The Derivative - Hedge Fund Quants Unlock the Power of Dispersion: Versor's Unique Cross-Sectional Relative Value Approach

Episode Date: July 25, 2024

In this episode of The Derivative, host Jeff Malec sits down with DeWayne Louis and Nishant Gurnani from Versor Investments, a quantitative investment firm celebrating its 10th anniversary. The discussion delves into the personal backgrounds and journeys that led to Versor’s founding, from their work providing efficient exposure to traditional hedge fund strategies through risk premia type approaches to new methods such as their unique GETT program which goes long/short more thana dozen global stock index futures markets.   The conversation explores the evolving skill sets required in the hedge fund industry, including the importance of math, quantitative finance, and AI/machine learning expertise. Versor's investment strategies are examined, focusing on their use of alternative data sets and innovative approaches to alpha generation, such as their cross-sectional relative value strategy which looks to capture dispersion in global equity markets through more than 30 alpha forecast models across short, medium, and long term time frames.     Join us for this comprehensive look into the quantitative investment strategies and innovative thinking that have helped Versor navigate the competitive hedge fund landscape over the past decade. Chapters: 00:00-01:38=Intro 01:39-12:22= Summer in the City and starting a Quant firm 12:23-25:10= 3 broad strategies, cross-sectional portfolios & positive convexity 25:11-42:57= GETTing into the strategy, dispersion across markets & capturing market dislocations 42:58-54:26= Alternative data sets & natural language processing 54:27-01:01:38= Using A.I., generative models & compliance constraints 01:01:39-01:10:51= Long short equity, use of data & the next 10 years From the episode: Has Trend Gone Flat? Return Convexity in Trend Following (whitepaper) Versor10: A Decade of quantitative research in 10 whitepapers Bastian Bolesta on The Derivative episode Salemn Abraham on The Derivative episode Follow along with Versor on Twitter @VersorInvest, look them up on LinkedIn - DeWayne | Nishant and check out their website versorinvest.com for more information Don't forget to subscribe to ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠The Derivative⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠, follow us on Twitter at ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠@rcmAlts⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ and our host Jeff at ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠@AttainCap2⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠, or ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠LinkedIn⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ , and ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Facebook⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠, and ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠sign-up for our blog digest⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠. Disclaimer: This podcast is provided for informational purposes only and should not be relied upon as legal, business, or tax advice. All opinions expressed by podcast participants are solely their own opinions and do not necessarily reflect the opinions of RCM Alternatives, their affiliates, or companies featured. Due to industry regulations, participants on this podcast are instructed not to make specific trade recommendations, nor reference past or potential profits. And listeners are reminded that managed futures, commodity trading, and other alternative investments are complex and carry a risk of substantial losses. As such, they are not suitable for all investors. For more information, visit ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.rcmalternatives.com/disclaimer⁠

Welcome to The Derivative by RCM Alternatives, where we dive into what makes alternative investments go, analyze the strategies of unique hedge fund managers, and chat with interesting guests from across the investment world. Hello there. Who's ready for the Olympics? I was just out in Colorado Springs and did the tour of the Olympic Training Center. It's pretty cool. So I'm full on ready. I'll be from the new breakdancing to the surfing and everything in between. And when will they have e-gaming or quant contests? Because the duo we have on from Versa Investments today would surely compete. We've got Versa's founding partner, Dwayne Louis, and head of Futures and FX partner, Nishant Gurnani joining us, taking us through a brief history of Versa from the days at Invescorp to time as ARP and not to their risk premium routes to now Versa and their new program trading long short 18 plus stock index futures around the world. We get into

AI, alternative data sets, and how a mean reverting model can actually exhibit positive skew. Some nerdy stuff. Send it. This episode is brought to you by just plain old RCM, not the managed futures group, not the execution desk, not the ag unit, just plain old mothership RCM. From clearing to execution, portfolio construction to outsourced fund ops, China to Nebraska, RCM has you covered for all things futures and derivatives. Learn more at rcmalts.com. All right, everybody, we are here with Nishant and DeWayne from Bursar. How are you guys? We're doing well. Great to be here.

Thank you. Excellent. We just spoke quickly offline that it's duane louis he's french haitian not lewis i get that right indeed i got that right like like louisiana louisiana i love it um and so your parents were born in ha? My father was born in Haiti. My mother was born in Jamaica. Okay. You get back to Egypt's places?

Yeah, I have more time in Jamaica than Haiti. Haiti tends to have some challenges from time to time. So growing up, always had a little challenges when we planned to travel back. And obviously, it's gone through some challenges from time to time so growing up always had a little challenges when we plan to travel back and obviously it's going through some challenges now i hear you uh and nishant what's your what's your backstory got anything half as interesting not not as interesting as for sure i'm i'm of indian origin i grew up in connecticut um grunani is a very standard indian type of name nishant is a little bit more unusual there aren't that many chance out there but uh you know it's uh it's a

pretty it's pretty nice nice name to have it keeps me unique i love it and now you both are in new york in manhattan that's right that's right we're both actually feet sitting feet away from each other but in different rooms i love it i know one day we'll figure sometimes we have people in the office or in the city and we're trying to do the podcast i still haven't been able to figure out how to do it live so we'll literally be like yeah four rooms away from each other doing the podcast but one day we'll get the technology um and where's verser's office right there in midtown yeah we're in brian park we just got the bank of america building right behind me right here so perfect it's like

a nice day there and you guys aren't guys who head out to the hamptons or whatever during the summer you're stuck there not i i don't done stay around we stay around most of the time dwayne you get out you know yeah the northeast is nice in the northeast so not the hamptons we um i spent a little bit of time in martha's beard sometime in august but uh we do most of the summer we're here all right um so let's jump into the verser background your guys personal backgrounds how you got into Versa? Who wants to jump in there first? Let's do it.

