Adding another USD short
“Hike up your skirt a little more”
It’s funny how some bands from 90s are still relevant and some bands that were equally popular in the 90s (like Dave Matthews Band) have completed faded from relevance.
See analysis at bottom of page.
Short USDJPY at 161.82
Stop loss 163.17
Buy 30JUL USDMXN put fly
17.40/17.15/16.90
1X2X1 for 30bps off 17.52 spot
First up, an interesting observation from Tim Power of Spectra FX. The 1-week, 10-delta butterfly in USDJPY is doing its thing as we are in what could be described as a pegged currency/jump diffusion model where USDJPY either does nothing, grinds higher very slowly, or collapses on intervention. It’s a weird situation where if you’re short (we are!) then you are kind of hoping for USDJPY to go up and hit the MOF tripwire so you can collect your 3.5% in one day instead of waiting forever for it to cooperate via some sort of organic drop in US yields or the broader USD.
You can see that once the intervention happens, the butterfly normalizes and USDJPY becomes a real currency pair again. This kind of stuff is important to understand when you are trading a currency pair as it gives you a sense of the future distribution of returns and risk management challenges.
While I would much prefer the instant gratification that would come from MOF intervention, I do also believe there is an organic path where USDJPY goes lower on its own, and that would probably start with a weaker-than-expected payrolls figure.
The last three payrolls releases have averaged +155,000 jobs and my guess is that some mean reversion is in order as the gravitational pull of a lower breakeven has to kick in at some point. This is a low conviction call as there is too much noise in NFP to have a strong view on the miss. There is nothing nefarious going on in the job market, but you can’t have a 20k breakeven, 150k job growth, and a flat UR forever. Something has to give. If it’s the UR (lower), that will embolden the bulls but my guess is that in a no hire/no fire job market, the last three months of NFP growth probably overstate what’s going on.
I continue to believe that the USD is nearing a top, but I am trying to be thoughtful about entry points and timing. I have the short USDJPY and will add short USDMXN via options today as I feel that the technical and seasonal set up there is nice. We are entering the best three-week period for short USDMXN today, and it did a nice test of the 200-day moving averages. It’s also less sensitive to NFP risk because an okay jobs market in the U.S. is generally perceived to be good for Mexico as well. A rip-roaring NFP would not be good for short USDMXN, and there is some USMCA risk, but when all is said and done, I think USDMXN will be lower a month from now.
We never got to the 200-day (17.78), but with corporate month end USD buy flows already petering out and USD long positioning across the board getting a bit feverish, I think the timing is good here. My view is that we grind lower back towards the big base at 17.15 so…
The trade I like is: Put fly
Buy 30JUL 17.40/17.15/16.90 1X2X1
You are paying 30bps to make a maximum of 144bps. Pretty good leverage for a not huge move. If USDMXN collapses, you will lose money which is obviously the risk of doing the fly. 30JUL is one day more than one month, but I picked that date to capture FOMC (29JUL). This is more of a carry / grind trade than a strong directional one.
As mentioned yesterday, most of the corporate USD buying is done, I believe. That’s why EURUSD is rallying. It has absorbed copious selling.
Here you can see the weird calendar for next week as Friday is the July 4th sub-in holiday and so we get NFP Thursday and then an early close.
Have a grunge-y weekend.
The bands of the 90s: Peak 90s popularity vs. popularity now
I agree with most of this except Goo Goo Dolls too high and Hootie too low vs. what I observe.
Methodology: The current-popularity axis comes from kworb.net, which aggregates Spotify data; I used each band’s biggest song by daily stream rate, since that shows what people actually play now rather than just rewarding long-tenured catalogs. The 1990s axis is my own estimate from peak-era sales, chart presence, and cultural ubiquity — informed judgment, not one clean dataset, so those values could reasonably shift a few points either way. Two caveats: “popularity” here means streaming, which penalizes album-and-touring acts like DMB and Hootie whose fame wasn’t built on hit singles; and the snapshot is current to mid-2026 (Goo Goo Dolls’ “Iris” only surged recently), so it would look different in a year.
I got this idea from Alvin.
