210 Systematic Investor Series ft. Rob Carver – September 17th, 2022

[00:00:00] Introduction

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 strategies. Welcome to the Systematic Investor series.

[00:00:24] Niels

Welcome, and welcome back to this week’s edition of the Systematic Investor series with Rob Carver and I, Niels Kaastrup-Larsen, where each week we take the pulse of the global market through the lens of a rules-based investor. If you’re new to the show, I hope that today’s episode will trigger your curiosity enough to check out the back catalog and listen to the past episodes that you may have missed.

Like my conversation with Rich last week, where we did quite a deep dive on how to use an ATR based stop-loss strategy, and much more. I would also encourage you, if you’ve missed the Wednesday episode, to go back and listen to Oded Galor, professor in economics at Brown University, discussing how we can explore and uncover new ways of thinking about the economy, based on his book, The Journey of Humanity, The Origins of Wealth and Inequality. It’s a really great conversation and very eye opening when it comes to the topic of inequality. 

It is great to be back with you, Rob, after a little while, how are you doing? How have things been? 

[00:01:20] Rob

Yeah, I’ve had a good summer.

I was actually over your side of the channel, in Germany, in Switzerland, Austria over the summer, which was nice. I’ve had quite a nice relaxing summer. Of course, the news here in the UK has been dominated by the death of our Queen, who died about a week and a half ago now.

I think it’s a kind of funny situation to be in, because obviously it’s been a very long time since we’ve had a monarch die. I think the royalty in the UK is a bigger deal than a lot of other countries that still have Kings and Queens.

So, for example, in some of the Scandinavian countries, they have Kings and Queens, but it’s not such a big part of the national fabric. So yeah, it’s been a bit weird actually, a bit of a weird time. I think we’ve got the funeral on Monday, and I think after that we’ll probably feel a sense of closure and be able to move on. Yeah, it’s been a strange week and a half, at least from my perception. 

[00:02:23] Niels

Yeah, indeed. I was actually spending the last week in London, so I definitely sense the mood that you’re talking about. Also, it was very impressive to see all the people lining up to go past the Queen’s coffin and pay their respect with long, long, or many, many hours of queuing up.

But of course, the Brits are very good at queuing up. 

[00:02:47] Rob

I was going to say, that’s really two of our great national institutions, the monarchy in the queue, there in one place. Yeah, in a way, a very meaningful tribute to the Queen to have us doing our national sport on a kind of grand basis.

[00:03:04] Niels

So, you mentioned the Scandinavian Queens and Kings, actually in Denmark, my birth country, we were actually celebrating our Queen, her 50th year, or anniversary on the throne as Queen. Yeah, those are very strong ladies around in that role. In terms, speaking of que, we’ve got, queued up, quite a wonderful line of questions, very long, because it’s been a little while since you were last here.

We also have got a few topics that are fun to dive into as well. As always let me just acknowledge and thank every one of you who took the time out to leave a rating and review on iTunes and Spotify. It is so nice to see them. I certainly read all of them. So, I’m very grateful for that.

Now, if you were (or if you are) a traditional long-only investor, perhaps the best thing that can be said about this week is, it’s over. Traders came into the week, optimistic with hopes sustained from the August employment situation. Many economists anticipated that the CPI would finally show a down tick, and from the relentless pressure on interest rates, and that would finally moderate. 

That is not exactly what happened. Instead, the CPI printed an unwelcome up tick across most categories and reinforced the need for higher interest rates. While the Fed sat on the sidelines, post report bond traders acted decisively, pushing the two-year note, 20 basis points higher on the day and, ultimately, ending the week 32 basis points higher. 

The worse than expected inflation report towards the equity narrative, that the Fed is close to an equilibrium point in overnight rates had Equity investors interpreting that as meaning that they are close to their first rate cut. All of which was seen as a reason to buy stocks, but instead the S&P 500 tumbled more than 5% for the week.

Then after the close of the bail on Thursday, FedEx, preannounced their earnings to disclose an absolutely abysmal quarter based on what they termed as a global slowdown in economic activity. According to Deutsche Bank, or one analyst at Deutsche bank, I should say, the FedEx miss (relative to their expectation) was the worst in 20 years. The market hopes that FedEx was the exception, but as they say, where there is smoke. Next week brings more data from the housing market, and more and, importantly, the FMC decision. 

Another 75 basis point rate hike is very likely with the post meeting press conference offering clues as to what we should expect will come next. If we see another 75 basis points in November, at that meeting, this will bring the overnight rate corridor to 4%.

Speaking about inflation and interest rates. Let me just mention that one of our previous global macro guests, Jim Bianco, was out saying on Macro Voices, this week (I should say, for context, that they were discussing a lot of the historical issues regarding the lack of inflationary process we saw for the past couple of decades), but he went on (and I’m just quoting a little bit from the conversation), he went on to say, so those were the bedrocks of why we had no inflation for 40 years and they’re all disappearing. That’s where I think people are struggling. They all say, no, this is just a one-time thing because of the pandemic. We’re going right back to the 2020, or 2019 situation. But I don’t think we are, I think we’ve got to start to realize that there are secular changes that have come into play, and yes, we might not see a persistently 9% inflation world.

If we do, we’re in a world of deep hurt, but we might be in a persistent 4% to 5% inflation world. That’s not a good place to be because that means neutral on interest rates is 4%, it’s probably 5%, or maybe 5.5%, and restrictive on interest rates is something higher than that. Those numbers, the markets are just not ready to accept. 

By the way, all things that the Gold Box has been arguing for, like inflation and geopolitical tension, have all come to pass. They have been absolutely right. But gold is now in a bear market. So, it just goes to show that you can be right and still lose money. That’s another reason why trend following doesn’t believe in prediction. 

Anyways, let’s talk about what has been standing out for you. I don’t know how much you’ve actually been able to follow the markets on your holiday, but I am curious to know what’s been going on from your radar’s point of view. 

[00:07:50] Rob

Yeah, it’s funny actually, there’s a guy I follow on Twitter who’s also a systematic trader and he actually recently put a series of posts up saying, oh, wow, apparently there are big things going on in the market. I hadn’t even noticed because trading systematically, unless you actually are following the business use (and I obviously haven’t been), you don’t necessarily notice big things going on if they’re not directly affecting your positions, and sometimes even if they are directly affecting your positions.

Anyway, if I look at the last three months, which is roughly how long it’s been since I’ve been on the podcast, then basically in a kind of modest draw down and then hit a high-water mark in early June, dropped a bit in the first half of June, and I’ve pretty much been flat lining ever since.

Over the last three months I’m down at 5%. If I just call a quick look at where those losses have come. I’ve lost some money in crude oil, gas, and wheat, soybean, soy oil. So, I’m in a little bit of a reverse of some of these very strong commodity trends that we saw in the first half of the year.

If I look at the winners, actually my biggest winner, interestingly, is the Brazilian/U.S. Dollar rate. But I’ve also done well in Yen/Dollar, and Euro/Dollar. There was good performance there from currencies. Also, some positive performance from bonds and fixed income. I’ve got a little report that I do, which looks at market moves over the last three months and normalizes and by stand deviation, just to get a feel for how surprisingly they are if you like.

Then it’s a nice little report, because I scan quite a big range of markets. It sometimes shows up markets that you wouldn’t necessarily think about. So, for example, oats are actually the biggest loser over the last three months but also gasoline, and heating oil, red wheat, iron hasn’t done so well.

Now the best market, interestingly, that I track is Ethereum. There’s been the news recently about this big merge. We may get time to talk about that a bit later. We’ll see how it goes. So maybe some of that movement was an anticipation of that because that’s certainly… I’m not seeing Bitcoin up there, which is the other crypto future I track.

