Why Trend Following Isn't Crash Insurance ft. Philippe Jordan

[00:00:01] Philippe

There’s a period not so long ago, call it 10 years ago, where people were running around saying, trend is dead. And we found out that it isn’t. And the reality of trend is that it is a 0.5 Sharpe strategy. It does play a role in investor portfolios, that is valuable. It does diversify but it’s not going to pay for your crash. That’s not what it does.

[00:00:32] Intro

Welcome to Top Traders Unplugged. In markets success doesn’t come from predicting what happens next, it comes from being prepared for what you can’t predict.

In each episode we go deep with some of the world’s most thoughtful minds in investing, economics, and beyond to understand how they think, how they prepare, and how they decide, and the experiences that shaped how they see the world. No noise, no short-cuts, just real conversations to help you think better and invest with confidence.

[00:01:05] Niels

Philippe, if money flowing into a strategy helps create the returns that attract the next wave of investors, how can we tell whether we found a lasting edge before those flows reverse?

[00:01:21] Philippe

Ooh, well, that depends what frequency you’re talking about. If you’re talking about something that’s operating at two to three or four days or a month in frequency, that would have a pretty high impact. If you’re talking about something that’s operating over four months to six months in frequency and does have turnover as opposed to none, I would think that doesn’t have much impact. If you have no turnover at all, I think that would have a lot of impact.

[00:01:54] Niels

Yeah, absolutely.

Well, today, Alan and I are joined by Philippe Jordan, president of CFM International. And for those who may be newer to the show, CFM is one of the leading quantitative investment firms in Europe and in the world, with about 36 billion in assets under management.

Philippe, before we get into how CFM invests, I’d like to understand how you came to see the market this way. You joined CFM in 2005. What brought you there? And which experiences, early in your career, had most shaped you to decide that CFM would be a good opportunity for you?

[00:02:33] Philippe

Good morning, Niels. Thanks for having me. Hi, Alan. I’m delighted to meet you and have this conversation.

Well, I have to dig back deep into my memory for this. Yeah, I joined in ‘05, but I started doing business with CFM in ‘91. So, I was one of CFM’s brokers and I became a friend, a fast friend of Jean-Pierre Aguilar, the founder, who was a terrific guy, who unfortunately left us in a tragic accident over a decade ago, now, doing something that he loved, which is flying airplanes that don’t have engines, otherwise known as gliders. He had a real quality to him in terms of charisma and entrepreneurship.

And his vision, which I came to share ultimately was, look, there has to be a better way to trade and invest in markets than reading newspapers and developing narratives. And that should consist in using data. And back in those days, data was a very simple form of data. It was price-based data and computers to analyze that data, extract the signal, and then executed faithfully because we didn’t have electronic systems. So, I execute all these signals manually in the markets on behalf of CFM. That approach had merit, but it also had a lot of drawbacks.

Trend following is an interesting strategy. It has a very robust Sharpe, but it doesn’t have a Sharpe of 1.0. It has a Sharpe, over the long term, of probably 0.5, which, over the long term, is very good. That’s actually better than the stock market over the long term (not the stock market in the last 10 years or 20, but the long, long term).

And then Jean-Pierre had this sort of increased vision which is, well, yeah, I need something else to sit beside this trend following strategy. And I need to work with scientists that are going to allow me to gain insight into things like statistical arbitrage, volatility, arbitrage, any number of other techniques that are essentially extracting data using a scientific framework and then executing it systematically. Systematically, still, in the ‘90s, means with hands, telephones, faxes, it’s still very manual.

And I followed this evolution, as a broker, for five or six years, and then as a fund of fund manager looking at CFM, and then at Credit Suisse I was doing spin out; spinning out people like Alan Howard and CQS. We were working on spinning out… not spinning out… but seeding a number of strategies at CFM with Credit Suisse Capital. That didn’t happen because Credit Suisse, being Credit Suisse, was unstable. And so that mandate evaporated, unfortunately.

And we had crossed what I call the threshold of velocity for that business because, prior to 2000, that business was executed by people like me, manually. Post 2000, this business went fully electronic. So, people like me were out of the picture. And all of a sudden, this business is being implemented by execution algos. And so, it becomes much more scalable. You’re able to trade at many different frequencies that were not accessible when people like me were executing these signals for people like Jean-Pierre.

And finally, Jean-Pierre ran two businesses. He had a hedge fund and he had a company that designed systems on behalf of banks, middle office systems, option systems. And there was some compatibility between the two businesses. There were some interesting synergies because there were some engineers over here that knew how to build systems and a hedge fund over here. But there was also a duality of purpose, and I didn’t really like that.

And he called me one day and he said, well, I’m selling the engineering business and I’m going to focus entirely on the hedge fund business. And that for me, in combination with all the other things, made CFM, in my view, a very compelling firm for the future. And I wanted to join that and help build it.

[00:07:34] Niels

Yeah, just out of curiosity, obviously, I mean, you and I, and Alan for that matter, we’re all there in the ‘80s and the ‘90s before things got systematic, electronic, and so on, and so forth. Was there something in particular, you remember, that you kind of had to unlearn when you started working with the scientists back then about how markets work or something that we had all come to believe, but maybe we realized maybe it’s not so.

