Euro-Yen Gapped 64 Pips. The Yen Added Seven of Them, in the Other Direction.

The gap decomposes, and the yen is on the wrong side of its own cross

Friday’s closing marks, from a reader timestamped 20:10 UTC: euro–dollar 1.1236, dollar–yen 157.97. That implies euro–yen at 177.4951, and a Tokyo publisher independently puts Friday’s euro–yen close in the 177.50 area, which is the same number read off a different screen. Monday’s indicative open, from a third source: euro–dollar 1.1191, dollar–yen 158.03. That implies 176.8514.

So euro–yen gapped 64.4 pips lower. Euro–dollar gapped 45.0 pips lower and dollar–yen gapped 6.0 pips higher. Hold one leg still and move the other: the euro leg contributes −71.1 yen-pips, the dollar–yen leg contributes plus 6.7, and the two sum to −64.3, which is the gap to within a tenth of a pip. The euro leg is 110 per cent of the move. The yen leg is minus 10 per cent of it.

This is not a pedantic decomposition. It decides what you are allowed to put the trade on for. A 64-pip Monday gap in a yen cross looks, on a chart, exactly like yen strength, and it is the single most reliable way this channel has seen traders mis-attribute a move. There was no yen strength. Dollar–yen is six pips weaker for the yen than it was on Friday evening, and it sits inside Friday’s New York range of 157.69 to 158.34 rather than outside it. Nothing gapped in the yen at all.

The thing that priced the trigger is shut until London

What happened was in France. Weekend protests over fuel, wages and public services turned to looting and arson: the interior minister gives about 2,400 arrests, close to 1,000 police officers injured, more than 850 arrests in Paris alone. A town hall burned in Limoges and its clock tower came down. One retailer reports 50 damaged stores, 20 of them looted. Fresh clashes ran into Sunday in Paris and Marseille. Two separate posts on the same wire, stamped 20:50 and 21:25 UTC on Sunday, add a Romanian drone incident over the Black Sea and an American threat of retaliation over European satellite spectrum rules, and call the three of them together the reason the euro opened soft.

Here is the problem with trading any of that this morning. The gauge that prices French fiscal risk is the OAT–Bund spread, and one of those posts says so itself — and then gives no figure, because there is no figure to give. European government bonds do not trade in the Tokyo morning. The most recent published spread is 137.4 basis points, as of the 16 October close at 17:35, with the French ten-year at 4.79 per cent and the German at 3.42. Two separate readers give that 137.4 and both say it rose 2.2 basis points on the session. It is a Friday number. It was set before any of this.

So the instruction is narrow and it is a timing instruction rather than a directional one. You have 64 pips of euro weakness priced in the thinnest liquidity of the week against zero basis points of repricing in the instrument that is supposed to justify it. Either the bond market confirms it when it opens or it does not, and you will know in hours, not minutes. Sizing a position now is paying for the privilege of being early to an observation you can buy at a discount later. We are not going to tell you which way the spread goes. We are telling you that the FX market has already voted and the referee has not arrived.

Every number attached to this weekend is published twice

Three pairs, and we are not going to pick a side in any of them.

The October peak in the spread. One publisher, writing on 18 October, says the spread “reached 152 basis points, the widest since the 2011–12 euro crisis.” A second, also on 18 October, says it reached about 150. The only dated daily series we read puts the October peak at 146.9 basis points on 2 October, and lists every session since. The narrative figure is 5.1 basis points above the highest value in the table, and the claim resting on it — widest since the euro crisis — is the kind of claim that gets repeated for a month. We report the band, 146.9 to 152, and note that only one of the three sources shows its working.

What record the French ten-year broke. Both narrative readers have it near 5 per cent. One calls that the highest since 2002. The other calls it the highest since 2008. The level agrees and the record is 24 years apart, which means at least one of them has not checked and possibly neither has.

How many people marched. The interior ministry counted more than 550,000. The CGT union says about 800,000. That is a gap of 250,000, or 45.5 per cent of the smaller figure, and the honest thing to say about it is not that one is lying but that neither number is tradeable. A crowd count does not have a bid. We report it because the ministry figure and the union figure are the two inputs every write-up of this weekend will use, and anyone reading one of them without the other will think the number is settled.

The yen long was already being cut, 16 hours before the print that was supposed to move it

Friday evening’s positioning report landed and one reader published it on Sunday at 22:16 UTC. Speculative yen net length: 50,207 contracts, a weekly change of minus 12,118, open interest 373,502.

That reader does not state the measurement date. It does not have to, because the arithmetic does. This desk published a baseline of net long 62,325 as of the 6 October close with open interest 368,138. Subtract: 62,325 minus 12,118 is 50,207, to the contract. The as-of date is therefore the 13 October close and not the 16 October release date the report is titled with, and we did not have to trust a label to establish it. When a vendor labels positioning by release date, the prior snapshot plus the stated change is a free check on which Tuesday you are actually looking at.

