2024's Biggest Investment & Trend Following Insights (Part 1)

[00:00:01] Intro

You’re about to join Niels Kaastrup-Larsen on a raw and honest journey into the world of systematic investing and learn about the most dependable and consistent, yet often overlooked investment strategy. Welcome to the Systematic Investor Series.

[00:00:22] Niels

Welcome and welcome back to this week’s edition of the Systematic Investor Series with Katy Kaminski, Cem Karsan, Rob Carver, Mark Rzepczinski, Richard Brennan, Alan Dunne, Nick Baltas, Andrew Beer and I, Niels Kaastrup-Larsen. And as you can tell from this introduction, today and next week will be very special episodes because it’s that time of the year where all nine of us get together for one big conversation and debate.

So firstly, let me start by thanking all of you for making the time for this extended recording today, which I really have been looking forward to. And of course, for all the time and energy you have put into making the weekly episodes that we have published this year. It means a lot to me, and based on the feedback we get, I know it means a lot to our community.

We are recording today on December 4th and this conversation will be split into two parts and published on December 21st and December 28th. We’ve got a great lineup of topics that we shared with each other beforehand, and to mention just some of the themes that we will cover today and next week, in part two, here are some of the headlines.

• The biggest investment mistakes of 2024
• The US dollar under the new Trump administration
• The efficacy of replicating trend following returns
• The meme bubble
• Managed Futures Beta does it exist?
• The return dispersion between CTAs this year
• The role of social media in raising AUM
• Are systematic strategies becoming overcrowded? and
• Are markets becoming less efficient?
So, as you can hear, we have a really packed agenda. So, let’s dive straight into it.

The format of the conversation today will be that each of us will select a topic and then it’ll be open to comment by the rest of the group. But generally, we are going to keep it a little bit fluid and see how we get along.

And since it’s polite to let the ladies go first, why don’t you, Katy, share your first topic and feel free to direct it to who you would like to have some comments from, and then we can discuss it afterwards. And if someone else has some strong views, please feel free to jump in. Otherwise, I may prompt one or two of you. So, Katy, over to you.

[00:02:34] Katy

So, I thought, you know, the best way to start off is to start thinking about what happened in 2024 and sort of some of the interesting themes that we’ve seen in trend following. I think, for me, we usually review the CTA style factors that we design at the end of the year, to kind of give a sense of sort of what themes within trend following worked well.

And when I think about 2024, I really think about it as a year where the market moves, particularly in fixed income and currencies, have been driven by expectations for monetary policy. And trend flowing is a strategy that does really well when we have trends.

And honestly the only big trend we had this year was equities and a couple esoteric commodities. So, when I looked at these CTA style factors, what we see, which is really fascinating is that big and slow wins the day. So, what do I mean by that?

It’s that the slower trend factor, so being slow, our slow trend factor is actually up 12%, and a 10 vol fast factor is down 10% or 11%. And that’s a big divide. So, it means anybody who listened to short-term noise, at all, got head faked. So that is one big issue is that this big divergence in speed of which trends worked.

Secondly, large markets, bigger, larger markets like you know, S&P, etc. actually worked quite a lot better for trend this year. So, our size factor, you know, also suggested larger markets did better.

Another area which is interesting is that correlation was actually negative, so, negative correlation factor. But the really big story is big and slow did well. Anything faster didn’t do as well.

And so that sort of is consistent with the macro environment that you see right now, which is that monetary policy expectations drove a lot of noise. Also, the US election, which suggests that, you know, the big trend was equities, and depending on what you selected, that would have changed.

So, I think I might turn this to someone like Nick, who’s done a lot of work on factors, and sort of see like what he’s seen and if he kind of agrees with me on that, and other factors he’s looking at to kind of distill 2024 in a nutshell for trends.

[00:04:59] Nick

Yeah, thanks. Thanks, Katy. So, I mean you made a number of interesting points. I would maybe start with the August event, you know, at least as far as I stand here, this is one of the big debates that I have had over the last three, four months with some of our clients and investors.

And the way that I look into this V shape was that it was so quick that in isolation you could do two things. Being super slow, navigate through, happy days, you recover by month end, or being extremely fast and managed to capture a bit of the downturn as well as a bit of the recovery. And there is this bitter spot in between whereby you’re a bit fast but not too fast, you’re a bit slow but not too slow, that by the time you’re short, the market recovers and you basically get a double down.

And the way that I’m looking into that in isolation, as an event, is that I use this I guess analogy where back in the college days, when do physics and you have the soldiers that basically synchronize as they cross the bridge and then the bridge is falling. It’s pretty much this bitter spot that made trend followers in that event underperform, at least the medium-term.

And by being slow and by being just passive on it, you’d navigate through. Now do we learn anything out of it? Personally, nothing, in the sense that we know exactly how we’d operate in this environment. And I don’t think we should just go back and rejig the models as to how we would run the programs.

But it suffices to say that there are very few trends that have done the job year to date. But beyond that it was like a year of whipsaws. And I’m pretty sure Andrew will have some good views on this one.

Beyond that, I would say reversion dynamics played out quite well, and I’m talking about cross sectional reversion dynamics. You know, various people here, I’m sure Rob has been following skewness dynamics that have helped quite substantially in various parts of the year, both in commodities as well as in the equity markets, specifically around August.

