So, I actually am a little bit afraid that we will be essentially getting more people saying, oh, CTAs are a Sharpe 0.5 strategy, and that, to me, is actually a danger. I think CTAs are a wonderful strategy. It has, you know, positive skew, positive alpha. We can create beautiful portfolios. And I'm kind of afraid, actually, of that happening. Imagine spending an hour with the world's greatest traders. Imagine learning from their experiences, their successes and their failures. Imagine no more.
Welcome to Top Traders Unplugged, the place where you can learn from the best hedge fund managers in the world so you can take your manager due diligence or investment career to the next level. Beforewe begin today's conversation, remember to keep two things in mind. All the discussion we'll 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. Here'syour host, veteran hedge fund manager, Niels Kaastrup-Larsen.
Hey everyone, and welcome to another edition of Top Traders Unplugged, where today I'm joined by no less than three wonderful guests from the world of systematic trading and trend following, namely Graham Robertson, Rob Carver and Yoav Git. Now,for those of you who are wondering what the connection is between our guests, well, they all used to work together at AHL, so this will be the second time I sit down with three people with routes leading back to AHL.
The first time, of course, was in fact with the founders of AHL when we sat down in Abbey Road Studios, in London, to record a conversation for their 30th anniversary. A three-part conversation I hope you may be inspired to check out after you've listened to today's episode. Gentlemen,welcome to the show. How are you doing? Where do I find you? Graham, where are you today? Niels? Yes, Hi. I am sitting in London. It is, I think, about the fourth or fifth absolutely glorious day we've had.
People are sitting by the river, enjoying the sunshine. Spring is in the air. I'm looking at two wonderful former colleagues here and very much looking forward to the conversation. Me too. Rob, what about you? I recognize the place, but of course people can't see you. That's where I usually do my podcasting from. So, it's the shed, the glorious shed in my garden, which is probably about 30 miles east of where Graham is and about 30 miles southeast of where Yoav is.
So, we're kind of forming a geographical triangle. Theweather here is also glorious, and, in fact, I've just been for a very pleasant bike ride and feeling refreshed and very much looking forward of the conversation. Fantastic, I like the preparation that goes into this, Yoav. We've kind of had an indication now from Rob where we might find you. But for those who may not be that well versed in UK geography, where about… So, yeah, I'm actually in Cambridge.
I moved to Cambridge as an undergraduate and fell in love with it. So, I've stayed in Cambridge ever since. My wife is a Cambridge researcher as well. I’m enjoying the sunshine and very much looking forward to discussing CTAs with my former colleagues. Yes, well, me too, me too. I was just saying to Yoav, when I heard he was based in Cambridge, that I don't know if you've watched this series on Netflix.
I think it's on Netflix or Apple TV, actually, it's probably Apple TV called Prime Target, but it's recorded in Cambridge and yeah, it's pretty good, actually. I was more of an Inspector Morse fan, that's all I'm going to say. Yeah, I watched this crime series called Grantchester, which is also, obviously, set in Cambridge, which is a bit more gentle and old fashioned, I suspect, but still quite nice. Fair enough.
Now,since Rob and Graham are well known to our audience, I think it would be polite to ask you, Yoav, maybe to give us a little bit of background to your road into trend following, maybe how you ended up at AHL, and what you've been up to since leaving AHL. Oh, thank you. So, I've actually had quite a diverse career. I've been an academic. I started my life as a maths lecturer at Cambridge and also, I was an associate research fellow at Imperial. Then I went to the sell side.
Iwasa risk quant at Credit Suisse, and then moved to the truly dark side and joined Brevan Howard as head of Research of Development in Israel. And that really started my journey into systematic. And I moved in 2007 to Winton, also an AHL child, and was head of fixed income in Winton. Andthen I joined my two esteemed colleagues here in AHL in 2011, I think, firstly looking after the fixed income portfolio in the Dimension product, which is the multi-strat arm of AHL.
And then in 2013, when Rob very kindly decided to leave AHL, I took over his job as the head of fixed income for AHL as a whole, and then later, as head of innovation in AHL. After that, I actually moved to become an independent PM in another macro hedge fund called Element, which specializes in macro and volatility. And since 2022, July 2022, I've been building my little baby called SAFI, which is a fixed income CTA. Fantastic.
Good stuff, all right, okay, so just full disclosure for our audience. The topics today were, generally, conceived between the three of you. People will be able to tell, afterwards, whether they conceived before you started having pints or after you started having pints. Oh, come on. It was coffee. It was over coffee. It was coffee bar. Very civilized. Okay, fantastic. Oh, then we'll definitely know. But let's just kick off with kind of the first.
And obviously, with a little bit of background, you know, you ended up working together at AHL. I'm kind of curious, in one sense, kind of how you work together, but also, you know, what were sort of the main themes back then that you were working on, and maybe some of the breakthroughs that you discovered while working together at the time. I can kick that one off to start with.
So, I guess my first recollection was really I was tasked with looking after the fixed income side of things when we were split by research topic. And I have great memories of sitting in meetings with fixed income meetings led by Rob. Itwas, at first, really cutting my teeth in CTAs; how we think about things, how we think systematically, which was very different from me to the environment I'd come from.
So, I’m very much indebted to Rob for teaching me in the basics, if you like, in the essentials of what I know. Andthen, as you have alluded to, Yoav picked it up when Rob left. And I think, as we'll find out today, Yoav's take on many things is unusual, different, and certainly worthy of thought. Yeah. What about you, Rob? They all talk about you as being the great master here. Well, no, I mean, I think, let's be honest, it's probably a case of being in the right place at the right time.
So, as Graham sort of hinted at, AHL prior to 2011 had, what I call, a sort of strategy focused split. So,there was a team running the trend following, which is obviously a very big team. Then there was a dream team running relative value. I was running a systematic, global macro team, and so on. And then they decided it would make more sense to organize things around asset class. And I was put in charge of fixed income. So,I just happened to be in the right place at the right time.