[00:08:06] Philippe

I would say that there was a need for a point of convergence, and I’ll explain that. If you’re viewing the world through the prism of statistics, and Gaussian distributions ,and fat tails, etc., what happens, happens (to a certain extent), and you have to live with it. The path to deal with the negative outcomes is to further diversify your book as much as you can, which is a very rational approach and it makes sense. However, from time to time, history does actually happen, and the outcomes can be absolutely extreme, and you need to do things that are not statistically driven and are not systematic.

And I had to evolve my view from, ‘well, you need to develop an intuition about what your risk is, do something about it all the time’ to, ‘no, most of the time you don’t if you have a properly diversified book, in terms of strategies per correlations, tail structure, and something that’s fairly well ballaced.’

And then they had to come a bit to me, which is, well, sometimes really bad things do happen that have high impact that can kill you when you’re in this particular business, and that can be counterparty failure, that can be airplanes that crash into buildings in New York City and suspend trading for a week. There are things… In the case of airplanes crashing into buildings, you can do things post not pre. But in the case of counterparty management, there’s things that you can do pre that matter and will settle outcomes favorably.

In the case of crowding, also, the first case of crowding in our business, that I remember vividly, is the quant crunch of ’07, in August of ‘07. That was quite clear to me that it was occurring. Meaning if you go back to July of ‘07, you’re in the first phases of the big deleveraging of mortgage books by Wall Street banks. And you’re in the first phase… that’s ‘07 or ’08? Help me here.

[00:10:51] Alan

It would be ’07, yes.

[00:10:52] Philippe

It’s ‘07. Yeah, it is. ‘07. It’s so long ago.

So, in July, we’re going through this mortgage deleveraging of bank balance sheets and it gets acute. Right? And when that happens, people that run trading rooms tend to run around the trading room and want balance sheet back because at one stage they can no longer liquidate what they’re liquidating because everybody’s trying to liquidate, but they do need balance sheet. And if you run around the trading room, you go see the macro guy and you say, well, how much balance sheet do you have on? He might have half a billion dollars worth of directional risk with a lot of VaR, but not a lot of balance sheet.

And you keep running through your departments, you’re going to meet Mr. Fixed Income Relative Value, that has a lot. And you’re going to meet Mr. Statistical Arbitrage, that has a lot. And you’re going to ask them for balance sheet back, which will produce a negative effect on portfolios that are highly diversified but are long and short equities. And at first cut, people that run businesses statistically will look at any drop as an opportunity to add risk until their own risk systems saturate and you get a negative feedback loop that gets amplified with every day that goes through.

And we made the decision collectively, but I was certainly a big proponent of it, of getting out of the way. Of course, that’s only half the equation. You get out of the way, that’s great. But then do you get back in September fast enough to make up for how quickly positions are being rebuilt and P&L is being rebuilt, which we did, but it wasn’t perfect. But that’s to illustrate the fact that we now have a point of convergence where we, today, have a group of, I think it’s 11 people that meet every two weeks to discuss potential unknown unknowns, terrorists that can affect our business, our portfolio, our business, every part of our business; operationally, portfolio wise, regulatory wise, all sorts of dimensions away from simply portfolio risk.

[00:13:23] Niels

Now you mentioned that CFM began back in 1991. That means, first of all, congratulations on 35 years anniversary. That’s a big achievement.

Now, today you run much broader strategies compared to kind of the managed futures maybe space that you started in. Which would you say, which decisions most change the firm’s direction over that 35 year period?

[00:13:56] Philippe

I’ll take you back again to that statement I made about trend following. The firm started in ‘91, took off like a rocket in ‘91, ‘92 and ‘93 and then ’94, I don’t know if you remember this, but there was something called… was it Tesobonos? It was the Mexican fixed income crisis in ‘94.

[00:14:22] Niels

Yeah.

[00:14:22] Philippe

Which created a flight to quality into US treasuries and then a big cycle of down rates after a big cycle of up rates.

[00:14:31] Niels

Correct.

[00:14:31] Philippe

Well, that event was very painful for CFM and its trend business. And Jean-Pierre had the insight and the entrepreneurship to say, well, I need to find other strategies that share this scientific framework and approach to data, but I need other strategies that are fundamentally decorrelated from this one. And I need to start building bricks away from trend in order to have something that’s much more robust.

I definitely bought into that. He hired Jean-Philippe or got into a consultancy agreement with Jean-Philippe Bouchaud who brought that scientific framework, Marc Potters who did join CFM, and then Laurent Laloux and then, etc., etc.

[00:15:23] Niels

Alan, why don’t you take us into the topics where you want to go?

[00:15:29] Alan

Yeah, well, maybe, I mean, you’ve touched on a little bit about the kind of scientific approach which obviously is central to the CFM success and background. But maybe if you could just give us a sense on, practically, when a team of people who have backgrounds in physics and applied mathematics and things like that, when they get together and look at markets and think about markets, presumably it’s very different than how say economists or people who just started in the trading floor think about markets. So, I’d be curious to get a sense on your take on that and what’s different about that kind of hard sciences application to financial markets.

[00:16:14] Philippe

Yes, you’re right, Alan, but we have gotten to a point in the road where we actually hire people now that have domain expertise in economics and macroeconomic variables, etc.

[00:16:26] Alan

Eventually you needed some, yeah.

[00:16:28] Philippe

We do have people that have actual expertise and intuition that participate. But you’re right.