And in the early part of the year, volatility selling kept on being a good rewarded exposure, less so the second half of the year. Maybe Cem has a good number of views. But I would say, overall in the systematic community, it hasn’t been the stellar year but it wasn’t a bad year either. It was not that great for trend if I were to be pretty honest, but in a multistrat diversified context it was a relatively okay year.

The last thing I would say is that the appetite has been and remains pro carry, maybe a bit more conservative following August. I mean, obviously following the November, and in elections, the story was very short-term hype, and so on, and so forth. There are more things to discuss later on about single stock equity momentum. But maybe I pause here.

So, I know I called on a number of people here. I mentioned Andrew, I mentioned Rob, I mentioned Cem. Guys, whoever wants to, just jump in. That’s my view.

[00:08:02] Andrew

Sure, well, as you know, replication tends to be a bit slower than the overall industry because we don’t have stop losses, and vol controls, and other things that kick into it. So, if you think about something that happened with whether the economic data, or Powell, or the bank of Japan, you know, what happens is we’re looking at a recent window. If that happens in the middle of the week, by the time we’re rebalancing, we’ve got a few fresh data points, but a bunch of older data points from the prior regime. So as this space is de-risking, we’ll de-risk a little bit slower.

And it happened exactly as both Katy and Nick described, which was, you know, when you hit that inflection point, we underperformed in the initial period, but the act of not de-risking as fast as, I think, some of the funds who have more of these faster models built into what they were doing, by the end of the month we’d recovered.

And then again, to follow on, on Katy’s point, you know, and I think that the language that I use is a little bit different because we’re not doing the same kinds of depth of research in terms of individual factors within it. It just felt to me that simple worked a lot better this year, that it was that the major markets tended to be more stable.

And I was talking to somebody who’s at one of the large managed interest funds, yesterday, and they said basically that this to them is an incredibly unusual year in that it’s been like taking two steps forward and two and a half steps back across a ton of different trades. And so, I’d be really interested to hear, as the conversation goes on, what people’s experiences have been around that.

And, I’ll pass the baton.

[00:09:39:00] Cem

So, I think I’m going to hop in here and actually give a little background. I like to talk about the why when we look at trend following. A lot of that, I think, sits here in the volatility space where I kind of tend to float around and live in. We have had two massive factors, I think, that have led to that mean reversion and negative correlation that was kind of referenced.

One, it was an election year and there was massive liquidity pumped into the system, as is often the case during election years. That’s very vol suppressing, supportive for markets, obviously. That emboldened the second factor, which is vol selling. There’s been a massive amount of vol selling at the index level in the S&P 500 throughout the year.

Another important factor there is the incredible adoption of structured products. We, in the last two years of structured product issuance have gone from approximately $500 billion globally to $1 trillion per year of structured product issuance. So, that is dramatic. You pair that with ETFs tied to yield going from, in the last two years, $25 billion to about $200 billion.

There’s a significant, significant uptick in volatility selling across the board. That is pinning the index. That is why, even in August when we had a massive vol blow-up, it reverted so quickly. That vol blow-up really came from overseas primarily. The vol was incredibly well supplied and the quick reversion was really tied to that. So even when we did get the vol blow-up, it reverted very quickly because in the US, index vol supply was incredibly well supplied.

Again, we reviewed this before, but that also leads to negative correlation. By definition, if the index is well supplied and pinned, and idiosyncratic risk still exists, things have to start moving away from each other by definition. So, that’s what’s really been driving the low correlation as well.

So, I think that’s what’s been contributing to these factors in this kind of environment that we’ve been in. Now, the question is, how much of that will continue next year? The structure product issuance, I believe, is a secular trend. That is not going away. So, there is some level of index supply that is continually increasing. I think that’s what speaks to why, if you look at a chart, correlation has been secularly increasing really since 2017 (the correlation breakdown, I mean), and so, I would expect more of that.

That said, in terms of liquidity, which is the other factor, I find that will be much less reliable given some of the other macro factors we’ve been talking about. But I’m sure we’ll get to that later in the show.

[00:12:44:00] Niels

Yeah, that’s great. So, let me jump in here because I want to keep it to about sort of eight, nine minutes per topic. So, I think this was a great start. And since Cem, you’re kind of the next one that I wanted to bring up one of your topics. I’ll leave it to you to bring it up and, and we’ll see who’ll jump in and respond or comment on it.

[00:13:07] Cem

Yeah, I’d love to kind of go for a little more of a macro tilt with this one. I’m really thinking, and don’t want this to be construed as political in any way whatsoever, but I think we would be remiss if we didn’t at least cover the new coming president, and the election. I think there is a geopolitical question out there.

Last time we had Trump in the White House, there was a generally very pro Russia stance. that seems to be continuing based on the appointments we’ve seen; at least open to working with Russia and less anti Russia, I guess I would say, here in the US.

The last time it was very, let’s say, anti-China. There was a very strong kind of line held with China throughout the administration. Since Trump has left, we have seen a joining of forces between Russia and China; bilateral documents have been put out in 2022. It’s very clear the coordination between those two countries.

I think one of the biggest geopolitical questions, biggest questions about where things are going, lies in how Trump will deal with China. Does his tone on China change? Yes, there’s still a strong line with tariffs, et cetera, rhetorically, but will the actual final push be more strong against China?

And I think that’s a really big question geopolitically. If the number one and number two economies can get over that divide, that will change a lot of things globally. I’m curious to hear people’s thoughts.

