225 Systematic Investor Series Group Conversation Part 2 – December 31st, 2022

[00:00:00] 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.

[00:00:23] Niels

Welcome and welcome back to this week’s edition of The Systematic Investor Series with Rob Carver, Cem Karsan, Mark Rzepczynski, Alan Dunne and I, Niels Kaastrup-Larsen where each week we take the pulse of the global market through the lens of a rules-based investor. Today, we’re back with Part 2 of our year-end special episode, where we all get together for one big conversation and debate.

As you know, this podcast series is all about voicing our differences on the one topic that brings us together, namely systematic investing. We’re recording on December 6. This conversation and what you’re hearing right now is the second part of our group conversation. So if you missed Part 1, make sure you go back and listen to that first. 

Last week, the conversation was more focused on key events relating to 2022, what stood out and what we learned from it, as well as the differences between explicitly long vol strategies and the long vol nature of trend following, among other things. Today, we’re going to broaden out the conversation a little bit. 

The first topic, selfishly, is from me. I wanted to bring up diversification as a topic. This might be a little bit outside your area, Cem, but I actually think it’ll be interesting to hear your thoughts as a non-trend follower, so to speak. So let’s see how we do that. 

Let me kick off explaining what I was trying to get to. Within trend following, we cherish diversification in many different parts of the investment process. And one of them, of course, is the number of markets we trade. This has often been a question that we’ve debated. And the question is, are more markets really better? Now, I know that Rich, unfortunately, is not here to defend his view. But I’ll do my best to provide an unbiased context.

In simple terms, if we trade, for example, 50 markets with a certain level of risk per trade, or we could trade 200 markets. If we’re using the same total risk, we would essentially trade one quarter of the risk per trade in order to get the same total risk. So far, I think that’s pretty uncontroversial. Now, if we trade 200 markets as opposed to 50, one of the thoughts I had was, can we or is there a way to be sure that we can find four times as many outliers because we’re obviously only taking a quarter of the risk per trade? So in order to get to the same opportunity set, so to speak, we should have more outliers, we should have four times more. Now, that’s part one of my questions, whether there is any thoughts on that. 

The other part is this thing where Rich takes it a little bit further away. He says he prefers to design his trend-following strategy to focus on hunting for outliers, these sort of 5% or so of really big market moves. I guess if you do so, you may implicitly leave a lot of profits on the table from the big trends that are good, but not the extreme outlier trends. I think that’s also a fair expectation. 

Again, I’m interested in your view on this as well, Cem, but I was trying to think of a volatility analogy, and obviously, probably have not come up with a very good example. But I wonder if it’s like buying a put or call option far out of the money. So sure, you’re going to make a lot of money if you’re right and you get this massive outlier move, but you’re also going to see a lot of those trades turn out worthless if the move just isn’t big enough. I don’t know whether that’s a good analogy or not. 

So a long-winded question, what’s your view about these extreme outliers with payouts compared to trying to capture maybe a bigger part of the distribution, but where you may also get a smaller payout on the extreme outliers, because you may do other things in order to catch a broader set of trades. 

So, I’m going start off with you, Rob, if you don’t mind, because obviously, you’re in the system design kind of realm. What are your thoughts?

[00:04:49] Rob

I think I can address the first question quite easily, and actually, the options example is an excellent one. The key thing about buying out of the money options is there’s a huge difference in magnitude between the premium you’re paying out and the potential return you could get. It’s a highly positively skewed distribution. And that means that even if you have to buy loads and loads of these things, you’re shelling out loads and loads of premium, sure, but you’re still going to get those big outliers arriving in your portfolio. The upside isn’t really going to be harmed by the fact that you’re having to buy more of these things because there’s this huge difference between the magnitude of the potential return and the premium you’re paying out. 

Now, in the trend space, because we can do this mapping between a trend and a long volatility strategy, they’re sort of the same thing. We say this quite a lot that their payoff structure looks very similar and their return profile is very similar as well in the sense that they’ll often bleed money for long periods of time, which is like your out of the money options are not paying off. And every now and then, you get a big spike and you get your outliers and that’s great. There is something you can do with your trend strategy, which is something I do, to make it even more like the option strategy, and that’s to have a variable forecast. 

To think of it in simple terms, it could be something where as a trends strengthens, you build your position and you make your position bigger. If you do that, that means that the difference in magnitude between a market where not really very much has happened (so it’s just facilitating long/short, long/short), you have a very small position each time, you’re gradually bleeding away money, sure, but it’s not really hurting your overall portfolio. Then, all of a sudden, it goes completely stratospheric. 

As it’s going hyperbolic and stratospheric and reaching for the moon or whatever, you’re building your position more and more and more, and that means your potential return is increasing and increasing. On that very small number of markets, you will end up with these huge returns compared to small losses and everything else effectively. So that’s one way to make your trend-following strategy a bit like holding a portfolio of out of the money straddles or whatever.

Personally, I think that the value of extra diversification is worth it. If all of these things had a straightforward linear payoff, then obviously, mathematically Niels, you’d be right. Buying four times as many markets means, yes, you’ve got four times as many chances as buying the winning lottery tickets if you like. But the potential return is going to be divided by four because the risk on each is a quarter. So net-net, you don’t really gain anything. In practice, it’s not like that because of this massive difference between effectively what is a bit like an option premium and what is a bit like an option payoff. As someone who hasn’t traded options in 20 years, I’m very much aware that I’m not an expert in the field and I look forward to hearing from our in-house options expert on what he thinks of my terrible analogy.

[00:07:55] Niels  

Exactly. Cem, what are your thoughts? 

[00:07:57] Cem

No, I think you got it right. Look, I think the first part of this conversation, you led off with The Rise of Carry. I think that’s such an appropriate way to lead off the whole conversation. We’ve been in this environment for some time now, which has been more certain outcomes, more leverage, more momentum, and yet more short-term tail, more extreme short-term tail. And that speaks to that skewness and that change in the distribution. 