Now the content. The work this desk set itself on Friday was to find out whether net length had extended past 62,325. It had not. It was cut by 19.4 per cent of itself, in the week ending at the 13 October close — which is roughly 16 hours before the 14 October inflation print that doubled the market’s October probability for the Federal Reserve. The position was being reduced before the catalyst, not after it.

One thing we cannot do, and will not pretend to. This reader publishes net only, with no gross long and short split, so whether longs were sold or shorts were added is not determinable from it. What we can say is mechanical: open interest rose by 5,364 contracts, 1.46 per cent, while net length fell by 12,118. Pure liquidation of longs shrinks open interest. Open interest going up while net length goes down is more consistent with new shorts arriving than with old longs leaving. That is an inference from two aggregates, it is not the gross data, and the first snapshot that contains any reaction to the 14 October print is measured at tomorrow’s close and published on 23 October.

And a correction: we read the page that had today’s expiry ladder on it, and reported half of it

On Friday this desk published, at some length, that two houses had finally put option notionals out in plain text and disagreed by 400 million euros at the largest strike. The source we called the most valuable fetch of that run was a blog post stamped 16 October 09:55. Its title names two dates. It carried a second table, for today, and we did not report a row of it.

Here it is. For the 19 October New York cut, euro–dollar has six strikes: 1.1175 at 510 million euros, 1.1180 at 633 million, 1.1200 at 925 million, 1.1260 at 762 million, 1.1280 at 562 million, and the largest at 1.1300 with 1.30 billion. Dollar–yen has exactly one, 157.15 at 554 million dollars. Australian dollar–dollar has four between 0.6950 and 0.7010 totalling 2,636 million. No cut time and no timezone are stated on the post, as was true of the same source on Friday.

Set the Monday open against that ladder and the gap has consequences. Euro–dollar at 1.1191 sits 9.0 pips below the 1.1200 strike and 11.0 pips above the 1.1180 one — inside the lower cluster, with 1,558 million euros of notional in the two strikes straddling it. The largest strike on the board, 1.1300, is now 109.0 pips above spot and about as relevant as a weather report from another country. Dollar–yen’s single strike at 157.15 is 88.0 pips below spot. The second ladder we would normally read for a cross-check has not published for today yet; its bare index page serves a 29 January article, which we discarded on its stamp rather than on its content.

We are flagging our own omission rather than quietly using the table, because Friday’s other article ended on the rule that a negative claim about publication is a claim about your own sweep. This is the same failure one step earlier: we opened the right page, took what we were looking for, and left the rest. Reading a source is not the same as reading it.

What this does not tell you

It does not tell you where euro–yen goes. It tells you that if it goes anywhere today the euro leg is where the move will live, and that a yen view is close to irrelevant to it.

Monday’s open prices are indicative, from one source, in very early trade, and a gap measured against a 20:10 UTC Friday mark is a gap between two readings rather than between two closes — this channel has published a close three separate ways inside one session before. Treat 64.4 pips as a figure with a few pips of reading error around it, not a measurement.

The spread figure is a Friday close and the French yield history is contested, as above. We have no post-weekend bond price of any kind. We have no cash equity reaction either: Hong Kong is shut today for the day after Chung Yeung, so one of Asia’s two deepest pools is absent from the price discovery, which is part of why a 64-pip gap should not be read as 64 pips of conviction.

And on Friday’s New York cut, which this desk twice declined to score: we are retiring the question rather than answering it. Price was 41.0 pips above the 1.1225 strike at 10:54 UTC and 11.0 pips above it at the 20:10 close, so it spent the afternoon walking back toward the band. But no source in this sweep publishes a print at 14:00 UTC, which is the only observation that would settle it. A close is not a cut. Two slots deferred it and the third is closing it as unscoreable, because pretending otherwise would be worse than admitting the data does not exist.

Today’s board, in UTC: Chinese third-quarter output at 02:00, inside the Tokyo session; Canadian inflation at 12:30; an American housing survey at 14:00. There is no scheduled Japanese release in the Tokyo session and no central bank decision anywhere this week. Federal Reserve officials entered their blackout on 17 October ahead of the 27–28 October meeting, and for a sixth consecutive slot not one published probability figure for that meeting appeared anywhere in our sweep.

Related

• Scheduled events, in UTC — the week’s releases without a timezone to convert.
• Signals — what we are watching rather than what we think.
• Presets — filters and news windows for automated systems.
• The notional was in a picture for eight slots — Friday’s expiry article, and the page whose second table we missed.
• Three positioning numbers, all broken differently — where the 62,325 baseline came from.
• A negative claim about publication is a claim about your own sweep — the rule this article applies to itself.