Into the Nosebleeds

Short volatility strategies are reflexively bullish and will return this year in force, but in ways that the market is not yet familiar with. Dozens of “short vol” expressions exist, but dispersion strategies are set to dominate equity volatility desks and thus equity markets. This is one of the several dynamics that have created extraordinary concentration in stock indices, whose impressive returns have been driven by MegaCap Tech like Nvidia & Microsoft.

Properly understanding this trade will make concerns over euphoria seem subjective. An outlook that's based on positioning or sentiment (everybody being very long the same things) is changed when you realize that those very concentration risks are supportive of many dispersion strategies, paradoxically creating concentration elsewhere. Likewise, thinking about risks to dispersion strategies, of which there are several, is crucial to timing a sharp and aggressive market correction.

The infamous February 2018 “Volmageddon” volatility spike actually did force short vol traders to flee, but the recent July/August volatility spike saw the opposite – flawed volatility strategies were discarded and imperfect ones refined, inviting more money to be thrown at it. At the end, I share exactly which themes and which names are most likely to see bids.

This trade is key to understanding why markets are the way they are today and, more importantly, where they’re headed in the coming year.

– Mike on X

The short volatility trade means a lot of different things, including but not limited to what everyone knows it as ("shorting VIX"). It's been around for years, infamously blowing up exactly seven years ago in what vol desks know as Volmageddon. The trade has been widely and accurately credited with supporting bull markets, both in individual stocks (“single names” like AMZN) and the indices they are a part of (like the S&P500). The reason is because volatility strategies, as their name implies, generally don't have a strong bias on the price. Instead the focus on volatility nearly eliminates directional risk, creating hedging flows that are generally self-fulfilling, something I'll briefly touch on in a moment.

On the other hand, when short vol strategies appear to implode (the notorious example is when the VIX 'fear gauge' spikes), that often feeds directly back into equity markets with forced selling:

Volatility traders think of volatility as an asset itself – something adjacent to but independent of whatever it's measuring – and vol is said to be ‘sold’ by selling options contracts. As most of you’ll know, options are equity derivatives: if I buy a Feb 220 AAPL call option, that gives me the right to buy 100 shares of AAPL for $220 on or before the contract expires on the third Friday in February. Most individual traders who buy call options will do so with the goal of getting cheap leverage for a bullish bet. Keep in mind that most people who do that don't consider volatility at all, which is basically the premium that they will inevitably be overpaying for.

Buying a Feb 220 AAPL put option would give me the right to sell 100 shares of AAPL for $220 on or before contract expires on the third Friday in February. Similar to calls, most individual traders who buy puts do so with the goal of getting cheap leverage for a bearish bet – doing that is what I'd call buying a lottery ticket.

Here, using the AAPL options, $220 is the strike price, and by buying a call or a put I'm buying the right to exercise the buy or sell trade for $220 (i.e., at the strike).

It should be obvious that the minimum a 220 call option on AAPL stock trading at $223 should cost would be $3, because otherwise it's a free lunch: if it only cost $1, I could buy it, immediately exercise my right to buy 100x AAPL for $220 before selling those 100 AAPL shares for $223 – a guaranteed profit. That $3 is the intrinsic value, and it’s already baked into the price of the contract as it's said to be $3 in-the-money (ITM), though most would say it's close enough and call it at-the-money (ATM):

If the market were 100% certain that AAPL would be at $223 by the date the option expires, the 220 AAPL call option would cost $3 and not a penny more. And indeed, that’s the contract’s value the instant before it does expire. But in reality, the market is never 100% certain of anything, and the options dealer who sold the contract demands compensation for his uncertainty. That manifests as time value – a seller would charge more for a 220 AAPL call option expiring in July than he would if it were set to expire next Friday. Along with uncertainty about what can happen over more time, there’s the other more obvious dimension of uncertainty about what the price can do, measured by the volatility or implied volatility.

Implied volatility (IV or "vol") is imperfect information derived from the stock’s options market. It’s important to note that vol does not determine option prices, option prices are what determine vol. When investors are buying options, which drives their prices up, implied volatility increases. When they are selling options, which drives down prices, implied volatility decreases.