In those environments, it’s probably good to be concentrated, less diverse outcomes, with more the Fed controlling a very controlled system in a certain way, but within that, having big blow-ups and then continuation of a more severe certain outcome. If the Fed loses control, if rates go higher, if there’s less leverage, it’s a more two-sided market. That’s a different distribution, as we kind of talked about last episode. And in those environments, there’s more uncertainty and there are more two-sided outcomes, and you probably want to be playing 200 names. You probably want to be out there in different places in the market with more options on top of that. 

As I mentioned, less liquidity means higher risk premia, more opportunity in different places, more fluidity. So, that also increases the opportunity in different places. It’s a more complex market in that world and more open for opportunity, and there you want to really diversify. So, I would argue it depends.
It depends on the environment, but I do think we’re getting to an environment where you’ll probably be better served by exposing yourself to a broader diversification of opportunities. I think particularly here on this turn of this regime shift, I think the opportunities are probably wider and more varied and there’s more just fishing in the little pond that we’re fishing in as well, which opens the opportunity for putting more lines out there.

[00:10:05] Niels

Sure. Mark, Alan, any of you want to sort of dive in on this?

[00:10:10] Mark  

Well, as I thought more and more about this whole outlier problem, I actually think of it in two parts. One is that outliers are time-specific. So in some sense, I could have outliers over a two-week period, I could also have outliers over a year period. There’s a fundamental difference between the two and how you capture an outlier over a two-week period is different than the way you capture an outlier for a year. So oftentimes, when the conversation about outliers is had, we say, “Look at this market, it has moved up by 150%.” Well, that could be a large number but you have to look at it in the context over the timeframe it did it. 

Sometimes if a market moves by 10% over two weeks, that could be a bigger outlier. So sometimes, if you’re capturing for long-term outliers or the large dispersion, you’re going have to use one type of model, a long-term trend model. If you’re looking for outliers over a short horizon, then you’re going to have to have a different model. 

The second part is always, it’s not so much the size of the outlier, it’s how much of it you actually capture. I don’t know how many times I’ve gone to clients and they look at a chart and they’ll say, “Look, there was a large move. Any idiot should have made money.” And I’m the idiot who did not make money over that period. Because maybe I was using a faster model and so I could have been in and out of a position a number of times. 

So one, you need to have outliers or you need to have big moves. But the question is, do you actually capture it? So if you give back half the move, I don’t know, is that something to be proud of if there’s 100% move of a given market over a period of time and they only capture, let’s say, less than 50% of that? Is that something to be proud of? So I think that you always have to put this outlier view in context. And by definition, the more markets you trade, at some point, you’re going to need to dissipate the value of any one of those outliers that you might capture.

[00:12:34] Niels  

Alan, I don’t know what your allocator eyes have seen over the years. I’ve always been reasonably vocal about saying that I see no evidence when I look at track records from various managers, those who trade hundreds of markets and those who trade the classic 50 or 60 markets. I don’t see any evidence that 200 markets are better than 50, 60 markets. Of course, I’m not an allocator like you.

[00:13:04] Alan  

From first principles, you would say more markets is more diversification, more opportunities. You would say it should be better. The idea of trend following is it’s a small edge in any individual market, but a big part of the benefit comes from applying it to many markets. Obviously, you get diminishing returns over time as you add more markets. Obviously, there’s costs, there’s liquidity as you move into the alternative markets that you have to be cognizant of. The question is, how much do you lever up for having those markets and have you got correlation risk? So I don’t know. I think I haven’t looked at it to say that the more diversified managers have done better. I think it’s interesting. 

If you think about it, there’s nearly three philosophies. At the other extreme, you’ve got maybe trend replication trading only 10 markets, you have a lot of the traditional CTAs maybe trading 50 to 80 markets, and then you have a few managers trading 200, 300, I think even more – maybe, 400 markets. So, there are different philosophies on this. 

I think maybe one thing to think about is the trend replication philosophy is about capturing big macro shifts saying, if there’s big changes in bonds and equities in the major markets, it’ll capture it. Whereas if you’re trading 300, 400 markets, you’re capturing much more idiosyncratic moves in individual markets. 

The nature of the return profile may be more driven by what’s going on in individual markets; in energy, in commodities for those versus the smaller number of markets. It’s going to be more influenced by macro risk. So I don’t know. I couldn’t say one is better. In theory, more I would have said is better but if there’s diminishing returns to more markets. 

I think the second thing I would just say is this whole discussion around outliers, it’s obviously something Rich talks about, Jerry Parker a bit. It’s almost like a different philosophy; that there’s two different philosophies. One is a belief in trends, and serial correlation and running trend-following strategies on that basis. And then a second one is that you get these periodic moves in markets. Every once in a while, you get these big outlier trades and you’re constructing a system to capture that. The latter seems more like a tail risk, kind of up long vol type strategy, as Rob was talking about. 

So I think it’s interesting that the people who end up running trend following, maybe from different philosophies, are, ultimately, similarly positioned but kind of driven by different beliefs. One is an academic kind of [belief where] there is serial correlation evident in markets, so let’s construct trend following from that basis. The other one being that periodically, you get these blow-ups in markets, outlier moves, so let’s construct systems to capture those. So, that’s the interesting aspect from my perspective.

[00:16:05] Niels

Sure. I just wanted to add to my observation about not seeing any evidence of one being better than the other. What I will say (and I’ve mentioned this before), and that is that I do see them being different; meaning that clearly, some of these funds doing 300, 400 markets will have performance at different times compared to others. That is, of course, clear.

[00:16:27] Mark 

 So Niels, one more on this issue is that there’s a difference between how many you trade versus how many you search on. In some sense, I think everyone would say that you’d like to search as great a number of markets as you can but you want to trade less. So, you search more, trade less. Ultimately, when you look at all of the risk premium analysis cross-sectionally, this is what they do. 