So that’s kind of interesting, but also a couple of U.S. stocks. You may think, stocks have done very badly, but interestingly, a couple of sectors have done very well. So, I’ve got U.S. utilities and U.S. discretionary consumer are doing very well. I think, as we look at the big market moves, we don’t often see what’s going on underneath that.

So, FedEx is probably a good example, because I would imagine that FedEx was a company that did very well during the pandemic, because we were all staying at home, probably getting a lot of stuff delivered to us. The delivery companies were probably doing very well. In the same way that firms like Peloton were doing very well during the pandemic, and Netflix, and then afterwards have suffered as we’ve reverted back to our previous habits of actually going out and buying stuff, rather than just getting it delivered. So maybe that’s the FedEx story and maybe there are still sectors of the economy that are going to do well.

In terms of, if you’re a stock picker, then these times when the environment is cycling from one sort of state to another, are often times to pick up really good value stocks, because different sectors go in and out of fashion as the economy cycles.

So yeah, so I guess the main market news for me, over the last few days, has been the Pound. The Pound actually dropped to a low, it’s not been at 1.16, let me see if I can get this right, 85 years – since March 1985, it actually dropped briefly below 1.14 – a big figure. Although I think it closed slightly above that yesterday. 

In 1985, so Neils, you were a little bit older than me, but in 1985, I was just… Actually, in March of ‘85 I was just a little 10-year-old boy. You can try and imagine what that looked like. We talk about returning to an environment of higher inflation, and the early eighties and seventies obviously were a time of higher inflation, but they were just a very different world.

I had a quick look, for example, at the music charts in 1985, just in March 1985. So, to just quickly run you through the top five, just to give you a feel for how different life was, So we had Private Dancer, by Tina Turner up there; The Heat Is On, by Glenn Frey, which was a great song; One More Night, by Phil Collins; Material Girl, by Madonna, and number one (when the Pound was lasted this level against the Dollar, was Can’t Fight This Feeling, by REO Speedwagon. I can see some nostalgia creeping into your face as you hear me name those songs Neils.

The other thing that happened roughly around the time when the Pound was last at this level, was the Oscars. The Oscars are quite interesting because do you know, in 1985, which film was the standout performer at the Oscars? 

[00:12:59] Niels

Oh, that is a good question. No, I don’t actually remember.

[00:13:05] Rob

I’ll give you a clue. It’s a film about a German musician. So it’s Amadeus. The film Amadeus. 

[00:13:18] Niels

Which, there’s a reason why I don’t remember that. I don’t think I saw it. 

[00:13:20] Rob

Exactly, I don’t remember it being an absolutely brilliant film. In fact, the main thing I remember about it was the theme tune, which was this kind of rocked up Mozart song, which, if you don’t know that song, then go and look it up on your favorite music streaming service. Don’t blame me if you hate it, because it really is an ironically poor eighties song. I was very surprised to see that the film was actually better than, for example, The Killing Fields, which is another film in those Oscars. It was actually a much better film. So maybe with a pattern of this level where we’re reaching back to a time when music in my opinion was better, but I’m not sure about the films. The quality of the films, maybe they weren’t so good.

But anyway, back to the serious business of trend following. So, one thing I like to do is look at my current risk. My current risk is still pretty low. I’m very much in this position of having made a lot of money in the first few months of the year and then have a draw down; reduced positions as trends become unclear. I’m husbanding my capital if you like. So, my risk is currently 7% annualized standard deviation, versus a target of 25%. So that’s obviously quite low. My biggest risk is probably in bonds. So, I have a reasonable size short-bond position. I also have reasonably sized long positions in AGs and oil and gas, although not as strong as before. I’ve got a sort of modest position, short metal, short vol. There are not really very clear messages coming out of the portfolio at the moment, because there aren’t really any strong trends out there. My biggest shorts in U.S. 20-year bonds, my biggest long is in gas, Henry Hub gas.

The kind of story of rising interest rates, and increasing inflation is still there but not in such a strong way, at least in the way my portfolio’s currently positioned. So yeah, very much a time of wait and see, I think for my portfolio, at least there’s nothing in there, that’s a big, strong kind of outright position.

I don’t actually, for example, have a position on in Pound/Dollar. So, no strong opinion there, or you could argue that I’m being patriotic by not betting against my country at this time of national mourning. 

[00:15:35] Niels

Yeah, thanks for that sort of historical walk down memory lane, that was interesting. I guess it goes to show, when you mentioned that last time the Pound was here, was 30 years ago, or thereabouts or maybe even, yeah, more than 30 years ago. It just goes to show, I think, that markets tend to go in cycles. The other thing that I think is interesting is that, as far as I’m aware, if you think about socialnomics, I think they believe certainly that the mood of the time is reflected in the art. In the music, in the films that were popular at the time. That kind of tells you where we are heading now. I have no idea who was last winning the Oscars, or the charts in this year. I don’t really follow it that closely. But maybe there’s something to it. Obviously as trend followers, we clearly believe that markets will continue to have these bigger cycles.

Maybe this is actually something that is interesting. I think a lot of people (and just to digress a little bit from our normal agenda), I think a lot of people, because of trend followers generally quite strong performance in the last say a couple of years, but in particular in 2022, I think a lot of them feel that maybe they missed the move if they hadn’t been invested in trend following.

It’s certainly a question that I have come across recently. But on the other hand, going back to the cycles, if we believe in cycles and a lot of money has been made from being short bonds and fixed income by trend follows for sure. But if we believe in cycles, and we believe that the last interest rate cycle literally just ended like a year ago, after 40 years of heading down. This really is just the beginning then. 

So, in that sense, you could say no, absolutely not. It doesn’t suggest that you’ve missed anything. On the contrary, we could be looking at quite a long period of time like Jim Bianco is saying, much higher interest rates, much higher inflation, which at least historically has given more opportunities for people who are betting on divergent market conditions, rather than the stable carry regime. I don’t know if you have any thoughts about that.

[00:18:08] Rob

Yeah, I find it hard to imagine that we’d get to the levels of interest rates that we saw in the early eighties, because I think there is a difference in that central banks have learned and changed their behavior since the 1970s.

So, the oil shock of the 1970s was something that had never happened before. It happened after something else had never happened before, which was a free floating of currency, post Bretton Woods. So, the world was in a unique situation and economies and central banks did not have the tools or the understanding to cope.

Now they do have a better understanding, I think, of how the economy works, and better tools, and new tools like QE, or QT, which is the reverse of it, which wasn’t available in the 1970s. Now those tools aren’t of course necessarily that effective. I think increasing interest rates, in the face of a supply side shock is to an extent (if you excuse my phrasing here), pissing in the wind. But it’s still better than blind panic and doing nothing, which is, I think, what would’ve happened before.

The other thing that’s potentially interesting, I think, and this is perhaps more notable in the UK, but I did feel like during the 10 years of austerity under the Cameron government, so from 2010 to about 2020, the Bank of England and the government were fighting each other. The fiscal policy was extremely aggressive. They were trying to cut the deficit very, very sharply. At the same time, monetary policy was extremely open. They were doing QE and trying to get money into the economies. They were kind of fighting each other. 

I think now we’re almost seeing the reverse, in that the Bank of England is desperately trying to tighten interest rates at the same time that the government, under our new leader, Liz Truss, has suddenly embarked on this massive and not very traditionally conservative, plan to spend huge amounts of money on providing households with the ability to pay these high energy bills.

So, there’s a huge, massive kind of fiscal deficit that’s going to be run over the next, at least certainly over the next year or so. At the same time, the Bank of England is frantically raising rates and tightening monetary policy. So again, we have these two things fighting each other. That’s an interesting thing, because although we do have these new tools and stuff, and we have the independence of central banks, which is great, But sometimes the independence of central banks means that the central bank can be desperately trying to do something at the same time that the government’s trying to do the exact opposite.

