TTU137: David Gorton, CIO at DG Partners

[00:00:06] David

We’re always concerned that when markets lack direction and there isn’t a fundamental driver, that returns are going to be lackluster in those types of environments and that people can get bored of trend following, and go back to your earlier point, well, you had one great year, a couple of good years here and there and then another banner year, maybe we’ll just wait two or three years and then they forget. I think it’s important that there’s an allocation and a dialogue, with investors now, that we don’t slip into too choppy markets for too long.

[00:00:43] Intro

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.

Before we begin today’s conversation, remember to keep two things in mind. All the discussion we will 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 product before you make investment decisions.

Here’s your host, veteran hedge fund manager, Niels Kaastrup-Larsen.

[00:01:38] Niels

Hey everyone, and welcome to another edition of Top Traders Unplugged where Alan Dunne and I are joined today by David Gorton, Founder and Chief Investment Officer of DG Partners as part of our mini-series focusing on the one investment strategy that beat everything else in 2022, namely, trend following and managed futures more broadly.

First off, David, it is great to have you on the podcast. Thanks so much for joining us today. We really have been looking forward to our conversation. I hope you’re doing well. I guess you’re in London today.

[00:02:05] David

Yeah. I’m in London. Very nice of you to invite me. Thank you.

[00:02:10] Niels

Absolutely. Now, before we dive into all the different topics that we were going to talk about, it would be nice to set the stage for our conversation a little bit and maybe give the audience a bit of knowledge, a bit of background to your firm and to yourself and also maybe share a couple of highlights in terms of the strategies you focus on and where the business stands as we have entered 2023

[00:02:33] David

Yeah, certainly. Well, DG Partners was founded in 2002. We spun out of J. P. Morgan, where I ran one of the prop divisions running prop trading for J. P. Morgan in Europe. We’d always traded various systematic strategies, but we started trend following in 2006 (May, 2006, actually), and then following quite a good 2008, we decided to launch systematic as a standalone business.

At that point, we formed a joint venture with Brevan Howard and we formed a company called BH-DG Systematic Trading, and it’s the BH-DG Systematic Trading that’s the focus of our business today. We manage around $3 billion as a firm, two and three quarters of it in trend following.

[00:03:20] Niels

Fantastic. Great. That’s a great background. I appreciate that. Now, we have different topics for today and what we normally do is we jump back and forth between Alan and I in terms of bringing up the first topic. So Alan, as we normally do, why don’t you start off with where you want to go.

[00:03:39 Alan

Great. Thanks, Niels. Morning, David. It’s interesting. A lot of people we talk to come from very much either a quant background and did a lot of research in the markets and then got into trend following. Obviously, you come from prop trading and there’s a big macro flavor running through your relationship with Brevan Howard. So how did a macro trader end up being such an advocate for trend following?

[00:04:05] David

That’s a very good question. Listen, lots of strategies and all different strategies have got their merits. I’ve always had a bent towards having researchers work with me and we’ve built lots of models, factor-based models in the past trying to describe yield curves, trying to describe currency movements, vol surfaces, and so on. But none of those were really robust. They could never really become independent standalone trading strategies. They tended to be much more as inputs to a macro or V-type trading strategies.

I guess back in the old days, we kept hearing what CTAs were doing and we thought we’d quite like to work out for ourselves perhaps rather than listen to rumors on the street. That led us to develop what I call our core trend or our classical approach to trend following, which is very much moving average and oscillator based. After some testing, we launched that as a standalone strategy in 2006. I realized that if you look back over long, long periods of time, the returns in trend following are very robust.

There’s a persistent inertia in markets, if you like, which doesn’t realize itself every year, but over periods of years seems to be rarely stable. If you can try and capture that in a consistent manner, you’ve got something that’s really quite useful in a variety of different portfolios.

[00:05:27] Alan

Obviously, you’ve been around a lot of traders being at prop desks. Do you think a lot of macro traders are trend followers themselves? Obviously, everybody’s trying to capture the same moves.

[00:05:40] David

Yeah, I remember talking to a really great investor, from a very premier shop, some years ago when we started this trend following business. He said, “I never invest in trend followers, per se.” He said, “I invest with Louis Bacon and Paul Jones.” He said, “They’re basically trend followers, but much more fun to talk to.”

[00:06:05] Alan

That’s probably true.

[00:06:08] David

I thought that was absolutely right.

[00:06:11] Alan

You touched on that inertia in markets and how it’s there over time but not all the time. What’s your thought about why that is the case? When we ask people why trend following works, you hear about different speeds of reactions, etc. But then you’ll have two or three-year periods where trend following doesn’t seem to be working and those biases are still there. So when you think about it, how do you think about it?

[00:06:37] David

Well, markets ultimately go from one unstable equilibria to another unstable equilibria. They don’t move in an immediate step from one stable position to the next stable position. So it really depends on market news, macroeconomic factors that are driving the individual commodity or asset class. Sometimes, there’s really fundamental things happening like last year. Markets were free to move and adjust. You can contrast that with what we call the dead decade.

I heard other people in the industry or you guys call it the Dark Age, which might be nicer, but that horrible period of financial repression sucking volatility out of markets, a deliberate policy tool to reduce risk premia and encourage financial flows. It was an example of a period, a prolonged period in fact, where trend following, I wouldn’t say it didn’t work. There were still positive returns, but they weren’t as spectacular as they’d been in the prior decade, and hopefully, they won’t be as good as the coming decade.

[00:07:48] Alan

You mentioned how you’ve done work with different researchers over the years, factor-based models and stuff like that, and it sounds like you’ve got very much a focus on pure trend. Is that a philosophical perspective?

[00:08:03] David

Yeah. I think, going back to that dark age period, it became clear from 2010 that the authorities’ desire was to suck vol out of markets. That’s very bad for us because two things happen. Markets are not really involved and you tend to get high amplitude sideways markets (which are a disaster typically for trend followers), or you get low amplitude, sharpe reversal type markets, which again, are not really very good. I think in that early period, it became obvious that we should be trying to do something, perhaps, to adjust our models because the regime was quite different.