So, let’s say I’m looking at value trades. If someone said, “I don’t trade value on everything. What I do is if I look at the Russell 2000. What I’ll do is I’ll rank-order all of the stocks based on some value criteria, and I’m going to go long with the ones that are the best value and short the ones that are least value,” in some sense, I search more but trade less, so I’m conditional on a sort of phenomena. 

What you have to say is that if you’re looking for outliers, you may not always want to trade everything, but you want to search for everything, and there’s a distinction between those two activities.

[00:17:48] Niels

All right. Moving on to a topic from Dublin. Feel free, Alan, to expand on this. Of course, one of the biggest changes we’ve seen, in the last year, that I do think has an effect on probably most markets, is the fact that we’ve had this massive change in interest rate levels. These went from pretty much zero to, now, where we’re at 3 to 5%, depending on which country you’re looking at. How does that change strategies like volatility and trend following? That’s kind of how I assessed your question. Alan, do you want to add something before we go around?

[00:18:20] Alan  

No, I think it’s a big jump in the rate environment from 0% to 5%. You will get a reassessment of asset allocation. One of the things that I would say is that in the low-rate environment, maybe going back 12, 18 months, if you were making a case for managed futures strategies, the low returns in traditional assets was a key part of it. Now, we’re into a higher rate environment. So to what extent does that change the case from an asset allocation perspective?

[00:18:58] Niels  

Sure. Rob, any thoughts on that?

[00:19:03] Rob  

Something we’ve already touched on is that actually, when interest rates go up, your returns from having money in a CTA also increase. One of the great things about futures trading, unless your leverage is excessively high, is that you’ll mostly have a bunch of money sitting around doing pretty much nothing. So you’d actually now be getting a few percent in annual interest on that, which you wouldn’t have got a year ago. That actually feeds directly into your expected returns on that side as well.

I’ll be honest, I’m a bit of a skeptic of coming up with numbers and saying, “Oh, yes, over the next 10 years, stocks will return 7%, bonds return 3%, CTAs will return 8% and therefore, this is the optimal asset allocation.” I find that very hard to do, frankly. Within my own personal portfolio, I pretty much just keep the numbers fixed, to be honest. 

You could argue that there should be perhaps more of a tilt towards bonds because yields are higher; potentially a tilt towards equities because valuation ratios have been looking nicer than they were. They got pretty crazy, particularly in the US. Things like the Shiller PE were just really massively high. But it’s not stuff that’s going to move the needle very much. We are talking about reallocations of perhaps 5%, 10% at the absolute maximum. I’d almost question whether it’s worth doing that stuff, to be honest. 

So, yeah, skeptic sitting here, and frankly, not bothering to do anything with his allocation.

[00:20:47] Niels  

Cem, from a vol point of view, a change in interest rates like this?

[00:20:53] Cem  

We’ve done a recent study. You and I briefly talked about it on another podcast recently. I think this is an interesting place to dive in on it. Basically looking during secular inflationary regimes, how do different asset classes perform? And that’s not just how do they go up/do they go down? What does that shape of the distribution look like? I think, what’s the personality of these different asset classes? 

So again, when I talk about vol, most people think vol as an asset class. For me, vol is not an asset class, it is the underlying distribution of every asset class. If we speak to those distributions on how they are different in inflationary regimes, again, that’s my opinion, based on some of our macro work that we’re likely to see here for the next 5, 10 years. How do different things perform and how’s that different from the last 40? 

What we learned and I’ll kind of break this into two pieces to start, is what is the short-term volatility of each asset class, and then what is the longer-term trend and the longer-term vol? I’ll break this into five asset classes: equities, bonds, FX, commodities like oil (I will use oil as a proxy), and then gold. 

If we look at those five asset classes (and I’ll try and be quick about this), what we find is that risk premia go up because there’s less liquidity and short-term volatility goes up for equities, for gold, for treasuries, for FX. Ironically, the one that it doesn’t go up for is oil and commodities, which is kind of counterintuitive, and we’ll talk about that. Then directionally, what we see is in nominal ways is that equities actually flatline (in nominal terms) because of these first and second-order effects which we reference.

When I say FX, I’m going to talk about the dollar. We had dollar strength. We get oil secular strength, longer-term strength, treasury decline (as we would expect), and then gold strength. So most hard assets (and I include dollars in that even though they’re not a hard asset) have a positive right trend as you would expect. Treasuries, being the opposite of yields, have a negative trend. Equities, ironically, don’t have that  as a risk asset, they go more sideways. 

So what does that mean and what’s been priced in? One, we’ve seen an increase in vol, in treasuries, in FX, in commodities, in equities, but not so much in gold (which I think is interesting). When I look through this, we’ve seen an efficient response in FX. We’ve seen a relatively efficient move in treasuries. In terms of the vol markets and the expectations of those markets, those areas are less interesting now for me. 

Again, we were very actively saying people should get into what was cheap FX vol and cheap treasury vol, a year and a half, two years ago, but now those markets have adapted. Gold has been a counterintuitive move. Gold vol hasn’t gone. I think that’s a very interesting place to actually now step in, both in terms of more positive trend as well as in terms of buying vol. 

Long-dated calls look very, very cheap in gold, which I think is an interesting kind of thing. Now on the other end of that, oil. Oil vol has gone up, commodity vol has gone up. Historically, during these periods (and I’m going to dive into this topic a little bit more), you actually see vol compression in these periods. Again, I know that’s counterintuitive knowing that we have an OPEC crisis, etc.

Longer-term vol gets compressed, and even medium-term vol. Why? Because much like the Fed kind of controlled the FX and treasury markets for 20, 30 years, because it was this 10,000-pound gorilla, the big bully in the room controlling those markets and controlling, ultimately, the vol in those markets and creating a certain kind of trend. 

