Stocks have been movin’ around more lately. Maybe you noticed.
Late last month, US equities finally snapped out of a somnolent lull that set in during summer’s dog days. It’s been a bit of a roller coaster since.
Why does that matter? Who cares? Aren’t roller coasters fun?
Yes. Or so I’m told. Roller coasters are fun, unless of course you get thrown off one and end up paralyzed as a result.
When it comes to equities, volatility’s good for an adrenaline rush, but from a modern market perspective it can set in motion an all-too-familiar “doom loop.”
The figure above shows you the spike in one-month realized vol over the past week. It’s 100%ile on a one-year lookback.
That’s a problem, or it can be, because to quote Nomura’s Charlie McElligott — this is his catchphrase — “volatility’s your exposure toggle.”
The most obvious example of that “toggle” dynamic’s the target vol universe. When realized vol rises — i.e., when the distribution of close-to-close spot outcomes expands — those funds mechanically cut exposure.
The figure above shows you the rolling seven-session “sell” impulse (and yes, I cherry-picked seven to get the most dramatic result for the current de-allocation). At $137 billion through Monday, the implied selling was a record for any rolling seven-session window going back at least to January of 2020 on my model.
Note that the modeled pace of selling can peak before modeled exposure bottoms. After another $9 billion in (estimated) selling on Monday, exposure was around $116 billion versus roughly $219 billion at the close last Monday, the day before Election Day.
The chart below shows you modeled exposure for this Monday (so, yesterday) along with the forward-looking scenarios for today (Tuesday).
The sensitivity analysis underscores why elevated vol’s an issue: Regardless of the directional move (i.e., up or down on the index), the model’s selling. A 2% SPX decline today would entail $5.8 billion of selling, while a 2% rally would imply $5.6 billion of selling. The only buying scenario shown is a flat close.
Of course, those aren’t the only flows in the market. Just because target vol as a monolith might’ve been a net seller on Monday isn’t in any way incongruous with (and certainly doesn’t preclude) a huge rally like the one we got on November 9. Indeed, one of the key points to grasp is that large moves in either direction can create a sell impulse from the read-across for trailing realized vol.
With that in mind, let’s talk about the rally on November 9. Note that skew and put skew steepened dramatically into last week’s fireworks.
The figures above show you skew (a proxy for downside protection demand) and put skew (a proxy for outright crash protection demand) for the semi ETF and the Nasdaq 100 ETF.
Who’s on the other side of those bought puts? Well, dealers. So, any explanation for Monday’s raucous rally isn’t complete without a mention of the chain reaction across dealer flows from a market that was over-hedged and started to run away to the upside, with implied vol collapsing along the way.
“The real world implications of a ‘VIX 28-to-VIX 20’ move over ~2.5 sessions are massive on market-maker books, which had become suddenly short puts and OTM puts,” McElligott wrote Tuesday, calling the vanna impact on Monday “absolutely massive.”
“As spot began to move higher in the morning and vol got completely scorched as all those puts lost beaucoup delta, dealers had to buy back a brick of futures into [a] virtuous feedback loop as vol gapped lower,” he went on, delineating the domino effect.
If your question’s whether that means a meaningful portion of the November 9 equity bounce was mechanical, Charlie’s answer is “yes.”
[Editor’s note: The charts here, and the models they reference, give you a sense of what’ll be in the new product I mentioned in the last Weekly. To briefly recapitulate: The product will be a standalone .pdf, sent out daily with updates on vol metrics, systematic strategies and so on, based on my new dashboards and models. It won’t be an “extra” editorial, beyond automated explanations for that day’s changes on the metrics. For early adopters, this’ll be $5/month and I’ll freeze that price for 18 months.
Importantly (and I really can’t emphasize this enough): This will not affect the on-site reading experience in any way. Again, this is a standalone .pdf, sent to subscribers directly from me each evening, and it won’t feature any substantive additional market color from me. If you don’t want it, you won’t even know it exists, and you won’t be missing any editorial color. I promise. Also, I’m not going to “sneak” it into the article feed, and at no point will anyone be “required” to buy it.
With that commitment on the books, reach out ([email protected]) if you’re interested in receiving it. I’m tentatively calling it “Heisenberg Report Analytics.” The $5/month “early enrollment” offer ends Sunday. I plan to start sending it next week. If you’ve already emailed me about this, there’s no need to email me again. I have a list of people who’ve already indicated their interest.]