Consider an important event is unfolding tomorrow, like an earnings call, and there's a wide range of uncertainty. Investors may buy options to hedge against adverse moves related to the event. But as soon as said event subsides, unless the adverse moves are realized, vol is crushed as demand for options stalls and investors sell their hedges. If you compare two different stocks that have the same price, and one's ATM options are trading at a much higher price than the other stock of the same expiration, you know that the higher price indicates higher vol. Thus, vol reflects the actual bid for the option – the actual demand – which in most cases is used for protection. Hence why vol (and vol indices like VIX) is called a 'fear gauge'.

Both time to expiry and implied volatility are also fed into the option’s value, which is the only reason the Feb 220 AAPL calls cost about $9 – along with $3 in intrinsic value, the other $6 is its extrinsic value – time and implied volatility. That means every options trade is expressing an outlook on the volatility –  option buyers think the market is underpricing vol, option sellers think vol is overpriced – even though this is lost on most individuals who trade options, and probably why most lose money.

The purest way to short volatility is to short a straddle – shorting both a call and a put option of the same (usually at-the-money) strike. The purest way to go long volatility is to buy a straddle:

But selling options on their own – 'naked' selling – exposes the trader to directional risks. What if I sell a straddle and the market proceeds to sprint far away? Even though I should be compensated for that risk because I sold the IV, the potential loss is unlimited – am I just f-cked? Short vol strategies mitigate these directional risks with dynamic hedging – buying and selling to keep neutral the position's exposure per $1 move up or down in the stock. But that concept is, for now, beyond the scope of what I wanted to discuss.

Because the point is not to go any further into the very technical (and frankly very boring) details of these derivatives and how they're priced. The point is to look at how short vol strategies can and likely will continue to affect equity prices. There’s one expression of short vol that has reemerged as a dominant theme, growing bigger-than-ever over the last couple years, and it’s called index dispersion.

The definition of dispersion is the process of distributing something over a wide area. When it comes to index dispersion, that means distributing the underlying “single names” within a stock index over a wide range of performance. In other words, dispersion implies a breakdown in correlation (when half of the index goes one way, and the other half goes the other way). The lower the correlation, the higher the dispersion.

A handful of stocks (like the 'Mag 7' & 'Fab 4') carrying index returns higher, even as entire sectors head lower, is a market that's high in dispersion – ideal conditions for the trade. Here's 2023, a year where correlation plunged (the actual S&P500 index returned 25% – right in the middle):

Noel Smith describes dispersion using the analogy of dropping a handful of marbles:

The simplest way to explain how dispersion trades work is by selling volatility on the index and buying volatility in the single names. This trade is based on the assumption that index vol is often overpriced relative to the vol of its constituents. The equivalent would thus be selling straddles on S&P500 futures and buying straddles on NVDA, MSFT, AAPL, AMZN and to a lesser extent the other 496 stocks listed on the S&P500:

Why would less correlation make for a more ideal environment, then? Remember, because most dispersion traders are long equity volatility and short index volatility, the key is for the equity volatility to, in a sense, "cancel each other out." A $1 trillion volatile swing up is, to put it a different way, "neutralized" by 100, $10 billion volatile moves down (neutralized as far as the index is concerned).

Imagine one big name, let's say NVDA, were to miss earnings and lose 5% in a few minutes, single-handedly shedding $150 billion off the S&P500 index. That's called idiosyncratic risk, and would be a positive for long NVDA vol since the options market wouldn't have priced in a huge flock of NVDA longs eagerly buying put option protection (buying vols).

How is this scenario positive for short S&P500 vol? It's not, unless that $150 billion drawdown were to replace itself elsewhere on the S&P500, which could mean back into the Mag 7 like it did last year, 200 smaller names like it mostly did in 2023, or some combination of the two. It often does as a result of those dynamic hedging flows – hence the trade is said to be reflexive.  Fear in one part of the index creates equal and opposite excitement in different parts, and the result is for the actual index to be "pinned":

Cem Karsan is not a savant (to me he's less of a market wizard and more of a marketing wizard), but he explains this "if one dollar goes one way, another dollar goes in the opposite way" concept well:

The way dispersion looks most likely to play out in 2025 is one where concentration in MegaCap Tech and the Magnificent 7 gets rebalanced across the entire AI theme:

