MXN, LDP, and weekend reads

Let’s talk about MXN

Sometimes I will point something out, even though I cannot explain it. Today, I do that with MXN. We had the choppy price action on the day of the FOMC, but then the day after FOMC, we saw a widespread USD selloff as risky assets rocketed across the board. Overnight, we saw a reversal of some of that USD weakness as the BOJ sounds dovish, the market is short too many USD (and long too many bonds), and the LDP election next week looms as a random JPY risk.

With bonds unable to sell off post-Fed, it makes sense that the funding currencies would get slammed. On the other hand, one would not expect MXN to underperform in an environment where commodities are ripping, cyclical FX is strong, and equities are blistering hot.

Top right quadrant shows currencies that went up both days (BRL and GBP, but barely). Bottom right quadrant shows currencies that rallied 18-19SEP then sold off overnight (the majority). Bottom left quadrant shows currencies that sold off during both sessions. JPY and CHF make complete sense. But there is MXN. Odd. I suppose one could argue it had overshot after the monster buy the rumor / sell the fact trade on pension reform. But still. Pretty strange. Explanations welcome.

---

LDP presidential election next week

A few more things about the LDP race that will determine a new Japanese Prime Minister next week.

1. Not many people are talking about it, still. I find that surprising given the key role that PM plays in swaying BOJ policy. The story for the JPY and JGBs looks different if Ishiba wins vs. if Takaichi wins.
2. The result tends to come out in the NY afternoon on the day of the vote in Japan. That means that around 3:00 p.m. on 27SEP we should get a result. Not guaranteed, but the prior vote results all came out around this time (2 p.m. to 4 p.m.).
3. The 2021 LDP vote result was a complete surprise. Taro Kono was a massive favorite in the polls with steady support around 43%/47% vs. the next candidate, Fumio Kishida. The polls were wildly off base, obviously, as Kishida was polling around 20% and proceeded to squeak out a win in the first round and then crush the runoff. He won the second round 60/40.
4. The 2020 LDP vote result was a bit of a surprise, too. Ishiba, who was polling just behind Suga, got crushed. He was polling around 30% and ended up getting just 13% vs. Suga’s 70%. Note that Ishiba is back for another attempt after bowing out mid-race in 2021.
5. This one looks like a 3-horse race. Koizumi (neutral for JPY), Ishiba (bullish JPY), Takaichi (bearish JPY).

The BOJ’s hawkish policy stance is weakly held. Takaichi could make things interesting. If you want to do an option for this event, you need to buy the Monday (30SEP) to capture the result.

If you want to get fancy, and you think USDJPY stays bid next week… Here’s an idea I really like. Very few people like doing these trades, but they can be incredibly juicy sometimes.

Sell Friday 146.25s and buy Monday 146.25s for around 6bps. The idea is that we rally towards 146 next week as the market is caught long bonds, and strong consumer sentiment and non-dovish Fed talk will push yields and the USD higher. Corporate month end on Thursday, and three bond auctions are also USD-bullish factors next week. Then, you pin the 146.25 strike 10 a.m. Friday and own a nice lotto ticket. Obviously, at such a low price, there is incredible leverage here (20x maybe in best case?), but many things need to go right.

People sometimes underestimate these trades because they look fanciful. But that’s why they are cheap. The idea is that you have a view that the path of USDJPY is not random. There is a greater chance it rallies next week and then if you are close to 146.25 on 10 a.m. next Friday, you have the chance to trade it and capitalize on the weekend gap, whichever direction. There are many ways to lose, of course. USDJPY doesn’t go up. USDJPY goes to 148. Koizumi wins and USDJPY flatlines. Etc. That’s why the trade doesn’t cost much. Nice leverage / optics but also positive EV, I think.

If you like the idea, you can also do a few different strikes like 146.00 146.50 or 147.00 etc. so you don’t have all your eggs in one basket.

---

Calednar

Forty percent of the time I type “calendar” it comes out “calednar”. The template for this weekly calendar is called: calednar.xls. Anyway. Some tier two central banks make rate announcements next week and we get feedback from the Fed (Bostic, Goolsbee, et al.) on the contested fiddy beeps.

Bostic at 8 a.m. Monday could easily be market moving and I would expect the USD to have some tailwinds next week with three auctions and two consumer confidence numbers. With gas prices crumpling, stocks and crypto at the highs, and a small drop in general malaise out there (opioid deaths finally peaked, for example)… Maybe the vibecession is over?

---

Final Thoughts

I’m still having problems with overeager spam filters. If you got this am/FX but have not been getting it lately, please let me know because that would mean that whatever I did to fix this issue today might have worked.

Weekend homework:

1. The podcast will be ready around 10 a.m. NY today. We have a brand-new YouTube channel, so you can watch a video version. Or go here (after 10 a.m.) to get the audio version.
2. 165-pager on the path from LLMs to AGI by one of the early employees at OpenAI. Provocative and alarming, but also educational and compelling. Situational awareness by Leopold Aschenbrenner. I am linking to the pdf because this is a free essay; I’m not violating Leopold’s copyright. Thanks to BB for sending me this awesome paper.
3. Kyla Scanlon on the “leaked” Mr. Beast PDF. The PDF itself is linked in her piece. It’s cool.

May you spend the weekend solving Fermi’s Paradox.

By Demetri Martin

good luck ⇅ be nimble