← Top Traders Unplugged
Loading summary
Transcript68 lines
- [00:01]
A
Welcome to Top Traders Unplugged in markets. Success doesn't come from predicting what happens next. It comes from being prepared for what you can't predict. In each episode, we go deep with some of the world's most thoughtful minds in investing, economics and beyond to understand how they think, how they prepare and how they decide and the experiences that shaped how they see the world. No noise, no no shortcuts, just real conversations to help you think better and invest with confidence.
- [00:36]
B
Welcome and welcome back to this week's edition of the Citamatic Investor series with your git and I, Nils Kastor Blason, where each week we take the pulse of the global market through the lens of a rules based investor. This week we have a very special guest, namely Rob Croce, who is a portfolio manager in the Quantitative Research and Investment division at Fidelity Investments and who is here to discuss a recent paper that he co authored as well as some of the other topics we have lined up today. Jor Rob, it's really great to be with both of you this week. Hope you're doing well. Where are you connecting from? Let's talk about.
- [01:10]
C
So I'm actually, I'm actually connecting from Israel. I'm visiting my parents. So enjoying enjoying the summer here. Very hot. And enjoying the fact that there is actually a lull in the hostilities between Iran and the US and longitudinal. May, may it continue?
- [01:29]
B
Absolutely. Rob, where you where are you dialing in from today?
- [01:32]
D
Beautiful Boston where you can see we have a, a lovely gray sky and about 45 degrees outside. Perfect Boston springtime.
- [01:40]
B
Fantastic. Looks very nice, I have to say. Now before we get into any of the topics, since it's your first time on the show, Rob, why don't you share a little bit with the audience about your background, your journey to where you are today and maybe also what got you fascinated about trend following.
- [01:57]
D
Yeah, life is so path dependent. I'm a recovering economist. I don't know if I'm embarrassed to say that, but I'm definitely a recovering economist. I did grad School Econ PhD at Ohio State and in my last year of graduate school I did an internship at a public investment plan Texas teachers and I was in their strategic research group for a summer and I saw how big institutional investors think about quantitative investing and I was hooked. And that led to my first job at a small firm in Houston. And at that job we were an outsourced CIO for a public plan as well as a direct fund of funds manager. When I started, one of my colleagues was in Manager due diligence. And he asked me if I knew anything about managed futures or trend following. And I said, no, what's that? And he said, well, basically you buy things that are going up and sell things that are going down. And I thought that was ridiculous. Right. Like in economics, like everything's already priced in and markets adjust immediately. And so, you know, but I went and I started playing with data and I went back to him very excitedly and said, look at this. Buying things that have been going up seems to work. And, you know, that was the beginning. And that firm decided to build out a direct asset management business. I got to be on the ground writing the code and learning by doing, which I think is a great way to do it. Managed futures was one of the first things that we worked on because the view of the CIO at that firm was that it's one of the most impactful things you can add to a portfolio from a diversification perspective. So worked on it there and it's been a theme in my work ever since. I've been in a couple of firms since then. I joined Fidelity about two and a half years ago and working in the same space.
- [03:59]
B
Absolutely. And I know of course that today you're here to share your own views and thoughts and not those necessarily of Fidelity Investments, if I'm not mistaken. Anyways, great to hear and thanks for the, for the background. We've got a really strong lineup of topics, as I mentioned already. One of them is a paper that you co authored, Rob, but before we dive into any of those, I'm always curious that sort of anything else other than that, that might have been on your radar in the last few weeks. So you're going to come to you first. Anything that excites you at the moment other than trend following?
- [04:38]
C
Yeah, absolutely. I'm from Cambridge, so I've got to talk about two undergraduates, a guy called Liam Price and another guy called Kevin Barretto, who is a third year undergraduate in Queen's College in Cambridge. And between the two of them, they're not really expert mathematicians, but they managed to crack a problem that has been outstanding for 60 years to do with the density of prime sets in sort of the natural numbers. And that was part of a bunch of problems that Paul Erdos, who was a Hungarian mathematician, a very prolific mathematician, he not only solved lots of problems, but he also posed the collection of problems. And this problem has been eluding a lot of mathematicians, including like Terence Tao and like amazing mathematicians for 60 years. And Liam and Kevin solved it by guess what? Asking ChatGPT5 whether it has how to solve gave them, it gave them not a full answer, but it gave them sort of the path to go down and they've just published it and very well done to both of them. Really amazing. But it look like vibe coding is one thing, but vibe mathing is also, is also a thing and it's, it's really, it's really impressive to see, to see that.
- [05:58]
B
Wow, that is amazing. I thought when you brought up Cambridge, I thought we were going to talk about any boat race or something like that that they may had won recently, I don't know, but apparently not something much more important. Rob, what. What's been on your mind recently?
- [06:14]
D
Yeah, so, you know, I. One of the things that I tend to do is think about unorthodox ideas and one of the things that's very unorthodox in what we do is the idea that you can do something other than pseudo out of sample testing. At the same time, when we're running out of sample test, we've also run many, many back tests in our lifetime and as a result it's very hard to run a true out of sample test. So what if we wanted to use information from a full sample of data to make inference? How can we do that without polluting an investment strategy or cheating? If I flip a coin 20 times and it comes up heads every time, can I infer that there's something about that coin? There's a one in a million chance of that happening without there being a bias in that coin. So can I make some inference if I observe that some commodity spread, for example, has been up, has a hit rate of 100% over the last 20 or 30 years, one spread in a particular market, let's say natural gas in November for a March contract, can I make some inference from that? And natural gas is a natural example because we know that there's a stock out risk premium there. But maybe as quants, as trend followers, we're a mile wide and you know, not as deep as we are wide. You know, we go for breadth and so maybe there's a way that we can use data to learn something about the fundamentals of a commodity market, for example, and make some inference that we could use for, let's say possibly risk management or something like that. Would you really want to be long March natural gas in the early winter? I don't know. There might be a risk management thing you could do with that type of information.
