John Marshall, Head of Derivatives Research, Goldman Sachs

[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. It was a pleasure to welcome John Marshall, Head of Derivatives Research at Goldman Sachs, to the Alpha Exchange.

[00:00:25] Our conversation explores a number of critical topics, starting with the meaningful growth of equity funds deploying options as part of a risk management overlay. John describes how covered call ETFs and systematic vol-selling funds have quietly reshaped the supply-demand dynamics for index optionality.

[00:00:45] He makes the point that this cohort, unlevered, yield-focused, and largely buy-and-hold, is proving more resilient than the vol-selling programs of past cycles, with implications for both market stability and the vol-risk premium. Next, John shares his team's efforts to find what he calls asymmetry alpha in the options market, focused on event-driven, catalyst-based trades at the single-stock level.

[00:01:11] We learn that option pricing is increasingly being informed by company-specific fundamentals. John explains how his team connects metrics like free cash flow yield, return on equity, and event-driven catalysts to the pricing of volatility and skew. Rather than relying solely on historical vol or peer group comparisons, this approach seeks out asymmetries in option markets that are grounded in the evolving health and risk of individual balance sheets.

[00:01:41] John argues that these additional company-specific variables are often overlooked by traditional volatility frameworks and, as a result, can help identify mispricing in the tails, informing more precise use of calls, puts, and risk reversals. I hope you enjoy this episode of the Alpha Exchange, my conversation with John Marshall. My guest today on the Alpha Exchange is John Marshall.

[00:02:07] He's the head of derivatives research at Goldman Sachs, a firm he's been at for 23 years. John, it's great to welcome you to the podcast today. Thank you very much for having me. Looking forward to this discussion on some of the interesting work you guys are doing on the research front and derivatives, especially with regard to event-driven trades and kind of how the market prices events. You guys have done some really interesting stuff there.

[00:02:35] One of the ways I'd love to get the podcast underway first is to learn a little bit about your background. You started at Goldman in a tech research capacity right around the advent of the or maybe the unwind of the tech bubble. I always say that as part of markets, you're kind of colored by your experiences with market moves, risk events, and sometimes when you start matters.

[00:03:05] For me, I got kind of underway a little bit before you, but no sooner did I land in equity derivatives at Lehman Brothers that Asian contagion hit in 97 and then LTCM blew up in 98 and then the tech bubble. So I got a lot of risk events under my belt in a short period of time. Tell us a little bit about your career trajectory as you got your start and then as you transitioned into this derivatives research role. Yeah, absolutely.

[00:03:33] So I started within research and for the 23 years I have been in research. I started on the tech team covering electronics manufacturing services or at least on that team. And I also found that a lot of options were traded on these names. And so after three years of building models, looking at fundamental research and understanding it from the fundamental side, I moved over into derivatives three years later.

[00:04:01] And about one third of all options were traded in tech at the time. So that gave me quite an advantage. And studying the electronics manufacturing services area gave me a lay of the land, how all the tech companies were connected and how when one company would report, it would affect the earnings or the outlook for another company. And that was a really formative time for me and really understanding markets.

[00:04:29] Now, as you mentioned, I started in March of 2002. And of course, in October of 2002 was when the market bottomed. So that was a rough first six months. But it really showed me how volatile markets can be and gave me a quick dose of understanding how little I knew and how much there was to learn.

[00:04:52] But also how there are unexpected things that come out in the market, things that even the most experienced strategists don't see coming. And you have to reserve something in your view for those unknown unknowns throughout that happen over the course of time and not be too confident that you do know everything about a particular name or a market.

[00:05:17] There's an old quote from Voltaire that I often use, which is doubt is an unpleasant condition, but certainty is absurd. And, you know, we really have to just respect the degree to which there's so much we don't know. As you sort of described your early days, you said a couple of things there that I'd love to maybe just do a juxtaposition on, which is this idea of the interconnectivity, especially of the tech sector.

[00:05:45] You know, the low correlation of 2023, 2024, you know, certainly into, let's call it March of this year before the tariff tantrum occurred, has been really an interesting dynamic. And at least in tech, it's so different than, you know, 99, 2000, 2001, where Intel's earnings meant something for Microsoft.

[00:06:13] There was just a tremendous degree of correlation across those stocks. The tech as an asset class was a thing. And here you have these giant idiosyncratic moves in NVIDIA, in Apple. And so they lead to much different correlation, much lower correlations. I'd love for you just to reflect on then versus now and kind of what the differences are. Yeah, I think there's a couple of items that are interesting here.

[00:06:41] The first is the advent of Reg FD. When I started in 2002, that was before the really big regulatory changes in 2003. And over time, what I have seen is companies increasingly being diligent about only giving information at major public events or releasing that information to the public all at the same time.

