Shutdown ending
The shutdown is over, the House will vote to end it tomorrow, and 13NOV25 should be the reopening. If you use the 2013 analog, or listen to GS, MS, and Jefferies, you would expect September NFP to come out two or three business days after reopening, which puts 18NOV as the likely date with 17NOV as a lower, but not zero likelihood. I suppose Wednesday is possible, too.
Given we are starved for data, I would think this is a great date to own and I would be buying USDJPY or fixed income vol for 18NOV expiry if you can. Once they announce the date of NFP, it will be too late, so you need to take a shot before the announcement. Granted, September payrolls feel a bit dated at this point, but if that number were to print negative, the reaction will be significant. In contrast, given the weakness apparent in various series, the market will probably mostly ignore a strong number and look forward to the October and November data to be released (maybe) 05DEC.
In FX, we had an unusually large move lower in EURCHF on the back of some tariff back and forth and AUDNZD had just its second down day in the past 14. Ironically, this comes as metals explode higher. When AUDNZD trends up hard like this, then has a down day, it’s a bit of a bearish setup because AUDNZD is fundamentally a mean reverting currency pair and so once momentum cracks, even a bit, it often stalls or goes lower. Filtering for “Was AUDNZD 2-day move UP at least 10 days in a row + today AUDNZD is lower than yesterday” the results are:
These are not dramatic results, but do indicate a mild bearish expectation from here that makes me feel a bit less bad about the AUDNZD short from 1.1494. There’s a tiny negative trend in the data as the lookback is to 2000 and the win rate for AUDNZD in the full sample is 48.9% because AUDNZD has gone from 1.27 to 1.1550 in that period.
Earlier I said: “AUDNZD is fundamentally a mean reverting currency” and I think most people would nod in agreement with that statement. But there is a super simple screen I use to confirm whether my suspicions on various currencies are true. I would consider USDCAD to also be mean reverting, for example, but not USDJPY. To test it, you can run a simple test of the strategy: If it went down last week, go long / if it went up last week, go short. Here are the results. I include SPX to show what a super trending product looks like.
Hypothesis confirmed. USDJPY is the trendiest, while USDCAD is the least trendy. AUDNZD and EURUSD are not particularly trendy either. This sort of test shows broad characteristics of the asset, but can be regime sensitive. USDJPY trended hard around the GFC and Abenomics, for example, while it was mean reverting 2022 to 2024 due to intervention and large corrections in the up trend.
ADP
We had an interesting information arbitrage today as Bloomberg did not publish the weekly ADP number. It’s a new series, and as the market is absolutely starved for data, it matters more than it might otherwise. The figure came in negative, which is weak. It doesn’t look like much of a delay, but USDJPY sat around for three minutes after the release before people started to hear about it.
I am sure Bloomberg will have it next week. Here’s the ADP chart:
This adds to concerns about the jobs market on top of Michigan unemployment expectations (see chart via UBS), JOLTS, Challenger (see am/FX: Challenging Jobs Data), ISM Employment, Conference Board data, etc. This further boosts the importance of the next few NFPs.
I am grudgingly starting to believe that the US jobs market may not be creating enough jobs to break even. This despite break-even NFP collapsing down to something around 30k/50k.
Final Thoughts
CRWV and NBIS are the two superstars of the neocloud hyperscaler universe and both threw up bricks on earnings in the past 24 hours. CRWV is lower, while NBIS is slightly higher, much to my surprise. Softbank selling its NVDA stake to raise cash is another of the many question marks around the AI trade right now as a buoyant fundraising environment would not have required them to raise cash via liquidation of assets. Short the AI complex is hard work but still looks logical to me. The margins and profitability of CRWV and NBIS are not as great as expected, and they require mega cash to manage huge expenses and large debt loads. More issuance and secondaries coming from these guys.
I find that Gemini is just as good as ChatGPT and at some point, I wonder if Google will just crush OpenAI with some kind of ridiculous pricing structure to try to squeeze them before OpenAI can monetize via pay-to-play query results and click-to-pay advertising.
Crazy times.
This Q3 letter from Horizon Kinetics is chock full o’ zesty tidbits.
Have a reflective day.