[00:15:01] Mark

Well, I’d like to ask a simple question because I enjoy, always, the discussion on geopolitics with friends and other professionals. But as trend followers and as quants should we really care? So, how do you incorporate geopolitics in a trend following model?

And in some sense, if I look at my models, if someone is pro Russia or anti Russia, does it really matter if I’m a trend follower? So, there’s what I do as a profession, and what I discuss as, we’ll say, a market observer, and those are very different. How do you square those two alternatives?

[00:15:54] Niels

Alan, do you have any thoughts on this in terms of the global macro side? You obviously do a lot of the global macro conversations on the podcast.

[00:16:01] Alan

Yeah, I guess that the only observation is, you know, you’ve got different personalities at play. As the administration comes in, you’ve got Trump himself, Scott Bessent as Treasury Secretary, Lighthizer is trade.

You know, they all seem to have different noises, particularly around trade and FX, which also influences China more generally. So, you know, Lighthizer very much in favor of getting the deficit down in favor of A weaker dollar, which was Trump’s position historically.

But then as Scott Bessent has the job as Treasury Secretary, he’s much more market friendly, and his job is obviously to make sure the deficit can be funded, which arguably requires a stable dollar. So, I don’t know. What the answer is, I don’t know.

But I do think a lot of this is going to depend on the interplay with those three personalities and who becomes the more dominant voice. I think the way the market is trading more recently is that there’s more optimism around the deficit with the appointment of Bessent. Yields had been rising and have steadied since that appointment. So, I think that the hope in the markets is he’ll be a steadying perspective with respect to tariffs and the dollar and deficit reduction. And that’s probably the narrative the market is going with for the moment.

[00:17:21] Niels

Nick, I’m going to turn it over to you to raise one of your topics, one of your questions, and if you want to direct it to some of the people on the call.

[00:17:31] Nick

Okay, yes, thank you. Thank you, Niels. So, look, the point that I want to bring here is maybe a bit broader than trend following, but I think it’s very relevant for the broader business, both systematic as well as more discretionary, and that has to do with crowding. I know we’ve discussed this topic and people keep on discussing it every other year that you see either some sudden reversion or some massive outperformance. So let me share a few anecdotes.

One of the best performing strategies year-to-date, that’s outside of trend, is single stock momentum. And I’m talking about, now, a relative value trade between winners and losers in the single names. This is not the market beta. This is not equity outperformance that we have seen in the trend following camp. This is purely the best that keeps on outperforming the worst, in a relative sense.

Just to put things in context, I was looking at some numbers. You can look into single stock momentum in a variety of ways. I’m looking into one implementation. It doesn’t really matter. For the last 10 years, just to put it in context, I have seen calendar year returns between -7% and plus 17%. Okay, this year it’s 44%. This is an order of magnitude greater. And it has been as such throughout the year.

And I remember speaking to some hedge fund clients back in Q1, Q2, and the big debate was across the following lines. How much of that is concentration? How much of that Is crowding positions? How much of that is going to be unwound? When the unwind is going to come? What are the catalysts there? How is the gross leverage in the system? How do hedge funds get positioned?

But calling the shots is extremely impossible. And we know the story and we’re not going to go around, you know, why is it hard for crowding to be predicted when it’s going to blow up?

So that’s point number one whereby great performance came on the back end of the mega caps outperforming, and the tech frenzy, and the AI, and whatever you want to call it. That’s statement number one, whereby crowding is there, has been there, and has delivered performance year-to-date and it’s still performing. That’s statement number one.

Statement number two, you can look into maybe the space that I’m much more involved in, which is the quantitative investment strategies, that is, again, broader than trend following, that’s a QI that comes out of banks.

There were a number of reports coming over the summer. There was one article in Bloomberg that was discussed quite substantially talking about, oh, there’s mass utilization of those strategies and maybe there is some sort of concentration risk coming along. Which, by the way, I definitely don’t see the data. I don’t see the growth necessarily, even if the growth has been there. But we can debate upon it.

Step number three, I’m going to August now, the reversion came, oh, it’s CTA’s, oh, it’s vol scaling, oh, it’s parity.

There was another FT article that followed after that. It was talking about oil volatility being elevated because of CTAs. And at the same time, you look into some of those active ETFs on the option side, you know, there’s so much proliferation of call overriding in those active ETFs and obviously, in terms of assets, they have grown substantially.

So really my question to the group is the following. Are those concerns legit or overstated? Is it really that we come up with a concern for the sake of coming up with a concern? Because it’s easy to point the finger. And I know we’ve discussed it so many times, but this year, every other month I keep hearing it in a different context.

[00:20:58] Niels

Maybe Katy would kick it off.

[00:21:02] Katy

Yeah, I love that because I love some of your papers on co-movement, and actually the co-movement factor that we look at for trend was not really a big factor this year. It was really a speed factor that was the biggest differentiator. And so, I think it’s also an interesting point that you bring up the active ETFs.

I mean, I think that’s been one of the biggest trends that we’ve seen. We wrote an interesting piece about managed futures ETFs, but that is only just the tip of the iceberg of how much active ETF volumes have gone up by orders of magnitude.

So, I think I actually agree with Nick that there’s a lot of finger pointing whenever something goes wrong. There has been a lot of coordinated positioning, but the coordinated positioning has worked and has continued to work. So, if you use it, use the crowding measure and you reduce your positioning, it would have been a very negative experience for you this year.

