If Trend Following was an Olympic Sport… ft. Mark Rzepczynski

[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.

[00:00:16] 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, it is great to be back with you this week. How are you doing?

[00:00:37] Mark

Not too bad, not too bad. Middle of the summer, you know, will probably sort of save my town. I think more people are going out on vacation, or they should be starting in August, and the place empties out. So, that’s when I like my town the most. It’s relatively empty.

[00:00:53] Niels

It’s pretty empty here in Zug at the moment, I have to say. But you know what day it is? It is the day where the Olympics start, which is very exciting in many ways.

And so, I obviously hope you are ready for it. I don’t know if you are interested in these types of events, but it inspired me a little bit to bring up a kind of a surprise question for you, Mark, because I know you’re up for it. So, the surprise, the silly and surprise question is, if trend following was an Olympic sport, which one would you say it is?

[00:01:33] Mark

Ooh, that’s a good one. So, if it was the winter Olympics, then it would be, you know, those guys who do this ski jumping where it could be, you know, either graceful or the agony of defeat and a crash, because we see these ups and downs with managed futures. Now, in summer Olympics, it’s not really, it’s harder to say, but I think it’s closer to more long distance running as opposed to a sprint.

[00:02:10] Niels

You know what? It’s so funny. I mean, when I thought of this silly question a few minutes ago, I didn’t necessarily have a good answer myself. So, I asked ChatGBT, because I thought, who else to ask?

But actually, you’re spot on. So, ChatGBT, and it was kind of a test to also find out, does it really understand trend following? ChatGBT came back and said, well, if trend following was an Olympic sport, it would be likely to be akin to marathon running, and here’s why.

And they go through four points, but it ends up summarizing and saying, in essence, the qualities that makes a successful marathon runner, endurance, consistency, adaptability, risk management, and strategic planning are also essential for successful trend following in financial markets. I thought that was pretty good.

[00:02:59] Mark

Well, I’m amazed that that’s what ChatGBT came with, and I’m also amazed that I sort of matched.

[00:03:09] Niels

It was completely unfair to do this on you, but you did very well, and it certainly proves that you know your trend following, Mark, for sure.

Anyways, as always, what I’d love to hear, sort of more broadly, actually, before we dive into our topics, is just kind of what’s been on your radar recently, just if something has come up that you find interesting.

[00:03:16] Mark

You know, the topic that I found most interesting, and this is in the last week, was from a good friend, very good investor. We share conversations at least once a week. He said, have you seen this paper on activist treasury issuance and monetary policy?

It’s a piece that has come out from Hudson Bay Capital, and the basic premise is that the activist treasury issuance of more bills than coupons has had an impact of providing quantitative easing that’s offset what the Fed has been doing. So, we’ll sort of say the Fed and the treasury, so they’re following offsetting policies. Fed trying to tighten financial conditions, and the treasury, through its issuance, if it issues less coupon bonds, then that means that those coupon bonds become scarcer, prices go up, and at the same time, this is that they’re providing more treasury bills, which is near money or as close to money as you can get.

And so, it’s actually following an offsetting policy, and the numbers are staggering in terms of the size of the difference of issuance. And so, you have to ask the question, well, why is the US economy doing so well? And why did the transition to lower inflation take so long?

And why are financial conditions still relatively positive in the US economy? And this tug-of-war issue between treasury and Fed seems to make sense, and it’s an area that I want to explore more, and I know that it sounds like you’ve had other conversations and other podcasts, but I think this is a very intriguing issue.

[00:05:27] Niels

Well, I did notice that, not the last episode, but the week before, Alan was speaking to Adam Posen, who obviously used to be a central banker himself, and he was quite taken back by the fact that the US didn’t issue very long-dated bonds when they had the chance a few years ago, like we saw in Austria and maybe in other countries. When interest rates were so low, why didn’t they lock in some long-term financing? And there was a bit of conversation about that, which I actually thought was quite interesting.

But maybe for those who may not be following this so closely, Mark, what is the case they make for issuing shorter-term debt being more accommodative rather than long-term debt?

[00:06:19] Mark

Well, there’s a couple things going on. First, there’s a greater demand for treasury bills because of the way we’ve changed the regulations on money market funds. That came around in 2016, so there’s been this natural stronger demand for shorter-term instruments, so the treasury might actually add more issuance.

But when the treasury issues bills, bills can be posted as collateral for lending. It’s collateral that’s needed by a lot of institutions, and so it actually has the characteristics of being money. So, if you more bills, it’s as if you’re issuing more money.

Alternatively, is that if you stop issuing coupon bonds, coupon bonds, because they have longer duration, are risky assets. Because they’re risky assets, there’s a limited ultimate demand or aggregate demand for those coupons. And so what happens is that treasuries have been bid up because there’s been less coupon supply available.

And if that occurs, then long-term interest rates are going to be lower, which is good for financial conditions. So, when you think about it, the supply can have an impact on the shape of the yield curve. It can also have an impact on the amount of risky assets outstanding, and it also can have an impact on the amount of near money that’s available.

And so, what happens is that the treasury usually has a rule of thumb to say about when we issue debt, we’ll do about 15% in bills and the rest in coupons. They’ve done numbers that are much higher on the bill side. Because of that, that has actually had a big impact on the supply relative to demand along the yield curve. And that has an impact on financial conditions.