One thing we looked at was speeds, because clearly, if markets are going to be less volatile, your expected return on any given trade is lower, therefore, your transaction costs are higher. You should trade more slowly. Also, trends are harder to discern over longer periods. For short periods, if you look at the JGB, you don’t really see it’s gone anywhere. It’s only when you look back over longer periods, you can discern the trend. So we thought about that and how far one should go in slowing when you talked about speeds.

The problem is, if you slow too much, you’re going to lose all your skew, because when markets turn, you’re going to be stuck for too long in the wrong trade. So we, philosophically, don’t want to try and maximize profits in a month or a quarter or even a year really. We want to be in this business. We’ve been in it now for 17 years. We want to be in it for the next 17 and produce stable, robust profits going forward.

[00:09:45] Alan

Maybe one final one on this section before we move on. Is it harder as somebody who has been a prop trader, macro trader and you have views on the market and the world? I suppose the trend follower of philosophy is really not to have a view on any market but have a view that this is how markets behave, and that’s kind of the edge. Whereas, obviously, you must always have a view on, “Do I like this position or not?” Does that make it more difficult?

[00:10:14] David

Yeah. It’s purely systematic. We don’t trade. Obviously, we can bring, or I can bring, some knowledge to the table when there are market blow ups and we’ve got to decide on risk limits and what products we’re trading, but also philosophically because in that dead period, there’s a lot of pressure on trend followers to do other stuff. Every investor that came in said, “Yeah, you’re a trend follower, but what else are you doing?” We did try and look at alternative sources of alpha.

The problem that we had or that I had with them was that they tended to have two features. One was that they were factor-based. The problem with factor-based models is they work very well until they don’t. You can predict the dollar-yen, with 10-year yield spreads, relative inflation rates, and the current account deficit. That might work very well for six months or a year or longer and then explain 70% or 80% of the market’s moves, and then suddenly, it doesn’t. Therefore, it’s not robust, so you can’t just leave it to go. The second set of models tend to be mean reversion or relative value type models.

The problem with them is that when trends really start to happen, you’ve got the wrong position and you lose all your skewness. So not only are we trying to source positive alpha, we also want to have positive skew. In 2015 or 2014, I made a really slightly controversial decision where I said we’re abandoning these new sources of alpha and the pursuit of them per se. We want to be purely systematic, purely trend following, and we’ve just got to be a better trend follower.

So we’ve got to look at what are our weaknesses with our classical trend following model or maybe just think blue sky about how we should trade trends. That’s, in fact, what we did. That went down like a dead duck, honestly, with all our investors or potential investors. They thought I was bonkers.

[00:12:14] Niels

I want to jump in on that. This is fascinating, David, because it’s actually some of the narrative I’ve used myself in terms of the weakness of trend following. So I’d love to hear your thoughts about what were the weaknesses that you identified and how did you try to improve those without giving any secret sauce away of course. But how did you think about it and how did you think you could possibly become a better trend follower without leaving the space so to speak?

[00:12:43] David

Yeah. If you think about the key features of trend following, it’s in all markets all the time. Classical trend picks up on trends very early, and when trends build and develop quickly, that’s a fantastic position to be in. But in building what we call a response function, the way that you’re going to react to the market’s price movements and volatility on a daily basis, for any individual market or sector, you data mine. That’s where you start typically modeling.

So you might say all bond markets and look at them and use all that data and conflate it in all kinds of different ways and try and sort out the signal from the noise, and then try and use that signal in production to be an element of your portfolio. The fact is that if you look at US equities or US bond markets or fixed income markets generally, they’ve basically gone up for the last 40 years. So any form of data mining will naturally give you a bias to want to be long in preference for being short. Bull markets traditionally go up the elevator and bear markets come down the escalator. So, by the time you’ve gone short, you’d be too late.

So there’s a skew, if you like, in the way that you would position. I think I said earlier that one of the features of the markets was that they were becoming more sideways for prolonged periods of time. Therefore, being in all markets all the time felt like a weakness. So we said let’s think about not new sources of alpha, but being a better trader, a better trend follower. Could we develop a conceptual approach to trends, i.e. one that wasn’t focused on data mining at its foundation?

In doing so, could we produce something that then only sought to be in the meat of the trend, recognizing that we would miss out on the profits in the early stages? So it might not be as efficient, if you like, but would allow us to stay out of markets that were just chopping sideways. So that was where we started. We’ve now run two broad approaches to trend following. One that’s conceptual and one that’s the original classical trend, if you like, or core trend.

[00:15:06] Niels

Okay. Almost sounding like, again, without giving any secret sauce away, that the newer approach, let’s call it that, does not necessarily need to wait for the price to go against you before it starts getting out. It sounds like you are okay, as you say, with capturing the meat of the trend, but you don’t want to give back too much.

[00:15:27] David

Yeah, that’s right. It does give back obviously. The edge of this stuff is it is not persistent. But certainly, when markets become overly exuberant, it’s out of 100%. In classical trend following, as you know, trends get quite aggressive if they move to very strong momentum signals and move away from their current trend along the way. We’re naturally taking profits.

[00:15:56] Niels

I want to go back to another thing you mentioned, David, and that is you talked about this funny story about your investor saying that it’s much more fun to talk to Louis Bacon and so on and so forth. Of course, he’s right. Narrative does play a role with investors, I think. One person that we like to quote a lot because he does produce a lot of narrative is Cliff Asness.

He talks about last year, and we bring this up with everyone because we do think it’s interesting to hear our different perspectives. But he wrote many papers last year, one of them was this idea or talked about this idea that maybe trend followers had become too concerned about the Sharpe ratio, and therefore, we’re putting other stuff in the models that maybe we shouldn’t. Now, you’ve made it very clear that you’ve kind of doubled down on trend or pure trend, so that kind of answers that question. But maybe then it’s a broader question for you and that is, do you think we as an industry have become too focused on Sharpe?