It was a very diverse asset class because as well as, obviously, bond futures, interest rate futures, we also had cash bonds, we had credit default swaps, we had interest rate swaps, we had some relatively unusual things which were in both the Evolution products, which is the sort of alt trend following product, alt assets trend following, and also in Dimension, which Yoav mentioned, which was the kind of things that weren't trend following. So, the alt strategy, if you like, view.
So, we had a very diverse and interesting group of assets and strategies to play with. AndI was very fortunate to work with some people who knew a lot about these individual buckets, people like Yoav, for example, who had exposure to fixed income outside the CTA world. They were able to bring in a different perspective as I had as well.
Soyeah, it was a very challenging time to be in that particular asset class because of course we had the QE, we had the taper tantrum and it was all terribly exciting. But yeah, a great bunch of people to work with, definitely. So, I think what was interesting at the time, also, was that liquid trend sort of stopped working for a while and that put a lot of pressure on the company because a lot of the earning was tied to the liquid CTA program.
And then they noticed that, actually, Evolution was still printing money very nicely. Andthere was a huge drive that I think Rob started leading in terms of increasing the AUM, increasing the diversity within the Evolution program. And my focus, when I was running the fixed income, was actually adding more diversity and really launching the Frontier AHL, Evolution Frontier product.
So,Andrea and Juliana were in the fixed income team, but we still led innovation into TBAs, into swaptions, into lots of very good, interesting things that just add diversity to your portfolio. So maybe a question for you, Graham, since you're still with AHL, it sounds to me… So, I come from the world where research is done on the portfolio, meaning the researchers will develop models that basically are just applied to all sorts of markets.
It sounds to me a little bit like you have a slightly different approach here, or maybe I'm misunderstanding it, that you kind of focus on asset classes, develop models for specific asset classes and maybe not so much models that are universal. Is that correctly understood? I think I would probably say it a little bit differently. I mean, you've obviously got the two threads, you've got the model research and the markets research and both are important.
ButI think you have hit the nail on the head there. What was really key at the time and, really, at the time that the three of us were working together was that traditional trend was struggling. Things were range bound, you could argue. Central bankers were all saying whatever, and equities were rebounding as a result. It was a difficult environment for trend following. The Fed put was really in its nascent beginnings.
Andwhat we certainly noticed, and as I said, you have alluded to it, is that there were a lot of markets out there that were far less affected by that. They were generally harder to access, they were generally OTC markets. And this, in a nutshell, was the Evolution program. So,I would have said, at that point, the research was very much focused on markets. Markets were clearly different. Researching new markets was genuine value add.
And that really helped in that I think people call it the CTA winter now, that diversification in terms of price drivers markets was really helpful to us. Yeah. Well, before we leave this topic, which is a little bit of a background in a sense to how you guys work together, this is one of my small contributions to the questions and that is, is there something that most people who have worked at AHL you think might have in common? It could be like a belief or certain way to think about markets.
Onething I have noticed they have in common is that when I walk around the offices of AHL in London, they all come out with very nice merch. They wear the same nice, you know, vests and T-shirts with Man AHL on. But other than that, what would you say, what's common about people who leave AHL because it's such an important institution for our industry in Europe? I've got a take on it which the others might disagree with.
I mean, obviously everyone who works from AHL is very good looking and very smart. But actually, in all seriousness, AHL tends to employ people who are quite nice, which can be unusual in finance, and who are good working together and are very collaborative. And it's the kind of place, at least when I was there, and I think it's probably still the case, where people are very open about ideas and research and where there's a lot of kind of open discussion of ideas and things.
And so, I think AHL tends to attract and breed people who are kind of very sort of openly intellectually curious. Andthat's the one thing I've certainly taken with me sort of away from that is, in the last 10 years, when I've sort of been doing my own research mainly, I sort of feel like the AHL trained me to be like a really good researcher. And in terms of being open to ideas and sort of in touch with being curious.
And that's because we had these open discussions and it wasn't like the kind of place where everyone's sort of sitting in their silos, visually blinded and not knowing what's going on around them and not willing to speak up and kind of question ideas. So, that's my take anyway. We didn't have the merch back then, I have to say. That's a new thing, I think.
There are two things that I notice when I work with other researchers from other CTAs or other non CTAs as well is quite a bit of structure and willingness to think about the mathematics, to think about doing something properly. A lot of research is very hacky in many cases. Andyou get the discipline in AHL, you get the training into discipline about thinking the problem in a correct framework. And that has been very good. The other thing which I found is the open source in AHL.
I must admit I was infected by that at AHL. I wasn't a buyer in the first instance, but I think the willingness to give back to the community is something which is very important and it's something that I really value about AHL. Yeah, maybe I can just add one thing to that as well. Touching on what Rob said. I mean, I'm obviously in my 14th year here. One thing that struck me, back in 2011 when I started, was it was the most academic commercial institution I'd ever been at.
Obviously, I've got an academic background, but I'd been in various funds and various banks up till that point. And I just found that the academic open discussion, that environment was just really suited to broad thinking. And that's just been a great place to be. All right, let's jump into sort of the first real topic that you thought would be a good thing to discuss.
And it relates to this idea of whether to focus on a single asset class or whether you should focus more on sort of a multi asset class perspective. Andactually, where you are today, at Gresham, is a kind of a good example of a slightly unusual maybe setup. And that is where you kind of offer products that are asset class related. And then of course people can mix it. It's kind of building blocks.
Butone is focusing on one area, kind of the financials, and one is mainly the interest rate and currencies, and then the other one is on commodities. And actually, we had your former colleague, Scott, to talk about that on one of the other series. But this idea about pros and cons, looking at one or the other, who wants to kind of set the framework for this particular topic.
There is a tendency in CTAs to think of all assets are the same, all asset classes are the same, we're going to all put them in one big bucket and get on with it. Because mathematically these are all just prices on the screen. The real question is, are all asset classes equivalent? AndI think the most important place where they defer is not really about us, but about the utility they offer to allocators. We don't sell, we don't sell a CTA just as an absolute return, you know, pure alpha.