Well, the scientific framework, because I’m going to use the word, again, framework because this is not actual science. We have to make things clear. It’s the framework that’s being used. There are some elements that are very scientific when you’re looking at the microstructure of prices. But away from that, most things are within a framework.

Well, they think about robustness. They like thinking about everything in terms of, well, is it really robust? How can we test this assumption twelve different ways and reach the conclusion that it is actually robust? Then how should we look at a Sharpe? After a long process, it says it’s 1.5. Well, after doing this for many, many years, we know that even at the end of that process, that’s probably 0.7, not 1.5.

So, they look at the world with this jaundiced scientific approach that says prove it, prove it, prove it, prove it, and then prove it again. And when it’s proven, I’m going to divide the result by two and then let’s move on to something else that we think has correlation characteristics which are accretive to what we already have.

So, always approach the world with that framework. But at the same time, the world’s not static. One of the reasons this stuff works is not just the scientific framework, it’s the computing power. Right?

If I go back to the beginning, in order to execute stuff, Philippe Jordan had to actually use telephones and faxes back in the ‘90s. We’ve moved a long, long way away from that. Back in the ‘90s or even in the ‘00s, we were looking at price data and only price data. Well, we’ve moved a long, long way from that also; operating LLMs and all kinds of complex data-based systems.

So, it’s the scientific framework over increasing amounts of computing power per dollar spent. That’s the big leverage. Right?

So, every time we think we’ve reached a permanent plateau, in terms of computing power, well, we get surprised. This time it came from computer graphic cards for games, who would have thought? But that’s where it came from, initially, these new dimensions in computing power.

What’s the next one? I don’t know. But the trend for the last 35 years has steadily been up and has surprised everyone at every stage. There are many moments, in the last 30 years, where people thought that Moore’s Law would flatten out and it not only did it not flatten out, there are periods of acceleration. So, for businesses like us, it’s the scientific framework leveraged by increasing compute power per dollar spent.

[00:19:59] Alan

Interesting. I mean, obviously CFM, alongside a number of other well-known names, have been very successful in systematic and quant investing. Firms like Renaissance or in our space maybe other well-known names, or D. E. Shaw, or whatever, or QRT, all of these firms.

I mean when you think about CFM’s success and other firms like that, people always think in terms of the war for talent. And it’s about kind of finding people who are smarter that will unearth new alpha. How much of it is that or how much of it is, like as you mentioned, things like compute power, data infrastructure, obviously it’s all of that. But I mean can you put your finger on any one thing that you think that marks out the most successful systematic firms?

[00:20:51] Philippe

Well, as you said, it’s the whole package. But I think the most successful ones are the ones that adapt to the technological changes. The ones that know how to leverage new technology. And everybody leverages it slightly differently.

People who come to work at CFM are very self-selecting because we are, call it, one of the more extreme forms of collaboration, transparency, one P&L. On the other end of the spectrum (and this is not a bad thing, it’s just very different way of doing things) you have Millennium that’s operating pods and separate people that are operating within silos. The two different populations want two different things, and go to two different firms, and want different experiences.

People that come to CFM want to be successful. They’re ambitious, but they also want to collaborate. They want to be able to openly collaborate with colleagues, and share, and grow through iteration with other people.

So, I think what makes a firm successful is to have a strategy which is in line with its culture and to stay that way. And that doesn’t mean your culture stays static. It grows, but your strategy is in line with your culture.

I think most firms get into trouble when they start desiring what other firms are very good at and say, I should get some of that and I should get it fast. Well, you can’t because you haven’t spent 15 years hiring engineers that are thinking about Level 3 data in order books and trading in microseconds, and you don’t have that savoir-faire, and you don’t have that culture, That’s something you can only acquire very slowly and cautiously over time, if you’re going to make that change.

[00:23:38] Niels

Yeah, I mean, speaking of data and technology, before I get to my main question, I actually noticed something on your website that I had not come across before, but that’s probably my technical lack of knowledge. You mentioned you have 12 petabytes worth of data. I have no idea what a petabyte is. It sounds like a lot. How should we think about that amount?

[00:24:05] Philippe

I don’t even pretend to keep track with that anymore because it was already a lot at 3 and now it’s 12. I keep track of the costs of all that, not what it means.

[00:24:21] Niels

Fair enough. So, let me go in and kind of follow along what Alan was asking about. So, your Chief Product Officer, Laurent Laloux, mentioned that with the help of AI, he believed that maybe you could generate significantly more. I think an article mentioned 10 to 100 times more models. What has to change in your kind of standard of proof so that you don’t simply get better at finding patterns that won’t last when we start going down the rabbit hole of AI?

[00:25:05] Philippe

AI for us, at this stage, is a piece of technology that leverages a researcher’s time. So, we can now build agents that can troll the internet for papers and quickly run a few tests on these papers and tell us if we want to have a human being look at them or not. That kind of work. So, that was the first layer of work; leverages a basic researcher’s time.

There’s going to be a point where, I think, we’ll go to another level where we actually have agents, that we’re developing, that are starting to trade. We’re doing that to some extent and we’re going to do more of it. I don’t think that is peculiar to CFM. I think everybody is going to be doing that. I think everybody is going to be a manager of agents. So, you, Niels, you’re going to have agents that you’re going to develop and it’s becoming increasingly easy to do so. You can type what you want and Anthropic, or Google, or any number of these firms will then troll the internet on your behalf with that agent.