So yeah, I don’t think we’ll see interest rates of, I don’t know what, where U.S. 10-years got to, if you go back to 1982, whether they got to 20%; probably close to that. But that wouldn’t surprise me if we saw high single digit interest rates, certainly. That wouldn’t surprise me at all.

[00:20:13] Niels

I love it when you say that you describe how the new prime minister in the UK has come out with this massive spending spree at a time where inflation is at least 10% in the UK. Yet, you also suggest that we’ve learned a lot since the seventies. It doesn’t seem to me that we’ve learned very much.

[00:21:30] Rob

I guess what I’m saying is the central bankers and economists have learned, but governments still think that the best thing to do is just to buy votes. 

[00:21:38] Niels

Yeah. But having said that, on a serious note, okay, so what did they learn? Haven’t they been saying that inflation was just transitory for however long? This is the thing that, yeah, we’d like to think that they’ve learned something over these decades, but to me that’s just… We’re all human beings, and maybe it’s for political reasons that they couldn’t say anything else, or maybe they… (which I hope they didn’t really believe that it was transitory, when the whole world…) They could just open their Twitter account and have lots of clever people suggest that this doesn’t look transitory when you increase spending and deficits and blah, blah, blah. So, it is interesting. Actually, you don’t know what the future will hold. That’s the whole point.

[00:22:19] Rob

Yeah. It is interesting. 

[00:22:19] Niels

You don’t know what the future will hold. That’s the whole point.

[00:22:23] Rob

I know, and actually, you know what, in terms of going back to the 1970s, I don’t know if you remember that in the 1970s, I think it was Nixon who had this phrase, “Whip inflation now.” WIN, and he had little badges made up. Now, you come forward today and you’ve got Biden trying to pass laws to make inflation illegal. Maybe I’m being too optimistic, and certainly politicians still think that, by sheer force of will, they can stop inflation from happening.

[00:22:52] Niels

Well, anyways, from a trend following point of view, just to add to what you said, it certainly looks to me like there are still positive winds blowing in September. So far it looks like some of these macro themes that we’ve been positioned for, continue to play out. Certainly, fixed income looks like it was the best sector once again, in the last week or so, with both short and long-term yields moving higher of course, because of the CPI number that came out, but also the Dollar continues to march higher.

There’s no sign still really of any parabolic blow off. So, this could be continuing for a little while longer. I think with equities, I think it’s really down to your timeframe. I don’t think there’s any hard and fast rule as to whether trend followers are making or losing money at the moment. I also think it’s probably quite mixed when you look at different types of equity markets. Grains and meats probably contributed a little bit positively this week, and metals probably not so much.

I think the biggest pullback in performance, for trend followers, would have come from energies. I think you suggested that as well, Rob. They’re still under a lot of pressure, despite the fact we’re talking about an energy crisis. I do think, certainly, the longer-term trend followers are still holding onto some long exposure.

This week alone, we saw heating oil dropping 11%, so that obviously would’ve cost a little bit of money. My own trend barometer is around 41 at the close of yesterday. So that’s a neutral-ish level. I think yesterday, Friday, was probably mixed for CTAs. 

If I look at the number so far this month, BTOP 50 up 73 basis points, up 15.5% for the year; SocGen CTA Index, exactly the same, up 73 basis points for the month, up 22% for the year; SocGen Trend up 1.34%, up almost 30% for the year; and the SocGen Short Term Traders Index up about .5%, and up 11.6% for the year. Comparing that, we have the MSCI Index down 2.22% for September so far, now it’s down more than 20% again for the year; and the World Government Bond Index just can’t catch any break this year, down another 1.4% for the month; and the S&P down 2% for the month down 18.75% for the year.

All right, Rob, are you ready for what I can only suggest is a long list of questions from our community, which we appreciate. Some of them, of course, have been building up. They were asked a couple of months ago when you started your holiday. So, let’s see how we go, and we’ll keep some of the topics that you brought along to the end and see how far we go.

The first question here is from your Yorkim. The headline is “Expanded Universe of Markets and Dynamic Optimization.” Now you’ve explained this concept a while back. So, if people go back on the podcast and find your episodes, they will find quite a detailed explanation of what you did.

So anyways, here are some questions he says, “Rob found quite a few more markets that were available and tradeable as a consequence of dynamic optimization.” Here are the questions, “One, did the dynamic optimization in itself make more markets tradable? When comparing the same universe with and without the optimization, the optimized version performed better. The greedy algorithm is cool. Almost like searching for a twin portfolio but settling for a pretty good lookalike.”

The follow up question is, “Two, but what’s actually going on? What is the reason for this improvement? Is there other information in correlated markets that can be used in other markets? If so, then we are not only trading price itself – we are using intra-market as a signal. This is not cross-sectional momentum trend following, but something else.” 

[00:27:20] Rob

Okay. Okay. So, I think there’s a bit of confusion. When I said the optimized version performed better, it doesn’t actually perform better with the same amount of capital. The idea behind the dynamic optimization is, if you have fifty-hundred million dollars, you could trade a hundred futures markets quite easily – a hundred very liquid futures markets. I don’t have that kind of capital in my trading account.

I’ve got about a half million dollars, so, I could only probably. trade 20, 25, maybe 30 futures instruments. As we say on this podcast, over and over again (it’s probably the one thing that all of us agree on without question apart from the fact that trend following is the best strategy out there), the other thing that we agree on is that diversification across markets is really the way to realize additional returns and this is particularly true in trend following. Going from a single market to a hundred markets, in my research, I see something like a fourfold to fivefold improvement in risk adjusted return, which is pretty extraordinary. There aren’t many places where you can get that kind of improvement. So, the idea behind the dynamic optimization is to effectively say, given the limited amount of capital I have, how close can I get to that big portfolio, which I can’t get given the constraint that we can only buy (unlike stocks), we can only buy futures contracts in these sometimes very big chunky contract sizes. 

So, whether it’s better or not… The optimized portfolio (if I was to run it with the same amount of capital) would be the same. It would say yeah, you’ve got a hundred million dollars. You can actually hold the best portfolio. What it does do is it improves the performance versus just having a fixed set of say, 20, 25, 30 instruments. It allows me to capture (and let’s put a figure on it) say 90% of the improvement I could get from going from a 25 to a 100-market portfolio. It allows me to capture 90% of that improvement, roughly.

I can’t quite get it all because at the end of the day there might be a market that does particularly well in the one year, and I just can’t get a position in it. It’s, because of the capital constraint, so it never quite gets there. So, I agree that the greedy algorithm (which is a technique I use) is cool. It’s great and it’s quite fun. It’s a bit of a leap in the dark. I’ve been discussing it with some people recently because you do have to have a lot of faith in the fact that it’s working. because for example, I’ve got a very strong signal in gold at the moment, but no position in gold, but I do have a position in silver.

You can see, okay, that makes sense. Silver’s substituting for gold even though it hasn’t got such a strong forecast. You have to trust it really, and you have to just have to trust the implementation, and the code, and then technique, and trust it’s working.

So, the first question actually is a little bit wrong. I think it’s an assumption. So, when he says, what’s the reason for this improvement? As I said, there isn’t really an improvement, but he does bring up an interesting point, which is this idea of using information from other markets. 

One thing I’ve done before, and it is discussed in my new book (if I may mention that), is to say let’s not just trend follow, say S&P. Look at the S&P price, use that to forecast where the S&P is going, look at the FTSE, use that to forecast where FTSE is going, but to take all the equities together and construct a single (you can think of it as a bit like MSCI World in the sense that it’s a bit like an equity of all indices, but unlike MSCI World I don’t market cat weight, and I use volatility weighted as well). You can think of it as me constructing some kind of index of all equity futures, index futures, and then trend following that. 