[00:08:38] Niels

Absolutely. And it is interesting in a sense because, okay, maybe we do talk about this from time to time. We don’t talk a lot about it. And we certainly maybe don’t think about it so much that this is not the Fed doing it. This is the treasury controlling this. And so, they can kind of offset all the good work, “good work” that the Fed is doing, trying to bring down or being kind of restrictive in its policy. And then they can come along and offset some of those effects.

[00:09:09] Mark

Right. And so, when you think about while we’ve always focused in, more recently, on interest rates, there is the backdrop. The Fed is still doing QT, quantitative tightening, on the supply side on their balance sheet.

And so, every month when they sort of sell off more of their coupon portion of their balance sheet, okay, that’s like adding supply to the market. So, in some senses, if the treasury doesn’t issue coupon bonds, then in some sense it’s actually sort of offsetting the pressures that have been put in from the Fed when it actually sort of sells off their coupon securities as part of quantitative tightening.

[00:09:58] Niels

Yeah. All right. Well, I mean, on my radar, Mark, to be perfectly frank, it has to do with trend following, unsurprising, but it is “inspired” (maybe not the right word), but it is a little bit related to this somewhat relentless give back of performance that trend followers have experienced in the past three months.

Actually, it continued in July. Even after a very good start to the month, we’ve kind of completely given that back and we’re in the red, as an industry, for the month of July. And it’s going to certainly foster conversations with investors, potential investors, because these are the periods that they don’t like. Of course, the optimist in me will say, well, it also opens up for some great entry points because still, in terms of investors, there are still way too few who have exposure to trend following and those who have are probably not enough of it.

So, I don’t know if it coincides with whether there is a link to the fact that there seems to be a little bit more calm in terms of where inflation is going. Obviously, there are some numbers coming out today, I think, as we record, PCE, but inflation concerns are maybe not as high as they were before and rates are subsiding, to some extent. So, I don’t know if we’re sort of opening up the debate about, well, if everything calms down again, trend followers might have a bit of a struggle like they had in the 10 years leading into 2019 where everything was well controlled by central banks.

I’m not sure we’re there yet, but what are your thoughts about these things when we go through one of these periods?

[00:11:56] Mark

Well, this is the age-old problem with trend following. In a sense, that we’ve done a very good job, trend followers in general, of talking about crisis alpha, and very good job of talking about here’s the benefits of trend following when there may be a sell-off in other risky assets like equities. But there’s a cost with that and the cost with that of having low correlation is the fact that when there’s calm periods in the market, such as when we have interest rates, if volatility has come down, obviously with inflation – inflation, volatility, and uncertainty have come down.

And markets, in general, might be a little bit more sideways or they’re retracting from highs. You’re going to have bad performance. And then when people look at their portfolio and they see, look at all these things that are doing well, because a lot of alternative investments have done well so far this year, but then they look at trend following, they see it’s negative.

They’ll say, why do I have this in my portfolio? And what you find is that they’ll forget, fairly quickly, the benefits of when the market turns down. When they look at it today, over the last three months or for the year to date, and they say, how do I rebalance my portfolio? It doesn’t look like you should have those trend followers there at the bottom of the list relative to many other strategies at this current point in the year.

[00:13:27] Niels

Yeah. As you’re talking, I just see the screen lighting up here with PCE numbers coming in a little bit higher than what was expected. So anyways, an ongoing story for sure.

But it did lead us into talking a little bit about trend following. As I mentioned, in terms of trend following, we’ve had another week of the continuation that we’ve seen the last few weeks or a few months, actually. And we’ve had quite a few, almost like a series of reversals in different markets, different sectors in the last few weeks as well.

Certainly, this month so far, currencies have been, as far as I can tell, has been hit pretty hard, not least from the strength, the newfound strength of the Japanese yen. So that’s been a major challenge. But also equities, metals, energy, fixed income, not a lot of sort of consistency from a trend following perspective.

So, it is just one of those transition periods that we are going through right now. My own trend barometer closed yesterday at 36. That’s weak. It’s not super weak, but it is weak. So, it certainly suggests a negative month so far for the industry.

And if I look at the numbers, the BTOP50 index as of Wednesday night, it must be, was down 1.48%, up 5.47% for the year. SocGen CTA index down 2.19% in July, up 4.91% for the year. SocGen Trend index down 2.21% for the month and up 6.28% for the year. And the short-term traders index also in negative territory in July, down 1.17%, but also down for the year, down 12 basis points so far.

On the flip side, MSCI World, only down 60 basis points in July after having been up strong, of course. Still up 10% or so for the year. The S&P Global Developed Sovereign Bond index (which is a new index I had to shift to since they stopped publishing the World Government Bond Index) it’s up about 1% in July, but down 26 basis points so far in this year. And the S&P 500 is down 1% as of last night, up 14% so far this year. Anything else you want to add to the trend following part?

[00:15:49] Mark

No, other than the fact that, you know, three months don’t make a year.

[00:15:54] Niels

All right. Well, we will stay with trend following a little bit longer. We had a question in from Philip. Philip writes, I have a question for whoever you think would like to answer it. (And I was thinking of you, Mark.) You might have addressed this topic already in that case, apologies in the email.