Now, you have to have intelligent ideas for those agents to be performing the tasks that are going to make you better at what you do. But I think this is going to be like using the internet. Right? And when people talk to me about AI, what do you do with AI, etc., AI, AI, AI, it’s like, well, do you use electricity? Right? Do you use the internet? This technology is going to become totally ubiquitous for everybody, not just CFM. It will have an immense impact on many aspects of our society. And I think managing money is only one of them.

Of course, we have a technological edifice over time that’s very important and we have something that is extraordinarily important when you have access to AI is 35 years of data that we’ve collected of our own trades. So, we can train systems over a data set that is completely curated, super clean and under our control, and is our IP. So, we can train these sets with a sandbox that protects the IP from the world at large. But that’s very valuable when you can deploy these kinds of agents.

[00:27:49] Niels

Yeah. Is there a different level of proof? Meaning what would a model have to demonstrate before it’s allowed to trade real money? And in a sense, the difference between models that we used to develop with humans and researchers, to these models where maybe we know a little bit less about what goes on inside. And also, I’m curious about whether there is something you could talk about as an example as to where AI has already improved results after trading cost, beyond just the fact that they can work more quickly in the research team.

[00:28:29] Philippe

Well, real world examples I won’t give you because they’re new and they’re special as far as we’re concerned. But there’s no doubt in my mind that there will be more. I don’t think there’s much doubt in mind of research overall at CFM on that. There’s a tradeoff between robustness and diversification.

Where I’m going with this is the more models you have that you faithfully believe have some degree of diversification, both in terms of the data that you extract and the intuition that you think this is exploiting, the less you can be or need to be demanding in terms of robustness. If you’re operating a trend system that’s looking at two or three frequencies over time, well, you want a lot of robustness statistically if you’re going to deploy that, and that’s going to be your core strategy. If you have hundreds, if not thousands of different alpha engines, over half a dozen or a dozen different strategies. if you have an AI model that doesn’t have any robustness because you’re not looking at 100 years worth of data, you might be looking at 2, but you have an intuition, and you’re getting good results, and you’re going to make a risk allocation to it that is within less than 1% of your overall risk, then you’re playing diversification. You’re playing the technology through diversification.

[00:30:15] Niels

Yeah. In the notes we shared with your team before our conversation a few weeks back, they mentioned this ML lab that you have. And I was curious what it’s trying to solve and what could change, maybe from this work, in terms of how you invest?

[00:30:43] Philippe

Well, in our history, we’ve always had these kind of efforts where, first of all, we publish. Right? So, we’ve always had these academic efforts because we believe that, first, they keep us crisp regarding new technologies, new ideas, and they attract talent to the firm. So, by doing this, we do a couple of things. We attract talent people to the firm that are familiar with a cutting-edge technology and can inspire us to at least have a framework that knows about this. Right?

And so, they run seminars internally that are very stimulating for our researchers. They’re not looking specifically to create models that are accretive. They’re looking to inspire our researchers to do so. Okay.

And then there’s a second dimension which is these are great recruiting tools because people that come to firms like CFM like to have an interesting, intellectually stimulating environment in which to be. And so, when you have these labs or these efforts to be in contact with academia and maintain links, they view that as a huge plus. So, that’s why we do this. We’ve done this right through the 35 years on many different subjects at different points in time.

[00:32:17] Alan

Taking a step back, I mean, you mentioned the long run returns and mentioned the S&P over the last decade, two decades, it’s been phenomenal. And I’m just curious, the average investor might say, I’m in equities, maybe I have some other assets, I have a long beta portfolio. Sure, you’ve got all of these bytes of data, but at the end of the day, you’re generating similar returns to maybe not even beating the S&P. Obviously, it’s done in a different way.

But for investors who haven’t embraced hedge fund investing or who haven’t embraced systematic investing, I’m sure you speak to a lot of more sophisticated investors who already have. But for those who haven’t, what is the argument, do you think? Obviously, it’s about diversification, but what else would you say?

[00:33:13] Philippe

I’d say it three times diversification, diversification, diversification.

[00:33:17] Alan

Okay.

[00:33:19] Philippe

So, that’s what I’d say. And then, finally, I think we’ve crossed the point with investors, broadly speaking, and technological approaches to invest in it. AI has become so ubiquitous in people’s lives and people’s imaginations that what we do is no longer exotic, in general. I think that within probably seven years, it’s the contrary that’s going to be thought of as being exotic, people that don’t use extensive technology to manage portfolios. I think the world’s moving very fast in that direction.

I mean, for years, and years, and years, you’re right, Alan, it was a very difficult conversation, specifically with a class of people that I always thought would naturally be inclined to invest in systematic investing, which are Menlo Park based venture capitalists. So, these are people that invest in technology. Right? But their view of investing in technology is a bottoms up approach where you know the company, and you know the founders, and you’ve been through many rounds, and it’s essentially a very intuitive, human-driven process.

And when you talk to those people about systematic investing, which is a huge diversifier to what they do, they are still very reticent, and they view it as somewhat suspect. But away from Menlo Park, the rest of the world has moved on, I think, and within the world of institutions, and large family offices, etc., they’ve embrace systematic investing. And I think AI is going to do the rest.

