Dean: Hello, this is Dean Curnutt and welcome to the Alpha Exchange, where we explore topics in financial markets associated with managing risk, generating return and the deployment of capital in the alternative investment industry.

Welcome, my friends, to the machine. Whoops. I could not get that epic 1975 Pink Floyd masterpiece out of my head. Let’s try that again. Welcome to the Alpha Exchange podcast on volume and risk. What follows is some updated thinking about the world of market prices, how they are arrived at, what they imply, and the behaviors they both encourage and discourage. Before we dive in, please permit me a short detour with a word about MacroMinds, the nonprofit I launched in 2019. They say that diversification is the only free lunch in markets. Scatter your bets around and you’ll realize a reduction in volatility that helps you manage risk. That’s what’s been happening at a truly momentous scale in US equity markets. The one month correlation among stocks in the S&P 500 is, to quote Dean Wormer from Animal House, 0.0. But I’d argue that today’s index and the trillions of dollars that track it are enjoying an unsustainable run of never seen before levels of low correlation among stocks. That is just unsustainable. For me, a free lunch is not one that is here today but potentially gone tomorrow. Diversification is proven to be the most fair weather of friends disappearing when you need it most during a market shock as correlations reliably surge to one.

This reality doesn’t mean you shouldn’t seriously focus on diversification. You just need to be mindful of its limitations. The magical way in which stocks have been uncorrelated cannot and will not last. So about that free lunch. I believe a proper one ought to deliver benefits today, be durable to market shocks, and compound over time. I suggest that an investment in building relationships with other financial market professionals is just that. Let’s be honest, there is a ton of ground to cover in markets across asset classes, product types, geographies, and trading strategies. It may be the same in every other industry, but in finance it’s surely the case. Relationships with industry colleagues built on trust and mindshare are a source of capital that can yield valuable returns. Relationships are capital this idea that our industry is at its best when we interact in person is one of the reasons I founded macro Minds in 2019 as an initiative to provide support for organizations that focus on student education. Our symposiums bring investment professionals together to hear from some of the sharpest minds in finance while engaging with each other, sharing ideas and developing rapport. We do this against the backdrop of goodwill and inspiration that results from knowing that our time spent together is supporting promising students.

Our 2026 event is scheduled for June 4th in New York City. I feel so fortunate to have the opportunity to host it again. Our fifth event, and in the spirit of the Power of Compounding, it will be our best yet. More insights, more relationship building, more engagements, more results for students. Oh, and more laughter as well. Laughter, as Milton Berle once said, is an instant vacation. If you can’t tell by now, I’m excited. We’ve already got an incredible lineup of world class speakers who have taken up my ask to participate and are ready to contribute their insights. More on this to follow as we build out the program. The goal of the 2026 symposium is nothing less than a thought leadership convention, causes, content and collaboration. I’d love to have you and your firm involved in some way and I thank you for listening. Okay, now back to the free lunch and the topic of turbocharged diversifications that exist right now in markets. As mentioned, realized correlation among stocks in the S&P has been absurdly low. It’s a risk that has been and continues to hide in plain sight. I will argue that low correlation is the definitive risk in markets today.

It is the accident waiting to happen. We all know that Today’s S&P 500 has an epic degree of concentration. The combination of high concentration and low correlation is without precedent. Of course you need really low correlation among the MAG7 to get to such a low level of correlation in the S&P at large. That’s exactly what we see. And rather than take my word for it, hit up the Alpha Exchange Twitter account. The handle is alphaexllc to see this chart and much more for yourself. What we see in the chart confirms intuition, volatility and correlation are themselves highly correlated. Stocks become more volatile and more correlated at the same time. This hampers efforts to diversify through stocks alone. Now I’m no AI guru, nor am I at any risk at all of chatgpt psychosis. But I am good at asking questions. And what is strange to me is that the sagging MAG7 correlations are being realized even as the Super Caps are highly interconnected business wise. They are both customers and suppliers to one another, generating costs and revenues all also to one another. And more recently, they are actually investors in each other.