Maybe I'll start. So Versa, we are a quantitative investment firm. We're actually celebrating our 10th year this year. So 10 years ago, a group of us, a lot of firms that were led by Deepak Granati, who's our managing partner. I'll give a little bit about Deepak's background because it really gives a genesis story to how we came about. So Deepak spent about 20 years of his career at a place called InvestCorp, where he was the head of the hedge fund business. And InvestCorp, I'm going to say we, because four out of the five of us that started the firm worked together at Invescorp. At Invescorp, we had a hedge fund business that looked to invest in other hedge funds, looked to seed hedge funds,

and then looked to do some internal investment of hedge fund strategies. Deepak and others at Invescorp in the early 90s sought out to do a research project that was designed to better understand the quantitative drivers of various hedge fund returns. So you can think of this as hedge fund beta and examining the beta associated with hedge fund strategies. At Invescorp, that was a big part of understanding dynamics around when to potentially invest in hedge funds, understanding the dynamics around managers that were able to exceed or investment strategies were able to seed those systematic drives of hedge fund strategies, and then when to tactically tilt in and out of strategies. I've mentioned this research because as a quantitative investment manager that started 10 years ago, it's oftentimes important to note that

we started our mission or our quest in achieving returns in the quantitative investment space going back 30 years ago and really starting with the research that Deepak and others started in the best group. So 10 years ago, Deepak, myself, Ludger Henschel, Andrew Flynn and Nirav Shah were the founding partners and many other partners, including the shot. We started the firm and I've been at it ever since. Was there any bad blood or any leaving Invescorp where they said, you're welcome to go start your own thing, hang your own jingle? No, it was quite a amicable departure. So, you know, Deepak had finished 20 years at Invescorp at that time and how it works, and he continues to work at this juncture, is that, you know, all of his stock in the firm at Cliff Fest,

it was a friendly departure at that juncture. He had articulated he wanted to go off and do some other stuff, other things. He actually spoke to Invescorp about myself and others who left the firm about joining it. So it was a quite amicable departure when we started the firm and no bad blood. We continue to remain colleagues, friends with our colleagues today. Love it. And Nishant, Invescorp was mostly sovereign wealth money. Is that correct? I believe so. So I was not, to be clear, I was not at Invescorp at the time.

Oh, got it. One of the more notable parts of my background, if you can see me on the YouTube version of this or you'll see my last name is i'm deepak nishant gurnani um and so i all coming together now we got it yes indeed so i've been unofficially part of the founding uh in the beginning and then officially joined the firm in in january of 2020 um about around the time the derivative podcast actually started. Exactly. Right around the time the fireworks started. Was that so right after you joined the big crash, the COVID, everything? Was that?

Yes, it was. It was a very interesting time to start. Actually, you know, sometimes luck happens this way. Things were interesting from the get go. And, you know, it certainly shaped my experience at the firm. And what was your background before that? I studied math at Princeton, and then I studied statistics at UC San Diego. Having been grown up in Connecticut and obviously being the Bucs' son, I was always sort of exposed to the hedge fund finance setup. A friend of mine used to joke in college that normal people when listing alternative careers would say, you know, I might be a doctor or a lawyer or something, and I would just list alternative things within finance itself. So like a good budding quant, I studied math at Princeton. I got to spend two summers while in

college at world-class quant firms. I spent a summer at AQR, and I spent a summer at what used to be called SAC MultiQuant, but is now called Cubist. I spent a bunch of time in grad school working on bootstrap time series, for those people who are familiar with those types of techniques. And then I got a bit of a tech bug, and so spent a year in san francisco working for a fintech before finally coming back and joining bursar full-time and what what's your thoughts on how trying to form this question in my head but how necessary is that math background like do you need a math background you need a quantitative finance background do you need a math background? Do you need a quantitative finance background? Do you need a AI background?

Do you need a, right? It all kind of blends together these days. It all sort of does blend together. I think one of the things that we tend to emphasize is that it's a combination of all those three skills, a combination of good modeling skills, particularly because we're fonts and that's the language and skills that we use to identify alpha and other opportunities. But it cannot be done devoid of financial market expertise and so our head of research luker and the book have been writing white papers and doing internal research for a long time as wayne alluded to the alpha

project and so you really need this marriage of quad skills software engineering skills and and genuine market insights to find alpha in these competitive markets and do you think that those skills are becoming less so with AI and prompts and right programming right itself you know I think this is somewhat a contrarian opinion I don't think they're becoming necessarily less rare I think they've just evolved and and evolved and once we've given tools and data access to a bunch of people, the ability to find alpha just has to be more creative in different ways. So the fact that LLMs and alternative data sets are table stakes today just means that

using them, you have to be a lot more creative than you had to be five, seven years ago. At Versa, we started using alternative data sets around 2017. We've been using machine learning since the very beginning, and I've certainly increased that scope over time. And what we've seen consistently throughout is you just have to be more clever. Assume this is a competitive business, things that worked a few years ago no longer work.

Things that work now will probably not work in the future. And so constantly adapting and evolving is the way to go. What I would add is, allow me, probably not worked in the future. And so constantly adapting and evolving is the way to go. Right. It's like, you know, what I would add to that is, you know, when you look across the team and particularly the team that Nishant leads, increasingly we're looking for men and women who come from quantitative research-oriented backgrounds.

So not necessarily folks that grew up in financial markets, but what's important and I think will continue to be important for firms like ours is the fine men and women who have classical training in dealing with large, unstructured alternative data sets. So data sets that don't come in neat rolls and tables that were designed to use the financial markets, but data sets that might be a series of letters and words or data sets that might be unstructured quantitative data analysis, and then having the skill set of my brain jumps to like, if you're sitting around the dinner table as a kid and your dad's quizzing you on statistics or anything, was there any of that kind of stuff going on? I think there was some of that, but honestly, he was fairly hands off. There was no necessarily encouragement to pursue this.

I was given a lot of freedom. It's just something I always enjoyed. I think general discussion of markets was particularly pertinent because we were living it and I happened to grow up in a town that is very, very finance and hedge fund focused. And so it was pervasive, but there wasn't necessarily a push here that way versus pursuing something else. Duane, I'm curious, since you guys peeled out and started Bursar, what was the choice? Why did you guys go with, we're going to design and do our own quantitative models instead

of we're going to be a fund-to-fund or allocate? What was that decision like of, let's do our own models instead of pick out the best managers that we know. So I'd say, you know, going back in the history of Bursar, maybe I'll spend two minutes describing that. So when we launched the firm 10 years ago, we actually launched as a firm called ARP Investments, Alternative Risk Premium Investments. So going back to the research I described,