But we also say, well it offers convexity, it offers crisis alpha. And that relies on what the actual allocator holds right now. IfI'm a bond portfolio holder and I worry about bond convexity and twice as half of a bond, then what do I care about coffee earnings or coffee convexity or cocoa? I don't really want a portfolio where only 25% of my portfolio is in fixed income. So, I think it's important to give allocators what they need.
Ithinkit's really important to ask the question of how do we add value to the investors. And in the case of CTAs, I think we generally think about hedge fund as just a standalone absolute returns. But when we think about the crisis alpha it offers, we need to think about it in terms of the asset classes that the investor actually holds. And that's why we think this offers better building blocks to the investor. But it's also actually important in terms of implementation.
So,I think that alludes to your earlier question. Different asset classes actually are very different, and I think I'm going to refer to Graham. Graham talked about it beautifully three weeks ago, on TTU, you spoke about crypto and you and Andre put out a very good paper about why cryptocurrencies actually trend very well - because there is no valuation. So, I completely agree with that. This asset class has a situation where it actually trends very well.
Andalso, the dynamics of the way you do the risk control on it, the way you manage the position, actually needs to be different because it has enormous right skew. And the way that we implement trend in different asset classes, we can take the generalist view, but then we will be leaving money on the table. We will just not be able to harvest the full alpha in a particular asset class.
So,I think it's important to put the effort in each asset class to implement it correctly and to just get this extra alpha because we know that the average market Sharpe alpha isn't particularly high in trend. It's like 25 basis points. If we can get it to 30, that really adds value to the investor. So,we try to do that, giving to the investors what they need and also giving added value by being careful about the way we implement it.
So, before I come to you, Graham, maybe Rob, because I have a feeling you might see this slightly differently. So, I'm curious in your thoughts on this. I mean, I think Yoav's confounding a number of different things. So, one is, should a CTA be the sort of entity that decides what the correct weight is between bonds, equities, FX, crypto, whatever.
And I think it's valid to say if it's a standalone entity that you're buying as a standalone unit, then yes, the CTA should decide that because they can do the portfolio optimization. They probably have a better understanding of how those different building blocks work together. Butyes, if you're buying it for crisis alpha or tail hedging properties, then sure, it makes sense to give people the option to invest in the individual units. Fair enough.
A second question then is whether different assets trend better than others? As people probably know, I'm quite skeptical on that. I'venever really seen a strong statistically significant effect in performance between different asset classes. And frankly, I think crypto has not really been around long enough to give us that kind of statistically significant difference. For sure. Tobe honest, I'm quite skeptical on that. There are, you know, you do see differences in performance.
So, I'm just glancing at, sorry to plug it, but a table from my latest book. And I can see, for example, that equity trends (it’s sort of well known) doesn't do as well as, say, I don't know, trend in VIX or trend in fixed income. And maybe there are good reasons for that. Andmaybe, you know, things like, for example (and Yoav might get onto this), the fact that carry is naturally higher and fixed income is an explanation of that potentially, but I'm a very skeptical about that.
Andthen the final thing is, should you sort of trade crypto differently, for example, because it's got different, say, distributional properties? Well, frankly, I think that's the sort of thing you can deal with without a universal risk model that just takes higher moments into account. So,I don't think it's something you need to build a completely new hedge fund for just to trade crypto because Crypto's got the different higher moments to say equities or fixed income.
So, I kind of agree with some of what Yoav, I said, but a lot of it I'm disagreeing quite strongly on, frankly. Where do you sit on this, Graham? I think Rob's asked all the right questions and I'm looking forward to hearing Yoav's response. Fair enough. That's very fair. So, I think what we find is that the implementation actually does need to be different in terms of… Let's take an example. Interms of portfolio construction.
For example, when we run the commodity portfolio, then those assets are very naturally diversified, they're very much uncorrelated. And the way to achieve a diversification there would be something like a one over N, like an equal weight across the markets. Butthen when we come to fixed income, we have a much bigger issue because there's a very high loading on the Fed as a risk factor. So, in order to get a diversification, we need to do something a lot more clever. And that's just one example.
Butalso, in terms of risk management, this is not just about alpha. This is also about the way you risk manage a portfolio. In a normal CTA you would just take one over vol and be done with and be very happy with it. Butactually, if you think about a fixed income portfolio, it's got very strong left skew. And you think about the way that you manage risk in a multi strat, because a multi strat is essentially a collection of lots of pods which are negative skew.
And you have to have a much tighter risk control to ensure diversification, to make sure that you do not get hit when things go south. Andthat is something which is very important in the fixed income universe, but is perhaps less important in a naturally diversified portfolio in commodities, or maybe done differently in equities. I'm not arguing, necessarily, generally, that the alpha of the underlying market is different, although actually I think there is a difference.
Thereis a very strong difference between them. If you look at AHL and you ask them what has been your biggest earner for the last 25 years, about 70%, 75% of the P&L has come from fixed income. But even just the implementation, just the risk management execution needs to be tailored to the specific asset class and you need the accountability of the person who's running it to actually do it properly. Okay? So,if you think of execution, fixed income is naturally liquidity is very fragmented.
You have multiple bonds for the same issuer, whereas in equities you have just one stock for a company. And that makes a difference in terms of the way that you execute. So,in execution, in risk management, in portfolio construction, in all of those situations you actually need a specific implementation, even if your underlying rationale of trend, of diversity, are all the same across all asset classes. So let me just ask one thing, Rob, and then I'll let you jump in here.
You mentioned this thing about that most CTAs have made most of their money in the last 25 years in fixed income. So, I think that's pretty universal. ButI think that there really are for two reasons. One, fixed income is generally the largest sector in any CTA portfolio just because it's super liquid and there are many instruments we can trade across the globe. And secondly, we've had a very nice environment, overall trend, in fixed income for the last 30, 40 years.