[00:07:07] And what we saw over time, you know, that 20-year span was that increasing information content on events and less information content between events concentrated all that volatility more on event days. We saw this happen in the U.S. It happened a little bit later, but also occurred in Europe and then in Asia last.

[00:07:34] But as companies started to disclose more information on events and less information between. Now, fast forward to the 2023 time period, we had not only that focus of information being released on events, but a lot of themes that were happening that had winners and losers, you know, people who were out front of, you know, the development of AI and companies that were left behind in the development of AI.

[00:08:04] And I think that each time one had a success, investors would translate that to a loss for other companies. And that drove some of the dispersion, whether that was among sectors within the S&P, whether it be industrials or tech or healthcare. But even within the NASDAQ and within the technology companies, there were winners and losers.

[00:08:27] Are these companies not clients of each other in the same way that they were back in 2000 and 2001, you know, where the earnings, let's say, either for better or for worse of a certain company were just viewed as having implications for another company that was, you know, related perhaps in the supply chain? And, you know, some of those profitability dynamics?

[00:08:54] Is it just a different manner in which the ecosystem of these tech companies either work with each other or compete against each other now? Yeah, it's a great question. I think there are different waves of innovation that can lead to either more dispersed environments or more correlated environments.

[00:09:16] For instance, if it's a hardware discussion, as a lot of it was in the early 2000s, you know, building out fiber, building out infrastructure, their companies tend to be very highly correlated. And the electronics manufacturing services are clearly highly correlated with the original equipment manufacturers.

[00:09:38] Now, when you get into a technology transition, like we're seeing with AI, there's much more of a potential for one company to be a leader in an area and has consolidated. It could also relate to the increased consolidation of the largest tech companies. People often talk about the Magnificent Seven. And a lot of those functions are happening within a company rather than a dispersed larger number of companies.

[00:10:08] So as you look back on your time at Goldman, a lot of what you do and your team does is going to be dictated by what clients are asking. And so their challenges are going to evolve over time. You know, I can imagine there were times when it was all about macro hedging.

[00:10:27] Or there were other times when maybe rates were super low and we're trying to find some ways to, you know, the bondification of the equity market, trying to create yield. Reflect on just some of that evolution. You know, that can kind of get us into the here and now of the work that you guys are doing.

[00:10:47] But I'd love for you just to comment on, you know, a period of time when the demands from the client base were very unique relative to what you're seeing in the current day. Yeah, absolutely. So it tends to follow market cycles fairly well. For instance, when the economy is growing at a steady pace or even at a brisk pace, you tend to get more focus on what are single companies doing.

[00:11:17] You have time to explore the individual successes of companies and can make those individual investments, whether that be ahead of earnings events or ahead of analyst days. You're looking for who are going to be the winners and investing there.

[00:11:34] And so we spend a lot of time on single name catalysts during that type of environment and a lot on understanding where the balance sheets of companies can take them, whether they can buy back stock, whether they can pay dividends, whether they are a candidate to be acquired. And that tends to dominate when we're in an environment like a 2005 to 2007 time period, the re-leveraging part of the market.

[00:12:04] That has been consistent with the last couple of years where we had the AI trends and relatively stable economic environments. This contrasts greatly with the middle section, which would be, I'd say, somewhere from 2013 to 2018, where we saw a huge focus on extracting yield from the options market.

[00:12:31] This is when you saw large put selling funds be launched, ETFs that are ETNs that sold options on the VIX or sold VIX futures. Those were all about harvesting yield at a time when bonds did not give investors much yield.

[00:12:51] And those were being done in a levered fashion, which of course led up to the VIX spike in early 2018, when a lot of these products unwound all at the same time. But that was a much more macro environment where we needed to focus on the potential for volatility in the market and what that would mean for these larger yield strategies in the options market.

[00:13:19] Currently, how would you just describe the big picture of the interest from clients? Does it have more of a micro tilt to it? We're obviously coming through a short-lived but intense kind of macro risk off. Give us a sense as to some of the more recent interest from clients. Yeah, I would say that the most interest, and this is consistent with the last couple of years and changed a little bit in the last couple of months,

[00:13:48] but the interest has been in single name opportunities, what is driving the extremely high dispersion or low correlation environment, which is quite literally off the charts. It's such a low amount of correlation. And so investors are searching for catalysts. They're searching to understand what is driving the large moves.

[00:14:14] Some of that comes down to positioning and not just about the fundamentals of the companies. Are people getting too over levered or under levered or too short into particular events? And that is a key component of what's driving the stocks as much as the fundamentals sometimes. Now, I would say that there is focus on index options selling and the launch of products that consistently sell index options,

[00:14:44] whether that be on the S&P or the NASDAQ. That's something that when we talk with macro investors is the biggest focus. But I would say that at least half of our time is being spent on those micro catalysts and positioning there as well. Well, before we will spend the bulk of our time, I think, talking about some of the interesting work you guys are doing on the single stock catalyst side of things.