- [08:16]
B
Interesting. You have since your firm trades a lot of commodities. Any thoughts on this?
- [08:22]
C
Yeah, I think it's a really valid question and we've been struggling with sort of true out of sampling. I think in the situation where there is a physicality to the market, it is fair enough to use that information even to the, for the process of, you know, there is a natural flaw to the market. For example, a lot of the, a lot of the carbon markets will have an underwriting by the government about where the auction is. So like Californian auction for, for carbon. So being able to incorporate physical nature of what the market is doing is really important. I mean there are a few techniques I always, I gave a talk in JP Morgan many years years ago about how to detect when you're, when you are really your result is too good. And sometimes it is actually about forward looking. You actually use the system to say look forward and say if, if the effect is there, should I make more money by actually looking forward? So sometimes a lot of the time we kind of need to think about what makes my strategy work work and then try to design the test around what is the essence of what the strategy is meant to capture. And it's not necessarily a backtest, which is the thing, but it might be the physicality of the market, it might be a mathematical trick. And I think that's really important. I think we'll come to that actually in the next paper where the authors give one reason but actually I think it's another reason altogether.
- [09:58]
B
Any any further thoughts Rob, on, on that comment or.
- [10:01]
D
No, it's good to know. I'm not crazy, but I like the idea that it might be physicality but at the end of the day I think momentum and trend were detected in the data first and then explained. So with that in mind, I think we have to let the data guide us and get some conviction from it in a way that doesn't overfit.
- [10:34]
B
One of my topics are not that related to any of your very thoughtful topics but it is but some people I think would have said that knowing what's about to happen tomorrow will definitely lead to lower rates. And that is of course that Jay Powell is going to step down as chairman of the Fed after eight years. And although we are not here to get involved in politics, I do think it's worth marking this change and it obviously also comes at a very interesting time for the, for the world and for the US economy. So we'll definitely watch that with interest through the data, of course. But speaking about sort of interest rates also, I did see a headline. It could have been a day or two old and the 10 year JDBs are hitting 29 year highs. I thought that's kind of interesting. Not necessarily what we in the western part of the world are sort of noticing. I know you trade your view trade a lot of that JGBs. So maybe you want to comment on the fact that we're seeing these on you know, unusually high interest rates for four for Japan at the moment.
- [11:43]
C
It's not just Japan. We're seeing it in many places. I think you can buy U.S. treasury. There was another headline. You can buy U.S. treasuries with 5% coupon. That's not a we. You haven't been able to do that since 2008. We are getting, we're getting fiscal issues everywhere and it's very interesting actually that you saw equity recover from over April. We have seen equity recovers all the way to the, to the, to the top of the market. Whereas in the bond market we have seen a nice bounce when the ceasefire was announced. But then since then we've seen the Treasuries being dragged down, yields rising pretty much everywhere in the market. So, so it's been very stark the way that credit and equities have recovered as a result, you know, since the, since the crisis. But bonds are really very much at the bottom of the market.
- [12:40]
B
Yeah, absolutely. Now the other thing I had noticed that came on my radar is you know, obviously we've noticed when big players like Fidelity and blackrock and all of the, you know, these great firms step into the, to the ring of trend following. That's great news. I mean I'm sure it'll be good for everyone. But I also noticed in an email that since I must be on the email list of, of rituals, the large IRA in the US that that they're now coming out with a concentrated momentum strategy. This is I think for, for single stocks. And it's not the product I'm here to, to talk about or mention. It's just the fact that here's another firm moving in the direction of momentum and you know, kind of, you know, buying into what we've seen for, for decades that there is really, really good, a really good interest and I should say at least a business case, investment case for momentum. And going back to the point about, you know, we're now at all time highs and is this a good time to buy? And most people say ah, probably not, we should have bought three months ago or six months ago. Well, in our world it probably is a good buy to buy a Time to buy when we see new all time highs. So it is interesting to see more and more people getting into, getting into that. I don't know if there's much to discuss here, it's just something that I found interesting. What there is to discuss is of course always a little bit of trend following update for the month of May. We're kind of halfway now. It's been a strong start to the second quarter. April of course was, was strong. May continues as far as I can tell, maybe with a little bit of different leadership. We see a lot more in the medals as far as I can tell in terms of trends, not so much in energies right now. They're actually net, net a little bit down in terms of price for May and then of course equities as mentioned. I mean US equities making new highs, other markets making new highs, also Japan and I think maybe even some of the currencies might be beneficial to, to trend followers at the moment. Of course, as I've spoken about a couple of times in the last few weeks, I think from an exposure point of view what, what I find interesting right now is that trend follow generally are probably going along with the trend as they should, meaning they're shorting fixed income the last few months. So that'll be interesting to see whether there's going to be any follow through because that has been perhaps the most difficult sector for many trend followers in the last couple of years pretty much since the banger year of 2022. So in terms of my own trend barometer, it finished yesterday at 64. That's a very strong reading and that just tells those who follow it that there's a lot of breath. It's not about that that trends are any, you know, stronger, but there are more of them in the 44 market portfolio that it tracks. So that's a good sign. We want more breath for sure. In terms of performance, the numbers I'm going to quote is as of Tuesday evening for the industry and I, I imagine that yesterday actually was a pretty good month. Sorry, pretty good day for the industry as a whole. So beta 50 so far in May up 1.41%, up 11.10 for the, for the year. SoC gen CTA index up 1.48% up 11.85 for the year. SoC gen trend up 1.84%, up 12.16% for the year. And the Soc gen Short Term Traders index is up 1.15% in May, up 6.47% so far this year. In the traditional world, MCI is doing okay, up 2.61% in May, up 8.59 so far this year. The S P 500 aggregate, not S P 500. The S P US aggregate bond index is what I was meant to say is down 12 basis points in May, up 16 basis points for the year and the S P 500 is up 3.31% so far this month and up 9.2% so far this year. I would love to, I mean, however much you can comment, generally speaking, feel free to, to give your observations about the trend environment you have. Maybe you can, you can start if there's anything you've noticed about the environment.