That gets to about three quarters of the performance, so trend following and all of these things individually. So, what that’s telling me is that the majority of the kind of trend that’s coming in, when a market is trending, is a trend in the entire asset class. What’s left over is this sort of idiosyncratic noise, if you like. That’s quite good because it means, for example, that another way of performing well, if you haven’t got much capital, is to actually look at the trend for an entire asset class and then choose a couple of markets that you can actually afford to hold positions in to implement that trend.

You will still do better if you can trade everything of course, or trade more markets. Absolutely. So, he’s right to say that there is effectively intermarket information that you can use. The dynamic optimization does do that implicitly, but there’s actually an explicit way of doing it that is much simpler, which is just to effectively trend follow equities as an asset class, bonds as an asset class, and use that as part of your model.

[00:32:20] Niels

Okay. Cool. All right. Then the follow up question is about adding non-tradable markets. “Rob added markets that were not tradeable. I guess they were only used as signal generators. If the answer in two is that there is information outside the traded instruments that can be used to improve signals, then this seems logical.” Then he talks about, “Does ‘thin’ mean that it might be fruitful to add non-tradable non-futures markets time series?” 

[00:32:58] Rob

Yeah. So, what I do when it says non-tradable, basically, these are markets that don’t meet my requirements, and I’ve got two main requirements to trade a market, actually three. 

One is, am I allowed to trade it? So, for example, I’m not allowed to trade, say, the U.S. sector futures. I’m not allowed to trade them. Some people aren’t allowed to trade crypto futures in the UK. You have to be a professional investor on the method. There are maybe other restrictions wherever you live.

Second thing I’m looking for is volume. I want sufficient volume. So, for me, that’s at least a hundred contracts a day. And then I have a second filter related to what the volume relates to in terms of annualized dollar risk, as well. Then the third constraint is cost. So, markets that are too expensive to trade, on a risk adjusted basis, I won’t trade them. However, with all of those markets, I still capture prices for them, and I still calculate forecasts for them. So, they do go into this big dynamic optimization. But then the dynamic optimization knows that it can’t trade these markets and therefore will ignore them. 

So, what that means is that… I’m trying to think of a… Yeah, let’s take the U.S. stock sector futures. Information about what the U.S. stock sector futures are doing will allow me to form a better opinion of where the U.S. stocks are going generally than if I just looked at the S&P 500, because it’s just giving me effective implicitly. You can think of it as more bits of data, and more bits of data are always going to be good. That’s the effect that’s happening there. That’s exactly right. That’s why I’ve probably got a hundred markets I am trading or could trade; 50 or so that don’t meet these requirements.

The other reason I track these other 50 is that things change, right? Maybe a market that’s on the verge of being tradeable suddenly becomes a lot more liquid and then I will start trading that market. It’s quite a lot of work to get data and set markets up because I’m doing everything in an automatic way. It’s much easier for me, personally, just to have a lot of markets I’m not trading and be collecting prices for and just monitoring. Then, when they meet my liquidity requirements, I can just flick the switch and start trading them very, very easily. That’s the other reason why, even if you’re not using the same techniques I am to trade, it’s still worthwhile.

I think having a kind of bank of markets in your back pocket, if you like, that you’re getting data for (assuming there’s no actual cost to getting that data and assuming that there’s no extra work involved which, wouldn’t apply if you weren’t an automated trader), it’s definitely worth having a bank of markets that you’re monitoring for future trading potential.

[00:35:37] Niels

All right. Okay. Let’s move on to a couple of questions from Kristen. Kristen writes, “I’m a longtime listener of the Systematic Investor Podcast series. There are some valuable nuggets in every episode and your conversations have helped me a lot.” 

That’s very kind of you to say, Kristen 

[00:30:40] Rob

Neils, did you write this question? Come on, be honest. 

[00:35:58] Niels

No, I have a feeling that most of these questions came from some of your family members using different names. 

“A few episodes back Rob clustered the different futures markets according to their risk return characteristics, instead of a more classical approach: stocks, bonds, currencies. Which method and time frame did he use? Does he, furthermore, think such a classification approach makes sense if you are on a risk parity type strategy? Keep up the good work.” 

Before you answer that, there is a second question. Then you can deal with the whole thing. 

“Since Rob hasn’t been on the show after his holiday. (We’re changing that today.) I want to follow up my questions regarding Rob’s non-traditional classification approach. I would be interested in what Rob thinks about how frequent one should re-update the classification: every six months, every year, maybe even less frequent; or what his gut feeling or his back-test is suggesting how often assets change the characteristics and move from one bucket to another? If one uses any classification of the potential assets, in his trend following portfolio (or a multi-asset portfolio like a risk parity) to limit the risk to one sector or asset class I max percentage to energies. The update frequency of such classification can have a big impact on the portfolio versus a more classical approach where you just do it in these standard classifications: stocks, bonds, currencies.”

Anyways, those were the questions from Kristen. 

[00:37:34] Rob

Yeah, so I’m glad you remember the piece I did about the classification. I do think it’s quite a neat little thing. 

[00:37:35] Niels

Yeah. So funny, I remember. 

[00:37:47] Rob

There were some interesting results. Yeah. I use correlations basically. The first thing to say is what’s the best way of estimating correlations in the sense that you want to predict correlation. In terms of estimating, say, standard deviations, looking back about a month gives you the best forecast of future standard deviations. For correlations you want to be a bit slower. I use roughly about a six-month look-back for correlations. Then the actual classification was just using a standard machine learning clustering algorithm. To be honest, you’d get very similar results regardless of what method you use. The key thing is the correlations that are going into that clustering process. 

So, in terms of timeframes, yes, I’m using a six-month correlation. That implies that the correlations are going to change quite slowly. It also is going to take quite a lot of noise out of the process, because you can imagine if you were measuring correlations every month, obviously, correlations on a month-by-month basis might be changing quite a lot and you would see changes between groupings quite frequently. Over six-month movements, probably less so. I haven’t actually formally checked to see what the movement in between clusters is like. It wouldn’t surprise me to see movement, of course, because markets do change the character in the way they trade, and bigger trends happen.

So, for example, let’s take a very obvious example, which is the correlation between bonds and equities. Correlation between bonds and equities, for much of the last 20, 20 years, 25 years has been close to zero, actually, strongly negative. Over the last year or so it’s been positive. Both bonds and equities, last week, got actually hammered for example.

So that’s an example of where correlations have changed and that would’ve meant that the grouping, the kind of bond equity grouping, would’ve looked very different if you’d used correlations to construct some kind of clusters there. It does make sense for a risk parity type strategy. This is what it’s really made for, because what you are already trying to do is, with risk parities, is to identify what your independent buckets of risk are. So, this kind of clustering is a very natural approach for that, definitely. 

In terms of using it for some kind of risk-limiting process, in fact, I don’t actually do that. I don’t limit my risk in, say, bonds and equities. I limit the risk of my total portfolio at quite a high level. That approach does not make sense in trend following. Okay, I’ll say that right now. You shouldn’t do this. This is one kind of area where I’m very much a back-to-basics person.

The reason for it is this. If you think about, say, the last year, if you’d limited your risk in, say, the energy sector, you would’ve probably given up on quite a big proportion of those huge gains that we saw from the energy sector in the first few months of this year. One thing about trend following is that you do not know where the returns are coming from in the future, right? You very much don’t know. You’re placing your bets all over the roulette table. If you limit your exposure in one particular area, as the trend starts to strengthen, then you are very much giving up the potential upside that you’re going to get from those markets.

So, limiting your risk concentration in a single asset class is something I would stay away from. It might make sense to do so, but only at a very, very high level. In other words, obviously one disadvantage of having all of your eggs in your energy basket is that when the markets turn against you, you potentially lose a lot.

If you don’t have the stomach for those short drawdowns, maybe you’re in the wrong strategy, frankly, because it’s part of the game. There is an argument maybe for introducing a cap that doesn’t come in very often. It means you will give up some of your potential gains, but your downside will look a little bit better when those reversals happen. But generally speaking, I would avoid using this particular risk limit measure. 