We may have over the years, but let’s just do it again. Different trend followers will have different rules to include a new asset, but the gist of it seems, understandably, does it add enough value relative to transaction cost and opportunity cost of not holding other assets anymore?

So, usually not an obvious choice, only trade-offs. Now, his question is, or Philip’s question is, the opposite side of that. When to remove an asset?

He writes, I find it more difficult to decide when to remove an asset. Typically, one will not remove any given asset unless something bad happens or a new asset takes over from an older one. The most frequent example would be in FX, I guess.

What did trend followers choose to do with the Swiss franc when it was pegged? At least for this example, there was a clear decision by the Swiss National Bank. But when there was a soft peg or the fussy feeling that something is happening, it is hard to quantify exactly what to do. Some will argue that if vol is too low, then the asset is out. And other hints.

So, what are your thoughts on this conundrum for a trend follower?

[00:17:36] Mark

This is really a tough question. One is because, on one hand, you think that I’ll add an asset if it adds diversification, okay? But then you have to ask the question, what about performance? Past performance is not indicative of future success. And so, if, let’s say, a market has been trading poorly over the last few years, that doesn’t mean it’s going to trade poorly going forward. There was one argument I had with a large trend follower, and we had one market that was trading, and it was in a drawdown for quite a bit – five years.

We just had not made money from this market. And we sort of said, well, by this time, we say, like, we got to drop this out of the portfolio. And so, the one argument was that something’s broken, something’s wrong with this, let’s drop it out. And the other idea was that, no, it’s a diversifier, we don’t know what the future is going to hold, let’s keep it in. So, that’s one example.

The other example is let’s take the example of short-term interest rates, like euro dollars. I remember, and this was one of my great forecasts, let’s just say rates are never going to go below zero. So, therefore, and if we’re getting close to the zero bound, let’s of course drop that market, because it’s a one-sided market, we’re going to get blasted if let’s say that we go long at close to zero. Of course, what do we know what happened? Markets went below zero. So, there was a good reason to continue to hold that into the portfolio.

FX is another example. There’s a lot of central bank intervention. Central banks actually intervene because they want to punish speculators. So, should you hold and trade in trends of a market where you have a central bank that’s not a profit maximizer who wants to punish your behavior?

So, again, it certainly suggests that you should lower your exposure. I will sort of say that we always did a review of markets, and I still do the same when we look at whether a market should be added to a portfolio. And this is done on a regular basis where we’d say, is there a reason to drop markets?

One is that you want to have two-sided markets where you have hedgers and both speculators. If it’s pure speculative, then that could be a problem. Second, you’d love to have more retail as opposed to all professional traders. If it’s all professional traders, it’s going to be harder to make money.

Third, you sort of say, let’s look at minimum open interest and minimum volume. So, if I have a portfolio with a number of markets and the volume starts to fall below a certain threshold or open interest falls, I’d want to drop it out because then you could get more likely a jump risk which could work against you.

We’ll also sort of say, we want to have a market that has been established for a while. So, a brand new market that just started trading, we might want to avoid that. Again, that’s before we add as opposed to dropping it out.

And finally, we’d actually want to look at the tests for that market and we want to try to sort of see not so much if it makes money or loses money, but how it makes the money and lose the money. Is it all coming in with a single trade? Has there been a certain amount of choppiness in the behavior relative to the past? And that might have an influence on whether we should drop it.

Generally, I’d sort of say that you want to be very careful in adding markets so that then you don’t have to be posed with the problem of whether you should drop it later on. But this is a very tough question.

[00:21:46] Niels

Yeah. No, I think, I mean, I think the points that I would gravitate towards are obviously liquidity. That’s a clear warning sign.

If liquidity drops, you definitely want to get rid of the market, I think, in many respects. But I think the one about the volatility and all of those things, I think that’s easier to deal with because, you know, all managers, all large managers, they will look at volatility over different timeframes. They’re not going to just have a shorter term volatility measure that could, as we saw with the Swiss, the Euro Swiss and the Swiss franc, but also with shorter term interest rates, where clearly the volatility just got sucked out of that market for a while. And there you have to have some safeguards, some volatility flaws that doesn’t prevent you (is the word I’m looking for here), prevent you from sizing up and getting really badly hurt. So, I think there are mechanisms you can do.

But I think the bigger one, which is more challenging to do, but where I think we have a perfect example as a reminder for why you should be careful of removing markets just because they haven’t been profitable for a while. And when I say for a while, I mean, cocoa literally was not profitable, as far as I can tell, for the last 15 or 20 years prior to the beginning of January 2023. It was a tough, tough market to trade.

Of course, it would be profitable from time to time. So, I think there are two things that that example shows. One is that it can turn out to be incredibly profitable, and we know for certain managers, it’s made so much money for them that it’s the only market that they’ve really made their profits in.

But the other thing is, as you point out, is also, or allude to at least, is when does it make money? I mean, there’s nothing wrong with having a market that doesn’t really make money for you. If it makes money when a lot of other things are not making money for you in your portfolio, then it adds some diversification value.

So, I think a lot of trend followers will be aware of these things. That’s why I think that very few markets get taken out of portfolios, except for liquidity concerns. That’s a pretty clear black and white issue, I think, for most managers.

