Seven hours ago this desk published that euro–yen had gapped 64.4 pips lower at the Monday open, that the whole of it was the euro leg, and that the instrument which actually prices the cause — the French–German ten-year spread — could not trade in the Tokyo morning and so had no post-event figure. Both halves have now been marked to market, and in opposite directions. The spread still has no 19 October value and will not get one until European bonds open at 07:00, an hour after this slot fires. The gap, meanwhile, is gone: at the 05:15 snapshot euro–yen implies 177.5238 against Friday’s 177.4951, which is 2.9 pips above the close it gapped away from. The referee has not arrived and the match is over. If you were short the cross on the French headline, you have had five hours to be right and you were not.
We published 64 pips. The real low was 80, and here is the correction
Our Monday open came from an indicative quote with no time of day on it: euro–dollar at 1.1191, dollar–yen at 158.03, implying 176.8514 and a 64.4-pip gap from Friday’s implied 177.4951. The humility section of that article said to treat 64.4 as carrying a few pips of reading error. It carried sixteen.
A European-open wrap published 05:45 gives euro–dollar’s Monday range as 1.1181 to 1.1232. The low is ten pips below the indicative open we used, and 55 pips below Friday’s 1.1236. Hold dollar–yen at the open level and that low implies euro–yen at 176.6933 — 80.2 pips below Friday, not 64.4. The error is one-directional and the reason is structural: an indicative open is a quote somebody posted, not an extreme somebody traded. If you size a gap from an indicative print you will under-measure it, every time, because the print is not the worst fill.
Which makes the recovery bigger, not smaller. From the indicative open the cross has retraced 67.2 pips, 104.5 per cent of the gap we published. From the actual low it has retraced 83 pips. Either way, a weekend event that moved a major cross by somewhere between sixty-four and eighty pips had no measurable residue by the time London’s traders sat down.
The recovery is the same leg as the gap, and the yen went the same way twice
Decompose the move back the way we decomposed the move out. From the indicative open to the 05:15 snapshot, euro–dollar goes 1.1191 to 1.1230 and dollar–yen 158.03 to 158.08. The euro leg contributes 61.6 yen-pips and the dollar–yen leg 5.6, summing to 67.2 — the recovery to a tenth of a pip. The euro is 91.7 per cent of it; the yen is 8.3.
Now put the two decompositions side by side, because this is the part worth keeping. On the way out the yen leg was plus 6.7 yen-pips, against a cross that fell 64.4. On the way back it is plus 5.6, against a cross that rose 67.2. The yen leg pushed in the same direction, at almost the same size, through a sixty-pip fall and a sixty-pip rally. It was not a participant in either. Dollar–yen has spent the entire session between 157.78 and 158.13 — a 35-pip band, inside Friday’s New York range of 157.69 to 158.34 — while the cross built on it travelled a hundred and fifty pips round trip.
The operational form of that is a sizing rule and it has nothing to do with France. If your position in a yen cross is there because of a view on the Bank of Japan, a 150-pip round trip in which the yen leg never left a 35-pip band has told you nothing about whether you are right. It has only told you that you are carrying somebody else’s risk.
A second publisher priced the same weekend at 492 pips. The arithmetic says 64
At 03:50 a publisher described the same event as a gap lower in euro–Canadian dollar, from 1.6269 to a weekly-open low of 1.5777. That is 492 pips, or 3.02 per cent — roughly eight times the move we measured in euro–yen, which would be remarkable if it were a gap.
It is not. Take Friday’s euro–dollar at 1.1236 and dollar–Canada at 1.4135 from the same timed reader our other figures come from: Friday’s euro–Canada implies 1.5882. The Monday indicative legs imply 1.5818. The weekend gap in euro–Canada was 63.6 pips, or 0.40 per cent — within a pip of the euro–yen gap, because it is the same euro leg. The 1.5777 low is credible, 41 pips under the implied open. The 1.6269 is not a Friday close: divided by dollar–Canada it implies a euro–dollar of 1.1510, which is 274 pips above where the euro actually finished the week. It is a prior high doing duty as an anchor.
We are not calling the publisher wrong about France; its reasoning on the budget and the ratings review is better than ours. We are saying that a move quoted from a swing high rather than a close will be out by a factor of eight, and that the cheapest defence against it is the one multiplication we have been doing all morning. Two legs from one timed reader, multiplied, is a cross quote you can audit. A level lifted from a chart is not.