I think we identified through the use of our, you know, conducting the quantitative research on the system that drives the hedge fund returns. We thought that there were economically efficient ways for investors to get exposure to hedge fund strategies. So we launched in order to provide that. So explicitly looking to provide kind of the sort of exposures that one would get from a fund to fund. As we've evolved over the last 10 years and as we've looked to, you know, generate alpha through other aspects of markets, increasingly looking at alternative data sets, we've ventured into more hedge fund oriented strategies, more alpha seeking strategies, particularly strategies that level alternative data. We changed our name to be explicitly clear that we offer both risk premium and hedge fund strategies. But the genesis, going back to your original question, the genesis was explicitly

to provide investors economic, more efficient exposure to the sort of things that one might get from a fund to fund, one might get from a series of hedge fund strategies. And then I noticed you sort of avoided the word replication, but was it kind of trying to replicate those returns? I wouldn't say replicate. I shy away from the word replicate because replicate can often, one can use derivative to replicate the risk associated with hedge fund strategy, but not necessarily replicating the returns associated with hedge fund strategies. The strategies that we have employed from the outset were invested in the same financial, have invested in financial instruments that are used by hedge funds, and the returns are designed to mimic and potentially exceed the returns that one gets

from investing in hedge funds. That was the case at the outset. Today, we're explicitly looking to generate alpha and be top quartile in terms of return stream. Love it. And that is what we're going to focus on. So in terms of that alpha generation, how many programs do you guys currently have, which then we'll whittle down into the one we're going to focus on today. But so in terms of, so you're saying there's still a risk premium piece, which you guys are doing that piece. And then there's also the alpha generating piece. So maybe quickly, how many pieces are in the risk premium? And then how many pieces are in the alpha side? Maybe the best way for me to describe it, because the way the teams are divided,

it's the same investment engine that works on both and then we split it across. So there's three broad strategies that we provide. One strategy is quant equities. So quant equities, we invest in single name stocks on a market neutral basis, both in North America and the US, where within the quant equity program, we have some style based investing value, momentum, quality type signals. And then we have more esoteric or more idiosyncratic sources of returns that leverage some of the alternative data sets that we'll describe, hopefully, as we go through this conversation. The second group are futures and FX strategies. So when the futures and FX, where we'll spend some of the conversation today, we invest in futures contracts across all the major asset classes, equities, fixed income, currencies and commodities in styles that resemble momentum or trend investing in styles that resemble systematic macro investing and in styles that are more idiosyncratic. And particularly the GET strategy will describe that investment in equity index futures across both developed and emerging markets. And the last group of strategies are merger or equity event strategies. So we have a systematic implementation of equity events, predominantly

focused on merger arbitrage, where we've leveraged a proprietary database that contains every announced deal over the last 30 odd years, and then apply machine learning and AI techniques to understand the attributes of deals that lead to the increased probability of several events. So things like failure, success, competing bid, and then invest or not invest in deals and wait deals appropriately based on a probability of those events. Awesome. And so in each of those main buckets, there's both the risk premium and the alpha seeking. I get it. That's exactly right. And then Nishant, so you've been mostly focused on which part of that?

I lead the futures and FX research at the firm, in addition to being involved in the broader efforts on alternative data and AI in general. Within the Futures sleeve, we offer a GET strategy, which is the flagship Futures strategy. It's the Global Equities Tactical Trading Strategy. It's a strategy that came about as a result of internal research that we think is a very natural complement to a trend-falling approach. It provides the same type of positive convexity profile that trend has originally attempted to provide, but we've noticed has not provided as well, at least in the recent past years. And this strategy is particularly unique. We don't believe other folks do it the way we do, and we have some genuine insights there. And it was a great marriage between internal research efforts that were ongoing.

Libra and Deepak have written a bunch of white papers on this topic. Most recently, they've written one called, Has Trend Gone Flat? Return Convexity and Trend Following from two years ago. That explains the main idea. The main idea here really stands from, we started to notice very early on that the positive convexity profile that trend followers are supposed to provide, particularly in the SG trend and CTA indexes, started to disappear, particularly post-financial crisis. And we realized that we could provide a lot of return convexity by taking a cross-sectional relative value approach, as opposed to a

directional time series type approach. And that really was the genesis of the GET strategy, which replicates this cross-sectional market neutral type approach using equity index futures, 24 equity index futures across global markets, 12 in developed and 12 in EM. I want to put a pin in GET for one second. We'll come back to that. And just I'm curious on the white paper and what you guys were researching there. Right. My feeling and from people have been on this podcast and managers I talked to, a lot of that had to do more with the trend followers. filters adding carry adding some different pieces to survive those flat periods for trend so they kind of morphed so i don't i don't know if it was the trend signal or the space morphing or whether

you guys cared if you just said hey this is what's happening yeah so we we looked at we looked at both so we look at a style decomposition in that paper to sort of attribute where the risk can be uh attributed towards the trend type signals or, broadly speaking, the other bucket, which was certainly a lot of carry and volatility. But even within the trend filters, the trend signals themselves, we found, and this is not unique to us, other people have also seen this, that the balance between looking at long-term type trend signals versus short-term signals has deviated over the years. If you do this attribution on the SG trend index,

for example, you'll see in the paper that medium term sort of entirely disappeared. And you're certainly hit the mark on this and other guests of yours have said the same is these were competitive pressures that inevitably caused people to change the character's profile. And ultimately we didn't wanna do that. We wanted to change the character's profile. And ultimately, we didn't want to do that.

We wanted to provide clients that positive convexity because that's what they really expected from these products. And they didn't necessarily deliver that. And look, it's a function of an adapting market for those. Right. I would say in the past on here, if you're given the choice change your stripes or the leopard changes spots or go out of business you're going to change your spots right so i think that's mainly what's going on there but then curious

to me that you guys ended up at that cross-sectional relative value piece instead of saying let's create a more trendy piece that's not doing all these things that take that trend component away from. How did that logic come about? Yeah, so we were really looking at trying to identify. So some of this, Dwayne is very humble about it, but there was a lot of push from Dwayne's side for us to come up with innovative products that were really differentiated in the marketplace that other people didn't necessarily have. We've been running a flagship trend risk premium product since inception that has done very well. But we really wanted and we're looking for something very differentiated. One of the advantages of our firm is we do have this open research framework. So we talk across teams. And so a lot of the cross-sectional approach was initially inspired by the quant equity

stat R type strategies that we do. And so one of the couple of differentiators that we have on our relative value approaches, commonly relative value approaches, particularly in the future space, are done over a few contracts. We do it over as large a cross-section as possible. We do it on a 12 developed markets and 12 EM equity index future markets basis. And then in addition to that, what was super interesting about doing it this way is it allows us to construct signals from both bottom-up single stock level up to the index level, as well as top down from the country level to help

forecast the cross-sectional predictability. So, Duane, that was as simple as raising your hand or someone on the team was like, hey, what if we can provide the same trend profile, but with this new model that's different than anything else that's out there? Yeah. So maybe taking a zoom out for a second. And I would say some of the conversations that we've had with clients, you know, from the very outset of a business,

we've worked with public pension plans and, and a number of public pension plans in the U S where funding status may vary, but the major concern for public funds is managing around their funded status and paying attention to drawdowns in their portfolio. So the predominant risk in any public puncher plan is equities. So I think the challenge as we were engaging the plans and the consultants, was focusing on mitigating risk. So risk mitigation, and then by various terms, crisis risk offset are terms that are oftentimes used

in the public pension community and some of their consultants to describe the objective of a portion of their portfolios, the alternative portion of their portfolios. I think the frustration that we identified amongst the public plan community is that there are some strategies, alternative strategies, that do well during a protracted drawdown. So there's a drawdown that persists for several months, maybe a year or more. There are strategies like trend that do a very good job of protecting against that.