The recent development over the weekend with DeepSeek makes no better example of this, where AI-dispersion has thrived: NVDA and other AI hardware (AMD, ONTO) & power (VST, CEG) names lost several hundred billion dollars in market cap overnight (sending equity vol soaring), the S&P500 index remained steady in the day session (crushing index vol), and each of these less concentrated names within the broader theme found strong bids as AI becomes less of a monopoly:

1. Snowflake (SNOW) – Provides a cloud-native platform for data storage sharing and analytics at scale. Their platform becomes increasingly vital as AI applications require more diverse and scalable data access.
2. Confluent (CFLT) – Offers a managed Apache Kafka platform for real time data streaming and processing. Their platform is essential for integrating AI models into streaming environments.
3. GitLab (GTLB) – Provides an integrated DevOps platform for software development testing and deployment. Enables seamless integration of AI models into development processes.
4. JFrog (FROG) – Manages software artifacts and ensures secure delivery in CI/CD pipelines. Ensures reliable updates and version control for AI components.
5. ServiceNow (NOW) – Delivers cloud-based workflow automation across IT HR and business functions. Makes AI-driven automation accessible to various business functions.
6. Elastic (ESTC) – Offers a search and analytics platform for real-time data insights, particularly in cybersecurity and business intelligence. Supports data-driven AI systems with powerful search and analytics.
7. SAP (SAP) – Provides comprehensive ERP software for managing business operations and improving decision-making. Makes AI-powered features accessible to enterprises of various sizes.
8. HubSpot (HUBS) – Offers integrated marketing sales and CRM tools for SMBs. Makes sophisticated AI features accessible to smaller businesses.
9. Klaviyo (KVYO) – Specializes in e-commerce marketing automation through email and SMS. Enables predictive customer behavior analysis and hyper-personalized messaging at scale. Makes AI-powered marketing automation accessible to e-commerce businesses.
10. Zeta Global (ZETA) – Provides marketing technology for personalized campaigns and audience segmentation. Optimizes marketing strategies through automated data analysis and precise recommendations. Enables sophisticated AI-driven personalization for businesses of all sizes.
11. Five9 (FIVN) – Delivers cloud-based contract center solutions for customer engagement. Makes AI-driven customer service solutions cost-effective for businesses.
12. Freshworks (FRSH) – Provides CRM and IT management tools focused on SMBs. Automates customer support and IT workflows for increased efficiency. Delivers affordable AI features to growing businesses.
13. Meta Platforms (META) – Operates social platforms and develops AR/VR technologies. Delivers tailored content and ads in real-time while fostering AI innovation through open source frameworks like PyTorch.
14. UiPath (PATH) – Provides robotic process automation (RPA) solutions for task automation. Enables cognitive automation for complex tasks requiring decision-making. Makes advanced AI tools accessible for enterprise automation.
15. Pegasystems (PEGA) – Offers low-code BPM and CRM software for enterprise workflows. Makes sophisticated automation accessible at scale.
16. Dynatrace (DT) – Provides observability and monitoring solutions for IT systems. Makes advanced observability accessible and affordable.
17. Procore (PCOR) – Delivers cloud-based construction management software. Makes AI-driven tools accessible to construction firms of all sizes.
18. FiscalNote (NOTE) – Provides tools for monitoring policy legislation and regulatory changes, enabling prediction that makes advanced monitoring accessible to smaller organizations.
19. RELX (RELX) – Offers analytics and decision tools for legal scientific and risk management. Makes advanced analytics accessible to legal and research teams.
20. CCC Intelligent Solutions (CCCS) – Provides software solutions for automotive insurance, claims processing and fraud detection.
21. Cognizant (CTSH) – Provides IT consulting and outsourcing services for digital transformation. Makes AI implementation accessible through cost-effective frameworks.
22. TaskUS (TASK) – Offers outsourced content moderation and customer support services. Makes AI-driven services accessible to clients with smaller budgets.

... as the rest of Mag7 outperforms NVDA, keeping the S&P500 index stable and "in check"...

Consider that whether the S&P500 goes up or down this year is not factored into how well the trade does, as many of those 22 adjacent names listed above are too small to have a jarring upside effect on the index. As disagreeable or unrealistic the AI theme taken as a whole may seem, with MegaCap Tech valuations already "in the nosebleeds" and equity markets top heavy, this is objectively how hedge funds will be retrenching their short volatility trades in 2025.

Time will tell if it was a bad idea to do so.

– Mike on X