[00:24:08] Mark

Right. The cocoa example is probably the poster child of something that was choppy. There will sort of say that you have governments that have commodity boards that actually can manipulate price.

So, there’s a whole set of reasons for why you wouldn’t want to trade cocoa. And then along comes this big supply dislocation. And if you didn’t have cocoa in your portfolio, I think you could sort of really see the difference.

I follow some indices like the old Managed Futures Index from Credit Suisse, and it doesn’t include cocoa. It’s, I think, a very well-constructed index, but it underperformed relative to a lot of CTAs who did have cocoa, and that was just because of this explosive return.

The other thing you find is that when volatility gets lower, okay, because let’s say maybe it’s, I don’t want to say it’s manipulated, but there’s some reason for why there’s less reason for markets to move. Sometimes people say, if I’m volatility targeting, I will increase my exposure in those markets that have low volatility. And then if, let’s say, there is a spike in volatility, I could be wrong sided on that market.

That would be the example of a manipulated market like the Swiss franc. Volatility got lower, so what would you do?

Well, you know, I can’t make the same amount of money. I should increase the size of my positions. And that was the worst thing you could possibly do when, let’s say, the peg was readjusted.

[00:25:52] Niels

Yeah. And I think this is where experience comes in, right? Because I think most experienced managers would have a volatility flaw knowing full well that you shouldn’t be fooled by the fact that it can go to very low levels of volatility for a year, maybe two, but it doesn’t mean that that is what you should size your position after.

But let’s not get into that because the list of topics that you brought along is, as usual, very long and extensive and interesting, of course. So, we’ll see how many we can get through in the next half an hour or so. But the first one kind of relates to what we’ve talked about, right?

Because you wrote a paper not long ago in the Journal of Investing called I’ve Never Seen a Bad Backtest. And I think it got a lot of exposure, impressions, and so on and so forth. So, of course, I’d love for you to talk through your thoughts about that.

[00:26:47] Mark

Well, the reason why we’re talking about this, I think that you had a question about whether people actually believe backtests and how valuable are they. And I probably would sort of say that for someone who has done a lot of backtesting and have presented it both as a manager with experience where we’re already running money as well as being a relatively new manager, I’ve gone through the throes of having people say, I don’t care about your backtest. I’m not going to invest money in you, and I don’t like it.

I’ve also been as an investor, into different managers, and I’ve had to look at backtests. I would sort of say that they’re very valuable. That doesn’t mean that that should be the reason to invest, but they’re very valuable, not so much from a performance and sharpe ratio, because I’ve never seen a bad backtest in terms of performance and sharpe.

But when you talk to the manager and you say, tell me how you went through and did your backtest, tell me the assumptions you used, tell me how you accounted for transactions, you get insight into the manager in terms of how careful he is with his research, how careful he is at being able to account for all of the realism in the market.

Also, people sort of say, well, I can come up with a model, and it won’t take me too long, and I’ll show that it’s a fantastic performance. That’s not what you’re looking for when you look at a backtest.

What you want us to understand is, how does someone build a system, not a model? And there’s a difference between the two. I can always come up with a good model.

The question is, can I take that model and then apply it across a large number of markets, build portfolios, and then build a system? And finally, when you think about backtests, what you have to do is move from backtest to what I call the factory. And so a backtest, in some sense, is like the scientist who’s in the laboratory.

But then, when you actually want to build the system that you’re going to trade, then you have to take off your scientist hat, you have to put on the hat of an engineer, and then you have to say, I am building a factory. And a factory is about a system that has to look at what the inputs are, what’s the output, when is the timing, what are my costs, how do I form execution?

And when I talk with people about their backtest, or show backtest, I want to talk to them about how we move from that back test and how it’s realistic, so it can then be exploited as a factory.

[00:29:51] Niels

Yeah, no, I like that analogy. And I think that is actually an underrated point about what you can use a back test for.

I think you’re absolutely right that most people will simply just take it from a performance point of view, and maybe a drawdown point of view, and so on and so forth. But everything that goes into creating is actually the really interesting part. So, I think that might surprise a few people.

[00:30:20] Mark

If I could tell you one small tale, in terms of the differences of opinion about backtest, I was working in midtown Manhattan. I went to one manager, showed him some backtested numbers because we were using a new model. He looks at me, he is like, look, I’m never going to invest with anybody who doesn’t have a sharpe ratio above two.

So, ours was, you know, the back tested number was not above two, clearly. So, we left the office, you know, sulking down the street. We have to go to our next meeting.

We go to the next meeting and he said like, wow, this is it. I would never invest in anything that has a backtest that’s above, you know, 1.25. I can’t believe that you could produce those kinds of numbers.

So, in the same midtown location, two different offices, one person said, I’ll never invest with someone who has a sharpe ratio that’s below two. And then another one said, I’ll never invest with anybody who has a sharpe ratio above two.

[00:31:36] Niels

Above 1.25, you said?

[00:31:39] Mark

Yeah. So, We’ll sort of say that there’s a lot of different opinions on that. I would say it’s the process of how you do your research, which is more important than the actual numbers you generate.

[00:31:55] Niels

Yeah. Yeah. Very true.

The next point you brought along is more related to systematic trend, as far as I can tell. And again, has something to do, I guess, could be related to, well, it is to do with systems building and so on and so forth. Simplicity versus complexity.

Where do you want to go with that?