[00:42:03] Niels

Yeah. It is actually an interesting discussion because it’s certainly something that I’ve come across as well in terms of conversations. As much as I agree with you on that, and obviously depending on how you’ve designed your overall trend following system, intuitively (and we know that some of the bigger managers are looking in that direction and that is), should you, at some point, cap your overall risk to say fixed income or et cetera, et cetera. Just going back to the previous… It would have been interesting to see if, for example, in terms of moving from one cluster to another, whether crypto has moved into tech stocks, I don’t know.

[00:42:45] Rob

Yeah. Yeah. I think when I did my clustering exercise crypto sat with the metals. 

[00:42:51] Niels

Okay. 

[00:42:53] Rob

Which, you know… 

[00:42:54] Niels

Gold? 

[00:42:55] Rob

Yeah. The other place you might have expected to see it was potentially an FX, but that’s where it sat. So, that was what it was certainly most correlated to when I last did that exercise. Obviously, the market environment is a bit different now and it may well be moved somewhere else. It may well be with equities. More close to equities, that wouldn’t surprise me at all. 

[00:43:14] Niels

Okay, we’re going to move to the Scandinavian countries. We’re going to have a question from Christopher and he’s writing, “Hello and greetings from Sweden. Thanks for a wonderful podcast. I’ve learned a lot listening to all of you. This question is specifically for Rob. I’ve read both Systematic Trading, and Smart Portfolios, and implemented a lot from both books in my portfolio. I have a risk weighting of 60% equities, 24% trend CTAs and 16% bonds and give each mutual fund an average forecast based on trend indicators calculated in Excel.

I also use volatility to adjust positions based on a 20-week rolling standard deviation, and in the end, normalize all positions so they add up to 100%. My question is how to treat CTAs within this framework. 

To me, it doesn’t make sense to trend follow the CTAs. They are already doing that. Nor [does it make sense to] volatility adjustment based on recent vol. Should I just leave their forecast as is and allow the allocation to CTAs to grow as equities and bonds get reduced, and vice versa, allow the CTA positions to be reduced if, for example, equities rally and get a strong forecast?” 

Then there’s a bonus question, if there’s time. “If you already have a government bond allocation, and because a portfolio invitation was forced to decide between allocating to investment grade corporate bonds or high yield bonds, what would you pick?”

Alrighty, thanks Christopher. 

[00:44:46] Rob

Yes. He’s clearly a very diligent student of my work and I’m very impressed with… 

[00:44:52] Niels

Or highly paid. Could be both. 

[00:44:55] Rob

Or a highly paid stooge who’s been paid to ask this question. Yeah, definitely. Either way, it’s a great question.

 Just to be clear about what’s going on here. If you think about a classic kind of simple, long-only bonds equity strategy, which you can input nicely with the mutual funds or ETFs, what you do is you can look at the trend in equities, the trend in bonds, and adjust. Let’s say you start with a 60/40 portfolio to adjust your 60/40 portfolio we put more inequities when equities are rallying versus bonds and vice versa.

And I do something exactly like this myself in my long-only portfolio. You won’t be surprised to hear I practice what I preach. Now, the awkward thing is when we add another asset into this, like CTAs, as he says, it’s not clear that you should use the same approach with CTAs.

This is one of those questions where there isn’t really a right answer. I can tell you what I do myself because I look at everything in risk terms, rather than cash terms. I actually look at the current risk of my CTA bucket portfolio, if you like. Obviously, the amount of capital that’s allocated to it because it automatically adjusts as I lose or gain money in my account.

Broadly speaking, that’s not going to change very much. It just gets a little bit lower when I’m in a drawdown, at the moment I’m in an 8% drawdown or something like that. So, it’s 8% lower than normal. It moves within quite a tight range. Broadly speaking, effectively, I treat my CT allocation as fixed. That seems to be the simplest thing to do, especially as if you’re doing these viral mutual funds, there are potentially likely to be exit penalties, bid off costs, when trading in and out of things. Those are more likely to be higher in your CTA UCITS type fund than in bond ETF or an equity ETF, which you could trade very cheaply. 

The simplest and probably the cheapest thing to do is to is just to keep that 24% constant and almost just ignore that, just leave that in a corner and then just focus on the rest of the portfolio, the other, what was it, 76% that you’ve got, and then just do your trend following strategy with that. I would advocate that approach. You could argue that you can start looking at things like the autocorrelation of the CTA returns and trying to time them, put money in and out. There’s weak evidence out there. Winton, for example, put papers out about this effect.

I would be surprised if you could overcome the transaction costs, because it is quite a weak effect that’s going on there. So, just keep it simple. And I’d also say that it’s nice to see someone coming on with about a quarter of their portfolio in CTAs because we always say what everyone should be doing is saying no, not 60/40 but actually, taking some of that money away from the 60/40 and putting a reasonable size chunk of it in CTAs. 25% feels to me like a pretty reasonable size allocation. So that’s good to hear. 

[00:47:58] Niels

Yeah, no, it’s a great, it’s a great start.

I say that in joking because I know you didn’t listen to last week’s episode that I did with Rich. We talked about that earlier. Actually, Rich had done some calculations where he had been in and done some math and optimization, et cetera, et cetera. Actually, the optimal allocation that he came to instead of being a 60/40 type allocation, it turned into a 40/60; to be more specific, a 64% allocation to the SocGen Trend Index, I think it was, or whatever. 

[00:48:30] Rob

That wouldn’t surprise me. That wouldn’t surprise me. But you’ve got to bear in mind that a theoretically optimal portfolio (I’m not going to tell you the exact number because it’s personal information), but my allocation to CTAs is nowhere near that high. Even though I don’t have to pay fees, and I have complete faith in my strategy, there are tax reasons why that’s not the case; why it can’t be as high as that. But because I can’t trade CTAs in my tax-sheltered accounts, which make up a…

[00:49:05] Niels

Is that the account you have?

[00:49:09] Rob

No. No. Definitely not.

It’s certainly not the case that all of my non-tax-sheltered funds are in a CTA type strategy. Even me, who is pretty gung-ho about how great a strategy is, I’d be uncomfortable allocating that much of my money into a leveraged trading strategy. Ultimately, there are risks involved with that, no matter how good it is or how diversifying it is. 25%, that’s higher than a lot of institutions. I think we’d be a much bigger industry if everyone put 25% in CTAs that’s for sure. 

[00:49:47] Niels

Yeah, absolutely. All right, that was a great question. 

Now, we move on to a couple of questions from Eli. He writes, “A couple of questions for Rob. Number one, I purchased Rob’s books on systematic trading, but I found much of the instructions for setting up and testing rules (without overfitting), too complex to understand, possibly due to some statistical terminology I’m not familiar with. Are there resources where a complete beginner could learn about relevant statistical concepts, as an intro to Rob’s book? 

Question number two, I’ve heard the topic of trend following on single stocks discussed often in the podcast, the consensus seems to be that even if you include these in your portfolio, you shouldn’t exclusively trend follow single stocks. This is because there’s too much correlation between single stocks. 

To quote Rob in a past episode, the average correlation between stocks and the S&P 500 is 0.6. My question is, why not build a trend following portfolio of single stocks which are optimized for low correlation. For example, you would begin with a list of all liquid stocks (which are in the thousands) and use some software to select whichever combination of 30 that has the least correlation – meaning within that 30, each stock would have a low correlation to the other 29. Using this method, you would end up with a sufficiently diversified portfolio.” 

I’ll let you start with that because there’s a long follow-up question, but I’ll let you start with that. 

[00:51:14] Rob

Yeah, I do sometimes say to people, you could have a look at my third book, Leveraged Trading, which is much more of a beginner level. That might help, but to be honest, that doesn’t really go into what I’d call statistical terminology, and concepts, and stuff like that. Yeah, it’s a tricky one to answer because, like a lot of people, I learned about this stuff at university. There’s not like a single textbook I could point to and say, have a look at this.