[00:15:10] But one of the charts that you presented as part of your research is on the growth of these, you know, call them risk managed or income oriented ETFs. I thought it was a really nice way of portraying it. So maybe you can show that chart and then walk through, you know, just describing what you see as the trends.

[00:15:34] I'm always interested on the supply demand dynamics in optionality, right? The price is just a function of where people met. And sometimes there's a shortage of supply and sometimes there's a excess of supply and it clearly matters, right? And so with this new advent of both ETFs and mutual funds, I'd love to explore that. Just like the market clearing price of, you know, we'll start with the S&P.

[00:16:02] That's probably where most of these funds are targeting their exposures. Do you want to show that slide? Yeah, absolutely. So over the last 15 years or so, part of our role here is to help clients think about what products to launch. And that's typically driven by two factors.

[00:16:26] One is where is there excess demand or supply in the financial markets for a particular type of trade? The classic one is investors typically want to buy puts to protect their portfolio and puts tend to get overpriced. So some investors will then come in and provide that liquidity and over time can outperform by providing liquidity that the market needs.

[00:16:51] The other factor in deciding what strategy to launch is what type of exposure to investors need. We were talking about income before. Many investors would like to be involved in equities but gain income. And one of the most popular ways to do that is to sell calls. That wouldn't necessarily be a rush to extract alpha from the market,

[00:17:14] but could be to change the shape or the distribution of returns to be something more like what the client wants. What we're showing in this chart here is we've aggregated over 300 products that have been listed in the market and cataloged the publicly available information on their strategy. So we can assess when we aggregate all of this into the different types of trades that ETFs and mutual funds do,

[00:17:44] and we can see how that adds up. And what you can see immediately from this picture is there's a variety of strategies, whether it be calls or puts, and some of the strategies do more than one of these. But the largest number on this screen is the negative $83 billion in AUM that sells one month 0% to 5% out of the money calls. Now, part of this is on the S&P, part of this is on the NASDAQ,

[00:18:14] but this is the bulk of the AUM or the largest portion of the $200 or so billion in AUM is selling calls as part of their strategy. What this does is it takes advantage of all of the investors out there that are buying calls in order to gain upside exposure. This offers liquidity to those investors systematically over time.

[00:18:39] And it creates a product that is exposed to the equity market, but has a capped return in exchange for income. And that is something that it turns out many investors want. They would like an income generating equity as they go into retirement, perhaps. And that has grown pretty significantly. We have not, while this chart only goes through March,

[00:19:05] we have not seen a significant pullback in the assets under management of any of these funds. And that shows the resilience of those types of funds, even in volatile macro environments. So I would love to get your take on this, maybe just stepping back and asking you around the vol risk premium. So we talk about selling upside calls.

[00:19:29] You're the provider of insurance historically, whether you're Geico or Goldman Sachs, ought to get paid for bearing jump risk, right? Bearing uncertainty risk. And so we see consistently in not just the equity market, but basically every financial market, there is a premium of implied to subsequently realized vol. It goes up, it goes down. We know about the negative skewness of the return profile of the S&P.

[00:19:57] So when you lose, you lose fast and by a large amount. But over time, there is a measurable vol risk premium. I would love just to hear how you think about it. Are you on the side that this is a valuable thing for investors to harness and the time series of it? We're talking about these new funds with new trading tools at their disposal. Zero DTE.

[00:20:25] As an example, this stuff wasn't around when you and I first started. So the evolution of the vol risk premium in light of new instruments delivered into the market. So it's an open-ended question, but I just would love to hear you think about it. If we were back in 2006, 2007 and looking at the history that we had at the time, 20 years of history of volatility trading, it was clear that there appeared to be a volatility risk premium.

[00:20:54] And there were fewer investors that were set up to sell options who could understand the risks and could create a strategy and harvest that volatility risk premium. In the mid-2010s, there were many strategies launched that could harvest that volatility risk premium.

[00:21:17] And I think that the launch of many of those strategies, including the VIX selling strategies, helped to compress that volatility risk premium. It works if I were to take a quick step back. When an investor sells an option, they sell it to a market maker. That market maker often will warehouse that option and will trade the underlying security to replicate the other side of that.

[00:21:44] If an investor sells an option to a market maker, the act of delta hedging of that option actually compresses realized vol. And so in 2017, we got realized volatility of around, it was single digits on the S&P, really low.

[00:22:03] And my belief is that was related to all of this index option selling through VIX and through other products that really not only compressed option prices, but compressed realized volatility. So if one was looking at implied versus realized, you would have still seen a spread that looked like it was harvestable, but that the act was actually depressing, was keeping that spread alive because of the delta hedging.