[00:32:16] Mark

Well, this was another question that you said you had from our readers when they talked about, well, it’s relatively easy to build a model. What’s so difficult about this?

When you think about it, building a model (and this is related to our backtest) is relatively simple. But it’s the actual construction of the system, which is where the problems come in. And what you have to think about in this, is that any portfolio you build has to be related to three components.

One, what is the process for return generation? Two, what is the portfolio construction? And three, what is the risk management?

So now that I’ve done more evaluation of managers, when someone comes in and they sort of talk about that, well, I’m a systematic trend follower or systematic managers, then what I really want to do is I’ll hear the pitch, in terms of the total system that they’re presenting. But then I want to sort of say, let’s break it down into those three components. Tell me how you generate signals for your longs and shorts. And in some sense, , if it’s more regression based or it’s a model based, then there’s a potential for parameter instability.

In some sense, you actually like simple rules. Second is that then you say like, well, how do you construct the portfolio? Because that’s where you’re getting the only free lunch available in finance, which is diversification.

And third, you’re going to say, what’s the risk management? The risk management is, if you think about it, is that what do you do when something goes wrong? And so, the equivalent would be, is the Mike Tyson quote that I like, he says, like, ‘everybody has a plan until they get punched in the nose’.

So, if you get punched in the nose, what are you going to do to protect the principle of investors? And, and when you think about it, that market goes against you, you start taking losses, that’s the equivalent of getting punched, punched in the nose. Well, then the response, in most trend followers and a lot of managers, is that I have a stop loss. I walk away.

So, in some sense, that’s your fail-safe. A good model should incorporate risk management in the portfolio construction and decision rules.

You determine when to get long, but then you also have a determination of when to exit. So, the stop loss is only there because something has gone wrong, but there’s nothing wrong with having that. But that’s the idea of your protection.

So, when you think about systematic trend following, they say, is it easy? Well, it may be easy for any one part, but you have to bundle it together as a system.

[00:35:22] Niels

Yeah, very true, very true. Good points.

Now we’re going to jump a little bit away from trend. Something you’ve been writing about a couple of times this month because you write a lot of blog posts, more blog posts than most, I think. I noticed you wrote more than one per day in June alone.

So that’s pretty, pretty amazing. Anyways, the question you pose, and you can take us into this topic, and that is whether people’s investors should prefer stocks or bonds.

So, tell us a little bit more about that.

[00:36:00] Mark

Well, there are two things that have really been driving this and that is, one, what’s the environment right now? Because we’re seeing an environment where stocks seem to be overvalued. So, when we look at it on a PE ratio, certainly, when you look at the mag six or seven, depending on what number you want to have, they’ve well overextended.

So now people will sort of say, well, what should I do for the second half of the year? Should I get out of stocks, which seem to be overvalued? But then at the same time, I’m hearing that the Fed is actually going to be lowering rates, possibly as early as September, which means that, well, if there’s lower rates, I should hold my equity performance. And on the other hand, people say, well, if I’m going to have lower rates, lower inflation, slower growth, and they’re lowering rates because there’s slower growth and lower inflation, I should hold bond. This should be an easy trade.

So, which one of those two dominate? We’ll sort of say that both of them are, it could be compelling reasons to invest in both stocks and bonds. When you find that that has a problem that both of them might be, then you have compelling reasons.

You could have a situation that they could become correlated. And if they become correlated, well, then your overall portfolio will actually increase in risk. So, the question comes in as, well, if I can’t answer the valuation question, then maybe I should just follow the trends in price and let the market tell me or disclose to me what I should do.

And that’s probably the conclusion that I came to is, well, I can find compelling reasons to get out of stocks and hold bonds. And there are also reasons for why, especially in a lower interest rate environment, bonds will do better, but stocks maybe will do even better than bonds. Perhaps the best way to approach it is to look at this from a trend following perspective and let the prices, themselves, tell you what you should do on terms of your action.

[00:38:17] Niels

It’s funny how all roads lead to trend following, Mark, but I’m not going to disagree with that point. Interestingly enough, you mentioned that if the correlation between stocks, bonds go up, I noticed that the one year rolling correlation is actually at the highest it’s been for a very, very long time, at the moment, which again maybe people haven’t been too worried about that because both stocks and bonds, up until recently, went higher, at least for a while.

But it certainly exposes their portfolio from a risk management point of view compared to what they may think they have in the portfolio.

[00:38:58] Mark

This is an area that I think for a lot of seasoned professionals, it’s not an issue. But I think for a lot of investors still, is you focus on volatility, volatility; you look at the volatility of every market. You also have to be looking at this in terms of a covariance matrix. We have to look at the off diagonal, which is the covariance or the correlations.

And if correlations change on you, and if correlations, especially if you have the classic 60/40 portfolio, if it turns positive, you’re really getting get a lot more volatility in your overall portfolio. Okay. And then you sort of say like, I actually took lower return by holding bonds because it was a great diversifier.

Now, if the correlation in stocks and bonds goes positive, then why am I holding bonds if, let’s say, their overall return may be less than equities? So, then you have to say, well, then I’ve got to really start to think about valuation because then it’s a relative return situation, not a diversification question.