Now you have OPEC and oil producers who are, in a time of resource scarcity probably during an inflationary period, much more in control than maybe they were during other non-inflationary periods. This can drive a put underneath the market for commodities like oil. That put, much like the Fed put, can compress volatility in that space. That’s counterintuitive but in a period of inflation, we get resource scarcity. 

Hard assets and the holders of those hard assets have more power, have more centralized power, and have the ability historically to control the downside in those assets. That reduction of making it from a two-sided market into a one-sided market (much like these other asset classes when the Fed was in control), now provides a put to that market, historically. We’ve seen some of these effects already coming into the market. Not just OPEC, but with the SPR draining, now there’s a $72, $70 put where they’re refilling the SPR. 

OPEC Plus has come in and defined that they’re going to defend certain levels, because they can. Because they can and they have more power in this type of market. That should compress commodity vol, not just in the short-term, but over longer timeframes. It should create a very secular move higher in oil over the long-term as well. It creates almost like an equity-type distribution to commodities in this environment, whereas we’ve had that for equities for some time now, ironically. 

So, that would be a sale of vol there, right? Whereas I’m saying buy, you want to buy commodity. So, gold vol, you want to really sell commodity vol. I think that’s one area where it’s been really overdone, if you look at kind of a secular timeframe. 

Lastly, in an inflationary period, equities, as I referenced on the last episode, there’s an argument because you have a much more two-sided market, even though short-term vol will be higher and you’ll have a higher risk premia, that longer-term vol mean reversion and longer-term vols are largely at sale during inflationary periods on what is, otherwise, a nominal asset. It’s those first-order effects that a lot of those gold and oil and whatever benefit from. The second-order effects, which is the reduction in demand, the reverse TINA effect, all these other things take money away from risk assets and deleverage the system and really pull demand away. 

So, it’s a much more two-sided market, much more mean reverting market where vol compression is much more likely there. So, that’s what we see out of the distributions. We can spend a whole hour going through all of these and get into more detail, and maybe we’ll do that on another podcast. There’s some places that have completely overreacted. There’s some places that seem normal, but there are real opportunities on a macro scale, and the market is still really adjusting to what, I believe, is a regime shift.

[00:27:47] Niels 

 Sure. Mark, from your point of view, this massive change in interest rates, what effects do you think it may or may not have?

[00:27:57] Mark  

Well, of course, it’s going to have an impact. But let’s talk about how that impact actually translates to trend models and pricing in general. So when you have an increase in interest rates, what you usually find is that there’s an increase in interest rate volatility. So, the level of rates tells us something about rate volatility. What you find out is that when the level of rates goes up, and if volatility was still low, people could adjust. But let’s go back to first principles. 

When you price any asset, what you do is you discount cash flows. If the discount rate becomes more volatile, then prices become more volatile. So there’s going to be more dispersion in return. It’s sort of interesting that, even if you look at the last couple of days, instead of people focusing on the impact, on the numerator, on earnings, everyone talks about what’s the terminal rate that the Fed is going to do. So, it’s what’s the volatility and the discount rate that’s going to be driving performance in equity markets and bond markets.

Second of all, you find out that, as rates get higher, now you’re going to get more dispersion in expectations, which opens up a whole new set of trading. So, before when we had rates very low and we had strong forward guidance (basically trading back in euro, dollars or short-term interest rates), that business was dead. But now, all of a sudden with rates higher, now people start to say, well, what’s the probability of rates going to be above 5% versus below 5% after March? It opens up a whole new set of trading that we didn’t see before. 

So one, higher rates affect discount rates, which then translates to an impact on all asset classes. Two, higher rates lead to a dispersion in expectations. We now can have more trading because people have differences of opinion on what their level of rates should be going forward. So that’s the real impact of what’s going to happen. 

When you talk about a secular change, you look at the post-Great Financial Crisis, which we often talk about as the Lost Decade For Trend Following. That’s because if rates are flat, there’s no change in the discount rate for a lot of asset classes, you’re not going to get the same kind of dispersion. Now, all of a sudden, we have rates going up, and we go back to that first principle of discounting. All of a sudden we’re going to get more dispersion in pricing, which can now lead to greater opportunities. So that’s what’s really going on, I think.

[00:30:56] Niels  

Sure. Fair point. All right, let’s move on to a topic that came from Rob. Again, one of those that only had two words in it. So, you might want to have to set the context a little bit here, Rob. The words you gave me were AI and Trading – so, artificial intelligence and trading. Enlighten us, where we’re going with this, Rob?

[00:31:20] Rob

Yeah. AI is another thing that I’m a well-known skeptic of, at least in relation to the financial markets. I guess I have this view that as most people are terrible traders, there’s no reason why a computer trained to behave like a human should be a great trader. But literally in the last week, there’s been a real game changer in AI. There’s been a release of this thing called GPT-3, which is a new advanced natural language model. If you go on Twitter, it’s full of people posting conversations they’ve had with it, and it is quite extraordinary the stuff that this thing can do. 

So it was a random example. If you say something like, “Summarize for me this Shakespeare play but using the voice of Groucho Marx,” it’ll do that. It’s quite incredible, really. People have also said that it can do stuff like, for example, replace Google search and write code for you. There are massive caveats around that. One is that the training set for this data only goes up to 2021. So don’t try asking anything recent, it won’t just know about it. 

The other thing is it just generally does get stuff wrong. It’s sometimes spookily right, but then it will make a mistake, it’ll get something wrong, which is fine if you’re just playing with it. If you generally think it’s going to replace Google search, then that’s potentially dangerous. And the same goes when it is writing code as well, actually. It often produces perfectly good code. But then the code will have an error in it or a mistake or won’t compile. At the end of the day, all this is doing is, effectively, saying, “Well, given this string of characters, what string of characters should come next?” Given the training set, which is effectively the whole internet, it’s not really thinking in any key way. 

Anyway, having said all that, I do wonder whether there’s going to be a new rise in people saying that there’s going to be all of these brilliant AI techniques that are going to revolutionize the development of systematic trading models. This is a kind of hype that’s been an awfully long time coming, because I still didn’t see any evidence of this. 