- [17:08]
C
I think what's interesting actually is on the dollar, US Dollar, we've seen quite a differentiated response because although we have seen the dollar slumping back and a lot of EM currencies quite recently, the gold hasn't tracked it. So a lot of people were trying to explain gold as a result of the weakness in the dollar, but actually gold hasn't recovered as much even though the dollar has slumped down. One of the things that we have seen is actually in India. India is a big consumer of gold because of the energy crisis. They are having trouble with FX reserves and they are trying to cut down on gold imports. Indian families are really traditionally very much investing in gold and it's actually a big consumer of gold in the market. So it's very interesting that the way that the crisis in the energy market is projecting actually on different segments of the market in a different way. So I think gold trend is actually quite interesting.
- [18:13]
B
I think what you're referring to was actually almost made it to my radar and that is that they have imposed a tariff.
- [18:18]
C
Exactly.
- [18:19]
B
No less. I think it's 15% on gold and silver imports. So. Yeah. All right, good stuff. Any thought from your side, Rob, or not too much on sort of the environment at the moment?
- [18:31]
D
I mean the most counterintuitive thing relative to the average investor is thinking about rates is just how bearish trend followers might be in rates today. Just given what we've seen in the data recently, it's definitely a counterintuitive of position.
- [18:52]
B
All right, we've got a couple of papers. The first one is one that you brought along your. So I'm gonna ask you to. To tell us more about it. It's called Betting Against Beta. It's by two people from aqr, Andrea Fascini and Lesser Hair Peterson, a fellow Dane. Actually, as I can tell, I was
- [19:15]
C
going to mention that yes, absolutely.
- [19:17]
B
Well that's always appreciated for sure. So why don't you tell us about what it's about and we'll try and when I say we, it's the royal we meaning Rob. We'll try and add some clever commentary to it.
- [19:33]
C
Okay, so a little bit of historic background. Lasse Pedersen came up with a host with aqr, came up with a host of papers really giving away a lot of the alpha in trend following in relative value carry in carry. And one of his and a part of that host of papers that he came out with in 2012, 2013, betting against beta was one of them. And it came up with a really interesting factor which is trying to explain why it seems to be that assets with equities with high beta will underperform equities with low beta to the sort of the S and P. And it's not just in equities and it's not just in the S and P. What is very interesting about this paper is that the effect that they're going to describe is prevalent first of all historically. So we're talking about out of sample backtest. It's been prevalent in US equities from the 1920s. It's been prevalent in international markets, international equity markets, but also in the bond markets, in the credit markets and a little bit in commodities, but not really, not very, not very big. And I think it's a very interesting proposition. So they try to give it a reason and we'll come back to why they are saying but let's first of all think what is the effect? The effect is that if I were to take, if I will go to go along half the S and P, say the ones with the low beta and then I would marry it with a short in equities which have got high beta on them. So it's like you're trying to create a factor which is beta neutral. You're trying to remove the index. This seems to make money historically. This seems to be. Well first of all by construction seems to have a zero beta but more importantly seems to make money very consistently across factors and across asset classes. So really impressive. And by the way AQR has continued to publish that factor on their website. So we now have 13 years of out of sample on the beta factor and long made continue. It is actually very strong. Okay. And I remember my, my shock when I first, when I first discover it and thinking hang on a second, does this make sense? And it's sort of, it seems natural in a way that we do Know that, for example, low volatility stocks tend to outperform high volatility stocks. There is a, there are a few ETFs out in the market which you offer. You like low volatility exposure. So the beta, if you think of it as a little bit of a form of risk, so it's chimed it's time. True, But I think the more I dug into it, I thought the more natural it became. So if it's okay with you, I'm going to try and explain why I think this is making money.
- [22:43]
B
Okay.
- [22:43]
C
Okay. So one of the things to realize is it's not necessarily that the, the stock itself is underperforming. It's just that the low beta portfolio, I can think of it as like the long portfolio and the short portfolio. The portfolio that you are long is a much higher quality portfolio. So we are used to trend following when we try to create a portfolio of a collection of sort of alphas of risk streams. And one of the things that we try to do is we try to do as trend followers is to grab lots of different uncorrelated bets, uncorrelated assets. So we have commodities, we have equities, we have effects. And the same effect is in a way is prevalent in the equity market. So if I were to try and create a long portfolio and ask you which, which portfolio will be a better sharp. So is it going to be the one where you've got a lot of idiosyncratic stocks or is it the ones that, you know, all the, the stocks are very highly correlated to each other. Then you would be very quickly say, well, actually I would prefer to have the high idiosyncratic risk and low on the beta. And in fact, that's kind of the way that the factor works, because in a way you need to be long one, because the long portfolio of low beta is so diversified, you have to buy essentially $2 of low beta stocks for $1 of high beta stocks. Okay. So you actually net long. Notional terms, you are, you are net long. Okay. But more importantly, the portfolio that you along is very high quality. It's got a high diversification because the stocks that you preferred are actually low beta. So it kind of makes sense. And the authors then try to explain this through why this effect is actually happening. Why is it the case that high beta is underperforming and they're coming up with an explanation to do with the cost of getting leverage, the cost of financing. So what they're saying is that high beta stocks are more wanted by investors who have capital constraints and therefore they want to get the higher beta. And that explains why this they perform worse. I'm not sure I'm a big buyer on the explanation. I have to say one of the things that they, they do, they do go through a lot of exercise in trying to sort of show that at periods where there is more constraint on, on leverage and on finance the betting against beta factor is more profitable. But I'm not sure if that is actually an explaining factor rather than at the same time. So that's the paper in general. But I mean I would open it. I've spoken long enough. Let's have Rob comment as well.
