Is There a Plumbing Problem in Equity Correlation?

[00:00:01] 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. January is behind us and we can now assertively say that with history as a guide, there's a 77% probability that the market will be up this year.

[00:00:29] The January barometer, as some call it, tells us that a positive start to the year in the SPX very likely leads to gains for the entire year. I googled what percentage of years is the stock market up and I was told 69%. With this supplemental stat in hand, I'm not so sure there's conditional information in said January gauge, but a solid start to 2025 for the S&P nonetheless.

[00:00:55] Up 2.7% in January on a realized vol of 14, the index performance hid some of the robust action beneath the surface. A new record for the largest single one-day loss of market cap was set by NVIDIA, losing $700 billion on a 17% down move in response to the news around DeepSeq.

[00:01:16] As I said before, a company with north of $3 trillion of market cap running at a 50 vol is easy come, easy go with respect to dollars at risk. Taking this a step further is NVDL. That two-times levered ETF built atop NVIDIA. This beastly product sported a market cap north of $6 billion heading into DeepSeq weekend. And on that same Monday, it had no choice but to shed 34%, twice that of its master.

[00:01:46] Mission accomplished. This podcast reflects on and asks you to think about risk. Risk is a lot of things, but in markets, I think risk is ultimately about being wrongly sized. And that, in turn, is about underestimating the volatility that you may experience, especially on down moves. The truly mark-to-market insensitive investor is indeed a rare find.

[00:02:10] The rest of us care about the prices flickering in front of us, and when they flicker lower, we may be called to action. So here's the TLDL. Too long, didn't listen. Your podcast host believes markets are riskier than the conventional stats suggest they are. It's been a great run in the U.S. equity market, but at this point, diversifying assets ought to be pursued. I continue to highlight three of them, gold, Bitcoin, and broad market put spreads.

[00:02:39] As we get going, let's do a brief review of January 2025. It brought three full days when the stock market was closed. Adding to the New Year's Day holiday and MLK Day was the Carter funeral observed on January 9th. You've got to go back to 2007, when the NYSE closed on January 2nd to honor the passing of Gerald Ford to find three New York Stock Exchange closures in a month.

[00:03:06] By the way, while Polymarket covers quite a wide range of bets, some quite creative, I must say I don't think Obama and Trump will be yucking it up together at Jimmy Carter's funeral could have been thought of. January brought fireworks to the market and sadly fires to L.A. One of the quotes I've always found instructive in the realm of market risk is from podcast guest and former LTCM partner Victor Higani, who said that, quote,

[00:03:35] financial market insurance is not like hurricane insurance. Exposures within markets are interconnected and risk can be amplified when positions are crowded. One investor's need to unwind a trade quickly could become a headache for you if your exposures share some linkage. Hurricane and catastrophe insurance, more broadly, is different. As bad as those fires were, they couldn't ignite some natural disaster somewhere else around the world.

[00:04:03] Mother Nature keeps no ledger of the insurance written to protect against her. That said, it's clear that insurance against weather and other catastrophic events has been and will continue to be repriced, perhaps sharply higher. As L.A. firefighters had a water shortage, I got more than my personal fill as I had the pleasure of having an old radiator give out, flooding a carpet at home. As Johnny Utah said to Bodie in Point Break, I thought you'd go quietly.

[00:04:33] My radiator did not get that message. The renter doesn't pay for the plumbing problem is one of my little sayings in espousing for the value at the right price of optionality in markets. The long option holder risks only the premium and can walk away if and when things turn ugly. As I was trying to get this radiator shut down, handy is not a word that comes to mind in my self-assessment, by the way. I couldn't help but think of how apt this saying was for me in the moment.

[00:05:02] The renter doesn't pay for the plumbing problem. As I stare at the trusted VCA page on the Bloomberg terminal, I see one-month implied correlation on the S&P at 11. I wonder if there's a plumbing problem afoot in equity correlation. To start, let's acknowledge that this is another period when the corrals ain't correlating. One-month realized correlation is hovering around zero again.

[00:05:28] This is the result of unique bursts of large daily moves in single stocks that are not common to other stocks on that same day. I posted a chart on Twitter to bring this statistic to life, showing just how much of the volatility in the S&P is a result of the stock movement and how little a result of the stock co-movement or correlation.

[00:05:50] An index that averages the one-month realized volatility of NVIDIA, Apple, Amazon, Google, Meta, Microsoft, Broadcom, and Tesla registered 47 for January. As I mentioned, the S&P realized just 14. That's a huge spread. What's happening? Take two stock pairs as an example of the low correlation. Apple to NVIDIA and Meta to Microsoft.

[00:06:15] Microsoft fell 6% on January 30th and Meta was up 1.6% on that same day. How about NVIDIA's previously mentioned 17% dump on January 27th? Apple was up 3% that day. This is low correlation on steroids. As Joe Biden would say, come on, man.

[00:06:36] If you've paid attention to this podcast, you will note my strong view that recency bias is a real thing in markets, yielding the tendency for us to over-extrapolate what we experience in the here and now. As we observe and enjoy these unbelievably low correlations, we forget about the very real idea that stocks can often become highly correlated.