As always, I remain open minded and I’d be curious to know whether the other guys on the panel have a similar kind of “grumpy old man” opinion like me or they’re like “Actually, you know what, the pace of change on this thing is so incredible, it’s not unrealistic to think that in a couple of years time, we’ll be at the point where we can literally say, ‘Model, here’s data’ (Go away. Make a cup of coffee. Come back) and have a completely brilliant trading model, not over fitted, ready to go. I’m very skeptical we’ll get there in my lifetime, even. What do you guys think?

[00:34:01] Niels  

Well, it sounds a little bit like all the promises that were made about self-driving cars. But just to be clear here, Rob, it’s not Elon Musk? Who’s behind this new model?

[00:34:17] Rob  

I think he may actually be one of the co-investors. 

[00:34:21] Niels

I think that’s an important piece of information to know.

[00:34:27] Rob  

I should say upfront, I’m pretty skeptical about everything that Elon Musk has ever been involved with as well. So there may be a theme here, but on this occasion, my skepticism is not purely Musk-based. I have other reasons to be skeptical.

[00:34:39] Niels  

Okay, fair enough. Cem, skeptic, not skeptic?

[00:34:43] Cem  

Look, given enough time, I think evolution is a powerful force and it’s exponential and we are heading towards a period where I think our brain and the way it processes, in my opinion, there’s a lot of things we don’t know [when it comes to] a very sophisticated computer. Can we create something that ultimately is an improved version of that computer? Yes, in my opinion. I think we have biases and things that have been ingrained over long evolutionary periods that probably aren’t suited for the current system. And that’s well documented by all these biases that we all know about. 

So, if we can take those biases out and have a more kind of open way to model, a more open system to model things, yes. Now, we are still so early in that process, and that doesn’t happen over weeks, months, years, even. It happens over decades and centuries. But it’s accelerating, and so my view is, yeah, I think AI, you can have smarter people, you can have better computers, you can have better systems. And I think my view would be that over longer timeframes, that these things are net positive. I think that’s probably well-proven in data. 

There’s a reason certain players in the market have been able to get certain returns over certain periods of time from a quantitative approach. But that’s my opinion. Again, these are big, big questions, right? Nobody’s got the market cornered on these and it definitely is a place where opinion matters.

[00:36:35] Niels  

Mark, you write a blog post every day. I’m sure AI has been part of your many hundreds of posts. What are your thoughts?

[00:36:44] Mark

Well, we’ll sort of say the idea of AI has been oversold. The impact of data science and machine learning has probably been undersold. What I mean by that is that I sort of came of age in more of the early ’80s graduate programs. So we always thought in terms of econometrics. There was no data science that we often thought about. 

When you think about the advancements of supervised, unsupervised learning, different techniques to try to be able to extract information from data, this has been a huge boom for trading, for strategies, and for just insights about data. And that has had a really large impact on trading and on how people view opportunities. Whether that can be converted into an artificial intelligence that will be able to be a perpetual money machine is a completely different question. 

So let’s go back to your self-driving car story. I sort of believe that we probably will have self-driving cars, but probably, if I got into self-driving taxi, I’d still probably want to have someone behind the steering wheel to watch so that occasionally, if it looks like we’re headed for a brick wall, it might actually hit the brake and override the system. So we’re not going to eliminate people. 

What you find out also, when you talk about artificial intelligence, is that your model or the artificial intelligence you build is a function of the builder. Who is the builder? He makes a tremendous number of choices on what he plans to add, what he doesn’t add, what emphasis he places. So, anyone who thinks that you could just build a model with AI technology and sort of say that you’re going to get a great solution is mistaken. It takes a lot of work and effort to understand how you build a model, what assumptions you make, and what output you’re actually going to get. So it’s a work in progress.

[00:39:05] Niels  

Absolutely. Any wise words from Dublin, Alan, on AI or would you rather stay out of it?

[00:39:09] Alan 

 Nothing much to add. What Mark said, in particular, just strikes me. We had, in the CTA space, a lot of machine learning systems, programs really being heavily marketed around a time when more traditional techniques weren’t working back in 2018, 2019. You don’t hear as much about it of late. I know of individual managers who have incorporated their machine learning techniques into systems evidently successfully.

So yeah, huge advances in computing power. You can see the potential so I can’t say no, but it does seem to be a cyclical thing that it comes on the radar every so often. I have an open mind but nothing to add to what you guys have said.

[00:40:04] Niels  

Yeah, no. I think I would agree with that. I do remember, very early on in one of the first, maybe first 100 episodes that I recorded, I did record an episode with someone who truly believed that they had found the holy grail with machine learning and AI and had gotten the best professor to build this. It all looked fantastic but within a year of our recording, it had blown up. 

[00:40:30] Mark  

Niels, one thing you should always talk about is that the evolution of strategies is driven by two factors: the cost of transacting and the cost of computing power. So when you go back into the early ’80s, you’d say why were a lot of people long-term trend followers? The reason why they were a long-term trend follower was not because it may have been the holy grail solution. It was because the cost of trading was really expensive. 

If you actually realized that you got cheap rates if you did $80 a round turn, which is just crazy. So, no one was going to run a fast system at $80 a round turn. Second is that if you look at some of the old boxes of computers, you sort of say you couldn’t test a lot of strategies. You didn’t have the computing power to do it. So, the evolution of trading strategies occurs because the cost structure of the businesses has changed, and that’s what’s really driving a lot of innovation in strategy development.

[00:41:46] Niels  

All right. We’ve got a couple of topics left for this year-end group conversation. I’m going to go with Mark’s one, and then I’m going to round off with a final one. 

Mark, you brought up the topic of liquidity. Interestingly enough, we already mentioned, in our conversation, the LME. I read this week that they had come out saying that their intervention, apparently, in the nickel market was necessary to avoid a $20 billion margin call that would have sent the market into a death spiral and threatened the exchange’s own survival. That’s what they said about that. 