- [25:44]
D
So two counterpoints to the interpretation that you just shared. So it seems to be the case in Treasuries as well. If you were to look at us 2 year versus 10 year it seems like the 2 year probably has better risk adjusted returns in the long run. Those are not a difference in quality. I don't think question about the paper. I don't recall whether they ran a control with quality as a factor on the right hand side. Do you remember?
- [26:15]
C
Yes. No. They've got three factors but I don't think quality is one of them. So I'm not sure that. That, that's, that's. But that's. You're right actually I'm not an equity expert. Yeah that that would have been very useful as a measure. But it's interesting and I think you, you may be right on the, on the Treasuries. I think one of the issues I have with the, with the disinterpretation is, is that if you really want risk there's plenty of risk in the ETFs, plenty of risk in, in futures to actually get yourself exposed to that. It doesn't seem to me that if I wanted more risk in equities my natural choice would be high beta. I would definitely want high volt asset and I would want maybe sort of a 2x ETF or the future but it's not clear to me. And again indeed in other currencies where you would have in other equity markets where you would have expect finance to be even more constrained than the US you don't see that effect even stronger. So it's not necessarily that we see this explanation but I mean to give them the credit they do go through a collection of looking at the TED spread as a proxy and trying to correlate that to the performance of the betting against beta. I think there are a lot of really nice observation that Came from that paper really nice. Which are not mathematical, but the way that they view, they look at private equity and they see one of the things that private equities do is invest in idiosyncratic risk and leverage it up, which is exactly what this betting against beta facto does. Similarly, they very cheekily download Warren Buffett's portfolio, Hathaway Berkshire, and they notice exactly the same thing that he would like to invest in idiosyncratic risk. I've got to say that's completely unsurprising. Berkshire Hathaway, basically, that's what he made his money. He's taking big bets on idiosyncratic risk. But I think they've done a lot of very good work. And personally, I think that that setup of picking a collection of very diversified assets, actually, I mean, we know we do that in trend following as well, will definitely on the long run will tend to outperform a collection of sort of high beta alpha returns.
- [28:50]
D
I think I buy the leverage aversion argument. If I think about how an institutional investor may think about a portfolio or a more conventional investor leverage sounds scary and dangerous potentially. And if I think about how large bond portfolios are likely managed, they probably use more duration when they want more risk, and often by going out the curve.
- [29:20]
C
Yeah, I'm not saying it's false, but I'm not saying it's false, but I am saying that it's not necessarily the whole story on. On that factor. But I think it's very interesting that this essentially a portfolio construction technique which is hedged to the first factor, is something that has really stayed very consistent all the time at multiple asset classes and out of sample. So well done. To Pedersen, do you know or do you remember your.
- [29:55]
B
When how far did they go back initially in terms of data? I'm just thinking the tier, the period that we are analyzing, Certainly the last 13 years, we know equities have gone one way pretty much all the time. Is that, is that something that we should take into consideration? That maybe this is partly because of the environment we're in?
- [30:15]
C
So. So absolutely. And in fact, every period that you've looked at, essentially equities has gone up. And one of the things that the betting against beta to remember is that notionally you're always long because the beta, in order to balance the beta exposure, you will always be long more the high quality, the low beta portfolio and short a little bit less in the high beta portfolio in order to get to the beta to zero. So certainly if you think the average there is an average earning on every equity, which is something to do with GDP growth, then that will perform overall. But the important thing is that you've constructed something that also has relatively low correlation to equity. So certainly this is kind of a way of harvesting idiosyncratic, like equity growth. They've gone back all the way, I think, to the 1920s, 1926 or something. Actually, the effect is not that Strong. Before the 1940, 45. They divide the period in the table, I think, going back to sort of before. Before World War II, the end of World War II, and. And then the effect is less, less significant then, and it becomes more. More significant. And then they, they in international markets, I think they start a little bit earlier, I think in the 80s or the early 90s. But I mean, the effect is that that equity has been very strong throughout the period and that factor stayed true. And I think kind of, I understand, you know, we understand a little bit about why it's making. And they don't try to hide it. I mean, if you look at the formula. Do you go for the formula? It's exactly that. What they expect to make is exactly the difference in the betas in terms of the. If you look at the formula, it's got beta high minus beta low divided by the residual volatility that remains, which is really proportional to beta high times beta low. So the math they're doing is perfectly correct. The model that they're setting up is perfectly reasonable, as Rob has pointed out. But I think people need to understand is like the way how does that. How is that relevant to our universe is also you always try to create a higher portfolio of sort of diversified trend. And maybe there should be a factor called betting against trend beta, which is, let's go long, the trend in the very diversified portfolio and short in the less diversified trend portfolio. So that I think, and I'm going to make the assertion without having a shred of evidence on it, that that will actually, actually do, do very well as a relative value trend portfolio.
- [33:05]
B
Betting against trend. That doesn't sound great in my ear. So I'm not sure I'm going to spend much time
- [33:12]
C
not betting against trend.
- [33:13]
D
I know.
- [33:13]
C
No, no, no, definitely.