The other question I wanted to ask you is, how do we change the narrative so that people don’t look at Sharpe on an individual line item but actually look at what we, as trend followers, can do for their portfolio Sharpe, which really is what the Sharpe was invented for and what people should be concerned about.

[00:17:33] David

Yeah. Sharpe is one measure of performance, isn’t it? But from a personal perspective, I wouldn’t necessarily focus on Sharpe. I don’t want to put my money away and make 2% with a Sharpe of 5. I’d much rather make 10% with a Sharpe of 1 and 20% with a Sharpe of 0.5. So it’s quite a restrictive kind of measure. I think investors, I think they do increasingly understand that traditional assets, fixed income, property, primarily equities, demonstrate negative skew.

They do generally well, but their worst months are much, much worse (and their worst years) than their best ones, typically. Classical trend following tends to have the opposite profile. Although if you look at the so-called SocGen Trend Index, it doesn’t demonstrate much of a positive skew any more. It used to. That probably speaks to some of the bigger players perhaps employing other types of strategies. I don’t know what they do and I don’t know them very well. I’ve certainly never met Cliff Asness nor indeed read any of his papers, but I’ve obviously heard of him.

So if he’s saying that trend followers have positive alpha and that positive alpha tends to be concentrated in periods where traditional assets are doing badly, then I’d agree absolutely with him.

[00:19:11] Niels

The other thing he talked about (and I think not just Cliff Asness but many people talk about) is this, “What is the mandate?” What are trend followers? What is the role we have to perform in a portfolio? He talks about a dual mandate. One is we have to produce absolute returns and then also we have to provide “crisis alpha”. Now, I’ve been on record, for many years, saying I was very excited about the term crisis alpha when it came out because it was so simple for people to remember.

So when they went into an IC committee saying we’re going to invest in some crisis alpha strategies, people kind of almost understood what it was. Later on, I found that actually it was a little bit counterproductive because then I had to argue every time there was a three-week wobble in the equity markets whether that was a crisis or not.

[00:20:04] David

I would agree with your investors. I think it’s a nice catchy term, but it’s probably not quite right. Why? Because it sounds to me like crisis alpha is almost like saying it’s a tail risk hedge. It sort of takes you on that continuum. I understand a tail risk fund or hedge is like an insurance premium that you pay out each year or each quarter. When the stock market has a really bad quarter or the bond market has a really bad run, it’s going pay off. So you expect a pay away.

The fact is that, as a trend follower, in the first quarter… Last year was a classic example of this. ’21 was a huge bull market in equities. Q1 ’22 was a dump and we all lost money(I presume) in equities, or we certainly did. So the turning point of a trend… If you’re trend following, you’re almost certainly going to be losing. The idea of crisis alpha, that the crisis just begins and tanks, is a difficult one to argue. I try to stay away from that tail risk hedge crisis alpha and I prefer to describe it as a risk mitigator.

So while we’re trying to produce positive alpha that’s got stable, non-correlated returns to equities primarily and fixed income secondly, we would hope that in sustained periods of risk-off, that we could produce positive returns. One of the questions in the old days you used to get, or criticisms of trend following was, “Let’s look at your P&L breakdown. You guys just made money being long fixed income.” I would say that’s true, but fixed income is rallying. I could be at home washing the dog or I could be long on fixed income.

So we want to be long on fixed income. The question then is, can you make money if fixed income ever goes down? The answer is we’ve never really had a great experience. We’ve got no statistical evidence. I can’t prove it to you that we could, because we weren’t around in ‘94, which was the last real, horrible market. Last year, of course, we did.

I think what traditional investors found was that bonds are a very nice hedge for equities when equities are going down. However, equities are not a very nice hedge for bonds when bonds are going down. That was the lesson of last year. I think trend following was able, in the broader sector, to produce pretty stellar returns in a bear market in fixed income.

[00:22:52] Niels

Now, before I turn it back to Alan, it kind of leads me to two follow-up questions of things that I’m hearing right now when speaking to investors and potential investors after a year like 2022. I get the distinct feeling that they continue, for whatever reason, to look at trend following as a trade, meaning they think about now, “Oh, did I miss it?” So that’s kind of the first thing that comes when they start talking.

The other thing that I hear is that if trend following did so well in 2022, surely it must be going to give back a lot of those profits in 2023. So I’d love to hear just how you would have that conversation with people and try to perhaps argue it slightly differently to what they are saying.

[00:23:44] David

Well, it’s a very, very sensible question. If you look at the profile, you’d think that’s right. In my mind, it’s very difficult to time trend following. You really want to be investing in a trend follower that hasn’t really lost money very often and produces steady alpha and then still does really well when you need them. I can remember in that dead decade someone coming to me and said, “We’ve looked at our returns. Every time we got up 10%, we took profits and switched the model back on when we were back to plus one. Our return profile is this beautiful straight line.”

I just said it’s a good job. I’m the Chief Investment Officer and you’re not. Because that is a catastrophic way to approach things. That just speaks to, having done so, the following year of course, returns were much higher than 10 and you’d never have got back in and you’d be missing 100% return by now (whatever it is). I’m just trying to speak to the point that it’s difficult to time it. As to the question of ’22 was great, and therefore, ’23 will be bad, although ’09 wasn’t necessarily terrible but it did pre-stage that dead decade.

I think there were some quite significant differences between 2008 and ’22. 2008 was a financial crisis. The whole financial system was blowing up. We had two huge high-profile bankruptcies, trading had seized up, people were talking about not being able to get money out of your ATMs, the interbank lending system had stopped. Afterwards, we all saw the LIBOR arguments and everything else. Interestingly, we had this huge move in fixed income. Who’d have thought we’d have a 400 basis point up shift in short rates pretty well? Nothing broke.