Ifyou look at fixed income returns for CTAs, and I'm speculating, but I'm kind of looking at what I can see from my vantage point. Fixed income hasn't made a lot of money for the last couple of years. So,I wonder if your data set is (and I shouldn't argue with someone who's got many more degrees than I have), is it a little bit optimized to say, yeah, fixed income has been the best in the last 25 years, so we should focus on that.
So, I actually don't necessarily think this is just fixed income. And even with fixed income it's not at all times. But I think it's also the argument that the reason why fixed income has performed well is because we've seen a secular trend in rates coming down. That sort of expired around 2010, I would say. So, sometime after GFC, rates have really come all the way down. So that trend has stopped, essentially.
Andyet, since 2010 to 2025, so the last 15 years, fixed income has been very consistent, not just in terms of consistency over time, so, different periods in time, also different speeds. It's actually the most consistent across all asset classes. So,I think it's nicer to look at FX and see where it isn't working and to understand the difference.
Because when you look at FX, FX has not been such a good trend follower, especially in the DM universe, because you have seen no interest rate gradient. Okay.The way I think about what causes trend is sustained imbalance, the sustainable imbalance in the asset class. So, what I mean by that, if I think about crypto, this is in a permanent state of imbalance. There isn't any situation where I say this is where the true value is up 20%, down 50%. What's the true value? We don't know.
And that allows big trends to form. WhenI think about fixed income, I think about interest rate gradient. It's like a temperature gradient between two places where I've got two players, two central banks who are keeping a metal bar, at two ends, at two different temperatures, and that causes sustained and sustainable cash flows, essentially, from one side to the other.
Andthose central banks don't mind that because one of them is worried about… Japan is worried about boosting the economy and adding liquidity and Australia might be worried about inflation. So, they don't mind that carry trade, and that allows for sustained flows, and it allows for sort of positive auto correlation, and that's what we see in the data. So,we understand that. We understand that quite well. So, in FX when you see Euro/USD rate you see literally no trend.
Trend doesn't really, particularly, work very well. But when you see EM currencies where you see very strong carry, you see actually very strong trend as well. So,it's not that a particular asset class is always the better, but there are certain imbalances in the economy, and if we can keep them sustained then that generates long term trends which is what is very useful for trend followers. I'm a bit skeptical about Yoav’s theory, I'll be honest, but it's a key theory so I'll leave it.
Let him have that. I'm just kind of circling back to the points I made before. I still think that I accept execution. Yeah,but you wouldn't have the same person executing cash equities as you would have executing cash bonds in a CTA. So, I think that's a sort of a reason to kind of go completely on a single asset class fund.
Andagain, I still think that there's a lot of work you can do with generalizing your portfolio optimization risk models so they can cope with these characteristics of different asset classes and things like different levels of diversification without again needing to completely isolate it as a single asset class. But anyway, we should probably move on.
Well, I mean it's kind of related because you have always talked about emerging market currencies, for example, are different than developed market currencies, and so on, and so forth. So, it kind of comes to the next point which is, you know, liquidity and how that plays into different kinds of alternative markets and whether or not, as some have argued, that less liquid markets, more alternative markets, sort of trend better.
Anyways,love to hear… I don't know who, of you, wants to take the lead on that, but it's certainly an important topic. Okay, so I think liquidity is important in terms of capacity, and in terms of our ability to achieve diversification in our portfolio. I don't think, necessarily, that liquidity is an explainer of performance per se, except possibly in commodities, but then it's not the way that Rob measures it.
So,Rob, if you don't know, Rob wrote a beautiful blog, just recently, looking at liquidity and seeing if it correlates with performance, and it doesn't. I pretty much agree with his conclusions there, but I think he's measuring the wrong liquidity. So,for example, in the fixed income markets, the thing I think which is important is about that imbalance. And that imbalance is to do with central banks and nothing to do with the liquidity of the underlying market.
WhenI look at equities, I think the thing that really harms equity is valuation and the fact that value traders inject negative auto correlation which makes it very difficult for trend followers to harvest trend in equity. In commodities, we actually think about imbalance in the way that we look at supply and demand imbalance. It takes time for supply to meet demand.
So, if energy requirements in the US, because of AI, grows 5% next year, it will take time to build power plants, and to get to upgrade the electricity grid, and price needs to meet that demand. There,I think it's very important to think about the liquidity in terms of what is the size of the physical market versus the size of the financial, the financial size.
So, it's looking about comparing sort of who are the hedgers, who are the ones who actually invest in building power plants versus the size of the financial market that we're trading. So, there I think liquidity takes an important aspect of where the alpha is coming. Butgenerally, we would like to think about liquidity as a way for us to manage diversity. And it's a restriction for us in terms of being able to actually diversify the portfolio properly.
And you just have to either accept that you cannot put a lot of risk into those small diversifying markets or restrict your own capacity to allow yourself to get the diversity and to access those small markets. If you think about the participants in markets, I mean, obviously, liquidity, size of markets, but who are the players in that market? Does that make a difference, particularly on the fixed income or alternative market side? So, certainly in terms of execution.
So, easily the way that you get better execution in fixed income is really understanding who is the one who is actually holding the inventory. Who are the biggest players in that market? If you just trade with the big banks, then you're just going to have a lot less liquidity and when the crisis comes, you have issues.
AndI think that's work that both Rob and I did in AHL, and we're doing elsewhere, in terms of being able to access multiple brokers and also understanding, but also in terms of adapting your trading system to account for that.
Soyou know, SAFI was designed from the ground up to be able to, you know, when you trade a 10 year Mexico bond, you need to be aware of that you are able to accept fill, which is partial fill because your counterparty doesn't have the inventory or you are able to accept fill in a different bond. It's not just a 2033 maturity, or the 2034 maturity, or the 2035 maturity. So,accessing liquidity in different asset classes is different.
And in fixed income though, it's actually quite an acute problem when it comes to cash bonds. I think trend following is interesting, because I think when people think about liquidity in trading, they're sort of thinking about things like spread trading or RV where you have this cycle where people rush into a market, and spreads compress, and people need to pump in leverage to get the same yield. And then, you know, everyone's then rushing for the exits at exactly the same time.