[00:06:59] We are left vulnerable to the expectation that the U.S. equity market's ability to diversify itself via ultra-low correlations will continue far into the future. While diversification is a risk management discipline as old as time, it's mostly a passive exercise. Don't be overly concentrated in any asset class, industry, or single security, says the Investments 101 textbook.

[00:07:25] But in today's world of ultra-high finance, diversification is a risk exposure deemed tradable via the derivatives market. The price of options on stocks like NVIDIA and Apple and Microsoft relative to those on the S&P 500 are how the market prices correlation and the resulting degree of diversification. Multi-strat hedge funds and pod shops are actively taking risk here in many forms.

[00:07:51] In the options market through variance and correlation swaps, through new age QIS exposures. To be sure, the backtests of said trades look pretty, pretty good, to borrow from Larry David. Let's do a quick review of some option pricing outcomes during January 2025. At the end of 2024, one month implied vol on NVIDIA was 45. One month realized vol for the month of Jan 25 was exactly twice that, 90.

[00:08:20] 45 vols of realized excess risk premium. For both Apple and Microsoft, this was nowhere near that size, but realized was 8 vols over implied. For the S&P, we ended 2024 at 13.5 implied vol and realized was 14.3 for the month. If you've had the, quote, dispersion trade on, that is, long single stock volatility and short index volatility,

[00:08:48] these huge divergences have made it work. But as De Niro said to Pacino in that famous diner scene from Heat, there's a flip side to that coin. Carry trades like dispersion sometimes sow the seeds of their own demise. As such, a concern should be that winning trades keep attracting capital, even at levels that offer increasingly smaller margins of safety.

[00:09:14] As low as implied correlation is, it clears the market because realized correlation is just about zero. The old adage, past performance is no indication of future results, does not apply to the backtesting community. For them, past performance isn't just a thing, it's the only thing. Let's run through some numbers. 2022 was the last year of correlation of any consequence.

[00:09:39] This was a global central bank catch-up trade to the inflation that did not prove transitory. Realized correlation for that 12-month stretch was right around 50%. Unsurprisingly, just as 2023 was set to begin, the market-priced one-year implied correlation on the S&P at 54%. Not a ton of risk premium to that which had just been realized,

[00:10:02] but at least a number that was relatively typical for the S&P in the 63rd percentile over the last decade. 2023 would prove the start of breathtaking declines in realized and, as a result, implied correlation levels. For the full year, realized was 22%. In response, implied correlation, which started the year at 54%, would fall dramatically to 34% by the end of 2023.

[00:10:31] And in a stunning second act of anti-connectedness, realized correlation on the S&P would register just 12% in 2024. This is certainly oversimplified, but if you sold correlation in 2023 at 54%, you made 20 points. If you sold it in 2024 at 34%, you made 22 points. These spreads are not insubstantial. Trades that work well invite capital.

[00:10:59] Post-2023 and 2024, the backtesters are surely working overtime to produce glossy charts illustrating the bountiful carry to be had in harvesting correlation risk premium. They may have some bells and whistles to add to trade construction, especially in the realm of bank-generated QIS, where there are versions of this strategy that number in the hundreds.

[00:11:23] Their success can be self-fulfilling, as fresh capital directed at selling correlation means that more index vol is sold and single stock vol bought. But at the end of the day, successful trades generally accrue sponsorship. Will there be a plumbing problem in correlation? Who knows? But the market is certainly priced as if a flood of co-movement simply cannot happen. I like buying things near zero implied probability.

[00:11:50] Ultimately, low correlation is the market assigning less value to an index option relative to the stocks that comprise that index. The market expects the zigging and zagging of stocks to continue. And even as it will pay up for single stock vol ahead of earnings, it translates little of that into boosting vol at the index level. To be sure, this view has proven correct. I'm just not convinced it will continue. And this brings us back to the unique properties of market-based insurance.

[00:12:20] Option prices respond not just to new information. They also respond to the implementation of and especially the unwinding of risk. Was the VIX spike of August 5th of last year an event that led to NVIDIA and Apple suddenly being identified as considerably more correlated to one another? No, not really. Rather, a sharp unwind in VIX futures led to mark-to-market losses for a subset of investors,

[00:12:48] including those running dispersion books that created further demand for S&P vol as those types sought to limit the damage. The VIX surge on August 5th was far greater than that of the vols on the mega cap stocks that comprise the S&P. That's what a correlation on wine can look like. Risk-taking in cat bonds doesn't create a natural disaster. Risk-taking in markets can be the source of a disaster.

[00:13:13] For me, the often circular interaction between market prices and the products built within the market has always been worth thinking about. This, quote, endogeneity of risk has reared its head many times. 1987, 98, 2008, 2018, 2020 are a few examples. Market prices not only respond to developments on the economic, monetary, and geopolitical front,

[00:13:39] but are also impacted by the presence, especially of price-agnostic trades, that result from products and strategies that respond mechanically to the movements of the market. Let's explore this further and highlight everyone's favorite product to both love and hate, daily resetting leveraged ETFs. I just Googled, quote, this is not your father's ETF market, and nothing came up, so I'm taking it.