We could talk about, perhaps, that later. It’s not really a systematic subject. But nonetheless, nothing broke and we got inflation. We hadn’t really seen any upsurge of inflation for many years. Given that we got an inflationary environment, rates had to go up and the so-called Fed put in the market, where stocks have a wobble and suddenly there’s a financial crisis somewhere, and the central banks have to intervene and bail the market out and bail the market participants out. The only place we saw that actually was in the UK, in October, September-October when the UK pension system was under severe pressure.

So I think there’s a stark contrast really between ’22 and ’08. Therefore, when you look at the following years, (and I’m not predicting that ’23 will be a great year for trend following because that’s not my job), but I do think over a multi-year period, it will be quite good because that Fed put doesn’t exist and won’t exist again until there’s a real Armageddon, I don’t think, or indeed, until inflation gets down and they can be a bit more active. Markets are going to be much freer to find their own equilibrium.

The first thing we talked about really was markets going from one dynamic equilibria to another. So if markets are freer to move and we’ve got new factors, new macro factors like inflation in there, that’s a pretty big deal. It probably means that markets will take some years to move around. So I’m hopeful, yeah. Quietly optimistic, not necessarily for this quarter but for the immediate period ahead.

[00:27:21] Alan

Well, maybe just to pick up on a few of these points and move into research as well, but you talked about this conceptual trend and capturing the meat of the trend. There is an alternative classic view about that you get these over extensions at the end of a trend, and often, it’s drawn back to the example of when you’re up 10% and the investor wants to take profits, but actually, at that point, you’ll tend to have the risk on and it’s either going to extend and be a really good trade or not.

By focusing on that conceptual trend and the meat of the trend, in that component, do you give up a bit of skew but you still get it in the core trend? How do you think about that balance of capturing the meat but not giving up the skew?

[00:28:08] David

Yeah. It depends, really, what you define as exuberance. Clearly in classical trend, as I said earlier, I don’t know what everybody else says but in our form of classical trend following, there’s a point of the trend momentum where we don’t want to have such a big position for a given vol level. Our position wants to be coming down. The trend could never get so strong, I don’t think, that we would come down to zero. But there’s certainly profit taking in there.

But when you’re focusing then on the meat of the trend, you know you’re missing the beginning, which means you’ve got to be more aggressive about your entry point when you go in. You can’t scale in and feel good the way you will with classical trend following. Classical trend following statistically finds a way to mimic human nature. When you and I talk about a stock today after this meeting, it’s got to be a great business. You tell me it’s £1 and it’s going to £5. Maybe on Monday, I’ll look at my Bloomberg and look it up and I see it’s trading at £3. I’ll go, “That was a genius, but I’ve missed it.”

Even though you told me it’s going to £5, I’m not going to buy it. But if it’s trading £1.30, £1.30, I’ll call you back and say are you so sure about this. Maybe I’ll buy a little bit. A couple of weeks later, if it’s trading at £2, I’ll probably buy more because I believe you. That’s kind of where classical trend mimics the market in a way. This is the conceptual model if you like. The conceptual really is about not using data in it’s entry point because it’s not in markets all the time. It just wants to be in them in the thick, misses that beginning of buying at £1.10, £1.20 and building up.

Therefore, it’s got to go almost all in at £2 or £3 hoping for the £5. Once the market then extends, and we have various measures to see how far we think it’s extended away from some form of norm, yes, you may miss a complete explosion in getting out completely. So you would give up. We’ve been running the system now for since 2015 and its skew is pretty much the same as classical trend.

[00:30:28] Alan

You mentioned the period where you thought you would try to be a better trend follower, and obviously did a lot of research around what are the possible avenues around that and you came up with conceptual trend. What about some of the other things? We talked about non- trend, and you said for various reasons, factor models were not a good fit. Some people talk about speed. You said you maybe trade slower. Obviously (and from the perspective of the crisis alpha risk mitigation), there might be an argument for trading faster at times. Is that something you think about? What’s the research shown on that side?

[00:31:10] David

When we change anything in a model, we do it extremely cautiously, in a very small way, typically. We put what we call conviction trend, the conceptual trend. It only started in a very small way and it took years to build up to be pretty well half of the portfolio. But equally, even relatively modest things like changing speeds needs to be done quite cautiously. So we kind of understood or we felt that slower trends would be better, but we didn’t want to slow out all the way down, because we couldn’t predict when the regime was going to change.

So there was some minor slowing, which I think was sensible. Then, as the regimes began to change once again (now a couple of years ago), we started to see the change with the COVID crisis, a very rapid regime shift. If you think in March 2020, we had a bear market in the first quarter and we were at new all-time highs by the end of the second quarter. Two complete regimes within six months. It was quite extraordinary. So yeah, that would bias you to be a little bit quicker, but our moves are very modest.

What we did think about, actually, was that with a classical approach of trend following, if you had 100 markets and they were all equally liquid, you would start by giving them all 1% each of your capital, and your positions would just build up from the strength and quality of the momentum signal in each of those markets. So you’d be trading big positions in markets that have strong momentum and small positions in markets that have weak momentum. That’s kind of a fixed approach to your capital allocation.

When you run a conceptual model, and you’re not in all markets all the time, that doesn’t work because you might only be in 60% of the market so they might well be fixed income and the dollar. So you can’t rely on some long run correlation coefficient to size your positions because the portfolio is diversified, you’ve got some benefit there you need to be taking care of and understanding your risk. So you have to have a dynamic approach to risk management.

We started to think, as well, a little bit about whether we can be more dynamic in terms of that fixed allocation in classical trend. We’re not overly dynamic, but each quarter, we’re informed by a signal to noise ratio system essentially, that we call trend efficiency. It just guides. So, if we’re 25% in equities, 25% foreign exchange, commodities and interest rates. We might be guided to have 20% in equities and 30% in fixed income, but we wouldn’t move much beyond those parameters.

So it’s a relative modulator that tries to address another weakness. It’s a weakness and a strength. The fixed allocation is both, but it’s a feature and maybe we could just improve a little bit on that.

[00:34:26] Alan

So is that allocating more risk to the markets that have been trending recently? Is that it in simple terms?