Whereas, you know, trend is less dramatic in that sense. Imean,you know, we're not all trading in a way that…, In a sense, actually, the more the market goes up, the more opportunity there is for the market to go up more. That's the way we're thinking. So, there's not this direct inverse correlation between how much money we've made from the market so far and therefore how many people are piled into it and therefore the returns going forward.
So,I think the kind of alpha liquidity story, in a market that you're trading for momentum, is probably different than it is in a market that you're trading sort of in a reversion. And so perhaps something that is interesting is what the kind of proportion of players is in each of those markets. So, I would imagine in single stock equities there's a lot more people doing stat arb type strategies and playing reversion trades.
Andthe same would be the case in fixed income sort of curve trades and things like that than there would be in, say, I don't know, pick a random commodity market like gold, or crude oil, or something like that. So that's probably an element that confuses the picture slightly. AndI think the way that we think about liquidity should be very different from the way that a lot of people in the hedge fund industry are interested about liquidity because we're looking for different kinds of things.
I'm not sure if I've articulating myself very well, but there's a bit of nodding going on. So, it's making some sense, I think. I would agree. I mean, certainly one of the things that we've noticed, I think over the years, is liquidity is a term that's often misused. You can talk about the size of the market and people will call that liquidity of the market, but it's not necessarily the same thing.
The way the market is made up here, the industry players dominating it, what are the transaction costs like? All of these things can affect the way a market trades and performs. And it's an important point. It's not just a simple how liquid is the market point. I mean, even actually, just in futures, there are two measures of liquidity, right? There's volume and open interest.
And you can have markets with incredibly high volume, I think Korean stock indices, for example, and quite low open interest. And then you can have markets with huge open interest and quite low volume. And that would be true of interest rate futures markets tend to look like that. So, you know, both markets can appear very liquid, but actually their liquidity is in very different ways. Inone place you've got a huge stock and a small flow.
In the other case you've got massive flow and a small stock. So that's going to give you a very different indication of what the liquidity really is. And it'll depend on whether you're in the sort of position building stage or whether you're in the liquidation stage of your trend.
Which leads me to ask, in thinking about liquidity, so we've heard obviously in the very liquid futures markets, even that can be difficult to sort of completely get a handle on, but probably easier than in the alternative markets. I mean, if you deal with markets where you're relying on counterparties and all of that stuff. You mentioned this thing about if something goes sharply in one way, maybe there is very little. So how do you even estimate liquidity?
How do you even think about liquidity in markets that are like off exchange, and so on, and so forth, and you don't have to give any secrets away here just generally how should people think about that? So, we do have some surveys which give you overall liquidity, and I think things after GFC have become a lot better in terms of reporting of transactions. So that's available these days.
So,certainly, when you started, in the early to 2010, you had very little indication and you really relied on your brokers to give you an idea about what the liquidity in the market. It's a lot better now because most of the transaction report in the good old days interest rate swaps were bilateral, now a lot of those clear on an exchange. So,there's been a long road to actually get more visibility on liquidity in the market.
But I think we find the relationship with our brokers is still extremely valuable in terms of understanding what's going on. In terms of actually managing crisis and when things go wrong, I think the trick is really to have a properly diversified portfolio before the crisis happens.
Ithinka lot of the time, if you're in the very liquid futures CTA your mindset is, well, if there is a high volatility I'm just going to close the position, I'm just going to trade, volatility is going to spike, I'm going to go down. The signal is going to be… The problem that you have is that on day one you're just stuck with the position, and you have to manage that risk.
So, really, when you construct a portfolio, you just have to be a lot more mindful about risk management than you are in perhaps a market where you think you can get out of it very quickly. I want to transition to the next topic just being mindful of time already.
You know we've discussed it, certainly Graham and Rob, over the years, but maybe it could be very interesting to hear your thoughts initially, Yoav, on this topic and that is why CTAs are kind of maybe not represented, as well as we would all love, in portfolios, and why you think that might be, and I’m curious to know whether you have a different thesis than what we've discussed over the years.
So, in many ways, if you are sitting here mathematically and you do the maths about improving Sharpe, we are completely underrepresented. And I think Graham wrote a beautiful paper, a very short paper, about what is the right allocation if you want to add it to a 40/60, but I completely get it. I completely get it. And I actually think that CTAs are very difficult to allocate to and the reason is because we are neither fish nor foul.
Andthe point is that if I'm looking for convexity to my underlying portfolio, then CTAs are a very crude, fuzzy way to get convexity because it is really not tailored to my underlying portfolio. I think it would be much better to have, say, an ETF which is replicating put on the Treasury. That will be a much better tool for an allocator to get that convexity into his book.
Andif you try to convince your investment committee that you want to buy CTAs because of convexity and crisis alpha, and then a crisis happens, and we don't actually deliver because some other asset class behaved differently, and then you look like an idiot in front of your investment committee, that's it, it's game over. So, it's not a very precise way to provide crisis alpha.
Andconversely, it's also not very good in terms of sitting inside an absolute return bucket because you will be that kid in class that nobody wants to talk to because your Sharpe is simply not good enough. If you're sitting in an absolute return, then we are looking for returns.
Andif your Sharpe is 0.7 and the other kids in the class have a Sharpe of 1.5, and if you put together private credit, and private equity, and multi-strat, and long/short equity, and that thing has a sharp of 2 and you're sitting on a sharp of.07, then adding risk, adding a CTA is actually very detrimental in terms of dollar terms. Now, it will be great from a Sharpe point of view, and if you could leverage then that will add value. But most allocators just don't have that way of leveraging.
They are just stuck in terms of, you know, what they have. Andalso, you put this 0.7 strategy in and every other year you have a very difficult conversation about why is this guy the lowest performing sort of member of the class? Why is he still in the class? So, I completely get why we're not getting allocated.