[00:22:32] So fast forward to now, what I think we're seeing that's different this time from the 2015 to 2017 is the volatility selling that's happening now is in unlevered form. The funds that I was showing and the positioning that I was showing before, the average delta or the average market exposure of those funds is 0.6.

[00:22:58] So they behave as if it's 0.6 times the daily move of the market. And so these are unlevered strategies done by investors that are typically buy and hold, as opposed to prior cycles where 2007, 2008, or in the 2018 timeframe, where those were levered sellers of volatility. And I think what we've seen through this recent environment is that that has more staying power,

[00:23:28] that these investors are in it for the type of return distribution that they signed up for. They're not actually trying to greatly extract the volatility risk premium. Now, I think that has a couple of other implications, one of which is that I wouldn't expect it to reverse so quickly as it did in 2007, 2008, or in 2017, 2018.

[00:23:55] And so far in the last couple of months, we've seen action that is consistent with that, that they're not reversing quickly. But I do think that the more people that are seeing this opportunity to sell options to generate income is eroding that volatility risk premium. And I no longer spend a huge amount of time trying to develop strategies which are specifically about harvesting the volatility risk premium,

[00:24:24] because I don't know how much longer that's going to be available. And I go into other strategies, really thinking about asymmetries, which combines the volatility and the direction move. And that's where I think that the bigger opportunity exists. We're going to pivot to the interesting work you guys are doing on that more catalyst-oriented, single-stock-oriented segment of the market.

[00:24:52] Just on the big picture of the pricing of, let's just call it index options, S&P index options, we've got this thing called the skew as well. And I think even in your work you referenced it, there's implied skewness and realized skewness. And just as there's a VRP, there's almost a skew risk premium too, right? So of course there ought to be a skew. The Black-Scholes world is not a real world.

[00:25:19] There are big downside asymmetries that exceed those to the upside, but the pricing of the skew over-accommodates for that, right? That the put costs more in vol terms than it, I don't want to say ought to, but at least that empirically it is realized. How do you think through the skew part of things?

[00:25:43] Is that an institutional demand for downside versus the upside supply through overwriters? How do you kind of explain the excess and skew? Yeah. Some of the work that we've done, depending on how you look at it, the realized distribution has as much skewness as the implied distribution on some of our analyses,

[00:26:08] which blew my mind when I first did that work back in 2012, 2013. And we also divided into another fun word, kurtosis, the fat tails, and also find that the realized distribution has significantly fatter tails, just as the implied distribution prices in.

[00:26:32] And so I'm less convinced than other market participants that there's a significant skewness to be harvested or a significant, and perhaps the kurtosis is even underpriced. So let's shift to the single stock realm. One of the first things that you've done, maybe just start from 30,000 feet, is you're connecting economic data,

[00:27:00] but also corporate profit-specific data, company-specific variables to vol surfaces or how option price is clear. I think that's super interesting. And, of course, there's the Merton model, which is tying together the creditworthiness to metrics like implied vol. The stock is an option on the firm. The debt holder is short the put.

[00:27:28] But you've got it to sort of some really specific variables. So just maybe, again, start from the big picture of how you thought about approaching this, connecting these two things, corporate variables and vol dynamics, and then we can kind of get into some of the specifics of what you guys have found. Yeah, absolutely.

[00:27:51] So the approach that we came at this several years ago was looking at not just the expected return or the expected volatility, but what clients really want to know about is what's the probability of a significant up move and the probability of a significant down move? Because most people trade puts and calls naked.

[00:28:18] You're buying a put or a call or you're selling a put or a call, and they're not actually buying volatility or selling volatility. And so we developed a framework. From the macro perspective, we call it GSEQ move. And I'll show a quick picture here, which may be worth a thousand words.

[00:28:36] Really, what we're doing is, just as you mentioned, using some of these macro variables to estimate the probability of a 5% up move in the S&P over the next one month, or a 5% down move in the S&P over the next one month. We're estimating them with the same data, but estimating separately that upside and downside,

[00:28:59] because there are certain variables that are much more valuable on the downside and certain variables that are much more valuable on the upside. And what we found is that when you compare these to option prices over time, and there are big gaps between the economic environment's perspective on asymmetry and the option market's perspective on asymmetry, that leads to opportunities, opportunities to buy or sell calls or opportunities to buy or sell puts. So there is a lot in this chart.

[00:29:29] So let's maybe take it slowly. So we're going back quite a ways. We've got almost 30 years of data. And certainly the blue, dark blue, is a market price. Yes. I believe it's kind of a binary option, struck out 5%. Exactly. Okay. And then so let's just take the Goldman Sachs GSEQ move.