So, I think I have to agree with Nick that it’s not that clear to answer that question. And that sort of concentrated positioning, it’s still a question that’s not really resolved, considering what has really worked. , I think maybe turning it also to the vol now we can turn it to Cem for the vol question. So, I haven’t seen any indications on my side either. That’s what I’d say.

[00:22:24] Cem

Yeah, much like we saw this summer, you know, you had the slowest period and the greatest amount of vol compression May, June, July, which led to the bigger expansion. So, the more vol compression you have, I think 2017 was the lowest realized and implied volatility in 120 years of history 30% because of the amount of vol selling that was happening on the S&P 500, which led to the Volapocalypse. That’s how this works.

The more vol compression we have, the more concentration we get, the bigger the reaction ultimately is. That that’s how it works. I would say positioning, particularly of dealers that we watch, is very correlated to trend. So, the more you have a trend and things work, the more people pile in and the more concentration happens. The more that happens, the more likely you are to get potential energy building – a bigger blow-up. So, I’ll kind of leave that there.

We are in a period of significant vol compression happening currently. It really started post-election. It will probably go through January. I would be hesitant after January 20th. It also correlates with the new presidential cycle and new things coming up. So, something to be thoughtful of.

[00:24:00] Niels

Good stuff. All right, well, let’s move on to something maybe a little bit different. Andrew, which of your Topics do you want to start out with?

[00:24:10] Andrew

So, we’ve talked about the fact that it’s sort of a year where major markets have worked better and where slow moving has worked better. One of the questions that I’ve had, so, when you look at the industry structure. As you’ve had things like replication, QIS, trend following products, kind of simpler ETFs and other things have come in, there’s been this idea that, okay, you’ve got this kind of core exposure to the space, like a beta to the space (and I don’t even think we need to go that much into even how to define the beta). But the implication of it is that if you can get something like that in a reasonably efficient way, and you have the fee budget, why not allocate your dollars to people who are doing really, really, really esoteric things?

And so, one of the things that I’ve just seen from afar is that if you look at a hedge fund like Florin Court, who has to put up astonishing numbers since it came out of AHL, basically focusing on alternative markets, and Man AHL’s Evolution fund (which obviously invests in a ton of different markets), these guys have put up unbelievable risk adjusted returns. And I guess I’d be very interested in the group because 20 years ago I started a commodity-based firm where they found the best opportunities in the semi-liquid markets. Like it was the markets where you were in essence coming in as a liquidity provider, but often it wasn’t about trend following, it was about mean reversion, and information asymmetries, and non-economic players who were driving prices like in the power markets. But I’d love to hear from the group in terms of when you look outside the major markets, do you see strong evidence that they trend more?

Do they go from 1 to 20 as opposed to going from 1 to 6? And is that more predictable and implementable subject to all the trading costs?

Because it turned out, at least anecdotally, it looks like it’s been a very, very tough year for some of these guys and their strategies where they’re suffering historic losses. So, I’m here to learn from the group.

[00:26:11] Rob

Can I jump in?

[00:26:13] Niels

Absolutely.

[00:26:13] Rob

I think at one point I was managing 60% of Evo, so I feel like I’ve got some… I’m also good friends with the guys at Florin Court as well.

So, I’ve kind of looked at this in the past. So, it’s sort of obvious why a less liquid asset would be more profitable in a mean reversion strategy. Because if you just think about mean reversion as kind of like market making at a slower time frame, then if you’ve got anything with kind of bigger spreads, it is going to be kind of wandering around. We’re not sure exactly where the kind of true value is, so it wanders up and down.

And if you have mean reversion around that kind of true value, it could be quite a profitable thing to do.

It’s less obvious to me what the drivers would be for why slow trend following would make money in less liquid markets. We kind of have to get into stuff about oh, it takes longer for investors to kind of digest information because there’s fewer of them looking at the asset and they’re just slower moving. So, if you believe the theory of trend following is that it’s around people digesting information slowly, and asset prices adjusting more slowly, then that kind of makes sense. I’ve never really liked that theory.

If I actually look at the empirical evidence, I’ve never really actually seen any evidence that less liquid markets do trend more. What tends to happen is that sometimes you see studies that are, not deliberately cherry picked, but what people do, sometimes, they do this analysis and because the less liquid markets have often got less history, it just so happens that the less history coincides with the period when trend following generally was better.

It just looks like they’ve done better because you’re comparing like a 50 year track record for something like, I don’t know, treasury bond futures that have been around for a long time, with bad periods and good periods of trend following, with a shorter time period for something that’s done very well, but it’s just been a time period when trend following has been particularly good.

Once you take that effect out, I’ve never really seen any statistically significant effect that it’s less liquid. This is purely in the futures space. So, I’ve seen people do stuff in crypto. I’ve seen people say, well actually, in crypto it’s the reverse – like less liquid coins seem to trend worse. In stocks, I think the evidence, again, is mixed. But certainly in futures I’ve never personally seen this. I don’t really buy the kind of explanation as to why it makes sense. For me the benefit of alts is purely that you get more diversification. It’s as simple as that.

[00:28:31] Niels

Yeah. Anyone else want to venture into this? Rich, do you have any thoughts on alternative markets versus core markets?

[00:28:39] Rich

Yeah. So, I think some of the smaller players have done particularly well this year. And this is on the back of some of these idiosyncratic outliers, you know, cocoa, coffee, OJ, Bitcoin, some of the yen pairs like the pound/yen. So, I think that there are certain trend followers who are focusing in some of the sort of smaller markets, OJ for instance, that sort of thing. And they definitely have been showing these significant trends.