Frankly,it's very difficult for us to live in the absolute return and we need to get our Sharpe up to sort of 1.2, 1.3 to sort of be in the same class as the rest of absolute return providers or do something better and much more tailored on the convexity and crisis alpha. Graham, can you help Yoav out with this? Look, I think the learned former colleague has given a great answer and it's quite technical when it comes to that. And larger investors get it.
Larger investors tend to have an allocation to trend following thick and thin. And then there's an element of family offices where this definitely comes into play, for example. I guess my bugbear with a question is even at the wealth end of the spectrum and I just think back to, I think it was a couple of weeks ago, there was an article in the Financial Times about how should investors deal with volatility? There's volatility in the horizon, how should investors deal with it?
I was shocked that trend following wasn't even mentioned. Andthat to me, I mean, if I had one crusade that would be it is to get trend following more established in real people's portfolios. I mean, don't get me wrong, I think it's a mistake and they should allocate more to trend following. But I think we should think about essentially providing a better solution that will make it easier for them to allocate to.
And that's important in terms of, let's say, higher volatility in terms of being able to give them better capital efficiency. I mean in fairness, I remember, you know, back in the day, and I'm sure it's still the case, AHL was offering high volume share classes to institutional investors who understood the product, and that's what they wanted. The problem is, I guess not that many people necessarily understand that high vol is are good thing.
And actually, the really horrible funds are often the ones with low vol and relatively high fees as a proportion of vol. We'veall been banging our head against the wall for a very long time on this, and we often say things like, oh, we need more education. I like Yoav saying, well actually we just have to live with it and change the product. Maybe that's the case.
Imeanit does worry me slightly because I guess one thing you could obviously do to make something kind of more tail protecty would be to speed it up and give it more positive skew. And then that would be probably be at the cost of alpha because you'll be moving out of the sort of area where trend following works really well.
and you'd make it look a bit more… Itwould become more like a tail protect product except instead of using options and implied vol, it's using realized vol and using futures as a kind of synthetic option. So that's maybe a valid thing to do. But it's very, very different from the kind of product that I personally think that this industry should be at about. Yeah, I mean, CTA is an amazing… Trend following is an amazing strategy.
It's one of the only strategies which has, you know, the positive skew and it's not actually harvesting liquidity premium. So, it should really be allocated to a lot more than it is. But we just need to make it easier for them. Yeah, you know, I'm kind of sitting here listening to the three of you, and I'm feeling I'm not quite fully in agreement, which is a little bit odd.
When I hear that maybe we should give people more vol, and that's going to solve the crisis, I'm thinking, no, no, no, no. Irememberthe times when we had more vol in our industry and that actually didn't make it institutional at all. The institutions probably were the cause of overall deleveraging of products once they started investing in around year 2000. Sothat to me suggests, at least, that managers are being told not to go crazy on vol.
And the fact, nowadays, most managers have probably two levels of vol that they offer of the same strategy. So, I mean, investors can get whatever they prefer. And in all my conversations, whenever I suggest, you know, would you prefer a higher vol strategies? They'll say, yeah, it'll be great, but we're just going to give you a smaller allocation. So, I don't think that's going to move the needle net-net for our industry.
Theother thing that I'm a little bit worried about, which is always something that I've mentioned, despite the fact that I really like the person who came up with and coined the crisis alpha, namely Katy, and that is, I don't know that we can actually claim that we are crisis alpha in the sense that we are not designing… Well, some may be, of course, but generally speaking, I don't think we, as managers, are designing our products to make money specifically when there is an equity crisis, I
mean, as far as I can tell. Now, the fact that we have delivered good returns during those crises, that's a different thing. Andthat speaks to the strategy overall and the diversification in the portfolio. But I don't think we should… Personally, I don't think we should sell ourselves as a crisis alpha strategy because I think we are much more than that.
Andto me it comes back to this delivering this uncorrelated return stream which I think is so much more valuable if people understand what uncorrelation really is. Anyways, sorry to jump into your session here. Education to me is still incredibly important. Helping people understand how we perform, the way we perform. It's not just about equities. That would be the one thing I think that's been missing.
And I think just to make it current with what's going on right now in the world and maybe the amount of fixed income we need to be issuing to pay for those changes, maybe people need to start thinking about trend following in the context of a fixed income portfolio, not just an equal portfolio. But that's just my own view here. Nowwe've already touched on this next topic that you brought along, and that's execution.
You've talked about, Yoav, specifically that you need to implement things differently. I don't know if any of you want to expand on this topic, but I see it on the list of things you wanted to bring up. Graham, do you want to..? I'll kick it off because when we were having our coffee session, no beer was involved at that point, my comment on the subject was that it wasn't typically an area I wanted to go to, execution.
It's the easiest thing for any manager to say, hey, we've got top notch execution. Andthen that sparked a nice healthy debate and that's where this came from. And it's probably good for me to hand over to Yoav at this point. Yeah, I published a blog about execution. I spoke about it in two conferences as well. It's actually an area where I think it's basically a missed trick.
So,on the one hand, if you ask most CTAs what's your cost, and you do the cost analysis, and it will come back to saying, oh, it cost us maybe 5 basis points of Sharpe per year. So, that's really not a great area where we want to spend a lot of effort, frankly. We just want to streamline execution. We want to make sure that we don't do any mess up. We don't forget to roll the contract or something like that.
So,it's actually very much that we think about it as a production line where we receive an order, we execute an order, we move on to the next order, and we bear the cost because the cost is only 5 basis points point. And we are happy to move. And in that respect execution doesn't really matter to most CTAs. Okay,now certainly as you move to the more illiquid or slightly more expensive markets, that is becoming a lot more important.
So ,if my raw Sharpe is, you know, X and now it cost me 30 or 40 basis points of Sharpe per year to execute, then it's actually important, and we start thinking about can we actually save an execution? Andthen you realize that the way that most CTAs execute basically gives no opportunity to actually improve your alpha. Not to reduce the cost, but to improve your alpha by looking at short-term execution.