- [33:15]
B
Fair enough, fair enough. Actually, on this slightly related topic about strong beta and all of that stuff, on Wednesday next week, we will actually publish a very interesting conversation that I just recorded with, with Harry Christlin and, and Jim Kasang where Harry is talking about a new paper that he published with Mike Green and a guy called Stefan Sturm, who you Know, has written a lot about, you know, the, the effects that passive investing is having on, on cap weighted indices and so on and so forth and yeah, so somewhat related but also very different but interesting nonetheless. Anything else we want to talk about in terms of this paper? I'm excited about the next paper. I'm hobbies as well, but I, I think I have more to say about that than, than beta stocks. And it is of course your paper that you co authored Trend Following Crisis alpha. Does it come from beta timing or market selection? So Rob, why don't you take us inside kind of the, the reason that you wanted to publish in the first place, kind of the thoughts behind it and, and, and frame the paper and take us into the page heap as well.
- [34:28]
D
Great, thanks. So actually betting against beta is an interesting segue because one of the things that we concluded in the paper is that if you're going to tilt a trend portfolio and keep the defensiveness of it, one way to do that might be to have beta neutral tilts within each asset class. And so the paper started from an observation probably about 10 years ago that what we were looking at the times when trend really did what we think of a prototypical trend following strategy when it did really well in the bad environments for risk assets. And what were the positioning of those strategies likely like through the lens of a simple model? And what we found is that the trend following sort of crisis alpha, delivering on crisis alpha seemed to occur when all of the positions within an asset class were pretty aligned. You were long haul all of the equity indices or you were short all of the equity indices, you were long all of the bonds or short all the bonds or mostly we started to ask ourselves, well, those cross sectional positions that are embedded in trend strategies, what's the value of those and are they also providing potential crisis alpha or not? And so we wanted to sort of explain that or examine that sort of rigorously. And the way that we did that in this paper is we basically built a simple trend following strategy and the positions were risk weighted and signal weighted. And then what we did was decompose that simple trend following strategy into beta timing exposure. So if you think about a trend following strategy, it's got a bunch of positions, they're long or short and they're risk weighted. And we can actually use a formula that looks like a linear regression formula to decompose those exposures into the beta component and then the residual, which is sort of a long, short tilt relative to the beta component. So the beta can be long or Short, basically, as the average of the trend positions in the asset class are long or short, and then the residual is tilting within the asset class. And we built the model in the paper to basically have it be a volatility neutral tilt. So you're long and short the same amount of volatility within each asset class in the till portfolio. And what we found was pretty interesting, and it was that the characteristics. So you've now got these decomposed portfolios and you can look at the historical return of them all. Sort of the portfolios are dynamic, but we're just examining the historical characteristics of the performance. And what we found is that the stereotypical crisis alpha returns seem to have come from the beta timing. Not a huge surprise. Trend following is in large part a beta timing strategy. And so most of the volatility of this simple model trend strategy was in the beta timing element. And less risk was taken in the residual component. But also the conditional equity correlations were more negative for the beta timing than they were for the relative value. And so we go a step further in the paper and we ask the question, well, what if we do something different in the relative value component of trend and don't use trend to do relative value? What if we use something like carry, and this is related to betting against beta? So if you want to do any of these signals, you might consider doing them relative to the riskiness. So carry over volume is likely the signal, and it's the signal we use in the paper to build these carry portfolios within each asset class. And they're largely beta neutral. They seem to behave in a way that's not related to the underlying asset class or to trend. And so when we do this within each asset class, what we find is that actually carry seems historically to have done a little bit better than trend. And so it's an interesting source of return. Carrie seems to have done better than historically, better than trend on average, and certainly better than the relative value component of trend. And so what if we swapped out the papers asking, what if we swapped out this relative value component and put it in some other relative value component that maybe could add to returns over time? The rationale for this was really the observation that trend can be quite episodic, as we know, the reason possibly that it's not a larger asset class or a bigger part of portfolios is that episodic behavior when stocks or the spy are up a huge amount every year for five or more years. It's hard to remember why you have an alternative in your portfolio. And so for that reason, it may be useful to try to embed something that could help in environments where risk assets are doing quite well. And so that was the rationale for using carry in the example. But, but the point is really that we may be able to improve on straightforward trend following strategies by having beta neutral, or at least market neutral tilts within each asset class where market neutral means equal volatility long and short. And so what we found when we wrote the paper was that the defensiveness is preserved even if you put something like carry in it. The condition for that being true is that the portfolio that you use to tilt has to be market neutral in volatility space. That's how you preserve the defensiveness of trend while also potentially adding value to it.
- [40:32]
B
What are your thoughts you have?