So, I don’t know whether I need to write another book. 

[00:51:53] Niels

I was just going to suggest that. 

[00:51:54] Rob

Why not Systematic Trading for Dummies, or something. Although I couldn’t call it that because of course the, For Dummies is a copyrighted term. But yeah, I guess I’m going to have to think about that.

I might have to write an article on my blog with some suggestions. I’m not sure I’m the right person to write a beginners guide to a financial statistics type book. Maybe I am, but yeah, I think you should probably start with Google until, and unless, I can think of a better answer there.

The second question is an interesting one actually, and you can say it’s not really a question about trading single stocks, because you could use this approach with futures. You could say there are 150 liquid futures markets in the world, let’s pick the ones with the lowest correlation and trade those.

Given that we’re always banging on about how diversification is brilliant, but also the fact that you can’t necessarily trade 150 markets unless you’ve got a pretty institutional size portfolio (which not everyone listening to the podcast certainly has) then maybe that would be a good thing to do. There’s a couple of problems with that. One is that (as we save pretty frequently), correlation is a linear measure of co-movements, but trending is an inherently non-linear strategy. It’s about the things that are happening in the non-normal part of the distribution.

Just because things have got a low correlation doesn’t necessarily mean that they’re actually going to provide you with a maximally diversified portfolio. The other thing to say is that a lot of the time (and I touched on this earlier when I was talking about the dynamic optimization), the reasons why you get such a huge increase in improvement when diversifying with trend following, that huge improvement is actually much greater than you would expect if you looked at the correlation numbers. So, if you look at the correlation, you’d expect your performance to be maybe two, two and a half times better. It’s actually close to four, five times better. That’s the nonlinearity effect coming in there. 

Basically, the problem is that the smaller your portfolio of assets, regardless of whether they’re single stocks or futures, the less likely it is that you are (and the terminology I like to use), that you bought the winning lottery tickets; that you’ve got one of the lottery tickets that actually is going to win money. You may pick 30 stocks that have very low correlation and be very happy with yourself but, just by bad luck, it’s one of the other 2000 liquid U.S. Stocks that does really well, and actually makes all the money and you just don’t happen to have it.

So yeah, it’s a nice idea. But I think if you were to form a portfolio of stocks that was large enough to give you a reasonable shot at picking up decent trends, you’d probably that basket was much more correlated than a basket with the same size number of futures.

The other problem, of course (and again, this touches back on something I discussed earlier), is correlations change. Just because two stocks have got a zero or even a negative correlation over the last six months doesn’t mean that’s likely to remain the same in the future.

Stocks, which have very low correlations, in something like the U.S., do so because there’s some really weird idiosyncratic thing going on with that particular stock. That’s a situation that won’t hang around forever. So, for example, I would imagine that the correlation of Twitter with the rest of the U.S. Stock market has been pretty low over the last few months, because Twitter’s in this weird takeover situation with Musk who may or may not buy it. Therefore, the price of Twitter is not driven by the same factors that are driving the rest of the S&P 500 but driven by how people think that the legal case in Delaware is going in as much as it will force Musk to buy Twitter or not, which he doesn’t really want to do anymore.

So, that last example is of a weird idiosyncratic relationship that’s created a temporarily low correlation. If and when this situation is resolved, either Twitter will be back to being a public company without Musk buying it (in which case its correlation will then jump to being much higher again, with the rest of all the U.S.). It’ll be trading more like normal U.S. Stocks. Or, of course, it will just disappear entirely and become a private company, if he’s forced to buy it – whatever those two things happen. So yeah, it’s a nice idea but I can’t help feeling it won’t work as well as you might expect.

[00:56:42] Niels

We’ve talked a lot about this in the past as well. As much as we want to be quantitative in everything we do, there are certain things that we are not so quantitative about in our research process. I think, to be fair, part of that is also market selection. I don’t think that there’s anything wrong with sometimes applying a little bit of common sense in your “research process” or market selection process as well. I think your answer also included that, so to speak.

Anyways, there’s a quick follow up question. And that goes to… No, not maybe a follow up… is it a follow up? Yeah, I think it is a follow up. 

[00:57:23] Rob

It’s a different topic, really.

[00:57:24] Niels

It’s a little different topic, but it is a follow up. It’s from Eli as well. 

“I’m very excited for Rob’s return next week. I’ve already sent a couple of questions for him, but another one occurred to me. I totally understand if this one has to wait for a different time, blah, blah, blah okay, that’s fine. I’m very interested in Rob’s continuous moving average crossover system where the size of the position is determined by the distance between the fast and the slow-moving average.

There is an upper bound which indicates a maximum position and anything under that maximum is calculated as a percentage of that maximum. How do you determine the distance which would indicate a maximum position? For example, suppose you have three crossover combinations to 5 day and 20 day, 10 and 40, to 20 and 80.

Obviously, you would need to use a different distance in each of these three. Would you perhaps do as follows: calculate the daily standard deviation for some look back (say 25 days) then for each crossover system multiply the standard deviation by the square root of the slower number. The result would be used as a distance indicating maximum position for each crossover set. Is that a valid method?” 

[00:58:38] Rob

So, my position size is actually determined by the distance between the two moving averages divided by the standard deviation. I’m scaling by the standard deviation. The way I actually do it is I measure the absolute size of that number over lots of different futures markets to avoid overfitting. Then my maximum position is effectively determined, almost arbitrarily. I basically say I never want to take more of a position that’s more than twice that average. So, if you think about it in risk terms, I go from 0% to a 100% risk in an average sized position if the forecast gets stronger. In other words, the distance in the moving averages gets bigger; the trend is getting stronger, or it’s been around for longer; then that will increase to 200%, but I don’t go any higher. I don’t give a higher position than that. 

There are a couple of reasons for that. One is to avoid concentrating too much risk in a single instrument. The second reason is that actually, if you empirically look at what happens when those forecasts get even bigger, there tends to be mean reversion. In other words, if you think about very fast trading, what you’re seeing here is a dead cap out. It’s an effect where a very strong downward movement is typically followed by an upward movement, particularly in risk assets, like stocks. It’s a reverse and something like volatility, of course (like the VIX). I think about slower trading than what you’re seeing is effectively a very strong trend that will eventually become exhausted. That’s the story behind that effect. 

So, he’s talking about taking the distance between moving averages and then multiplying by a standard deviation, blah, blah, blah. I think, in terms of the units he’s using, it’s right. He’s not dividing by a standard deviation, so it makes sense to multiply by one. Obviously, the distance in the moving averages can get bigger and bigger the longer your look-back is. If you’ve got a short look-back, the distance isn’t going to be that big. 

Then he’s saying multiply by the square root of the slower number, again, using the square root, the-look back, again, that makes sense quantitatively, if you like. What I’m not really convinced about (because I have never done this myself) is whether the scaling is going to be correct. So, I don’t know whether the upper bound that will give you would make sense. It sounds like it’s a reasonably sensible method, but the thing I would emphasize with anyone who’s trying something new is to back-test it, see what that looks like. See what that maximum position looks like. 

If you are hitting that maximum position, say, 10% of the time, yeah, that’s probably about right. That’s probably what you want to do. If you’re hitting that max position 90% of the time, that maximum position is way too tight. If you’re never hitting that maximum position, it’s not doing anything, you’re wasting your time. You’ve got an increased complexity in your system that’s not actually having any effect whatsoever in your trading patterns. 

Yeah, so it sounds like Eli is reasonably quantitatively switched on. 

I’m going to just keep using that word now. I’m sure they’re capable of doing that exercise. That’s what I do is a kind of sense check. It sounds to me like the units are correct, but I don’t know whether the kind of absolute value that would come out of that process would be about the right number. The only way to check that would be to actually test it and see what it looks like. 