So, you know, at the moment I’m of the opinion that we’ve got this major decoupling going on financially around the world and I think that we are seeing the rise and rise of these idiosyncratic outliers. So, the dominant markets, I think, tend to be much more correlated than some of the alternative assets.

And I think, as Rob said, I think you do get diversification benefits in some of the alternative assets. There is less chance for them to be significantly correlated with each other. I think as assets mature, naturally there’s more participation in there, more players coming in there, more mean reversion, all of the different institutional players come in there.

I think they become far more noisy. I remember old Perry Kaufman had this proposition where he was suggesting that infant markets tend to show this prevalence towards trending conditions. As they mature, and more and more participants come in of different flavors, all of their impacts start creating far more noise in the market.

The market becomes naturally more efficient as time goes on with more participation going in there. I know that Florin Court sort of specialize in this very sort of very alternative markets, you know, China and all of these sort of things. But I’ve certainly seen it in some of the smaller liquid assets in the futures this year.

And, you know, I’ve seen this significant dispersion in the trend following group because of the likes of Mulvaney, I think, is pulling out probably around a 60% return this year if things are going according to plan. It’s about 60% now I think year-to-date.

I think some of the what I sort of refer to as classic trend following, I think they’ve done very, very well this year. And you know, I put this down to generally a significant regime shift that occurred from about 2020 onwards.

This change from the quantitative tightening to the quantitative easing or the quantitative easing to the quantitative tightening. I’ve just seen this big boost in all of these idiosyncratic outliers coming from everywhere. So yeah, that’s what I’d like to say about those alternative markets. I think they do offer opportunities but you’ve got to be very diversified to capture them.

[00:31:34] Niels

All right, good stuff. Let me just jump in here before we go to the next topic.

And that is to say I think what Andrew raised is very interesting and your response from you, Rich, in particular was also interesting because I’m not sure it actually covers completely what might be going on because one of the debates we’ve had is how many markets should you be trading?

And so, I think this year we actually have a situation where some of the smaller commodities, as Rich point out, have moved, but the ones who trade 300 or 400 or 500 markets actually their performance is not making any real benefit of that. And that’s simply in my view because one market doesn’t make a big impact anymore when you trade that many markets. Whilst the names that Rich was bringing up that have done well, actually those managers have been trading very few markets.

Let’s jump in with Alan and see what you have in store for us this year in your first round of topics.

[00:32:39] Alan

Thanks Niels. As you are not going to introduce any of your topics, I was going to steal one of yours, if that’s okay.

[00:32:45] Niels

That’s fine. Yeah, please do so.

[00:32:47] Alan

You referenced, are markets more or less efficient?

So, this year one of the more interesting research papers was from Cliff Asness who had a blog, kind of an elongated blog post, suggesting that markets had become less efficient over time. And I guess it very much is an equity perspective.

The evidence he presented was looking at valuation factors for different stocks and how high valued stocks had traded at a higher premium to lowly valued stocks for longer periods of time than you would have expected historically. And we saw this back in 2000, around that time, with the dot com bubble. But we’ve seen it again, now, in the last number of years. The valuation has come down a bit, this differential between value and I guess it’s kind of a value growth differential.

But his interpretation of it is that, due to various factors: indexing, the low rate environment which you previously had, and possibly the impact of social media blunting the wisdom of the crowd (the fact that everybody’s responding to the similar factors now and people are responding to independent factors), that the market has become less efficient over time.

And the implication for quant strategies is that you may go through longer periods of drawdowns but ultimately there should be good opportunities because of these mispricings, but longer periods of drawdown as you suffer that the market, I guess not adjusting the valuations. It is very much an equity focused view.

So, I wanted to ask people, one, does that idea of markets being less efficient, does that resonate at all? Is it just a narrative or do they think there’s something in it? And then secondly, more from a futures and FX perspective, is there any evidence that we’re seeing similar dynamics there?

So maybe Mark might be a good person to go to in the first instance, but anybody else happy to dive in?

[00:34:50] Mark

Well, I think the most important thing to remember, when we talk about market efficiency, it’s always a joint hypothesis. It’s a hypothesis of whether the markets are efficient and the model you use to determine whether there’s efficiencies. So, it’s easy to do this in equities because you could sort of benchmark against valuation. So, tell me what the valuation for corn, or soybeans, or cocoa, is and then I could tell you whether it’s going to be market efficient or inefficient.

Unfortunately, we don’t have good valuation for non-equity markets. You might sort of say we’re better at, you know, fixed income. But generally, go back to the old argument of Samuelson, that we could be micro efficient but we’re macro inefficient. And for a lot of asset classes, the reason why trend following is effective is because we don’t have good valuations. So, oftentimes you have to use price as your measure of where valuation is. So, I don’t know if markets are any more or less efficient relative to the past when you get beyond equities.

[00:36:00] Niels

Andrew.

[00:36:02] Andrew

So, I think one of the issues that I often have with academic research, and the reason I actually didn’t go do a PhD when I had the opportunity to do this 30 years ago, was that I think there’s often a disconnect between just analyzing the topline data in terms of the differences between growth and value and other things, and a more nuanced understanding of actually what’s happening underneath the hood of these companies. And I think it’s difficult to…

So, when I hear these long-term studies, to me I think about how different it is, how different the equity markets are today than they were 30 years ago. You know, back then, in the early 1990s, you couldn’t get smart people to go into manufacturing. Now the very, very smartest people are working for tech companies. I mean, they’re smarter than the people on Wall Street.