- [40:36]
C
I like the paper. I really like the paper. You know me, I've been banging the drum between, between the sort of the beta, the bitterness of macro trend, what I call it. So the way I see it is that a lot of what trend is about is about sort of macro level trend which expresses itself in multiple assets. And I think anything that sort of highlights that is a good thing and the paper goes about it in broadly the right way. So I think the observation that about 75% of the risk is in sort of essentially beta timing is valid, very valid and very important for us to understand why trend works. In fact, that's part of the reasons why we get good skew in alpha sort of crisis alpha. And that's really interesting. I found a very interesting observation that they make in terms of the beta timing correlation to equity in different asset classes. So a lot of the time we think about trend as oh, it's got crisis alpha mechanically. Why? Because if equity were to go down, we would also be shorting equity and we'll be making money. So a lot of people think about crisis alpha as that sort of mechanical. If the market goes down, trend followers will go, will go short and that will give you that negative correlation. What is very interesting in the paper, you look at the Exhibit 9, they look at the correlation for all asset classes. If I look at commodity trend okay, or bond trend okay, which has got no explicit exposure to equity, overall, it has negative correlation to equity. So CTA have negative correlation to equity not because of mechanical one, because you get the mechanic, you get negative correlation from other asset classes as well. What you're seeing here is that relationship between equity performance and volume. So what we normally, I think we believe we see is that trend following is a Little bit like buying straddles in each of the assets that you trend and therefore you are likely to make more money when there is volatility in the market because then your straddle becomes more valuable. And we also know that equity has got negative correlation to volatility. That's why the Saber model, that's why volatility is skewed. And that kind of expressed itself in the data. And I found it really a great observation by that team, by the Fidelity team, in terms of that negative correlation, which is not for mechanical one, but actually a long term relationship between the volatility and equity in the markets. So that was a really great observation that they made. I've got to tell you, thinking about the difference, the way that they constructed the long only beta versus the way that, that we will take into account when we construct correlation might be a little bit different. So for example, if I were to look at developed market bonds, you will find that Italian bonds and Canadian bonds behave very differently to the US Treasuries. So you can think of Italian bonds are actually risk on, whereas Japanese bonds are risk off, for example. Okay, so sometimes actually the correlation between the assets creating a single index, for example to commodities, probably not what I would have done. I think part of the reasons why they did FX developed markets and FX immersion markets separately is because these two asset classes, assets actually behave very differently. And the main factor is not necessarily a simple single main factor which affects these assets. But I think it's a great observation, it's a great paper. So one thing I wanted to ask Rob is when you look at the breakdown between beta trend and relative value trend, one of the things you notice is that relative value trend is actually very valuable. If the signal is strong on one asset within the asset class, it is likely to perform better than an asset with a weaker trend. And it still has negative correlation to equity. So it still has a little bit of sort of crisis alpha, which is a really nice feature. Why would you discard it completely? I mean, one would imagine that you should be able to combine it, probably leverage it up, because it's actually performing very well.
- [45:05]
D
You're 100% right. For the purposes of the paper, we wanted to just show that you could do something quite different and keep the defensive characteristics. In practice you have a lot of choices. You could choose to size it differently, you could choose to keep the native amount of relative value trend in the portfolio and just overlay something else over it or anywhere in between. But we just wanted to show that you could do something super different and keep the defensiveness as a result of that.
- [45:36]
C
By the way, your portfolio is probably more leveraged than a normal trend portfolio. If I think about the leverage of a trend portfolio, it's Normally something like 3.54x in terms of the risk. I presume once you put in a higher allocation to relative value strategy, it will increase, increase in terms of leverage in terms of trading costs as well. I presume could be.
- [46:01]
D
I think it cuts both ways because on average you're actually reducing diversification by tilting within each asset class. So I think that would push back on the amount of leverage you would need to implement. Just thinking about how the portfolio would be constructed, the, the relative value tilts applied to the asset class would actually reduce, would reduce the amount of diversification if it's very directional within each asset class. So I think it could go either way.
- [46:33]
C
Oh, I see. So Instead of holding $2. $1 in treasuries, $1 in gilts, you will be holding $2 in US treasuries.
- [46:43]
D
If you think about how mechanically you would apply an overlay to a directional portfolio, I think it could shake out that way in some cases.
- [46:52]
B
I just had a quick question actually, and that is, I mean, you mentioned, Rob, that trend following is known for kind of the episodic return streams that we have. And maybe that is one of the reasons why, you know, not more people have it in their, in their portfolio. And however, you know, Carrie is also known for looking great most of the time and then suddenly it has a really horrific period. And I guess maybe my, my question, and this is not necessarily, it's just an observation that I, I know that there's a lot of people who want to kind of remove the bad stuff from trend by adding other stuff to it. But you know, any, any thoughts on, on just other, other ways to improve trend or do we have to add other factors or signals?
- [47:47]
D
I'm sure that there are a lot of ways that strategies can be, you know, these are research topics we look at every day. We haven't stopped researching how to improve trend following. We were just trying to make this case that the, the defensive benefits of it can be preserved and you could potentially add other stuff to it without losing the thing that people invest in trend for. I was trying to be unorthodox and say you don't have to be pure trend to get the benefits.
- [48:17]
B
And I think more and more people are starting to talk about it in this way. It's really these regime changing, adaptable sort of Properties that trend following has, that really is sort of the essence or the benefit that people get from the approach. I guess it's also what you find.
- [48:36]
D
The other thing that stands out is that the prototypical behavior of carry that you mentioned and the one that many investors have been taught to believe is that carry always does poorly in risk off environments. And if you build a portfolio that's. So if you do em, carry emfx, carry just an EM and you go equal volume long and short, we don't see as much of this sort of bad crisis performance in the model in the paper as you might expect. That's another takeaway. If you build the cross sectional factor truly market neutral, truly beta neutral, you might get a different outcome.
- [49:23]
B
Yeah. And Easter, have you, did you find that there's a tipping point where yeah, you can do this much non trend and then you really need to stop or was that not part of the study?
- [49:37]
D
We wanted to just pick nice round numbers and put them in there. For sure you could run some type of analysis of where it becomes a carry portfolio. So I think that's straightforward to look at, but we didn't do it.
- [49:51]
C
Yeah, I think what we're doing in terms of portable alpha, essentially what we're doing here is taking a collection of strategies. So trend as a crisis alpha, we have relative value carry, which we know which actually Pedersen wrote a paper about that as well in 2012, 2013. And so whether the number is 1.68, whether it's 1 whatever, depending on the implementation, that's up to debate. But that's another form of alpha or at least sort of a risk portable portable alpha that you can add to your portfolio. And the question is how do you, how do you build them up together in a way that for, from an investor? The reason why investors actually like it so much is because it gives them, if you like, sort of more risk for what they are paying. Right. So in the same way that we said we investors would like to have sort of a beta, high beta stocks, they might say, well actually I can have 100% of trend but also on top of it another 30% of another strategy, be it relative value carry or be it sort of, you know, long bonds, be it, be it sort of long, long, long equity actually is the, the thing that most investors actually seems to be interested in these days that just gives them more, more risk. For investors that can do leverage. Right. We can do in our, in our universe it's, it's relatively easy to do. But that's exactly why, you know, the beta factor is what, why investors actually like it.