AndI think the issue is that people don't realize that the execution in CTAs don't actually lose money by delaying. Execution loses money by crossing the spreads by actually taking liquidity from the market. And that's where that 5 basis points comes, the liquid CTAs and 30 basis points for the less liquid CTAs. Andif you lag your strategy, suppose you say, okay, I'm going to lag my signals by a day, you find that actually your alpha doesn't decay that much.
You will maybe lose one or two basis points. So, there is a great opportunity here to say, well, actually, we're going to give the execution desk more time to avoid crossing the spread. Crossingthe spread is a Tobin tax that you pay. And the moment you trade and you take the liquidity, that's when you pay it. And that's where your main costs come from.
So,actually, when you do the analysis, it's a lot better to say to the trading desk, okay, maybe you can take three, four days to execute, do partial execution, just be patient, wait for the market to come to you. The trick here is not that I'm having even a particularly good trading strategy in the short term, it's just avoiding crossing that spread. Takingthe liquidity from the market can make a very meaningful impact on the P&L.
And if you ask, if you want to find out which CTA actually cares about execution, you should really ask them, you know, do you pay performance fee to your execution desk? And very few, because most people, most CTAs will treat their execution desk as execution monkeys on the production line. Whereas,what you find is that if you do give the desk more time to execute, you will actually get a performance up kick. It's actually very interesting what the dynamics are.
You have to be aware that it slows down your trading strategy. But the funny thing is that most CTAs do it anyway. MostCTAs will have a signal which is a raw signal, the raw alpha, and then they will say, well, this specific market is very expensive, so we will slow down the signal. But then they will insist that the execution desk trades it within half an hour or maybe half a day. So,what you have is a very funny situation.
You will take a signal and you will sit on it for five days because you are buffering it, and then insist that the execution desk executes it within half a day. Andit makes, I think to me, a lot more sense to say, well, why don't we give the execution desk the raw signal and then say, but make sure that you trade slowly, and be patient about it, and avoid paying this Tobin tax. And we find that that actually saves a lot in terms of costs.
So,I think this industry, certainly in the liquid end of the market, doesn't pay a lot. But I think there's a missed trick here. So just for me to fully understand, Yoav, so, you think this is mostly important when it comes to sort of the more illiquid market, right? So, you don't have to wait five days to execute a German bund future. I think that you will be able to save more by waiting for the execution of a five-year bond future than you will by losing alpha.
So, what you will find is that the cost of waiting… sorry, the benefits of waiting far outweighs the loss of alpha, even for those markets. I mean also the other thing is size. So, I just quickly did a back of the envelope calculation, and my training costs are only 4 basis points of Sharpe (go me), half of that's commission, which for an institution is negligible. So, only two basis points of that is bid/ask.
So,I could only save two basis points a Sharpe by putting any effort into my execution, which you know, is probably not worth doing. I guess I do like the idea, and I agree with you, I mean, I do remember, certainly back when I started with AHL, we did make improvements, and maybe Graham can comment on how things have changed since then, if he's allowed to. But no, he's not allowed to, which is fair enough.
But,certainly the way that it used to be done a long, now this is a very long time ago, this was when I started with AHL. So, this is certainly not how things are done anymore. But basically, the trading desk would be given a window of an hour and told that the order has to be executed in an hour. And then, in an hour there would be another order coming. And yes, they had to very much get the order done by the end of that sort of period.
And I think we did move a little bit more in the direction of what Yoav says, although I'm pretty sure they were not allowing those orders to go beyond the end of the day. So,I think, certainly there's a lot more value in doing it with the weird kind of OTC bond than there is the bubble because, yeah, you might save a couple of basis points maximum. It's not really worth doing.
Theother thing is to think about what you're effectively doing then is essentially your hedge fund has become a CTA plus a trading desk with a horizon of a few hours. So, you've kind of bolted on a sort of separate alpha strategy which has got an horizon of a few hours perhaps, or a few days. That'sgreat if you have the skills and the technology to actually make money at that frequency. So, if you've got traders who actually are able to do that.
And if they're experienced traders in an OTC market, then yes, almost certainly they can. Butthat's not necessarily given that you'd be able to just magic up this extra alpha necessarily. I mean, for me personally, if I was trading a weird OTC market and I was paying 30 basis points of Sharpe in costs, I do one of two things. I'd either slow down the signal even more or I'd stop trading it.
Tome, that 30 basis points is probably way too big a bridge to sort of get over just by improving my execution or allowing my execution traders a bit more leeway in execution. Yeah, makes sense. I want to cap this topic and then move on to the next one just in the interest of time. It's somewhat related. You put down research technology and so on and so forth, love to hear your thoughts about it. But actually, it's also somewhat related to execution.
Because when I did with Alan, our big series where we interviewed all the CTAs in the SocGen CTA index and we talked about research, a lot of the big firms actually said, yeah, we probably spend 35% to 40% of all our research just on executing. So, I mean, it is interesting. Anyways,research and technology and systematic trading, who wants to kick this one off? I think out of us three, I'm actually the most technologically oriented.
I remember going to the AHL tech team meetings they would give every week. There would be a talk given by some of the techies. And I was pretty much the only one from the sort of the research team who actually joined. They hated me because my views were always like, you're doing things wrong here. Your API is different. You should do that. You should do this.
Ithinktechnology can be a really great enabler in a CTA in terms of being able to facilitate the thoughts that the researchers have and making it happen quickly and efficiently and cleanly. And I think a lot of the time the technologist would also even notice issues that you haven't noticed, because these guys are clever guys. These are not just code monkeys. They are people with PhDs and people with market experience.
AndI think technology, when it is really married to the research, when you work side by side, is really a great boost to doing what we do. Can you perhaps comment on, I think you said at the beginning that you were a great fan of open sourcing. I guess there's potential conflict to an industry which is famous for its IP and its secrecy with, at the same time, embracing open source. Absolutely.