But where I think some of our clients have been frustrated is in trend as a group, trend following strategies as a group, their ability to mitigate risks during sharp drawdown events. So that was led us to kind of go down the path of let's think about convexity over multiple different time horizons. So let's think about providing positive convexity when there's an extended drawdown like 2022, but also think about what it means to provide convexity and provide diversification in a period where we have a sharp drawdown that lasts for three or four months, Q4 of 2018 or the first quarter of 2020, and think about designing a strategy that responds well to those shorter episodic drawdown events or those long, more protracted drawdowns. Looking at convexity across multiple different time horizons.

Getting into the GET model, 12 developed, 12 emerging. So that's essentially the entire universe of the futures indices. There might be a few others. Yeah. So these are global equity index futures. And so these represent, in our opinion, a great liquid global coverage of indices. So it goes from ones that are more liquid contracts like S&P 500 and FTSE all the way down to some of the least liquid contracts

we trade are WIG20 in Poland and FTSE KLCI. And the investable universe has been selected through a lot of research on our process to be in such a way that we can see enough dispersion among the country equity index futures, but not so much that there's a material dislocation that is unrelated to each other. So it's a fairly fine-tuned process of finding a cross-section that moves reasonably well together. And the split between developed and EM is done because we believe those two profiles have very different risk characteristics that we want to encode in the portfolio construction process. And then let's, so let's dive into the strategy a little. Is it, are you trying to have one signal where this is kind of a voting machine and you have out of those 24, 20 long, four short,

and it's giving you that symbol or each is its own trade, so to speak, in a pair? So the way we think about it is we think about it in terms of a market neutral cross-sectional portfolio. So we have over 30 alpha forecast models that we split into three broad buckets. So the first bucket is what we call price dislocation type strategies. These are short term that effectively are trying to make bets on relative value movements across the entire cross section. So I want to know if FTSE KLCI is going to outperform WIG20, which is going to outperform

GSC40 and take long exposure on half the 12 short on the rest so that we remove the market beta exposure. The short term signals try to exploit price dislocation, stuff that can broadly be categorized under deviation from law of one price. These tend to be the ones that provide the most positive convexity and tend to do the best during periods of market stress and volatility

when there's a lot of inelastic flow that goes into these contracts. So for example, in the recent past, we've seen a lot of movement in the European indices due to the French snap election. The week of the Indian election, we saw some big moves based on expectations for the BJP to win versus that not being the majority it was. And so we're really trying to exploit those types of moves in the short-term price dislocation signals. The second category is what we broadly call medium-term or under-reaction signals. These are more momentum-like type effects that tend to

manifest themselves less on the order of days and weeks and more on the order of a couple of weeks to a couple of months. And these tend to be price momentum type things. But we've also spent a lot of time on economic signals, looking at economic data that we collect, now casting a bunch of forecasts and surveys that we have to come up with some market view for each equity index to exploit this sort of medium term momentum effect. And then the last bucket is what we consider the valuation long term type signals. These are very bottom up. This is one of the places where we do construct measures from each individual stock level all the way up to the equity index market, just getting a good sense of value. And they tend

to be a couple of months on average in terms of turnover. And what we do in our portfolio construction process is we combine the signals across all these three large buckets. We do give a higher weightage risk weighting to the short-term dislocations strategies, because that's what provide us this great positive convexity that tends to do well during these times of market stress. But the medium and long term, while also providing some positive convexity, also help with modulation of the strategy performance during times of less volatile periods. So, for example, a lot of the beginning of this year, we've seen VIX at a very low level throughout.

It hasn't broke 20 this year to date. Those other signals tend to kick in during those times. And to echo what Dwayne said, they help sort of provide the profile where the strategy can do well in periods of slow drawdowns, equity market drawdowns, as well as these more volatile shock type drawdowns. And then each of those models, or let's just start one at a time. So in the short term, is it always paired off? So is there always a short each model? So there are always six long and six short, which across each sex are out.

So there's a market neutral. There's no exposure taken in any. There's no directional exposure taken in any of the markets. Right. But your quote unquote market neutral layers, some sort of fictitious global market. Yes. It's the market neutral with respect to the 12 developed indices and the 12 EM

industries, each of which are, are market neutral individually. Yeah. So it's, it's market neutral, like versus the MSCI EIF or the Emerging Market Index? Yeah. Ostensibly, yes. But in practice, what we're doing is we've defined our 12 developed equity index futures universe. We've gone long 12 of those, long six of those, short six of those. Same on the EM, which it provides as as you described it

got it um and so how do you differentiate that for a minute versus traditional long short equity right which is in single names so that's a huge component right there but what else do you find of like in the profile that differs so one of the things that's an advantage of doing it in the equity index futures contracts is it's the right level of resolution that allows us to get this global profile for a lot of the em contracts uh it's non-trivial it's not impossible to trade the underlying stocks even though we have a lot of data on those things it's pretty hard to go out and trade FTSE, KLCI, individual stock within that index. And so equity index futures was sort of the perfect resolution of liquidity and global coverage to express this global equities tactical allocation. And that's what sort of really motivated this setup of focusing on the equity index futures, but then utilizing components from the individual stock level, the options market,

intraday data, anything that we can to come up with a view for that particular equity index market. And then Dwayne, how did you talk with the customers who I'm assuming, but maybe not, you could correct me if I'm wrong, but some sort of like another long, short equity thing? We've got enough of those. We've seen a million of these. What's different? Oh, you've got some fancy factors, right? I feel like there was long, short equity fatigue there for a little bit. So how did you get over that yeah yeah i think you know what i would say one of our initial challenges and with the strategy we've been running now for uh seven years our initial challenges that we've had is where does one put this in their portfolio yeah right so as you alluded to you're referencing it trades equity-like instruments um but their futures contracts so is it a is it a managed future or is it

long-term equity strategy? So what we point to is some of the artifacts of the returns, but I think how we articulate the value-added strategy in the opportunity set isn't what we're doing. Intuitively, what we're trying to exploit are dislocations that are caused by the buying and selling behavior between two distinct market participants. So Nishant alluded to things that we do over the shorter term, over the medium term, over the longer term. But the interesting thing about equity index futures is that there is an underlying. There are multiple underlines.

