271 Systematic Investor Series ft. Mark Rzepczynski – November 25th, 2023

[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 Mark Rzepczynski and I, Niels Kaastrup-Larsen, where each week, we take the pulse of the global markets through the lens of a rules-based investor. Mark, wonderful to be back with you on this holiday week. How are you doing?

[00:00:39] Mark

I’m doing well. Yesterday was Thanksgiving in the United States. I will say it’s my favorite holiday. It’s hard to commercialize just getting family together, having a good meal and saying what you’re really thankful for for the last year. So this is my favorite holiday.

[00:00:55] Niels

Yes. I’m not sure you’re thankful for the trends we got so far this year, but maybe something else I imagine.

[00:01:02] Mark

Yeah. It’s been an ouch on a number of different levels I think, not only for trends, but systematic long-short, equity, systematic macro. A lot of managers have had some speed bumps here over the last year.

[00:01:17] Niels

Indeed. I have a feeling we’re going to be talking about that today. But before we do, I’m always curious, on a more serious point other than the turkey, just to know what you’ve been focusing on, maybe you’re paying attention to, outside the topics we’re going to be talking about in the last couple of months since we last spoke, anything in particular?

[00:01:38] Mark

Well, it always comes back to the topics we’re going to discuss, but I’ll say macro, macro, macro. I think especially with November and the last month, we really had a strong macro rally. There are companies that are called zombie companies in the United States, really poor companies, that in a strong macro environment, they’ll start having strong rallies better than good companies because they think that the good macro environment will lift all boats.

So you’ve got to focus in on how you discount macro information over different timeframes. That’s really been the focus of a lot of my attention or what I’m thinking about in the last month or so.

[00:02:20] Niels

Yeah, that’s a very interesting topic, actually. Now, before we get to all of those topics that you brought along, as usual, just a few words about trend following, where we are so far this month, so far this year. I do know we’re going to be taking a little bit more of a deep dive into some of this in just a few minutes. Actually, because of the holiday in the US, it has been quite a fairly peaceful and quiet week from a performance point of view. Not a lot has moved since last week.

My own trend barometer, I had a look at it this morning, and it finished yesterday at 25. That’s one of the lowest readings we’ve seen so far this year. And all of the year (people can see it on the website of course), if you look at the last 12 months, most of the time has really been spent at 40 or below, which really indicates that, even with this simple tool, the environment has not been conducive for trend.

I think we’ve had a few days above 50, which is really required for it to be called a good environment for trend following. So I think that oddly enough, this little tool is just displaying very well the challenges that trend followers have had this year and why the performance so far is negative.

Speaking of performance, BTOP 50 is down 2.7% so far this month and it’s down 0.9% this year. The SocGen CTA Index down 3.3% so far this month, and about the same for the year. Trend Index down 4.67%, down 3.67% for the year, and the Short-Term Traders Index down 1.18% for the month and down 3.65% for the year. So even though it’s been challenging, I will say I think that overall, the industry is holding up pretty well after a super strong 2022 of course. So maybe not surprising that there’s a little bit of giveback so far this year.

There’s not much giveback, so to speak, on equities. They are up strong in November. Up 8.67% for MSCI World, up 15.6% so far this year. World Government Bond Index having a good month. Up 2.37% for the month and just made it into the black territory so far for the year, up 1.15% so far. And the S&P 500 Total Return Index up 8.81% so far, very strong, and up 20.44% so far this year.

All right, Mark. Let’s turn the attention to your topics. They’re really good, very interesting. The first one really is something because, I think if we look at the trend year so far, there’s been a couple of defining moments. There’s been the bank scandal and there’s been that CPI number from last week or whenever it was earlier this month.

Those two months have been quite defining for the whole performance year; so March and November so far. You have highlighted some of these events in your notes to me. So why don’t you just talk a little bit about what is actually going on in the markets.

[00:05:30] Mark

Well, actually, I’ll probably say that there are three defining moments: I’ll say the bank debacle in the first quarter; then in the end of the second quarter, I think, the Fed raised rates. We had this different view on where the macro economy was coming in at the end of June, early July, which then spilled over into the Jackson Hole view that rates were going to be higher for longer. But this last month has been really an interesting period.

We had the big bond rally; 50 basis points in a month for the 10-year. We had a big macro stock rally; 45 points in the SPY. This is based on good macro news. Good macro news comes in a number of different forms. But this is without Powell suggesting that there’s going to be a cut. What we talked about in my initial comment, what am I studying?

Well, here we have a big macro rally. Especially on individual stocks, you get some very strange behavior where good stocks don’t perform as well as bad stocks. The good macro environment lifts all boats. The bad companies are now saying that’s where you want to put your money. But what you do, when you look at this rally in the environment, we’ve seen that CTAs have cut their equity shorts by 80%.

I was reading something from UBS that that means that’s equivalent to about $60 to $70 billion of flow that has changed. They’ve cut their bond exposure 35% to 40%. It’s another $50 to $60 billion. They’ve cut her dollar exposure and then there’s also been a big squeeze in the equity markets, especially it’s affecting long-short systematic equity managers.

So short positions have been cut, and this is a reflection of the fact that the names that people thought were bad are now good. We’ll say it’s the rise of the zombies in the equity market. All of this has led to a tremendous amount of flow, and that flow difference is now starting to have a greater impact on prices. It tells you, you’ve got to follow the flows.