It's certainly, you know, my observation has been it’s a great way of attracting the best technologists. Yeah. Let me give heads up to Gary Collier, who started the open source in AHL. And I told him, like, you know, you're nuts. You're taking our code and you're putting it out there. And I was wrong. And it not only improved the way that we code, but it also improved the quality of candidates that were attracted to AHL. Theability to convince people, this is what we're doing.
this is really exciting, this is really great code. It really fires up people and I think really improved the quality of the technology. Andsometimes an open source project like ArcticDB, you know, spins a life of its own. And then we have James Monroe and the team in AHL, which are running essentially almost a separate business coming out of that technology, which was originally used to speed up our access of data in research, but has grown way out of it.
Now it's being used by Bloomberg, by BitQuant. It's a great piece of technology. Therecould be some improvements, by the way, but I have these discussions with James and the team directly. But I think open source is good. In fact, we open source a lot of the basic foundation layer of what we do. Sothat is actually the way that you attract talent in a very competitive industry in technology. So, I've written sort of two versions of the training system I use myself.
The second one, I decided to open source. And it's definitely the case that the discipline of knowing that lots of other people are going to look at your code, it makes you more likely to write good quality code and test and document it properly. So, that's a plus one from me on that. Right, so the last two points I'm going to try to combine and then I have a third one I wanted to quickly ask you about, more for sort of my own interest here.
So, you know, we've obviously been very focused on the fact that you all worked together at AHL and, and kind of the evolution you've seen through those eyes. So, I'm interested in any kind of final thoughts that you have on that. If there's a memorable story, that's fine as well. Butthen I'm also very curious as to who came up with this heading, My Love for You Is Like A Magnitude 6 Earthquake. Your Sharp Ratio Makes Me Tremble Inside?
So, I should explain that this is a bit of an inside joke because we were talking about ChatGPT, and LLMs, and so on, and so forth, and I think it was Graham said, oh, Rob, all you need to do is record the conversation and then just put it into ChatGPT and it will try to put a transcript for us. And I said, oh, I'd never do anything like that.
So,what I thought I'd do is I'd put that at the end as a sort of Easter egg, sort of to make them think that maybe I had got ChatGPT to do it, because of course that's the sort of nonsense that you'd get if you used an LLM. I would say to that is, I know my wife never listens to these things, so I'm perfectly safe. But ChatGPT and LLMs are magnificent for writing Valentine's poems. You can tell it whatever you want, you know, it'll make a quant or a CTA manager even sound romantic.
I advise everybody to give it a go. And that's where that came from, Graham's comment that he used ChatGPT to write his Valentine's Day cards. So fantastic. Good stuff. But I know you're not going to believe that we weren't drinking when this happened. Honestly, we were all completely sober. Well, I mean, it could be Irish coffee you were drinking, you just said coffee. But you never know. Anyways,any final thoughts that you have before I ask the final question, which is not on your list.
Is there anything that you want to kind of leave the audience with in terms of your journeys, and so on, and so forth? I would just say it's a very small community. I think it's just fantastic that, you know, certainly the three of us, now the four of us, keep up on a regular basis. Long may it continue. Yeah, absolutely. All right, well let me throw it in then, the last question.
So,I think we've obviously been going through the whole sort of a lot of the evolution, in terms of trend following systematic strategies. We've also been curious a little bit about why is it under allocated after all these years with all the evidence we have. So,let me just ask you one final thought and that is Fidelity and BlackRock are now announcing that they are moving into our space. They're launching trend following products.
Doyou think this is the stamp of approval for trend following that we've been waiting for, for five decades? Is this what it takes, firms that size, the ability to plot it into model portfolios, et cetera? Is this a turning point? Imean,obviously pure speculation, but do you think this could be a turning point? It's kind of depressing when it's like two years after they got into crypto. So, we're less respectable than crypto. It doesn't really affect our life.
So,the way we think about CTAs, I think, it's similar to you in terms of it turns alpha and we're in the business of providing alpha and creating a very high Sharpe offering for the clients.
I think that it actually will be very difficult world for incumbents to really offer, in that capacity, with the capacity of BlackRock and Fidelity, and the AUM that they are targeting to really… And,I mean, it will be great to get allocation to the CTA industry as a whole, but I think it will be very difficult to create a properly diversified product at the AUM that they will be targeting.
So, I actually am a little bit afraid that we will be, essentially, getting more people saying, oh, CTAs are a Sharpe 0.5 strategy. Andthat to me is actually a danger. I think CTA is a wonderful strategy. It has positive skew, positive alpha we can create beautiful portfolios, and I'm kind of afraid, actually, of that happening. Very interesting. Yeah. Any thoughts on your side, Graham? Again, my perspective, if the more people have these strategies within their portfolios, the better.
Yeah, well, let's see about that. In any event, we'll know in five, ten years time. We’ll all be working for BlackRock or Fidelity. Well, we'll still be sitting here talking about the wonderful things we can about systematic trading. Anyways, gentlemen, this was fantastic. This was fun. It was fascinating. Great to hear your perspective. So,Rob, Graham, Yoav, I really appreciate you coming on and sharing these ideas and some context for our industry and I enjoyed it.
I'm pretty sure all the audience as well, listening to us today. enjoyed it. Andif you want to show your appreciation for the three gentlemen here, go and leave a top-notch rating and review on your favorite podcast platform because it certainly is a nice way to show your appreciation for the work that they put into putting this episode together. So, from Rob, Graham, Yoav and me, thanks so much for listening.
Welook forward to being back with you on the next episode of Top Traders Unplugged as we continue our journey into the CTA industry. And in the meantime, of course, you should go and check out the show notes for this episode and all the other resources you can find on our website. Andmaybe I should, because I always forget. There is a new version of the Ultimate Guide to the Best Financial Books on my website right now.
It is, now, more than 500 books that you can get your hands on in that guide. So go check that out. And in any event, take care of yourself and take care of each other. 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 iTunes and subscribe to 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 in iTunes. It only takes a minute and it's the best way to show us you love the podcast. We'll see you next time on Top Traders Unplugged.