[00:29:55] And let's just start with the most recent, because there's a big divergence there, right? So your calcs, your variables, leave you with a 35% probability of a 5% up move, much higher than the markets. Give us a sense as to what is going in there to create that big dislocation. Yes.

[00:30:17] And so the four variables that are most prominent in this model are ISM new orders, U.S. capacity utilization, another macro variable, free cash flow yield on S&P 500 companies, and return on equity for S&P 500 companies.

[00:30:37] And what is happening more recently is that return on equity is really high, and that tends to suggest an unusually large probability of upside. What's happening on the downside with our model, which is suggesting a 26% probability of a 5% down move, it's also suggesting an elevated level of probability of down move.

[00:31:02] This is driven by unusually low free cash flow yield of companies. So when companies have low free cash flow yield, that tends to mean they're not generating as much cash, that cash is not piling up on their balance sheet and providing the security that it would to that company if it were a high level of free cash flow yield.

[00:31:25] And so both of these fundamental variables are going in the volatile direction at the same time, and that is suggesting that options in general are inexpensive. And I can't help but look at always the 2020 experience, and so there's two giant divergences here, and these are the same variables. Yeah.

[00:31:51] So just walk through, especially on the downside, it's just fascinating. You've got almost a 0% for your models, and the market's at an enormous, obviously vol got to 80% for a short period of time there. So not surprised on that front. But your model not showing anything is super interesting. Yeah, and it comes from the variables interacting with the coefficients in the model. This is a logistic regression model.

[00:32:21] And this was giving us a great signal telling us that, well, the cash flow of companies was doing just fine because of the actions of the government. While elements like the ISM was not doing well, the rest of the variables were relatively stable. And so that just mathematically gave us that outcome.

[00:32:46] One of the things that we have to watch out for in a model like this is when you have an environment that has never been experienced before, you have to be cautious in applying these particular coefficients and using the model framework. We had other ways that we were looking at the market as well at that time. So we didn't rely on this one metric. But clearly, if we had, we had been better off because we did get that significant rally.

[00:33:15] And it turned out call options were grossly underpriced at that point. And what you just said there, I'd love to just follow up on, which is knowing that, you know, this time sometimes is different, right? That market conditions change 2020 brought on a policy response, the force of which we've never seen before. With, you know, with, you know, lasting implications in some ways. But, you know, that was a part of the risk dynamic.

[00:33:45] I had Cliff Asner speak at this charity event that I host, Macro Minds. And he was talking about, you know, circa 2020 and 2021 being someone focused on the value factor. And, you know, you've got stocks like Zoom and Peloton are at all-time highs.

[00:34:03] And you've got lots and lots of data historically telling you to stay with your model, but fighting the urge to recognize that, boy, things have really, really changed. And so as you deal with data and know that market structure changes and relationships between companies and maybe the broader economy change, how does that kind of figure into utilizing data, being careful? I think you maybe implied it there.

[00:34:32] The overfit is something you've got to be mindful of. Walk us through some of that tension as a researcher. Yeah, it's a great question. And something in particular in the last two years with the focus on AI and machine learning, those have become more important tools in our toolkit.

[00:34:52] And what I find with some of those frameworks is they're even more focused than other regression frameworks on not relying on outliers. Whether you're using something like a random forest framework or others. And we do this much more in our single stock work where you have hundreds, even thousands of underliers that all have rich sources of data and you can draw on them.

[00:35:19] I think it's particularly important to be chopping up data into smaller time increments, into, you know, we're constantly dividing up the universe by the alphabet or by sector and understanding which relationships seem truly universal and which relationships have faded or grown over time.

[00:35:48] And that is absolutely critical for understanding how to use the tools in actual investing. And something that we found it fruitful to be moving on beyond some of the typical regression frameworks. Well, one of the things about markets is that, you know, no free lunch is probably the most intellectually honest starting point.

[00:36:15] And, you know, quote, free lunches or the perception of them create crowding in a hurry and it undoes it. Right. That's just the competitive forces. But one area of your research is on something you believe is the least crowded area of alpha, which you call asymmetry alpha. So high level. What does that mean? And then we can dive into some of your interesting results.

[00:36:41] Yeah. So the GSEQ move that we just looked at is certainly related to this topic and was the first work that we did on separating that upside probability from downside probability. We also do this on single stocks.

[00:36:57] We have a seven variable model that's also a logistic regression framework that looks at things like free cash flow yield, dividend yield, whether the stock is heavily owned by ETFs, whether it has an earnings event coming up.

[00:37:11] And in that type of a framework, we're able to, on the micro level, distinguish between stocks that have high upside asymmetry, low upside asymmetry, high downside asymmetry or low downside asymmetry separately.