[00:35:32] Niels

Can I jump in here, Alan, just because I think it’s an interesting point? I certainly remember, also back in the early ‘90s, that when we came out talking about CTAs and trend following, people thought of it as a black box, and it was very scary, and all of those things. And I completely agree with you, Philippe, that I think people have started to realize and kind of embrace the benefits that using rules, using systems bring to an investment process.

However, your colleague Jean-Philippe Bouchaud, who has also been a guest on the show, takes it to, I think, the next level because now (and this is not new, but some people will not necessarily have heard the conversation), he’s arguing, of course, that it’s the buying and the selling that happens that has a lasting effect on the markets. It’s not necessarily what we used to think was driving market action. So, if an investment committee, today, takes this view seriously, what should they, in your opinion, do differently?

[00:36:51] Philippe

Well, you should diversify away from trends. Trend is one strategy. Because what you’re describing… I mean, why do trends function? Right? Trends function because people follow price points, prices in markets every day. Right? So, momentum. It is one of our strategies, but it is not our only strategy.

We have other strategies that are not trend-based. The S&P’s magnificent success, or the US equity market’s magnificent success is, in large measure, due to, yes, flows. It’s not new. So, flows do matter, reversal flows do occur and flows do have impact. I mean, to me it’s not really a revolutionary concept to think that flows actually impact prices.

[00:37:48] Niels

I agree. Although I think the world has been pretty slow at adopting some of the views from people like Mike Green about the impact of passive investing, for example, which probably poses some risk down the road if those flows start to reverse at some point.

[00:38:08] Philippe

Absolutely. Especially they’re in forex ETFs on top of it.

[00:38:12] Niels

Right? Yeah, exactly. I mean, the other thing I’m curious about, when you go out talking to people, and it’s all about kind of placing something for them to think about. I mean, to a large extent, people always worry about things that can go wrong, so, typically risk that would get their attention. I’m curious, from your experience, and we know a lot of people, as Alan indicated, a lot of people just look at the equity market, said, yeah, it’s done really well, so I’m just going to stick with that or something else, it doesn’t really matter. I mean people said the same about bonds only a few years ago.

So, looking at today’s market environment, what do you think people, whether they’re institutions, whether they’re individuals, what do you think are some of the risks they are likely to underestimate in today’s market environment by predominantly just looking at historical data?

[00:39:16] Philippe

Everything that can go wrong will go wrong. It’s a good rule to start with. What can go wrong today that is specific to today’s circumstances? Well, you pick your plat du jour, right, between geopolitics in Russia and the Ukraine, or the Middle East, or the election process in various countries.

But I think there’s one risk or two which have a long-term quality to them and the risk is not showing up in the short-term or is not priced properly: demographics in the West. So, it’s going to be hard to change this trend of people having less, and less, and less children. We live in societies that have been built for and on behalf of a middle class that has children. We no longer do that. So, we’re going to have a different social bargain over time. And we’re rapidly running into the retirement age of folks like me and maybe a little older than me. I’m the last of the baby boomers in ‘64. But we’re hitting that wave.

[00:40:44] Niels

We’re all getting there.

[00:40:45] Philippe

And that is going to, I think, change the social contract and all these convulsions that we’re seeing between generations, and convulsions regarding what people perceive as being inequality. And when I say perceive, I mean you take a country like France right now that’s going through one of these social convulsions that the French are so good at. One of the great recriminations is inequality. While the Gini factor in France peaked in 1970 and has steadily been going down since and has plateaued. And France is one of the countries with the lowest levels of inequality of the G7. It is very good at redistributing wealth. Yet we have these anxieties, and what are they due to, and how will we solve for that?

I think these are long-term risks. But long-term risks have a way of showing up somewhere short-term and being quite consequential. So, that’s something I think is severely underpriced and will shape markets in a structural fashion for the next 10, 15 years.

[00:42:08] Niels

Alan, I would like, and maybe you could follow up after this little intro. I would like to maybe talk a little bit about trend or something that’s trend related because what you also do well at CFM is you publish a lot of papers. And another previous guest of ours, your colleague, Philip Seager, recently published a relatively short note about alternative markets. And this is a topic that we’ve been discussing many, many times on the podcast.

I remember the narrative back in the, say, mid-2010s when, perhaps, traditional trend followers were going through a quiet period. Along comes these alternative market focused trend following strategies delivering very impressive returns. And the narrative was, back then, well, alternative markets, they trend better, so, that’s kind of the future. And a lot of money was heading in that direction. Today, maybe we look back at the last five years and we say, well, maybe that didn’t really hold up in terms of them trending better.

But I’d love to hear your thoughts about alternative markets. Also, to you, when does it become too alternative (if you know what I mean) for you, and just kind of hear your thoughts on this debate.

[00:43:37] Philippe

Well, we’re all for finding markets that offer low correlation to the existing mix of things that we trade. If we think that correlation is persistent and true, that’s one. And then two, that they’re operationally sound. Even if that correlation were true, if we have to take risks in markets that are operationally complicated, illiquid, and difficult to process, that’s a very large ask.

So, we went through the exercise of just running through what people were calling alternative markets and we found that viewing them through the prism I just described, the only true alternative markets were the Chinese markets. The Chinese contracts were truly

diversifying and had correlation/decorrelation characteristics to our existing mix. That they were deeply liquid and were complicated, because nothing’s easy in China, but they were operationally sound to access. The other ones, things like electricity, etc., did not meet the thresholds that merit investment.

