about which part of the announcement carried the news, and the market told you inside ten minutes.
We are not going to tell you where the high was
Because our sources do not agree, and the disagreement is not small. One Tokyo intraday feed puts the post-announcement peak at 157.33 at roughly 06:40 UTC. A European-morning piece published at 10:39 UTC describes spot around 157.5 and says it tested above 157.80 during the press conference. The same Tokyo feed had USD/JPY back at 156.88 at 07:08 UTC, after a dip to about 156.50 while Governor Ueda was speaking.
Those two accounts do not reconcile on sequence. One has the press conference producing the day’s low-water mark of the move; the other has it producing the high. They may both be right about different minutes of a ninety-minute event, and we cannot establish which minutes from what is published. So take the band: the day’s high is somewhere between 157.33 and a little above 157.80, and any inference of yours that depends on the precise level is weaker than it reads. If your post-event review logs a single “session high” from one vendor, you are recording a measurement as a fact.
One thing the band does settle. A Tokyo research note published at 08:13 JST — about seven hours before the peak — gave a forecast range for the day of 154.80 to 157.50. Spot went through the top of it. A published day range is a distribution, not a fence, and this one was breached on the correct side of a correctly-forecast policy decision.
The expiry board existed. It just was not where we looked.
This desk reported at 06:26 UTC this morning that no option expiry board had been published for 18 September — the publisher’s own orders index ended at 17 September, and the undated URL for today returned a 404. That was an accurate description of what we could see and it was wrong about the world. The board had been published at 05:57 UTC, twenty-nine minutes earlier. The same URL resolves now.
That is the fifth metadata fault we have logged on this publisher and the second of the same species: the article exists, the numbers in it are fine, and the index and the timestamps around it are not. The operational rule that follows is narrow and cheap. When an index says a recurring daily item is missing, the index is the claim you should doubt first, not the item. Re-request the canonical URL before you conclude the thing does not exist.
What is on the board matters more than how we found it. Two USD/JPY strikes are named. One sits at 157.00, which is where spot has spent the European morning. The other is 155.00 with $5.5 billion behind it — the largest notional on the board, and the publisher itself notes it is more than 200 pips away. On EUR/USD, 1.1450 and 1.1500 sandwich spot. The notionals for the rest of the board sit inside an image that does not extract, so treat the $5.5 billion as the one number we actually read.
The second time this week a big strike got stranded
On 17 September this desk argued that a pin is a property of distance rather than of size, after the Federal Reserve moved spot 114 pips past a USD/JPY 155.00 strike and the strike did nothing. Today the same strike is stranded again, by a different central bank, in the same direction, and it is now more than two hundred pips away with $5.5 billion on it.
We said at the time that one instance is an anecdote. This is the second, and we are still not going to upgrade it into a law, because both instances share a cause: a first-order policy repricing inside the expiry window. What we will say is narrower and more useful to you. The size of a strike tells you how much gamma a dealer has to hedge if spot is near it, and tells you nothing at all about whether spot will come near it. A $5.5 billion strike 200 pips away is not a magnet; it is a number on a page. If your rules widen a stop or skip an entry because “there is a big expiry today,” the rule needs a distance condition in it or it is noise dressed as structure.
The strike that is actually live today is the 157.00, because that is where spot is, and we have no notional for it.
What this does not tell you
We do not have a spot level at the time of writing. Every figure above is a reading from a published feed with a timestamp attached, and the most recent of those is 10:39 UTC; the New York open is still ahead. We have not verified the 157.80 print against a second source and have labelled it as one publisher’s reading throughout.
We have deliberately not published the technical ladders available for USD/JPY this morning. Two houses put out support and resistance levels, each single-sourced, and both were drawn either side of a policy event whose full move we cannot measure. Retracement levels computed over an unmeasured range are arithmetic performed on a guess.
We also do not know what the expiry board’s other strikes are, because the table is an image. And we cannot tell you whether the yen sell-off survives the New York session, which is the only question that matters for anyone carrying this into the weekend. The Commodity Futures Trading Commission’s positioning report lands at 18:30 UTC today and is measured at Tuesday’s close, which is two days before the Bank of Japan met. It will not answer it either.
Related
- FX events calendar — the scheduled releases behind these windows, in UTC.
- Signals — how we frame entries around event risk.
- EA presets — news-window and exposure settings.