If you’ve got a Bloomberg terminal, hit up the stupendous function called splc, a not so easy to remember mnemonic for supply chain analysis. When we run it for Nvidia, we see that its five biggest customers are Microsoft, Meta, Super Micro, Amazon and Alphabet. Together, these five companies are responsible for half of Nvidia’s revenue. I guess we all know who’s picking up the tab at the next CEO dinner. Half year revenue from five clients. This makes the S&P concentration look modest by comparison. Microsoft alone is just under a fifth of Nvidia’s revenue, and Nvidia constitutes 47% of Microsoft’s costs. And yet the two month realized correlation of daily returns for these two beasts is 35%. For comparison purposes, Wells Fargo, JP Morgan and Citibank are all in the neighborhood of 75 to 80% correlated. And that’s in a sideways market. Meta and Nvidia are similarly entangled. Meta is almost 10% of Nvidia’s revenue, which constitutes 25% of Meta’s costs. The two month realized correlation among the two 14%. The customer supplier overlap has recently taken on a new dimension. Investments Nvidia recently invested $100 billion. That’s $100 billion in OpenAI. While the disclosure is not direct, it is suggested that OpenAI may account for 15 to 23% of Nvidia’s revenue.

A recent piece in Fortune magazine is entitled and asks the following provocative questions. Nvidia’s 100 billion OpenAI investment raises eyebrows in a key how much of the AI boom is just Nvidia’s cash being recycled? It goes on to say there are so many interlocking rings of circularity where Nvidia has invested in a company such as OpenAI that in turn purchases services from a cloud service provider that Nvidia has also invested in, which then also buys or leases GPUs from Nvidia. In a recent Monetary Matters podcast, Rob Arnott, founder of Research Affiliates, makes some comparisons to the loans that Cisco made to telecon companies that purchased its networking gear. I put purchased in quotes. Were these really bona fide sales that could be counted as revenue circa 2000? Arnot also tells us that the concentration of the S&P, the top five currently constituting 28% of the index, peaked at just 17% during the dot com bubble in 1998. LTCM was dubbed the central bank of volatility. The fund’s enormous leverage made its way into all corners of the market, serving to depress risk premiums as it hoovered up nickels around the world. As the compensation from traditional strategies thinned out, they got themselves into risk arbitrage, taking an indexation approach of doing every deal in impressive size.

They did so via equity swaps, renting dealers balance sheets and paying handsome financing fees to them for the service. Those profits were some part of the profits engine that allowed the dealers to take on more risk. The dealers got larger and larger. You see where this is going Back to correlation. As I’ve said, you simply can’t trust correlation. One of the biggest blind spots for investors is the expectation that this once in a lifetime run of equity market diversification via these low correlations will continue. History simply suggests otherwise. Let’s dig in on this a little bit and talk about the dispersion trade first. While I do not like their VIX Premium product fees one bit, I do appreciate the CBO’s VIXification of just about everything. Name an underlying and the CBOE has VIXED it. Credit volume has the VIX IG EMVOL has the VX EEM Index Bond Vol. Vxtlt Russell Vol’s got the RVX index, Vola Vol. The VVX Brazil volume the VX ewz. You get the picture. So I was pleased to find the VIX EQ index, the S&P constituent volume that is one month weighted average single stock volume for members of the S&P.

They apply the VIX methodology to a series of single stocks weighted based on the S&P weights and out pops a number that is what the dispersion trade buys against selling S&P. Vol. The dispersion trade for review is pretty simple. Buy a series of single stock options Nvidia, Microsoft, Apple, Amazon, etc. And recoup much of the premium by selling S&P volume against it. When you get highly idiosyncratic moves in the stocks that do not translate into large moves in the index, you capture profits on the single stock volume you own while not losing as much on the index volume or short call these anti correlation profits. When an event like the tariff tantrum hits in April and you get successive 5% down moves in the S&P, things do go awry. The dispersion trade leaves you short a burst of highly correlated moves in stocks that do result in the index moving a lot. Highly correlated stock moves like death, taxes and giant US treasury auctions are a certainty in life. There will surely be another correlation episode. It’s not if, but when. That reality is clearly not good for the dispersion trade, but the profits that can be banked during interim periods when correlation is low can make up for those unpleasant spikes.