You know, the idea of conglomerates was a terrible idea because corporations were terrible at making capital allocation decisions. Now you have these leading companies, part of what’s driving them is that they have taken over the capital allocation process. They’re just better at it than stock pickers. So, Amazon is a retailer. Then they jump into the cloud business. Now they’re jumping into AI. This is what these companies are doing.

And so, I think that’s always the challenge when you’re trying to draw these very, very long-term conclusions based upon just sort of the topline data. I don’t have a view on whether the markets have become more efficient or not.

I do think a lot of the allocation process that I saw in the early 1990s, where you could look at an industry, and you’d have a bad company that was bad 10 years ago and was still bad today and was priced less than everything else, and you could have good companies. It was just a much more slow-moving process where stock pickers could… There were more opportunities to kind of buy the good and short the bad, or buy the bad and short the good, depending upon what you thought.

Whereas, when I look at, for instance, the average activist target today, they’re generally pretty great companies, maybe with a not terrifically wonderful structure, but you’re not talking about the golden toilets, the extra fleet of planes that they don’t need. It’s not the low hanging fruit anymore. So, from a capital allocation and stock investment perspective, I think that side of it has in fact become more efficient.

[00:38:20] Cem

I’d love to hop in here and add a couple two cents. I think those are interesting points for sure. When I think about markets, I think this idea of efficiency is about fundaments. That’s the weighing issue, that’s the long term where should things in theory be valued? But the voting machine, the short-term supply and demand in markets I think often takes us away from that. Like that’s the big idea.

The question is why is the voting machine, why is supply and demand becoming more frequently out of line with these fundamental valuations? My simple answer would be the inputs to supply and demand, those flows are increasingly uncorrelated, unrelated to fundamentals.

I think there are big structural flows that happen, these days, that are a function of both central bank liquidity, treasury liquidity, global moves, options and volatility, structured product issuance, a lot of things that really, really affects the day to day, week to week, month to month movement and can trend these corners of the market. Add in the Wall street bets in a kind of meme crowd, and just simply animal spirits having probably a bigger effect, more retail elements in there, passive investing. I mean, there are so many flows that have very little to do with fundamentals these days.

And so, it makes sense to me that we could really get further and further away and be more “inefficient”. There was a time when the majority of the flows were people who were making fundamental bets. That’s not the case anymore. That’s not the majority of supply and demand. So, I think it makes sense.

[00:40:28] Niels

So maybe we’ll just jump into the next round, which is. Rob, what’s your first topic of this year?

[00:40:38] Rob

I’ve put something on the kind of initial discussion we had around artificial intelligence. I noticed there’s a question from a listener about artificial intelligence as well. So, I’m going to wrap those two together.

So, my first one is quite simple, which is does anyone else actually use AI in their kind of job in any serious way? Personally, for me, the answer has been no. I’ve not found a serious use for it. But I’m curious as to know whether I’m, now, whether I really am just a grumpy old man and an outlier. And you guys are all using AI in very sophisticated ways, which would be lovely to hear.

The second question from the listener, which you can maybe also address, is whether (and it kind of links back to this idea of market efficiency), you know, if we open up with just AI, and sort of everything being done in a very sophisticated AI way, then effectively markets will become perfectly efficient and there’ll be no need for any of us to exist. So, is AI at the point where we can… Is anyone actually using it? And in the future, is it just going to completely drive us all out of business? Those are my two questions.

[00:41:40] Niels

I know that Rich uses AI, but probably not for trading. So maybe we’ll see if there’s anyone who uses it in their trading or system development or anything like that.

Otherwise, we can obviously get a, a quick comment from Rich how he has incorporated AI in his world. Anyone who wants to comment on this? Mark?

[00:42:02] Mark

Well, I think that’s the large challenge. And I guess I’d sort of say that one of the questions I had was, what was the most interesting piece of research or what was the more important thing you learned? So, I’m learning a lot about machine learning as opposed to making a distinction between that and AI. But it is really hard to translate that into better models.

So, there are a lot of techniques. There are high barriers to entry because you have to learn a lot of math and statistics to be able to get to the point. Then you sort of generate the models. You look at what the impact has relative to what your existing simple models may be or alternative models, and you find out that it might be sometimes marginally better, sometimes it does marginally worse. So, I guess the jury is still out on exactly how you can use machine learning to develop better models than what you might already have.

And this is a constant battle, I will sort of say 20 years ago, you’re constantly looking for, is there a better way to do trend following? And you use it against your existing benchmark and you find out that the existing benchmarks, which are oftentimes very simple models, are fairly effective. It’s hard to beat what you may have been using for a long period of time.

[00:43:30] Niels

Sounds good. Anyone else wants to jump in here, Cem?

[00:43:32] Cem

Yeah, I think the way we tend to use it is not for intelligence as the name implies, but really as workers. It’s a way to go aggregate large sums of data and maybe do very basic things with it, prepare, get wireframe code in very basic ways. Again, it is a way to replace the low-end work, not a way to solve big quantitative problems at the end of the day. But that can make you more efficient and lower your costs. And I think that’s how we tend to use it.

[00:44:19] Andrew

So, we don’t use it, which makes me feel old and curmudgeonly as well, other than basically trying to find out if I can find any cool information on anybody I’m about to go meet. But basically, so I’ll just share an anecdote, which is a friend of mine is involved in the hospital system in New York. And health systems have closed information ecosystems because they don’t share the information. And so, he, early on, started trying to train AI models on their internal health data.