I know that, obviously, the whole nickel market had massive effects with trades being canceled, the market being closed for a few days. Cem and I have also talked about liquidity changing in the options markets recently. I’m going to start with maybe Mark. Do you want to say anything on the topic briefly before we hear what the other guys have to say about liquidity?

[00:42:53] Mark  

We’ve had this underlying view by the Fed. They said, “We’re going raise rates until something breaks.” What is the most likely to break in 2023? My view is liquidity. Liquidity sort of scares the heck out of me, because when you think about it, it’s that trend followers are liquidity takers. Generally, they’re not liquidity providers. So, my lifeblood of being a trend follower is that there has to be somebody to take the other side of the trade and my ability to be able to take out profits or extract money is based on liquidity. 

If there’s no liquidity or liquidity changes, it’s the death knell, not only for trend following, but for a lot of other strategies. We’ll probably sort of say that more so than leverage (because you can say, “We can get changes in leverage and we will adapt to it”), a fundamental decline in liquidity will be a death knell to many hedge funds, to many strategies, and is probably something that will break in 2023 that will have a major impact. 

The one example, at the extreme on liquidity, was when they set the Blackstone Real Estate Fund where, basically, they limit the liquidity of how much can actually leave the fund. So we have a tremendous number of ETFs that are giving liquidity to investors for illiquid assets. That’s going to come back to haunt a lot of people. So, this issue of liquidity is something that is pervasive across all markets and it is fundamental to what’s going to happen in the next year.

[00:44:44] Niels  

Yeah. Alan, I’m going to allow you to jump in early on this one so you don’t have to just say I agree with everything that’s been said.

[00:44:52] Alan  

Now, I better have something to say on liquidity. 

[00:44:55] Niels

You don’t have to. It’s completely up to you.

[00:44:56] Alan  

I think two things. You’ve got liquidity, which is obviously an issue that’s always a concern in markets and is more of a concern now. I think certainly in the treasury market, there is a concern. We had the issues in March 2020 when basically the Fed had to intervene because you had massive basis trades from hedge funds. 

So we did see an incident where there an issue – a potential issue. All of these things seem to get resolved until the next crisis, but it did highlight a potential crack in the system. That reflected as well various kinds of changes in the market microstructure over the years. Dodd-Frank, less prop desk trading in investment banks (so, banks don’t hold inventory of assets), and all of this stuff, and then more high-frequency traders. These high frequency market makers tend to disappear as volatility picks up and then spreads widen out. 

We saw this kind of market action maybe a little bit in sterling recently when they had it set off. It had a big dive down on a Sunday night. So we do see these periodic liquidity issues. As Mark said, trend followers are price takers. It could be an issue in an individual market. Obviously, we don’t know, all of these things the treasury had been investigating, looking at the treasury market, serving banks, asking what they should be doing, etc. It’s hard to know. It is a tail risk out there. You could get caught on the wrong side of that. 

I think the second thing, in this broad category, I’m not sure if you can call it liquidity, but I think it’s linked to it, is the shadow banking system and everything that goes with that. Are there risks out there? As I said before, you had this kind of issue in March 2020 with hedge funds doing a lot of these basis trades. You’ve got a lot of exposure now out in the shadow banking system. 

The BIS had their report out yesterday, and there was a headline, I haven’t read it yet, but a headline about the non-bank system having $60 trillion in FX swaps unaccounted for. I’m not sure exactly what it was. There is this growth in the shadow banking system which is hedge funds, pension funds, or anybody who’s taken big, big, big trades and risk in markets and not a regulated credit institution. But there remains the question, are banks levered into these shadow banks as well in some shape or form? 

I think that’s the area where nobody really knows how that plays out. Could there be a problem there? Are banks going to be exposed by their prime brokerage divisions, etc.? So, yeah, I think it is a black swan, it is something that, if you were writing a list of risks in markets at the moment, a blow-up in the shadow banking system will be one, and that having an impact on liquidity in some market at some point in time. It certainly is a risk, yeah.

[00:48:26] Niels

Yeah. I know Cem is itching to say something, but before he does, Rob, liquidity, any thoughts on this topic?

[00:48:34] Rob

Yeah. Government bonds are interesting if you read the excellent newsletter that Matt Levine, of Bloomberg, puts out every day, you had this (for a while) recurring headline, everyone is worried about bond market liquidity. There are a number of reasons for that. One was (as Alan said) regulation, making it harder for banks to warehouse stuff on their balance sheets, which as a market maker that makes life difficult. 

Another reason for it was the fact that there were just fewer bonds out there. Because of QE, the kind of free float, if you like, of the government bond market was relatively small. There weren’t many bonds to trade. That is one thing that is reversing now. We know that as the balance sheets of the central banks get reduced, I think there are going to be more bonds out there. 

Certainly, in the UK, we’re running a kind of reasonably expansionary fiscal policy, which means there’ll be more bonds issued. I think bond market liquidity, at least, will probably be okay. As a futures trader, I kind of take liquidity as an input into my system rather than worrying about it too much. So, I just say if the market is not liquid enough, I’m not going to trade anymore. That’s my kind of pragmatic approach to dealing with changes in liquidity. But yeah, specifically in bonds, things may actually improve in the near future rather than get worse.

[00:50:01] Niels

Yeah. Interesting. All right, Cem, you’ve been waiting patiently to weigh in on this. I do want to encourage you to do it relatively short because we have one big topic left. But I know it’s a big thing for you.

[00:50:15] Cem  

Yeah. Liquidity is two sides of the same coin with volatility. Liquidity drives risk premia, it drives the shape of the distribution, as we were talking about prior. We’ve talked about this briefly, and I’ll try and be quick with this, but the S&P 500, particularly is a major source of liquidity and hedging, the amount of hedging in that space has helped compress vol and has been the source of liquidity. 