[00:34:32] David

Well, it’s not quite that they’ve been trending, but that they exhibit a higher signal-to-noise ratio. A trend might go from 100 to 400, but it might be going in a very noisy way. We’d rather be in a trend that was going from 100 to 150 but doing it one day at a time.

[00:34:53] Alan

So those markets are showing more favorable trending characteristics.

[00:34:58] David

They’re just less noisy.

[00:35:01] Alan

I suppose what I’m getting to is that by making that…

[00:35:04] David

It’s a relative measure. So it doesn’t really help you unless you’ve got the core alpha producer underneath it.

[00:35:10] Alan

Is there inherently an assumption that does matter, that persists for a bit of time that you’re going to get a slightly better return by upgrading it gradually. But presumably, it’s a fairly weak effect if you’re only making a small tilt. Is that fair to say?

[00:35:29] David

Yeah. Well, I think if you did the full tilt it recommended, or you conceptually came out with, you could have quite profound effects. I think I said at the beginning, we want to have stable, robust returns. If we were banging our risk allocations from 80% equities to 20% equities, quarter by quarter, there’d be a lot of binary market risk in there. That would completely undermine our longer-term goal.

[00:36:00] Alan

Okay. That’s interesting. You talked about generating alpha absolute return, but having this stability of performance is a secondary objective and the risk mitigation, as you say, it’s a secondary objective there in the background as well. Is that how you think about it?

[00:36:18] David

Yeah. I think that’s right. I think that’s right. I said we want to be in this business for the long run, so we’ve got to have a really stable, robust approach. It’s a truism of sure, you could ask me something about research, but I might preempt it and say that no researcher has ever brought me a model enhancement or new model that has not backtested well. My job and our risk committee’s job is to see the wood for the trees a little bit and apply a very healthy dose of skepticism to these things.

[00:36:57] Alan

Outside of that significant change, the development of conceptual trend, what are the other smaller incremental changes that the research has thrown up? Do you do an annual review of the speeds you trade at or is it the execution? What would you say is the focus?

[00:37:19] David

We do an annual review. If there are 10 speeds, we’re only going to move one. So at the margin, it may help. I’ll say quarterly, we look at our classical trend following and think about our four asset classes and where our risk should be. We want to add new markets as assets… One of the great benefits of a trend follower is there’s a machine so it can be in all markets.

The commodity trader doesn’t get tired, or is going through a divorce and needs some time off. It’s in the markets all the time and that’s really important. So you get this diversification benefit which is really, I believe, key to our return. As assets grow, we need to add meaningful markets that give us more liquidity, and hopefully, don’t compromise our diversification.

[00:38:15] Niels

Can I ask you a question about that, David? That is, first of all, I’m curious how many markets you trade today. But also, where do you think you can go to, in terms of markets and they’d still be meaningful, as you say, meaningful additions?

[00:38:30] David

Yeah, that’s a great question. It’s not one I can give you a very full answer to. I can tell you we trade 130-something markets right now, up from 110 or so a couple years ago. So, add them slowly and cautiously as assets have grown. Then there’s this issue of alternative markets. I suppose that’s slightly definitional. It depends what you call an alternative market. I think we’re offering daily liquidity, so we don’t want to put anything in that’s too esoteric or we might get stuck in for a period of time.

So everything’s got to meet quite strict liquidity requirements. Not that anybody really uses data liquidity. We generally have redemption subscriptions at month ends, like we used to. Nonetheless, it’s a feature and it’s one that we have to be very cognizant of because in a real crisis, I think that’s kind of important. So we want to add new markets. Some may call them alternative, some may not. We run them through and we test them to see any diversification benefits.

Clearly, you’re unlikely to find a new German bund contract tomorrow that you’re not trading, something without profound liquidity, or a new dollar-yen. So everything, by definition, is going to be a lower rank liquidity bucket. So we don’t want to go down to trading potatoes in Norwich, but we need to find things with sufficient liquidity, but recognize that they’re probably less liquid, and therefore, we could have a lot less in them than we can in the main parts of the book.

[00:40:18] Niels

Now, we’ve talked about some quite interesting things I find, because you talked about your journey to make trend following better that has taken you into thinking about entries differently. I think that’s interesting. We don’t come across that so often. You talked about this new conceptual or conviction trend, where you definitely also are dealing with your exits differently. I think that’s very interesting. You’ve also talked about that you needed to rethink correlations, because here you have a part of your portfolio that is not always in the market.

So there’s one thing that I think that I would love to ask you because I think that’s maybe the fourth thing that we all look at when we design these systems, and that’s volatility. Have you thought differently about volatility? Has the dark 10-year period taught you anything about maybe we need to look at volatility differently? I’d love to hear your thoughts on that.

[00:41:21] David

Actually, we didn’t look at volatility that much in the dark period. We did look at it at the beginning of the dark period, I suppose, but not because we knew we were in a dark period. We did have some quite serious brains try and rebuild our vol estimators. They spent a year or so doing it, and in the end, they beat our original vol estimator by a point zero something. The added complexity was not worth the change.

I think in general, when you think about systematic strategies or any kind of systematic trading across the whole spectrum, you can either be at one end of the spectrum where you innovate all the time. I guess, your pitch is, “We’ve got the cleverest guys. We’re constantly doing stuff. Just trust us.” Or you say that, “We’ve got a really stable approach to this, and while the past is no guide to the future, this is what we did then, and this is still what we’re kind of doing now.”

If you are the latter of those, which we are, then we know that we need to research and innovate because in order to stay the same, you must change. But equally, we’ve got to do that really cautiously. Vol is obviously a massive component of our portfolio construction. So the answer is yes, we did have a whole big multi-year project in it, but that was some time ago, and maybe it’s something we should look at again.

[00:43:03] Niels

Okay. Another thing that we’ve heard from people that we talk to that they spend a lot of time on (actually some people say that’s the biggest part of their research efforts now) is execution. I think a lot of people will be aware, of course, that the more AUM you have, probably the more efforts you need to spend on your execution. But do you think there’s also a limit to how much you can do in that area before it really doesn’t matter anymore so to speak?