And if you’re very, very wonky and have read Kahneman and Tversky’s early research, they actually started talking about how bad doctors were at reading radiology reports, basically X rays and other things. And so, what they found when they did this study, using again proprietary internal data, was the AI models were exactly as bad as the doctors in terms of interpreting the data. So, it wasn’t any step up in function.

My guess is, if you were to draw that analogy to the investment world, and also given the staggering terribleness of the data that we have out there, and the proliferation of beta, and people trying to game the system, my guess is we’ll look back in three or five years and people will say it’s an incredibly cool idea, but it seems to act in an unbelievably stupid ways a lot of the time. And really, it’s just going to be picking up and amplifying on the biases that we know exist. And I think one of the beauties of trend following, the way it’s done, is human beings are terrible market timers and it turns out that these models are pretty good at it. And that’s a real differentiating factor that I suspect is not going to go away anytime soon.

[00:46:00] Niels

Interesting. Good stuff. All right, let’s jump on to the next topic which is going to come from you, Mark. Where do you want to go?

[00:46:08] Mark

Well, everybody at the end of the year, they wanted to sort of look forward and sort of say these are my predictions for the next year and I want to take a chance and let’s look backwards. And in backwards I say like, what was the biggest mistake that you made in 2024? And given that mistake, what did you learn that you promised you won’t do again going forward in 2025.

So, you know, oftentimes they make the joke is that, well, mistakes were made. So what were the mistakes that were made?

And I probably, I’ll start off with the one that caught me by surprise and we’ll just sort of say from a person who looks at both price trends and macro trends, the one that really sort of like, you know, caused me the biggest concern was the divergence between Fed behavior in September and then the bond market.

So, usually, you always sort of think that if the Fed is lowering rates, well then I’m going to sort of see that across the board and there’s going to be this parallel shift down of rates in some sense. And I think even the Fed believed that, well, if they lower rates then the longer end will also come down and will relieve some of the pressure for mortgage rates. So, that was what I would call the biggest mistake I saw, which was conflating macro trends with market trends.

The question is, is this a big enough mistake that it’s going to cause a change in my thinking in 2025? I’m not sure, but I’d be interested in what other people’s mistakes were for 2024.

[00:47:56] Niels

Yeah, so would I, actually. Anyone want a venture something that they got very surprised about? I wouldn’t necessarily call it a mistake.

[00:48:03] Katy

I would agree with you, actually. My prognosis that we were going to have a steeper yield curve and, well, I was thinking to predict that again. But I think the biggest mistakes are people trying to anticipate Fed policy and sort of having sort of swinging back and forth between expectations has been the biggest challenge this year, I think for trend and also for a lot of investors, just trying to anticipate long term expectations in the short term.

[00:48:34] Niels

You know Mark, I was told early on in my career that the beauty about trend following is that it never makes any mistakes, it’s never wrong because we’re not trying to predict anything. So, I think for me I’m just going to leave it like that.

[00:48:45] Mark

That’s a good one. Yes. Yeah, it is what it is. Prices are primal. You know, it’s just that you’re not making any forecasts. All you’re doing is following what prices tell you. So that’s how you can’t make a mistake.

[00:49:00] Rich

I made a mistake in 2024. I complained about my wife’s spaghetti bolognese. That was probably the worst thing I did.

[00:49:09] Rob

I increased my allocation to trend following in, what turned out to be pretty close to a local maximum. So, the system did everything perfectly even when I didn’t.

[00:49:21] Niels

Anyways, let’s move on. Thanks for that Mark. Rich, you’ve been patiently waiting and it’s way over midnight where you are right now. So, we really appreciate that and we are very excited to hear what topic you’re going to bring up.

[00:49:35] Rich

I think this is for broad discussion here. It’s talking about dispersion this year. So, you know, I jumped onto IASG just before I came on tonight and I noticed that year-to-date, with the trend following community, it’s ranged basically, year-to-date, between maybe 70% for the year down to minus 5% for the year – but massive dispersion this year.

We’ve also seen dispersion over the last few years. I just wanted to open up the conversation and debate, you know, what are the reasons you guys think are the main reasons for this dispersion in returns? And this also creates problems in relation, talking about Andrew’s thing about trend following beta. Is there such a thing, especially with such large levels of dispersion.

We’ve got some of the benchmark indexes like SG Trend index which are, you know, they’re clearly not representative of the broader community of trend followers. They’re more the sort of the very big end of town in the trend following space. So yeah, what are some of your guys thoughts about the main causes for this level of dispersion, for instance in 2024?

[00:50:49] Niels

I would love to hear from Nick and Katy so why don’t you jump in first. Katy?

[00:50:53] Katy

Yeah, I mean I think return dispersion is sort of always underestimated by clients in our space because we all have very high correlation, but the range of outcomes are very large. And so, this year, in particular, we saw much larger risk dispersion and especially based on time horizons.

And I mentioned that earlier, and that is kind of puzzling to me because when you look at that from a trend followers perspective it’s really sort of just a fact of the data. But if you step back and look at the markets, my interpretation of that is that we were in a period where the long-term trends seem to be more clear but shorter-term undulations and sort of expectations and digesting all of this, what does it mean to be sort of moving into a new interest rate regime? What does it mean to sort of figure out the change and shift of regime in the US politics as well?