[00:45:04] Niels

No, absolutely.

Alan, let me head back to you.

[00:45:08] Alan

Yeah, well, maybe related to that topic, but more generally as well, obviously, you talked about kind of changes in the markets, execution, etc. Everything has gone electronic. Also though, we’ve had changes in the market microstructure; new players, pod shops, new products, more retail participation, zero-DTE options. All of this which has shaped the market somewhat, and obviously, I imagine, creates new opportunities and maybe changes the return profile of some strategies.

How do you think about that? Or is that a specific thing you look at and try and think about or is that just kind of reflected in the price action which you primarily analyze?

[00:46:01] Philippe

Well, we tend to, I mean, historically, the microstructure of markets we look at through the prism of execution and costs and friction matters. And the faster you trade, the more that matters. We have never been in the super high frequency business, so we’ve operated our businesses as fast as (for alpha purposes) hours, days, weeks. And most of our businesses are really in much longer frequencies.

So, we look at the microstructure to understand the cost of executing our business at large. And we think of ourselves as being able to execute efficiently and competitively, but we’re not trying to beat the XTXs of the world, and the Citadel Securities of the world. So, we’re on the other side of that line.

Are we taking a hard look at how order books function? Yes, because that definitely has a big impact on prices that you’re going to pay to execute your book. And friction is everything in our business. And there are two ways of dealing with it, trade longer frequencies or understand the microstructure better if you know you’re executing very fast for alpha purposes.

[00:47:34] Alan

Interesting. And I mean, you say that in relation to not being very fast. And obviously Niels is talking about some of the papers that you guys have written and we talked about on the podcast, the paper that Jean-Philippe wrote with colleagues around fast trend following and how the vol adjusted tick size, seemed to be an explanatory factor in, I suppose, the efficacy of fast trend following. Which, again, is another, I suppose, market microstructure change. But I mean, is that your decision not to get involved at the very fast end of markets? Is that just reflecting it requires more cost, more technology, more access, etc., or is it more of a philosophical thing?

[00:48:24] Philippe

I’ll go back to something I said. One of the keys to our success, or other of my peers’ success, is to have a strategy that’s aligned with your culture. And our culture is not an engineering culture to make money. We have engineering solutions to deploy models that have been analytically built to make money in markets, which is a very different mindset. Right?

If you’re extracting alpha using engineered solutions, that’s a very different thing than extracting alpha with an analytical algo that was built from an analytical insight as to how prices evolve in a particular market. An engineered solution is about proximity, it’s about speed, it’s about levels of compute that you can deploy. Those are engineers that are thinking about how to get the best out of the market structure itself.

We think about the market structure, but we think about it in terms of what does it cost us to trade, to clear the market structure? And we want to be efficient at it. We don’t want to extract alpha in a process. We’d like to, but it’s not who we are today. Right?

[00:50:01] Alan

Yeah. Okay. And on the risk management side, I mean, you’ve touched on how you have this team that meets every couple of weeks. So, it sounds like maybe you’ve learned from past crises. I imagine a firm that’s been in existence 35 years has that history. And obviously, you do all of the data analysis, look at the past.

But as you’ve said yourself, in markets, literally anything can happen. And whatever. You said, whatever you don’t want to happen will probably happen. You can diversify, but do you think you need to have that capacity to override systems to make discretionary interjections? Do you think it’s all about taking slightly less risk than the model suggests you should? Or what is the solution to managing those major tail events?

[00:50:48] Philippe

So, there are many levels to that. There’s the major, major, and we talked about one of them which was the quant crisis. But then you have idiosyncratic events. How much risk do you want to have in exposure to the Brazilian election as an idiosyncratic risk? Do you want full risk or do you want to have half risk or quarter risk? The same could be true for a Japanese election or US election, anything that could polarize the markets in a very digitalized fashion.

So, we make decisions that range from the mundane like this, to, ah, do we need to reduce risk across our entire statistical arbitrage book in the case of the quant crunch? And it runs the gamut, but the idea is to try and do this as infrequently as possible. We are never thinking in terms of, oh, well, maybe we should have more risk to this, because right now the circumstances seem favorable to what we’re doing. So, we want to run at as fixed a risk regime as we can. And we will, at times, purchase insurance, whether it’s idiosyncratically for a given circumstance in a given market or overall in the portfolio if we think something large enough is perhaps occurring that will have consequential effects on the entire portfolio.

[00:52:21] Alan

So, you could actually go out and buy options or some explicit insurance?

[00:52:29] Philippe

No, when I say we’re buying insurance if we reduce risk…

[00:52:33] Alan

Okay, so, by risk reduction.

[00:52:35] Philippe

By risk reduction, that’s how we’re buying insurance.

[00:52:38] Niels

This may sound like a naive question here, Philippe, but is this something that you can somehow quantify your concerns for certain events and thereby take it into account in the way you test things historically? Because otherwise it may be difficult to rely on a backtest if it doesn’t have any insight to all the manual things that happen along the way, if you know what I mean.