- [51:21]
B
Absolutely. No, it's great, great paper. Thanks so much for, for sharing that Rob. But we're not quite done yet because there was another presentation you have that you brought along that is really relevant for our industry for sure. But I will have to defer to you for the, for the details in terms of that. But it is a presentation given by a general called Rob Al Green as far as I recall. Yes. So what was so exciting about that?
- [51:54]
C
So I think we talked last time I was on ttu I talked about the way that we trade in, in the fixed income portfolio during the crisis, do you remember we had an issue which we have both the long and the short were trying to reduce size, whereas actually what you were trying to do is to reduce the, the first factor, your duration straight away during the crisis. So I've been thinking about multi asset execution and Gresham, we went to a talk by Rob, he gave a conference talk about how he's doing execution. He's running a company which is doing execution algos and he's talking about the things that he's is looking at. And I think first of all it's important for the listeners to understand how do they measure what a good execution is and a bad execution. So what happens? They get given an order and they get given a time frame to execute it. So let's call it, you have to execute it during a day. And in the simple case, suppose you need to execute a certain amount of U.S. treasuries and the question is what is the best way to execute it? And there are two things that are in conflict. You kind of want to be passive. You want the market to come to you because that will reduce your bid. Ask spreads that you're paying. On the other hand, the more you hold on to the order, the more sort of risk you're accumulating in a sense that you're increasing your volatility of your tracking error to some benchmark. So these days most execution algorithms are benchmarked against either their priorities price that they were given when the order was first received. So the price right now is 100. Can you please make sure that you don't stray away too much from 100 or something called the vwap or a twap, sort of an average price over the period at which you're executing. And that's kind of how you benchmark yourself. And I think Rob's observation, which is really interesting in his first, I mean he made this Observation many years ago is that most CTAs actually execute in a way which is less optimal. So what we normally do in a cta we would receive an order and we start by being passive and we are passive and we are passive until we run out of time, at which point we become less passive and then we run out after more time and then we become even less passive and we become quite aggressive and oh my God, the two hour window to execute is almost up and I'm doing the rest of my order at market prices. Okay. And Rob's observation is that it's exactly the opposite, right. The problem is that when you start running your order, you accumulating a lot of risk. You have a lot of risk on your hand because you've got 100 orders to execute, okay. By the time you're finishing, you actually have don't have that many orders. You're not accumulating volatility that much. So actually his observation, innovation, and I think he built his company based on that, is that what you should do is you should start immediately, cut the risk that you're holding in the order very quickly so you're more aggressive to start with and then once your 100 lots have reduced to 50 lots, now you can be more passive. So you can be more passive towards the end. So it's a very interesting observation that from the risk reward pattern from how much volatility versus how much passiveness you you have, it's actually better to schedule it the other way. Now these days you have lots of multi assets that you need to execute. For example, we have a relative value carry strategy that is sending an order and saying can I please be long US treasury, but long the two year Treasury. So I'm long the 10 year and short the two year. For example, I don't know who might be running relative value carry strategy, but I suspect somebody might be. Okay. And then Rob was talking about the difficulty that you have in maintaining the same framework of reducing the risk that you're accumulating while executing essentially pair orders or executing multi asset orders. And I think that what was very interesting and a really interesting observation that he made is first of all, the correlation between the assets are important. So if you're doing relative value, if you're doing a long long soybean and short oil, these assets are relatively detached. So it actually doesn't matter how you the order at which you execute, you can move to almost executing them independently. But as the correlation between the markets you have a higher correlation. You have to Start thinking about the spreads. You can take one size of the position, but you can't take the other side of the position. And that's actually makes sense because otherwise you expose yourself to beta that we've talked about and then your risk actually goes very, goes high. And the last observation that he made, which was really, which we saw, in which we saw in March, was that sometimes those markets will be open at different times or they will be the trading hours that you have will be at different times. So for example, if I looked at the fixed income portfolio, we talked about Japan and we talked about the US and we talked about, so what you, you might have a situation where your portfolio is short Japanese bonds, but going into the crisis, but long U.S. bonds. So now when you have a pair order, you might have a situation where one of your order is at a certain volatility profile during Asian trading times and then you have US times. So these are not solutions. The presentation doesn't give you the answer necessarily, but it talks about the considerations that you have to think about when you are trying to manage that portfolio over time. So I thought that was really interesting and really chimes with a lot of the things that we saw during the March crisis.
- [57:49]
B
Yeah, I'd love to have any thoughts you have, Rob, on this issue, but also like to ask questions, both of you. If, if, if your firms are at a stage where you basically let the systems go directly to exchange and execute orders, which some firms do, others don't. But what, what are your stand on?
- [58:09]
D
I'd like to comment on the presentation. I, we think a lot about execution and two things occurred to me as I was reading this. You know, one of the things that we have varying degrees of conviction in is the correlation estimates we have across markets. So I would not use this approach. Just, you know, just let it go and figure out, you know, how, how to execute. I think in the other part of the question is how much conviction we have in the views that are implied by what we're trading today. So if what we're trading today is just a modest rebalance based on what our models are telling us about risk and trend. Trend doesn't have a super high information coefficient. How strong is this view? If it's not that strong, do we really want to be aggressive ever? I would start there and then in cases where we're thinking about risk management in a crisis environment, I think those could be handled separately. Of course, if one market's open and we have high conviction that in this environment the correlation between a And B is high and one's open and one's closed. You can reduce risk by intentionally choosing the most available and cheapest to trade asset and then rebalancing around that position.