[00:37:29] And, you know, for the, for the options traders that are listening to this, it quickly can go to, if you have separate decision-making on puts and calls, you can do eight different types of strategies, whether that be buy calls, sell calls, buy puts, sell puts, buy straddles, sell straddles, risk reversals or callers. And so the, the modeling that we do uses that same signal that we saw in EQ move.

[00:38:00] Well, it's twisted a little bit to fit into the single stock realm in order to understand which of those eight strategies is optimal for a given stock at a given time. Some of the work that we're doing now that has basically, as, as you mentioned before, client demands lead us to do the work that we do is integrating positioning into those types of models.

[00:38:27] As you mentioned before, crowding is a big topic that people focus on. So we know that we need to not only know the drivers of that fundamental asymmetry or the environment asymmetry, but also how are people positioned in that name and how does that affect the realization of some of that asymmetry?

[00:38:49] And one of the areas that you guys have focused on is catalyst driven option buying. And you've got a result here that, you know, I wouldn't have expected. We talked about the volarisk premium, which is, look, insurance is great to have, but it can be costly.

[00:39:08] I would think with a known catalyst on the calendar, the options market would kind of see that buyer coming and, and the, you know, you might get the move, but it would be insufficient over time to pay you back for all the premium you expended. Right. And so you've found some interesting stuff about, I think, catalysts in general, maybe earnings is the most prominent one. Walk through that, that area of research for you guys and what you found.

[00:39:39] Yeah, I agree. It's kind of counterintuitive that buying calls ahead of single stock events tends to be profitable or has in history. And I think there are a couple of factors at work here. The first one is stocks tend to go up over time and that tends to benefit call buyers.

[00:40:00] And in particular on events where stocks tend to be volatile, more of that return happens on these events than happens between the events when there's no information being released. And so I think by strategically buying calls ahead of particular events on single stocks, this gives you an advantage in participating when the information is being released.

[00:40:29] We've also found that this is a particularly historically profitable strategy around analyst days. And again, those even have an added benefit of analyst days tend to not be as followed by options traders and the options markets. And therefore, there's a rationale that you could get an underpriced option because it has less people paying attention to that event.

[00:40:57] But the event is still an important release of information. One caveat to this, and for all people who have done option strategies and paying attention to selling out of the money calls as a profitable strategy on the market over time, what we're focused on is at the money option performance. So if you had a $100 stock, it would be a 100 strike call.

[00:41:26] This tends to be a less popular option to own. Most people, when they buy options, they tend to buy out of the money options, which have a low premium and a really high leverage. And so that tends to be a bit more crowded. The at the money options that have high stock exposure or high delta, that tends to be not as popular a trade.

[00:41:47] And that's one of the reasons why I think this artifact exists in the data where the call buying or at the money options ahead of events has been typically profitable. Well, I saw in your work that you had cited that the earnings day vol was roughly 4x the non-earnings day vol. I was very happy to see that because I had ballparked it myself and I said 3 to 5x. Yeah. You're right in the center. So I had good intuition.

[00:42:18] So stocks go up in general. Is there an earnings day premium? Do stocks do better on earnings days is the first question. Is there a realized excess return of just buying on the earnings day? Does that tend to do better? Yes.

[00:42:37] And that tends to be because the companies are growing and companies are making more money each year and they tend to do better than analysts expect. More companies tend to beat earnings estimates from consensus than tend to miss. And I think that that is part of that.

[00:42:59] But one of the questions that we've been asked by CIOs and hedge fund board members is about not just the excess return of stocks around earnings events, but the return even after adjusting for volatility. And that, I think, is an important component because, sure, you may experience a bigger return on the earnings event.

[00:43:25] But if it comes with twice as much or four times as much volatility, maybe that's not worth it. We did an exercise last October of what if you removed companies from your portfolio that were reporting earnings? How would that affect the volatility of your portfolio if you took them out for the week of earnings and then put them back in? And how would that affect the return?

[00:43:49] And it turns out it barely affected the volatility of your portfolio because that's very idiosyncratic returns and doesn't affect your portfolio volatility that much. But it significantly reduced your earnings potential or your portfolio performance by removing those stocks. So I think that tradeoff is really in favor of higher profits from owning those stocks around their earnings events, even despite the excess volatility.

[00:44:18] And so as you talk about the option side of things, as you said, you have an opportunity to buy an at-the-money call. You could gear it up and buy an out-of-the-money call. Are your results implying that the way the market has the out-of-the-money call priced relative to the at-the-money call is not – it's just not jiving empirically with what happens?

[00:44:42] In other words, I think you're saying I'm better off buying one at-the-money call versus three or four equivalent out-of-the-money calls. Is that the conclusion or what you see? Yeah, I think that that is a fair conclusion from our work. It could be related to the greater difficulty in hedging the out-of-the-money options.