[00:43:33] David

We work quite hard on our execution too. It’s our fiduciary duty and our firm desires to be as much alpha as we can. We want to pay as little execution costs as we can. Yeah, there are generally declining returns to all investments, so I’d agree. We’ve got quite a long way down the refinement of our execution process.

[00:44:01] Niels

One final thing before I hand it back to you, Alan. This is another topic that came up during last year, not that it was new. It’s been in there for quite a while. It’s this CTA replication or trend replication. We obviously knew about this quite a long time ago. Some people tried, failed, and maybe they were building too simple models. Maybe they didn’t know enough about real trend following frankly, but just wanted to come out there because it was popular after the GFC and they could use the argument of saying, well, it’s so simple that we can do it for a much cheaper price.

Now, I have no idea what fees you charge, by the way, David, so maybe I’m setting myself up here for failure, but anyways.

[00:44:46] David

We’re very reasonable, Niels.

[00:44:49] Niels

Okay. I guess my question is the new breed of replication strategies are focused predominantly on regression, meaning they don’t actually have their own model. They’re just trying to estimate what exposure you and the other members of the SocGen CTA Index have in their portfolio. I’m not a quant, but I have this inner feeling that there are some risks in doing that. They say, “Well, if you pick one manager, there’s a manager risk.” Yeah. Of course, there is. But I’m thinking there must be a replicator risk somewhere so maybe you can help me identify where these risks might lie.

[00:45:28] David

That’s a really good question. I can remember when we first started DG Partners and we were trading macro LV, LV macro, really. Some very sophisticated investor came into it a couple of years later. He had quite a good track record. The two PhDs came in and they said, “We’ve managed to replicate your trading. We can break it down to these three factors.” I came out of the meeting distraught. I thought, “That’s it! I’m out of business.” Anyway, luckily, their model, like all my factor models, didn’t work when regimes changed.

So I think there’s obviously some danger in someone replicating an index without really understanding what’s driving it. Here, we’ve got all these teams of people seeking to move forward in a nice gradual way. I don’t know how other people do it, but certainly, we do. I think if you can find someone that can consistently beat the index and you like the management and you like the stability of the firm and you are looking at that at the net of fee basis, it strikes me as being insane to invest in the replicator.

[00:46:48] Niels

Yeah, Alan. Where do we go next?

[00:46:52] David

Sorry, it might be a bit strong.

[00:46:54] Niels

No, no. Well, the thing is, I agree with you, David. But I have also said that there is one thing about replication that I kind of like, and it’s because they can package it in ETFs, which we would probably never do really because then we have to disclose our positions every day. So most likely our friends and peers wouldn’t do it.

Having said that, if they don’t blow up and they can get something that’s close and it allows many more small investors to get exposure to trend in their portfolio, all power to them. That’s what I like about it. But I am worried that it’s not going to pan out too well at some point and then it might hurt the whole industry.

[00:47:39] David

That’s a great point. Yes, it’s a very good point. I agree.

[00:47:42] Alan

Maybe just moving on to talk a little bit about risk management and a few topics around that. Again, coming back to the kind of discretionary imports. Obviously, this is something that managers have different thoughts on. Obviously, there will be some instances where there may be a big event coming up. From your perspective, what are the kind of areas where it’s reasonable, as a trend follower, to make some kind of discretionary adjustment if any?

[00:48:17] David

Yeah, that’s a very good question. Our systematic approach, the risk management is endogenous. It’s within the models. It’s not run by stop losses or anything like that. There’s a consistent and rigorous approach to going in and out of risk with various price data. I think we talked earlier about building a portfolio up. In building a portfolio up, you need to think about the implicit correlations between the positions that you have. If you want to be stable, you probably want to use a reasonably medium-term correlation measure.

I think for us, it was 2010 where the light bulb went off. It was the lead up to QE 2 in that summer. Everyone was speculating that Ben Bernanke at the time was going to launch this big program. The dollar started going down aggressively, equities rallied, bonds rallied and commodities were rallying. So our measure of value at risk (to pick one measure of risk management) was high, and we were making a fortune, by the way. Trends followers in general were having a good run.

But actually, we weren’t measuring the risk in the way that the market was positioning itself, because any correlation factor that we had was probably wrong, because the markets actually were demonstrating a correlation factor of 1, or very close to 1. So that made us think that we should have more rigorous exogenous limits on the model. The key one really is to think about a stress test for us that says, what kind of horrific loss in a short period could we survive, as a business, and would investors find tolerable given our return profile?

So we pick a number and we say, okay, if each sector does x, y, and z and that x, y, and z are worse than they’ve ever done in a given day, and all four sectors go against us. Not only that, but all the products within those sectors go against us by that amount (because some may be long, some may be short), we can’t lose more than y. That is a nice, easy conceptual way to put an exogenous limit on that stops your portfolio not recognizing short term, big correlation changes, for instance.

Then when you get to specific events, we have quite a detailed meeting pretty well every week with (what we call) the risk committee, where we look through all the dynamics of the market, all our positions, lots of different risk metrics, lots of different ways of slicing and dicing the portfolio. We also look at all the fundamental risks. We look at those every day.

If, for example, we were taking the view that it was likely that Russia, this time of the month, January ’22, that Russia was likely to be invading Ukraine, even though we were short the ruble, which probably would have been the right trade in the short term, it wasn’t really a market for us. So we’ve got to make those decisions. That’s kind of a discretion in the model, I suppose.

[00:51:39] Alan

You mentioned that if everything reversed at the same time and that kind of exogenous limit (which I guess, to limit your loss to that level, you’ve got to be able to get out), so there’s kind of an inherent assumption around liquidity. What’s your thinking around liquidity? You do see in the media sometimes people saying that liquidity isn’t as good. We had crazy moves.

You mentioned the UK in straight markets and in cable, around a time and going back to 2020 treasuries, bid asks widened that substantially. You’ve been in markets multiple decades. What’s your read on the liquidity in markets now versus the past?