We’ve had all kinds of volatility increasing around the center of which is equity domestic US equity vol. And that vol has been kind of that Dutch boy with his thumb in the dike holding back the water on the other side. We had a market that was essentially 100% equity exposure. Think of it as a 30 delta put in the market. So 70% beta, 0.7 beta to the market vol structure. We’re still there at 0.7 beta, but instead of 100% equity exposure on a 30 delta put, we broadly have 70% equity exposure and no put. That’s the same beta. 

So, to a lot of people out there that are just looking directly, they think, “This market is about the same leverage as it was.” The reality is there’s just a much fatter tail. People are not as hedged for a downside move if and when it comes. This is that cyclical effect that we see people crowding into vol, crowding out of vol. We’re in the zero percentile because of all the liquidation of vol, because it hasn’t been performing and people are probably not hedged in the domestic equity space. 

That is a transition from a room full of pillows, where something happens, you get knocked over, you’re okay, to a room full of an inch of gasoline. That doesn’t mean something is going to happen, but the distribution has changed in the short term. There is a fat tail developing because there’s just less liquidity. That liquidity is not just broad liquidity, it’s liquidity on the tail. That is where the liquidity matters most. That’s where the most leverage is in the system. That’s where things get ugly. That’s where everybody tries to get out at the same time, and that’s where the liquidity is the poorest. We’ve had good liquidity on the tail this year. We see that structurally diminishing. 

That is, again, the Dutch boy with his thumb in the dike. He’s tired, he wants to go home (we’ve referenced this before). That’s really bad for liquidity across the market, like broad liquidity. I think that is something we’re likely to see in 2023 as a different type of move coming in, a second move phenomenon, because where the vol has been well supplied, that has been holding it all together as things fly around around the center is now actually weakening. And I think that’s an important thing to note in the short-term liquidity. 

I think this next move will release risk premia higher. I think they’ll secularly stay higher after that. Much like as a tornado comes through town, that hurts risk premia sellers when it happens. Afterwards, there’s a massive opportunity in insurance, generally. I think we’re going to enter a period, because of these other long-term trends I’ve talked about with vol, where after this next move happens there will be a big opportunity to harvest risk premia into broadly, as a secular trend, for the next 5, 10 years and really take advantage of some of these trends.

 But in the short term, I think liquidity is the word of the next year, and something, particularly at the center, that hasn’t broken yet. I think, at the center, it is breaking, and I think that’s important to know.

[00:53:48] Niels  

Yeah, absolutely. It’s certainly a topic we will stay on top of as we move into 2023. 

Speaking of 2023, that is going to be my last topic. We’re going to round off this conversation, two-part conversation with a glimpse into 2023. You all had different type of ways of phrasing it. Mark wrote, “What will be the big risks and opportunities for 2023?” Also, I can’t remember who wrote this, “What should allocators do in 2023 with their portfolios?” It could have been you, Alan. And also geopolitical events, so to speak. 

I’m going to kick it off selfishly, and just say a couple of things that I would worry about. I’m not so sure whether it’s a 2023 event. It could be building up during that year. It could be something that we don’t see until 2024. But I can’t help noticing some of the risks we’ve seen already this year in the pension fund system. We saw it in the UK with the gilt, we’ve seen, now, the largest public pension plan in Denmark losing 47% of their reserve capital in the first nine months of this year.

 I think this is much more critical because of the shifts we’ve talked about already with secular inflation and higher interest rates, potentially. And also, the other thing I worry about could be, if we do see inflation rise again next year, if we do see interest rates go up and no pivot, I wonder if the market could basically lose confidence in the policies of the central banks. I think that can also be a massive event for the market. 

So is it 2023, 2024? I have no idea. But those are some of the bigger things that I would worry about. Anyways, Alan, what are your thoughts as you look into the new year?

[00:55:56] Alan  

Yeah. I think I mentioned shadow banking, which includes pension funds, obviously. That’s pension funds, hedge funds, insurance companies, all of those are part of that ecosystem. So in terms of risks, I think that’s one thing I would be focused on. I think as well, and kind of linked to that, t a lot of these have gone heavily into privates. How does that play out? At some point, do we get a remarking of assets? It just doesn’t appear to have happened yet. 

We heard about the Blackstone story. Part of the reason that they had the outflows was because REITs generally were down in value and their fund hadn’t marked down yet. So, people took advantage of that. Obviously, we don’t know how public markets will trade from here. But at some point, if we did see this markdown in privates, how that impacts all of those kinds of investors who’ve gone so heavily into that sector. 

I think from an economic perspective and markets generally, the thing I’m looking at is the lags. As we said, we’ve had this jump from 0% to 5%, nothing’s broken yet. That doesn’t mean we’re not going to see a nonlinear reaction at some point. That could be next year. It could be that we see the data… ISM services came out strong, you could have a couple of months of strong numbers, and it keeps the Fed going, but then you ultimately get a nonlinear response to the higher rates. 

I think back to the financial crisis, and back then, a lot of people had these adjustable-rate mortgages in the US. So when rates started to go up, you didn’t get much of a reaction to higher rates. But then you reached kind of a tipping point where you got a nonlinear response to higher rates. So I think that’s a possible scenario if I had to make some kind of projection from the economic perspective. That could be the scenario. But who knows? As we say, all of these are just stories that we tell that might play out. But I definitely think that the private sector is going to be tested through this next downturn or through this part of the cycle.

[00:58:15] Niels

Rob, that’s been two negatives? Do you have a positive outlook maybe for 2023?

[00:58:21] Rob

  You know me. I’m always positive. Yeah. Unfortunately, I think I’m going to spin the theme of it further, really, because I do think that the one thing that worries me is things that we don’t know about. So think back to ’07, February ’07, how many people knew about things like CDO Squared and leveraged MBS and stuff. People just didn’t know this stuff existed until it blew up in our faces. 