And if we could look at the buying and selling rationale of both parts of those underlying, we're able to identify really interesting trading opportunities. So a stock investor. So as we're explaining, a stock investor typically buys a stock because it's seeking to outperform an index. I own Microsoft because Microsoft, I expect Microsoft to outperform the S&P 500. If I buy and sell an equity index future, it's not because I have a view on a particular security, it's that I'm trying to manage an exposure. So the rationale behind kind of a top-down allocator versus a bottom-up investor is very different. And what we express, we articulate to potential clients is that because of this difference in the rationale behind why a person buys Ether security, it leads to dislocations in their behavior, particularly during periods of market stress.

So I think we do a reasonably good job of articulating that and articulating why that should persist, particularly in times where you need it most. So what we're really looking at is dispersion across these markets. This dispersion tends to increase as volatility increases. As volatility increases, it tends to coincide with equity markets drawing down. And that's where we tend to do well, particularly on our short-term signals. And that was going to be my next question. Is it fair to kind of look at it as a long dispersion trend, which it seems like you just said yes? So I want to also just define very clearly what we mean by the dispersion, because I

know it certainly has different meaning in the options space. What we're really interested in is in relative value movement among the country equity indices. One of your previous guests has described the world equity markets as this global relay race. And that is certainly the characteristics of these things, where the S&P 500, if you think about it, is sort of leading a lot of behavior. There is some beta component exposure across all equity index markets that leads them to move in specific ways. So yes, we are trying to exploit dispersion,

specifically dispersion amongst the performance of equity index markets globally, which often tend to manifest during times of high vol periods, but also are a function of economic performance in these different markets that we trade over the longer term period. So my brain was going to a second of like the S&P 500 is comprised of many of these stocks that are also in some of those foreign indices. Is that fair to say? Right. So it's almost when you're buying the S&P and selling those foreign, you're kind of have a exposure. Yeah. A dispersion trade. Yeah. Not with options, not the volatility of them, but.

That's exactly right. And that's what makes this a particularly interesting strategy and also one that's non-trivial to implement. So one of the reasons we look individually at market effects at the stock level is because we are trying to understand these global movements versus the idiosyncratic component, which is what we're actually trying to exploit. Because as a result, what you alluded to, even from a macroeconomic perspective, a lot of FTSE, so FTSE 100, a lot of the company's revenues in the FTSE 100 are derived from external trade as opposed from local UK market trade. And that has an explicit component to that. One of the things that you were also referring to maybe two questions ago that Duane was talking about with regards to long short equities, where does this sort of fit in? Because of this macroeconomic approach, we do think it fits more naturally on the future side. It also was designed from the future's

investors' perspective. This is something we do emphasize heavily that we came upon this because of our own internal research. It was not something we just sort of found. Questions were asked around how do we get this positive convexity? We found the cross-sectional way of doing it was great. We looked at doing it in different asset classes.

We found equities in particular was a great place to do it. You could get this profile and you had this advantage of, this is unique to equities as an asset class, having stock level data that you can do bottom up and top down, which is very unique. You can't do that in FX markets. You cannot do that in fixed income. You cannot do that in commodities to the same level. And so that's where that came from. Some nat gas traders would disagree with you, that they can look at all the different hubs and do it.

But I understand what you're saying. And then, so is there a typical trade? Is it typically long, more developed and short, or it just varies depending on what's going on? It will totally vary depending on what's going on. And it's a function of the market dynamics that are taking place. Because we take a little bit more risk exposure to our short-term strategies, a lot of the short-term price dislocations tend to lead, but are certainly not dominating these effects. We will see things, it will go from profiles of HAV being long, CACS40 because the short-term dislocation suggested as such,

to then going short CACS40. And even within regions within the 12 developed in EM, so if you just look at European indices, it's not uniform. They'll necessarily go long all Europe and short sort of North America. It will take different positions within those as well. And then maybe the French election would be a good example of, it sounded like you were saying you use FTSE, but I'll try and put an example. You can correct me of like, okay, we're seeing

that the French market is selling off because this French election, but inside of that market are these 50 stocks that really don't have much to do with the French economy. So we want to go long that sell the other, get market neutral in order to take advantage of that dislocation. Yeah. To be clear, we're not taking positions in the individual stocks. Yeah. To use the French election, a more interesting example is, if you look at the, sorry. Yeah, no, save me. CACS 40, with respect to the other European equity indices behavior during this French election time. So FTSE MIB, CACS, FTSE MIB for Italy, CACS 40 for France, IBEX 35 for Spain, what we saw was reflective of what perceptions people had,

which was we had the snap election in France, and suddenly there were concerns around Euro region stability in general, and we saw IBEX 35 fall tremendous. And what we're trying to do is capture these types of dynamics where in times of stress, yes, all instruments are correlated. And so they'll move in the same direction, but they're not going to drop to the same extent. And some of our signals will do this just by looking at futures technical data. underlying stock components and saying, oh, actually, this is really being driven by one stock or a couple of stocks or the index is unusually imbalanced compared to historical norms, it's going to snap back. It's not going to move as much as another where we've seen sell-offs broadly across

all the stocks in that particular index. And so on a relative value basis, we should probably go along the one that didn't see as much internal dispersion in that index, as opposed to the one where we saw a lot more dispersion. And then it's... Maybe just philosophically... Go for it, Dwayne. Go for it. I'll say maybe philosophically just engaging with some of our clients. We work with a number of public pension plans that might have a beta overlay in their portfolio.

So we know in conversations with these plans and as Sean and his team can observe, there are periods of market stress. There tends to be rebalancing activity that occurs in the futures market then moves to cash. So there's a market that says, Macron calls a snap election, there's a market movement across Europe. We noticed that certain institutional investors or investors will use a futures market to do these beta rebalancing exercise. Again, it's this top-down beta rebalancing exercise before doing bottom-up exercise moving to cash.

That creates this buying or selling pressure in the futures market that we can exploit, that we're seeking to look. So look at dislocations across markets and exploit those dislocations cross-sectionally. And are you saying there the institutional will come in, hey, I want to lighten up. I'm going to sell the futures. And then over the next few days, I'm going to buy, I'm going to sell my individual holdings and buy back the futures to get to where I want to be.

But they can do it quickly and as a proxy in the futures first. Certainly. and as a proxy in the futures first. Certainly, there's some of that, but there's also some of just passive fund flow activity that we can track using alternative data sets where the amount of passive investment that's tracking some of the more larger European indices far exceeds that for some of the smaller European indices that we're also interested in. You mentioned the alternative data set.