[00:52:27] David

Liquidity is always a chief concern. It should be for any risk manager, not just systematic but for us all. Clearly, systematic strategies are quite an important part of markets. Some people say up to 10% of assets or whatever. Post-financial crisis, there has been a sort of disintermediation of the banking system. Lots of risk is placed in end investors’, if you like, hands, and there may be a false sense of liquidity in their holdings. We have to be cognizant of that, I think, thinking about the bid offers.

If you stress test markets, to your point, you can’t just stress test the move. You’ve got to think about what’s your likely bid offer in there as well. So, you could do that, you could just add some to the stress test. So, I would say liquidity is fine right now. We haven’t found any difficulties at all really last year. We closed some pretty scary moments in UK fixed income, in particular, in that period. But as I say, when things start to move profoundly like that, we are naturally reducing risk in an absolute sense.

[00:53:41] Alan

You talked about the 2010 period and underestimating the correlation or that kind of idea. What about the 2018 type scenario where equities rallied very substantially, say, 2017 and it was a low-vol environment so people were naturally getting levered up in the trade and then obviously we had the sharp reversal in 2018. So, was that something that the model just sailed through mechanically or did you have to make any adjustments for an unusually low-vol environment? I suppose, in relation to your interest rate markets going back over the last decade or so up until the last year, how did you think about all of that?

[00:54:26] David

You’re testing my history now. 2018 was Volmageddon, right?

[00:54:31] Alan

Volmageddon. Exactly. Yeah.

[00:54:32] David

Volmageddon. That was really February, March, whatever the thing was. Yeah. We were very concerned about the persistent decline in volatility in equity markets. I think the VIX got down to almost single digits or did get down to single digits. Yeah, it did make us think quite hard, and yes, we did have to think very hard about that as a risk committee, yes, and we managed to get through it.

[00:54:59] Alan

Okay. Niels?

[00:55:01] Niels

I was just going to jump in. I want to get back to you, Alan, but I just want to jump in on this point about liquidity because I think with you, David, we can throw a few things at you given your background that may not be strictly how we as trend followers and systematic managers think about it. But we talked about liquidity, we talk about it most of the time, 99% of the time, I think, we talk about it as; can we execute our trades, the bid offer spread, the volumes, etc.?

One of our previous guests on the podcast in a different series, Michael Howell from CrossBorder, he talks about liquidity in a slightly different way as well. He talks about market liquidity and then he talks about funding liquidity. He talks about funding liquidity as being; is there enough liquidity in the system to basically fund what corporates are doing, and so on and so forth. He’s talking about this currently, whereby it used to be, can you fund new projects? Can you do CapEx?

But now, what it’s more about is actually is can you refinance? When you get to your refinancing, you refinance your debt, and there’s a lot of it so to speak. I’m kind of thinking, well, there’s clearly a link between the two. So I’m just curious whether this is also a kind of risk, even though it may not impact right now, today, how you’re executing your next bund trade.

But is this something that, given what we saw, I guess, in the gilts markets (and God forbid, it happened in the US Treasury markets), then that probably would have been a completely different scenario. But is this also kind of a risk that you as a trend follower, a systematic trend follower, have to conceptualize a bit somehow?

[00:56:53] David

Yeah. I think we do. I don’t think there’s any immediate action. If I put my systematic hat down and just talk to you in the round about that because I think it’s an interesting point that the chap brings up. One way we’ve thought about it, at least, is to say, listen, over the last couple of years since the COVID crisis, let’s say the Fed’s balance sheet blew up by $5 trillion and basically brought mortgages and primarily US Treasuries. So typically, the buying of those securities, had they been issued to the free market, would have been held on bank balance sheets and individual balance sheets.

So when rates rise, there’s a huge amount of pain in the financial sector. Even if they held to maturity books in banks, there’s this capital hit at the front end, and there’s a funding hit to their P&L, which means there would have been more panic because they would have been paying in derivatives if they couldn’t sell the securities, or liquidating the risk somehow. That would have been a big stress on market liquidity. I think we’d have seen more profound moves.

But the fact was that in that incredible period of virtually zero interest rates across the curve, all that fixed rate lending was essentially being done by the authorities, who of course, are not accruing a loss, and therefore we’re not seeing the stress from them in the marketplace. So you’ve got the fixed rate borrowers are fine and then the fixed rate lenders are not taking the pain that they would ordinarily have done with a 400-basis point or 300-basis point rate move.

I’m no expert in private equity and so on, but I think there’s been obviously a huge amount of capital inflows into that sector. When they do these buyouts, I assume that they do staged funding, so they have some fixed for 10 years, some for 5 years, some for 1 year, some for 2 years. So the pain of the higher rates is only feeding in slowly. Because the first tranche is maybe they are at 10% for a year, or whatever, and then now they’re having to pay 10% of their funding costs. Its’ gone up 400 points or 500 points or whatever the spread is widening.

As time goes by, that’s going to come into the market and create a new stress. So, it’s obvious for us to be cognizant of the big picture, but it explains a bit why, actually, in that huge bear market, outside of that one period with gilts, everything was perfectly tradable, reasonably tradable. There were very few air pockets.

[00:59:39] Alan

Yeah. Well, this is coming back to the point as to what you talked about at the start, how rates have gone up 4% or 5% and nothing has broken. Obviously, the banking sector hasn’t taken on that risk in the same way. But it’s kind of curious, because obviously, we had QE. What did QE do? It kind of pushed investors into taking on more duration, more risk. So you would have thought you would have seen pain for investors, but we haven’t really seen that yet either. Are you surprised by that?

[01:00:18] David

Well, that’s the only way I could really explain it. I think we did see pain in some end holders like pension plans and big institutions generally, and we’re seeing pain now, if you read about it, in real estate funds and this sort of thing with liquidity, to your point about liquidity problems there. But that’s the only way I can rationalize it at the moment.