[00:08:05] Niels

Yeah. It also tells me, when I hear those numbers (and I have no idea where UBS gets these numbers from; are they just looking at the Commitment of Traders report or whatever they’re doing), it kind of implies that trend followers would have been short equities overall. I don’t see that, certainly not from the people that I know best. I don’t see that we were massively short. We certainly weren’t short all markets, and actually, probably net-net not even short at all.

Obviously, this situation does create dispersion because some managers would have had maybe shorter-term timeframes and they could well have been short across many different markets. So these are interesting events, and I guess that’s part of why people tend to choose two or three different trend followers, because one trend follower is not the same as the next one.

[00:09:01] Mark

Right, timeframe matters. Now, in terms of where people get these numbers, a lot of large banks have now tried to say, what’s the size of the CTA market? What they’re trying to then determine is what does that translate into the size of trading in given areas? Then they’ll overlay a trend model on top of that to try to determine what they think are the flows.

More than one bank has been trying to reverse engineer what they think CTAs are doing and then writing reports to talk about when flows will change. Are they successful at that? Are they useful? I follow some of that because it’s indicative of some flow, but as you rightly mentioned, it’s hard to say exactly how much flow is occurring in the markets given the different timeframes that the CTAs use.

[00:10:01] Niels

Clearly, it’s not just Andrew who’s trying to reverse-engineer what CTAs are doing. But having said that, do you think they just do it because they want to have some information they can send out to their clients saying, oh, they’re going to be reducing by X amount of billions so I better buy now. Is it just for generating turnover at their own shop and generating commissions? Why do we have this, in my opinion, slightly unhealthy interest in what CTAs are doing to that degree?

[00:10:38] Mark

There has always been this flow interest. Even in the late ’90s or mid ’90s, some of the big banks were constantly trying to say, can they ascertain what the large trend followers are doing as a mechanism of flow. I think that they have a lot of the quant talent that they have in-house and they’re building more of these alt risk premium models. A lot of those alt risk premium models are based on momentum or trend. You can go to a bank right now and you could do a swap that will get you a trend model.

So given that they’ve built that infrastructure, now they’re saying if we could do this in-house, can we also then try to ascertain what the large CTAs are doing? So it’s a way to drum up more business, to get some more dialogue, to maybe get people to actually trade more. This has been an ongoing theme for a long time but especially picked up pace in the last few years as they’ve done more trend swap business.

[00:10:38] Niels

Okay, fair enough. Now, you’re also talking in your notes that you are seeing some differences between the economic data and the data that’s coming from these surveys, and that may be causing some confusion in the markets. Can you elaborate a little bit on that?

[00:12:02] Mark

This is something that when I started out, I said I’ve spent a lot of time on macro, macro, macro. Well, what you have is you have economic data that you could call it the real data. That’ll be the information that is generated by the government. Then there’s also survey data. This is, let’s say, the Conference Board surveys, University of Michigan survey.

So it’s a survey of individuals or investors to say, what do you think about expected inflation? What do you think about growth? What do you think about whether we’re going to be in a positive environment or more negative environment? Usually, those two move closely together. They’re highly correlated. Survey data may be a little bit more noisy but it does give you a good indication of what may be happening in the real economy.

If you look at more recent data, there’s a huge divergence between the real data and the survey data. I think you could even see this in the US, probably also in Europe, where the economic data seems like it’s not too bad in some places, especially in the labor market. But if you survey individuals, they say, I’m not feeling really good about the economy.

So given there’s a big disconnect between the real data and survey data, that’s causing a lot of confusion in the marketplace and that means it’s harder to determine what’s going on or where the economy will be going. A perfect example is the leading economic indicators are down for 19 months in a row. This is probably the same as 1994, 2008. And yet, you’re seeing this.

Actual inflation numbers are coming down. They’re nowhere close to the 2% target. But if you look at over the last two years, it certainly isn’t the expectation that it was going to be high for 5 plus percent. Survey data, people are still fairly negative. So there is this disconnect, and I think that has an impact on what people do with their decisions. It’s harder to get trends. There’s more uncertainty.

Now, even though there’s more uncertainty, we’re not seeing it in some of the data. The VIX index is down, as you probably follow that. The MOVE index is down. Uncertainty is down. But at the same time, it’s harder to explain what is going on in the real economy and how that’s translating into market prices.

[00:14:39] Niels

I understand that a lot of investors, that are not trend followers and rules-based and price-driven, will look at this data and make decisions on it. I understand that. I also get your point that there is this difference. I guess one of the things that I’ve been wondering a little bit about has been after the last CPI number in the US. Now, that comes from the BLS, some people call it the BS, but let’s call it the BLS.

So you have this inflation; this “official inflation number”. When you dive into it, the numbers look really weird, like a 34% drop in health insurance. I don’t think anyone has really experienced that in their own pocket. There was some other really funky stuff going on. Also, the way they allocate, they allocate twice as much to alcohol than they do to health insurance.

Again, I’m not sure that’s really what’s going on in the real economy. But then I heard, just looking at that, some people mentioned that you have these other data “providers.” I thought there was one called the C-H-A-R-D (or something like that) survey, or whatever it was, where they sample I think 500 products that most people buy in their household, or maybe it was 50 or 60 products, and they just look at that.