[01:02:13] Niels

Okay. Let’s move on. We have a quick question from Duky and Duky is interested in how you would determine what you would be looking for in order to decide or determine how much capital you can run in in your system.

He talked about if your capital goes up by quite a large number. I don’t think that the number he quotes is necessarily realistic. So, I think it’s more about rephrasing the question into more of a general question about what you would be looking for. I guess it goes to capacity, what’s the capacity of a strategy, so that we don’t overstep that? So how do you think about that? 

[01:03:02] Rob:

We can think about capacity in a couple of different ways. So, one is costs. If you are trading one contract at a time, you’re unlikely to be carrying much in the way of costs, because you’re almost certainly going to (worst case scenarios) get filled at the best bid or the best offer. If you’re trading a thousand contracts at a time, clearly then it’s very unlikely you’re going to get filled there. Also, you’re going to have what the market structure people call a permanent impact on the price as well.

If you’re putting in a thousand lot orders every five minutes, then the market’s going to be, hang on a second, there’s a big player here. They’re going to start pushing the price away from you in advance of what you’re doing. Your effect on the cost is going to be much higher than you would’ve anticipated.

The other thing is volume. These things are related, but not necessarily. If you are a big percentage of the volume of a market, it’s quite likely that you’re also going to be paying higher costs. But in terms of open interest (the relationship to an open interest and volume is not the same for every market), so you can think about markets where the open interest is quite small but there’s a huge volume. I call these gamblers markets. There’s not a lot of people with real positions; there’s a lot of people turning over speculative positions. 

In a market like that you may think that your position is not that big because you’re trading quite a small percentage of the volume, but you could be half the open interest. That’s potentially a dangerous position to be in if the market moves sharply against you, because all of those speculators will just disappear, and you’ll be stuck with half the open interest and not be able to do much about it.

That’s normally a second order problem. So, you normally run into problems with costs before you run into problems with being too big in proportion to the open interest. So, let’s just focus on the costs. 

If your portfolio is small enough, you can basically run what… Indeed, where we’ve been talking earlier about the problems of having a portfolio that’s too small. So actually, at the moment, my portfolio’s effectively too small in many ways. It would take quite a lot of additional capital. If I had, say, a hundred times more capital than I’ve got, which would put me up to 50 million, which is a sort of a small to medium sized institutional sized capital. I could still be trading in exactly the same way I’m trading now, but I would be able to trade all the markets rather than having to do dynamic optimization. So, my performance would improve, but I wouldn’t be doing anything fundamentally different.

Basically, there’ll be certain markets that were too expensive for me to trade, but that cost has nothing to do with the size that I’m trading at, they’re just too damn expensive anyway. There’ll be certain markets that are too illiquid to trade. Again, it’s nothing to do with size, they’re just too damn illiquid. They just don’t even meet a very basic minimum floor. 

The interesting thing is that when you start to go beyond that, so, if you go from, say, $50 to a $100, to $250, to $500 million, and upwards to that (back in the day I was at a hedge fund that was $30 billion, at one point, while I was there). Then you really have got to start worrying about this stuff. So, what will happen, gradually, as you move up is that there are less liquid markets where your costs are going to start to look too high. The market where you could do one lot a day, that was fine. Although it wasn’t very liquid, that one lot a day was fine. Now you’re trying to do 10 lots, 100 lots a day in a market where the volume is only, say, 500 lots a day, you are going to be having a big impact on that market. 

So, you’ve basically got two main options that you can follow. One is you can slow down your trading. So, in that particular market, you say I’m going to just slow my trading right down. That will deal with the cost issue, but it won’t deal with the open interest issue because you will still be, potentially, quite a large proportion of the open interest even if your volume’s lower and, therefore, your costs aren’t too bad.

The other thing you could do of course, is reduce your allocation to that market. The main thing that will happen, assuming you use that second approach which is probably the more sensible of the two (because slowing down can only get you so far and doesn’t deal with the open interest problem), effectively your portfolio will gradually shift from being one that’s maximally diversified to one where… Ultimately if you imagine you’re this kind of unicorn, $100 billion CTA (no one’s that big, I don’t think at the moment), your portfolio’s going to be dominated by the big liquid futures markets. Your allocation to the smaller markets is going to be (just rounding out), it’s just going to be tiny. 

When I was at AHL running the fixed income portfolio, we had more money than we would like in, say, the U.S. Treasury markets, because we were so big that we couldn’t put as much as we would like in a much smaller fixed income market, like I don’t know, five-year interest rate swaps, for example. That could only ever be a tiny, tiny part of our market. Whereas, for someone who is running, say 5% of that; a 20th of that sort of size of capital, they will be out of a maximally diversified set of instruments.

The interesting thing about futures is there is kind of a real sweet spot that doesn’t really exist in equities in the same way. There’s this real sweet spot of somewhere between $50 to $250 million, where you can trade all the futures you want to trade. You’ve got reasonable management fees coming in to bring in reasonable resources, but you’re not yet at the point where you have to start un-diversifying your portfolio because you’re just too big. That’s the sweet spot to be in, in that sense. 

[01:09:06] Niels

All right. Good stuff. Let’s move on to two questions from Adam. 

Adam writes, “Hope you’re doing well, great show as always. I have a question regarding running multiple systems, multiple different systems, and all parameters together. I understand that there are two different approaches to this: combining signals from different systems to form a one main forecast signal, which tells you the confidence you have in a trade or therefore the position size that should be taken,” and then he says, “Rob’s approach (if I understand it correctly; so you have to address that), splitting risk across multiple systems, which all trade independently according to the rules of each specific system. What do you and your guests consider the pros and cons of each?” 

What are your thoughts? 

[01:09:50] Rob

I think you should kick off on this one, Niels. I feel like I’ve been talking a lot. 

[01:09:55] Niels

I think, from a simplistic point of view, you could say that having them run independently. I think for most people who want to do this themselves, that’s probably not a bad way. You would have to do your thorough back-testing, et cetera, et cetera. In a sense, your total exposure to each market will reflect the confidence that you have, overall, using different types of approaches. 

Then, Adam, you write about, “combining signals from different systems to form one main forecast signal.” Yeah, there probably will be some differences to doing it the other way. It also depends on how you do your risk management.

If you have multiple ways of generating a signal, at the end of the day, whether you say it’s a different, separate systems and you form it into one signal, or you have different systems doing their own thing but getting into, again, one exposure at the end of the day, I wonder how much of a difference it means, at the end of the day, depending on how you manage the overall risk of your portfolio?

If you’re doing this more on an individual investor basis, to some extent maybe running the systems separately, it also might be easier to monitor what the actual risks are in your trading program as a whole. 

[01:11:35] Rob

Yeah. I guess that the advantage of doing things separately is… 

[01:11:41] Niels

It’s more transparent.

[01:11:42] Rob

It’s more transparent, right. So, you can very easily say oh, this system did these trades and therefore this is its performance, blah, blah, blah. There are some difficulties with that as well. So, you have to be very careful, for example, especially if you’re running an automated system that you don’t have potentially two systems trying to do the opposite trade at the same time, because that’s actually illegal.

It would be stupid if you had one system that bought, paid the bid off, and in five minutes another one came in and sold, and paid the bid off. Of course, if those two trades had been netted, then that would’ve been an advantage. 

[01:12:14] Niels

I was just going to add to that. When I’m talking about these things I’m thinking about an end of day system, whereby, if you do run separate systems, before you generate the order, it would combine the order. So, you don’t do silly stuff like that. I agree with you. 

[01:12:30] Rob

Yeah. Yeah. Just taking a step back, if you combine things at the position level, which is what you’re talking about, and I think that’s probably the latest I’d want to combine anything. Yeah. There’s the advantage of transparency. 