But we do know that Quote correlations go to one as they say in a big risk off. For the dispersion trade, the biggest defense is the entry price. That is if you put the dispersion trade on at a high enough implied correlation, you give yourself margin of safety. And that brings us to today’s pricing of this trade, which has very very little of that. We can see that by comparing the VIX EQ index to the VIX index. As I showed in the chart I posted on Twitter, you are paying on the order of 20 + Vols more for one month single stock volume versus index volume. That’s how you arrive at these very low levels of implied correlation. Now as with most everything in options markets, pricing is justified by carrying one month realized correlation. Dean warmer style as I said, is basically zero. So this trade, despite being put on at very thin margins of safety, is carrying just fine, thank you. The epic and magical internalized diversification on the S&P matters not just for the mathy nerds that constitute the dispersion crowd. The low level of realized correlation is a massive depressant on overall S&P realized volatility.

Move realized correlation back into the 20s and 30s and that alone adds 4 to 6 vols to S&P 500 volume. And we know of course that stocks become more correlated and more volatile at the same time. It’s just how markets and the economy work. A while back I created a little framework that summarized the factors driving equity implied volatility. I called it the five Cs carry, credit, calendar, concern and Capital. Each of these played a role in explaining why options prices were where they were. Of the five carry is the most prominent. Show me an assets realized volume and I’ve got a pretty good idea where it’s implied will be certainly not perfect. And that’s where the other four Cs come in. The marginal price setter of an option is beholden to his or her capacity to capture the swings in the underlying. When realized volume sputters, the replication portfolio that trades in and out of the underlying fails to deliver enough profits to support option prices. There is a pull lower in option premiums as a result. While they can’t last forever, there are dislocations between realized and implied. We saw this in the period leading into the US 2024 election when event risk kept volume super elevated on days on or after November 5th.

That’s the calendar aspect of the five Cs today is also one of those divergences. First, let’s set the stage by reviewing the substantial decline in realized volume. Our market stability index, which captures the number of days the S&P fails to move by 75 basis points or more in either direction over a three month period, is running at 84%. That’s more than four. In five days, the S&P fails to move up or down by 75 basis points. The flip side of this is the Gamma index, which shows the losses accruing from owning weekly S&P straddles. A third stat is that S&P3 month realized volume down days is just 7.6%. That, my friends, is incredibly low. It’s clear realized Vol is highly muted. What is strange, however, is the lack of impact that this huge decline in realized has had on implied volume. The second VIX future, for example, is barely lower over the last three months, even as realized volume down days has fallen by 24.4vols. The normal gravitational pull has not exerted itself. And it’s not like the US has an election coming up in November. While the forward looking outlook is, as it almost always is, littered with monetary and macro uncertainties, we cannot point to a on the calendar source of risk that explains the stickiness of the VIX.

Attribution is a most popular sport in high finance. We all search for the proximate cause of change, or in this case, divergence. I showed on Twitter recently the dreadful performance of three popular VIX ETPs, VIXy, VXX and UVXY. They all simply buy and roll VIX features. They rallied sharply during the April tariff tantrum, but have lost consistently ever since, much like the Gamma index. The losses, however, would be even greater still had the realized volume shortfall been more effective in pulling implied volume lower. Curiously, even as these funds have had negative returns, they’re gaining assets. And that’s because they are seeing substantial increases in shares outstanding, picking up $1.25 billion in AUM. This would be the capital component of the 5Cs. That is, when there is capital moving into or out of an asset, it can have some impact on the clearing price. In this case, it can be argued that the demand for VIX futures is keeping the VIX high, even as the significant shortfall of realized volume tries to pull it lower. As you know, I think our need to connect the dots in markets can be a bit much. Some moves are just noise, things just happen, and it’s difficult to explain why.