That inflation number, according to this article (it was written by a previous guest on the podcast), was up like 9% year over year. So clearly not the numbers that the officials would like us to believe in terms of inflation.

You clearly mentioned some surveys that are a little bit out of sync. But have you also looked at inflation and whether you think that what we’re being told from the officials are in any way, shape, or form close to what the real inflation is? I would imagine that’s pretty important for people who think that bonds are going to have a big rally from here or people who think that they’re not.

[00:16:46] Mark

The inflation numbers are head-scratching in the sense that there are different groups like the Shadow Open Market Committee where they look at inflation, but they look at it based on the basket of goods that were applicable 10 years ago or 20 years ago; what we use as hedonic pricing, let’s say we try to account for quality differences. So if a price goes up but you get more quality, then it’s possible that the price has gone down.

If you take the basket from 20 years ago or 30 years ago and you say I run that forward, inflation is running at a much higher number. So the perception that people have is that inflation is a lot worse. Certainly, if someone said inflation was 6%, 7% last year and now it’s down to only 3%, you should feel good about that.

But you’ll say, well, I’ve just had a debasement of all of my wealth by 6% last year and another 3%, I’m not getting that money back. So the surveys are saying people don’t feel any wealthier or they don’t feel any better just because inflation is now at 3%. Now, what does that mean for modeling in the systematic space or what does that mean for trend following?

Well, if you can’t really put a lot of stock in the government data because it may be, I’m not going to say manipulated, but because the way it’s collected and the way it’s presented may not give you a true story of what people are facing when they go out to the store and buy, then you’ve got a discount that. Which means that then you’ll say the only alternative is to just follow prices themselves.

But those prices are still being driven by perceptions in surveys, perceptions in the data that’s been presented. So you can’t really separate these two, but it does tell you that I’ve got to sometimes then go with just what the market is telling me as opposed to what the government is telling me because there’s this disconnect.

[00:19:02] Niels

I agree with that. Of course, I guess what I’m noticing is just that the market is reacting quite strongly to these official numbers. So it does have a spillover effect, of course, as to how we react as price followers on these things.

[00:19:25] Mark

You make the good point that on the one hand, you say I shouldn’t put as much stock on this inflation number given because of the way it’s calculated. But then we have a CPI print and the market goes crazy. There seems to be still putting a lot of value on those macro data.

Even though on the one hand you could say, I don’t believe these inflation numbers, I believe that inflation may be higher, but at the same time, if I ignore the print and the CPI announcement date, then that’s going to add a lot of risk to my portfolio because I got to take into account what the market is telling me in terms of the data.

[00:20:14] Niels

I think that’s still the strength of what we do, to a large extent, is that even if the news doesn’t fit into our positions, we just follow along. Anyways, you did talk also a little bit about it, and I guess we kind of started on that discussion in terms of price signals at the moment, where you do see some longer-term macro signals that are mixed, but then you have some short-term positive signals overall.

So is that from your own modeling that you see this dispersion at the moment? I guess it ties into one of the first things we talked about, and that is that we will be seeing some dispersion in manager returns in a year like this.

[00:20:56] Mark

Well, it’s personal modeling but then also, if you look at the overall modeling across many firms, both in what they write and what we see in the data, we still see that there’s a potential for a soft landing or recession in 2024. Now, we go back a year ago, it was going to be the second half of 2023. So it’s being pushed out in the future. The joke would be, yes, eventually, we’re going to get this recession. So I could say I’m forecasting a recession, I’m not forecasting the time of when it will occur.

That being said, we’ve got some good macro data in the short run. We’ve got some good prints and information. But at the same time, underlying if you look at a longer term, there’s still the threat of a recession in the first half or sometime in 2024. So the issue, when you think about macro data, is that how do you discount the long run versus the short run?

So I get a good CPI print that has a reaction, a positive reaction in the marketplace, but there’s other data that may be more negative. So how do I discount that?

[00:22:15] Niels

All right. Then you brought up an interesting point that I really can’t wait to dive into. The heading of this topic was really, what are investors asking in meetings at the moment? I’m curious to hear what people ask you in those meetings. We can swap a few stories, perhaps.

[00:22:36] Mark

Sure. Especially as you get closer to the end of the year, we’ll say in another probably two, three weeks, visiting clients or talking to clients is going to shut down. You need to get ready for the holidays. They’re going to go in the bunker to make their final end of year decisions. So I’ll say November Pre-Thanksgiving is a period that you want to try to go out and be a little bit more aggressive in seeing clients.

So you go into these meetings, and I always think it is really useful because if you talk to smart clients, they’re going to make you smarter. But you say, well, what are they really asking for? They’re getting very specific. How did you handle the early part of November when a lot of systematic managers got hit hard at macro? What are you doing about it? How did you avoid it? And then they’ll be specific in terms of how do you protect yourself from big reversals?

When you started to talk, at the beginning of our discussion, we had the first quarter debacle of banks. We probably had the end of second quarter short squeeze situation which carried over into the view on rates being higher for longer, and now this, we’ll call it the macro rally going on. So they ask to say, how do you protect yourself from that? That’s the number one question we hear from a lot of investors.

[00:24:07] Niels

Yeah. How do you protect yourself from that, Mark?