Positioning said euro shorts. The ladder we can read is three days old
The positioning snapshot behind this weekend is dated, and the date is not the one on the report. Speculative euro net short stood at 137,573 contracts, a weekly change of minus 38,241, open interest 874,541. The report is titled by its release date; the measurement is the 13 October close, which we established on Saturday by subtracting the stated change from our own prior snapshot in the yen row of the same table and landing on the published figure to the contract. So: 38,241 contracts of fresh euro shorts went on in the week to 13 October, five days before the headline, and the headline’s move was bought back inside five hours. That is the shape of a crowded short being fed, not a position being established.
The expiry ladder is where this slot is structurally weakest and we would rather say so than pretend. The board we hold for today was published on 16 October, three days in advance, by a house that posts tomorrow’s ladder alongside today’s. The house that publishes live had not put out a 19 October board when this slot ran, and a third publisher’s page for today returned a client error on two attempts. The companion piece this morning explains why that is a clock problem rather than a sourcing one.
What the three-day-old ladder says: 925 million euros at 1.1200, with six euro–dollar strikes totalling 4,692 million and the largest clip, 1.30 billion, up at 1.1300. Spot was 19 pips above that 1.1200 strike at the Monday low and is 30 pips below it now — it has been on both sides of the only near-the-money size on the board inside one session, which is precisely the condition under which a 14:00 cut matters and precisely the condition in which a three-day-old notional is least trustworthy. Dollar–yen has one row, 554 million at 157.15, 93 pips away and not in play.
Two of the three numbers we published about the weekend already have a rival
This morning we published, from two readers, that the weekend produced about 2,400 arrests and close to 1,000 injured police. A third reader, publishing at 05:45 with more of the weekend behind it, says about 1,800 arrests and more than 500 police, gendarmes and firefighters injured. That is 600 arrests, a third of the smaller figure, and a factor of two on the injuries. The turnout pair held: 550,000 by the ministry against 800,000 by the union, now two readers each.
We are not adjudicating. We are flagging that the figures we put in print seven hours ago are the high end of a range, that the spread on the injury count is a factor of two in a day-old number, and that nothing in this paragraph is tradeable — an arrest count does not have a bid. What has a bid is the spread, and the spread has not printed. Friday’s close was 137.4 basis points on two readers; a third now says “around 140” for the same week’s end, which is a rounding rather than a dispute but is a third version all the same. The German ten-year is quoted this morning at 3.42 per cent, unchanged from Friday, with Bund futures up two ticks. No publisher in our sweep quotes a French yield at all today.
The calendar makes the next three days the test rather than this morning. A French budget vote tomorrow and a ratings review on Friday are both single-sourced in our sweep and both land after the spread reopens. Today’s own scheduled business is Canadian inflation at 12:30, now on two calendars, and Chinese third-quarter growth that already printed at 4.5 per cent against 4.4 expected, with retail sales at 1.1 against 0.5.
What this does not tell you
The 05:15 snapshot is one reader, published at 05:45, and both of its legs are quoted to four decimals while our Friday baseline is a 20:10 mark rather than a close. The 2.9 pips by which the cross is above Friday is therefore inside the noise of the comparison and should be read as “flat”, not as a gain. The 67.2-pip recovery and the 104.5 per cent are robust to that; the sign of the last three pips is not.
The 80.2-pip figure is a construction, not a print: it pairs euro–dollar’s session low with dollar–yen at the open level, and nobody publishes the cross at the instant of that low. The true extreme is somewhere between 64.4 and 80.2 and we cannot narrow it further from published data. What we can say is that our 64.4 was the floor of the range, which is the direction that matters for anyone who sized from it.
On positioning, the open-interest change by currency is not broken out for the euro the way it was for the yen, so we cannot tell you whether those 38,241 contracts were shorts added or longs sold — and we have not inferred it. Nor do we know what happened to the position after 13 October; the first snapshot containing any of this weekend is measured tomorrow and published Friday.
And the honest limit on the whole piece: a closed gap is not a resolved question. The spread opens in an hour, the budget vote is tomorrow, and a market that bought the dip before the bond market could vote has taken a position on the vote rather than avoided one. We are not telling you which way it goes. We are telling you the FX market has already answered and the instrument with standing to overrule it has not yet been asked.
Related
- Event calendar — the French budget vote and Friday’s ratings review are the two dated items on this story.
- Signals
- EA presets — a cross whose base leg sits in a 35-pip band is a filter question, not a direction question.
- Euro–Yen Gapped 64 Pips. The Yen Added Seven of Them, in the Other Direction. — the article this one marks to market.
- The Notional Was in a Picture for Eight Slots. Two Houses Published It and They Are 400 Million Apart.