[01:00:44] Alan

Just one topic before it comes to time. I wanted to get your perspective. It was around just the role of trend following in a larger portfolio, in a broader portfolio. If you were moving away from DG Partners and you got a CIO role at an endowment, what would you be thinking is the right asset allocation and the right allocation to trend within that?

[01:01:15] David

You guys might be better at this. I think if you run lots of modeling, you’d end up with quite a big allocation to trend following. I would suggest that that’s probably not as far as I would go as an endowment. But you would think you’ve got to have a fairly healthy allocation, somewhere around you know the 10% mark.

[01:01:32] Alan

Okay.

[01:01:33] David

5% to 10%.

[01:01:34] Alan

Is that just the behavioral? I think if you run the numbers, it might be somewhere closer to 30% or 40%. Maybe even higher.

[01:01:43] David

That’s where we got it. But I think that would be crazy. I as a CIO of an endowment couldn’t do that. Because in the end, I would invest in real GDP in a way, because I want to get my long run liabilities matched. But equally, I’ve got to recognize that just being long of real assets has some dangerous potholes in the road. Last year was one of them. So using some capital, perhaps, more usefully, in something like trend following rather than levering long your fixed income portfolio, might be a better strategy.

[01:02:24] Alan

Yeah. It’s interesting, even from the perspective that we may be going into a different world, higher inflation. As you said, more macro opportunity could be more challenging for traditional assets. Obviously, as you say, you could have a more persistent positive correlation between bonds and equities. From that perspective, isn’t that an argument for an even higher allocation, would you say?

[01:02:50] David

Well, listen, we just want to be the best trend follower that we could be. The markets will unfold as they do.

[01:03:01] Alan

Niels?

[01:03:01] Niels

Yeah. Speaking on that and kind of dreaming a little bit that maybe in the next few months, people or investors will realize that maybe they should have a bigger allocation, perhaps not 30% or 40% but just have a bigger allocation because of course all the evidence tells them that they should. We’ve talked to some managers that track what they call trend crowding.

It brings me to just a topic before we wrap up in terms of maybe industry capacity. You can talk about your own capacity if you want, but also industry capacity. How do you see that? I’ve done this for more than 30 years and I still have no sense that there’s ever going to be too much money in trend following frankly, but anyways, are there certain things where you would say we could actually end up in a situation where there’s just not enough “capacity” where it would affect trend following as a strategy.

Because that’s another thing that people often throw at us is, oh, well, too many people are doing that, so clearly, you can’t keep delivering the returns you did back in the ’70s, in the 80s, in the 90s. Yet we do, of course.

[01:04:18] David

Well, conceptually, you’ve got to think that the more cash goes into any strategy, the returns from that strategy are likely to diminish within reason. You could argue they might go up in the very first instance because you’ve got better systems and you can invest more and so on. But once you get to a critical mass and you grow, you would think that any strategy, be it some kind of micro long-short in technology stocks, the more cash that goes into that or the more that goes into event driven (or anything else) returns, you’d expect to diminish. I think that’s a reasonable expectation.

Where we stand exactly in that cycle, right now in trend, is it’s likely that more assets I think will come in than will leave. I think those sorts of diminished returns are relatively minor given the big picture returns on offer. We’re not yet at some kind of tipping point where trend followers are just too big for the markets. We’re still at a good distance from that.

[01:05:27] Niels

Yeah, I agree with that. I just want to wrap up with two small simple questions. I say it’s simple, but I think they are.

[01:05:28] David

I hope so.

[01:05:29] Niels

The first one is one thing that we’ve asked everyone because we think it’s kind of funny to hear the different answers. Of course, some of the answers are the same, but anyways, let’s go for it. That is, what’s the one thing you hear about trend following that you disagree with the most?

[01:05:56] David

One thing that can rile me a little bit is when someone says are all trend followers of the same. Going back to your point about investing in the ETF, I think it’s true to say that unless you’re 70% or 80% correlated to the trend following index, you’re probably not a trend follower. So we’re going to be more correlated than perhaps some other… The dispersion is likely to be less wide in trend following than it is in many other strategies.

But that doesn’t mean that there isn’t dispersion and that there aren’t better managers and worse managers. I think all the efforts and things we’ve spoken about today, we haven’t just made a model 20 years ago, switched it on and left it. I think it’s important that not only are some changes made, but they’re transparently made for investors so they’re not surprised with what’s happening. They’re kind of brought along on the journey.

[01:06:53] Niels

Absolutely. Yeah. Final question for you, David, and that is, looking into 2023 and beyond maybe, given the fact of what we’ve gone through now with 2022 and so on and so forth, what are you most excited about? It could be anything. It could be from your business perspective, from a market opportunity perspective, but also, if there’s anything that you’re a little bit maybe concerned about.

[01:07:23] David

Listen, one can never rest on your laurels. It would be unlikely that we’re going to have a year as solid as last year as an industry. So I’m not overly concerned. I think from our business perspective, we’re in quite a nice place. Hopefully, we’ll raise some more assets this year. We’re always concerned that when markets lack direction and there isn’t a fundamental driver, that returns are going to be lackluster in those types of environments and that people can get bored of trend following and, to go back to your earlier point that, well, you had one a great year, a couple of good years here and there and then another banner year, maybe we’ll just wait two or three years, and then they forget.

So I think it’s important that there’s an allocation and a dialogue with investors now, that we don’t slip into too choppy markets for too long.

[01:08:19] Niels

Great place to end our conversation. David, this has been delightful. Thank you so much for being on the podcast for sharing your thoughts and insights with us. We hope we can do this again sometime in the future. To all of you listening today, I hope that you were able to take something from today’s conversation onto your own investment journey. If you did, please share these episodes with your friends and colleagues.

From Alan and me, thanks so much for listening. We look forward to being back with you on the next episode of Top Traders Unplugged as we continue our deep dive into the CTA industry. In the meantime, go check out the show notes for this episode and all the other resources you can find on our website. Not least, as usual, take care of yourself and take care of each other.

[01:09:00] Ending

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 on 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.