[00:45:04] You have more complex freaks, potentially more gap risk if you're in that out-of-the-money option, which has a lower premium. And that could theoretically kind of excuse that. But empirically, how you just discussed it is what we're seeing.

[00:45:24] So one of my stories, many war stories about putting trades on with clients is I had a client come – it's a while ago, but the stock got completely crushed. And as happens when a stock gets cut in half, the vol went to 120. And a client came in to sell just some truckload of downside puts in the stock.

[00:45:50] And, of course, the realized is through the roof and the implied is 90 or something like that. And I just realized all at once that this was not a fair fight, that your trader is looking at the HVG screen on the terminal. And you can't help but price anything off – you're pricing it off of realized. You don't know much about the stock.

[00:46:10] And I concluded that in some ways the best vol trader is someone who's really in the weeds on the fundamentals but also has an appreciation for option pricing and so forth. And that's my way of just wanting – asking you to explore the work that you're doing connecting these single stock fundamentals to vol. I think that's just a super interesting area that you guys are working on.

[00:46:37] So we talked about the more macro variables, PMI, capacity utilization and so forth, return on equity. Walk us through the single stock tie-in to vol. I think you said there's seven variables in one of these models. Let's talk about that. Yeah.

[00:46:55] Given my experience over the last 20 years in derivatives research, one of the things that I have pretty good confidence in is that market makers in the options market, particularly the single stock options market, they are doing everything they quantitatively can to keep all of these option prices in line.

[00:47:18] In line with realized, in line with peers, in line with everything that you could program a computer to do. So therefore, what I need to do as a strategist is understand what could they be missing that is systematically really interesting and could have systematic implications.

[00:47:40] The first one that I think is the most important, and we see this across macro, micro, cross-sectional, is free cash flow yield. And I know I mentioned it before, but I think it bears going into a little bit more detail. Free cash flow yield is really the first derivative of leverage. If you're constant, if you have a, let's say, a 10% free cash flow yield, you're constantly piling 10% of your market value onto your balance sheet.

[00:48:09] If you don't do anything else, that's going to deliver your company. Of course, you may take other actions that are chosen by the management team to buy back stock or to pay a dividend or to spend on R&D or to make acquisitions. But if you don't do anything, if you just pause and let things happen, you will deliver. We think of it as downside protection in a down environment.

[00:48:34] Now, when an options trader prices one stock with high free cash flow yield and one stock with low free cash flow yield, you know what they do? They do the same thing. They look at realized vol and implied vol and peers. And that relates to whatever environment that you are most recently in.

[00:48:53] And it completely ignores the potential for the environment to change, particularly in the down direction and the environment to, if it gets worse, the company with high free cash flow yield will keep piling that cash on their balance sheet. And the company with low free cash flow yield will find it very difficult to borrow in order to grow their business or in order to shore up their balance sheet.

[00:49:20] And the path of those two companies is completely different in an environment if it changes to be a worse environment. And so what strategies like the one that we discussed, what they really rely on is understanding how the environment could change. And so those strategies may keep up with other strategies during normal environments and then have outsized returns when the environment changes.

[00:49:50] And that's really what we're trying to capture with some of these variables. So free cash flow yield sounds like the most important one. And what would be the next two or so that are part of this model? Another thing is knowing whether there is an event that is happening. So we actually have three different models that have been trained with the same data.

[00:50:13] And one of the models is for months in which there is an earnings event reported that's going to be reported. Another one is if there is an analyst day that's going to be reported. And then a third model is for no event during that month. Why that's very important is we also know how much a stock moves on its earnings event. Right.

[00:50:39] So if you're an options trader and a stock tends to make huge moves on earnings, that's always going to be in your three month realized volatility that you're using to price an option. And that will lead you to overprice options in non-earnings months because that is always in your realized volatility window.

[00:51:01] However, if you treat the stock with a big earnings day move differently in earnings months and non-earnings months, that will lead you to be sensitive to knowing how much of this stock's volatility is concentrated around earnings versus being dispersed all the way throughout the quarter.

[00:51:19] And there are companies that you're constantly receiving data on whether it be an autos company where you get monthly auto sales or retail companies when they give you more frequent updates. And then you have maybe a tech company that is completely quiet during the quarter and gives all their information on the event, on the earnings event. And so the model is smart enough to treat those situations differently.

[00:51:45] And it might lead you to be more likely to sell options on a big earnings mover in a non-earnings month. And would you say that there's a subset of factors that tends to be better at explaining up shocks versus down shocks? You know, are some of the factors more get out of the way versus get involved? Yeah, that's absolutely true.