[00:40:16] Niels

Yeah, yeah. Again, actually, that is a very good segue perhaps to the next point you brought along. It’s something that Alan and I briefly touched on last week, but I think you might have some different perspectives.

I’d love to hear your opinion about it, but it is a paper that came out as a kind of a joint piece between the Sovereign Wealth Fund in Singapore, GIC and JP Morgan on the topic of building portfolios. So, can you talk us through, from your perspective, why this is a good paper to read?

[00:40:56] Mark

Well, it was a great paper because it was just about to sort of say that their framework between a large Sovereign Wealth Fund and JP Morgan on how they look at alternative investing hedge funds. Okay. How do you classify them and how do you sort of determine how do you compare one strategy versus another?

When I first heard it, I said like, okay, this seems really interesting, and I want to have a broad base of interest. How do people compare different hedge fund styles? So, then I started reading this, and as I read more and more into this, I said like, this is actually probably one of the best explanations for why you should have trend following that I’ve ever seen. So, it wasn’t the intent of the paper.

It wasn’t your objective. But when you looked at all of the evidence that they have, that they actually came to the conclusion is that probably the one thing that you need to have in your portfolio is managed futures.

[00:42:06] Niels

So, can you visualize that? Can you visualize that? Because that is interesting. I mean, you’ve read a lot of papers in your career.

So, what is it that makes their argument so compelling from our point of view?

[00:42:21] Mark

They first sort of said that all strategies, all hedge fund strategies can actually be put into four buckets or four types. We’ll sort of say (and we’ve had this discussion over and over again), managed futures, all managed futures are not trend following. So, they can be classified differently within that strategy categorization. So, first they said there are four different types.

There’s are loss mitigation strategies, equity diversifiers, and then, we’ll call it, equity compliments, and equity substitutes. Okay. So then from that, they say, well, to look at those four classifications, we have to look at three characteristics.

Okay. And the characteristics they look at, they say like, well, one, you want to look at what’s the correlation with equities. Okay. So, a good diversifier is, is going to be someone that has low correlation, right? An equity substitute would probably have high correlation and something that has low correlation would be closer to an equity diversifier.

So, in that perspective, we could classify all hedge funds based on their correlation. The second characteristic is their drawdown mitigation, and this is relative to equities. So, if you have an equity substitute, okay, and it has a high correlation with equities, it’s probably not going to provide a lot of drawdown mitigation. Okay. So that’s the second characteristics we’re going to look at.

The third characteristic is that after we account for beta, okay, what is the alpha that’s left over? Is it an alpha generator? Okay. So, I said, seems to make sense.

We’ve talked about each one of these three individually, and they say, well, let’s put it all together and let’s look at what are the strategies that actually could give you all of those characteristics? And you say like, oh, lo-and-behold here, low correlation – trend following seems to do that. Okay.

Check. That’s check one. So, if we look at the spectrum across correlation, you know, trend following is probably on the low end.

Okay. Let’s look at draw down mitigation. Okay. Now that’s sort of the crisis alpha. Let’s take a look at what are all the different hedge fund strategies there. Lo-and-behold, trend following seems to be giving a drawdown mitigation. Check on that.

Let’s look at alpha. And we sort of say like, well, no one has probably sort of really talked about trend following as an alpha generator, but at the same time we do use the term crisis alpha.

You know, I think that it’s a mixed metaphor, but that’s a different discussion. But what they did is they looked at the beta line and then they looked at all the different strategies and they said, well, does it generate positive or negative alpha relative to its beta. And lo and behold, managed futures is probably number two. Number one was global macros. But you’ve got positive alpha in that sense.

So, then you sort of say like, given their characteristics of what you want to look for and the characteristics of what you think you should have from hedge funds, you say, you’d like to sort of see, and you don’t get it for everything, but you want to have low correlation, drawdown mitigation, and alpha.

And if you look at what is the strategy that seems to fit within that? It’s probably trend following on all three of those counts. So even though they didn’t say that this is our objective with the paper, this is the conclusion that I drew.

[00:46:22] Niels

Of course. Absolutely. And now I haven’t read the paper myself, but if they didn’t draw that conclusion, that seems pretty clear when you explain it like that. What was their own takeaway, so to speak? And, and did they have a different agenda, so to speak? Because again, yeah, no, there are sovereign wealth funds out there, for sure, who have large allocations to trend following and CTAs. We know that for sure. But JP Morgan is not a firm that strikes me as being a big promoter of trend following. They may have some exposure to it, so, what did they plan to get out of the paper that was different to your conclusion?

[00:47:07] Mark

Well, I think that their intent was not to sort of make this a paper about trend following or the benefits for managed futures. I think that they were just sort of saying we’re developing a framework. Our framework is saying that we’re going to categorize all hedge fund strategies according to risk mitigation, diversifier, compliment, substitutes.

But then the data sort of spoke for itself, like, well, if you want an equity substitute, you’re not going to pick managed futures. Okay, if you’re looking for risk mitigation, definitely you want to try trend following.

So, I think they’re trying to sort of say, we’re trying to have an agnostic view. We just want to provide a framework and then it’s up to you, as the reader, to sort of determine what you want to try to get from this. And too, depending on your perspective, you’ll get a different conclusion, but, you know, I think that it just jumped out at me, given that when I walked in to do this read, I didn’t have a view that this is what I thought this was going to be a paper about trend following. It’s only at the end that I said, you know, this is a great story. It’s a great argument for why you should hold trend following.