So, I think about it as one data point. This year is one data point, one data point of return dispersion where returns dispersion was very high based on the macro environments. And then we have other years where return dispersion is not as high. So, I think it’s just really a function of what trends worked and sort of at what frequency for me.

But return dispersion is always sort of an issue in our space because trends are very idiosyncratic, and they follow time series patterns that depending on your parameterization can be very different.

[00:52:24] Niels

So, any other one, Nick?

[00:52:27] Nick

Look, I mean maybe I’m going to say very similar things. In my mind managers differ for three plus one reasons, and I had a conversation with Andrew, about a month ago, on this one.

Number one is speed. Katy mentioned it. I mentioned it earlier. Number two is universe. We kind of touched upon it on the alternative markets. Number three is purity of trend or other stuff. Call it reversion, call it carry, call it whatever you want to call it. And the plus one that I’m going to just add is more discretionary de-risking.

I can easily see how in August between the NFP and Bank of Japan over that weekend, I cannot give zero probability on a systematic manager basically calling the shots over the weekend. I’m actually de risking. So, if you break everything down, the dispersion this year, I think it primarily comes from the speed. I agree with Katy on this one. And that’s back to the point I made about the August event in the very beginning of the podcast.

Number two, I think the universe could be a reason unless you’re an alternative markets trend follower, which is kind of very different. Some markets could, for example, have helped at times. So, I think it can be a factor.

I think purity is a factor of dispersion this year. I genuinely feel like, for example, cross sectional reversion dynamics played out super well both this year and last year. And the question then becomes, do trend managers differ because they don’t do trend? Maybe.

So, these are the three plus one factors that I have in my mind and the last one is not even systematic. Right. The discretionary de-risking.

[00:54:02] Niels

Sure.

Andrew.

[00:54:05] Andrew

So, I think back to Katy’s point. I think the dispersion space is always, I think the standard scenario is that dispersion is really wide. And so, you know, when we were looking at the space back in 2015 for the first time, we were considering should we just build our own trend following models, and we’re seeing what the banks were doing with their QIS models.

And I actually wrote a little paper on looking at all the different bank models that I could find. We were sort of stunned by the level of dispersion with everybody who’s walking in and saying we’re going to do kind of a simple trend model.

And when we started to build it ourselves, we found that small changes in parameters would not only result in big near-term dispersion, depending on which markets we’re overweighting, underweighting, how we’re sizing positions, but those often could extend out over time. You could go through these long periods where, if you were overweight currencies or something during a particular period, it would have a big impact.

So, to me, the reason we ended up kind of doing this kind of very, very simple top down replication, which is arguably not trend following to begin with, was just because we found it to be somewhat more stable and more adaptive as the industry changes. So, I’m surprised actually, when you say that the dispersion is much wider this year. I don’t see it that much in terms of, I mean I do see dispersion, but I don’t see it meaningfully larger than other years. But it doesn’t surprise me given all the factors that Nick is saying.

[00:55:29] Niels

Mark, you had a brief comment as well, I think.

[00:55:31] Mark

Yes. When I think of dispersion, I first sort of so say like, how do you characterize trend followers? And I’ve always used this sort of what I call a three factor model. I call it STM – style, timing, and markets.

So, once you put people into a timing bucket, a market bucket which, you know, with a sort of set style you use, which might be pure trend following or trend following plus something else, a lot of that dispersion goes away. So, part of it is a classification problem. And if you better classify the names you’re looking at, then you’ll see that there’s probably not as much dispersion as what you would see when you look at the overall set of performance in the market.

[00:56:21] Niels

Good stuff.

I actually had one more comment I want to throw in one that is maybe not often mentioned and that size. I wonder if larger managers this year just simply couldn’t get enough exposure to the markets that moved, like coffee and cocoa. Because if you have some kind of limitation on the total number of contracts you hold, then for a multibillion dollar fund, I imagine that you’re going to be hitting that number way much sooner than you would want in order to have a meaningful allocation to some of these markets. So, I think that might actually also be a factor. But many people don’t want to talk about the size issue, so to speak.

On that note, let’s wrap up part one of our year end group conversation. We hope that you’ve enjoyed it as much as we did making it for you. And if you want to show your appreciation for all the hard work that the amazing cohosts put into making these episodes each week, I would encourage you to head over to Apple Podcast or Spotify or wherever you listen to podcasts and leave a nice rating and review. We really do appreciate all of them.

And next week we’ll be publishing the second part of our group conversation. So, we hope, of course, that you’ll be back for more, and to hear our 2025 outrageous predictions.

From all of us at Top Traders Unplugged, thanks so much for listening. We look forward to being back with you next week. Until next time, happy holidays, and take care of yourself and take care of each other.

[00:57:51] Ending

Thanks for listening to the Systematic Investor podcast series. If you enjoy this series, go on over to iTunes and leave an honest rating and review and be sure to listen to all the other episodes from Top Traders Unplugged. If you have questions about systematic investing, send us an email with the word question in the subject line to info@toptradersunplugged.com and we’ll try to get it on the show.

And remember, all the discussion that we have about investment performance is about the past, and past performance does not guarantee or even infer anything about future performance. Also, understand that there’s a significant risk of financial loss with all investment strategies, and you need to request and understand the specific risks from the investment manager about their products before you make investment decisions. Thanks for spending some of your valuable time with us, and we’ll see you on the next episode of the Systematic Investor.