This was exactly the same with the LDI crisis in the UK a few months ago. That was just a completely unexpected thing that no one seemed to know was there. Maybe only a few people knew about these things and none of them predicted what was going to happen. So, that stuff does worry me. 

I think going back to our early conversation about crypto, what’s been staggering is the number of relatively respectable PE funds that invested in FTX, for example. I think that out there, there’s a lot of people sitting on leveraged bets. It’s the classic when the tide comes out, you can see who’s been swimming naked. The tide is really starting to recede and I think we’re going to see a lot of leveraged balance sheets kind of blowing up here, there and everywhere with people who’ve bought stuff in this desperate hunt for yield when interest rates were low. They’d just gone out and sought returns where they could get them and they’ve leveraged them up to juice up their returns. That’s not an unexpected thing, because it happens all the time. 

But as you said, it’s the private markets. There are things that aren’t easily visible and observable, and it’s not until the explosion happens that we can see them happening. And these sort of so-called Black Swan events are, by definition, really, really hard. Apart from, of course, investing in a tail protection fund, a reputable one, it’s really hard to know how to kind of quote-unquote “insure” yourself against them. So yeah, it does worry me that there’s probably a lot of these bad stories going to come out in the next few months or even a couple of years.

[01:00:22] Niels

Yeah. Mark, before we get to Cem, who will have the final word, what are you looking for in the new year, Mark?

[01:00:30] Mark

What I love is the fact that there is dispersion, and dispersion meaning that there’s plenty of differences in points of view. Let’s take the optimistic point of view. If you look at Fed forecasts, a lot of forecasts from Wall Street economists, they’re arguing that interest rates and inflation is going to come down in the second half of the year. So from an optimistic point of view, you’d say this could be a great round-trip environment. We made money as your rates went up and inflation went higher, and then we’re going to figure out how to make money when it reverses and goes back the other way. So that’s the great optimistic round-trip story that will sustain trend-following for 2023. 

On the negative side, what we’ll sort of say is, I’ll call it the Cockroach Story. Usually, if you’ve found one cockroach, usually you can  rest assured there’s more than one out there. So, if we have one crisis, we’re going to have multiple crises. And that’s still likely going to be on the negative side that we’re going to have one blow-up, and that’s going to lead to others that we’re going to see. Part of that blow up risk, there’s the cockroach of behavior and there’s a cockroach of policy. My prediction is that whoever is in charge of government policy, whether fiscal or monetary, they will make mistakes. We’ve already seen the mistakes. Will they ever admit that they’ve made policy mistakes? The answer is no. But we know that they’ll occur and from that, there will have to be adjustments and those adjustments create trends and those create opportunities.

[01:02:10] Niels  

Absolutely. All right, Cem, the pressure is on you. Final word, outlook for 2023 as we wrap up this amazing conversation.

[01:02:21] Cem

Yeah. I’m going to try and tie it all together. Rob mentioned you can’t predict the unexpected, and we’re likely to get more unexpected. But as we referenced earlier in the conversation, you can predict the shape and the distribution broadly if you get these things right. 

I think it’s important to note that we saw during other periods of inflation a clustering, a clustering of things, a fatter tail, less liquidity, more risk premia (as we mentioned). But less liquidity also means things are changing fast and there’s less ability to absorb them. So, you get more of the unexpected. So, expect the unexpected. That sounds like we’re making no prediction. That’s not true. We are talking about the shape of outcomes and how one should be generally positioned as a function of that. 

Mark mentioned this cockroach theory. During periods of high inflation, you get resource scarcity. Because there’s less liquidity, power matters more, centralization of power. We’re transitioning (this is what happens during inflationary periods) from a time of cooperation (where corporations are getting all the money, corporations are global, and there’s cooperation), to a time of, hey, there’s scarcity of money, there’s scarcity of heads commodities, of labor, and these things drive competition and more of unexpected power shifts. That leads to war, historically, that leads to commodity crises, that leads to funding crises. 

Resources are more important. In the reserve currency, that’s a major source of power, you have leverage. If you have commodities, that’s a source of power, it’s a source of leverage. If you have labor, that’s a source of leverage, that’s a source of power. These things are fundamental truths and important and they have less importance in time of cooperation when we’re working across borders. But this is a time of nationalism, protectionism. We saw this in the ’60s and ’70s. This is part of what drove the Vietnam War. This is part of what drove the OPEC crisis. This is part of what drove labor rights and protectionism. These are things that we’re likely to see in the periods ahead. Again, it speaks to the shape. We’re likely to get debt crisis and things along those lines as well. More recessions, more tails more often, a more leptokurtic fat tail distribution. 

If I was to button it up and talk about what the trends are, those things will ultimately lead to more inflation, more labor rights, more protectionism. More competition ultimately means less free market economics, more protectionism, like I said, more labor rights and more populism I think is the word of the day. Those things ultimately mean higher interest rates for a more secular time and more short-term crises.

[01:05:19] Rob  

Wow, can I just say I think we’ve got a winner for today, which is using the word leptokurtic gives you 500 points, which puts you at the top of the leader board. 

[01:05:30] Niels  

Absolutely. Absolutely. 

On that note, we’re going to wrap up this Part 2 of our conversation. We hope that you enjoyed last week’s Part 1, as well as today’s episode. If you did, as usual, we would so much appreciate it if you would head over to iTunes, Spotify, Amazon, leave a rating and review so that more people can find the podcast and make sure you follow the podcast so that you know every time we publish a new episode.

Next week, we’re going to be back with our usual format. Make sure you join us and send your questions to info@toptradersunplugged.com. We’ve got some great midweek episodes also coming up that I’m already aware of. So make sure you catch them as well.

From Rob, Cem, Mark, Alan and me, thank you ever so much for all of your support in 2022. We love you, we appreciate you, and we look forward to being back with you next week. Until that time, take care of yourself and take care of each other.

[01:06:30] Ending

Thanks for listening to the Systematic Investor podcast series. If you enjoyed 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. 

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.