[00:48:39] Niels

Yeah, I like that. Hopefully the paper will give some inspiration, to all the people who, will be diving into it. Now, from what the heart is full of, Mark, to something slightly different, namely a book that you came across, which incidentally, I came across as well.

I haven’t read the book, but I’ve heard an interview with one of the authors, Philipp Carlsson-Szlezack. I’m not entirely sure how you pronounce that, okay. The other author is Paul Swartz, and the book is called Shocks, Crisis and False Alarms: How to Assess True Macroeconomic Risk.

And interestingly enough, I have suggested to our friend and co-host, Kevin, that he invite these authors to his series, which is all about authors and new books. But we’ll see if that works out later in the year. So yeah, it was interesting when you put it on the agenda thing.

Yeah, I want to talk about this. I’d love to hear what your thoughts are.

[00:49:57] Mark

I’m not going to steal the author’s thunder if, let’s say, you could get them onto your podcast, which I think would be a great idea. But, you know, I think that they said this is it.

Well, what we really have to do is we have to worry about how do we think about shocks and crisis? And that sometimes we expect that there should be a shock or a crisis, but then it turns out to be a false alarm. And the problem comes in that we often use judgment that is wedded to a specific model or school of thought.

Okay. Perfect example is, that we had a big increase in money from QE. How many people thought that we were going to have rampant inflation right after the last great financial crisis? And it just never occurred. If anything it’s that we are fighting an issue that we couldn’t even get up to 2% inflation. So, if you were wedded to a specific model or thought, then you actually could have a false alarm and you weren’t able to truly assess macroeconomic risk.

Similarly, is that we have a lot of macroeconomists or doomsdayers that make their bones by suggesting that the world is going to come to an end. And what we find out is that, for a lot of things the economy has actually done better than what we expected.

Look at COVID, you know, I think that a lot of people thought that the global economy had come grinding to a halt. There’s no question. It was an extremely sharp reduction, but I think what was really surprising was how fast we came out of that.

How fast to do it – there was a reversal in that. And so, I think the argument was that you shouldn’t put stock in a specific model. You should ensure that you have a debate about assessing alternatives and that you shouldn’t be swayed by sort of doomsday thinkers.

And what that really sort of came out, to us, is that when we talk about this from the context of our podcast, that I really go back to the basics to say that prices are primal and use prices as your sort of your null hypothesis. I like to sort of say, start with a null and then maybe if you need to deviate, you could think about it, but sort of let the prices speak for itself, to say, this is what the market is telling me. And then from there I can sort of say either I could make an adjustment, or I just follow what the prices are going to do.

And, and I think that they would sort of say, don’t chase the numbers. You know, let’s say we get the PCE that came out today. There are a lot of people are going to spend a lot of ink telling you about PCE, when in reality what you don’t want to do is, what’s the price action that you see in markets in the next couple of days? And that’s going to be what the real driver is.

So, I think it was a refreshing, different view on macroeconomics. It doesn’t say that you should be a trend follower, but it also sort of said, is that you should be a healthy skeptic of whatever economists that you might be following or you read about.

[00:53:08] Niels

Yeah, no, I think that’s perfect, actually, that’s a very good, appetizer or teaser for if we’re able to get one of the authors on the podcast later in the year. I thought actually, as well, when I heard an interview with them, that it was an interesting topic and a different take.

Now, Mark, as always, we are close to running out of time, but we still have so many topics left. So, I’m going to let you choose which ones. Maybe we can tackle another two before we wrap up for this week. So, maybe you choose two that you think are really relevant, interesting, from where we are right now.

[00:53:57] Mark

Well, I hate to rain on the AI, we’ll call it machine learning, parade, but I think that’s worth spending some time on, especially when we talk often about systematic model building, we talk about trend following, we talk about quant work.

There are some interesting papers that have come out that give us a reason to pause about some of our thinking about machine learning and AI. One is that there is a piece (and you can get it from my blog) that they looked at LLM (large language model) models, and they tied it to time series.

So, they say, if I use an LLM model, and I apply this in addition to time series modeling, can I get improvements? And what they find out is that’s not the case.

Well, I thought that everybody is spending time on LLM models. Everybody’s working on this. The tests that they performed are very specific, but it did do a comparison at a number of different approaches.

And they said that LLM models may not be able to add as much value as what you think. I think that what I’m hearing from a lot of researchers is that they’re spending a lot of money on large language models. I’m not sure exactly what the output they’re getting or whether it’s been successful or not.

So, I think it just causes a little bit of a pause to say that model building is a lot harder than what you may think. Second is that there has been another paper that looked at machine learning across the business cycle. And they say like machine learning does pretty well during the up parts of the business cycle.

But then when you hit, the down drafts, when you hit a recession, actually the market performance gets a lot worse. And when you think about it, if you’re training a model, you have only a few periods of recession. And then most of the periods are during expansion.

You should expect that it’s going to be trained on what the most data you have. And so, it might actually do a lot worse during a recession. Again, you know, time to pause and sort of say that maybe machine learning doesn’t have all the answers just yet.