But the sharp increase in assets on long VIX futures products seems a reaction to the explosion of volume in early April when The VIX surpassed 50. There are only two precedents to this. The GFC and Covid and both saw the VIX reach 80. Perhaps it’s the case that the massive legion of financial advisors started putting VXX and its brethren into client portfolios during that time period. So let’s put all this together. Both S&P implied volume and implied correlation are low, but their realized cousins are even lower. In fact, as I said, with the VIX index trading at 16, it could justifiably be 12 or 13. Given the meager level of trailing realized, especially on down days in the S&P, let’s split the difference and instead of pricing a one month 97% of spot put on the S&P at 16 volume, let’s use 12.5 volume. The cost of that put is reduced by almost half. That is quite a bit. So it’s no slam dunk here for me to buy volume, neither through index puts and certainly not through VIX calls. It rarely is the entry price for insurance matters a ton.

I’m left continuing to favor the put spread instead of the outright put. There’s a real cost reduction that’s achieved by selling off the bottom leg and set against today’s high volume risk premium, it feels somewhat necessary. The exercise at hand is to recognize the vulnerability that arises from the combination of high concentration and low volume and low correlation. It’s especially important to recognize this vulnerability given just how much implicit and explicit benchmarking there is against the S&P. In our industry, there could be upwards of $20 trillion of assets benchmarked to the S&P. With half of that passive indexation, there are clearly countless trillions of dollars of assets managed that may not identify the S&P as a specific benchmark, but even still find it very difficult to stray too far from it. For a money manager to be underweight, Nvidia poses existential career risk. The stock rose 239% in 2023, 178% in 2024, and is up a meager 37% this year. It’s added $4 trillion in market cap over this short time horizon. These numbers are really difficult to get your arms around. To re quote from the Fortune magazine piece how much of the AI boom is just Nvidia’s cash being recycled?

So what to do? Stress test your portfolio against a joint shock higher in both realized volume and realized correlation. If S&P realized volume quickly goes to 20 and correlation to 40, how protected are you? If you haven’t guessed by now, I’m a fan of a put spread overlay to protect against drawdowns. Yes, it imposes a performance drag when markets go straight up, but I consider it Sleep at night insurance well worth paying for. You may not find room to pay for explicit market insurance, but you also have to have assets that have different distributions than US Equities. Gold fits that bill. I’ve discussed the idea that gold exhibits low correlation to the S&P overall, but most importantly, the correlation remains low even when the S&P tanks. That is really important. And gold, I would argue, is an asset that defends against US Government chaos, something we have more than enough of with no obvious end in sight. Let me conclude with a few closing thoughts on the first rule of risk management. Don’t die. Effective risk management requires a healthy imagination and a willingness to carefully evaluate blind spots.

The devastation of the GFC resulted in part from a collective breakdown of our ability to imagine the consequences of housing prices falling by just a little bit, even as it was obvious that underwriting standards had eroded markedly and the real estate market had surged at a rate without precedent. Why did this happen? While of course, no one wants a good party to end prematurely, there is another critical factor at work. When something has failed to occur for a very long time, we lose our capacity to appreciate its return. Instead, we extrapolate the present, what we see and experience now, well into the future. In 2006, that was housing price appreciation. In the aftermath of large scale drawdowns and spikes in measures like the VIX, a consistent realization by investors is that the degree of software sameness in assets was underestimated. It took us until 2008 to recognize that the substantial run up in housing prices was linked to a common underlying driver, the vast supply of mortgage credit. There was a hugely underappreciated source of correlation that failed to make it into how securities and risk scenarios were modeled today. Amidst these record low levels of correlation in the S&P, are we similarly missing a hidden yet shared connection that exists in the ecosystem of companies all engaged in the pursuit of AI riches?

Is the stunning wealth already generated being recycled in the same way that mortgage credit was recycled in 2006 back to Nvidia and the $4.2 trillion of market cap it has added in less than three years? Minsky once said that, quote, success breath breeds the disregard for the possibility of failure. In the context of risk taking and the amazing run in US Tech, it feels like wisdom worth keeping close at hand. I thank you for listening and wish you a wonderful week.

You’ve been listening to the Alpha Exchange. If you’ve enjoyed the show, please do tell a friend. And before we leave, I wanted to invite you to drop us some feedback as we aim to to utilize these conversations to contribute to the investment community’s understanding of risk. Your input is valuable and provides direction on where we should focus. Please email us at [email protected] thanks again and catch you next time.