[00:24:11] Mark

This gets to be the issue where we start to get into philosophical differences. This opens up the Pandora’s box which is ongoing.

[00:24:25] Niels

Let me interrupt you here. That’s very impolite to do, but just to set the scene, because of course a lot of people will be asking that question. But if it’s a meeting with a systematic trend follower or systematic manager, in general for the most part, I wonder whether that’s a very good question to ask in the first place.

It assumes that we have some kind of magic wand whereby we can be always ahead of the market and when something happens out of the blue, like a CPI print or SVB going down over a weekend, we can somehow magically have the right positions on the next day even though markets were going in the complete opposite direction going into these events.

So I just want to preface with that because I think a question like that is kind of difficult to tackle in a nice way. If you think you can do something to avoid these small swans of some sort, then you can’t really be systematic, in my opinion.

[00:25:34] Mark

Well, you can be systematic because when you built your model you had to probably account like, how do I minimize drawdowns? So when you think about any kind of model building from its first principle, there are a couple of things you always want to try to do. You’re going to say, how do I ensure that I protect principles? So that means how you minimize drawdowns.

You want to try to say, how do I maximize my Sharpe ratio? Okay, I want to do that. I want to be able to say, how am I going to be uncorrelated with some of my peers if possible? Or how do I be uncorrelated with the marketplace? That’s the third. Fourth, you could say, specifically, what do I do to make sure that I’m going to get some low correlation on the downside.

When you start with a model, it’s almost as though you start at the end result and then, in some sense, try to work your way back. That’s not the way everyone does it, but I think that’s one way to do it. You say, what am I trying to create? And then you say, how do I do that? When they ask the question, how do you protect yourself from big reversals, they may not say, we’re not expecting you to change who you are, but how have you built a model, in the first place, to protect yourself from these large events?

The answer can’t be that we don’t. That’s not the right answer. But in some sense, and I think that we’ll get to this in a little bit more detail, every model or every firm has personality. It’s a strange thing to say because you think that when I build a systematic model, it’s it’s cold, it’s very analytic, it’s very rules-based. But even when you do that, a model has personality, a firm has personality.

When they ask the question, how do you protect yourself against big reversals, or the bigger question when investors say, what’s your edge, what they’re really asking for is, what’s your personality? Now, that’s a strange word to use, but I think that it gets to the heart of the issue. You probably have had guests on here, when you talk to them, the way they build models, it reflects their personality.

[00:28:10] Niels

I agree with that of course, that we all do things differently even if we may be categorized as being a certain kind of manager. So I agree with that. But on the other hand, there are certain things that you cannot model your way out of. This goes to a question that comes up every time we have “a difficult month,” whether it was Volmageddon February 2018; whether it was COVID, whether it was actually Thanksgiving two years ago, oil dropped 12% the day after or something like that, which was a major pain.

Of course a major pain as a one-day event, but it was the beginning of one of the strongest periods for trend followers that happened exactly on that day. So I think the true answer is, yes, there are certain things we can do to minimize the impact on some of these events, even though each event is different. So of course, you can’t give any guarantee.

But there probably are certain things you can do unless of course it’s over a weekend, SVB, etc., and even a CPI print. These are very difficult things to take account of. Anyway, in my experience, the true answer is, if someone asked you, can you do something about these unpleasant days or couple of days, couple of weeks, I think the answer is yes.

However, there’s a cost to doing it. I’m sure all of our peers in this industry will have looked at this and they will have concluded that, if we try to make the journey a little bit less painful during certain events, it’s just going to cost us down the road in terms of long-term performance.

So this is what I meant by when I hear the question being raised in that way: what do you do to protect yourself? I think the honest answer is there are certain things that we have to accept as part of this journey, and it could well be a couple of bad months or a couple of bad days. If we were to do something to improve on that, it’s just going to cost you significant amounts of performance compounded over 10, 20 years.

[00:30:48] Mark

When talking to clients, there’s a sequence of almost three questions. So the first question is, what’s your edge? Sometimes, that’s hard to answer because you don’t know what everybody else is doing. So, in some sense, talking about edge is validating what you do. Can you explain what you do in a good, clear narrative? Can you, in some sense, show your personality?

Then the next two questions that most investors will ask is, tell me the periods when you’re going to do well. What’s the environment you’re going to do well? And then the next question is, what are the environments that you do poorly? Sometimes, depending on your structure, if there’s a big macro reversal and it occurs in a very short time horizon, and let’s say my time horizon for my model is fairly long or longer, like an intermediate trend follower, then you’re going to have a situation.

There is going to be a disconnect between the two. There’s nothing you can do about it at some level. You could talk about diversification, you can talk about other things you can do. But at the heart of the issue, you can say, here are the environments I’m going to do well and I know myself, and here are the environments I’m not going to do well and I know myself, and in some sense, as you put it, there are certain risks I have to take in those bad environments in order to be able to exploit the good environments.

[00:32:28] Niels

When people ask that question, they ask the manager in isolation. So they say, oh, you’re down 5% this month or you’re down 10%, whatever. In March, you were down 10%. Not great, right? But we’re really looking at this in isolation because in March, equities and bonds were up strong. In November, equities and bonds were up strong. So when you look at it on a portfolio level, actually, there’s not a lot of damage.