[00:14:05] I think it's a pretty apt description given the action in products like MSTU, the two-times levered product built atop MicroStrategy. Oh, sorry, I mean strategy, as Michael Saylor has just renamed his Bitcoin buying engine. More recently, NVDL has been a fascinating one to watch. This is the Granite Shares 2X long product written on NVDA. What a beast this thing is.

[00:14:31] It closed Friday, January 24th, with assets under management of $6.6 billion. On Monday the 27th, following the DeepSeek news, it traded 92 million shares, five times the normal volume, as it fell 34%. Note that for NVIDIA, Monday's volume was only a bit more than twice the daily average volume. While NVDL is massive because there's so much volume already in NVIDIA,

[00:14:59] the extra volume that NVDL imposes on its underlying stock through the end-of-day rebalance isn't terribly large, maybe 6% or so additional volume from that big down day. But let's step back and appreciate that there is a, quote, company with a now $5.5 billion market cap with a single holding in leveraged fashion running on a 170 vol that is being traded at a furious pace.

[00:15:30] It has already suffered a fair amount of, quote, vol drag this year, as NVIDIA is down 3.3% and NVDL is down 14%. Lastly, and you know we had to get here, there are options on NVDL. Markets change over time, especially as financial innovation is never ending. Understanding the role that products can play is important. Back in December, I did a series of podcasts on the leveraged ETF complex

[00:15:59] that I believe remain very much worth a listen. Last point on financial products providing feedback to the market. It would be inappropriate to allow February 5th to come and go and not mention the seven-year anniversary of XIV Day. On that Monday in 2018, the XIV and SVXY were effectively wiped out, down 95% in a single day.

[00:16:24] The combined value of the ETP complex went from $3 billion to around $150 million. On the Friday, February 2nd of 2018 close, XIV and SVXY were both short approximately 200,000 VIX futures or $200 million of Vega. 95% of that short volatility risk was covered on the following Monday, likely in the minutes leading up to the 415 futures close.

[00:16:51] Just get it done, the XIV was told on the afternoon of February 5th, 2018, with respect to buying VIX futures. This drove the massive move higher in VIX futures from $4 to $4.15 and was a significant contributor to the S&P futures sell-off in this time period. An accident hiding in plain sight for months and months during the extremely low vol period that preceded it. This is a lesson on how financial products can impose themselves on markets.

[00:17:21] It's also a strong reminder of the limitations of market liquidity, especially when trading must happen at a certain time no matter the price. Note, the Sharpe ratio of the XIV was well north of 3 in 2017. Repeating one of my old sayings, risk on and risk off are curious cousins. If you've made it this far, you are a loyal listener. At the risk of repeating myself, which I do quite a bit, just ask my kids,

[00:17:50] I believe now is a critical time to think about unwelcome scenarios and ways to insulate your portfolio from them. As I surveyed the damage to my carpet from the radiator that ran amok, I asked my plumber if I should replace the other three. In a thick Irish accent, he told me, No, Dean, you don't want to do that. You'd be replacing so many things in the house. You just have to let them fail and then deal with it. Interesting advice and certainly not a sales pitch.

[00:18:19] There's absolutely a thing called being too cautious and or over-insured. The quote, always on hedge in markets would fit this category. The seller of that insurance is in the business of making money and a risk premium comes with the territory. My recommendation then is a small ball implementation of insurance. You need not take out the no deductible gold plated policy that costs you an arm and a leg.

[00:18:46] Instead, take a look at two and three month put spreads on the S&P and triple Q. I like striking them 5% out of the money on the top side and then down 20% on the bottom strike, creating a 95-80 put spread that is priced very reasonably. Alongside this, I think finding diversifying assets is an absolute must right now. The stock market has found a way to magically diversify itself via these absurdly low levels of realized correlation.

[00:19:14] Will this continue? Perhaps. But there is potential for a plumbing problem in correlation. We have to be careful not to be seduced by the tremendous diversification already delivered and bank this as a permanent feature of risk. Rather, find assets that are just a bit different. For me, those are gold and Bitcoin. In my recent podcast, Digital Gold and Actual Gold, I ran through the reasoning.

[00:19:41] While we must still classify Bitcoin as a risk asset, gold especially has proven highly durable to S&P risk-off events. Both Bitcoin and gold share a characteristic that sees them enjoy jump risk to the upside. Stock-up, vol-up assets, as I call them, are hard to come by. Maybe this is how the notion of scarcity is expressed via price moves. In contrast, as the U.S. runs tremendous deficits,

[00:20:08] gold and Bitcoin are conceivably becoming expressions of investor concern that the exorbitant privilege enjoyed by the United States has a downside to it. While you are looking for broad equity market put spreads, also check out call option structures on both the GLD and IBIT. With that, I will bid you a farewell for now. Keep the feedback coming. I love hearing from you. I'm looking forward to engaging with all of you on the wonderful world of market risk in 2025.

[00:20:38] Until next time, be well. 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 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 feedback at alphaexchangepodcast.com. Thanks again and catch you next time.