[00:53:07] Philippe

I’m not sure I know what you mean but, in any case, we do have a vast database of all the things that have happened against them. And we test against five sigma events across the portfolio all the time. I’m not sure that was your actual…

[00:53:23] Niels

No, I mean, I think that’s fine. I mean that’s fine. It’s just that you often hear people say that well, if you start making discretionary decisions in your trend following strategy, you can’t use your backtest for anything because it doesn’t know about all your decisions that weren’t part of the rules. That’s what I was thinking of.

Maybe it’s too limited due to kind of a trend following system. But I get what you mean in terms of stress testing, and so on, and so forth. That is of course, something we…

[00:53:51] Philippe

But let me push it into another dimension. I mean, whether it’s relevant statistically or not that your counterparty is going to fail on you, it doesn’t matter. You want to get that risk down if you want to move it somewhere else. So, in some circumstances it’s irrelevant whether you’ve…

[00:54:16] Niels

Whether you can test for it or not, in that sense. Yeah. that’s true.

Well, one thing, just staying on trend for a little while longer, something I picked up on when you and I talked briefly last week, which I thought was very interesting. I mean, your trend program along with all your other programs have done really well. So, congratulations on that. When you kind of briefly explained it to me, I was kind of thinking, it sounds like you’ve evolved trend following quite a bit actually. And maybe even too, from my recollection with our conversation with Philip Seager a few years ago, would you classify you doing trend differently maybe to the traditional trend follower?

[00:55:03] Philippe

Well, we’re not doing 1991 trend, that’s for sure.

[00:55:06] Niels

Correct.

[00:55:07] Philippe

We’re doing the advanced version 35 years later that’s informed by the fact that we also run very large statistical arbitrage books that have very sophisticated PCA techniques that are deployed, and other strategies. So, it’s trend informed by other strategies. But it’s definitely trend. But it’s kind of neat. It’s trend on trend itself and then all the derivatives thereof. Right? It’s all of the principal components that are not fully explained by the first component.

[00:55:54] Niels

Yeah, one thing, with the success you’ve had comes new clients, comes new assets. And we have obviously seen, I know most of your assets are in strategies defined as multi-strategy. And of course, in the last few years, multi-strategy has really been the buzzword, and a lot of money has been raised. I noticed recently in the last week or so that, was it Millennium that had like a US$30 billion subscription demand, but they could only offer US$20 billion, etc., etc. These are huge amounts.

How do you think about capacity in all of this? And would you even consider, I know you mentioned earlier that you have a collaborative approach and it’s not a trading people or teams or whatever. But would you even, at some point, consider saying, well, maybe we need to find people outside CFM to take on some of the responsibility of trading all the assets that clearly is coming our way or your way, I should say, how do you think about all these things? It’s a positive problem to have, of course.

[00:57:06] Philippe

Yeah, it’s a caviar problem, but nonetheless. And we’re going to go back to a statement I’ve made many times because I believe in it, which is your strategy has to be in line with your culture. At this stage this is not in line with our culture and we also have plenty to do with the capacity that we have in trend, the capacity that we have in Cumulus (which is the second derivative of Stratus, the excess capacity of Stratus). We have plenty of capacity in our systematic global macro business. So, we’re going to focus on executing that capacity. And perhaps once this is executed, we can ask ourselves questions about what’s next, and within the realm of our culture or within 5 degrees of it that we can develop. But we’re not at that inflection point.

[00:58:05] Niels

The other thing I was thinking of is you go out and you talk to some of the largest, most sophisticated investors in the world and I was curious about what you kind of need to understand about their portfolio, about what they do in order for you to figure out how you can best help them. What are the things that you look for in the conversations that you have with these types of investors?

[00:58:36] Philippe

Diversification and alignment of purpose. So, diversification is always first and foremost. I mean do we bring true diversification to their existing portfolio or not? In most cases, yes, because there’s not many things that multi-strat businesses are correlated to. And then there’s alignment. Alignment means, is there an aligned understanding of what we do and what we can provide and what we can’t do?

And what I mean by that, there’s a specific tranche of time, post what people call the QFC in Australia, which is the 2008 crash, and call it 2015, where people have been marketed CTAs as crash insurance and they came to believe that trend following could perform this sort of crash hedge. And it can, but it’s frequency dependent, meaning the frequency has to match the frequency of the actual crash for that to happen.

And it’s unlikely, over time, that you’re going to be able to do that at a cost-efficient price point. Because if you’re running strategies at that frequency, usually you’re not making any money unless one of these tail events happens. So, that’s what I mean by alignment, that there’s a good understanding of what can realistically be expected of what we do, and is that valuable to the investor or not?

[01:00:20] Niels

Yeah. Alan, I’ll let you pose another question. If you want. And then I’ll let actually Philippe bring up anything that we haven’t touched on that we should have touched on. But Alan, is there anything towards the end that you would love to ask Philippe?

[01:00:37] Alan

I mean, just in terms of the direction of CFM from here, obviously what you see with a lot of other firms is it becoming more of a solutions provider, a platform? Is that the direction for you as well, in terms of allowing clients to pick strategies to build custom portfolios, or not, or what’s the next evolution for your firm?

[01:01:01] Philippe

Well, I think there’s a measure of that. We now have Stratus, the flagship, Cumulus, which is the excess capacity of Stratus. We have systematic global macro, Discus, which is multi-strat futures, trend following, and trend following beta capped. And if you want to pick between these building blocks, perfectly happy to do that. So, we do have bricks and people can put those bricks together in various ways. But I wouldn’t call us a solutions provider in the sense that we’re going to tailor that to the millimeter and carve out models for people, but within the constraints of those bricks, yes, we will do things that people find useful.