- [59:36]
C
I'm a big fan. I came onto the postcard, my first postcast here. I talked about multi day execution and saying you want to be passive as long as possible. Right? So I'm a big fan of that, especially in the alternative markets. I think to answer your question, Nils, I, I do want to answer your question. A lot of the markets that we trade are OTC and they are not, they don't, they don't necessarily trade on the exchange. And you can't really automate that. Certainly in the case, even in the case where we trade commodities which do have, which do settle on the exchange, let's lot of the time the, for the alternative market, the liquidity will be sitting off exchange. So that presents itself. I mean certainly, you know, Gresham as a whole has got a liquid commodity business as well and that one trades straight away to the exchange. It's all very, it's all very automated so to speak. But when it comes to markets in certainly in commodities, in freight, you just have to, you really have to speak to the brokers and execution really needs to be handled. I think Rob's observation is spot on. And in fact I wrote to Rob and I said, I think your correlation estimation, the fact that you're kind of assuming the correlation is static, I think is an issue and I think, but in some way it actually helps you. So it's exactly during the time of crisis what we observe is that the first factor actually becomes very dominant and it's not surprising, right? So if you like, there is a, there's a spike in correlation and we see that across, we've seen that in March where actually different, the liquid markets become actually better substitute for the less liquid ones. Because normally you might say the correlation between US Treasuries and I don't know, Zar swaps or Taiwanese dollar swaps, the correlation is actually fairly low. In fact, the whole point that we are in the position is because the correlation between those assets is low. That's why we, we look at alternative markets. Okay. But during a crisis that correlation begins to spike. And that's the time exactly when you think, hang on a second, I want to risk manage that, that, that position. But I think your, your observation that it's that there needs to be a sort of some sort of intelligent decision whether you actually want to manage it. Asset by asset and the correlation, whether the correlation is static or whether it is a crisis period where you think actually this is not an alpha, this is a risk reduction exercise and therefore I'm going to manage it using the more liquid assets. I think it's something that you will find most discretionary macro traders would have done. I mean if I was sitting at a macro shop during COVID I was sitting in various macro shops in the past where there was a crisis and that would be the natural thing what a macro trader would do. It would pick the first most liquid assets, be it the S and P, be it crude, be it US Treasuries to reduce its overall duration or its overall beta to equities and then we'll work around it as the view becomes more nuanced, right in a way that oh, which equity during COVID is going to suffer more? Which countries are suffering more from COVID Then it becomes more. But the first order of the day is actually helped by the rising correlation between the assets that you're trading.
- [62:56]
B
Any final thoughts from you, Rob?
- [62:58]
D
Or the other thing that stands out as important is that that correlation estimate we're using to execute could potentially be stale at the exact time when we need it to capture the fact that there's are opportunities to hedge in more liquid markets. The other thing is transaction costs go up a lot when volatility goes up a lot. So accounting for all of these things at the same time is super challenging. But I really like the exercise of thinking about how we can find an optimal path to execution. Lassie Peterson wrote a great paper about this as well.
- [63:35]
C
Yes it did. Although actually if you're talking about correlation, intraday correlation, I think Neil shepherd wrote in his time using the intraday correlation as actually as a risk gauge. So what you find is that when there is a spike in correlation, that's generally a risk off signal because it's an indication that the market is becoming dysfunctional as we've seen. So. But yeah, I'm 100% with you. It's not a solved problem by any means.
- [64:07]
B
Good stuff. I think we'll leave it there for today. We've covered three different topics and really appreciate that. And of course people should go and follow both yours and Rob's work and see what else they get up to and I'm sure they can get a copy of of the paper you, you talked about today, Rob, by him making contact with you. Let me ask you on that note, is there any particular place people can approach you on to get a copy of the paper. Is it published on your website?
- [64:42]
D
Sure, the papers@institutional.fidelity.com Fantastic.
- [64:46]
B
Thanks so much for that. And I'll also mention it in the show notes. Next week I won't be here, but Moritz will be sitting in for me as I head back to HQ in Sunny Florid and he'll be joined by Nick Bolters. So no doubt that will be another fun and insightful conversation. So make sure if you have any questions for Moritz or Nick that you can send them to us. As usual, infotoptraders unplugged.com and I'll do my best to make sure it gets covered next week. With that from Rob Yorov and me, thanks ever so much for listening. We look forward to being back with you next week. And in the meantime, as usual, take care of yourself and take care of each other.
- [65:28]
A
Thanks for listening to Top Traders Unplugged. If you feel you learned something of value from today's episode, the best way to stay updated is to go on over to your favorite podcast platform and follow the show so that you'll be sure to get all the new episodes as they're released. We have some amazing guests lined up for you, and to ensure our show continues to grow, please leave us an honest rating and review. It only takes a minute and it's the best way to show us you love the podcast. We'll see you next time. Next time on Top Traders Unplugged. This podcast expresses the views of its hosts and the guests appearing on the podcast as of the date of its recording, and such views are subject to change without notice. Top Traders Unplugged do not have any duty or obligation to update the information contained herein. Furthermore, Top Traders Unplugged make no representation to its accuracy and it shall not be assumed that past investment performance is an indication of future results. Moreover, wherever there is a potential for profit, there is also the possibility of loss. This content is not is made available for educational purposes only and should not be used for any other purpose. The information contained in this podcast does not constitute and should not be construed as investment advice or an offer to sell or a solicitation to buy any securities or related financial instruments in any jurisdiction. Certain information contained herein concerning economic trends and performance is based on or derived from information provided by independent third party sources. Top Traders Unplugged may believe that the sources from which such information are obtained are reliable. However, Top Traders Unplugged cannot guarantee the accuracy of such information and has not independently verified the accuracy or completeness of such information or the assumptions on which such information is based. This podcast, including the information contained herein, may not be reproduced, copied, republished or posted in whole or in part in any form, without the prior written consent of top traders.
- [67:00]
C
Unplugged.