So let's dig into that. You guys are big on that. And then we can also touch a little on the AI piece. But sure. Talk to us a little bit about alternative data sets. It seems cliche at this point of like hedge fund using alternative data sets. But what's your unique take on that? Yeah.

So for us, the first thing that we really like to emphasize, which is that for quants, using alternative data is a very natural part of the alpha generation process. What we're looking for is data sets that are repeatable, that are consistently released on a regular basis that can allow us to forecast the returns of various instruments that we're interested in. My sort of contrarian view is the term alternative data

tends to get used and bandied around a lot. Certainly the amount of alternative data available to us in 2024 far exceeds that what we've seen over even the past few years. That doesn't mean the amount of alpha has grown exponentially as well, right? It's just not the case. And so for alternative data for us, we really employ our scientific alpha hypothesis driven approach where we ask really hard questions about does this data set

even have any causal relationship between the instruments we want to trade and the data set itself? This is not a data mining exercise. There needs to be a relation between these things. And that's a human decision of, yes, there's a relation. Yes. So we have statistical ways of talking about it as well, but there needs to be some sense of the thing that this data set we found affects company revenues because if it doesn't, there's no reason.

One of the things that people underappreciate is no alternative data set just exists because it's out there. The fact that data is being collected on a daily basis by somebody, by a lot of programmatic systems, and then being cleaned in a specific way, and then even sold for others to use means that a human at some point thought it was worth collecting in the first place. And our job as financial investors is to figure out whether that particular data set has a causal relationship with the thing we're trying to forecast. So let's take the canonical example that everybody loves to use, which is credit card data. Credit card data is something that we use and certainly others do as well. And there is good reason to do that. There are some non-trivial number of companies in the S&P 500 who derive material revenue from credit card sales. But you have to ask questions about, OK, what are those companies? What are the companies that you're not going to get in that universe?

And then in addition to that, what you end up having to do is you end up having to be a little bit more clever about how you utilize the types of alpha that you can derive from that data set. So maybe a few years ago, it would have been reasonable to take rolling window statistics of forecasts of KPIs that were fairly simplistic and make genuine money doing so. That is no longer the case. You need to combine these forecasts with a bunch of other alternative data forecasts. You need to ask good questions about coverage. One of the challenges that is underappreciated on the alternative data format is operationally, it's very difficult to deal with alternative data. Quants are best suited to deal with alternative data because they're used to running large market data systems on a daily basis with vendors sort of dropping feeds and changing panel sizes.

And there's a lot of alpha in overcoming even those operational challenges, which means that you can do useful things with the data that you get out of that process. What's one of the most unique or esoteric or weird data sets that you use that you're allowed to share? So I would put it differently. So rather than it being a unique data set that nobody else has heard of, there are very sensible market intuition based combinations one can do that potentially will add interesting alpha to your strategy. So I'm going to give you an example of one that's relevant to the modern day is the result of LLMs and natural language processing getting a lot better is that you can now extract interesting text alpha from smaller texts that you were unable to previously.

So job descriptions, where job descriptions themselves are not these long, large texts that have a lot of information in them, they're sort of fairly short, but you can extract information from the job descriptions itself. So no longer are you looking at just jobs, growth and number of change in employees for a particular company, you can also look at the content and the language that they're using of the hiring profile. And then that can become an interesting signal. So a lot of it is about,

as has always been the case in finance, clever insights into existing things that other people aren't necessarily noticing, as opposed to some unique data set that nobody else has access to. That's, I'm going to say, from a lot of experience working with data sets, that's largely that those don't exist. It's about being more clever in utilizing the data that is available, broadly speaking. Are you guys bigger NBA or NFL fans? I'm going to put it in some sports lingo.

I'm going to say NBA. Dwayne? fans i'm gonna put it in some sports lingo i'm gonna say nba all right i'm a football guy uh all right we'll do we'll do but right the nba you used to have points rebounds assists winning margin who knows now there's like points with the closest person against the closest guy guarding them, clutch points, all this different data. In the NFL, it might be yards after contact, yards between the tackles, contested catches, all that stuff, right? Versus, so that to me is kind of what you're saying of like, we're just expanding the data set inside this universe versus,

oh, we're getting points, rebounds and assists from the South African league or the right, the Thailand league or something and be like, oh, we have this new data set, these new points. So is it is it you're saying it's more of the like, we're expanding and doing different analysis on the current data set of companies or revenues and earnings and customers versus like, we're grabbing this data from out of nowhere? Yeah. In 2024, I would say it's a bit of both, but it's definitely more on the side of being more clever with data sets that are out there. Again, there's been a tremendous explosion of data availability. If you go back, so actually, let me preface this thing. There's been a data explosion specifically for developed markets.

If you try and get a lot of the cliched alternative data for even something as simple as European countries or South Africa, it's very hard, if not impossible, to get existing interesting data sets that you may be familiar with, but are extremely hard to acquire for EM countries. Within developed countries, US specific, but also the UK, it's really about being clever and more insightful about using a bunch of the different data sets that you have available to you. And is it better- And I would say, just going back

to the analogy real quickly i think what we spend a lot of time when the shot in his tent spends a lot of time is around like framing a problem so the nba analogy is it is to best find the best point guard for your particular team is using data sets that are available globally to help us identify the best point guard right that won't just be assists anymore. It'll be these different pieces. Exactly right. Exactly.

And how it relates to your team, right? So we have specific problems that we're trying to solve. We'll go out and seek data sets to help us solve those problems. But the key there is around framing the problem it is that we're trying to solve. I love it. And then does this create a bit of a moat for you guys, right? Of like, this has to be very expensive to get all these alternative data sources.

And right, is that increasingly so in the machine learning and everything you have to put on top of that? Like, what does it look like in terms of expense to run this? And scale you need to run it? Yeah, I think it certainly is. But to use another sports analogy, I like to think about it in terms of Formula One. So Formula One, the research span between the top teams and the middle teams is a factor of 10 difference usually between your Ferraris versus everybody else. But that doesn't mean the middle teams can't compete. So there is a minimum spend of running

a large, sophisticated quantum infrastructure. That's just the case. We've been early users of AWS since the beginning. All of us pay large AWS bills, large cluster compute, and large bills for the alternative data. But it's also about being more clever. One of the things that people underappreciate, this is why I have this contrarian view that alternative data is not, a lot of people might say alternative data doesn't work. I strongly disagree with that because we've experienced otherwise. But I also know from our experience that there are alternative data sets that are publicly available that are expensive, but not absurdly so that if you're clever, you have the right insight, you think about this

causal relationship that you're actually trying to model, you can actually make good money on without sort of spending millions and millions of dollars every year. I had the someone, I don't know if it was on a podcaster, but he was telling me that the hedge funds were right. They were photographing using satellite imagery to see how much oil was in storage. And then BP like covered the oil storage. And then they started using infrared to measure how much. And then BP painted it with infrared resistant paint. So it's like, I don't know why BP was fighting against it.