In fact, I would argue that probably a portfolio even with a healthy allocation to trend, which is never going to be big enough to be meaningful with most investors, but you know it can. But even if you had a 25% allocation to trend, you’re probably still going to be up in March and be up in November.

Because this year, interestingly enough, stocks and bonds to a large extent have been positively correlated, but we’ve had very strong negative correlation to equities and bonds so far this year.

[00:33:30] Mark

Well, this gets back to the old adage that you really don’t have diversification, not unless you have some losers in your portfolio. At any given time, if everything you’re doing is making money, then you really don’t have diversification. From a portfolio level, if I’m making money in stocks and bonds, and let’s say I have a trend follower and I’m not making money there, what you’re really saying is that I really do have diversification.

With my partners, we use an ensemble approach. We have a combination of 23 different strategies, we’re using 18. At any given day, a number of them are going to lose money. I expect that. If it’s not happening, then I’ve done something wrong. The same, in a sense, applies from a higher portfolio level, which you just mentioned.

[00:34:28] Niels

Back to the person asking the question, what can you do about these drawdowns that occur from time to time? The question is, why would you want me to do anything about them?

[00:34:46] Mark

They’re going to ask the question because in some sense, there are two levels. If they’re talking to you, and you’re a trend follower, that means that they’ve already decided that they want to do trend following to some degree. Now the question is, which one should they choose? So there’s the absolute question, do I want this type of strategy? Then there’s the relative question, which one of the different choices I have?

So if I’m a systematic manager, if someone’s talking to me, they’ve probably already decided that they want to have some systematic exposure. Now, the question is, who should you choose? So now you have to say, why are you different from somebody else and what do you do differently in those extreme events? If you say, here’s what I do or don’t do when there are large macro reversals, which we had.

Then you say, now, in exchange for that, where I might do poorly in November, here’s the other environments where I’m going to do better. In some sense, I have to take some of that risk, like early November, in exchange for something that I’m going to get over the longer run.

[00:36:01] Niels

Thanks for allowing me to be a little bit of a devil’s advocate here. Of course, we know that is the question that people ask and we strive to come up with reasonable answers of course, not necessarily the answers I just gave. Anyways, these periods get a lot of attention from investors of course. But you brought up a topic about actually, there’s also things that don’t get a lot of attention from investors. Can you enlighten us on that?

[00:36:32] Mark

Well, a couple of things that don’t get attention, first is the attention of personality, which we talked about. I think that tied with personality is I think that type of knowledge that often is given with a manager. We’re going to talk about it, attention itself, in just a second. So there are two types of knowledge that you have. There’s explicit knowledge, the knowledge that you can get from a book you’ve been reading or a classroom, and then there’s tacit knowledge.

The tacit knowledge is the knowledge that is hard to convey, that you gain from experience. So in some sense, when people say, tell me about what makes you different, I would like to say that I have more tacit knowledge. Other managers that have been around for decades also have more tacit knowledge. This is the knowledge that is hard to put in a book but actually can be added into a model that will add value.

We’ll say explicit knowledge is when you buy a cookbook, there are all the rules. But the tacit knowledge is that knowledge that a chef has. So you could read the cookbook, but that doesn’t mean you’re going to replicate what you get in the restaurant. The chef is what matters. I think that oftentimes, when we talk about managers, we don’t put some thinking into who the chef is when we think about systematic managers, and that does matter.

[00:38:16] Niels

I don’t know if you’ve thought about this, and maybe this is where we’re going but I think it’s actually a very valid point and I think there are certain things where you just think, the way that he or she thinks about this problem led to them making some slightly different choices that actually could make a really important improvement to what they do in our space for example, I imagine so.

But we also know the power of narrative. We know the power of storytelling. So how do we take this slightly elusive, not specific data that you can’t really put it in a number; that he’s a 7, and the other ones are 5. Well, you can’t really say that. So how do you best explain this, which I think is an interesting observation, this type of knowledge that is somewhat elusive?

[00:39:19] Mark

This is the real art of marketing and going out and seeing clients, being able to give a clear narrative. We’ll talk about this in a very broad sense in just a second, but part of the narrative in systematic trend following has been where they say we’re uncorrelated. Then the narrative was we’re crisis alpha. The narrative has been honed a little differently more recently.

Part of that discussion with investors is that they want to try to understand your narrative and then can you be able to really clearly explain who you are and how you differentiate yourself? Now, one of the behavioral biases that we find, or a bias that we often see with many people, is that there’s the illusion of explanatory depth.

When people say, do you know how electricity works? You can say, yeah, I understand it. And then you say, can you really explain how it’s done? And then they’ll say, yeah, there’s the power plant somewhere and then I plug in my computer and it works. We think we know how things work when we really do not. So part of the discussion of narrative, when you go see clients, is that they’re trying to see what’s your explanatory depth, and your job as a narrative is to try to provide that.

I think from a very broad sense, let’s talk about how that narrative or explanatory depth has been applied especially in momentum and trend following. I think that this is where we get into this whole issue of attention. One of the things I was doing is that I try to keep up on academic literature. If you look at just the last issue of the Journal of Finance, for example, these papers have already been around for a while by the time they actually get in a journal.

But the explanation that is being given now for why momentum, why trend following works and I think a lot of systematic models in general work is what we call the inattention problem. Because markets or individuals in markets, the investors, are inattentive, it causes a slow speed of adjustment. The slow speed of adjustment means that there could be trends in the markets, there can be momentum, and that this is what those systematic managers are trying to exploit.