There are some disadvantages. It depends on how complicated your systems are. For example, I do these things around risk management that make most sense at a total position level, or at a total forecast level. If I was to do those individually for each system, I might get some weird interactions and I’d be doing things more than once. It would be a bit more difficult for me to think about in that sense. I think it’s a lot down to personal choice. To be honest, I don’t think it’s one of these things that people should spend too much time worrying about. There are bigger fish to fry. 

[01:13:16] Niels

Yeah, absolutely.

All right. We’ve got only a few more questions left. [There’s a] question from Michael, it’s super short. “I’ve noticed that your (and that is you, Rob) trailing vol is down to 7.1%, the lowest since your system started in 2013. What’s behind this?” 

[01:13:34] Rob

Weak forecast, basically, yeah, just not very strong trends over the last few months in many markets.

The other thing that could be driving its correlation, it might be that I just happen to have a set of positions on that are hedged, if you like, because that would also reduce my volatility. Just looking at my positions, they’re a bit of a mishmash, I don’t really see anything there that’s hedged per se. So yeah, I don’t think it’s being driven by some kind of weird hedging effect. I think it is just weak forecast. There’s just not very much going on there in the markets, at least as far as my system’s concerned. 

[01:14:16] Niels

Yeah. All right. Last question is from Toby. I say Toby. It could be Rob. I’m not sure. 

[01:14:22] Rob

This one is suspicious, actually. Yeah. Is this Toby from Australia? I think I know who this is. 

[01:14:30] Niels

Yeah. It’s from Sydney. So, it would be Australia. Yeah. 

[01:14:33] Rob

So yeah, know who this is. Hi Toby. 

[01:14:34] Niels

Yeah. Okay. Fair enough. 

[01:14:36] Rob

It’s not me. It’s not me. It is somebody I know quite well. I honestly didn’t ask him. 

[01:14:40] Niels

You must be on page eight now on acknowledgement for your new book. Just going to Top Traders Unplugged in terms of all the plugs we’re doing for your book. But anyway, as always, I will, of course, provide the question.

“Hi Rob, you have an excellent book and blog. Thanks for all of your contributions. Could you spend some time on the podcast explaining your new book?” 

Oh, my God. 

[01:15:07] Rob

It is very suspicious, isn’t it? It really is. Explaining the new book… Okay, the first thing to say is the book is actually doing well in terms of its construction at the moment. Neils normally asks me to…

[01:15:20] Niels

I didn’t ask you this time because I knew this question was coming. 

[01:15:21] Rob

You knew this was coming, exactly. It’s just about to go to the typesetting stage. So all of the letters and the words are in the right place. It’s just a question of making it look nice. That, apparently, will take quite a long time because the book isn’t due to be out until April next year. This is, apparently, quite a slow, cumbersome process. But then it is quite a big book, and quite technical, with lots of equations and graphs and stuff. So, maybe that’s to be expected. 

Yeah, it’s coming out in April and my main contribution is pretty much done. I’m going to have to do some final proofreading, but that’s pretty much it. I’ve got to also put together a set of resources for the book, so, spreadsheets and Python code that people can use. That’s all I’ve got left to do really. 

In terms of describing the book, it’s called Advanced Futures Trading Strategies, so that kind of gives it away. It’s not really what you might expect in terms of being, here’s strategy number one, here’s strategy number two, here’s strategy number three, and they’re different ways of trading the market. 

So, for example, if you’re familiar with Perry Kaufman’s big book whose name has temporarily escaped me. I’m just trying to have a quick look on my shelf to see if I can see the title. 

[01:016:34] Niels

Don’t worry about it.

[01:16:35] Rob

So, Barry Kaufman’s written this very large book about trading strategies, and that’s not specifically for futures, I should add. It’s a great book though. 

[My book] is not like that at all. It’s written more in a way to both build up people’s understanding of the future’s markets, but also to explain how effectively it’s a journey. [ It starts with] owning futures and not doing anything very much with them, which exposes you to the characteristics of the underlying assets. So, for example, if you take an asset like equities, you’ve got negative skew, you’ve got reasonable returns over pretty good Sharpe ratio long-only (but that may not be likely to be repeated from now onwards). You’ve got fat tails and so on, and then explaining what happens as you layer that with different, more complex trading strategies. It effectively explains the transformative nature of trading strategies. So how does trend following affect the underlying asset? How does carry affect the underlying asset? So, the first part of the book is very much like that.

It goes into a lot of things. So, it’s not just about trading strategy. Actually, in a way (hopefully) it explains the way of the futures markets and what’s going on underneath them. So, these ideas about things I’ve talked about, in this episode, the fact that a lot of the trends in futures could be explained if you look at the asset class level, and here’s a strategy that happens to capture that effect.

So, I would hope that people wouldn’t expect it to be, oh, here’s 30 different trading strategies. I think it’s much (this sounds quite pompous to be honest), but it’s much more than that. It’s really about getting across to people the kind of intricacies and what’s going on underneath the hood in the futures markets and how trading strategies seek to exploit that. So yeah, the first part of the book goes from very, very basic, just literally buy and hold a single contract up to the last strategy in part one which is a mixture of trend following at different speeds and carry. 

Then the second part of the book goes into more advanced stuff about trend following and carry. So, for example, what happens if you trend follow the spot price rather than the back adjusted futures price? Can you, or should you try and time trend and carry? How do different volatility environments affect trend and carry? 

Then the third part of the book is, what I call, other directional strategies. So, it’s things like this asset class trend stuff, and mean reversion within asset classes, cross sectional carry, so on and so forth. 

Then the final part of that book goes into detail about dynamic optimization. So, if that’s something you want to understand more about, that’s covered there.

Then the fourth part of the book is about faster trading strategies, so more intra-day, basically mean reversion. Now we know mean reversion is a very dangerous strategy. It’s a convergent rather than a divergent strategy. So, I introduce this very dangerous strategy and say, don’t trade this. 

The next chapter I explain how to make it safer by introducing a trend following overlay and a volatility scaling overlay. Then part five is about relative value. So, it’s about trading spreads, butterflies, both across instruments, but also calendar spreads within instruments. Moritz, for example, has had a lot of success with that, so that’s potentially something people find quite interesting, I think. 

Then the final part, I call Tactics Rather Than Strategy, because it’s about things like, when should you roll? [It’s about] things that seem like really simple questions: which contract should you hold? When should you roll it? How should you manage the cash in your portfolio? How should you do your execution? How should you do risk management? So, it’s these other things that could almost fill a book in themselves, but sort of the non-sexy part of futures trading.

So yeah, I’ve tried to do almost a brain dump of everything I know, or think I know, about futures trading specifically, and the sort of research I’ve done over the last, crikey, I think it’s been nearly nine years now since I started by myself.

I tried to get an awful lot into that book, but hopefully in a way that kind of leads people along a journey and gives them a much deeper understanding of this interesting and weird world of futures trading. 

[01:21:20] Niels

Yeah. Sounds exciting. And obviously not that long to go before we get it out and you can all …

[01:21:28] Rob

If I’m allowed to make one final plug, you can, if you go to my social media or website, there is a link there to pre-order the book. So, get in there now.

[01:21:038] Niels

We’ve been going for a while. I don’t know if you want to go into some of the topics.

No, I think you’re right. Let’s just leave it with all the questions. It’s just kind of a Q and A session today, which is great, since it has been a few months since you were last here. So, I hope everyone got the answers they needed. And of course, as always, you could send in questions to the upcoming guest on info@toptradersunplugged.com.

Next week, I’m joined by Mark for another fun and insightful conversation. No doubt he will have brought some great topics with him as well. But do send your questions if you have anything specific you wanted to bring up.

From Rob and me. Thanks ever so much for listening. We look forward to being back with you next week, and in the meantime take care of yourself and take care of each other. 

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.

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.