But right. It's like a never ending battle of like, OK, once we it's there for don't know why bp was fighting against it but right it's like a never-ending battle of like okay once we it's there for an instant and now it changed and it's something else yeah and that's what look that's also what keeps these things interesting i think people underappreciate the alternative data sets there is this real challenge of getting it and making it consistent and clean it's not like major data, which shows up on a regular basis, and you don't experience outages, knock on wood, most of the time, minus this S&P 500 index being down a couple weeks ago, if people were following that for 30 seconds. And yeah, it leads to that. One of the challenges with commodities, because you alluded to that, is a lot of the commodities data sets are not collected on a point-in-time basis because the major commodities companies don't need it on this

historical time series point-in-time basis. And so the effort we also do is just collecting these data sets for a couple of years, constructing our own proprietary databases that are now perfectly point-in-time. And that's the moat for us, that combination of external stuff that we've gotten, but also the internal stuff that we've spent a lot of time and a lot of man hours carefully collecting, cleaning, and utilizing in interesting ways. We want to expand on how you're using the AI, if you like that term. Sure, sure. I like the term AI.

I just like to always clarify that AI is the broad term referring to the building of a system that can mimic or exceed human behavior. When finance, when people are really talking about AI, they're saying machine learning, which are models built using large data sets. The one that most people are familiar with these days is ChatGPT, which is a natural language processing model that uses a large amount of text data in order to come up with interesting general responses to you typing in who's the best basketball player of all time, for example. Michael Jones. On a firm approach, one of the things, you and I are in agreement on that answer.

I'm sitting here in Chicago. If I didn't say Michael Jordan, my house would explode. I'm originally from Chicago. I would 100% Michael Jordan all day. Yes, I love it. Die Hard Last Dance documentary watcher. I watch it once every month. Wow, I love it.

But to answer your real question with the application of machine learning, one of the things that we've seen over the years internally is that our comfort and expertise in applying it across the domain space has increased over time. So using machine learning today is table stakes. If you're working with text data, there's no way to work with text data and not use an NLP model.

These can be models that you've either built yourself internally or taken off the very high quality open source environment that exists, where you can take a very high caliber, large language model off Hugging Face and utilize it and start building something very meaningful. In the initial days, we've seen an ability for these types of techniques to combine a bunch of forecasts in interesting ways. Increasingly, we're finding more and more use cases for them to make

the forecast in the first place and also on the portfolio construction optimization side. And that's been very interesting and sort of exciting for us. The challenge has always been is that there's a lot of noise in this data. And a lot of the ML techniques aren't naturally made for that. So the example I often like to give is ML techniques are often built on data sets where it's very obvious that the thing you're trying to predict is it's an image of a cat. Whereas in finance, there's so much noise, even in setting up the target level.

Did the S&P 500 move 2% because there was some genuine expected return difference or it was just variance due to external noise? So there's both noise in the thing you're trying to predict and your predictors that you constrain. And so being really scientific, being really statistically focused on your research methodology is what you need to deploy these methods at scale and do so successfully.

And it's been a long journey for us to do so. And we're very happy with the progress we've made, and we'll continue to. But it seems like you guys don't necessarily hold yourself out as like an AI hedge fund and that kind of thing. It's just, and to your point, it's just table stake now. It's like what the best firms are using as part of their research process. Yeah, that's certainly the case, but there's, we do focus on it extensively because there's such a large variety of techniques that you can utilize that provides a lot of scope for you to be creative and interested. So for example, let's just take the neural networks, which is one subset. Within neural networks themselves, there's a wide variety of ways of architectures and

models one can use. It far exceeds anything else that a classical econometrics approach-based way would have done. And so we generally believe there's alpha in ML, not only because of what the techniques can do, but because the space of techniques is so large that I can go and find work that other people don't have. So for example, the paper that created the attention mechanism, which is the heart of the transformer that gets used in chat GPT and so on and so forth. I had presented a paper of myself in 2017 at an NLP conference, and there was a little bit of

buzz around that paper, but people weren't sort of fully aware of the implications of that. And so what's really cool about a lot of the ML literature that comes out is there's stuff being published that people don't realize that, oh, that's actually the next best thing. And for us as investors, we might find something useful there that we can use to potentially get that better forecast that leads to more alpha for our clients. With the ML. You're using ML to research ML papers. So my most expended idea is the following,

which is the LLMs are generative models. You can certainly use generative models to generate text that another human can understand, but you can also encode them with an interesting signal grammar so that you can get them to generate sensible signal ideas expressed as value time series minus moving average over X look back. Time series minus divided by standard deviation. And so what these tools help you do is they help you sort of speed up the process of doing the core research that we do. That's how we really look at it.

In addition to just using them throughout our strategy. Right. It's like you have a room of 100,000 smart people and 1,000 managers managing those smart people. Yeah, the way I also think about it is, today, anybody at the firm, if you told them that we were going to take away access to the LLMs that we provide them, there would be a big hue and cry

because everybody uses it on a daily basis, whether you're writing emails or you're writing code or you're researching signals or you're trying to synthesize the literature that's out there to identify what ideas may be promising. All of us use it constantly. Is it all in-house? Do you have worries of using ChatGPT as a crude example of,

like, help me write this model, and some of your IP is leaking out there into the universe? Yeah, I think there's certainly some of that. We do have some good compliance constraints around what we can and cannot do, particularly with respect to code automation LLMs. One of the things I alluded to earlier, which font firms are in a better position for is because there's a lot of open source, large language models available, one can potentially build your own using off the shelf and make them internally complete. But these are all nascent efforts. It's not something that is sort of systematized yet, but that's the direction we're seeing this space go to deal

with those IP concerns. Shifting gears a sec. We were talking about, it seems to me a lot of the signals you were talking about is more not contrarian necessarily, but counter-trend, right? Like it's, hey, this is out of whack. We're going to buy it expecting it to come back. But that seemed to be against the positive convexity, right? So it seems

like you're taking some, what I would call like negative skew and even classical long short equity would be kind of negative skew, right? Like it's generating small consistent returns and risking some big blowups. So how do you, right? They seem contrary to each other. Like you're buying dips, selling rallies, but it's also positive convexity. How do those two things happen at the same time? Yeah. So I think the way we're doing that is pressing these views across multiple time horizons.