They’re trying to exploit the problem of market inattention. I’m not saying this is laziness. In some sense, efficient markets or rational expectations say that this shouldn’t occur. But when you look at reality, it seems as though we get a lot of evidence that it actually is occurring; that there is a slow response to information. Now, how do we look at this or what do we see?

Well, one study has shown that when you put on a screen a list of stocks, if you put a lot of winners, over the last two weeks, and then a random stock, most people believe that then that random stock would also be a winner. So they’re not focusing a lot on inattention. They just say, if it’s clustered with these other names, they must be related. What we’re also finding, in terms of inattention, is that they’re finding out that some of the smart money is actually sort of dumb.

In particular, there’s a study that looked at institutional money managers, and those would be ones who are running portfolios that are above half a billion dollars, what they find out is that the smart money has very good skill on their buying decisions, but they don’t show a lot of attention to their selling. Their selling is almost as if it was random. So, again, smart money is having dumb behavior because there’s a lack of attention to the selling decisions.

We also find this in, for example, earnings expectations. Earnings expectations adjust slowly over time because there’s a slow response to information. We can see that in how data gets updated. We’re seeing that that actually will then lead to slow adjustments in prices because if your expectations are slow to adjust, then prices will be slow to adjust.

[00:44:17] Niels

Since you and I started in this industry, the amount of information we get is obviously, I don’t know how many times, it’s more today than it used to be. That kind of bodes well, I imagine, for what we do because it’s impossible, essentially, to keep up with all that information. So as you said, it may not be laziness, but it takes time to go through all of that.

[00:44:44] Mark

What we have found or what researchers have found is that probably the speed of adjustment has improved. So whether it’s Bloomberg, other news sources, we probably adjust quicker to information in 2023 than we did in 1993, certainly more than 1983. So if we look over the span of time at speed of adjustment, we’ve gotten more attentive than what we were before because we have more tools. We have better tools to be able to assess information.

At the same time, we’re still not there. We still have inattention problems, and that actually leads to the opportunity for you to use systematic approaches to make money. So in some sense, when you think about let’s go back to the systematic narrative (this applies to trend, momentum, or any manager), you say, what do I do, what’s my edge? My edge is that I’m better able to take in information, assess that information quickly, and then generate a decision based on it. I call it the return factory.

The return factory is more efficient than let’s say a discretionary person. Because of that, that’s where we create edge or we create opportunities. Some markets are less efficient or have less attention than others. You look at commodity markets, we have a lot of natural hedgers, but that hedging behavior is almost slow to react. So people don’t put out a full hedge. They may take time to adjust to new information. So that creates the opportunity for trends.

If you look at equities, you look at the range of stocks, probably the speed of adjustment is pretty fast for those top 10, the top 7 stocks; the big mega caps. You go below that, you don’t have probably as much attention. Especially going to the small caps, there’s probably even less attention and there’s more opportunity for you to exploit that.

[00:47:04] Niels

It’s very fascinating to me what you’re saying. Of course, I can’t dispute the fact that we may be faster to react today than we were 20, 30, 40 years ago. I think that’s hard to dispute. But that suggests, to some extent, and I think that has been part of the narrative by some participants in our industry, that that bodes well for shorter-term models and that they should do better than longer-term models, but it’s just not what the data shows. I don’t think that we can say that longer-term trend following is doing worse in any shape or form than shorter-term models.

So have you thought about that slight contradiction? I think people have been swayed by the narrative saying that clearly must mean that shorter-term models, that’s where it’s going. I’ve heard that even from some of our friends in the industry, but it’s just not what the data shows. So how do we square those two things?

[00:48:15] Mark

That’s a very interesting idea. I probably would say I’m in your camp. We should expect the faster models to do better than they have and so you say there’s something wrong with this idea. Because if you believe that speed of adjustment is increasing, faster models should be doing better. So one of these papers that is looking at inattention and especially what they found is that the response of expectations is faster when there’s higher volatility.

Then what he did is he looked at longer-term and shorter-term models, and he said, shorter-term models do better in high volatility environments because then the speed of adjustment is faster in high volatility, which is a little bit counterintuitive, but that’s what we find. Expectations of earnings actually is more responsive in higher volatility environments.

So what he’s saying is that what we see is that in a high vol environment, short-term does better than long-term. But over longer-term horizons that are normal environments, long-term would do better than short-term. So I think we need to do more research in this area about the switching between time lengths. But we will say that if you had one single approach, intermediate trend following will be better.

But if you look at conditional on the environment, we also know that when there’s a spike in volatility, you’re probably going to get, I don’t want to say a momentum crash, but you’re going to get more reversals. You get penalized, to some degree, in the longer-term models in a high volatility spiky environment. So now the question that comes in is, how do we then deal with shorter-term and how do we sort of blend the two together?

Now, some may say, I don’t think it can be done, so therefore, this is the risk I’m going to have to take. But the data suggests that flipping between trend lookback base, on the vol environment, may have some benefit.

[00:50:34] Niels

I was just going to say that I’ve actually come across that narrative as well, where some people have in their models some kind of switching between time lengths based on volatility. But then again, when I look at their performance, I don’t really see anything that suggests that statistically, that is true. Well, that it has worked, let’s put it that way. it would almost be the holy grail of investing if you could somehow switch ahead of time so that you always had the optimal time frame. That would be fantastic. Maybe with all this AI, you never know.