[00:52:14] One of the great examples of that is we have a service called the Short Squeeze Manager, which has been all the rage for the last few years. But the model really focuses on retail activity and options activity as measured by the volumes relative to the size of the company or the market cap.

[00:52:38] And what we have noticed is these activity indicators of different types of investors, retail and shares, options, and then hedge funds and options. If you can have indicators of those, a two-week average of those activity estimates gives you really great insight on the potential for an upside move over the next one week.

[00:53:07] So we use those variables to estimate the probability of a one standard deviation up move over the next one week to help investors get out of the way of those waves of activity. And interestingly, on the downside, this has very little implications. These same variables are not very useful to explain downside volatility.

[00:53:31] That's a completely different type of function that would involve more of these fundamental variables that we were talking about before. Not so interesting. I think back to the meme episode. It's a crowding in, right? It's a gigantic stock up, vol up episode. And following the volumes, right?

[00:53:53] When the call volumes suddenly exceed the put volumes and volume itself is extraordinarily high relative to history, there's probably some signal there or something you're supposed to at least watch pretty carefully. That's super interesting. And then just on maybe in that same category of thinking about crowding, you guys publish lots of different baskets. You've got the GS VIP long and VIP short.

[00:54:21] Are there any characteristics of those baskets from a vol or a skew perspective that might be interesting to comment on with respect to crowding? One of the types of strategies and cohorts of stocks that we watch very closely is the most shorted basket. So you can look at, and it's publicly available data, which stocks have high short interest.

[00:54:52] And that can give you an idea of when these squeezes are happening that may be painful to investors in the market that are shorting stocks. And that can really inform your view of what type of environment that you are in. And that can help understand where the opportunities might be. So last little topic is just around something you referenced. You said trained the model. And that's a new term.

[00:55:21] You know, we're not really an Excel spreadsheet so much anymore. There's a revolution in the availability of data, the technology that we can use to harness it. We've got large language models. How has that changed your process? And then in terms of what your team is working on, give us a sense as to sort of, you know, we've talked a lot about some of the interesting stuff on the single stock front and the catalyst front.

[00:55:47] But this data availability and ability to analyze it is a real new and powerful thing. I'd love to learn about how that makes its way into your process. Yeah, it's a great question. And frankly, the most important thing for our team really over the last year. And what we have found and what I have found personally is that while I typically was not a programmer,

[00:56:16] you know, I have done programming in the past a little bit, but it's not something that I do every day. What the large language models have enabled me to do is become a programmer, to go in and ask for it, to write code, to build models that I have the structure of in my head. And I know what clients do and why they do it and what variables are likely related to that.

[00:56:45] And it becomes a tool that enables me to put that on paper and to put that into programs that I can run in very quick fashion. Whereas it was just not really possible a couple of years ago.

[00:57:01] And so what I think we're going to find is that people in finance that have decades of experience are now going to be able to use some of the tools that specialists may program or a quant specialist. You know, as long as you know, as long as you know, the type of model that you want to use and you have experts that can kind of check your work at the end,

[00:57:27] you are going to be able to do amazing things with the knowledge that's in your head that you wouldn't have been able to do even three years ago. And so that's been the biggest revolution for us. It has expanded the types of models that we can explore, the speed at which we can explore them.

[00:57:46] And even more recently, performing some of that analysis inside the large language model framework or interface has made things even faster. So I'm very excited about the next year. We've mostly focused on equities and equity options. One thing about our market is it gives you a ton of data, listed options.

[00:58:11] And sister markets are more OTC, but increasingly there are abilities to trace prices in a way that you couldn't. And so are there cross asset, you know, other asset classes for which a lot of this analysis, you think there's going to be a similar focus for you and your team? You know, where are the applications to different asset classes?

[00:58:36] For me, the exciting element is where I can join fundamental analysis with the other analysis that already exists. Trusting, as I mentioned before, within the options market that there are market makers out there keeping everything in line.

[00:58:56] That is actually a pretty important assumption and allows me to jump straight to using other types of data that I think have implications that no one else is using. With some of the other asset classes, I think that there's a huge amount of analysis that has been done, but I'm not sure it's keeping everything in line.

[00:59:19] For instance, credit, far less liquidity than the equity market and some major bond issues, traded, appointment only. And so I don't know whether I have the confidence that some of those are the work, the basic work in a highly liquid environment has been done. And so it may be that I can let add less value in those using the fundamentals, if that makes sense.

[00:59:48] Well, John, this has been a pleasure. I really enjoyed the conversation. It's been great to explore some of your work, both on the kind of macro side, but also really the micro side, which I don't think gets as much attention in a lot of ways. So your focus on catalyst-driven option trading, I think, is super interesting. So thank you so much for being a guest. Thank you very much for having me. I enjoyed the conversation as well.

[01:00:38] Thanks again and catch you next time.