[00:56:26] Niels

And also I would just add that our industry is one of few industries that really have gone through many “technology revolutions”. I mean, from the introduction of the internet and all of these things. And every time, at least in the 30 plus years that I’ve been doing this and I’m sure people would have heard the same before that time, people were quick kind of state that this would have a negative impact on trend following.

Why wouldn’t it? I mean, suddenly all everybody has all this information. Now everybody has all this power of technology, power of some sort, but of course it, so far at least, has not proven to be the case.

One should never say never, but I think there’s a lot of… experience is hard to backtest. Let’s put it that way. I’m sure also an AI would find it hard to put on the experience, the real experience that many of these firms, our colleagues, in the industry have gone through.

[00:57:32] Mark

There’s no question that there’s, there’s a reason to do a lot of work on new techniques and there will be ways that you can exploit markets using new techniques. But the answer is not going to be digital. Like, we just need to switch, and it’s an on off switch that we should go from one set of models to a new set of models.

So, I’ve really spent a lot of time thinking about this. And I said, why do we search for models and why do we go through this? And I go back to sort of the idea of information theory. When I think about this, all the prices that we have, they’re public information, okay? You had some great announcements this morning. It’s public information.

In efficient markets public information, all that information is going to be discounted in prices almost immediately. So, what is the purpose of building models? And the purpose of building models is to take public information, which could be past prices or macro data, and then we want to turn it into private information.

Private information is that we contort it, we mold it, we switch it into something that we find useful, that’s private, that other people don’t have. And then I could sort of use that private information to tell me something about what might be going on in market prices today, which means it could be, I’m trying to, I’m not saying it’s a forecast. I could just say, I’m doing a better job of being able to explain where prices may be headed.

So why do we want to use a new technique? Well, a new technique is another way in which we could take the public information that everybody has and mold it into some private information that might give me some insights that other people don’t have. So, in an efficient market, we’re constantly looking to try to say like, how do we sort of create that private edge?

And we sort of say, if everybody’s using the same data, so for example, we were talking about cocoa markets is that everybody has cocoa prices. So, what can we do with those cocoa prices? That’s different than the next person that could tell me something that other people don’t know.

And that’s the process we’re going through, but because it’s hard often, and we’re all using the same information to start with, it’s not clear that just because we want to contort it into some private information, it’s going to work.

[01:00:16] Niels

Yeah, actually that’s a great way of explaining it. Actually, I think.

[01:00:20] Mark

The final thing that I want to talk about is, and I’ll leave this as a conversation that we should have more of, and there’s been some research done, and we’ve talked about this, and other people have talked about this before, and this is in the behavioral economic area. It’s called the law of small numbers.

The law of small numbers says that we create what we call the gambler’s fallacy, or the hot hand problem is, is that we often extrapolate from a small sample on what we might think what will happen in the future, as opposed to using a very large sample, okay.

A perfect example of that the hot hand problem was, well, if you have a basketball shooter and let’s say he’s made three, three point shots in a row, so what do you think we should do? And you say, “Hey, give him the ball.” He’s got a hot hand, you know. Let’s extrapolate from that small sample and assume that he’s going to continue on and do the same.

Okay. Now, on the other hand, you have the gambler’s fallacy. This is that, let’s say I’m playing a game and I’m flipping coins and I’ve had five heads in a row.

What do you think I should do on my next bet? So, a lot of people would sort of say like, well, by definition, if there’s been five heads in a row, it has to come up tails next, because, you know, it’s a coin. It should be 50/50 over the long run.

So therefore, I should, you know, mean revert yet when you say what’s the chance of having a tail after five heads in a row? It’s still 50/50, it’s still a flip of the coin. So, what happens is, the problem in behavioral finance has been, we sort of say like, I could use this gambler’s fallacy in a hot hand problem to explain why I should extrapolate. It also could explain why I should mean revert.

So, there is some mathematics about like, whether you assume that you know the distribution or not. If you know the distribution, then you probably, you’re always going to be a mean reverter. If you don’t know the distribution, then you might sort of follow the hot hand.

But when you think about all of our trend following and all of our models, is, is that oftentimes we have to make a lot of decisions off of very limited data. So, we’re constantly faced with the law of small numbers in a lot of the decisions you make.

So, when we think about your trend following model, or you think about any quant model, I ask that all of your listeners, between now and our next call, to say, think about your sample size, think about the distribution and see whether you think that you might be falling into the problem of the law of small numbers.

[01:03:30] Niels

That’s a great way to end the conversation. As always, Mark, really insightful and some different things I think we’ll leave the audience with this week. Now wrapping up this conversation, of course, hoping that everyone enjoyed it.

Please take a few minutes to go and leave a rating and review or share the podcast with your like-minded friends as it is simply the best way for us to grow the podcast. That is from being recommended by someone who’s actually listening to it. So, we would greatly appreciate that.

Next week I’m joined by Cem. No doubt we’re going to have a conversation about the crazy world we live in right now.

And also maybe as a follow-up on the conversation we had with MacroAlf, last week, or this week, actually, I think it was released. So, lots of good stuff to dig into. And of course, as usual, if you want even more trend following information, we are publishing quite a lot on the website every week.

We do a summary of the world in trend, so to speak. And we just published our monthly update that Rich and I work on together. So, there should be plenty of stuff to keep you entertained for sure. Besides the Olympics, of course.

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

[01:05:12] Ending

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

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