That’s another narrative, and I’m being a little bit cheeky now, because again, that’s another trend, so to speak, and that is that AI is going to resolve all these problems and make everything easier and better, and so on and so forth. But I do find that people are often swayed in our industry and with many things by the latest, I wouldn’t call it gadget, that’s not what I mean, but the latest innovation perhaps we could say, even within investing, even within trend following.

You come up with something and say, yeah, well, now we can do all of this. You automatically think because it is new, it must be better. But I really encourage people to go back and study the long-term track records of some of these managers, because I actually think that that may not actually be the case in our industry in particular, so to speak, where some of these more universal, time-tested principles that we adhere to actually are far superior, or at least superior, than most of this new, innovative stuff that comes out.

[00:52:31] Mark

I couldn’t be more in agreement. There are fads in fashions and research. So when you get back to when we were talking a little bit about personality, part of the personality is how you account for the different fads or fashions that you see currently. But two interesting pieces of research that were also recently published in the Journal of Finance is that perhaps it gives us heart that we should go back to being traditionalists, but it also incorporates some of the new ideas.

One of them is that it focused on just looking at the direction of stocks, not the momentum. Can I just be able to get the direction right? I’m trying to forecast the direction, not looking at momentum, but then I try to measure the likelihood of direction. If you create a long-short portfolio just based on direction, the simplest of all things is that when you think about most trend followers, they’re just directional guys.

If the price is above the moving average, all I care about is direction. I’m not caring about what’s the mean squared error of my forecasts, I’m just thinking about direction. That actually outdoes a momentum portfolio. So if you sort according to momentum, or you sort just on the likelihood of direction, the directional forecasts of stocks actually do better especially over the last 20 years, which then tells you keeping it simple, trying to figure out the direction is actually fundamentally the right way to look at things.

The other study that was also done, in terms of fads and fashions, was that some researcher used some machine learning. They used convolutional neural nets, CNN, not the news service, but it’s a technique. It looks at images, and so it’s trying to forecast an image, not a specific point estimate. So what they did is that they said, what is the image we’re going to look at? Let’s look at a sequence over time of open high, low, close.

When you think about it, for a lot of people who are technicians, the bread and butter is looking at a chart of open high, low, close. Now, the systematic managers say, how do I use that information in a very disciplined rules-based approach? But a lot of technicians say, hey, I look at the chart, I’m looking at the image, and it’s going to tell me something.

So lo and behold, we look at this machine learning technique, the convolutional neural nets. It’s looking at the open high, low, close over a period of time as an image and we try to make predictions based on that image and say, well, lo and behold, it actually does pretty well. It actually beat some simple trend models and some momentum models.

Now, I’m not saying that we should now go out and use this type of convolutional neural net, but it tells you that just looking at the image, looking at a lot of information to try to describe what’s happening in price, which is fundamental to what people have been doing for decades, actually does work.

[00:56:10] Niels

The great thing about this is that we won’t know the answer to this until many years from now, but it is interesting, and it probably has a lot to do with what we also benefit from, which is human behavior, that there is just this tendency to think and believe that something that’s newer must be better. But I wonder if there’s a day where we realize that actually, things that are all tried and tested are perhaps what they should be looking for. We’ll see.

We have about five minutes left or so, Mark, and you brought along actually more topics than we maybe can cover today. But is there one left that you want to dive into before we wrap up?

[00:56:58] Mark

I have two areas I want to look at. One is sort of like a plug for myself and my partners. We’re having a paper come out in Journal of Investing that we looked at on backtesting because it’s a relatively new manager. We go in there, we’ll show some live trading and then it’ll also show our backtested numbers. And then they’ll say, I don’t believe a backtest. Or you say, I’ve never met a bad backtest, is the answer that the allocators will give.

If they do see a backtest, they’ll say, my rule of thumb is I’ll discount that by 50%. So we tried to address that issue by trying to say, how do you do a good backtest that should be applied as a standard operating procedure internally? Also, how do you describe in a narrative, to someone who’s a skeptic, what you do as a backtest?

So we think that for anybody who’s building models, both internal or showing it to external investors, this backtest issue should be addressed. I think that that’s one thing that we’ve been spending time on.

The other is that we spend a lot of time, especially in the last few months on transaction costs and the total cost of slippage and such. I think that we’ve really come to a conclusion after we study and we look at our live performance versus theoretical, and we look at that gap between the two, because that’s what my true cost is. That’s actually a source of hidden alpha.

A lot of systematic managers probably don’t spend as much time talking about how do they minimize T costs, because for every dollar I save on costs, or every dollar I save on slippage, that’s money that’s going to the investor’s pocket. From your incentive point of view, that’s also going into my pocket. So that’s almost as though that’s a cheap source of alpha if I could do transaction costs better.

We spent a lot of time on that and I think that better execution, we talk about maybe a little bit about it, but I’ll rather probably say that that’s something that we get back to the issue of personality, edge. What makes you different? How you think about it, or how you try to minimize transaction costs is an important part of who you are as a manager.