I think that's one of the key things. And the other thing we've also seen is that the dispersion that we see during these times of equity market stress are less akin to the sort of stuff that you see in the long, short equity approach, where it's more an idiosyncratic driven. So don't forget, because we're dealing at the country level, there's this big macro latent factor that is moving things in a common space. That's what really prevents this negative skew. So for example, in March 2020, the alpha is not coming from this buy the dip sort of situation. It's really saying, okay, everything is going to go down. Can we figure out which ones are going to go down more in relation to the others as a function of price movement, economic composition, stock movement, valuation of the companies within the index? And that's where this alpha is coming from, as opposed to the long short equity, which is betting on some idiosyncratic thing happening, which during times of stress,

everything goes, it's the fan because those relationships no longer hold. It was fair to say Italy is not going to zero, which maybe things like that of like, it's at the index level, you have some protection. Yes. And they're not going to move wildly away from the S&P 500. And if the S&P 500 moves, it's not going to move wildly from global equity markets. But the flip side of that would be you might have to add some leverage to get the return you want if they're very even. But we'll leave that for the time being. Duane, you want to save us from this quant talk and bring us back to the pertinent points? You want to close this out?

No, I think we've gone down the point because it's relevant what we're doing. So I think when we talk about Versa, a big part of our genesis story has been our use of data in multiple different ways. So alternative data increasingly is an area of focus for us. When we engage with our clients, we're really thoughtful in how we try to engage with our clients, the idea behind the products that we've launched since inception until today. Everything we do will focus on quantitative

investment strategies. But it's really creating strategies, being thoughtful around the problems our clients are explicitly looking to solve. As it relates to get strategy, the challenge as we look across equity markets and we look across our client portfolios and investing in general is that we think and our clients seem agree, that the next 10 years are likely to be a little bit more challenging than the last 10 years. And in that sort of environment, making sure that one has strategies that are able to help diversify not only traditional risk factors that one might have in their portfolio, but also some of the alternative risk factors. We talked about trend and macro, all the managed future strategies. We think that those offer a fair amount of diversification, but they tend to be correlated with one another.

The ideas behind the things that we're exploring and the get strategies are meant to complement that, and I think we've done that reasonably well. Do you know off the top of your head, or do you guys have a target of what that correlation looks like? Right to trend, I guess? Yeah, it's around. around yeah dwayne go

ahead you know yeah it's about point point two so the correlation with trend is point two the correlation systematic macro implementation is literally zero got it and with equities it's probably close to zero yes maybe slightly negative all right so that's the key right there hey this is alternative piece it's not trend it's not long short equity something different um add it in there call it a day but i'm curious you said they're thinking the next 10 years will be worse than the past 10 the past 10 years was no picnic but do they have any insight into that or just like it's going to be, why is it going to be more difficult? Slightly more challenging in an environment where interest rates are hovering around zero and inflation is not at a contained weight. So I think there's lots more uncertainty,

right? The last thing you're certainly had a fair amount of uncertainty, but there was this constant around rates and this constant around kind of general inflation that I think it's gone away. Not completely gone away, but changed quite a bit. Outside the box question, do you guys ever think of peeling out the alternative data as like a separate product? I've always thought that would be a billion dollar business, right? If you're just leasing out all the alternative data sources to all the hedge funds it seems like all of them are doing their own work to gather and and clean and do all this where if you had one source that could get it all we'll partner on next time yeah i'll say we're not necessarily in the business of that i think i know

yeah the alpha is in the difficulty of putting that all together. Yeah, just when I look and talk to everyone, I'm like, it seems such a waste of time and resources that all these hundreds of groups are doing the same thing and spending all this money to get that done. But to your point, it's the same as why is everyone paying the CME

for market data and doing all this stuff? Because it's in the way you use it, not necessarily in how you get it. I think that's right. Awesome. Any last thoughts? We'll put the links to some of these white papers you guys mentioned in the show notes.

Also, you mentioned the foreign relay race. I think that was Bastion Balesta of Deepfield. We'll put a link to that pod in the show notes notes anything else you guys want to link to or let people know the website no we'll send a website we um i mentioned we're celebrating our 10 year so in the research we actually put together a book that's available in pdf format of uh kind of 10 research papers over the last 10 years that we think are most relevant to some of the things that our clients are facing. So range of costs,

managed futures, risk premium, value investing, merge arbitrage. I think folks will find that somewhat interesting. I'll go read it. What are the big 10-year

celebration plans that already happened? Is there a party? No, it comes up in the fall. So we're in the process of planning the official 10 years in October. So we'll sort that out in the next couple of months here.

That's just right to be a hedge fund in the last 10 years. How many hedge funds launch and make it 10 years? Probably 10%, if that? A few, I would suspect. When I first started, Nishant doesn't have any grays, but I certainly didn't have any gray in my beard or in my hair. They're coming. You just can't see them yet.

He's got the opposite problem. He's missing some up top there. My daughter says, she's like, oh, your blonde hairs are coming out when my beard gets a little long. I'm like, yes, we'll call them blonde. Well, thanks, guys. We'll look you up next time I'm in New York

or when you come back to Chicago at all and grab a... I should be there this week. I'll shoot you a line. I'm actually going to be there on Thursday. If you're around, I'll shoot you a line. Awesome. Appreciate it. Thanks, guys.

Okay, that's it for the pod. Thanks to Dwayne. Thanks to Nishant. Thanks to Nishant. Thanks to RCM for sponsoring and Jeff Berger for producing. We'll be back next week with some option trading folks from across the pond in dear old London. Cheerio. You've been listening to The Derivative. Links from this episode will be in the episode description of this channel.

Follow us on Twitter at RCM Alts and visit our website to read our blog or subscribe to our newsletter at rcmalts.com. If you liked our show, introduce a friend and show them how to subscribe. And be sure to leave comments. We'd love to hear from you. This podcast is provided for informational purposes only and should not be relied upon as legal, business, investment, or tax advice. All opinions expressed by podcast participants are solely their own opinions and do not necessarily reflect the opinions of RCM Alternatives, their affiliates, or companies featured. Due to industry regulations, participants on this podcast are instructed not to make specific trade recommendations nor reference past or potential profits. And listeners are reminded

that managed futures, commodity trading, and other alternative investments are complex and carry a risk of substantial losses. As such, they are not suitable for all investors.