[01:01:54] Niels

Yeah, I said before, Philippe, that there are quite a few things we could have brought up today. We’ve kind of tried to stay true to some of the topics we know from your team that they like, but also the things we found interesting. But I’m going to give you the last word anyways to bring up anything that you think is important that we didn’t touch on today. So, anything you want to highlight as we close?

[01:02:23] Philippe

Well, I mean, since trend was so prominent in this conversation and it’s prominent in your business, there’s a period not so long ago, I’ll call it 10 years ago, where people were running around saying trend is dead. And we found out that it isn’t. And the reality of trend is that it is a 0.5 Sharpe strategy.

And 0.5 Sharpe over a distribution of decades and centuries is going to have decades that look like 0.1 Sharpe or 0.2 Sharpe. That’s absolutely possible. And the flip side of that, it will have decades where the Sharpe is 1.2 and people will think that trend is a 1.2 Sharpe strategy in those decades. It is not a 1.2 strategy, it’s a 0.4, 0.5 strategy. But it does play a role in investor portfolios that is valuable, it does diversify, but it’s not going to pay for your crash. That’s not what it does.

[01:03:35] Niels

Yeah, I appreciate that. And of course, you and CFM authored a paper, I think it was two centuries. There’s been a few different number of centuries papers out there. I think yours was the two-century trend following evidence, which was another great paper. So, I appreciate you reminding people that when it’s the bleakest time for trend, it’s not the end of a trend, but when it’s the best time for trend, it’s probably not going to last forever either. So, always a good reminder.

Philippe, thank you so much for joining Alan and me today and for being so open about what you think about the topics that we wanted to discuss. This has truly been an enlightening conversation. I’m sure our audience has learned a lot from you.

And to everyone listening, thank you so much for spending time with us today. We very much look forward to being back with more conversations with some of the greatest minds and smartest people in finance. And until next time, as usual, take care of yourself and take care of each other.

[01:04:36] 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 your favorite podcast platform and follow the show so that you’ll be sure to get all the new episodes as they’re released. We have some amazing guests lined up for you and to ensure our show continues to grow, please leave us an honest rating and review.

It only takes a minute and it’s the best way to show us you love the podcast. We’ll see you next time on Top Traders Unplugged.

This podcast expresses the views of its hosts and the guests appearing on the podcast as of the date of its recording, and such views are subject to change without notice. Top Traders Unplugged do not have any duty or obligation to update the information contained herein.

Furthermore, Top Traders Unplugged make no representation to its accuracy and it shall not be assumed that past investment performance is an indication of future results. Moreover, wherever there is a potential for profit, there is also the possibility of loss.

This content is made available for educational purposes only and should not be used for any other purpose.

The information contained in this podcast does not constitute and should not be construed as investment advice or an offer to sell or a solicitation to buy any securities or related financial instruments in any jurisdiction.

Certain information contained herein concerning economic trends and performance is based on or derived from information provided by independent third-party sources. Top Traders Unplugged may believe that the sources from which such information are obtained are reliable.

However, Top Traders Unplugged cannot guarantee the accuracy of such information and has not independently verified the accuracy or completeness of such information or the assumptions on which such information is based. This podcast, including the information contained herein, may not be reproduced, copied, republished or posted in whole or in part in any form, without the prior written consent of Top Traders Unplugged.

FURTHER DISCLAIMER
ANY DESCRIPTION OR INFORMATION INVOLVING MODES, INVESTMENT PROCESSES OR ALLOCATIONS IS PROVIDED FOR ILLUSTRATIVE PURPOSES ONLY AND DOES NOT CONSTITUTE INVESTMENT ADVICE NOR AN OFFER OR SOLLICITATION TO SUBSCRIBE FOR ANY SECURITY OR INTEREST. ANY STATEMENTS REGARDING CORRELATIONS OR MODES OR OTHER SIMILAR BEHAVIORS CONSTITUTE ONLY SUBJECTIVE VIEWS, ARE BASED UPON REASONABLE EXPECTATIONS OR BELIEFS, AND SHOULD NOT BE RELIED ON. ALL STATEMENTS HEREIN ARE SUBJECT TO CHANGE DUE TO A VARIETY OF FACTORS INCLUDING FLUCTUATING MARKET CONDITIONS AND INVOLVE INHERENT RISKS AND UNCERTAINTIES BOTH GENERIC AND SPECIFIC, MANY OF WHICH CANNOT BE PREDICTED OR QUANTIFIED AND ARE BEYOND CFM’S CONTROL. FUTURE EVIDENCE AND ACTUAL RESULTS OR PERFORMANCE COULD DIFFER MATERIALLY FROM THE INFORMATION SET FORTH IN, CONTEMPLATED BY OR UNDERLYING THE STATEMENTS HEREIN. CFM ACCEPTS NO LIABILITY FOR ANY INACCURATE, INCOMPLETE OR OMITTED INFORMATION OF ANY KIND OR ANY LOSSES CAUSED BY USING THIS INFORMATION. CFM DOES NOT GIVE ANY REPRESENTATION OR WARRANTY AS TO THE RELIABILITY OR ACCURACY OF THE INFORMATION CONTAINED HEREIN.