[00:58:25] Niels

I agree with that. Just one comment and I’ll see where that lands. I would have thought that most billion-dollar-plus managers, they do spend time and think hard on how to minimize slippage. Whether your slippage is 50 basis points per year or 25 basis points, there is a limit. And deep down, I would think that a really good way of minimizing slippage is not to overstate your capacity.

[01:00:06] Mark

Absolutely. The issue of capacity is something that oftentimes, as a relatively new manager, people say, what’s your capacity? We throw out a number and I think that we’ve tried to systematically address this issue. But then you really say I’ve just thrown out a number, here’s what I think my capacity is. It might be $3 billion or such. Well, what’s going to happen when I get to $1.5 billion? I might have to rethink this because the market situation may change.

Now, getting back to the issue of transaction costs, there’s a limit to how much you can cut those because you can’t go below zero. So we spend some time on that with some of our conversations with investors. We’ll show them our actual transaction costs versus our assumptions, whether we’re beating this, and we try to put it in a nice chart. People will tell us, they say, wow, this is interesting.

A lot of our managers who come in here don’t talk about how they minimize transaction costs. I’m sure that you and everybody else is doing it internally, but whether they share all that information of how they’re minimizing that cost with investors, that’s a different story. So they’re looking at it, but are they sharing that information?

[01:01:32] Niels

Okay, well, not to leave people on a complete cliffhanger, I think we may come back to some of this. Just a couple of quick questions that I would love to know the answer to. One, what do you think a decent reasonable amount of slippage, defined by the price you want to execute at and the actual price you execute it at, should be for a manager? And two, what would be your top two things that you think managers should be doing or are doing to minimize slippage?

[01:02:10] Mark

Right. How you minimize slippage, one, is trying to understand the capacity of what you can be able to trade in a given market. In a prior life, we’ll say that oftentimes, when we wanted to do some large trades, it would take more than one day to put all of that on. We were trying to minimize the impact of what we were doing. Because when you think about it, if you’re a large CTA, and you actually push prices with your behavior because you’re large, then you’re going to use those prices in your model to try to figure out the next signal.

So you want to make sure that your behavior doesn’t actually distort prices because those prices are then going to be used for you to figure out your next signal. So the idea is to try to say, how do you minimize the market impact? I think that’s something that’s very personal, but at the same time is critical. So what’s the timeframe you use, how much volume you represent in a given timeframe.

Part of this when you break down your trading is that are you a market maker or you’re a market taker? So you’re a market maker is that you’re providing liquidity to the market, and that might be that you’re trading reversals. If you’re a market taker, then that usually is that you’re a trend follower, you’re a taker, so that’s going to have a different impact on slippage. So it’s understanding what your size is relative to a given timeframe. What’s the market you trade, whether you’re a market maker or a taker?

So those are all issues that you have to think about. How do you actually do this or how do you measure this? I think that most of our managers have minimized the cost of brokerage. So the brokerage costs, that’s not what we’re talking about. What we’re talking about is, how do you minimize the market impact? That may mean that you have to do what we call AB testing. You have to try different types of execution strategies and then see what the impact would be and compare the differences.

[01:04:25] Niels

Fair enough. Last question, which is one of my previous questions, I just want to see if I can pin you down here. You’ve been around for a long time. You probably may not remember, but you probably have a good idea of what the slippage was back in the John Henry days. But nowadays, things have changed a lot. Nowadays, if you were to look at a manager and you were only to judge them on their capabilities of execution, etc., what do you think is a reasonable number to have as “slippage”?

I know it’s defined in many different ways, but just as a simple thought of, well, it’s not quite where we wanted to execute our trade, it’s a little bit worse. So that’s how I define slippage.

[01:05:17] Mark

It should be less than 0.5%. You’re trying to pin me down and I’m not going to give you a direct answer because it depends on market to market. I will tell you this, I spend a lot of time thinking about this. Because when you ask me this question and I can’t give you a straight answer, it’s because it’s a complex question. So tell you what, come back with the same question on our next podcast and I’ll give you a set number. So how about that for a cliffhanger? I won’t give you a direct answer today, but I will the next time people listen in.

[01:06:08] Niels

Absolutely. This is very likely to happen, Mark. You have to remember to bring it up because I may have completely forgotten.

[01:06:16] Mark

Okay. I’m writing down in the notes because that means I now have to come up with a definitive answer. You’ve put me on the spot.

[01:06:22] Niels

I think you’re allowed to be given a range, but I’m just curious. Anyways, this was wonderful, Mark. As always, you bring up some interesting points and maybe topics that are not discussed in so many other places, and you break it down so it’s understandable. Coming back to this, having a clear narrative and all of that stuff, you do a wonderful job. So thank you for that. I really do appreciate it.

Next week, I’m joined by Andrew. So now we will be talking about maybe one of the new shiny things in our industry, namely replication. But there could be other things. It very much depends on what questions we get. So I would encourage you to send us some questions to info@toptradersunplugged.com and I’ll do my best to wiggle that into the conversation next week.

Of course, if you want to hear more of these conversations, it would always help if you would share the podcast with your friends and family and colleagues, as well as rate and review them on iTunes, Spotify, Amazon, wherever you listen to your podcasts. We really do appreciate that.

From Mark and me, thank you ever so much for listening. We look forward to being back with you next week. Until next time, as usual, take care of yourself and take care of each other.

[01:07:46] 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.